article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
uncovering the formation mechanisms of the most massive galaxies has a long history and has largely driven the field of galaxy formation and evolution .
traditional analyses of stellar populations in nearby ellipticals reveal that the bulk of the stars in the most massive galaxies formed early , within a few gyr of the big bang ( e.g. , trager et al .
2000 ) . observing distant galaxies
is complementary to nearby galaxy age - dating , as it allows us to directly observe galaxies while they are forming .
early detections of high redshift galaxies revealed that uv bright star forming galaxies at @xmath11 are dominated by relatively low stellar mass systems ( @xmath12 ) ( papovich et al . 2005 ) .
as the most massive galaxies today are @xmath13 , this implies that high@xmath14 galaxies grow by a factor of 5 - 10 through some process , such as merging ( e.g. , conselice et al .
2003 ; conselice 2006 ) .
it seems likely that at least some progenitors of today s massive galaxies are not bright in the uv , and a population of near infrared ( nir ) selected distant red galaxies ( drgs ) was described by saracco et al .
( 2001 ) and franx et al .
( 2003 ) , who suggested that some of these systems are the progenitors of the most massive present day galaxies .
these drgs are selected using a single color cut , @xmath15 , and thus require near infrared ( nir ) imaging to detect . while it has been proposed that drgs are massive @xmath16 galaxies , the bright end of this population
has yet to be studied , and thus we do not yet know the general characteristics of red galaxies selected in the near infrared .
we investigate this problem by examining the redshifts , masses , morphologies and spectra of drgs in the extended groth strip ( egs ) .
we study 1010 bright drgs with @xmath17 within the egs , and two other fields that are part of our nir survey with the palomar telescope .
we examine the structures and morphologies of drgs that are coincident between the acs imaging in the egs and our nir survey .
we find that a high fraction of drgs at @xmath17 are at @xmath18 , with the majority between @xmath19 , including 95 ( 10% ) with spectroscopic redshifts .
we present an initial study of these nir selected galaxies at @xmath20 and conclude that they are small , low mass , agn / star formation dominated galaxies .
furthermore , we find a diversity in the morphological properties of drgs at all redshifts , suggesting that there is a wide diversity of properties for galaxies selected with the @xmath15 color cut , perhaps even larger than that seen for traditional eros ( moustakas et al .
2004 ) . in this paper
we assume the following cosmology : h@xmath21 = 70 km s@xmath22 mpc@xmath22 , @xmath23 = 0.7 , and @xmath24 = 0.3 , and use vega magnitude units throughout .
the data used in this paper originate from multiple data sets partially associated with the all - wavelength extended groth strip international survey ( aegis ) ( davis et al .
the main source of data is from the palomar observatory wide - field infrared survey ( powir ) ( bundy et al .
2005a , b ; conselice et al . in prep ) , while other sources include deep2 spectroscopy ( davis et al . 2003 ) , one orbit hubble space telescope ( hst ) imaging using the advanced camera for surveys ( acs ) of the egs ( lotz et al .
2006 ) , and an x - ray survey of the region ( nandra et al .
2006 in prep ) .
the powir data were obtained from september 2002 until october 2005 at the palomar observatory 5 meter using the wide - field infrared camera ( wirc ) . for the j and k data we cover a total of @xmath25 deg@xmath1 to 5-@xmath26 point source sensitivities of @xmath27 and @xmath28 . within the egs , where our highest quality data exist , we image 0.2 deg@xmath1 to a 5 sigma depth of @xmath29 and @xmath30 for point sources . to construct a sample of drgs which is nearly complete in the k - band , we limit our study of drgs to those at k@xmath31 .
we select drgs through the use of a @xmath15 cut . to supplement the existing 95 deep2 spectroscopic redshifts , we calculate photometric redshifts derived from brijk photometry using the annz method for @xmath9 systems , and hyper - z for those at @xmath32 .
the photometric redshift accuracy at @xmath9 for all galaxies is quite good , with @xmath33/@xmath34 = 0.05 , based on comparisons to @xmath35 spectroscopic redshifts .
the subset of 95 drgs with @xmath9 spectroscopy have a slightly softer accuracy of @xmath33/@xmath34 = 0.07 , with very few catastrophic mismatches .
we test our @xmath32 photometric redshifts by comparing to 44 galaxies in the steidel et al .
( 2004 ) bm / bx @xmath36 population , where we find a lower accuracy of @xmath33/@xmath34 = 0.22 .
we use the spectral energy distributions ( seds ) and spectroscopic / photometric redshifts to calculate stellar masses utilizing the methods outlined in bundy et al .
( 2005a , b ) .
our star formation histories are parameterized by a single exponentially declining star formation event .
we vary the metallicity , dust extinction and age of these various models to calculate the distribution of stellar masses using a chabrier imf ( see bundy et al . 2005a , b for a detailed discussion ) .
we find a total of 1010 drgs , brighter than @xmath37 , over our 0.7 deg@xmath1 ( 2520 arcmin@xmath1 ) palomar survey area , including two fields in addition to the egs . in comparison to previous work ,
van dokkum et al . ( 2006 ) studied @xmath38 drgs over 400 arcmin@xmath1 , fires 14 drgs over 100 arcmin@xmath1 , and papovich et al .
( 2006 ) 153 drgs over 130 arcmin@xmath1 , all with similar drg number densities at @xmath17 as found in our study . out of our 1010 nir selected drgs , we determine that 55 are stars based on their optical colors and unresolved structures .
the drg redshift distribution is shown in figure 1 .
a total of 95 of these drgs have secure spectroscopic redshifts taken with the deimos spectrograph as part of the deep2 survey ( davis et al .
this gives us a lower limit of @xmath39% on @xmath9 galaxies in pure photometrically selected drg samples to @xmath37 .
using these redshifts with our calibrated photometric redshifts at @xmath9 , we find that @xmath40% of @xmath17 drgs are at @xmath9 , but most of the drg population ( 64% ) is between @xmath19 .
objects at @xmath16 represent @xmath41% of the @xmath17 drg population , while @xmath42 of the population is found between @xmath43 .
this implies that bright drgs have a broad redshift distribution , which is also found in other recent studies ( reddy et al .
2005 ; papovich et al . 2006 ; grazian et al .
2006 ) .
we plot the stellar mass distribution for our sample in figure 1b . from this
we conclude that @xmath15 selected galaxies do not uniquely sample high - mass galaxies , with a low - redshift ` contamination ' rate approximately similar to what franx et al .
( 2003 ) estimate based on smail et al .
drgs however do select high redshift systems , generally at @xmath44 , and have a redshift distribution broader than the ero population ( moustakas et al .
a @xmath15 selection at @xmath45 measures approximate @xmath46 colors similar to dusty or old galaxies at those redshifts , based on 6 gyr single burst stellar population models . in figure 2
we plot the @xmath47 vs. @xmath48 diagram for all drgs at @xmath49 that have stellar masses @xmath50 .
the most massive galaxies with masses @xmath50 are shown as red boxes for @xmath51 systems , and blue circles for those at @xmath16 .
the green crosses represent galaxies with m@xmath52 at @xmath16 .
the average @xmath47 color for all galaxies at @xmath16 with m@xmath52 is @xmath53 , while the most massive systems with m@xmath52 have a mean color of @xmath54 .
although the average massive galaxy has a similar color as the drg limit , we find that a significant number ( @xmath55% ) of all massive galaxies at @xmath56 will be missed by the standard drg selection , in agreement with van dokkum et al .
these results are robust when using brighter magnitude cuts where our photometry is more accurate , and at redder color limits , up to @xmath57 .
these trends also remain after we select only galaxies that are redder than 1 @xmath26 of the drg limit , and after testing our selection through a bootstrap resampling .
that is , the diversity in redshifts / masses for bright drgs is not produced by photometric errors bringing up galaxies into the drg cut .
we find however that at @xmath16 the drg selection limit to @xmath58 does locate very massive galaxies , with an average stellar mass of @xmath59m@xmath60@xmath50 = @xmath61 .
the average drg at all redshifts has a stellar mass of m@xmath62 , with the spectroscopic sample at @xmath9 having an average stellar mass of m@xmath63 . for the remainder of this paper we focus on the drgs found within the egs region , where we have ancillory hst and x - ray data . due to the acs overlap with our nir imaging
, we can determine the morphological properties of 107 drgs at @xmath17 up to @xmath64 .
we utilize both apparent eye - ball morphologies and quantitative measurements of size and structure on the f814w acs imaging in the egs to determine what type of galaxies drgs are .
we classify by eye our galaxies into the types ( with the number of objects in each class ) : ellipticals / compact ( 61 ) , peculiars ( 31 ) , disks ( 0 ) , edge - on disks ( 7 ) , and those too faint to classify ( 8) following the method described in conselice et al .
the morphologies for the galaxies show a diversity of types , as can be seen in figure 3 , similar to the situation for lower redshift eros ( moustakas et al .
the most common type are ellipticals / compact making up 57% of the total drg sample , with most of these ( 80% ) compact .
about 7% of the sample are composed of edge - on disks , which are likely red in @xmath47 due to dust redenning .
peculiars account for the remainder ( 29% ) of the drg sample . at @xmath16
there is roughly a similar mixture of ellilptical and peculiars .
we have morphologies for 14 systems with spectroscopic redshifts , the majority of which are compact ( 4 ) .
we find very little evolution in the morphological distribution with redshift , with a slight increase in the relative number of peculiar galaxies at higher redshift .
we also utilized the revised cas system ( conselice 2003 ; conselice et al .
2004 ) to quantify the structures of the galaxies .
the cas ( concentration , asymmetry , clumpiness ) parameters are a non - parameteric method for measuring the structures of galaxies as resolved on ccd images ( conselice 2003 ) . as expected from the eye - ball classifications , we find a diversity of cas values . for the systems with spectroscopic redshifts at @xmath9
we find average values , @xmath65 , @xmath66 , @xmath67 , which is typical for nearby normal galaxies , and spirals with exponential profiles ( conselice 2003 ) .
we find nearly the same values for galaxies with @xmath9 photometric redshifts .
for the higher redshift sample at @xmath49 we find that the morphologically classified ellipticals / compacts are more concentrated with @xmath68 , with many systems at @xmath69 , similar in morphology to nearby massive ellipticals . furthermore , as we are examining @xmath32 systems in the uv , we are likely underestimating the rest - frame optical light concentration ( taylor et al .
2006 , submitted ) .
this is an indication , along with their high masses , that these @xmath49 systems are likely elliptical progenitors .
there are two caveats to using the f814w band acs imaging on these galaxies .
the first is that there are redshift effects which will change the measured parameters ( conselice et al .
2000a , b ; conselice 2003 ) .
systems at @xmath70 are also viewed in the rest - frame ultraviolet , complicating comparisons to nearby galaxies viewed in the optical .
there is evidence , however , that distant galaxies dominated by uv bright star formation look similar in the rest - frame optical and uv ( windhorst et al .
2002 ; papovich et al . 2005 ; conselice et al .
we have 95 confirmed high quality spectroscopic redshifts for drgs at @xmath9 .
about 7% of these systems are traditional eros with @xmath71 , while @xmath72 33% have red optical / nir colors , with @xmath73 .
the average magnitude of our spectroscopically confirmed lower redshift drgs at @xmath9 is @xmath59m@xmath74 , with an average color of @xmath75 , and these systems span the color - magnitude relation , suggesting a heterogeneous origin .
our lower redshift drgs are a mangitude bluer in observed @xmath10 color compared to the @xmath49 drgs .
they are also roughly a magnitude brighter than the deep2 spectroscopic limit , with an average @xmath59r@xmath50 = 23.3 .
the average half - light radii of the spectroscopically selected @xmath15 galaxies is 0.7@xmath761 kpc , with an average stellar mass of 10@xmath77 within the acs region .
the concentration indices of these systems are moderate , with a average of c = [email protected] , typical for low mass ellipticals , or disks ( conselice 2003 ) .
the systems at @xmath9 with photometric redshifts in the acs region have very similar masses , sizes and morphologies as the spectroscopic sample .
interestingly , none of these lower redshift drgs are face - on disks , which is the most common galaxy type at @xmath9 ( conselice et al .
2005 ) . to further investigate the nature of the @xmath9 objects we combine the spectra of all drgs with spectroscopy to produce a co - added non - flux weighted ( figure 4 ) spectrum .
this spectrum shows that the @xmath9 drgs with emission lines , host star formation , agn activity , as well as evidence for post starbursts with ages 1 gyr , revealed through strong balmer absorption lines .
features produced by star formation , including [ oii ] and balmer absorption lines , are seen , as well as higher excitation lines such as [ neiii ] and [ oiii ] . the average ratio of ( [ oiii]@xmath785007)/(h@xmath79 @xmath784861 ) @xmath80 after correcting for h@xmath79 absorption .
this suggests a mixture of star formation and agns ( seyfert 2s ) could be responsible for the lower redshift sources ( veilleux & osterbrock 1987 ) .
the preliminary x - ray chandra catalog of the egs reveals that a lower limit of six @xmath18 drgs have bright x - ray detections to a limit of @xmath81 erg s@xmath22 @xmath82 , all with photo - zs , and soft hardness ratios .
other sources could be obscured agn in moderate l@xmath83 systems .
we conclude that @xmath15 systems at @xmath9 appear to have average masses , a moderate concentration of light , and are host to star formation and agn activity , but otherwise are heterogenous .
with several large area and deep nir surveys coming online soon , such as ukidss and vista , it is desirable to understand the properties of pure photometrically selected nir samples , such as the drgs with .
our results suggest that bright drgs are a mixed population , as other smaller studies have previously found ( reddy et al .
2005 ; papovich et al . 2006 ; van dokkum et al . 2006 ) .
the differences between the original drg , franx et al .
( 2003 ) , study and ours is due to the depth of the surveys , with franx et al .
being several magnitudes deeper , although most drgs studied in detail thus far are at @xmath84 . because drgs have unique seds , their photometric redshifts are potentially harder to measure accurately , and these results , and others that utilize photometric redshifts , should be viewed as tentative until significant numbers of spectroscopic redshifts become available .
this is demonstrated by a high absolute lower limit ( 10% ) contribution to the entire drg population from @xmath9 galaxies , which tend to be small , lower mass galaxies with optical agn signatures .
the morphological mix derived in this paper is however robust , regardless of the redshift distribution .
the drg population is morphologically dominated by compact galaxies , with edge - on spirals and peculiars making up 36% of the population . in the future multi - object near - ir spectrographs
will be necessary to make definitive progress in our understanding of non - uv bright galaxies at @xmath16 .
we thank the members of the aegis , palomar and deep2 teams , particularly kevin bundy , for their many invaluable contributions to the surveys that have made this paper possible .
we thank chuck steidel for allowing us to utilize the spectroscopic redshifts from his bx / bm survey before publication , and casey papovich and naveen reddy for enlightening discussions .
this work was supported by a nsf astronomy & astrophysics postdoctoral fellowship , pparc and nasa .
alc is supported by nasa through hubble fellowship grant hf-01182.01-a .
bundy , k. , ellis , r.s . , & conselice , c.j .
2005a , apj , 625 , 621 bundy , k. , et al . 2005b , astro - ph/0512465 conselice , c.j . ,
gallagher , j.s . ,
calzetti , d. , homeier , n. , & kinney , a. 2000a , aj , 119 , 79 conselice , c.j . ,
bershady , m.a . , & jangren , a. 2000b , apj , 529 , 886 conselice , c.j .
2003 , apjs , 147 , 1 conselice , c.j . , bershady , m.a . ,
dickinson , m. , & papovich , c. 2003 , aj , 126 , 1183 conselice , c.j .
, et al . 2004 , apj , 600 , 139l conselice , c.j . ,
blackburne , j. , & papovich , c. 2005 , apj , 620 , 564 conselice , c.j .
2006 , apj , 638 , 686 davis , m. , et al .
2003 , spie , 4834 , 161 franx , m. et al .
2003 , apj , 587 , 79l grazian et al .
2006 , astro - ph/0603095 lotz , j. , et al . 2006 , astro - ph/0602088 moustakas , l.a . ,
2004 , apj , 600 , 131l papovich , c. , dickinson , m. , giavalisco , m. , conselice , c.j . , &
ferguson , h.c .
2005 , apj , 631 , 101 papovich , c. et al . 2006 ,
apj , 640 , 92 reddy , n.a . ,
erb , d.k . ,
steidel , c.c . ,
shapley , a.e . ,
adelberger , k. , & pettini , m. 2005 , apj , 633 , 748 saracco , p. et al .
2001 , a&a , 375 , 1 smail , i. , owen , f.n . ,
morrison , g.e .
, keel , w.c . ,
ivison , r.j . , & ledlow , m.j .
2002 , apj , 581 , 844 trager , s.c . ,
faber , s.m . ,
worthey , g. , & gonzalez , j.j .
2000 , aj , 119 , 1645 van dokkum , p.g . , et al .
2006 , apj , veilleux , s. , & osterbrock , d.e .
1987 , apjs , 63 , 295 windhorst , r. , et al .
2002 , apjs , 143 , 113 | we use deep and wide near infrared ( nir ) imaging from the palomar telescope combined with deep2 spectroscopy and hubble space telescope ( hst ) and chandra space telescope imaging to investigate the nature of galaxies that are red in nir colors .
we locate these ` distant red galaxies ' ( drgs ) through the color cut @xmath0 over 0.7 deg@xmath1 , where we find 1010 drg candidates down to @xmath2 .
we combine 95 high quality spectroscopic redshifts with photometric redshifts from brijk photometry to determine the redshift and stellar mass distributions for these systems , and morphological / structural and x - ray properties for 107 drgs in the extended groth strip .
we find that many bright @xmath0 galaxies with @xmath3 20.5 are at redshifts @xmath4 , with 64% between @xmath5 .
the stellar mass distributions for these galaxies is broad , ranging from 10@xmath6 , but with most @xmath7 systems massive with m@xmath8 .
hst imaging shows that the structural properties and morphologies of drgs are also diverse , with the majority elliptical / compact ( 57% ) , and the remainder edge - on spirals ( 7% ) , and peculiar galaxies ( 29% ) .
the drgs at @xmath9 with high quality spectroscopic redshifts are generally compact , with small half - light radii , and span a range in rest - frame optical properties .
the spectral energy distributions for these objects differ from higher redshift drgs : they are bluer by one magnitude in observed @xmath10 color
. a pure ir color selection of high redshift populations is not sufficient to identify unique populations , and other colors , or spectroscopic redshifts are needed to produce homogeneous samples . |
the subject of partial differential equations related to generalized derivatives is a somewhat recent and unexploited research theme , with connections with physics and probability . by a generalized derivative we mean , _ grosso modo _ , a radon - nikodym derivative .
in this paper , we are concerned with the following partial differential equation @xmath4 where @xmath5 is the one dimensional torus and @xmath6 is a right continuous and periodically increasing function in the sense that @xmath7 for every @xmath8 . or else , @xmath1 can be understood as the distribution function of a probability measure @xmath9 on the torus @xmath10 .
a function @xmath11 for which @xmath12 is well defined and differentiable may have jump discontinuities at the discontinuity points of @xmath1 , namely the points with positive @xmath9-measure . in the section 2.7 of @xcite ,
it was proven the existence of a unique weak solution belonging to the space @xmath13 , { h^1_w(\mathbb t)})$ ] for the equation , where @xmath14 is a suitable sobolev - type space which admits discontinuous functions .
we restrict ourselves to the class of measures @xmath9 for which the lebesgue measure is absolutely continuous with respect to @xmath9 .
our main result is a continuous dependence of the unique solution @xmath3 of with respect to @xmath1 . by means of a sequence of transformations ,
we not only solve the problem of continuity but we actually give an explicit construction of the weak solution .
the subject of dynamics related to generalized derivatives has connections with different areas .
for instance , the book @xcite studies one - dimensional markov processes whose generators involve radon - nikodym derivatives .
in fractal analysis we cite @xcite .
related to krein - fellers operators , see @xcite .
partial differential equations related to this operator @xmath15 naturally come out in hydrodynamic limit and fluctuations of interacting particle systems in non - homogeneous medium , see @xcite .
the equation is , for some cases of the measure @xmath9 , in correspondence with classical pde s .
clearly , if @xmath9 is the lebesgue measure , the pde is equivalent to the classical heat equation in the one dimensional torus .
additionally , in the case where @xmath16 , where @xmath17 is the lebesgue measure , and @xmath18 is the dirac delta measure at zero , the pde is equivalent to the following heat equation with robin s boundary conditions : @xmath19 .
\end{cases}\ ] ] this equivalence was showed in @xcite .
notice that the boundary conditions above represent the fourier s law : the rate of heat transfer across the interface between two media is proportional to the difference of temperature in each medium . in this case
, the rate is given by the partial derivatives @xmath20 and the difference of temperature is given by @xmath21 . in @xcite , it was also described the behaviour of the solution @xmath22 of the equation above as a function of the parameter @xmath23 . it is proven that when @xmath24 , the function @xmath25 converges to the solution of the heat equation with newmann s boundary conditions . when @xmath26 , the convergence is towards the solution of the heat equation with periodic boundary conditions .
our main theorem covers this last case with much more generality .
the outline of the paper is the following . in section [ s2 ]
we present heuristics on how one can deduce an equivalent equation for the pde and the proof s scheme about continuous dependence of solutions with respect to @xmath1 . in section [ s3 ] precise definitions and statements
are given . in section [ s4 ]
we transform the equation into a classical pde with continuous weak solutions . in section [ s5 ]
we deal with the continuous dependence of the equivalent version of by means of a careful analysis on its fourier transform .
some auxiliary results are left to the appendix .
in this section we informally discuss the subject of this paper and the proof s general idea .
all arguments ahead are of heuristic nature .
precise definitions and statements will be presented in the next section .
is well known that the heat equation may be derived from fourier s law , which states that the heat transfer @xmath27 is proportional to the negative gradient of temperature @xmath3 , or else , @xmath28 an argument about conservation of energy leads to @xmath29 where @xmath30 and @xmath31 are functions of the position @xmath32 . in physical nomenclature , @xmath30 is the heat capacity and @xmath31 is the thermal conductivity .
equation models the case where @xmath31 may be degenerate in the following sense : assuming for a moment that @xmath1 is differentiable , equation takes the form @xmath33 so @xmath34 is the inverse of @xmath31 and represents the thermal resistance .
fourier s law takes henceforth the form @xmath35 when @xmath1 is not differentiable ( possibly not even continuous ) , we shall interpret this equality with @xmath34 and @xmath36 as schwartz distributions , and @xmath27 a continuous function .
let us consider the case when @xmath37 , hence @xmath1 and @xmath3 must have both a jump discontinuity at @xmath38 . from , we see that the ratio between the size of the jumps is the heat transfer at that point
. this agrees with the robin s boundary conditions in mentioned in the introduction . keeping this interpretation in mind , we will reparametrize the interval @xmath39 $ ] in such a way that the thermal conductivity becomes constant , leading to a pde with the classical laplacian operator ( with possibly zero heat capacity at some points ) . roughly speaking
, we are going to stretch " the support of the singular part of @xmath9 with respect to the lebesgue measure .
for instance , as showed in the figure [ fig1 ] , the point @xmath40 in the left graphic is transformed into the interval @xmath41 $ ] in the right graphic .
being @xmath1 a strictly increasing function , it has a continuous left inverse @xmath42 \to [ 0,1]$ ] such that @xmath43 for all @xmath44 $ ] .
as we shall see , a function @xmath45 with discontinuities at the same points as @xmath1 may be represented as a composition @xmath46 for some continuous function @xmath11 . under the change of variables @xmath47 ,
equation becomes @xmath48 where , according to , the function @xmath49 plays the role of the heat capacity
. the general strategy will be to establish the continuity of the solution @xmath50 of with respect to @xmath51 in a convenient function space , and then to prove that the composition @xmath52 is the solution of .
this transformation puts together in the same space functions which are discontinuous in distinct sets . as usual
, the topology on the measures will be given by the vague convergence . and @xmath50 are related by @xmath53 . the grey line is a @xmath54-linear interpolation . ] observe that if @xmath1 has a jump discontinuity at a point @xmath55 , namely @xmath56 , then @xmath57 in the interval @xmath58 $ ] .
this corresponds to an interval with zero heat capacity and since reduces to @xmath59 , the temperature must be the linear interpolation of the values of @xmath50 at the end points of the interval , see figure [ fig1 ] .
this fact may be interpreted as an `` infinite dispersion '' phenomena : any initial temperature at @xmath58 $ ] is completely dispersed at any positive time , and the initial condition @xmath60 will be satisfied only when @xmath61 .
this justifies the second equation in . in principle
, the function @xmath50 is defined only for @xmath62 . extending @xmath50 as zero in the negative half line
, we can take the fourier transform with respect to time in , obtaining @xmath63 the term @xmath64 appears because @xmath50 is discontinuous at @xmath65 so @xmath66 has a dirac delta .
now this equation is uncoupled in @xmath67 so it may be viewed as a one parameter family of periodic complex ode s @xmath68 with @xmath69 as the parameter . at this point , the classical theory of ode s assures the existence of a unique solution that depends continuously on @xmath51 and @xmath67 .
the main difficulty here is to show that we are able to anti - transform @xmath70 with respect to @xmath67 without loosing continuity .
this is the subject of the section [ s5 ] .
to a probability measure @xmath9 on the torus we can associate a unique right - continuous function @xmath6 such that @xmath71 , @xmath7 for every @xmath8 and such that if @xmath72 $ ] represents an interval in the torus then @xmath73 ) = w(b ) - w(a).\ ] ] the function @xmath1 completely characterizes the measure @xmath9 . [ defw ] we denote by @xmath74 the set of functions @xmath6 as above associated to the probability measures @xmath9 on the torus @xmath10 such that the lebesgue measure is absolutely continuous with respect to @xmath9 .
it follows that @xmath1 is strictly increasing .
this restriction for the set @xmath74 is stronger than the one assumed in @xcite namely , that @xmath75 for every open interval @xmath76 .
see remark [ weaker_condition ] in the appendix .
we say that a sequence @xmath77 converges vaguely to @xmath78 if @xmath79 for every function @xmath80 .
it is well known that this convergence is equivalent to the pointwise convergence of @xmath81 to @xmath1 in the continuity points of @xmath1 , see for instance @xcite .
we use the notation @xmath82 for the usual inner product in @xmath83 .
also @xmath84 will stand for the usual periodic - sobolev spaces and @xmath85 for the space of periodic @xmath86-hlder continuous functions .
now we present the definitions concerning the generalized derivative as in @xcite and @xcite . for a function @xmath87
, we define @xmath88 as follows : @xmath89 if the above limit exists and is finite .
[ defdw ] denote by @xmath90 the set of functions @xmath11 such that @xmath91 } \int_0^y g(z ) \ , dz \ ; dw(y)\ ] ] for some function @xmath92 in @xmath83 and some @xmath93 , with @xmath94 one can check that the function @xmath92 , as well as the constants @xmath23,@xmath30 , are unique .
the first requirement corresponds to the boundary condition @xmath95 and the second one to the boundary condition @xmath96 .
define the operator @xmath97 by @xmath98 it is easy to see that @xmath99 a.e in the notation of .
[ weak_w_pde ] we say that a measurable bounded function @xmath100 is a weak solution of if for all functions @xmath101 and every @xmath102 , @xmath103 here @xmath104 denotes the function @xmath105 .
we are in position to state our main results : [ dw_parametrization_intro ] let @xmath106
. then there exists @xmath107 such that @xmath46 .
[ mainthm ] let @xmath81 , @xmath78 .
let @xmath108 , @xmath3 be the corresponding unique weak solutions of the equations @xmath109 and respectively , where @xmath110 and @xmath106 .
note that by proposition [ dw_parametrization_intro ] , @xmath111 and @xmath112 for some @xmath113 .
assume that a. @xmath114 vaguely ; b. [ compatibility ] the functions @xmath115 are uniformly bounded ; c. @xmath116 in @xmath117 .
then there exist continuous functions @xmath118 such that * the functions @xmath119 are the unique weak solutions of equation with initial conditions @xmath120 and @xmath11 , respectively ; * @xmath121 and @xmath122 ; * for every @xmath123 , the functions @xmath124 and @xmath125 in the topology of @xmath126 . in order to illustrate the range of applicability of our theorem , we present some examples .
denote by @xmath17 the lebesgue measure on @xmath10 and by @xmath127 the indicator function of a set @xmath128 .
[ lebesgueplusdelta ] consider the measure @xmath129 . in this case , the function @xmath51 from equation is given by @xmath130 } + { \bf 1}_{[\frac 34,1]}$ ] .
we observe that equation is the classical ( periodic ) heat equation for @xmath131 \cup [ \frac 34 , 1]$ ] while for @xmath132 $ ] it reduces to @xmath59 . as a consequence , if we assume that @xmath133 is continuous then @xmath134 for all @xmath135 . the function @xmath136 satisfies the robin s boundary conditions as in , with @xmath137 .
if instead we consider @xmath138 then the theorem [ mainthm ] guarantees that the solution varies continuously with respect to @xmath139 and @xmath140 $ ] .
the laplacian operator with respect to fractal measures was considered in @xcite where its properties of self - similarities are exploited .
consider @xmath141 where @xmath142 is the usual ternary cantor `` staircase '' function .
observe that if @xmath143 $ ] is the cantor set , then @xmath144 is a cantor - like set of positive measure and @xmath145 . consider the usual uniform approximation @xmath146 by piecewise - linear continuous functions , and @xmath147 . theorem [ mainthm ] is applicable to this situation , although the corresponding solutions of and are hard to describe .
[ piquito ] consider the measures @xmath148 whose vague limit is @xmath149 .
as in example [ lebesgueplusdelta ] , the solutions exhibit robin s boundary conditions at the points @xmath150 and they converge to the solution of example [ lebesgueplusdelta ] . in words ,
the two boundary conditions overlap in the limit .
[ piquitomalo ] regarding equation in the situation of the previous example , the functions @xmath151 are shown in [ fig2 ] .
assume that the initial condition @xmath11 is also as in figure [ fig2 ] .
as described in the introduction , for fixed @xmath135 the solution of is linear in the intervals where @xmath57 , therefore the solution @xmath152 do not converge to @xmath153 uniformly in @xmath154 , and the theorem fails . in this case the convergence is only @xmath155 in time .
this counterexample is not relevant to equation since the initial conditions @xmath156 do not converge at @xmath157 , but it suggests there should exist a compatibility condition between @xmath51 and @xmath11 in order to have the desired continuity .
this condition is contained in the requirement of theorem [ mainthm ] that @xmath158 which does not hold in this case .
the reader should compare this example with proposition [ dw_parametrization ] in the next section .
vanishes . ]
a strictly increasing ( not necessarily continuous ) function @xmath159 \to [ 0,1]$ ] has a generalized inverse @xmath42 \to [ 0,1]$ ] defined as @xmath160 some properties of the generalized inverse are listed in the appendix .
if @xmath78 then @xmath161 is an absolutely continuous function and thus it is the primitive of a non - negative function @xmath49 , see [ w_is_abs_cont ] in the appendix for a proof . in this section
we show that equation is equivalent to equation in the sense that @xmath50 and @xmath3 are related by @xmath53 .
the equation has to be regarded in the weak sense defined as follows .
[ weak_a_pde ] denote @xmath162 .
we say that a continuous function @xmath163 is a weak solution of if , for all functions @xmath164 , @xmath165 here @xmath166 denotes the function @xmath167 .
we first characterize the space @xmath90 as a set of functions composed with @xmath1 .
[ dw_parametrization ] let @xmath168 .
then there exists @xmath169 such that @xmath170 and @xmath171 .
take @xmath168 , which according to definition [ defdw ] is represented as @xmath172 } \int_0^z g(r ) dr \ ; dw(z).\ ] ] by the substitution rule applied to the integral with respect to @xmath1 above , @xmath173 taking the substitution @xmath174 , we obtain @xmath175 finally @xmath170 , where @xmath176 is clearly in @xmath177)}$ ] and @xmath178 . in order to show that @xmath179 is periodic it is enough to use property and the substitution rule . for the second statement
, the discussion in definition [ defdw ] shows that @xmath180 . keeping this equivalence in mind
, we see that the following proposition is immediate .
[ equations_are_equivalent ] let @xmath163 be a continuous weak solution of .
then the function @xmath53 is a weak solution of as in definition [ weak_w_pde ] .
the statement follows from proposition [ dw_parametrization ] and the identity @xmath181 which is a consequence of the substitution rule . in section [ s5 ] we shall prove the continuity of the solution of with respect to the function @xmath49 under the weak-@xmath182 topology . since @xmath183 ,
the weak-@xmath182 convergence of the sequence @xmath151 is equivalent to the vague convergence of @xmath184 .
next , we state a general result relating the convergence of @xmath184 to the convergence of @xmath81 .
assume that @xmath1 is a _
strictly _ increasing function .
a sequence @xmath81 converges vaguely to @xmath1 if and only if it converges pointwise in the continuity points of @xmath1 . in that case , the sequence of generalized inverses @xmath184 converges pointwise to @xmath161 and thus , vaguely .
the proof of the proposition above can be found in ( * ? ? ?
* proposition 0.1 , page 5 ) . in consequence , under the hypothesis of theorem [ mainthm ] , namely , that @xmath114 vaguely , we have that @xmath185 weakly in @xmath182 .
as mentioned in the introduction , equation is in correspondence with the family of complex ode s . in this section
our aim is to prove that solutions of are bounded by an @xmath186-function of the parameter @xmath67 .
then , considering @xmath70 as a function @xmath187 , this will allow us to take the inverse fourier transform with respect to @xmath67 .
a quick inspection to the easy case @xmath188 gives some insight on what we should expect .
the periodic solutions of equation @xmath189 can be described by taking the fourier series with respect to @xmath32 .
the transformed equation is @xmath190 and the solution is @xmath191 of course , @xmath70 can be thought as a function @xmath187 and we need to describe its behaviour as a function of @xmath67 .
first we observe that @xmath192 so we have a singularity at @xmath193 . to avoid this problem we require @xmath194 which guarantees that @xmath70 is bounded .
since the equation is linear with respect to @xmath50 , this requirement do not represent a restriction .
moreover , the solution @xmath195 converges to its average @xmath196 as @xmath197 , thus we need @xmath194 in order to expect @xmath50 to be an @xmath155-function of time . by cauchy - schwarz ,
parseval , and the inequality @xmath198 , @xmath199 thus , if we assume only that @xmath200 , then we can only expect that @xmath201 , which is not in @xmath155 . in order to obtain a sufficiently rapid decay with respect to @xmath67
we may impose the condition that @xmath202 , which means @xmath203 and yields @xmath204 however , this inequality only guarantees that @xmath70 is bounded by an @xmath155-function of @xmath67 and does not imply continuity in time of @xmath205 $ ] .
we observe that this bound can not be improved because the function @xmath50 has a discontinuity at @xmath206 .
we overcome this difficulty in the following way : subtracting to @xmath195 the function @xmath207 , ( here @xmath208 is the heaveside step function ) we obtain a continuous function .
the fourier transform in time of the difference is @xmath209 and one can easily verify that @xmath210 finally , the estimate applied to @xmath211 gives @xmath212 which yields the desired @xmath182 bound for @xmath31 , provided by the fact that @xmath213 .
consequently , the inverse fourier transform @xmath214 is continuous in time , for @xmath215 .
unfortunately , the estimate fails in the general case when @xmath51 is not constant . under additional assumptions on @xmath1 ,
this estimate holds and one can obtain solutions with improved regularity .
this is left for future work . for the rest of the section ,
all integrals are with respect to the lebesgue measure .
we will denote @xmath216 for @xmath217 and @xmath218 , we define the operator @xmath219 by @xmath220 here @xmath221 is identified with the closed subspace of @xmath222)}$ ] , consisting of functions @xmath70 such that @xmath223 and @xmath224 .
it is easy to prove that @xmath225 is a fredholm operator of index @xmath226 . saying that @xmath70 is a solution of equation is equivalent to @xmath227 . for @xmath218
, we define the seminorms @xmath228_a = \sqrt{\int a|u|^2 } \;\;\hbox { and } \;\ ; \|u\|_{{h^1_a}}= \sqrt{\|u'\|_2 ^ 2 + \int a|u|^2}\;.\ ] ] observe that if @xmath70 is bounded , then @xmath229_a \leq \|u\|_\infty$ ] .
[ equivalentnorm ] for any @xmath230 , we have @xmath231 where @xmath232 is independent of @xmath51 .
in particular @xmath233 is a norm .
observe that @xmath229_a^2 \geq \min\{|u|^2\}$ ] and @xmath234_a^2 \leq 2 \|u\|_2 \|u'\|_2 + [ u]_a^2\;. \end{split}\ ] ] by young inequality , @xmath235_a^2
\leq \frac 12 \|u\|_\infty^2 + 2 \|u\|_{{h^1_a}}^2\;,\ ] ] from what we get @xmath236 and finally @xmath237 [ t_inf_bounded ] let @xmath218 and @xmath238 .
then @xmath219 is an isomorphism of banach spaces .
moreover , if @xmath11 is bounded and @xmath239 , then @xmath240 and @xmath241_a$ ] . for the first statement , since @xmath242 is a fredholm operator of index @xmath226 it suffices to prove injectivity .
computing @xmath243 we note that @xmath244 implies @xmath245 .
this proves the first statement . for the second ,
notice that by hlder inequality @xmath246_a [ u]_a\,.\ ] ] hence @xmath239 put together with imply @xmath247_a [ u]_a \leq [ f]_a \|u\|_{{h^1_a}}\ ] ] and the result follows .
[ u_is_bounded ] let @xmath218 , assume @xmath248 and @xmath70 is the solution of .
then @xmath249_a\ ; .\ ] ] multiply equation by @xmath250 and integrate in order to obtain @xmath251 denote @xmath252 . since @xmath253 , @xmath254_a [ u - p]_a \leq [ f]_a \|u - p\|_\infty \leq [ f]_a \|u'\|_2\;,\ ] ] where we have used the poincar inequality in the last inequality of above .
notice that since the left - hand side integrals of are real , @xmath255_a \|u'\|_2\;.\ ] ] consequently , @xmath256_a\;.\ ] ] now we need to bound @xmath257 .
integrating each side of we get @xmath258 .
denote the ( real ) inner product of complex numbers by @xmath259 .
equation @xmath258 implies @xmath260 .
then @xmath261 we therefore obtain @xmath262 , leading to @xmath263_a\;.\ ] ] [ u_falls_fast ] let @xmath218 , assume @xmath11 is bounded and @xmath70 is the solution of .
then , for any @xmath264 , @xmath265_a\;,\ ] ] where the constant @xmath232 is independent of @xmath51 . by lemma [ equivalentnorm ] , it is sufficient to show that @xmath266_a\;.\ ] ] from we have that @xmath267_a [ u]_a\;.\ ] ] thus , @xmath268_a^2 \leq [ f]_a [ u]_a\;,\ ] ] which implies @xmath269_a \leq |\xi|^{-1 } [ f]_a \leq |\xi|^{-\frac 12 } [ f]_a\ ] ] because @xmath264 . on the other hand , from ,
@xmath270_a [ u]_a \leq [ f]_a^2 |\xi|^{-1}\;,\ ] ] implying @xmath271_a\;.\ ] ] putting together and , we obtain that @xmath272_a$ ] , finishing the proof .
[ u_falls_faster ] let @xmath218 , @xmath264 , @xmath107 and in addition @xmath273 , where @xmath92 is a bounded function .
let @xmath70 be the solution of .
then , for @xmath31 defined through @xmath274 holds the estimate @xmath275_a\;,\ ] ] where the constant @xmath232 does not depend on @xmath51 . moreover , if @xmath276 , then @xmath277 is bounded by @xmath278_a$ ] . replacing into equation
we obtain @xmath279 since @xmath280 , we get @xmath281 applying lemma [ u_falls_fast ] , we obtain @xmath282_a \leq c |\xi|^{-\frac 32 } [ f+g]_a\;.\ ] ] and follows . additionally , if we have @xmath283 , then @xmath284 because @xmath11 is periodic .
this permits to invoke lemma [ u_is_bounded ] proving that @xmath285_a$ ] . throughout this section
we assume the conditions of theorem [ mainthm ] , namely , a. @xmath114 vaguely , which implies that @xmath185 weakly in @xmath182 ; b. @xmath286 with @xmath287 uniformly bounded , as in proposition [ dw_parametrization ] ; c. @xmath116 in @xmath117 . the main result of this section is contained in the propositions [ v_of_a_is_continuous ] and [ v_is_the_solution ] that is , the solutions @xmath288 of converge to @xmath50 in the space @xmath289 for every @xmath123 . for simplicity
, we shall assume without loss of generality that @xmath290 [ u_of_a_is_w_continuous_pointwise ] fix @xmath238 .
let @xmath291 and @xmath292 .
in other words , @xmath293 is the solution of with @xmath51 replaced by @xmath151 .
denote @xmath294 and analogous definition for @xmath295 .
then , it holds the convergence @xmath296 uniformly .
recall from lemma [ u_falls_faster ] that @xmath297 since @xmath298_{a_n } \leq \|f_n + g_n\|_\infty$ ] is bounded , by lemma [ t_inf_bounded ] , the sequence @xmath299 is @xmath300 bounded , hence precompact in @xmath301 .
we shall prove that @xmath31 is the only @xmath301-accumulation point of @xmath299 .
let @xmath302 be such a point and @xmath303 a test function .
take a subsequence of @xmath299 converging to @xmath304 and call it again @xmath299 . multiplying equation by @xmath305 and then integrating by parts , @xmath306 adding and subtracting suitable terms , @xmath307 since @xmath308 in @xmath301 , @xmath309 , @xmath310 weakly in @xmath182 , and @xmath116 in @xmath117 , we can take limits obtaining @xmath311 therefore , @xmath312 .
then @xmath296 uniformly as desired .
next , we recall the classical definition of weighted @xmath155-spaces , see @xcite for details . for @xmath313 ,
let @xmath314 be the set of measurable functions @xmath315 such that @xmath316 .
it is a banach space with the norm @xmath317 .
consider the space @xmath318 with the norm @xmath319 [ u_of_a_is_w_continuous ] denote @xmath320 and @xmath321 .
denote similarly @xmath187 and @xmath322 .
then , for every @xmath123 , holds the convergence @xmath296 in the space @xmath318 .
clearly , @xmath323 and we intend to prove that the last integral goes to @xmath226 as @xmath324 .
since the functions @xmath120 fulfil the conditions of lemmas [ u_falls_faster ] and [ u_is_bounded ] , @xmath325 where @xmath232 is independent of @xmath326 and @xmath67 . for each fixed @xmath327 ,
lemma [ u_of_a_is_w_continuous_pointwise ] implies that @xmath328 as @xmath329 . then using the bound
we find that the integrand in is bounded by @xmath330 finally , we apply dominated convergence to conclude that @xmath331 as @xmath324 .
[ v_of_a_is_continuous ] following the notation of lemma [ u_of_a_is_w_continuous ] , denote @xmath332 , where @xmath333 is the fourier transform with respect to @xmath67 , and analogous notation for @xmath195 .
then , the convergence @xmath125 holds in @xmath334 .
the operator @xmath335 is continuous ( * ? ? ?
* thm 3.2 , pag .
47 ) so clearly it extends continuously to an operator @xmath336 . in consequence ,
@xmath337 in @xmath338 , as @xmath329 . on the other hand , @xmath339 and @xmath340 , where @xmath341 is the heaveside step function , which is not in @xmath342 but it is in @xmath343 .
lastly , since @xmath116 uniformly , it follows that @xmath344 in @xmath345 , and then @xmath125 in @xmath334 , as desired .
it remains only to show that @xmath50 is indeed a solution of the corresponding equation .
[ v_is_the_solution ] the function @xmath195 from proposition [ v_of_a_is_continuous ] solves the equation in the weak sense of definition [ weak_a_pde ] .
we take the fourier transform of @xmath50 with respect to time , which is the function @xmath187 from the previous lemmas .
we know that @xmath346 solves the equation @xmath347 since we do not know _ a priori _ if @xmath348 exists , we multiply the equation above by a test function @xmath349 and then integrate , obtaining @xmath350 by fubini theorem , @xmath351 which can be written in the form @xmath352 where @xmath353 and @xmath354 are continuous functions , and @xmath355 is a constant .
the identity above implies that @xmath356 is the weak derivative of @xmath357 and that @xmath358 .
finally , we infer that @xmath359 and the statement follows .
let @xmath195 be the function defined in proposition [ v_of_a_is_continuous ] , which by proposition [ v_is_the_solution ] is the weak solution of equation .
defining @xmath53 and recalling proposition [ equations_are_equivalent ] , we conclude that @xmath3 is a weak solution of as in definition [ weak_w_pde ] .
finally , proposition [ v_of_a_is_continuous ] gives the desired convergence , concluding the proof .
let @xmath161 be the generalized inverse of @xmath1 defined by @xmath360 which is an right inverse of @xmath1 , or else @xmath361 .
if @xmath362 is a continuity point of @xmath1 , it holds also that @xmath363 .
let @xmath370 $ ] be a measurable set with @xmath371 .
let @xmath372 denote the set of discontinuity points of @xmath1 .
clearly @xmath373 because @xmath372 is only denumerable .
since @xmath369 , by proposition [ pullback_measure ] , we only need to show that @xmath374 .
[ weaker_condition ] it is almost immediate that the converse of proposition [ w_is_abs_cont ] is also true . in @xcite , the condition @xmath369
is replaced by a weaker one , namely that @xmath75 for every open interval @xmath379 .
this is not enough to guarantee that @xmath161 is an absolutely continuous function .
for instance , the measure @xmath380 , which has a delta at each rational number @xmath381 is a counterexample because it assigns positive measure to every open interval , but @xmath161 is a cantor - like staircase function , thus not an absolutely continuous function .
t. franco , p. gonalves , a. neumann .
_ phase transition of a heat equation with robin s boundary conditions and exclusion process _ ,
arxiv:1210.3662 . accepted for publication in the transactions of the american mathematical society , ( 2013 ) . c. kipnis and c. landim .
_ scaling limits of interacting particle systems_. grundlehren der mathematischen wissenschaften [ fundamental principles of mathematical sciences ] , 320 .
springer - verlag , berlin , ( 1999 ) .
_ construction and generators of one - dimensional quasi - diffusions with applications to selfaffine diffusions and brownian motion on the cantor set_. stoch . and
stoch . rep . *
42 * , 93114 , ( 1993 ) . | we consider here a generalized heat equation @xmath0 , where @xmath1 is a finite measure on the one dimensional torus , and @xmath2 is the radon - nikodym derivative with respect to @xmath1 .
such equation has appeared in different contexts , being related to physical systems and representing a large class of classical and non - classical parabolic equations . as a natural assumption on @xmath1
, we require that the lebesgue measure is absolutely continuous with respect to @xmath1 .
the main result here presented consists in proving , for a suitable topology , a continuous dependence of the solution @xmath3 as a function of @xmath1 . |
it is well - known that due to the fading effect , the transmission over wireless channels suffers from severe attenuation in signal strength .
performance of wireless communication is much worse than that of wired communication . for the simplest point - to - point communication system , which is composed of one transmitter and one receiver only
, the use of multiple antennas can improve the capacity and reliability .
space - time coding and beamforming are among the most successful techniques developed for multiple - antenna systems during the last decades @xcite . however , in many situations , due to the limited size and processing power , it is not practical for some users , especially small wireless mobile devices , to implement multiple antennas . thus , recently , wireless network communication is attracting more and more attention . a large amount of effort has been given to improve the communication by having different users in a network cooperate .
this improvement is conventionally addressed as cooperative diversity and the techniques cooperative schemes .
many cooperative schemes have been proposed in literature @xcite .
some assume channel information at the receiver but not the transmitter and relays , for example , the noncoherent amplify - and - forward protocol in @xcite and distributed space - time coding in @xcite .
some assume channel information at the receiving side of each transmission , for example , the decode - and - forward protocol in @xcite and the coded - cooperation in @xcite .
some assume no channel information at any node , for example , the differential transmission schemes proposed independently in @xcite .
the coherent amplify - and - forward scheme in @xcite assumes full channel information at both relays and the receiver . but only channel direction information is used at relays . in all these cooperative schemes ,
the relays always cooperate on their highest powers .
none of the above pioneer work allow relays to adjust their transmit powers adaptively according to channel magnitude information , and this is exactly the concern of this paper .
there have been several papers on relay networks with adaptive power control . in @xcite , outage capacity of networks with a single relay and
perfect channel information at all nodes were analyzed .
both work assume a total power constraint on the relay and the transmitter .
a decode - and - forward protocol is used at the relay , which results in a binary power allocation between the relay and the transmitter . in @xcite ,
performance of networks with multiple amplify - and - forward relays and an aggregate power constraint was analyzed .
a distributive scheme for the optimal power allocation is proposed , in which each relay only needs to know its own channels and a real number that can be broadcasted by the receiver .
another related work on networks with one and two amplify - and - forward relays can be found in @xcite . in @xcite , outage minimization of single - relay networks with limited channel - information feedback
is performed .
it is assumed that there is a long - term power constraint on the total power of the transmitter and the relay . in this paper , we consider networks with a general number of amplify - and - forward relays and we assume a separate power constraint on each relay and the transmitter . due to the difference in the power assumptions , compared to @xcite , analysis of this new model is more difficult and totally different results are obtained . for multiple - antenna systems ,
when there is no channel information at the transmitter , space - time coding can achieve full diversity @xcite . if the transmitter has perfect or partial channel information , performance can be further improved through beamforming since it takes advantage of the channel information ( both direction and strength ) at the transmit side to obtain higher receive snr @xcite . with perfect channel information or high quality channel information feedback from the receiver at the transmitter , one - dimensional beamforming
is proved optimal @xcite .
the more practical multiple - antenna systems with partial channel information at the transmitter , channel statistics or quantized instantaneous channel information , are also analyzed extensively @xcite . in many situations , appropriate combination of beamforming and space - time coding
outperforms either one of the two schemes alone @xcite . in this paper
, we will see similar performance improvement in networks using network beamforming over distributed space - time coding and other existing schemes such as best - relay selection and coherent amplify - and - forward .
we consider networks with one pair of transmitter and receiver but multiple relays .
the receiver knows all channels and every relay knows its own channels perfectly . in networks with a direct link ( dl ) between the transmitter and the receiver
, we also assume that the transmitter knows the dl fully .
a two - step amplify - and - forward protocol is used , where in the first step , the transmitter sends information and in the second step , the transmitter and relays , if there is a dl , transmit .
we first solve the power control problem for networks with no dl analytically .
the exact solution can be obtained with a complexity that is linear in the number of relays .
then , to perform network beamforming , we propose two distributive strategies in which a relay needs only its own channel information and a low - rate broadcast from the receiver .
simulation shows that the optimal power control or network beamforming outperforms other existing schemes .
we then consider networks with a dl during the first transmission step , the second transmission step , and both .
for the first case , the power control problem is proved to be the same as the one in networks without the dl . for the other two cases , recursive
numerical algorithms are provided .
simulation shows that they have much better performance compared to networks without power control .
we should clarify that only amplify - and - forward is considered here . for decode - and - forward ,
the result may be different and it depends on the details of the coding schemes .
the paper is organized as follows . in the next section ,
the relay network model and the main problem are introduced .
section [ sec - pl ] works on the power control problem in relay networks with no dl and section [ sec - dl ] considers networks with a dl .
section [ sec - conclusion ] contains the conclusion and several future directions .
consider a relay network with one transmit - and - receive pair and @xmath0 relays as depicted in fig .
[ fig - network ] .
every relay has only one single antenna which can be used for both transmission and reception .
denote the channel from the transmitter to the @xmath1th relay as @xmath2 and the channel from the @xmath1th relay to the receiver as @xmath3 .
if the dl between the transmitter and the receiver exists , we denote it as @xmath4 .
we assume that the transmitter knows @xmath4 , the @xmath1th relay knows its own channels @xmath2 and @xmath3 , and the receiver knows all channels @xmath5 and @xmath6 .
the channels can have both fading and path - loss effects .
actually , our results are valid for any channel statistics .
we assume that for each transmission , the powers used at the transmitter and the @xmath1th relay are no larger than @xmath7 and @xmath8 , respectively .
note that in this paper , only short - term power constraint is considered , that is , there is an upper bound on the average transmit power of each node for each transmission .
a node can not save its power to favor transmissions with better channel realizations .
we use a two - step amplify - and - forward protocol . during the first step
, the transmitter sends @xmath9 . the information symbol @xmath10 is selected randomly from the codebook @xmath11 .
if we normalize it as @xmath12 , the average power used at the transmitter is @xmath13 .
the @xmath1th relay and the receiver , if a dl exists during the first step , receive @xmath14 respectively . @xmath15 and @xmath16 are the noises at the @xmath1th relay and the receiver at step 1 .
we assume that they are @xmath17 . during the second step
, the transmitter sends @xmath18 , if a dl exists during this step . at the same time
, the @xmath1th relay sends @xmath19 the average transmit power of the @xmath1th relay can be calculated to be @xmath20 .
if we assume that @xmath4 keeps constant for the two steps , the receiver gets @xmath21 @xmath22 is the noise at the receiver at step 2 , which is also assumed to be @xmath17 .
note that if the transmitter sends during both steps , we assume that the total average power it uses is no larger than @xmath7 . with this ,
the total average power in transmitting one symbol is no larger than @xmath23 .
clearly , the coefficients @xmath24 are introduced in the model for power control .
the power constraints at the transmitter and relays require that @xmath25 and @xmath26 .
our network beamforming design is thus the design of @xmath27 and @xmath28 , such that the error rate of the network is the smallest .
this is equivalent to maximize the receive snr , or the total receive snr of both branches if a dl exists during the first step . from ( [ x2-old ] )
, we can easily prove that an optimal choice of the angles are @xmath29 and @xmath30 .
that is , match filters should be used at relays and the transmitter during the second step to cancel the phases of their channels and form a beam at the receiver .
we thus have @xmath31 what is left is the optimal power control , i.e. , the choice of @xmath32 .
this is also the main contribution of our work .
in this section , we investigate the optimal adaptive power control at the transmitter and relays in networks without a dl .
section [ sec - pl - result ] presents the analytical power control result .
section [ sec - pl - discussion ] comments on the result and gives distributive schemes for the optimal power control .
section [ sec - pl - simulation ] provides simulated performance . with no dl , we have @xmath33 and @xmath34 . from ( [ x2 ] )
, the receive snr can be calculated to be @xmath35 it is an increasing function of @xmath36 .
therefore , the transmitter should always use its maximal power , i.e. , @xmath37 .
the receive snr is thus : @xmath38 before going into details of the snr optimization , we first introduce some notation to help the presentation .
@xmath39 indicates the inner product .
@xmath40 indicates the 2-norm .
@xmath41 indicates the probability .
@xmath42 denotes the @xmath1th coordinate of vector @xmath43 and @xmath44 denotes the @xmath45-dimensional vector @xmath46^t$ ] , where @xmath47 represents the transpose . if @xmath48 are two @xmath0-dimensional vectors , @xmath49 means @xmath50 for all @xmath51 .
@xmath52 is the @xmath0-dimensional vector with all zero entries .
denote the set @xmath53 or equivalently , @xmath54 for @xmath51 , as @xmath55 .
for @xmath56 , denote the set @xmath57 as @xmath58 , where @xmath59 is a @xmath45-subset of @xmath60 .
define @xmath61 , \bb=\left[\begin{array}{c } \frac{|f_1g_1|\sqrt{p_1}}{\sqrt{1+|f_1|^2p_0 } } \\ \vdots \\
\frac{|f_rg_r|\sqrt{p_r}}{\sqrt{1+|f_r|^2p_0 } } \end{array } \right ] , \ba=\left[\begin{array}{c } \frac{|g_1|\sqrt{p_1}}{\sqrt{1+|f_1|^2p_0 } } \\ \vdots \\ \frac{|g_r|\sqrt{p_r}}{\sqrt{1+|f_r|^2p_0}}\end{array}\right ] , \hspace{3 mm } \mbox{and } \hspace{3 mm } a=\diag\{\ba\},\ ] ] where @xmath62 indicates the diagonal matrix whose @xmath1th diagonal entry is @xmath42 . with the transformation @xmath63 , or equivalently , @xmath64 , we have @xmath65 where @xmath66 \left[\begin{array}{ccc}\frac{|f_1g_1|\sqrt{p_1}}{\sqrt{1+|f_1|^2p_0 } } \\ \vdots \\ \frac{|f_rg_r|\sqrt{p_r}}{\sqrt{1+|f_r|^2p_0}}\end{array}\right ] = \left[\begin{array}{ccc}|f_1| \\ \vdots
\\ |f_r| \end{array}\right].\ ] ] the receive snr optimization problem is thus equivalent to @xmath67 the difficulty of the problem lies in the shape of the feasible set .
if @xmath68 is constrained on a hypersphere , that is , @xmath69 , the solution is obvious at least geometrically . given that @xmath69 , @xmath70 where @xmath71 is the angle between @xmath72 and @xmath68 .
the optimal solution should be the vector which has the smallest angle with @xmath72 .
thus , we decompose ( [ opt - problem ] ) as @xmath73 since @xmath74 and @xmath75 , we assume that @xmath76 and @xmath77 . define @xmath78 for @xmath51 and , for the sake of presentation , define @xmath79 .
order @xmath80 as @xmath81 @xmath82 is thus an ordering of @xmath83 and @xmath84 .
define @xmath85 since @xmath86 , we have @xmath87 for @xmath88 .
thus , the feasible interval of the radius , @xmath89 $ ] , can be decomposed into the following @xmath0 intervals : @xmath90=[r_0,r_1]\cup [ r_1,r_2]\cup \cdots\cup [ r_{r-2},r_{r-1}]\cup [ r_{r-1},r_r].\ ] ] we denote @xmath91 $ ] for @xmath92 .
thus , ( [ opt - problem-1 ] ) is equivalent to @xmath93 we have decomposed the optimization problem into @xmath0 subproblems .
we now work on the @xmath1th subproblem : @xmath94 denote the solution of the inner optimization problem , @xmath95 as @xmath96 .
we have the following two lemmas .
@xmath97 for @xmath98 .
[ lemma - edge ] we prove this lemma by contradiction .
assume that @xmath99 for some @xmath100 .
we first show that there exists an @xmath101 such that @xmath102 .
assume that @xmath103 for all @xmath104 .
we have @xmath105 .
thus , @xmath106 this contradicts @xmath107 .
thus , there exists an @xmath101 such that @xmath102 .
define another vector @xmath108 as @xmath109 , @xmath110 , and @xmath111 for @xmath112 , where @xmath113 since we have assumed that @xmath99 and have just proved that @xmath102 , such @xmath114 is achievable . to contradict the assumption that @xmath96 is the optimal , it is enough to prove the following two items : 1 .
@xmath108 is a feasible point : @xmath115 and @xmath116 , 2 .
@xmath117 . from the definition of @xmath108 ,
we have @xmath118 since @xmath119 , we have @xmath120 .
also , since @xmath121 , we can easily prove that @xmath122 and @xmath123 .
thus , @xmath116 . the first item has been proved . for the second item , since @xmath124 , we have @xmath125 @xmath126 for @xmath127 . [ lemma - scale ] from lemma [ lemma - edge ] , @xmath97 for @xmath98 .
thus , ( [ inner - problem ] ) can be written as @xmath128 define @xmath129 .
it is obvious that @xmath130 for all @xmath131 .
in other words , to maximize the inner product , @xmath132 should have the same direction as @xmath133 .
thus , we only need to show that this direction is feasible for @xmath134 .
this is equivalent to show that @xmath135 for any @xmath134 .
we can easily prove that @xmath136 where @xmath137 $ ] for @xmath92 .
thus , for any @xmath138 and @xmath139 , we have @xmath140 .
hence , @xmath135 .
combining lemma [ lemma - edge ] and lemma [ lemma - scale ] , we have @xmath141 and thus @xmath142 we have solved the inner optimization of subproblem @xmath1 .
the solution of the @xmath0 subproblems can thus be obtained . for @xmath51 ,
define @xmath143 the solution of subproblem @xmath144 is @xmath145 .
the solution of subproblem @xmath1 for @xmath146 is @xmath147 that is defined as @xmath148 [ lemma - problem - i ] from ( [ inner - solution ] ) , subproblem @xmath1 is equivalent to the following 1-dimensional optimization problem : @xmath149 when @xmath150 , ( [ decom - problem-1 ] ) is equivalent to @xmath151 .
since @xmath152 is an increasing function of @xmath153 , its maximum is at @xmath154 . for @xmath146 ,
define @xmath155 we have , @xmath156 thus , @xmath157 if @xmath158 and @xmath159 if @xmath160 .
so , if @xmath161 , the optimal solution is reached at @xmath162 .
otherwise , the optimal solution is reached at @xmath163 . from ( [ zi ] ) , subproblem @xmath1
is solved at @xmath147 as defined in ( [ sub - solution ] ) .
now , we can work on the relay power control problem presented in ( [ opt - problem ] ) .
define @xmath164 as @xmath165 the solution of the snr optimization is @xmath166 , where @xmath167 is the smallest @xmath1 such that @xmath168 .
[ thm - main ] first , since @xmath79 , we have @xmath169 .
thus , @xmath167 exists .
also , since @xmath170 , and @xmath171 decreases with @xmath172 , we have @xmath173 for @xmath174 .
this means that @xmath166 is in the feasible region of the optimization problem .
denote @xmath175 note that @xmath176 .
since @xmath177 , @xmath178 is also a feasible point of subproblem 1 .
thus , @xmath179 due the optimality of @xmath180 in subproblem 1 .
this means that there is no need to consider subproblem 0 . for @xmath181 , if @xmath182 , @xmath183 and @xmath184 since @xmath185 , @xmath147 is a feasible point of subproblem @xmath186 .
thus , @xmath187 due the optimality of @xmath188 in subproblem @xmath186 .
this means that there is no need to consider subproblem @xmath186 .
thus , we only need to check those @xmath147 s with @xmath168 , and find the one that results in the largest receive snr . from the definition in ( [ sub - solution - x ] ) , this is the same as to check those @xmath164 s with @xmath168 .
now , we prove that @xmath189 if @xmath190 . first , from @xmath190 , we have @xmath191 since @xmath192 , we can prove easily that @xmath193 thus , we only need to check those @xmath164 s for @xmath194 and find the one causing the largest receive snr . from previous discussion , @xmath195 . define @xmath196 . now , we prove that @xmath197 for @xmath194 .
from the proof of lemma 3 , we have @xmath198 thus , the optimal power control vector that maximizes the receive snr is @xmath166 .
it is natural to expect the power control at relays to undergo an on - or - off scenario : a relay uses its maximum power if its channels are good enough and otherwise not to cooperate at all .
our result shows otherwise .
the optimal power used at a relay can be any value between 0 and its maximal power .
in many situations , a relay should use partial of its power , whose value is determined not only by its own channels but all others as well .
this is because every relay has two effects on the transmission .
for one , it helps the transmission by forwarding the information , while for the other , it harms the transmission by forwarding noise as well .
its transmit power has a non - linear effect on the powers of both the signal and the noise , which makes the optimization solution not an on - or - off one , not a decoupled one , and , in general , not even a differentiable function of channel coefficients . as shown in theorem [ thm - main ] and lemma [ lemma - problem - i ] , the fraction of power used at relay @xmath172 satisfies @xmath199 for @xmath200 and @xmath201 for @xmath202 .
thus , the @xmath167 relays whose @xmath203 s are the largest use their maximal powers . since @xmath204 , there is at least one relay that uses its maximum power .
this tells us that the relay with the largest @xmath203 always uses its maximal power .
the remaining @xmath205 relays whose @xmath203 s are smaller only use parts of their powers . for @xmath202 ,
the power used at the @xmath172th relay is @xmath206 , which is proportional to @xmath207 since @xmath208 is a constant for each channel realization .
although @xmath209 does not appear explicitly in the formula , it affects the decision of whether the @xmath172th relay should use its maximal power . actually , in determining whether a relay should use its maximal power , not only do the channel coefficients and power constraint at this relay account , but also all other channel coefficients and power constraints .
the power constraint of the transmitter , @xmath7 , plays a roll as well .
due to these special properties of the optimal power control solution , it can be implemented distributively with each relay knowing only its own channel information . in the following ,
we propose two distributed strategies .
one is for networks with a small number of relays , and the other is more economical in networks with a large number of relays .
the receiver , which knows all channels , can solve the power control problem .
when the number of relays , @xmath0 , is small , the receiver broadcasts the indexes of the relays that use their full powers and the coefficient @xmath208 .
if relay @xmath172 hears its own index from the receiver , it will use its maximal power to transmit during the second step .
otherwise , it will use power @xmath210 .
the bits needed for the feedback is @xmath211 where @xmath167 is the number of relays that use their maximal powers and @xmath212 is the number of bits needed in broadcasting the real number @xmath208 . instead , the receiver can also broadcast two real numbers : @xmath208 and a real number @xmath213 that satisfies @xmath214 .
relay @xmath172 calculates its own @xmath80 .
if @xmath215 , relay @xmath172 uses its maximal power . otherwise , it uses power @xmath210 .
the number of bits needed for the feedback is @xmath216 .
thus , when @xmath0 is large , this strategy needs less bits of feedback compared to the first one .
networks with an aggregate power constraint @xmath217 on relays were analyzed in @xcite . in this case , with the same notation in section [ sec - pl - result ] , @xmath218 and @xmath219 .
the optimal solution is @xmath220 @xmath221 is a function of its own channels @xmath222 only and an extra coefficient @xmath223 , which is the same for all relays .
therefore , this power allocation can be done distributively with the extra knowledge of one single coefficient @xmath224 , which can be broadcasted by the receiver . in our case
, every relay has a separate power constraint .
this is a more practical assumption in sensor networks since every sensor or wireless device has its own battery power limit .
the power control solutions of the two cases are totally different .
if relay selection is used and only one relay is allowed to cooperate , it can be proved easily that we should choose the relay with the highest @xmath225 we call @xmath226 the relay selection function since a relay with a larger @xmath227 results in a higher receive snr . while all relays are allowed to cooperate , the concepts of the best relay and relay selection function are not clear .
since the power control problem is a coupled one , it is hard to measure how much contribution a relay has . as discussed before , in network beamforming
, a relay with a larger @xmath80 does not necessarily use a larger power or has more contribution .
but we can conclude that if @xmath228 , the fraction of power used at relay @xmath45 , @xmath229 , is no less than the fraction of power used at relay @xmath230 , @xmath231 .
it is worth to mention that in network beamforming , relays with larger enough @xmath203 s use their maximal powers no matter what their maximal powers are .
actually , it is not hard to see that if at one time channels of all relays are _ good _ , every relay should use its maximum power . in this section ,
we show simulated performance of network beamforming and compare it with performance of other existing schemes .
figures [ fig - r2l2 ] and [ fig - r3l2 ] show performance of networks with rayleigh fading channels and the same power constraint on the transmitter and relays . in other words , @xmath232 are @xmath17 and @xmath233 .
the horizontal axis of the figures indicates @xmath217 . in fig .
[ fig - r2l2 ] , simulated block error rates of network beamforming with optimal power control are compared to those of best - relay selection , larsson s scheme in @xcite with total relay power @xmath217 , distributed space - time coding in @xcite , and amplify - and - forward without power control ( every relay uses its maximal power ) in a 2-relay network . the information symbol @xmath10 is modulated as bpsk .
we can see that network beamforming with optimal power control outperforms all other schemes .
it is about 0.5db and 2db better than larsson s scheme and best - relay selection , respectively . with perfect channel knowledge at relays , it is @xmath234db better than alamouti distributed space - time coding , which needs no channel information at relays .
amplify - and - forward with no power control only achieves diversity 1 , distributed space - time coding achieves a diversity slightly less than two , while best - relay selection , network beamforming , and larsson s scheme achieve diversity 2 .
[ fig - r3l2 ] shows simulated performance of a 3-relay network under different schemes .
similar diversity results are obtained .
but for the 3-relay case , network beamforming is about 1.5db and 3.5db better than larsson s scheme and best - relay selection , respectively . in fig .
[ fig - r2l2-dp ] , we show performance of a 2-relay network in which @xmath235 and @xmath236 . that is , the transmitter and the first relay have the same power constraint while the second relay has only half the power of the first relay .
the channels are assumed to be rayleigh fading channels . in fig .
[ fig - r2l2-distance ] , we show performance of a 2-relay network whose channels have both fading and path - loss effects .
we assume that the distance between the first relay and the transmitter / receiver is 1 , while the distance between the second relay and the transmitter / receiver is 2 .
the path - loss exponent @xcite is assume to be 2 .
we also assume that the transmitter and relays have the same power constraint , i.e. , @xmath237 . in both cases ,
distributed space - time coding does not apply , and larsson s scheme applies for the second case only .
so , we compare network beamforming with best - relay selection and amplify - and - forward with no power control only .
performance of larsson s scheme is shown in fig .
[ fig - r2l2-distance ] as well .
both figures show the superiority of network beamforming to other schemes .
the previous section is on power control of relay networks with no dl between the transmitter and receiver . in this section , we discuss networks with a dl . as in @xcite , there are several scenarios , which we discuss separately . in this subsection , we consider relay networks with a dl during the first step only .
this happens when the receiver knows that the transmitter is in vicinity and listens during the first step , while the transmitter is not aware of the dl or is unwilling to do the optimization because of its power and delay constraints .
it can also happen when the transmitter is in the listening or sleeping mode during the second step .
in this case , @xmath33 . from ( [ x1 ] ) and ( [ x2 ] ) ,
the system equations can be written as @xmath238 = \left[\begin{array}{c } \alpha_0\sqrt{p_0}f_0 \\
\alpha_0\sqrt{p_0}\sum_{i=1}^r \frac{\alpha_i|f_ig_i|\sqrt{p_i}}{\sqrt{1+\alpha_0 ^ 2|f_i|^2p_0 } } \end{array}\right]s + \left[\begin{array}{c } w_1 \\ w_2+\sum_{i=1}^r
\frac{\alpha_i|g_i|\sqrt{p_i}}{\sqrt{1+\alpha_0 ^ 2|f_i|^2p_0}}e^{-j\arg f_i}v_i \end{array}\right].\ ] ] using maximum ratio combining , the ml decoding is @xmath239 the optimization problem is thus the maximization of the total receive snr of both transmission branches , which equals @xmath240 first , both terms in the snr formula increase as @xmath36 increases .
thus , @xmath37 , i.e. , the transmitter should use its maximum power .
the snr optimization problem becomes the one in section [ sec - pl - result ] , in which there is no dl .
therefore , the power control of networks with a dl during the first step only is exactly the same as that of networks without a dl .
this result is intuitive . since with a dl during the first step only , operations at both the transmitter and relays keep the same as networks without the dl .
the only difference is that the receiver obtains some extra information from the transmitter during the first step , and it can use the information to improve the performance without any extra cost . for the single - relay case , it can be proved easily that to maximize the receive snr , the relay should use its maximal power as well , that is , @xmath241 . in this subsection ,
we consider relay networks with a dl during the second step only .
this happens when the transmitter knows that the receiver is at vicinity and determines to do more optimization to allocate its power between the two transmission steps .
however , the receiver is unaware of the dl and is not listening during the first step .
it can also happen when the receiver is in transmitting or sleeping mode during the first step .
in this case , @xmath34 and @xmath242 is given in ( [ x2 ] ) .
the receive snr can be calculated to be @xmath243 first , we show that @xmath244 should take its maximal value 1 , i.e. , the transmitter should use all its power .
assume that @xmath245 is the optimal solution .
define @xmath246 .
we have @xmath247 . therefore , @xmath248 .
this contradicts the assumption that @xmath249 is optimal .
define @xmath250 the receiver snr can be calculated to be @xmath251 for any fixed @xmath36 , we can optimize @xmath252 following the analysis in section [ sec - pl - result ] .
the following theorem can be proved .
define @xmath253 for @xmath88 and @xmath254 . for any fixed @xmath255 , order @xmath256 as @xmath257 for @xmath258 , let @xmath259 and define @xmath260
is defined as @xmath261 the receive snr is maximized at @xmath262 , where @xmath263 is the smallest @xmath1 such that @xmath264 .
[ thm - main - dl-2nd ] the proof of this theorem follows the one of theorem [ thm - main ] and the lemmas it uses . as discussed in section [ sec - pl - result ] , for networks with no dl , there is no need to consider the solution of subproblem 0 .
here it is different .
define @xmath265 .
if we denote the solution of subproblem 0 , @xmath266,0_r\preceq \hat{\by}\preceq\hat{\ba } } \frac{\left(a+\langle \hat{\bc } , \hat{\by}\rangle\right)^2}{1+\|\hat{\by}\|^2}$ ] , as @xmath267 , because of the existence of the dl during the second step , it is possible that @xmath268 . now we discuss the optimization of @xmath36 .
we first consider the case of @xmath255 .
for any given @xmath269^t$ ] , the @xmath36 that maximizes the receive snr satisfies @xmath270 .
thus , the optimal @xmath36 can be found numerically by solving @xmath270 .
it can be proved easily that @xmath271 when @xmath272 and @xmath273 when @xmath274 .
thus , the maximum of @xmath275 is reached inside @xmath276 .
when the power at the transmitter is high ( @xmath277 ) , the receive snr can be approximated by @xmath278 where @xmath279 and @xmath280 .
it can be calculated straightforwardly that for @xmath255 , @xmath281.\ ] ] and @xmath282 this is a quartic equation of @xmath283 , whose solutions can be calculated analytically .
note that @xmath284 when @xmath272 and @xmath285 when @xmath274 .
thus the maximum of @xmath213 is reached inside @xmath276 .
an approximate solution of @xmath36 can thus be obtained analytically at high transmit powers .
now we consider the cases of @xmath286 and @xmath287 . if @xmath286 , the system degrades to a point - to - point one since only the dl works .
thus , the receive snr is @xmath288 . for @xmath287
, we can obtain the optimal @xmath289 using theorem [ thm - main - dl-2nd ] .
thus , we obtain three sets of @xmath290 and @xmath289 for the three cases : @xmath255 , @xmath286 , and @xmath287 , respectively .
the optimal solution of the system is the set of @xmath36 and @xmath289 corresponding to the largest receive snr .
the power control problem in networks with a dl during the second step only can thus be solved using the following recursive algorithm .
1 . initialization : set @xmath291 , the @xmath0-dimensional vector of all ones , @xmath292 , and @xmath293 . set the maximal number of iterations @xmath294 and the threshold @xmath295 .
2 . optimize @xmath36 with @xmath296 .
denote the solution as @xmath297 .
we can either do this numerically or calculate the high snr approximation .
3 . with @xmath298 ,
find the @xmath289 that maximizes the receive snr using theorem [ thm - main - dl-2nd ] .
denote it as @xmath299 .
calculate @xmath300 .
4 . set @xmath301 . if @xmath302 and @xmath303 , set @xmath304 , @xmath305 , and go to step 2 .
find the solution of @xmath289 with @xmath287 using theorem [ thm - main - dl-2nd ] .
denote this solution as @xmath306 .
the optimal solution is : @xmath307 .
[ alg - pl - dl-2nd ] similarly , the distributive strategies proposed in section [ sec - pl - discussion ] can be applied here . in this subsection
, we consider relay networks with a dl during both the first and the second steps .
this happens when both the transmitter and the receiver know that they are not too far away from each other and decide to communicate during both steps with the help of relays during the second step . from ( [ x1 ] ) and ( [ x2 ] )
, the system equation can be written as @xmath238 = \left[\begin{array}{c } \alpha_0\sqrt{p_0}f_0 \\
\beta_0|f_0|+\alpha_0\sqrt{p_0}\sum_{i=1}^r \frac{\alpha_i|f_ig_i|\sqrt{p_i}}{\sqrt{1+\alpha_0 ^ 2|f_i|^2p_0 } } \end{array}\right]s + \left[\begin{array}{c } w_1 \\ w_2+\sum_{i=1}^r \frac{\alpha_i|g_i|\sqrt{p_i}}{\sqrt{1+\alpha_0
^ 2|f_i|^2p_0}}e^{-j\arg f_i}v_i \end{array}\right].\ ] ] similar to the networks discussed in section [ sec - dl-1ststep ] , the maximum ratio combining results in the following ml decoding : @xmath308 the total receive snr of both transmission branches can be calculated to be @xmath309 the same as the networks in section [ sec - dl-2ndstep ] , @xmath244 should take its maximal value , which is 1 .
that is , @xmath310 .
similar to the snr optimization in section [ sec - dl-2ndstep ] , for any given @xmath255 , the snr maximization is the same as the maximization of @xmath275 , which is solved by theorem [ thm - main - dl-2nd ] .
but due to the difference in the receive snr formula , the optimal @xmath36 given @xmath252 is different .
it is the solution of @xmath311 .
when the dl exists during both steps , the case of @xmath286 , whose receive snr is @xmath288 will never outperform the case of @xmath287 , whose receive snr is @xmath312 for some @xmath289 .
thus , the case @xmath286 needs not to be considered .
the power control problem in networks with a dl during both steps can thus be solved using the following recursive algorithm . 1 .
initialization : set @xmath291 , @xmath292 , and @xmath293 . set the maximal number of iterations @xmath294 and the threshold @xmath295 .
2 . optimize @xmath36 with @xmath296 .
denote the solution as @xmath297 .
we can do this numerically .
3 . with @xmath298 ,
find the @xmath289 that maximizes @xmath275 using theorem [ thm - main - dl-2nd ] .
denote it as @xmath299 .
calculate @xmath313 .
4 . set @xmath301 . if @xmath302 and @xmath303 , set @xmath304 , @xmath305 and go to step 2 .
find the solution of @xmath289 with @xmath287 using theorem [ thm - main - dl-2nd ] .
denote this solution as @xmath306 .
the optimal solution is : @xmath314 .
[ alg - pl - dl - both ] again , the distributive strategies proposed in section [ sec - pl - discussion ] can be applied here . in this subsection
, we compare single - relay networks in which the power constraints at the transmitter and the relay are same , i.e. , @xmath235 .
the channels are assumed to have both the fading and path - loss effect .
there are four cases : no dl , a dl during the first step only , a dl during the second step only , and a dl during both steps . in fig .
[ fig - triangle - network ] , we compare networks in which the distance of every link is the same , i.e. , the three nodes are vertexes of an equilateral triangle with unit - length edges as shown in fig .
[ fig - equallacteral - network ] .
we can see that the network with no dl has diversity 1 while networks with a dl and power control achieve diversity 2 .
the network with a dl during the first step performs less than 0.5db better than the network with a dl during the second step only , while the network with a dl during both steps performs the best ( about 1db better than the network with a dl during the first step only ) . to illuminate the effect of power control , we show performance of networks whose transmit power at the relay and transmitter are fixed . for the network with a dl during the first step only
, there is no power control problem since it is optimal for both the transmitter and the relay to use their maximal powers . for the other two cases , we let the transmitter uses half of its power , @xmath315 , to each of the two steps and the relay always uses its maximum power @xmath217 .
we can see that , if the dl only exists during the second step , without power control , the achievable diversity is 1 . at block error rates of @xmath316 and @xmath317
, it performs 3 and 6db worse , respectively . for networks with a dl during both steps , power control results in a 1.5db improvement . in fig .
[ fig - line - network - e2 ] and [ fig - line - network - e3 ] , we show performance of line networks with path - loss exponents 2 and 3 respectively . as shown in fig .
[ fig - line - network ] , the three nodes are on a line and the relay is in the middle of the transmitter and receiver .
the distance between the transmitter and receiver is assumed to be 2 .
the same phenomenon as in the equilateral triangle networks can be observed .
the network with a dl during both steps performs the best ( about 1db better than the network with a dl during the first step only ) .
the network with a dl at first step only performs slightly better than the one with a dl during the second step only .
but the difference is smaller than that in fig .
[ fig - triangle - network ] .
the performance difference between line networks with and without dls is smaller than those in equilateral triangle networks , and it gets even smaller for larger path - loss exponents .
this is because as the distance between the transmitter and receiver or the path - loss exponent is larger , the quality of the dl is lower .
therefore , the improvement due to this link is smaller . for both cases , power control results in a 1.5db improvement when the dl link exists for both steps and a higher diversity when the dl exists for the second step only .
then we work on the random network in fig .
[ fig - random - network ] , in which the relay locates randomly and uniformly within a circle in the middle of the transmitter and the receiver .
the distance between the transmitter and the receiver is assumed to be 2 .
the radius of the circle is denoted as @xmath318 .
we assume that @xmath319 .
this is a reasonable model for ad hoc wireless networks since if communications between two nodes is allowed to be helped by one other relay , one should choose a relay that is around the middle of the two nodes .
in other words , the distance between the relay and the transmitter or receiver should be shorter than that between the transmitter and receiver .
we work out the geometry first . as in fig .
[ fig - random - network ] , we denote the positions of the transmitter , the receiver , the relay , and the middle point of the transmitter and the receiver as @xmath320 , and @xmath321 , respectively .
denote the angle of @xmath322 and @xmath323 as @xmath324 and the length of @xmath323 as @xmath325 .
the lengths of @xmath326 and @xmath327 are thus @xmath328 and @xmath329 .
since @xmath330 is uniformly distributed within the circle , @xmath324 is uniform in @xmath331 and the pdf and cdf of @xmath325 can be calculated to be @xmath332 respectively .
define @xmath333 .
if @xmath334 is uniform on @xmath276 , it can be proved that @xmath335 thus , @xmath336 has the same distribution as @xmath325 .
therefore , we generate @xmath336 to represent @xmath325 . fig .
[ fig - random - network - p ] shows performance of random networks with path - loss exponent 2 and @xmath337 .
we can see that the same phenomenon as in line networks can be observed . with a dl at both steps
, the random network performs about 1db worse than the line network .
in this paper , we propose the novel idea of beamforming in wireless relay networks to achieve both diversity and array gain .
the scheme is based on a two - step amplify - and - forward protocol .
we assume that each relay knows its own channels perfectly .
unlike previous works in network diversity , the scheme developed here uses not only the channels phase information but also their magnitude .
match filters are applied at the transmitter and relays during the second step to cancel the channel phase effect and thus form a coherent beam at the receiver , in the mean while , optimal power control is performed based on the channel magnitude to decide the power used at the transmitter and relays .
the power control problem for networks with any numbers of relays and no direct link is solved analytically .
the solution can be obtained with a complexity that is linear in the number of relays .
the power used at a relay depends on not only its own channels nonlinearly but also all other channels in the network . in general , it is not even a differentiable function of channel coefficients .
simulation with rayleigh fading and path - loss channels show that network beamforming achieves the maximum diversity while amplify - and - forward without power control achieves diversity 1 only .
network beamforming also outperforms other cooperative strategies .
for example , it is about 4db better than best - relay selection .
relay networks with a direct link between the transmitter and receiver are also considered in this paper . for networks with a direct link during the first step only , the power control at relays and the transmitter is exactly the same as that of networks with no direct link . for networks with a direct link during the second step only and networks with a direct link during both steps , the solutions are different .
recursive numerical algorithms for the power control at both the transmitter and relays are given .
simulated performance of single - relay networks with different topologies shows that optimal power control results in about 1.5db improvement in networks with a direct link at both steps and a higher diversity in networks with a direct link at the second step only . we have just scratched the surface of a brand - new area .
there are a lot of ways to extend and generalize this work .
first , it is assumed in this work that relays and sometimes the transmitter know their channels perfectly , which is not practical in many networks .
network beamforming with limited and delayed feedback from the receiver is an important issue . in multiple - antenna systems , beamforming with limited and delayed channel information feedback has been widely probed . however , beamforming in networks differs from beamforming in multiple - antenna systems in a couple of ways . in networks , it is difficult for relays to cooperate while in a multiple - antenna system , different antennas of the transmitter can cooperate fully .
there are two transmission steps in relay networks while only one in multiple - antenna systems , which leads to different error rate and capacity calculation and thus different designs .
second , the relay network probed in this paper has only one pair of transmitter and receiver . when there are multiple transmitter - and - receiver pairs , an interesting problem is how relays should allocate their powers to aid different communication tasks .
finally , the two - hop protocol can be generalized as well . for a given network topology ,
one relevant question is how many hops should be taken to optimize the criterion at consideration , for example , error rate or capacity .
y. chang and y. hua , `` application of space - time linear block codes to parallel wireless relays in mobile ad hoc networks , '' in _ prof . of the 36th asilomar conference on signals , systems and computers _ , nov . 2003 .
y. hua , y. mei , and y. chang , `` wireless antennas - making wireless communications perform like wireline communications , '' in _ prof . of ieee
ap - s topical conference on wireless communication technology _ , oct .
y. tang and m. c. valenti , `` coded transmit macrodiversity : block space - time codes over distributed antennas , '' in _ prof . of ieee vehicular technology conference 2001-spring _ , vol . 2 , pp .
1435 1438 , may 2001 .
a. sendonaris , e. erkip , and b. aazhang , `` user cooperation diversity - part ii : implementation aspects and performance analysis , '' _ ieee transactions on communications _ , vol .
51 , pp . 19391948 , nov .
r. u. nabar , h. b@xmath338lcskei , and f. w. kneubuhler , `` fading relay channels : performance limits and space - time signal design , '' _ ieee journal on selected areas in communications _ , pp .
1099 1109 , aug . 2004 .
j. n. laneman and g. w. wornell , `` distributed space - time - coded protocols for exploiting cooperative diversity in wireless network , '' _ ieee transactions on information theory _
49 , pp . 2415 - 2425 , oct . 2003 .
m. janani , a. hedayat , t. e. hunter , and a. nosratinia , `` coded cooperation in wireless communications : space - time transmission and iterative decoding , '' _ ieee transactions on signal processing _ , pp .
362 - 371 , feb . 2006 .
k. azarian , h. e. gamal , and p. schniter , `` on the achievable diversity - multiplexing tradeoff in half - duplex cooperative channels , '' _ ieee transactions on information theory _ , vol .
51 , pp .
4152 - 4172 , dec . 2005 .
j. n. laneman , d. n. c. tse , and g. w. wornell , `` cooperative diversity in wireless networks : efficient protocols and outage behavior , '' _ ieee transactions on information theory _ , pp .
3062 - 3080 , dec . 2004 .
y. cao , b. vojcic , and m. souryal , `` user - cooperative transmission with channel feedback in slow fading environment , '' in _ proc . of ieee
vehicular technology conference 2004-fall _ , pp . 2063 - 2067 , 2004 .
p. a. anghel and m. kaveh ,
`` on the performance of distributed space - time coding systems with one and two non - regenerative relays , '' _ ieee transactions on wireless communications _
, pp . 682 - 692 , mar .
2006 .
a. narula , m. l. lopez , m. d. trott , and g. wornell , `` efficient use of side information in multiple - antenna data transmission over fading channels , '' _ ieee journals on selected areas in communications _ , pp .
1423 - 1436 , oct . 1998 .
s. a. jafar and a. j. goldsmith , `` transmit optimization and optimality of beamforming for multiple antenna systems with imperfect feedback , '' _ ieee transactions on wireless communications _
, pp . 1165 - 1175 , july 2004 . k. k. mukkavilli , a. sabharwal , e. erkip , and b. aazhang , `` on beamforming with finite rate feedback in multiple - antenna systems , '' _ ieee transactions on information theory _ , pp .
2562- 2579 , oct .
2003 .
s. ekbatani and h. jafarkhani , `` combining beamforming and space - time coding for multi - antenna transmitters using noisy quantized channel direction feedback , '' _ submitted to ieee transactions on communications _ , 2006 . | this paper is on beamforming in wireless relay networks with perfect channel information at relays , the receiver , and the transmitter if there is a direct link between the transmitter and receiver .
it is assumed that every node in the network has its own power constraint .
a two - step amplify - and - forward protocol is used , in which the transmitter and relays not only use match filters to form a beam at the receiver but also adaptively adjust their transmit powers according to the channel strength information . for a network with any number of relays and no direct link , the optimal power control is solved analytically .
the complexity of finding the exact solution is linear in the number of relays .
our results show that the transmitter should always use its maximal power and the optimal power used at a relay is not a binary function .
it can take any value between zero and its maximum transmit power .
also , surprisingly , this value depends on the quality of all other channels in addition to the relay s own channels . despite this coupling fact
, distributive strategies are proposed in which , with the aid of a low - rate broadcast from the receiver , a relay needs only its own channel information to implement the optimal power control .
simulated performance shows that network beamforming achieves the maximal diversity and outperforms other existing schemes .
then , beamforming in networks with a direct link are considered .
we show that when the direct link exists during the first step only , the optimal power control at the transmitter and relays is the same as that of networks with no direct link . for networks with a direct link during the second step only and both steps , recursive numerical algorithms are proposed to solve the power control problem .
simulation shows that by adjusting the transmitter and relays powers adaptively , network performance is significantly improved . |
in this supplementary material we provide the details of the model realization , and the proof of the generic theorem to determine the topology , i.e. the chern number of a chiral superfluid / superconductor .
generate a blue - detuned square lattice .
the incident polarization of @xmath134 is @xmath135 , and the @xmath4-wave plate changes the polarization of the reflected field to @xmath136 .
( b , c ) as illustrated for @xmath137 fermions , the lattice potential for the state @xmath138 is generated by @xmath139 .
( d ) the raman coupling is generated by @xmath134 and @xmath140 , which has a tilt angle @xmath8 with respect to @xmath9-axis in the @xmath9-@xmath141 plane . ]
the spin-1/2 pseudospin is defined by the two atomic states @xmath142 and @xmath143 of the atom @xmath137 .
they are then coupled to the excited states in the manifold @xmath144 and @xmath145 through various two - photon processes depicted in fig .
[ realization](b - d ) .
+ the electric field of standing - wave lights for the present realization can be written as : @xmath146 where @xmath147 , @xmath148 . in the last line of the above equation we have made the change of parameters @xmath149 and @xmath150 and a multiplication of an overall phase factor @xmath151 .
the rabi - frequencies of the transitions described in fig .
[ realization](b - d ) can be derived .
the light denoted by black line has the rabi - frequency @xmath152 .
those lights depicted as red lines and connecting existing states with @xmath153 have the rabi - frequency @xmath154 , and the one depicted as blue line is @xmath155 .
the phase @xmath156 are irrelevant and so are omitted here .
the quantities @xmath157 is the dipole matrix element @xmath158 , where @xmath159 and @xmath160 is the corresponding ground state and excited state , and @xmath161 is the spherical component of vector @xmath162 .
@xmath163 @xmath164 the lattice potential is generated by the two - photon processes which intermediate states are in the manifold @xmath165 ( described in the diagram fig .
[ realization](b , c ) by the process generated by black and red lines ) .
each process gives a contribution @xmath166 to the lattice potential , where @xmath167 is the corresponding rabi - frequency , @xmath76 runs through all possible atomic states and @xmath168 or @xmath169 if the intermediate state considered is in the manifold @xmath145 or @xmath144 . according to experimental data , for reference ,
see @xcite ) all different @xmath170 or @xmath169 corresponding to intermediate states of different @xmath171 are of the same order of magnitude ( order of thz ) and the differences among each group are negligible ( order of 100mhz to 1ghz ) .
the lattice potential of both spin is @xmath172 where we have omitted the constant term , and @xmath173 ( @xmath174 ) is the reduced matrix element between the total angular momentum @xmath175 and @xmath176 .
( the coefficients in @xmath177 can be calculated with the help of table [ dipolet ] in a straightforward way . )
notice that the coupling between spin - up @xmath178 and spin - down @xmath179 state through these processes are negligible since @xmath180 , where @xmath181 is the energy of the spin - up or spin - down state .
the raman lattice is generated via only one raman process ( described in the diagram fig .
[ realization](d ) ) .
the raman potential generated is given by the formula @xmath182 .
in this case , the generated potential is @xmath183 where @xmath184 .
the zeeman term @xmath185 is generated by the a small off - resonant in the raman process , where @xmath186 .
we derive the tight - binding model by considering the hopping contributed from lattice and raman potentials , respectively .
the lattice potential contributes the spin - conserved hopping terms as @xmath187 where @xmath188 , @xmath189 and @xmath190 , and @xmath191 is the wavefunction at centered at @xmath192 . on the other hand , for the spin - flip term induced by raman coupling , we have that the first term of provides spin - flip hopping along the @xmath10-direction of strength @xmath193 where @xmath194 .
the same term gives no contribution to the hopping along @xmath9-direction since @xmath195 is antisymmetric in the @xmath9-direction at the local minimum of lattice potential .
likewise , the second term of ( [ ramanf ] ) gives rise to an onsite zeeman term @xmath196 where @xmath197 .
this term gives negligible contribution to the hopping term since @xmath198 .
therefore , in the tight - binding model , the raman potential contributes @xmath199+h.c.\end{aligned}\ ] ] the total tight - binding hamiltonian reads @xmath200 , which can be simplified by applying the gauge transformation @xmath201 with the fourier transformation @xmath202 , we can finally get the bloch hamiltonian of the tight - binding model @xmath203 where @xmath204 and @xmath205 .
all the parameters are independently tunable except that @xmath206 and @xmath207 are related by @xmath208 .
it can be seen that the inversion symmetry is controlled by the tilt angle @xmath8 , and the gap opening at dirac points is controlled by the @xmath35 , the @xmath38-component of the @xmath2 field .
note that @xmath24 is induced by onsite spin - flip transition .
thus a small @xmath35-term in eq .
can generate a relatively large @xmath24 . for our purpose
, we shall consider a small @xmath35-term compared with @xmath34-term .
thus the @xmath2 field is mainly polarized in the @xmath36 direction .
an attractive hubbard interaction can be described effectively as @xmath209 .
we introduce three order parameters @xmath210 and @xmath211 when considering superconducting pairing , where @xmath212 and @xmath70 or @xmath71 .
if only one of them is non - zero , the bdg hamiltonian can be written as @xmath213 , where @xmath214 and the basis of @xmath215 is @xmath216 .
the pairing matrix @xmath217 is assumed to be real valued , which can be set via a change of overall phase factor of @xmath218 . here
, @xmath73 is the center - of - mass momentum of the non - zero pairing .
if there are more than one of them is non - zero , we need to fold the brillouin zone .
let @xmath219 , where @xmath220 and @xmath221 are coprime integers .
then the density of the bdg hamiltonian can be written as @xmath222 where @xmath223 , @xmath224 , @xmath225 and @xmath226 . the nambu basis in the folded brillouin zone , where @xmath227 $ ] , is @xmath228 in this basis
, the gap equation can be derived through diagonalizing ( [ bdgnotfolded ] ) or ( [ bdgfolded ] ) and then calculate the expectation values of the order parameters . therefore , from a particular value of @xmath229 the order parameters
can be solved self - consistently .
the bkt temperature can be calculated through the approximate bkt criterion @xmath230 where @xmath231 is the superfluid density , which is can be calculated by @xmath232 .
@xmath233 is the supercurrent density and @xmath234 is the center - of - mass momentum of two paired fermions when the system cut only one fs , which is of center - of - mass momentum @xmath234 .
@xmath233 is to be calculated by the variation of the zero - temperature energy , i.e. @xmath235 . the total energy @xmath236 can be calculated through diagonalization of the total hamiltonian in the nambu basis , by assuming the center - of - mass momentum of the pairing to be @xmath237 .
the numerical result is shown in fig .
[ fig : tbkt ] . ]
after presenting the generic result , we shall show step by step the generic theory that the topology of 2d chiral superfluid / superconductor can be precisely determined by knowing only the properties of the fermi surfaces and the topology of other normal bands away from the fermi surfaces . for simplicity , we do not consider the case that the normal bands barely touch the fermi energy .
we classify the normal bands into three groups : upper ( lower ) bands , which are those with energy above ( below ) fermi energy , and middle bands which are crossed by fermi energy .
let @xmath238 ( @xmath78 ) be the total chern number of the upper ( lower ) band and @xmath81 be the chern number of @xmath77-th middle band .
each middle band may form multiples fermi surfaces .
we denote the pairing of the @xmath239-th fermi surface , projected on @xmath77-th band , to be @xmath240 and its phase @xmath241 .
we denote @xmath242 to be the area enclosed by the fermi surface @xmath243 and @xmath244 to be the unenclosed area .
notice that among each of the regions @xmath242 or @xmath244 , the energy of the @xmath77-th band is of the same sign .
we shall show that in general , the formula for the chern number of the negative energy states , after considering the superconducting pairing , can be written as @xmath245 where the integral in the bracket is the winding number of @xmath246 .
the integral is evaluated in the direction satisfying right hand rule .
the phase factor @xmath247 and @xmath248 is @xmath88 if the energy of the @xmath77-th band is negative in the region @xmath242 and @xmath244 respectively , and they are @xmath249 if the corresponding energy is positive .
we have chosen a gauge here such that the phase of the berry connection of the @xmath77-th band is continuous throughout the regions @xmath242 . the case with an alternative gauge can be found below .
this quantity is non - zero if the system is topological .
first , we consider the case where there is only one band in the system and there is only one fs . so the berry curvature , and thus the chern number , of the normal states is zero since we can choose a gauge of real eigenvectors where the berry connection is identically zero .
+ we can calculate the berry connection , after considering the superconducting pairing , by direct calculation over each region @xmath250 . the density of the bdg hamiltonian and the berry connection can be written as @xmath251 where the @xmath252 sign denotes upper ( lower ) band .
@xmath34 and @xmath35 can be found by diagonalizing the @xmath253 : @xmath254
so the berry connection of the lower band is : @xmath255 since the chern number remains unchanged under all continuous changes , we can assume that @xmath256 and let @xmath257 .
we can also continuously deform the system so that the fs is symmetric . in this case
, the berry curvature is @xmath258 where @xmath259 denotes a function which valued to 1 inside ( outside ) the region @xmath250 and 0 outside ( inside ) , and the upper ( lower ) sign means that the region @xmath250 is of positive ( negative ) energy .
the chern number can be calculated through an integral of @xmath260 @xmath261 where @xmath262 and that the berry curvature localizes on the boundaries of @xmath250 . to generalize the equation ( [ sichern1 ] )
, we notice that when @xmath257 , the momentum modulation is unimportant .
thus the chern number of the lower band depends only on the pairing within each fs . therefore , the total chern number is the summation of @xmath263 , where @xmath264 is the winding number of @xmath265 , and @xmath266 is the pairing order of the @xmath76-th fs ; when the energy of the normal states is positive ( negative ) within the region @xmath267 , we take the negative ( positive ) sign in @xmath263 .
it is worthwhile to note that in general there maybe order parameters connecting two different fs ( like the bcs order @xmath211 in the main text ) . to simplify our discussion
, we assume that the system contains minimal order parameters to be fully gapped . without lost of generality , we can assume that the fs looks like fig .
[ gents](b ) .
( if we have @xmath268 inside @xmath250 , we can just flip the sign as indicated by ( [ sichern1 ] ) . )
folding up the brillouin zone , we will find that two fs coincides with each other and that @xmath269 is the subspace of the bdg hamiltonian in the folded brillouin zone , where @xmath270 is the two center - of - mass momentum of the fs and @xmath271 is defined from @xmath272 to @xmath273 , where @xmath221 is an integer number .
the minimal order parameters to fully gap the system can be @xmath274 , which we have discussed , or only @xmath211 alone .
when only @xmath211 is non - zero , the winding numbers are calculated for @xmath211 around @xmath74 and @xmath234 ( as indicated by the diagonal term in the @xmath170 matrix ) .
if there are many bands in the system , we call the chern number of the upper ( lower ) band be @xmath275 . these bands
contributes @xmath276 to the total chern number for negative energy states in the duplicated space .
the arguments in the one band system need generalization since the berry curvature of the normal bands are non - zero and an intuitive generalization of @xmath277 to @xmath278 , where @xmath279 is the eigenvector of the single particle hamiltonian , is now gauge dependent .
the single particle hamiltonian density @xmath280 have eigenvectors @xmath281 so that @xmath282 and the bdg hamiltonian density can be written as @xmath283 write the bdg hamiltonian density in the bases @xmath284 and @xmath285 , @xmath286 has diagonal entries with values @xmath287 and @xmath288 and the off - diagonal entries due to @xmath289 , i.e. @xmath290 .
let @xmath256 and @xmath257 , the upper and lower bands are diagonalized automatically .
the hamiltonian in the subspace of middle bands can also be diagonalized as follows @xmath291 where @xmath292 .
+ therefore , when the @xmath77-th middle band is duplicated and gapped by @xmath289 , the berry connection of the upper and lower band are @xmath293 there are three contributions when we calculate the berry curvature , which is the curl of this quantity : ( i ) the curvature from the normal band ; ( ii ) the extra curvature due to the mixing of two bands ( from the derivative of @xmath294 ) ; ( iii ) the curvature due to superconducting pairing .
after continuously deforming the fs so that it is symmetric , we can write the berry curvature of the lower band to be @xmath295 where the upper ( lower ) sign is taken when the @xmath77-th band has positive ( negative ) energy in the region @xmath250 , @xmath296 denotes a function which valued to 1 inside ( outside ) @xmath250 and 0 outside ( inside ) @xmath297 and @xmath298 ( @xmath299 ) denotes the berry connection ( curvature ) of the corresponding normal band .
+ the integral of berry curvature can be simplified to ( writing @xmath300 ) : @xmath301 notice that the third and fourth terms are gauge dependent , and they all add up to a gauge independent value .
we now specify the gauge to finish the calculation .
we require that : @xmath302 we call this gauge @xmath250 and for the corresponding eigenfunctions we call them @xmath303}^{(i_m)}$ ] .
then the phase of @xmath277 can be defined unambiguously .
( i.e. we define the matrix element in @xmath250 gauge : @xmath304}^{(i , j ) } = \left < u_{k[\vec{\mathcal{s}}]}^{(i)}\right| \delta(k ) \left| u_{-k[\vec{\mathcal{s}}]}^{(j)*}\right>$ ] ) .
the gauge merely means that the eigenvectors @xmath305 are continuous within the region @xmath250 .
if we have other choice of gauge , we count the number of singularity ( with the sign determined by right hand rule ) within @xmath250 .
each singularity contributes to @xmath306 extra winding of @xmath307 and is canceled by the subtraction of integral of berry connection from @xmath308 .
therefore , we know that the winding in different gauge can be related as @xmath309}^{(i_m ) } \cdot dk= \int_{\partial \vec{\mathcal{s } } } \nabla_k \theta_{k [ \mathcal{g}]}^{(i_m ) } \cdot dk + 2n_{s,\vec{\mathcal{s}}[\mathcal{g}]},\ ] ] where @xmath310}$ ] is the number of singularity inside @xmath250 in gauge @xmath311
. therefore total chern number of the negative energy states is @xmath312}^{(i_m ) } \cdot dk \right ) = n_l - n_u \pm \left ( n_f^{(i_m ) } - \int_{\partial \vec{\mathcal{s } } } \nabla_k \theta_{k[\mathcal{g}]}^{(i_m ) } \cdot dk - 2n_{s,\vec{\mathcal{s}}[\mathcal{g } ] } \right ) \label{mochern},\ ] ] for a general gauge @xmath311 .
note that each particular middle band @xmath77 can contribute to the chern number by @xmath313}^{(i_m , j ) } + 2n_{s,\vec{\mathcal{s}}_{i_m , j}[\mathcal{g } ] } \right),\ ] ] where the phase factor @xmath247 and @xmath248 is @xmath88 if the energy of the band is negative in the region @xmath242 and @xmath244 respectively , and they are @xmath249 if the corresponding energy is positive .
it is worth noting that as in this case , one more complexity arises . as in ( [ fbzchern ] ) , we consider order parameter that connect two fs : @xmath315 if @xmath316 and @xmath317 are of the same sign in the coinciding region in the folded brillouin zone , the argument in [ sec : omfs ] have no difficulties except the similar generalization to the solution of the problem of allotment of chern number and the problem of gauge .
if @xmath316 and @xmath317 are of opposite sign , the minimal order parameter set can not be @xmath211 alone ( see fig .
[ gents](c ) for illustration ) .
this is just the case for the model discussed in the main text , where @xmath211 alone can not gap the system .
we state without proof here that , if only @xmath211 and one of the @xmath274 is non - zero , the system can still be fully gap . without lost of generality ,
suppose @xmath318 is non - zero , the @xmath319 band connect with its hole partner directly and @xmath320 band connect through two intermediate steps . the chern number of this kind of interaction can also be calculated .
the direct pairing is straightforward , while for the indirect pairing ( the one containing two intermediate steps ) , we can count the winding number of the effective @xmath170 . in this case , the effective pairing between @xmath321 and @xmath316 are proportional to the multiplication of @xmath322 , @xmath323 and @xmath324 , therefore the winding of the effective @xmath170 is the sum of the winding of these three quantities . the problem concerning gauge and the allotment of the original chern number
can also be similarly investigated . | the recent realization of two - dimensional ( 2d ) synthetic spin - orbit ( so ) coupling opens a broad avenue to study novel topological states for ultracold atoms . here , we propose a new scheme to realize exotic chiral fulde - ferrell superfluid for ultracold fermions , with a generic theory being shown that the topology of superfluid pairing phases can be determined from the normal states .
the main findings are two fold .
first , a semimetal is driven by a new type of 2d so coupling whose realization is even simpler than the recent experiment , and can be tuned into massive dirac fermion phases with or without inversion symmetry . without inversion symmetry
the superfluid phase with nonzero pairing momentum is favored under an attractive interaction .
furthermore , we show a fundamental theorem that the topology of a 2d chiral superfluid can be uniquely determined from the unpaired normal states , with which the topological chiral fulde - ferrell superfluid with a broad topological region is predicted for the present system . this generic theorem is also useful for condensed matter physics and material science in search for new topological superconductors . _
introduction._the recent experimental realization of two - dimensional ( 2d ) spin - orbit ( so ) coupling for ultracold atoms @xcite , which corresponds to synthetic non - abelian gauge potentials @xcite , advances an essential step to explore novel topological quantum phases beyond natural conditions .
ultracold fermions with so coupling can favor the realization of topological superfluids ( tsfs ) ( similar as topological superconductors in solids @xcite ) based on an @xmath0-wave feshbach resonance @xcite , which are highly - sought - after quantum phases for their ability to host non - abelian majorana zero modes and implement topological quantum computation @xcite . note that a superfluid phase has to exist in 2d or 3d regime , so having a 2d or 3d so coupling is the basic requirement for such realization of gapped tsfs . while experimental studies of tsfs are yet to be available , different proposals have been introduced for rashba and dirac type so coupled systems @xcite , with bcs or fflo pairing orders @xcite .
when the fermi energy crosses only a single ( or odd number of ) fermi surface ( fs ) , the so coupling forces cooper pairs into effective @xmath1-wave type , rendering a tsf phase @xcite .
however , in the generic case it is not known so far whether there is a universal way to precisely determine the topology , e.g. chern numbers , of a superfluid phase induced in normal bands . on the other hand , to achieve a tsf phase integrates several essential ingredients , which may bring challenges for the experiment .
the minimal schemes of realization are therefore desired to ensure high feasibility of real studies .
in this letter we propose an experimental scheme to realize chiral fulde - ferrell ( ff ) superfluids , and show a generic theory to compute the chern number of tsf phases through the normal states .
the findings are of significance in both experiment and fundamental theory .
first , we propose that a new type of 2d so coupled dirac semimetal can be realized based on a scheme adopting only a single raman transition and simpler than the recent experiments @xcite . by readily breaking the inversion symmetry of the dirac system , we find that the ff superfluid phase can be favored under an attractive interaction .
moreover , we show a generic formalism for the chern number of a 2d superfluid induced in the normal states , with which the topological chiral ff superfluid with a broad topological region is predicted . generate a blue - detuned square lattice .
the incident polarization of @xmath2 is @xmath3 , and the @xmath4-wave plate changes the polarization of the reflected field to @xmath5 .
the raman coupling , illustrated in ( c ) for @xmath6k fermions , is generated by @xmath2 and an additional running light @xmath7 which has tile angle @xmath8 with respect to @xmath9-@xmath10 plane . ] _ 2d dirac semimetal._we start with the effective tight - binding hamiltonian @xmath11 on a square lattice , whose realization shall be presented below : @xmath12 where @xmath13 ( @xmath14 ) is annihilation ( creation ) operator with spin @xmath15 , @xmath16 is the hoping constant along @xmath17 direction , @xmath18 is the strength of spin - flip hopping , and @xmath19 denote the effective zeeman couplings . as described in fig .
[ fig : realization](a ) , the above hamiltonian describes a dirac semimetal if @xmath20 and @xmath21 , with two dirac points at @xmath22,0\right)$ ] .
the term @xmath23 breaks the inversion symmetry and leads to an energy difference between the two dirac points .
finally , a nonzero @xmath24 opens a local gap at the two dirac points . without loss of generality
, we take that @xmath25 and @xmath26 to facilitate the further discussion .
as we show below , the above hamiltonian can be readily realized in experiment .
the realization is sketched in fig .
[ fig : realization](b , c ) .
the ingredients of realization include a blue - detuned spin - independent square lattice and a raman lattice generated via only a single raman transition ( details can be found in the supplementary material @xcite ) . in particular ,
the two standing - wave lights ( red and black lines in fig . [ fig : realization](b ) ) of frequency @xmath27 form the electric field components @xmath28 and @xmath29 , and generate the square lattice potentials @xmath30 and @xmath31 , respectively , with @xmath32 and @xmath33 . for our purpose we take that @xmath34 is large compared with @xmath35 , namely , the field @xmath2 is mainly polarized in the @xmath36 direction . the @xmath37-phase difference between @xmath36 and @xmath38 polarized components is easily achieved by putting a @xmath4-wave plate before mirror @xmath39 , as shown in fig .
[ fig : realization](b ) .
all the initial phases of the lights are irrelevant and have been neglected . to generate the raman lattice another running light of frequency @xmath40 is applied along a tilted direction so that its wave vector @xmath41 ( blue line in fig .
[ fig : realization ] ( b ) ) .
@xmath42 together with the light @xmath2 induces the raman transition between spin - up @xmath43 and spin - down @xmath44 , as illustrated in fig .
[ fig : realization](c ) .
the generated raman potential takes the form @xmath45@xmath46@xmath47@xmath46 , with @xmath48 and @xmath49 .
the former @xmath34-term in @xmath50 leads to a spin - flip hopping along @xmath10 direction @xmath51 , and the latter @xmath35-term gives an onsite zeeman term @xmath52 , with @xmath53 @xcite .
note that a small @xmath35-term can induce a relatively large @xmath24 compared with hopping couplings , since @xmath24 is induced by on - site coupling . through a gauge transformation @xmath54
@xcite , we finally get the effective hamiltonian , with the parameters being @xmath55 $ ] and @xmath56 $ ] .
it can be seen that the inversion symmetry is controlled by the tilt angle @xmath8 , and the gap opening at dirac points is controlled by the @xmath35 , the @xmath38-component of the @xmath2 field .
the above realization is clearly simpler than the recent experiments on 2d so coupling , for it involves only one raman transition .
we note that a 2d dirac semimetal driven by so interaction , being different from graphene whose dirac points are protected by symmetry only when there is no so coupling , has not been discovered in solid state materials @xcite .
the high feasibility and controllability ensure that the 2d dirac semimetal proposed here can be well observed based on the current experiment . versus chemical potential for the band structures ( c ) and ( d ) at @xmath57 . in ( b )
an insulating regime can be reached .
the parameters are rescaled by taking @xmath58 , and we take @xmath59 and @xmath60 , where @xmath61 for ( a ) and ( c ) , and @xmath62 for ( b ) and ( d ) . ]
_ anomalous hall effect._when both the inversion and time - reversal symmetries are broken , the hamiltonian leads to anomalous hall effect , which reflects the nontriviality of 2d so interaction @xcite . fig .
[ fig : ahe ] shows the hall conductance @xmath63 versus chemical potential @xmath64 and the spectra in different regimes .
the hall conductance at zero temperature is calculated by ( @xmath65 is the step function ) @xmath66 .
\label{hallcon}\end{aligned}\ ] ] it can be seen that @xmath63 is always zero at @xmath67 or when @xmath64 is in the band gap [ fig .
[ fig : ahe](b , d ) ] . however , the nonzero @xmath63 for finite @xmath64 crossing bulk bands shows the nontrivial 2d so effect , implying that the superfluid phase of nontrivial topology may be obtained .
_ states of superfluidity._ the superfluid phase can be induced by considering an attractive hubbard interaction . the total hamiltonian @xmath68 .
due to the existence of multiple fss corresponding to different dirac cones , in general we can have intra - cone pairing ( fflo ) and inter - cone ( bcs ) pairing orders , defined by @xmath69 , with @xmath70 or @xmath71 @xcite .
on the other hand , since the inversion symmetry is broken , the bcs pairing would be typically suppressed .
the topology of the superfluid phase can be characterized by the chern number . however , to compute the chern number of the present system is highly nontrivial .
the reason is because the fflo order breaks translational symmetry can fold up the original brillouin zone into many sub - brillouin zones . in particular , if @xmath72 , with @xmath1 and @xmath73 being mutual prime integers , the system includes @xmath73 sub - brillouin zones .
moreover , if @xmath74 is incommensurate with original lattice , the system folds up into infinite number of sub - brillouiin zones .
in such generic way , it is not realistic to compute numerically the chern number of the system .
_ generic theory for chern number._to resolve the difficulty , we show a generic theorem to determine the chern number of a 2d system after opening a gap through having superfluid / superconductivity . for convenience
we classify the normal bands of the system without pairing into three groups : upper ( lower ) bands which are above ( below ) the fermi energy , and the middle bands which are crossed by fermi energy .
each middle band may have multiple fss ( loops ) , and we denote by @xmath75 the @xmath76-th fs loop of the @xmath77-th middle band .
let the total chern number of the upper ( lower ) bands be @xmath78 ( @xmath79 ) .
we can show that the chern number the superfluid pairing phase induced in the system is given by @xmath80 .
\label{eqn : generalgent}\end{aligned}\ ] ] here @xmath81 is the chern number of the @xmath77-th middle band and @xmath82 $ ] is the phase of the pairing order projected onto the @xmath75-th fs loop @xcite .
the integral direction is specified by arrows along fs lines in fig .
[ fig : gent ] ( a , b ) , which defines the boundary @xmath83 of the vector area @xmath84 in @xmath85 space .
the quantities @xmath86 are then determined by right - hand rule " specified below . the quantity @xmath87 ( or @xmath88 ) if the energy of normal states within the area @xmath89 is positive ( or negative ) , while @xmath90 ( or @xmath71 ) if the regions ( @xmath91 ) unenclosed by any fs loop have positive ( or negative ) energy ( fig .
[ fig : gent ] ) . with this theorem the chern number of the superfluid phase can be simply determined once we know the properties of lower and upper bands , and the normal states at the fermi energy , which govern the phase of @xmath92 . in the @xmath85 space .
the blue and green colors represent the closed fss , and the red color represents open fss . ] to facilitate the presentation , here we show the above theorem for a multiband system , but with only a single fs .
the generic proof can be found in the supplementary material @xcite .
we write down the bdg hamiltonian by @xmath93 where @xmath94 is the normal hamiltonian , @xmath95 is the pairing order matrix , and @xmath85 is the local momentum measured from fs center .
note that if fs is not symmetric with respect to its center , one can continuously deform it to be symmetric without closing the gap .
finally we can always write down @xmath96 in the above form to study the topology .
denote by @xmath97 the eigenvector of the normal band crossing fermi energy .
focusing on the pairing on fs , the eiegenstates of @xmath96 has the form @xmath98^t$ ] .
the chern number of the superfluid phase : @xmath99 , with @xmath100^\dag\nabla_{\bold k}\bigr[\alpha^{(i_m)}({\bold k } ) u_{\bold k}^{(i_m)}\bigr]+\nonumber\\ & & i[\beta^{(i_m)}({\bold k } ) u_{-{\bold k}}^{(i_m)*}]^\dag\nabla_{\bold k}\bigr[\beta^{(i_m)}({\bold k } ) u_{-{\bold k}}^{(i_m)*}\bigr].\end{aligned}\ ] ] let @xmath101 $ ] be the phase of order parameter on fs . taking that @xmath102 with @xmath103 ( without closing bulk gap ) , and with some algebra we can reach @xmath104 \label{eqn : gaugegent}\end{aligned}\ ] ] with @xmath105^\dag\nabla_{\bold k}u_{\bold k}^{(i_m)}$ ] being the berry connection for the normal band states . if we choose a gauge so that @xmath106 is smooth on @xmath107 , the two terms regarding @xmath106 in the right hand side of eq
. cancels .
we then reach the formula for the case with a single fs .
the proof can be generalized to the case with generic multiple fss which may be closed or open [ fig .
[ fig : gent](b ) ] , with multiple bands crossed by fermi energy , and with the pairing within each fs or between two different fss , given that the pairing fully gaps out the bulk @xcite .
since the result is not restricted by pairing types , the generic theorem shown here is powerful to quantitatively determine the topology of the superfluid phases .
on the other hand , it should be noted that this theorem is applied to judge the topology of the phase gapped out by small pairing orders .
starting from the phase governed by eq .
, when the magnitude of superfluid order increases , the system may undergo further topological phase transition and enter a new phase with different topology .
monitoring such phase transitions can determine the topological region of the phase diagram . _ phase diagram._ with the above theorem we can easily determine the topology of the superfluid phase .
note that for the present dirac system , a bcs pairing can not fully gap out the bulk but leads to nodal phases , while the fflo or ff order can .
when there is only one fs , the system with an ff order is topological since @xmath108 and @xmath109 , giving the chern number @xmath110 . in contrast , when there are two fss , from the same or different bands , one can readily find that the contributions from both fss cancels out and the chern number is zero , rendering a trivial phase .
this result implies that the system can be topological when it is in an ff phase . from the mean field results shown in fig .
[ deltacal ] ( a , b ) , we can see that the bcs pairing is greatly suppressed , and @xmath111 dominates over @xmath112 , with @xmath113 for positive @xmath64 . a rich phase diagram is given in fig .
[ deltacal ] ( c , d ) , where the topological and trivial ff phases , and fflo phase are obtained .
it is particularly interesting that in fig .
[ deltacal ] ( d ) a broad topological region is predicted when @xmath114 is away from @xmath115 .
the broad topological region implies that the upper critical value @xmath116 , characterizing the transition from topological ff state to other phases , is largely enhanced compared with the case for @xmath115 .
this is a novel effect explained below .
note that the superfluid order @xmath111 also couples the particle - hole states at @xmath117 .
increasing @xmath111 to @xmath118 closes the bulk gap at @xmath119 momentum , with the critical value being solved from bdg hamiltonian as @xmath120 where @xmath121 . for @xmath115
, we have @xmath122 and a small critical value @xmath123 .
this is because the gap closes at the right hand dirac point @xmath124 , where the original bulk gap less than @xmath125 before having superfluid pairing [ fig .
[ deltacal](e ) ] .
importantly , for @xmath126 , we find that @xmath127 , which is of the order of band width . in this regime @xmath119
is away from the right hand dirac point , and corresponds to a relatively large bulk gap before adding @xmath111 [ fig .
[ deltacal](f ) ] . as a result ,
a large @xmath111 is necessary to drive the phase transition , giving a broad topological region , as shown in fig .
[ deltacal ] ( d )
. dominates over others for @xmath128 .
( c ) phase diagram with normal - size topological region when @xmath129 .
( d ) phase diagram with broad topological region .
the values of the parameters in diagrams ( a , b , d ) are @xmath130 .
schematic diagrams showing the underlying mechanism for the normal - size ( c ) and broad ( d ) topological regions . ]
_ conclusion._in conclusion , we have proposed an experimental scheme to realize chiral fulde - ferrell ( ff ) superfluids , and showed a generic theorem to determine the topology of tsf phases through the normal states .
we start with a tunable dirac semimetal driven by 2d so coupling , which has not been discovered in condensed matter materials but can be readily realized with the current ultracold atom experiments , and show how a superfluid phase with nonzero center of mass momentum is typically favored in the system .
moreover , we have shown a generic formalism for the chern number of a 2d superfluid induced in the normal states , with which the topological chiral ff superfluid with a broad topological region is predicted .
our findings are of significance for both ultracold atoms and condensed matter physics , and also can be useful for material science in search for new topological superconducting states .
this work is supported by most ( grant no .
2016yfa0301604 ) , nsfc ( no . 11574008 ) , and thousand - young - talent program of china . 2 z. wu , l. zhang , w. sun , x .-
t .
xu , b .- z .
wang , s .- c .
ji , y. deng , s. chen , x .- j . liu , and j .- w .
pan , science * 354 * , 83 ( 2016 ) .
l. huang , z. meng , p.wang , p. peng , s .- l . zhang , l. chen , d. li , q. zhou , and j. zhang , nat . phys .
* 12 * , 540 ( 2016 ) .
z. meng , l. huang , p. peng , d. li , l. chen , y. xu , c. zhang , p. wang , and j. zhang , phys .
rev .
lett . * 117 * , 235304 ( 2016 ) .
j. ruseckas , g. juzelinas , p. hberg , m. fleischhauer , phys .
rev .
lett .
* 95 * , 010404 ( 2005 ) . k. osterloh , m. baig , l. santos , p. zoller , m. lewenstein , phys .
rev .
lett . * 95 * , 010403 ( 2005 ) .
x .- j .
liu , m. f. borunda , x. liu , j. sinova , phys .
rev .
lett .
* 102 * , 046402 ( 2009 ) .
n. read and d. green , phys .
rev .
b 61 , 10267 ( 2000 ) .
a. y. kitaev , physics - uspekhi 44 , 131 ( 2001 ) .
a. y. kitaev , ann .
phys .
( amsterdam ) 303 , 2 ( 2003 ) .
l. fu and c. l. kane , phys .
rev .
lett .
100 , 096407 ( 2008 ) .
j. d. sau , r. m. lutchyn , s. tewari , and s. das sarma , phys .
rev .
lett .
104 , 040502 ( 2010 ) .
r. m. lutchyn , j. d. sau , and s. das sarma , phys .
rev .
lett .
105 , 077001 ( 2010 ) .
y. oreg , g. refael , and f. von oppen , phys .
rev .
lett .
105 , 177002 ( 2010 ) . c. zhang , s. tewari , r. m. lutchyn , and s. das sarma , phys .
rev .
lett . 101 , 160401 ( 2008 ) .
m. sato , y. takahashi , and s. fujimoto , phys .
rev .
lett .
103 , 020401 ( 2009 ) .
s .- l .
zhu , l. b. shao , z. d.wang , and l .- m .
duan , phys .
rev .
lett .
106 , 100404 ( 2011 ) . c. qu , z. zheng , m. gong , y. xu , l. mao , x. zou , g. guo , and c. zhang , nature comm . * 4 * , 2710 ( 2013 ) .
w. zhang and w. yi , nature comm . * 4 * , 2711 ( 2013 ) .
y. cao , s. -h .
zou , x .- j .
liu , s. yi , g .-
l .
long , and h. hu , phys . rev .
lett .
113 , 115302 ( 2014 ) x. -j .
liu , k. t. law , and t. k. ng , phys .
rev .
lett . * 112 * , 086401 ( 2014 ) ; _ ibid _ * 113 * , 059901 ( 2014 ) . c. chan and x. -j .
liu , arxiv:1611.08516 .
f. wilczek , nat .
phys . 5 , 614 ( 2009 ) .
j. alicea , reports on progress in physics 75 , 076501 ( 2012 ) .
m. franz , nat .
nano . 8 , 149 ( 2013 ) .
s. r. elliott and m. franz , rev .
mod .
phys .
87 , 137 ( 2015 ) .
p. fulde and r. a. ferrell , phys . rev . *
135 * , a550 ( 1964 ) .
a. i. larkin and y. n. ovchinnikov , zh .
eksp .
. fiz . * 47 * , 1136 ( 1964 ) .
see the supplementary material for details of the realizatoin , and the proof of the generic theorem for computing the chern number .
the reference @xcite has been included .
s. m. young and c. l. kane phys .
rev .
lett . * 115 * , 126803 ( 2015 ) .
t. g. tiecke , feshbach resonances in ultracold mixtures of the fermionic quantum gases @xmath131li and @xmath132k .
phd thesis , university of amsterdam , 2009 .
n. nagaosa , j. sinova , s. onoda , a. h. macdonald , and p. ong , rev .
mod .
phys . * 82 * , 1539 ( 2010 ) .
s .- k .
jian , y .- f .
jiang , and h. yao , phys .
rev .
lett .
114 , 237001 ( 2015 ) .
g. bednik , a. a. zyuzin , and a. a. burkov , phys .
rev .
b 92 , 035153 ( 2015 ) .
t. zhou , y. gao , and z. d. wang , phys .
rev .
b 93 , 094517 ( 2016 ) .
the pairing can occur between two different fermi surfaces , say between @xmath75 and @xmath133 . in this case
the eq .
is still valid , but the integral will be performed on both fermi surfaces at the same time , see @xcite . |
although a systematic study of electromagnetic phenomena in media is not possible without methods of quantum mechanics , statistical physics and kinetics , in practice a standard mathematical model based on phenomenological maxwell s equations provides a good approximation to many important problems . as is well - known
, one should be able to obtain the electromagnetic laws for continuous media from those for the interaction of fields and point particles @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite . as a result of the hard work of several generations of researchers and engineers , the classical electrodynamics , especially in its current complex covariant form , undoubtedly satisfies dirac s criteria of mathematical beauty , being a state of the art mathematical description of nature . in macroscopic electrodynamics , the volume
( mechanical or ponderomotive ) forces , acting on a medium , and the corresponding energy density and energy flux are introduced with the help of the energy - momentum tensors and differential balance relations @xcite , @xcite , @xcite , @xcite , @xcite , @xcite .
these forces occur in the equations of motion for a medium or individual charges and , in principle , they can be experimentally tested @xcite , @xcite , @xcite , @xcite ( see also the references therein ) . but
interpretation of the results should depend on the accepted model of the interaction between the matter and radiation . in this methodological note
, we discuss a complex version of minkowski s phenomenological electrodynamics ( at rest or in a moving medium ) without assuming any particular form of material equations as far as possible .
lorentz invariance of the corresponding differential balance equations is emphasized in view of long - standing uncertainties about the electromagnetic stresses and momentum density , the so - called abraham - minkowski controversy ( see , for example , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite and the references therein ) . the paper is organized as follows . in sections 2 to 4 , we describe the @xmath0-complex version of maxwell s equations and derive the corresponding differential balance density laws for the electromagnetic fields .
their covariant versions are given in sections 5 to 9 .
the case of a uniformly moving medium is discussed in section 10 and complex lagrangians are introduced in section 11 .
some useful tools are collected in appendices a to c for the reader s benefit .
traditionally , the macroscopic maxwell equations in a fixed frame of reference are given by@xmath1@xmath2 here , @xmath3 is the electric field , @xmath4 is the displacement field ; @xmath5 is the magnetic field , @xmath6 is the induction field .
these equations , which are obtained by averaging of microscopic maxwell s equations in the vacuum , provide a good mathematical description of electromagnetic phenomena in various media , when complemented by the corresponding material equations . in the simplest case of an isotropic medium at rest , one usually has@xmath7 where @xmath8 is the dielectric constant , @xmath9 is the magnetic permeability , and @xmath10 describes the conductivity of the medium ( see , for example , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite for fundamentals of classical electrodynamics ) .
introduction of two complex fields@xmath11 allows one to rewrite the phenomenological maxwell equations in the following compact form@xmath12@xmath13 where the asterisk stands for complex conjugation ( see also @xcite , @xcite and @xcite ) .
as we shall demonstrate , different complex forms of maxwell s equations are particularly convenient for study of the corresponding energy - momentum balance equations for the electromagnetic fields in the presence of the free charges and currents in a medium .
we begin from a complex @xmath0-interpretation of the traditional symmetric energy - momentum tensor @xcite . by definition,@xmath14 = t_{qp}\qquad\left ( p , q=1,2,3\right ) \nonumber\end{aligned}\ ] ] and the corresponding momentum balance equation,@xmath15 _ { p}\label{momentumbalance}\\ & \ = \frac{\partial t_{pq}}{\partial x_{q}}+\frac{1}{16\pi}\left [ \operatorname{curl}\left ( \mathbf{f\times g}^{\ast}+\mathbf{f}^{\ast}\times\mathbf{g}\right ) \right ] _ { p}\nonumber\\ & \quad+\frac{1}{16\pi}\left ( f_{q}\frac{\partial g_{q}^{\ast}}{\partial x_{p}}-g_{q}\frac{\partial f_{q}^{\ast}}{\partial x_{p}}+f_{q}^{\ast}\frac{\partial g_{q}}{\partial x_{p}}-g_{q}^{\ast}\frac{\partial f_{q}}{\partial x_{p}}\right ) , \nonumber\end{aligned}\ ] ] can be obtained from maxwell s equations ( [ maxwellcomplexhom])([maxwellcomplexnothom ] ) as a result of elementary but rather tedious vector calculus calculations usually omitted in textbooks .
( we use einstein summation convention over any two repeated indices unless otherwise stated . in this paper ,
greek indices run from @xmath16 to @xmath17 while latin indices may have values from @xmath18 to @xmath19 inclusive . ) * proof . * indeed , in a @xmath0-complex form,@xmath20 due to an identity @xcite:@xmath21 taking into account the complex conjugate , we derive@xmath22 \label{facti}\\ & \quad=\frac{1}{2}\left ( \mathbf{f}\operatorname{div}\mathbf{g}^{\ast } -\mathbf{g}^{\ast}\times\operatorname{curl}\mathbf{f}+\mathbf{f}^{\ast } \operatorname{div}\mathbf{g}-\mathbf{g}\times\operatorname{curl}\mathbf{f}^{\ast}\right ) _ { p}\nonumber\\ & \qquad+\frac{1}{2}\left ( \mathbf{g}\operatorname{div}\mathbf{f}^{\ast } -\mathbf{f}^{\ast}\times\operatorname{curl}\mathbf{g}+\mathbf{g}^{\ast } \operatorname{div}\mathbf{f}-\mathbf{f}\times\operatorname{curl}\mathbf{g}^{\ast}\right ) _ { p}\nonumber\end{aligned}\ ] ] as our first important fact . on the other hand , in view of maxwell s equations ( [ maxwellcomplexhom])([maxwellcomplexnothom ] ) , one gets@xmath23 and , with the help of its complex conjugate,@xmath24 or @xmath25 providing the second important fact . ( up to the constant ,
the first term in the right - hand side represents the density of lorentz s force acting on the free charges and currents in the medium under consideration @xcite , @xcite . ) in view of ( [ factii ] ) and ( [ facti ] ) , we can write@xmath26 \nonumber\\ & \quad\quad-\frac{1}{2}\left ( \mathbf{g}\operatorname{div}\mathbf{f}^{\ast } -\mathbf{f}^{\ast}\times\operatorname{curl}\mathbf{g}+\mathbf{g}^{\ast } \operatorname{div}\mathbf{f}-\mathbf{f}\times\operatorname{curl}\mathbf{g}^{\ast}\right ) _ { p}\nonumber\\ & \quad=\frac{1}{4}\frac{\partial}{\partial x_{q}}\left [ f_{p}g_{q}^{\ast } + f_{p}^{\ast}g_{q}+f_{q}g_{p}^{\ast}+f_{q}^{\ast}g_{p}-\delta_{pq}\left ( \mathbf{f}\cdot\mathbf{g}^{\ast}+\mathbf{f}^{\ast}\cdot\mathbf{g}\right ) \right ] \nonumber\\ & \quad\quad-\frac{1}{4}\left ( \mathbf{g}\operatorname{div}\mathbf{f}^{\ast } -\mathbf{f}^{\ast}\times\operatorname{curl}\mathbf{g}+\mathbf{g}^{\ast } \operatorname{div}\mathbf{f}-\mathbf{f}\times\operatorname{curl}\mathbf{g}^{\ast}\right ) _ { p}\nonumber\\ & \quad\quad+\frac{1}{4}\left ( \mathbf{f}\operatorname{div}\mathbf{g}^{\ast } -\mathbf{g}^{\ast}\times\operatorname{curl}\mathbf{f}+\mathbf{f}^{\ast } \operatorname{div}\mathbf{g}-\mathbf{g}\times\operatorname{curl}\mathbf{f}^{\ast}\right ) _ { p}\nonumber\\ & \quad=4\pi\frac{\partial t_{pq}}{\partial x_{q}}+\frac{1}{4}\left ( \mathbf{f}\operatorname{div}\mathbf{g}^{\ast}-\mathbf{g}^{\ast}\operatorname{div}\mathbf{f}+\mathbf{f}^{\ast}\operatorname{div}\mathbf{g}-\mathbf{g}\operatorname{div}\mathbf{f}^{\ast}\right ) _ { p}\nonumber\\ & \quad\quad+\frac{1}{4}\left ( \mathbf{f}\times\operatorname{curl}\mathbf{g}^{\ast}-\mathbf{g}^{\ast}\times\operatorname{curl}\mathbf{f}+\mathbf{f}^{\ast}\times\operatorname{curl}\mathbf{g}-\mathbf{g}\times\operatorname{curl}\mathbf{f}^{\ast}\right ) _ { p}.\nonumber\end{aligned}\ ] ] finally , in the last two lines , one can utilize the following differential vector calculus identity,@xmath27 _ { p}\label{vectoridentity}\\ & \quad = a_{q}\frac{\partial b_{q}}{\partial x_{p}}-b_{q}\frac{\partial a_{q}}{\partial x_{p}},\nonumber\end{aligned}\ ] ] see ( [ a5 ] ) , with @xmath28 @xmath29 and its complex conjugates , in order to obtain ( [ momentumbalance ] ) and/or ( [ momentumbalancedreal ] ) , which completes the proof .
( an independent proof will be given in section 7 . )
derivation of the corresponding differential energy balance equation is much simpler . by ( [ maxwellcomplexhom]),@xmath30 due to a familiar vector calculus identity ( [ a1 ] ) : @xmath31 in a traditional form,@xmath32 ( see , for example , @xcite , @xcite ) , where one can substitute@xmath33 as a result , @xmath0-differential energy - momentum balance equations are given by@xmath34 and@xmath35 _ { p}+\frac{\partial t_{pq}}{\partial x_{q}}-\left ( \rho\mathbf{e}+\frac{1}{c}\mathbf{j}\times\mathbf{b}\right ) _ { p}\label{momentumbalancedreal}\\ &
\ + \frac{1}{8\pi}\left [ \operatorname{curl}\left ( \mathbf{e\times d}+\mathbf{h}\times\mathbf{b}\right ) \right ] _ { p}\nonumber\\ & \quad+\frac{1}{8\pi}\left ( \mathbf{e}\cdot\frac{\partial\mathbf{d}}{\partial x_{p}}-\mathbf{d}\cdot\frac{\partial\mathbf{e}}{\partial x_{p}}+\mathbf{h}\cdot\frac{\partial\mathbf{b}}{\partial x_{p}}-\mathbf{b}\cdot\frac{\partial\mathbf{h}}{\partial x_{p}}\right ) = 0,\nonumber\end{aligned}\ ] ] respectively ( see also @xcite , @xcite ) .
the real form of the symmetric stress tensor ( [ stresssymcomplex ] ) , namely,@xmath36 \qquad\left ( p , q=1,2,3\right ) , \nonumber\end{aligned}\ ] ] is due to hertz @xcite .
equations ( [ energybalance])([momentumbalancedreal ] ) are related to a fundamental concept of conservation of mechanical and electromagnetic energy and momentum . here , these balance conditions are presented in differential forms in terms of the corresponding local field densities .
they can be integrated over a given volume in @xmath37 in order to obtain , in a traditional way , the corresponding conservation laws of the electromagnetic fields ( see , for example , @xcite , @xcite , @xcite , @xcite , @xcite ) .
these laws made it necessary to ascribe a definite linear momentum and energy to the field of an electromagnetic wave , which can be observed , for example , as light pressure .
* at this point , the lorentz invariance of these differential balance equations is not obvious in our @xmath0-analysis .
but one can introduce the four - vector @xmath38 and try to match ( [ energybalance])([momentumbalancedreal ] ) with the expression,@xmath39 as an initial step , in order to guess the corresponding four - tensor form .
an independent covariant derivation will be given in section 7 .
* in an isotropic nonhomogeneous variable medium ( without dispersion and/or compression ) , when @xmath40 and @xmath41 the ponderomotive forces in ( [ energybalance ] ) and ( [ momentumbalancedreal ] ) take the form @xcite:@xmath42{c}\dfrac{1}{c}\left ( \dfrac{\partial\varepsilon}{\partial t}\mathbf{e}^{2}+\dfrac{\partial\mu}{\partial t}\mathbf{h}^{2}\right ) \bigskip\\ \mathbf{e}^{2}\nabla\varepsilon+\mathbf{h}^{2}\nabla\mu \end{array } \right ) , \nonumber\end{aligned}\ ] ] which may be interpreted as a four - vector energy - force acting from an inhomogeneous and time - variable medium . its covariance is analyzed in section 7 .
the @xmath0-linear momentum differential balance equation ( [ momentumbalancedreal ] ) , can be rewritten in a more compact form,@xmath43 with the help of the hertz symmetric stress tensor @xmath44 defined by ( [ stresssymreal ] ) .
a net force is given by@xmath45 _ { p}\label{netforce}\\ & -\frac{1}{8\pi}\left ( \mathbf{e}\cdot\frac{\partial\mathbf{d}}{\partial x_{p}}-\mathbf{d}\cdot\frac{\partial\mathbf{e}}{\partial x_{p}}+\mathbf{h}\cdot\frac{\partial\mathbf{b}}{\partial x_{p}}-\mathbf{b}\cdot\frac { \partial\mathbf{h}}{\partial x_{p}}\right ) .\nonumber\end{aligned}\ ] ] in this notation , we state the @xmath0-angular momentum differential balance equation as follows@xmath46 where the field angular momentum density is defined by@xmath47 and the flux of angular momentum is described by the following tensor @xcite:@xmath48 ( here , @xmath49 is the totally anti - symmetric levi - civita symbol with @xmath50 ) .
an elementary example of conservation of the total angular momentum is discussed in @xcite .
* indeed , in view of ( [ compactmomentumbalance ] ) , one can write@xmath51 which completes the proof
. * note .
* once again , in @xmath0-form , the lorentz invariance of this differential balance equation for the local densities is not obvious .
an independent covariant derivation will be given in section 8 .
with the help of complex fields @xmath52 and @xmath53 we introduce the following anti - symmetric four - tensor,@xmath54{cccc}0 & -g_{1 } & -g_{2 } & -g_{3}\\ g_{1 } & 0 & if_{3 } & -if_{2}\\ g_{2 } & -if_{3 } & 0 & if_{1}\\ g_{3 } & if_{2 } & -if_{1 } & 0 \end{array } \right ) \label{complexqtensor}\ ] ] and use the standard four - vectors , @xmath38 and @xmath55 for contravariant coordinates and current , respectively .
maxwell s equations then take the covariant form @xcite , @xcite : @xmath56 with summation over two repeated indices . indeed , in block form , we have@xmath57{cc}0 & -g_{q}\medskip\\ g_{p } & ie_{pqr}f_{r}\end{array } \right ) = \left ( \begin{array } [ c]{c}-\operatorname{div}\mathbf{g}=-4\pi\rho\medskip\\ \dfrac{1}{c}\dfrac{\partial\mathbf{g}}{\partial t}+i\operatorname{curl}\mathbf{f}=-\dfrac{4\pi}{c}\mathbf{j}\end{array } \right ) , \label{cmproof}\ ] ] which verifies this fact .
the continuity equation,@xmath58 or in the @xmath0-form,@xmath59 describes conservation of the electrical charge .
the latter equation can also be derived in the complex @xmath0-form from ( [ maxwellcomplexhom])([maxwellcomplexnothom ] ) . * note . * in vacuum , when @xmath60 and @xmath61 @xmath62 one can write due to ( [ b1])([b2]):@xmath63 as a result , the following self - duality property holds@xmath64 ( see , for example , @xcite , @xcite and appendix b ) .
two covariant forms of maxwell s equations are given by@xmath65 where @xmath66 @xmath67 and @xmath68diag@xmath69 the last equation can be derived from a more general equation , involving a rank three tensor,@xmath70 ( @xmath71 are fixed ; no summation is assumed over these two indices ) , which is related to the pauli - lubaski vector from the representation theory of the poincar group @xcite .
different spinor forms of maxwell s equations are analyzed in @xcite ( see also the references therein ) .
two dual anti - symmetric field tensors of complex fields , @xmath52 and @xmath53 are given by @xmath72{cccc}0 & -g_{1 } & -g_{2 } & -g_{3}\\ g_{1 } & 0 & if_{3 } & -if_{2}\\ g_{2 } & -if_{3 } & 0 & if_{1}\\ g_{3 } & if_{2 } & -if_{1 } & 0 \end{array } \right ) = r^{\mu\nu}+is^{\mu\nu}\label{complexq}\\ & = \left ( \begin{array } [ c]{cccc}0 & -d_{1 } & -d_{2 } & -d_{3}\\ d_{1 } & 0 & -h_{3 } & h_{2}\\ d_{2 } & h_{3 } & 0 & -h_{1}\\ d_{3 } & -h_{2 } & h_{1 } & 0 \end{array } \right ) + i\left ( \begin{array } [ c]{cccc}0 & -b_{1 } & -b_{2 } & -b_{3}\\ b_{1 } & 0 & e_{3 } & -e_{2}\\ b_{2 } & -e_{3 } & 0 & e_{1}\\ b_{3 } & e_{2 } & -e_{1 } & 0 \end{array } \right ) \nonumber\end{aligned}\ ] ] and@xmath73{cccc}0 & f_{1 } & f_{2 } & f_{3}\\ -f_{1 } & 0 & ig_{3 } & -ig_{2}\\ -f_{2 } & -ig_{3 } & 0 & ig_{1}\\ -f_{3 } & ig_{2 } & -ig_{1 } & 0 \end{array } \right ) = f_{\mu\nu}+ig_{\mu\nu}\label{complexp}\\ & = \left ( \begin{array } [ c]{cccc}0 & e_{1 } & e_{2 } & e_{3}\\ -e_{1 } & 0 & -b_{3 } & b_{2}\\ -e_{2 } & b_{3 } & 0 & -b_{1}\\ -e_{3 } & -b_{2 } & b_{1 } & 0 \end{array } \right ) + i\left ( \begin{array } [ c]{cccc}0 & h_{1 } & h_{2 } & h_{3}\\ -h_{1 } & 0 & d_{3 } & -d_{2}\\ -h_{2 } & -d_{3 } & 0 & d_{1}\\ -h_{3 } & d_{2 } & -d_{1 } & 0 \end{array } \right ) .\nonumber\end{aligned}\ ] ] the real part of the latter represents the standard electromagnetic field tensor in a medium @xcite , @xcite , @xcite . as for the imaginary part of ( [ complexq ] ) , which , ironically , pauli called an artificiality in view of its non - standard behavior under spatial inversion @xcite , the use of complex conjugation restores this important symmetry for our complex field tensors . the dual tensor identities are given by@xmath74 here @xmath75 and @xmath76 is the levi - civita four - symbol @xcite . then@xmath77 and both pairs of maxwell s equations
can also be presented in the form @xcite@xmath78 in addition to the one given above@xmath79 the real part of the first equation traditionally represents the first ( homogeneous ) pair of maxwell s equation and the real part of the second one gives the remaining pair . in our approach both pairs of maxwell s equations appear together ( see also @xcite , @xcite , @xcite , @xcite , and @xcite for the case in vacuum ) . moreover , a generalization to complex - valued four - current may naturally represent magnetic charge and magnetic current not yet observed in nature @xcite .
an important cofactor matrix identity,@xmath80 was originally established , in a general form , by minkowski @xcite .
once again , the dual tensors are given by@xmath81{cc}0 & f_{q}\medskip\\ -f_{p } & ie_{pqr}g_{r}\end{array } \right ) , \quad q^{\mu\nu}=\left ( \begin{array } [ c]{cc}0 & -g_{q}\medskip\\ g_{p } & ie_{pqr}f_{r}\end{array } \right ) , \label{pqblockform}\ ] ] in block form .
a complete list of relevant tensor and matrix identities is given in appendix b.
as has been announced in @xcite ( see also @xcite ) , the covariant form of the differential balance equations can be presented as follows@xmath82 \label{covariantenergymomentumbalance}\\ & \ + \frac{1}{32\pi}\left ( p_{\sigma\tau}^{\ast}\frac{\partial q^{\tau \sigma}}{\partial x^{\mu}}+p_{\sigma\tau}\frac{\partial\overset{\ast}{\left .
q^{\tau\sigma}\right . } } { \partial x^{\mu}}\right ) \nonumber\\ & \
= -\frac{1}{c}f_{\mu\lambda}j^{\lambda}=\left ( \begin{array } [ c]{c}-\mathbf{j}\cdot\mathbf{e}/c\smallskip\\ \rho\mathbf{e}+\mathbf{j}\times\mathbf{b}/c \end{array } \right ) .\nonumber\end{aligned}\ ] ] in our complex form , when @xmath52 and @xmath53 the energy - momentum tensor is given by@xmath83{cc}\mathbf{f}\cdot\mathbf{g}^{\ast}+\mathbf{f}^{\ast}\cdot\mathbf{g } & 2i\left ( \mathbf{f\times f}^{\ast}\right ) _ { q}\bigskip\\ -2i\left ( \mathbf{g\times g}^{\ast}\right ) _ { p } & \quad2\left ( f_{p}g_{q}^{\ast}+f_{p}^{\ast}g_{q}\right ) -\delta_{pq}\left ( \mathbf{f}\cdot\mathbf{g}^{\ast}+\mathbf{f}^{\ast}\cdot\mathbf{g}\right ) \end{array } \right ) .\nonumber\end{aligned}\ ] ] here , we point out for the reader s convenience that @xmath84 and , in real form,@xmath85{cc}\left ( \mathbf{e}\cdot\mathbf{d}+\mathbf{h}\cdot\mathbf{b}\right ) /2 & \left ( \mathbf{e\times h}\right ) _ { q}\bigskip\\ -\left ( \mathbf{d\times b}\right ) _ { p } & \quad e_{p}d_{q}+h_{p}b_{q}-\delta_{pq}\left ( \mathbf{e}\cdot\mathbf{d}+\mathbf{h}\cdot\mathbf{b}\right ) /2 \end{array } \right ) . \label{4energymomentumtensorreal}\ ] ] the covariant form of the differential balance equation allows one to clarify the meanings of different energy - momentum tensors .
for instance , it is worth noting that the non - symmetric maxwell and heaviside form of the @xmath0-stress tensor @xcite , @xmath86 appears here in the corresponding momentum balance equation @xcite:@xmath35 _ { p}+\frac{\partial\widetilde{t}_{pq}}{\partial x_{q}}-\left ( \rho\mathbf{e}+\frac{1}{c}\mathbf{j}\times \mathbf{b}\right ) _ { p}\label{momentumbalancedrealnonsymmetric}\\ & \quad+\frac{1}{8\pi}\left ( \mathbf{e}\cdot\frac{\partial\mathbf{d}}{\partial x_{p}}-\mathbf{d}\cdot\frac{\partial\mathbf{e}}{\partial x_{p}}+\mathbf{h}\cdot\frac{\partial\mathbf{b}}{\partial x_{p}}-\mathbf{b}\cdot\frac{\partial\mathbf{h}}{\partial x_{p}}\right ) = 0.\nonumber\end{aligned}\ ] ] at the same time , in view of ( [ momentumbalancedreal ] ) , use of the form ( [ 3denergymomentummaxwellhtensor ] ) differs from hertz s symmetric tensors in ( [ stresssymcomplex ] ) and ( [ stresssymreal ] ) only in the case of anisotropic media ( crystals ) @xcite , @xcite .
indeed,@xmath87 _ { p}. \label{divergencetwotensors}\ ] ] moreover , with the help of elementary identities,@xmath88 _ { p}=\frac{\partial}{\partial x_{q}}\left ( a_{p}b_{q}-a_{q}b_{p}\right ) \label{curlidentity}\ ] ] and@xmath89 _ { p } , \label{curlidentitytwo}\ ] ] one can transform the latter balance equation into its symmetric form , which provides an independent proof of ( [ momentumbalancedreal ] ) . the fact that maxwell s equations can be united with the help of a complex second rank ( anti - symmetric ) tensor allows us to utilize the standard sturm - liouville type argument in order to establish the energy - momentum differential balance equations in covariant form . indeed , by adding matrix equation@xmath90 and its complex conjugate@xmath91 one gets@xmath92 a simple decomposition,@xmath93 with @xmath94 and @xmath95 ( and their complex conjugates ) , results in @xmath82 \label{pqforce}\\ & \ + \frac{1}{16\pi}\left [ \left ( p_{\mu\lambda}^{\ast}\frac{\partial q^{\lambda\nu}}{\partial x^{\nu}}-\frac{\partial p_{\mu\lambda}}{\partial x^{\nu}}\overset{\ast}{\left . q^{\lambda\nu}\right . }
\right ) + \left ( \text{c.c.}\right ) \right ] = -\frac{1}{c}f_{\mu\lambda}j^{\lambda}.\nonumber\end{aligned}\ ] ] by a direct substitution , one can verify that @xmath96 ( an independent covariant proof of these identities is given in appendix c. ) finally , introducing@xmath97 we obtain ( [ covariantenergymomentumbalance ] ) with the explicitly covariant expression for the ponderomotive force ( [ pforce ] ) , which completes the proof . as a result ,
the covariant energy - momentum balance equation is given by@xmath98 in a compact form . if these differential balance equations are written for a stationary medium , then the corresponding equations for moving bodies are uniquely determined , since the components of a tensor in any inertial coordinate system can be derived by a proper lorentz transformation @xcite .
by definition , @xmath99 and @xmath100 where @xmath101diag@xmath102 in view of ( [ covariantembalancecompact ] ) , we derive@xmath103 as a required differential balance equation . with the help of familiar dual relations ( [ b04 ] ) , one can get another covariant form of the angular momentum balance equation:@xmath104 in @xmath0-form , the latter relation can be reduced to ( [ vectormomentumbalance])([tensormomentum ] ) . indeed , when @xmath105 and @xmath106 one gets@xmath107 + \frac{\partial}{\partial x_{s}}\left ( e_{pqr}x_{q}\widetilde{t}_{rs}\right ) \label{4dto3dangmom}\\ & \qquad+e_{pqr}\widetilde{t}_{qr}+e_{pqr}x_{q}\left ( x_{r}+y_{r}\right ) = 0,\nonumber\end{aligned}\ ] ] where @xmath108 is the familiar lorentz force .
substitution , @xmath109 results in ( [ vectormomentumbalance ] ) in view of identity ( [ divergencetwotensors ] ) .
the remaining cases , when @xmath110 can be analyzed in a similar fashion . in @xmath0-form
, the corresponding equations can be reduced to ( [ energybalance ] ) and ( [ momentumbalancedrealnonsymmetric ] ) .
details are left to the reader .
thus the angular momentum law has the form of a local balance equation , not a conservation law , since in general the energy - momentum tensor will not be symmetric @xcite .
a torque , for instance , may occur , which can not be compensated for by a change in the electromagnetic angular momentum , though not in contradiction with experiment @xcite .
let @xmath111 be a constant real - valued velocity vector representing uniform motion of one frame of reference with respect to another one .
let us consider the following orthogonal decompositions,@xmath112 such that our complex vectors @xmath113 are collinear with the velocity vector @xmath111 and @xmath114 are perpendicular to it ( figure 1 ) .
the lorentz transformation of electric and magnetic fields @xmath115 take the following complex form@xmath116 and@xmath117 although this transformation was found by lorentz , it was minkowski who realized that this is the law of transformation of the second rank anti - symmetric four - tensors @xcite , @xcite ; a brief historical overview is given in @xcite . )
this complex @xmath0-form of the lorentz transformation of electric and magnetic fields was known to minkowski ( 1908 ) , but apparently only in vacuum , when @xmath60 ( see also @xcite ) .
moreover,@xmath118 in the same notation @xcite .
the latter equations can be rewritten as follows@xmath119 \mathbf{v},\qquad t^{\prime}=\gamma\left ( t-\frac{\mathbf{v}\cdot\mathbf{r}}{c^{2}}\right ) , \label{lorentzcoordinatetransformshort}\ ] ] where @xmath120 in a similar fashion , one gets@xmath121@xmath122 as a compact @xmath0-version of the lorentz transformation for the complex electromagnetic fields . in complex four - tensor form,@xmath123
although minkowski considered the transformation of electric and magnetic fields in a complex @xmath0-vector form , see eqs .
( 8)(9 ) and ( 15 ) in @xcite ( or eqs .
( 25.5)(25.6 ) in @xcite ) , he seems never to have combined the corresponding four - tensors into the complex forms ( [ complexq])([complexp ] ) . in the second article @xcite , max
born , who used minkowski s notes , did nt mention the complex fields . as a result , the complex field tensor seems only to have appeared , for the first time , in @xcite ( see also @xcite ) .
the complex identity , @xmath124invariant@xmath125under a similarity transformation , follows from minkowski s determinant relations ( [ b18])([b20 ] ) . * example*. let @xmath126 be an orthonormal basis in @xmath127 we choose @xmath128 and write @xmath129 with@xmath130{cccc}\gamma & -\beta\gamma & 0 & 0\\ -\beta\gamma & \gamma & 0 & 0\\ 0 & 0 & 1 & 0\\ 0 & 0 & 0 & 1 \end{array } \right ) , \qquad\quad\beta=\frac{v}{c},\quad\gamma=\frac{1}{\sqrt{1-\beta^{2 } } } \label{lorentzboost}\ ] ] for the corresponding lorentz boost ( figure 2 ) . in view of ( [ complexfieldtransform ] ) , by matrix multiplication one gets@xmath131{cccc}\gamma & -\beta\gamma & 0 & 0\\ -\beta\gamma & \gamma & 0 & 0\\ 0 & 0 & 1 & 0\\ 0 & 0 & 0 & 1 \end{array } \right ) \left ( \begin{array } [ c]{cccc}0 & -g_{1 } & -g_{2 } & -g_{3}\\ g_{1 } & 0 & if_{3 } & -if_{2}\\ g_{2 } & -if_{3 } & 0 & if_{1}\\ g_{3 } & if_{2 } & -if_{1 } & 0 \end{array } \right ) \left ( \begin{array } [ c]{cccc}\gamma & -\beta\gamma & 0 & 0\\ -\beta\gamma & \gamma & 0 & 0\\ 0 & 0 & 1 & 0\\ 0 & 0 & 0 & 1 \end{array } \right ) \nonumber\\ & \qquad=\left ( \begin{array } [ c]{cccc}0 & -g_{1 } & -\gamma g_{2}-i\beta\gamma f_{3 } & -\gamma g_{3}+i\beta\gamma f_{2}\\ g_{1 } & 0 & \beta\gamma g_{2}+i\gamma f_{3 } & \beta\gamma g_{3}-i\gamma f_{2}\\ \gamma g_{2}+i\beta\gamma f_{3 } & -\beta\gamma g_{2}-i\gamma f_{3 } & 0 & if_{1}\\ \gamma g_{3}-i\beta\gamma f_{2 } & -\beta\gamma g_{3}+i\gamma f_{2 } & -if_{1 } & 0 \end{array } \right ) . \label{martixmuttiplication}\ ] ] thus @xmath132 and@xmath133 in a similar fashion , @xmath134 and@xmath135 the latter relations are in agreement with the field transformations ( [ paralleltransform])([perpendiculartransform ] ) . in block form ,
one gets@xmath136{c}f_{1}^{\prime}\\ f_{2}^{\prime}\\ g_{3}^{\prime}\\ g_{2}^{\prime}\\ f_{3}^{\prime}\\ g_{1}^{\prime}\end{array } \right ) = \left ( \begin{array } [ c]{cccccc}1 & 0 & 0 & 0 & 0 & 0\\ 0 & \cos\left ( i\theta\right ) & \sin\left ( i\theta\right ) & 0 & 0 & 0\\ 0 & -\sin\left ( i\theta\right ) & \cos\left ( i\theta\right ) & 0 & 0 & 0\\ 0 & 0 & 0 & \cos\left ( i\theta\right ) & \sin\left ( i\theta\right ) & 0\\ 0 & 0 & 0 & -\sin\left ( i\theta\right ) & \cos\left ( i\theta\right ) & 0\\ 0 & 0 & 0 & 0 & 0 & 1 \end{array } \right ) \left ( \begin{array } [ c]{c}f_{1}\\ f_{2}\\ g_{3}\\ g_{2}\\ f_{3}\\ g_{1}\end{array } \right ) , \label{emfieldstransformmatrix}\ ] ] where , by definition,@xmath137 as a result , the transformation law of the complex electromagnetic fields @xmath138 under the lorentz boost can be thought of as a complex rotation in @xmath139 corresponding to a reducible representation of the one - parameter subgroup of @xmath140 ( cyclic permutation of the spatial indices cover the two remaining cases ; see also @xcite . )
electromagnetic phenomena in moving media are important in relativistic astrophysics , the study of accelerated plasma clusters and high - energy electron beams @xcite , @xcite , @xcite , @xcite . minkowski s field- and connecting - equations @xcite , @xcite were derived from the corresponding laws for the bodies at rest by means of a lorentz transformation ( see @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) .
explicitly covariant forms , which are applicable both in the rest frame and for moving media , are analyzed in @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ( see also the references therein ) . in standard
notation,@xmath141 one can write @xcite , @xcite , @xcite , @xcite : @xmath142 , \label{3dmaterialeqs}\\ \mathbf{h } & = \frac{1}{\mu}\mathbf{b}+\frac{\kappa\gamma^{2}}{\mu}\left [ -\beta^{2}\mathbf{b}+\dfrac{\mathbf{v}}{c^{2}}\left ( \mathbf{v\cdot b}\right ) + \dfrac{1}{c}\left ( \mathbf{v\times e}\right ) \right ] .\nonumber\end{aligned}\ ] ] in covariant form , these relations are given by@xmath143 ( see @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite and the references therein ) .
here,@xmath144 is the four - tensor of electric and magnetic permeabilities and@xmath145 is the four - velocity of the medium ( @xcite , @xcite , a computer algebra verification of these relations is given in @xcite ) . in a complex covariant form,@xmath146 in view of ( [ covariantmaterialeqs ] ) and ( [ b1])([b2 ] ) , we get@xmath147 in terms of the real - valued electromagnetic field tensor . in practice
, one can choose@xmath148 for the real - valued four - vector potential @xmath149 then@xmath150 by ( [ fourpermeabilitytensor ] ) .
substitution into maxwell s equations ( [ maxwellii ] ) or ( [ maxwelli ] ) results in @xmath151 a_{\sigma}\right .
\label{covariantau}\\ & \qquad\left .
+ \partial_{\sigma}\left ( \partial^{\tau}a_{\tau}+\kappa u^{\nu}u^{\tau}\partial_{\nu}a_{\tau}\right ) ^{\ } \!\right\ } = -\frac{4\pi \mu}{c}j^{\lambda},\nonumber\end{aligned}\ ] ] where @xmath152 is the dalembert operator . in view of an inverse matrix identity,@xmath153 the latter equations
take the form ) and ( [ covariantauj ] ) , together with the gauge condition ( [ covariantaugauge ] ) , may be considered as the fundamentals of the theory @xcite .
our complex fields are given by ( [ qpmaterialf]).]@xmath154 a_{\sigma}-\partial_{\sigma}\left ( \partial^{\tau}a_{\tau}+\kappa u^{\nu}u^{\tau}\partial_{\nu}a_{\tau}\right ) \label{covariantauj}\\ & \qquad=\frac{4\pi\mu}{c}\left ( g_{\sigma\lambda}-\frac{\kappa}{1+\kappa } u_{\sigma}u_{\lambda}\right ) j^{\lambda}.\nonumber\end{aligned}\ ] ] subject to the subsidiary condition,@xmath155 these equations were studied in detail for the sake of development of the phenomenological classical and quantum electrodynamics in a moving medium ( see @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite and the references therein ) .
in particular , green s function of the photon in a moving medium was studied in @xcite , @xcite , @xcite ( with applications to quantum electrodynamics ) .
we follow @xcite , @xcite , @xcite with somewhat different details .
the substitution,@xmath156 ( a generalization of hertz s potentials for a moving medium @xcite , @xcite ) , into the gauge condition ( [ covariantaugauge ] ) results in @xmath157 in view of@xmath158 then , equations ( [ covariantauj ] ) take the form@xmath159 \partial_{\sigma}z^{\lambda\sigma}=-\frac{4\pi\mu}{c}j^{\lambda}. \label{dzjhertzequation}\ ] ] indeed , the left - hand side of ( [ covariantauj ] ) is given by@xmath154
a_{\sigma}=\left [ \partial^{\tau}\partial_{\tau } + \kappa\left ( u^{\tau}\partial_{\tau}\right ) ^{2}\right ] g_{\sigma\mu } a^{\mu}\\ & \quad=\left [ \partial^{\tau}\partial_{\tau}+\kappa\left ( u^{\tau}\partial_{\tau}\right ) ^{2}\right ] g_{\sigma\mu}\left ( \frac{\kappa } { 1+\kappa}u^{\mu}u_{\lambda}-\delta_{\lambda}^{\mu}\right ) \partial_{\rho } z^{\lambda\rho}\\ & \quad=\left [ \partial^{\tau}\partial_{\tau}+\kappa\left ( u^{\tau}\partial_{\tau}\right ) ^{2}\right ] \left ( \frac{\kappa}{1+\kappa}u_{\sigma } u_{\lambda}-g_{\sigma\lambda}\right ) \partial_{\rho}z^{\lambda\rho}\\ & \qquad=\frac{4\pi\mu}{c}\left ( g_{\sigma\lambda}-\frac{\kappa}{1+\kappa } u_{\sigma}u_{\lambda}\right ) j^{\lambda},\end{aligned}\ ] ] from which the result follows due to ( [ inversea ] ) . finally ,
with the help of the standard substitution,@xmath160 ( in view of @xmath161 we arrive at@xmath162 z^{\lambda\sigma}+4\pi\mu p^{\lambda\sigma}\right\ } = 0 .
\label{dzphertz}\ ] ] therefore , one can choose@xmath159 z^{\lambda\nu}=-4\pi p^{\lambda\nu}. \label{zphertz}\ ] ] here , by definition,@xmath163{cccc}0 & -p_{1 } & -p_{2 } & -p_{3}\\ p_{1 } & 0 & m_{3 } & -m_{2}\\ p_{2 } & -m_{3 } & 0 & m_{1}\\ p_{3 } & m_{2 } & -m_{1 } & 0 \end{array } \right ) \label{ptensor}\ ] ] is an anti - symmetric four - tensor @xcite , @xcite , @xcite .
the electric and magnetic hertz vectors , @xmath164 and @xmath165 are also introduced in terms of a single four - tensor,@xmath166{cccc}0 & z_{1}^{\left ( e\right ) } & z_{2}^{\left ( e\right ) } & z_{3}^{\left ( e\right ) } \\
-z_{1}^{\left ( e\right ) } & 0 & -z_{3}^{\left ( m\right ) } & z_{2}^{\left ( m\right ) } \\
-z_{2}^{\left ( e\right ) } & z_{3}^{\left ( m\right ) } & 0 & -z_{1}^{\left ( m\right ) }
\\ -z_{3}^{\left ( e\right ) } & -z_{2}^{\left ( m\right ) } & z_{1}^{\left ( m\right ) } & 0 \end{array } \right ) .
\label{ztensor}\ ] ] in view of ( [ azhertz ] ) , for the four - vector potential , @xmath167 we obtain@xmath168 and@xmath169 .\nonumber\end{aligned}\ ] ] then , equations ( [ zphertz ] ) take the form@xmath159 \mathbf{z}^{\left ( e\right ) } = 4\pi\mu\mathbf{p},\qquad\left [ \partial^{\tau}\partial_{\tau}+\kappa\left ( u^{\tau}\partial_{\tau}\right ) ^{2}\right ] \mathbf{z}^{\left ( m\right ) } = 4\pi \mu\mathbf{m } \label{zempm}\ ] ] and , for the four - current , @xmath170 one gets@xmath171 ( see @xcite , @xcite , @xcite for more details ) .
the hertz vector and tensor potentials , for a moving medium and at rest , were utilized in @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ( see also the references therein ) .
many classical problems of radiation and propagation can be consistently solved by using these potentials . in the case of the covariant version of the energy - momentum tensor given by ( [ fourenergymomentumtensor ] ) , the differential balance equations under consideration are independent of the particular choice of the frame of reference .
therefore , our relations ( [ qpmaterialf ] ) are useful for derivation of the expressions for the energy - momentum tensor and the ponderomotive force for moving bodies from those for bodies at rest which were extensively studied in the literature .
for example , one gets@xmath172 with the help of ( [ covariantmaterialeqs])([fourpermeabilitytensor ] ) and ( [ b10 ] ) ( see also @xcite ) .
in modern presentations of the classical and quantum field theories , the lagrangian approach is usually utilized .
we introduce two quadratic lagrangian densities@xmath173 and@xmath174 then , by formal differentiation,@xmath175 and@xmath176 in view of ( [ b3 ] ) .
the complex covariant maxwell equations ( [ covariantmaxwell ] ) take the forms@xmath177 and the covariant energy - momentum balance relations ( [ covariantenergymomentumbalance ] ) are given by@xmath178 \label{energymomentumlagrangian0}\\ & \quad+\frac{1}{32\pi}\left [ p_{\sigma\tau}^{\ast}\frac{\partial}{\partial x^{\mu}}\left ( \frac{\partial\mathcal{l}_{0}}{\partial p_{\sigma\tau}}\right ) + p_{\sigma\tau}\frac{\partial}{\partial x^{\mu}}\left ( \frac{\partial\mathcal{l}_{0}}{\partial p_{\sigma\tau}^{\ast}}\right ) \right ] = -\frac{1}{c}f_{\mu\lambda}j^{\lambda}\nonumber\end{aligned}\ ] ] and@xmath179 \label{energymomentumlagrangian1}\\ & \quad+\frac{1}{32\pi}\left [ p_{\sigma\tau}\frac{\partial}{\partial x^{\mu } } \left ( \frac{\partial\mathcal{l}_{1}}{\partial p_{\sigma\tau}}\right ) + p_{\sigma\tau}^{\ast}\frac{\partial}{\partial x^{\mu}}\left ( \frac { \partial\mathcal{l}_{1}^{\ast}}{\partial p_{\sigma\tau}^{\ast}}\right ) \right ] = \frac{1}{c}f_{\mu\lambda}j^{\lambda}\nonumber\end{aligned}\ ] ] in terms of the complex lagrangians under consideration , respectively .
finally , with the help of the following densities,@xmath180 one can derive analogs of the euler - lagrange equations for electromagnetic fields in media:@xmath181 in the case of a moving isotropic medium , a relation between @xmath182 and @xmath183 is given by our equations ( [ qpmaterialf])([4vectorpotential ] ) .
taking the real and imaginary parts , maxwell s equations ( [ maxwellii ] ) can be written as follows@xmath184 here,@xmath185 with the help of ( [ qpidentity ] ) and ( [ 4vectorpotential ] ) .
thus the second set of equations is automatically satisfied when we introduce the four - vector potential .
for the inhomogeneous pair of maxwell s equations , the lagrangian density is given by@xmath186 in view of ( [ covariantmaterialeqs ] ) .
then , for conjugate momenta to the four - potential field @xmath187 one gets@xmath188 and the corresponding euler - lagrange equations take a familiar form@xmath189 the latter equation can also be derived with the help of the least action principle @xcite , @xcite , @xcite .
the corresponding hamiltonian and quantization are discussed in @xcite , @xcite , @xcite among other classical accounts . in conclusion ,
it is worth noting the role of complex fields in quantum electrodynamics , quadratic invariants and quantization ( see , for instance , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) .
the classical and quantum theory of cherenkov radiation is reviewed in @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite .
for paraxial approximation in optics , see @xcite , @xcite , @xcite , @xcite , @xcite , and the references therein .
maxwell s equations in the gravitational field are discussed in @xcite @xcite .
one may hope that our detailed mathematical consideration of several aspects of macroscopic electrodynamics will be useful for future investigations and pedagogy .
* acknowledgements*. we dedicate this article to the memory of professor alladi ramakrishnan who made significant contributions to probability and statistics , elementary particle physics , cosmic rays and astrophysics , matrix theory , and the special theory of relativity @xcite .
we are grateful to krishna alladi , albert boggess , mark p. faifman , john r. klauder , vladimir i. manko , and igor n. toptygin for valuable comments and help .
among useful differential relations are@xmath190@xmath191@xmath192@xmath193@xmath194 ( see also @xcite , @xcite and @xcite . ) here , @xmath195 and @xmath196
in this article , @xmath75 and @xmath76 is the levi - civita four - symbol @xcite with familiar contractions:@xmath197{ccc}\delta_{\kappa}^{\nu } & \delta_{\lambda}^{\nu } & \delta_{\rho}^{\nu}\\ \delta_{\kappa}^{\sigma } & \delta_{\lambda}^{\sigma } & \delta_{\rho}^{\sigma } \\
\delta_{\kappa}^{\tau } & \delta_{\lambda}^{\tau } & \delta_{\rho}^{\tau}\end{array } \right\vert , \label{b01}\]]@xmath198{cc}\delta_{\lambda}^{\sigma } & \delta_{\rho}^{\sigma}\\ \delta_{\lambda}^{\tau } & \delta_{\rho}^{\tau}\end{array } \right\vert = -2\left ( \delta_{\lambda}^{\sigma}\delta_{\rho}^{\tau}-\delta_{\rho}^{\sigma}\delta_{\lambda}^{\tau}\right ) , \label{b02}\]]@xmath199 dual second rank four - tensor identities are given by @xcite:@xmath200 in particular,@xmath201@xmath202@xmath203@xmath204@xmath205 by direct calculation,@xmath206@xmath207 as a result,@xmath208 an important decomposition,@xmath209 is complemented by an identity,@xmath210 in matrix form,@xmath211 here,@xmath212@xmath213@xmath214@xmath215 where @xmath216diag@xmath217 is the identity matrix .
also,@xmath218 and@xmath219 other useful dual four - tensor identities are given by @xcite : @xmath220 in particular,@xmath221 and@xmath222 ( see also @xcite ) .
in view of ( [ pqdual ] ) , or ( [ b3 ] ) , and ( [ b22 ] ) , we can write@xmath223 or@xmath224 by ( [ b3 ] ) .
therefore,@xmath225 in addition , with the help of ( [ b3 ] ) one gets@xmath226 which completes the proof .
b. acosta - humnez , s. i. kryuchkov , e. suazo , and s. k. suslov , _ degenerate parametric amplification of squeezed photons : explicit solutions , statistics , means and variances _ ,
j. nonlinear opt .
phys . mater .
* 24 * ( 2015 ) # 2 , 1550021 ( 27pp ) .
biaynicki - birula i and biaynicki - birula z _ the role of riemann - silberstein vector in classical and quantum theories of electromagnetism _ , j. phys .
_ _ * 46 * ( 2013 ) , 053001 ( 32 pp ) .
b. m. bolotovskii and s. n. stolyarov , _ current status of the electrodynamics of moving media ( infinite media ) _ , sov . phys.usp . * 17 * ( 1975 ) # 6 , 875895 ; see also a revised version with the annotated bibliography in : _ einstein s collection 1974 _ , pp .
179275 , nauka , moscow , 1976 [ in russian ] .
v. v. dodonov and v. i. manko , _ invariants and correlated states of nonstationary quantum systems _ , in : _ invariants and the evolution of nonstationary quantum systems _ , proceedings of lebedev physics institute , vol .
71181 , nauka , moscow , 1987 [ in russian ] ; english translation published by nova science , commack , new york , 1989 , pp . 103261 .
a. einstein and j. laub , _ ber die elektromagnetischen grundgleichungen fr bewegte krper _
, ann . phys .
* 26 * ( 1908 ) , 532540 ; _ bemerkungen zu unserer arbeit @xmath227@xmath227ber die elektromagnetischen grundgleichungen fr bewegte krper@xmath228@xmath228 _ , ann
* 28 * ( 1908 ) , 445447 .
v. l. ginzburg , _ the laws of conservation of energy and momentum in emission of electromagnetic waves ( photons ) in a medium and the energy - momentum tensor in macroscopic electrodynamics _ , sov
. phys.usp . * 16 * ( 1973 ) # 3 , 434439 .
w. israel and j. m. steward , _ progress in relativistic thermodynamics and electrodynamics of continous media _ , in : _ general relativity and gravitation ( one hundred years after the birth of albert einstein ) _
( a. held , ed . ) , vol .
2 , pp . 491525 , plenum press , new york , 1980 .
s. i. kryuchkov , s. k. suslov , and j. m. vega - guzmn , _ the minimum - uncertainty squeezed states for atoms and photons in a cavity _ , j. phys .
b : at . mol .
* 46 * ( 2013 ) , 104007 ( 15pp ) ; iop select and highlight of 2013 . h. a. lorentz , a. einstein , and h. minkowski , _ das relativittsprinzip , eine sammlung von abhandlungen _ , teubner , leipzig , berlin , 1913 ; english translation in : a. einstein et al , _ _ the principle of relativity , a collection of original papers on the special and general theory of relativity _
_ , 4th edn , dover , new york , 1932 .
h. minkowski , _ die grundlagen fr die elektromagnetischen vorgnge in bewegten krpern _
. knig . ges .
gttingen , math .- phys .
( 1908 ) , 53111 ; _ the fundamental equations for electromagnetic processes in moving bodies _ , english translation in : _ the principle of relativity _
( m. saha , ed . ) , university press , calcutta , 169 , 1920 . | we consider a complex covariant form of the macroscopic maxwell equations , in a moving medium or at rest , following the original ideas of minkowski . a compact , lorentz invariant , derivation of the energy - momentum tensor and
the corresponding differential balance equations are given .
conservation laws and quantization of the electromagnetic field will be discussed in this covariant approach elsewhere . |
the frustrated quantum antiferromagnets are a subject of great research interest .
the frustration arising due to competing interactions or lattice geometry of a system , helped further by the quantum fluctuations , acts unfavourably towards classical magnetic order , and can give rise to exotic quantum paramagnetic ground states @xcite .
many different materials are known to display frustrated magnetic behaviour in a variety of ways .
for instance ,
cugeo@xmath3 @xcite , srcu@xmath4(bo@xmath3)@xmath4 @xcite , zncu@xmath3(oh)@xmath5cl@xmath4 @xcite , and cav@xmath6o@xmath7 @xcite , are some of the very well known materials of this kind . on the theoretical side , there is a lot of interest in exploring and understanding the exotic behaviour in variously motivated frustrated quantum spin models . the simplest way in which the frustrated antiferromagnets exhibit quantum paramagnetism is through the formation of pairwise singlets , called dimers , in the ground state .
the pioneering work of majumdar and ghosh showed this exactly in a spin-1/2 heisenberg chain @xcite .
similarly , there is a strong numerical support for the spin-1/2 @xmath8-@xmath9 model on square lattice to have non - magnetic columnar dimer ground state in the range of @xmath10 @xcite .
the dimer ground states have been shown to occur in many other frustrated models on different lattices @xcite .
a dimer singlet ground state can also be realized by taking the antiferromagnetic exchange between the spins of selected pairs to be stronger than the rest .
for instance , in fig .
[ fig : stg_model_fig ] denoting a heisenberg model on square lattice , a staggered pattern of thick nearest - neighbour bonds ( @xmath11 ) is taken to be stronger compared to the other interactions ( @xmath12 and @xmath13 ) .
the spin-1/2 case of this staggered - dimer model without @xmath13 has been studied by using the perturbative series expansion method @xcite , exact diagonalization @xcite , coupled cluster method @xcite , and quantum monte carlo simulation @xcite .
these studies found the staggered dimer ground state for strong enough @xmath14 , which undergoes a transition to the nel state as @xmath14 becomes weaker . a case for topological order in this model
has also been made on the basis of tensor network studies @xcite .
while the studies concerned with dimer states focus mostly on spin-1/2 problems , darradi _ et al _ made an interesting investigation of their dependence on spin quantum number , @xmath0 , in the unfrustrated ( @xmath15 ) staggered - dimer model @xcite . using exact diagonalization , coupled - cluster method and variational mean - field approach , they studied the model for different values of @xmath0 , and tried to find how the critical coupling for the transition from staggered dimer to nel phase changes with @xmath0 . based on their calculations for upto @xmath16
, they inferred that @xmath17 .
the triplon analysis , based on the bond - operator representation of spins @xcite , is nice analytical method for studying the ordering instabilities of the dimer singlet states . using the spin-@xmath0 bond - operator representation of kumar @xcite , in this paper , we do a triplon analysis of the frustrated staggered - dimer model on square lattice ( fig . [
fig : stg_model_fig ] ) .
we find the ground state of this heisenberg model to undergo continuous phase transition from the staggered dimer to nel or collinear state , and the critical couplings for both the transitions to scale precisely as @xmath1 .
this lends support to the earlier finding for @xmath15 @xcite .
interestingly , for strong enough frustration ( measured by the relative strengths of @xmath18 and @xmath19 ) , we also find the staggered dimer phase to survive in the large @xmath0 ( classical ) limit .
it looks a bit unusual , but we know of one other case ( the famous shastry - sutherland model @xcite ) in which the dimer ground state is exactly realized for arbitrarily large spins .
thus , by helping the singlet phase to survive in the large @xmath0 limit , the frustration seems to be differentiating the classical limit of a quantum case from the classical case itself .
but we also find that the expected classical behaviour is restored in the large @xmath0 limit even by an infinitesimally small axial single - ion anisotropy ( that is ever - present4sc(nh@xmath4)@xmath4 ] @xcite , nenp[ni(c@xmath4h@xmath21n@xmath4)2ho@xmath4(clo@xmath6 ) ] @xcite , nenc [ ni(c@xmath4h@xmath21n@xmath4)@xmath4ni(cn@xmath6 ) ] @xcite , csnicl@xmath3 @xcite , y@xmath4banio@xmath22 @xcite , pbni@xmath4v@xmath4o@xmath21 @xcite , and lifepo@xmath6 @xcite _ etc_. ] ) . through this study of the @xmath0 dependence of quantum phase transitions in the staggered - dimer model
, we present an interesting perspective on the change of behaviour from quantum to classical as the spin quantum number becomes larger .
this paper is organized as follows . in sec .
[ a quantum spin-@xmath0 model ] , we state the staggered - dimer model on square lattice , and carry out the mean - field triplon analysis of its purely heisenberg case . next , in sec .
[ model with single - ion anisotropy ] , we study the effect of single - ion anisotropy with increasing @xmath0 .
we then conclude with a summary in sec . [
summary : ] .
the hamiltonian of the spin-@xmath0 staggered - dimer model on square lattice depicted in fig .
[ fig : stg_model_fig ] can be written as follows .
@xmath23\end{aligned}\ ] ] as dark solid lines , @xmath12 as light solid lines , and @xmath13 as dashed lines .
the shaded region ( on the right side ) is the first brillouin zone of the staggered - dimer square lattice.,title="fig:",height=240 ] as dark solid lines , @xmath12 as light solid lines , and @xmath13 as dashed lines .
the shaded region ( on the right side ) is the first brillouin zone of the staggered - dimer square lattice.,title="fig:",height=240 ] here , @xmath24 is the position of a dimer , and the labels @xmath25 and @xmath26 denote the spins of a dimer .
the @xmath27 is the intra - dimer exchange interaction , @xmath12 is the nearest inter - dimer interaction and @xmath28 is the second nearest inter - dimer interaction .
all of these exchange interactions are antiferromagnetic .
the strength of the single - ion anisotropy is given by @xmath29 , which is taken to be a positive number . without the anisotropy term , for @xmath30 and a non - zero @xmath31 ,
this model is same as the @xmath8-@xmath9 heisenberg model on square lattice ( with @xmath32 and @xmath33 ) .
it has the nel ordered ground state when @xmath15 or sufficiently small .
however , for @xmath34 , it would form the staggered - dimer singlet ground state by suppressing the nel order . here , the dimerization is not spontaneous , but induced by the greatly stronger @xmath27 . besides a strong @xmath11 , the frustration due to competing @xmath12 and @xmath31 also facilitates disorder in the ground state .
but the anisotropy , @xmath29 , tends to support magnetic order .
clearly , here there is an ample scope for competition between the quantum disordered staggered - dimer phase and the classical nel or collinear ordered phases .
we study this competition as a function of the spin quantum number , @xmath0 , by doing the stability analysis of the staggered - dimer phase using of the bond operators .
the heisenberg case of @xmath35 with @xmath36 is discussed in the following subsections , and the model for @xmath37 is studied in detail in the next section .
the systems where the pairs of spins ( dimers ) act as basic units , the bond - operator representation of spins has been found to be very useful in doing calculations .
this approach was pioneered by sachdev and bhatt for studying quantum phase transitions in dimerized spin-1/2 quantum antiferromagnets @xcite . for the case of general spin-@xmath0 ,
the bond operator representation was derived by kumar @xcite .
the basic idea is to construct the spin states on a dimer , and associate with each one of them a distinct bosonic creation operator called bond operator . for a pair of spin-@xmath0 interacting via exchange interaction , @xmath38 ,
the suitable dimer states are the total - spin eigenstates of the pair .
that is , the spin singlet : @xmath39 , triplet : @xmath40 , quintet : @xmath41 , and so on .
the corresponding bond operators are defined as follows : @xmath42 here , @xmath43 denotes the vacuum of the bosonic fock space . in order to describe the finite - dimensional physical hilbert space of the dimer
, these bond operators are required to satisfy the constraint : @xmath44 , where ` @xmath45 ' is meant to denote the contributions from all the higher total - spin dimer states upto @xmath46 .
the exchange hamiltonian of the dimer can now be written as : @xmath47 \nonumber \\ & & + j^{\prime}\left[\hat{t}^{\dagger}_{m_{1}}\hat{t}_{m_{1}}+3 \qhat^{\dagger}_{m_{2}}\qhat_{m_{2}}+\cdots \right].\end{aligned}\ ] ] for the interactions between the spins of different dimers , one needs the bond - operator representation of the individual spin operators . at the very least
, it can constructed in the restricted space of singlet and triplet , and can also be extended further to include quintet or higher states , if necessary .
the bond operator representation of @xmath48 and @xmath49 in the restricted space of singlet and triplet has the following form @xcite .
@xmath50 here , @xmath51 and likewise for @xmath52 and @xmath53 , and @xmath54 is the totally antisymmetric tensor . the staggered - dimer _ heisenberg _ model , @xmath35 of eq .
( [ stg_mdl ] ) with @xmath36 , in the bond - operator representation turns into a model of interacting triplets and singlets , which is hard to make progress with . to simplify it
, we make the following standard approximations @xcite .
we treat the singlet bond - operators ( for the staggered singlet bonds of fig .
[ fig : stg_model_fig ] ) in the mean - field approximation as : @xmath55@xmath56@xmath57@xmath56@xmath58 .
it now becomes a model of interacting triplet excitations with respect to the mean singlet background . to keep it simple
, we also neglect the interaction between triplets , and impose the local constraint on the bond - operators through an average lagrange multiplier , @xmath59 .
the resulting hamiltonian is a soluble bilinear problem of the triplet bond - operators , referred to as _
triplons_. finally , by doing the fourier transformation : @xmath60 , we get the following effective hamiltonian of the triplon dynamics in the momentum space . @xmath61 \nonumber\\ & & + \frac{1}{2}\sum_{\bf{k},\alpha}\bigg\{\big[\lambda-\bar{s}^2 s(s+1 ) \xi_{\bf{k } } \big]\left(\hat t^\dag_{\bf{k}\alpha } \hat t_{\bf{k } \alpha } + \hat t_{-\bf{k}\alpha}\hat
t^\dag_{-\bf{k}\alpha}\right ) \nonumber \\ & & - \bar{s}^2 s(s+1)\xi_{\bf{k } } \left(\hat t^\dag_{\bf{k}\alpha } \hat t^\dag_{-\bf{k}\alpha } + \hat t_{-\bf{k}\alpha}\hat t_{\bf{k}\alpha}\right ) \bigg\ } .\end{aligned}\ ] ] here , the wavevector @xmath62 lies in the brillouin zone shown in fig . [
fig : stg_model_fig ] , @xmath63 is the total number of dimers , and @xmath64/3 $ ] .
the triplon hamiltonian , @xmath65 , can be diagonalized by the bogoliubov transformation for bosons , @xmath66 , with @xmath67\}$ ] .
the @xmath65 in terms of the quasiparticle operators , @xmath68 , has the following diagonal form .
@xmath69 here , @xmath70 , and @xmath71}\ge0 $ ] is the triplon quasiparticle dispersion .
the ground state energy per dimer is given by @xmath72&=&e_{0}+\frac{3}{2n_{d}}\sum_{\bf{k } } e_{\bf{k}}. \end{aligned}\ ] ] minimization of the ground state energy density , @xmath73 , with respect to the mean - field parameters , @xmath74 and @xmath75 , gives the following equations .
@xmath76 we determine @xmath75 and @xmath59 by self - consistently solving the eqs .
( [ sbar ] ) and ( [ lambda ] ) in the space of @xmath19 and @xmath18 . in the staggered - dimer phase ,
the self - consistently determined quasiparticle gap is non - zero .
however , for some critical values of @xmath19 and @xmath18 , this gap would vanish .
the contours of such critical points define the boundaries of the staggered - dimer phase . on the other side of this phase boundary
, the condensation of triplons at wavevector @xmath77 , given by @xmath78 , determines the antiferromagnetic order .
the vanishing of the gap also fixes the value of @xmath59 at @xmath79 .
for the purpose of calculation , we set @xmath80 , and vary @xmath12 and @xmath13 as free parameters . in the present case ,
we now describe the quantum phase diagram of the staggered - dimer heisenberg model on square lattice in the @xmath12-@xmath13 plane , as obtained from the mean - field triplon analysis . from our calculations , we find two different ordering wavevectors at which the triplon gap vanishes .
it is @xmath81 , causing nel order for sufficiently strong @xmath12 , and the other is @xmath82 that gives collinear order for strong enough @xmath13 .
we determine the phase - boundaries between the staggered - dimer and two ordered phases from the following equation . @xmath83&=&\frac{3}{2s(s+1)}.\label{sifnty}\end{aligned}\ ] ] the quantum phase diagram , thus obtained , is presented in fig .
[ fig : phs_s ] . in eq .
( [ sifnty ] ) , the interaction parameters , @xmath12 and @xmath13 , and the spin quantum number , @xmath0 , appear separately on the left and right hand side , respectively .
it clearly implies that the phase boundaries for all the different values of @xmath0 ( in the left panel of fig .
[ fig : phs_s ] ) will collapse onto two lines only ( one for dimer to nel and the other for dimer to collinear transition ) in the space of rescaled interactions , @xmath84 and @xmath85 , as shown in the right panel of fig .
[ fig : phs_s ] ) .
-@xmath13 plane ( for @xmath86 and @xmath36 ) .
the region above the upper phase boundaries is the collinear phase , below the lower phase boundary is the nel phase , and between the two is the staggered - dimer phase . for @xmath0@xmath56@xmath87 ,
the phase boundaries are @xmath31@xmath56@xmath88 and @[email protected]@xmath12 for nel and collinear transition , respectively .
right : the quantum phase diagram in the space of rescaled parameters.,title="fig:",width=297 ] -@xmath13 plane ( for @xmath86 and @xmath36 ) .
the region above the upper phase boundaries is the collinear phase , below the lower phase boundary is the nel phase , and between the two is the staggered - dimer phase . for @xmath0@xmath56@xmath87 ,
the phase boundaries are @xmath31@xmath56@xmath88 and @[email protected]@xmath12 for nel and collinear transition , respectively .
right : the quantum phase diagram in the space of rescaled parameters.,title="fig:",width=297 ] for @xmath89 , the quantum critical point for staggered - dimer to nel transition is given by @xmath90 .
for @xmath91 , the critical @xmath13 for transition to the collinear phase is @xmath92 . for @xmath15
, we can compare the results of our calculations with those of others obtained by different methods .
for @xmath93 , we get @xmath94 , while it is around @xmath95 from other numerical calculations @xcite .
take the values of darradi _ et al _
@xcite for @xmath96 @xmath97@xmath98 against ours of @xmath99 ( 0.5188 , 0.1945 , 0.1038 ) for @xmath100 and @xmath101 , respectively .
the agreement here is quite reasonable , and seems to get better as @xmath0 increases .
moreover , in the limit @xmath102 , when either @xmath12 or @xmath31 is zero ( unfrustrated cases ) , the ordered classical phases are the only possibility .
however , when both @xmath12 and @xmath13 are non - zero , the large @xmath0 limit turns out to be surprising , as discussed below . in the limit @xmath103 , the right hand side of eq .
( [ sifnty ] ) vanishes , which then gives two limiting phase boundaries : @xmath104 for staggered - dimer to nel transition , and @xmath105 for the transition to collinear phase .
a small difference in their slopes leaves a tiny region of staggered - dimer phase between two limit phase boundaries ( see fig . [
fig : phs_s ] ) .
it presents an interesting case of a heisenberg problem in which a quantum mechanical ( staggered - dimer ) phase , helped by strong frustration , survives the approach to classical limit .
it is possible that this tiny region eventually vanishes , if the corrections due to higher dimer - states such as quintets _ etc .
_ are properly included .
one natural and simple way by which the higher dimer states , say quintet , can participate in the effective low - energy dynamics is through the single - ion anisotropy , @xmath29 . in the next section ,
we investigate how @xmath29 helps in fully establishing the classical order in the limit @xmath103 .
before studying the staggered - dimer model including @xmath29 , let us first understand a single dimer with axial spin anisotropy .
consider the following hamiltonian for such a dimer . @xmath106
the presence of a non - zero @xmath29 causes splitting and mixing between the total - spin states of a dimer .
for instance , now the singlet mixes with the triplet as well as the quintet states .
this we can see by writing @xmath107 and @xmath108 in the less restricted ( compared to last section ) subspace of singlet , triplet and quintet states @xcite .
thus , in the bond - operator representation , @xmath109 and @xmath110 therefore , @xmath111 where @xmath112 , @xmath113 , and @xmath114 are normalization constants . notice the mixing between @xmath115 and @xmath116 in eq .
( [ eq : d - dimer ] ) , which in addition to the triplons , will most directly affect the dimer singlet state .
hence , we now write @xmath117 of eq .
( [ dimerd ] ) in terms of @xmath115 , @xmath118 and @xmath119 only , and ignore all the other bond - operators for simplicity . in this approximation , the @xmath117 takes the following form .
@xmath120 where @xmath121 .
the mixing of @xmath115 with @xmath119 in eq .
( [ bndhml ] ) can be diagonalized by unitary rotation of @xmath115 and @xmath119 to new operators @xmath122 and @xmath123 such that @xmath124 and @xmath125 . here ,
@xmath126}$ ] .
the @xmath117 in terms of the bond operators @xmath122 , @xmath118 and @xmath123 has the following diagonal form .
@xmath127\nonumber\\ & & + \frac{1}{2}\left(-\tilde{s}^{\dagger}\tilde{s}+\tilde{q}^{\dagger}_{0}\tilde{q}_{0}\right ) \times \left(\frac{2d}{\sin 2\theta}\sqrt { \frac{2{\tt n_q}}{3{\tt n_s}}}\right)\nonumber \\ & & -2d \left({\frac{{\tt n}_{t}}{2 { \tt n}_s}}+{\frac{{\tt n}_q}{3{\tt n}_t}}\right)\that^{\dagger}_z\that_z-\frac{d}{2}\left({\frac{{\tt n}_q}{{\tt n}_t}}+1\right)\that^{\dagger}_{\alpha}\that_{\alpha}. \label{dgnlzdbndhml}\end{aligned}\ ] ] @xmath561(left ) and for @xmath0@xmath565 ( right).,title="fig : " ] @xmath561(left ) and for @xmath0@xmath565 ( right).,title="fig : " ] the eigenvalues of the diagonalized dimer hamiltonian of eq
. ( [ dgnlzdbndhml ] ) as a function of @xmath29 are plotted in fig .
[ fig : energy_s1 ] . as the quintet state , @xmath128 , is much higher in energy , for a low - energy effective description , we can ignore it .
moreover , in the mean - field approximation @xmath129 is replaced by @xmath58 .
therefore , we get @xmath130\nonumber\\ & & + \left[-j^{\prime}s(s+1)+j^{\prime}+\frac{d}{2}\left({\frac{{\tt n}_q}{{\tt n}_t}}+1\right)\right]\hat{t}^{\dagger}_{\alpha}\hat{t_{\alpha}}\nonumber\\ & & + \left[-j^{\prime}s(s+1)+j'+2d\left({\frac{{\tt n}_t}{2{\tt n}_s}}+\frac{{\tt n}_q}{{\tt n}_t}\right)\right]\that^{\dagger}_z\that_z . \label{dgnlzdbndhml_fnl}\end{aligned}\ ] ] the bond - operator representation for @xmath131 and @xmath132 is also required to be reconstructed for this case . from ref .
@xcite , we know that @xmath131 and @xmath132 in the restricted space of @xmath133 , @xmath134 , @xmath135 , @xmath136 , can be written as follows .
@xmath137 after taking care of the mixing between @xmath115 and @xmath116 , and then using the mean - field approximation for @xmath129 operator and ignoring the @xmath138 excitations , we get the following expressions for @xmath131 and @xmath132 .
@xmath139 here , @xmath121 . using this representation
, now we can do the standard triplon analysis of the full staggered - dimer model of eq .
( [ stg_mdl ] ) including the spin anisotropy , @xmath29 .
the triplon hamiltonian for the @xmath35 of eq .
( [ stg_mdl ] ) , including single - ion anisotropy has the following form .
@xmath141 where @xmath142 , and the parameters @xmath143 and @xmath144 as functions @xmath0 , @xmath11 and @xmath29 are defined as , @xmath145 and @xmath146 . similarly , the spin anisotropy induced split effective chemical potentials are : @xmath147 , and , @xmath148 , and @xmath149 .
\nonumber\end{aligned}\ ] ] after bogoliubov diagonalization , we get @xmath150 here , @xmath151@xmath56@xmath152 for @xmath121 and @xmath153@xmath56@xmath154 are the quasiparticle dispersions .
the ground state energy per dimer is @xmath155&=&e^d_{0}+\frac{1}{2n_d}\sum_{{\bf k } , \alpha } e_{{\bf k \alpha}}+\frac{1}{2n_d}\sum_{{\bf k } } e_{{\bf k } z}.\end{aligned}\ ] ] minimizing , @xmath156 with respect to @xmath74 and @xmath75 gives the following mean - field equations .
( top - left ) , for @xmath157 ( top - right ) , for @xmath16 ( bottom - left ) , and for @xmath158 ( bottom - right ) .
the boundaries of the nel and collinear phases merge into a single line at @xmath31@xmath56@xmath159 for strong enough single - ion anisotropy ( @xmath29).,title="fig:",width=288 ] ( top - left ) , for @xmath157 ( top - right ) , for @xmath16 ( bottom - left ) , and for @xmath158 ( bottom - right ) .
the boundaries of the nel and collinear phases merge into a single line at @xmath31@xmath56@xmath159 for strong enough single - ion anisotropy ( @xmath29).,title="fig:",width=288 ] + ( top - left ) , for @xmath157 ( top - right ) , for @xmath16 ( bottom - left ) , and for @xmath158 ( bottom - right ) .
the boundaries of the nel and collinear phases merge into a single line at @xmath31@xmath56@xmath159 for strong enough single - ion anisotropy ( @xmath29).,title="fig:",width=288 ] ( top - left ) , for @xmath157 ( top - right ) , for @xmath16 ( bottom - left ) , and for @xmath158 ( bottom - right ) .
the boundaries of the nel and collinear phases merge into a single line at @xmath31@xmath56@xmath159 for strong enough single - ion anisotropy ( @xmath29).,title="fig:",width=288 ] @xmath160 \nonumber \\ & & + { \frac{1}{2n_d}}\sum_{{\bf k } , \alpha}\frac{\lambda_{\alpha}\bar{s}^2 \phi_\alpha \xi_{{\bf k}}}{e_{{\bf k \alpha } } } + { \frac{1}{2n_d}}\sum_{{\bf k}}\frac{\lambda_z \bar{s}^2 \phi_z \xi_{{\bf k}}}{e_{{\bf k } z } } \label { dmu}\end{aligned}\ ] ] to generate the boundaries of the gapped staggered - dimer phase , we track the quasiparticle gap by solving eqs .
( [ dsbar])-([dmu ] ) self - consistently for @xmath59 and @xmath161 .
below we discuss the quantum phase diagram thus generated .
the closing of triplon gap at @xmath163 = @xmath164 in the nel phase , and at @xmath163 = @xmath165 in the collinear phase occurs through @xmath166 , which fixes @xmath167 at @xmath168 . for positive @xmath29
, the @xmath169 always remains gapped for @xmath121 .
recall that , for @xmath36 , we always found the dimer phase , even for @xmath103 ( classical limit ) .
but now , for @xmath170 , we find that , above some critical value , @xmath171 at @xmath31@xmath56@xmath159 , the nel and collinear phase boundaries merge into a single line for any value of @xmath0 , and the staggered - dimer phase completely vanishes . in fig .
[ mergeline ] , we show this for @xmath172 , @xmath157 , @xmath16 and @xmath158 .
thus , a sufficiently strong @xmath29 will always make the dimer phase disappear for any @xmath173 .
we estimate the critical values of @xmath29 for first few values of @xmath0 , which are @xmath174 for @xmath172 , @xmath175 for @xmath157 , @xmath176 for @xmath16 , and @xmath177 for @xmath158 .
we note that for large @xmath0 , the critical @xmath171 becomes negligibly small , and the ground state is always a classical .
in fact , @xmath178 for large @xmath0 , as estimated from the data for upto @xmath179 in fig . [
fig : d_vs_s(s+1 ) ] .
this calculation clearly shows that , in the limit @xmath102 , an infinitesimal amount of single - ion anisotropy would in suffice to completely suppress the quantum mechanical dimer phase , and to fully establish classical order . as a function of spin quantum number @xmath0 from @xmath172 upto @xmath179 .
right : the log - log plot of the same data.,title="fig : " ] as a function of spin quantum number @xmath0 from @xmath172 upto @xmath179 .
right : the log - log plot of the same data.,title="fig : " ]
here , we studied the quantum phase transitions in a frustrated staggered - dimer model for arbitrary spin-@xmath0 on square lattice by doing triplon mean - field theory , without and with single - ion anisotropy . in the purely heisenberg case , we found two ordered phases , viz .
nel and collinear states , sandwiching the spin - gapped staggered - dimer phase .
the critical couplings for these transitions are found to scale precisely as @xmath1 .
interestingly , we also find that this quantum ( dimer ) phase , helped by strong frustration , persists even in the classical limit ( @xmath102 ) . in the presence of single - ion anisotropy
, we first reworked the spin-@xmath0 bond - operator representation to include the effects of singlet - quintet mixing on a dimer .
then , we used it to carry out the standard triplon analysis to work out the quantum phase diagram .
interestingly , for any @xmath180 , we always find a critical value of the anisotropy , @xmath171 , for which the dimer phase between the nel and collinear phases vanishes .
we also find that @xmath178 for large @xmath0 .
thus , an infinitesimally small @xmath29 would be sufficient to kill the quantum mechanical dimer phase and establish classical order in the limit of very large @xmath0 . through this model calculation
, we presented here an interesting perspective on the change of behaviour from quantum to classical in frustrated spin systems with increasing @xmath0 .
it would to nice to check this in other models , and if possible , by emulating similar models using cold atoms in optical lattices .
bimla danu thanks csir ( india ) for financial support during her phd when most of this work was done .
brijesh kumar acknowledges financial support under the upe - ii and dst - purse programs of jnu , new delhi .
50 c. lacroix , p. mendels and f. mila , springer series in solid - state sciences , _ introduction to frustrated magnetism _ , * 164 * , ( 2011 ) .
m. hase et al .
lett . * 70 * , 3651(1993 ) .
kageyama et al . , phys .
lett . * 82 * , 3168(1999 ) .
helton et al . , phys .
lett . * 98 * , 107204(2007 ) .
bert et al . ,
journal of physics : conference series , * 145 * , 012004(2009 ) .
taniguchi et al . , j. phys .
. jpn . * 64 * , 2758(1995 ) . c. k. majumdar and d. k. ghosh j. math
. phys . * 10 * , 1399(1969 ) .
s. chakravarty , b. i. halperin and d. r. nelson , phys .
b * 39 * , 2344(1989 ) .
s. sachdev and r. n. bhatt , phys .
b * 41 * , 9323 ( 1990 ) .
b. kumar , phys .
b * 82 * , 054404(2010 ) .
d. poilblanc , e. gagliano , s. bacci and e. dagotto , phys . rev .
b * 43 * , 10970(1991 ) .
j. oitmaa and z. weihong , phys .
b * 54 * , 3022(1996 ) .
b. s. shastry and b. sutherland , physica b. * 108 * , 1069(1981 ) .
d. j. klein , j. phys a : math .
gen . * 15 * , 661(1982 ) .
i. bose and p. mitra , phys . rev .
b * 44 * , 443(1991 ) .
b. kumar , phys .
b * 66 * , 024406(2002 ) .
b. danu , b. kumar and r. v. pai , epl * 100 * , 27003 ( 2012 ) .
b. kumar , phys .
b * 71 * , 174434(2005 ) .
a. gelle , a. m. lauchli , b. kumar and f. mila , phys .
b * 77 * , 014419(2008 ) .
r. kumar and b. kumar , phys .
b * 77 * , 144413(2008 ) .
r. kumar , d. kumar and b. kumar , phys .
b * 80 * , 214428(2009 ) .
r. p. rajiv singh , p. martin gelfand and a. david huse , phys .
lett . * 61 * , 2484(1988 ) .
kr@xmath181ger , j. richter , d. j. j. farnell and r. f. bishop , phys .
b * 61 * , 14607(2000 ) . s. wenzel and l. bogacz and w. janke , phys .
* 101 * , 127202(2008 ) .
j . jiang , phys .
b * 85 * , 014414(2012 ) . ching- yu huang and feng - li lin , phys .
b * 84 * , 125110(2011 ) .
r darradi and j richter and d j j farnell , j. phys . : condens .
matter * 17 * , 341(2005 ) .
s. a. zvyagin et al . , phys .
98 * , 047205(2007 ) .
j. p. renard et al . , j. appl
* 63 * , 3538(1988 ) .
m. orend et al . , phys .
b * 52 * , 3435(1995 ) .
w. j. l. buyers et al .
lett . * 56 * , 371(1986 ) .
guangyong xu et al . ,
b. * 54 * , r6827(1996 ) .
a. zheludev et al . , phys .
b. * 62 * , 13(2000 ) .
jiying li et al .
b. * 73 * , 024410(2006 ) .
alexey v. gorshkov et al .
lett . * 107 * , 115301(2011 ) .
s. gopalakrishnan , benjamin l. lev and p. m. goldbart , phys .
* 107 * , 277201(2011 ) .
j. simon et al . ,
nature * 472 * , 307(2011 ) .
joseph w. britton et al . ,
nature * 484 * , 489(2012 ) . | we study a frustrated spin-@xmath0 staggered - dimer heisenberg model on square lattice by using the bond - operator representation for quantum spins , and investigate the setting in of classical magnetic order with increasing @xmath0 , starting from the staggered - dimer singlet ground state .
we find the critical coupling for this quantum phase transition to scale as @xmath1 .
but , for strong enough frustration , we also find the quantum mechanical dimer phase to survive even for large @xmath0 ( classical limit ) .
the inclusion of single - ion anisotropy , however , expectedly suppresses the quantum phase , and helps in establishing the classical order ( nel or collinear , in the present study ) above a critical value that goes as @xmath2 $ ] for large @xmath0 .
thus , in the large @xmath0 limit , an infinitesimal amount of anisotropy would suffice to realize the classical behaviour .
* s * gbk october 12 , 2016 |
it has long been known that particles with non - zero spin can have `` toroidal '' electromagnetic form factors that are odd under charge conjugation @xmath2 , which implies that they violate either parity ( @xmath3 ) or time reversal ( @xmath4 ) , but not both symmetries simultaneously @xcite . the toroidal dipole form factor ( tdff ) , also called the anapole @xcite , requires spin 1/2 or higher , violates @xmath3 and conserves @xmath4 . the toroidal quadrupole form factor ( tqff ) , which requires spin 1 or higher , violates @xmath4 and conserves @xmath3 , and so on @xcite .
toroidal form factors produce no physical effects when the photon is on - shell , and correspond in a classical picture to fields within the charge distribution @xcite .
these features contrast with the more familiar @xmath2-even electric and magnetic form factors , which respect or violate both @xmath3 and @xmath4 simultaneously , and produce effects for on - shell photons .
the only form factors allowed for massive particles that are their own antiparticles are toroidal @xcite .
the toroidal form factors do contribute to the short - range interaction with a charged particle .
for nucleons and nuclei , in particular , they are in principle accessible via lepton scattering .
while there exist calculations of the tdffs of the nucleon and nuclei , there is apparently no calculation of a nuclear tqff .
the tqff of positronium was calculated in ref .
@xcite .
the aim of this paper is to provide the first controlled calculation of the tqff of the simplest nucleus , the deuteron , at low momentum .
the lorentz - covariant electromagnetic current of a particle with spin @xmath5 is described by seven electromagnetic form factors : charge , magnetic dipole , and electric quadrupole , which are @xmath3- and @xmath4-conserving ( @xmath6 ) ; electric dipole and magnetic quadrupole , which are @xmath3- and @xmath4-violating @xmath7 ; tdff , which is @xmath3-violating and @xmath4-conserving ( @xmath8 ) ; and , finally , tqff , which is @xmath3-conserving and @xmath4-violating @xmath9 .
we can write the spatial , @xmath10 component of the electromagnetic current as @xcite @xmath11 f_{p { \slash\hspace{-0.4em}t}}(\vec q^{\ , 2 } ) , \label{pctvff}\ ] ] where @xmath12 is a deuteron state with momentum @xmath13 and polarization @xmath14 in the rest frame , normalized so that @xmath15 , @xmath16 is the ( outgoing ) momentum of the photon , @xmath17 is the deuteron mass in terms of the nucleon mass @xmath18 mev and the binding momentum @xmath19 mev .
@xmath20 is the tqff , which is proportional to the proton charge @xmath21 and has dimensions of mass@xmath22 .
we express it in units of @xmath23 @xmath24 .
we denote the corresponding toroidal quadrupole moment ( tqm ) by @xmath25 . it can be viewed as an interaction of the deuteron @xmath26 with the second derivative of the magnetic field @xmath27 of the form @xmath28 where @xmath29 denotes the deuteron spin , and @xmath30 the anticommutator . using maxwell s equations to replace the curl of the magnetic field with a current
, we can trade the @xmath10 moment for a contact interaction .
for example , the @xmath31 interaction of a non - relativistic lepton of mass @xmath32 with the deuteron becomes a dimension - eight contact interaction , @xmath33.\ ] ] the first term is due to the lepton kinetic term and gives rise to a non - local interaction involving @xmath34 .
the second one comes from the interaction of the lepton spin @xmath35 with the deuteron @xmath31 form factor .
effects of a tqff on polarization observables in lepton - deuteron scattering have been investigated @xcite .
there should be similar effects in proton - deuteron scattering such as in the planned tric experiment at cosy @xcite , but there they are likely swamped by non - electromagnetic interactions .
we work in the framework of chiral effective field theory ( eft ) and take into account the dominant parity and time - reversal violation in and beyond the standard model ( sm ) of particle physics . @xmath3
violation is commonplace in the weak interaction of the sm .
@xmath4 violation , on the other hand , is small in the sm , which opens up the possibility that operators involving the sm fields but having dimension larger than four could be noticeable .
@xmath4 violation from the ckm quark - mixing matrix is suppressed with respect to other aspects of weak interactions by a small combination of matrix elements @xcite , @xmath36 .
moreover , it is loop suppressed in flavor - conserving quantities , such as @xmath4-violating form factors of the nucleon and nuclei .
this leaves the qcd vacuum angle @xmath37 @xcite as the potentially largest dimension - four source of such form factors .
however , the stringent experimental limit on the neutron electric dipole moment , @xmath38 fm @xcite , constrains it to @xmath39 .
therefore , we also consider @xmath4 violation originating beyond the sm at a high energy scale @xmath40 .
the dominant such higher - dimensional @xmath4-violating operators are of effective dimension six .
the tqff is in principle sensitive to @xmath31 physics beyond the sm .
however , the lowest dimension where we find @xmath31 operators is eight , which means that , in the simplest scenarios , they would be highly suppressed by the presumably high scale of physics beyond the sm .
discussions and references on @xmath31 interactions at low energies , including situations where they could be relatively enhanced , can be found in ref .
we focus here on what is likely to be the largest `` background '' in the deuteron tqff : the combination of @xmath8 from the ordinary weak interactions with @xmath41 from the @xmath37 term and from the dimension - six operators .
not surprisingly , we find a very small background value for the deuteron tqff , so that any experimental evidence for a nonzero tqff likely results from new @xmath31 physics .
our discussion is organized as follows . in section [ interactions ] we construct the effective chiral lagrangian for the relevant @xmath6 , @xmath8 , and @xmath41 interactions and currents involving nucleons , pions , and photons . in section [ calculation ]
we calculate the long - range contributions of these interactions to the deuteron tqff . in section [ discussion ]
we discuss our results and compare the deuteron tqm to its @xmath41 electric dipole moment ( edm ) and magnetic quadrupole moment ( mqm ) .
three appendices are devoted to details of our calculations . in appendix
[ app : dimsix ] the various @xmath41 operators are presented in more detail , and the orders of magnitude of their contributions are given in appendix [ app : ordmag ] . in appendix
[ app : ffs ] we give the expansion of loop diagrams that define the deuteron tqff .
at a momentum @xmath42 much below the characteristic qcd scale , @xmath43 gev , electromagnetic form factors can be calculated with low - energy effective field theories ( efts ) .
the most predictive such an eft is chiral eft ( for a review , see ref .
@xcite ) , a generalization to an arbitrary number of nucleons of chiral perturbation theory ( chpt ) ( for a review , see ref .
@xcite ) , where @xmath44 , with @xmath45 mev the pion mass . in this eft pion propagation
is included explicitly , and the properties and interactions of the pions are strongly constrained by the approximate chiral symmetry of qcd . for the nucleon , form factors can be calculated in perturbation theory as a systematic expansion in @xmath46 @xcite .
the @xmath8 anapole and the @xmath41 electric dipole form factor of the nucleon have been calculated to next - to - leading order ( nlo ) in refs . @xcite and @xcite , respectively . in nuclei , pions can still be treated in perturbation theory @xcite , but then the expansion is in powers of @xmath47 , where @xmath48 in terms of the pion decay constant @xmath49 mev . for observables involving momenta above @xmath50 , one - pion exchange needs to be iterated to all orders @xcite , which complicates renormalization @xcite .
however , light nuclei are dilute systems and , unless one is interested in form factors at high momentum , one can use a chiral eft with perturbative pions .
indeed , the @xmath2-even electromagnetic form factors of the deuteron , both @xmath6 ( charge , electric quadrupole , and magnetic dipole ) @xcite and @xmath41 ( electric dipole and magnetic quadrupole ) @xcite have been successfully derived in this eft .
the tdff of the deuteron has been calculated at lo in ref .
similar calculations could be performed for other light nuclei .
the relevant low - energy eft can be written in terms of nucleon , pion , and photon fields .
the nucleon field @xmath51 is an isospinor bi - spinor , with isospin @xmath52 and spin @xmath53 in the rest frame , where the velocity is @xmath54 .
the pion field @xmath55 is an isovector pseudoscalar , for which we choose a stereographic parametrization ( see , _ e.g. _ , ref .
@xcite ) of the coset space @xmath56/@xmath57 , where @xmath58 is the spontaneously broken , approximate chiral symmetry of qcd , and @xmath59 its unbroken isospin subgroup .
we define @xmath60 .
the photon field @xmath61 ensures electromagnetic @xmath62 gauge invariance , appearing in the gauge and chiral covariant derivatives @xmath63 and @xmath64n$ ] , and in the field strength @xmath65 .
we use the notation @xmath66 , where @xmath67 and @xmath68 .
the coefficients of interactions constructed with up to two nucleon fields are estimated , in the absence of other information from qcd , by naive dimensional analysis ( nda ) @xcite .
for multi - nucleon couplings the scaling of a coefficient on the various scales depends also on the number of @xmath29 waves the operator connects @xcite . in the following
we will need only a few terms in the leading pion - nucleon - photon @xmath6 chiral lagrangians , _ viz .
_ @xmath69 where @xmath70 is the nucleon axial coupling and @xmath71 a contact two - nucleon parameter , and @xmath72\right\ } s^\nu n + \ldots , \nonumber\\ \label{subptlag}\end{aligned}\ ] ] where @xmath73 and @xmath74 are , respectively , the isoscalar and isovector anomalous magnetic moments of the nucleon , and @xmath75 .
the @xmath31 tqff vanishes unless there is , in the eft , either a @xmath31 interaction or a combination of @xmath8 and @xmath41 interactions between the two nucleons .
@xmath31 operators in the eft lagrangian arise in two ways .
first , they represent dimension - seven @xmath31 operators in the quark - gluon lagrangian just above @xmath76 .
these dimension - seven operators in turn can have two origins above the electroweak scale @xmath77 .
on one hand , they can be generated by possible gauge - invariant dimension - eight @xmath31 operators , in which case they would be expected to be suppressed by four powers of the high , new - physics scale @xmath40 , that is , they would scale as @xmath78 . on the other hand
, they can arise from the interplay of @xmath8 in the sm and possible dimension - six @xmath41 operators , when one would expect the suppression scale to be @xmath79 rather than @xmath80 .
a second way to generate @xmath31 operators in the eft lagrangian is from @xmath8 and @xmath41 interactions in the quark - gluon lagrangian at low energy , when we integrate out non - perturbative dynamics on scale of order of the typical hadronic scale @xmath76 .
again here we expect a suppression of @xmath79 rather than @xmath80 .
if the new - physics scale is much higher than the electroweak scale , the contributions from @xmath8 and @xmath41 interactions are likely to dominate the @xmath31 interactions in the eft .
interesting scenarios in which this is not the case are discussed in ref .
here we are interested in the background to genuine @xmath31 interactions at the high energy scale . in this case , as discussed in app .
[ app : ordmag ] , the contributions from @xmath31 interactions in the eft are likely smaller than the long - range components from @xmath8 and @xmath41 interactions , which we can , and will , calculate .
@xmath8 interactions in chiral eft have been discussed for example in refs .
they originate at the qcd scale from four - quark interactions proportional to the fermi constant @xmath81 gev@xmath82 .
a dimensionless measure of the relative strength of the weak interactions at low energies is @xmath83 .
the most important interaction is the @xmath8 pion - nucleon interaction @xmath84 with @xmath85 .
the @xmath8 pion - nucleon coupling @xmath86 is not well - known . in lo of the eft with perturbative pions , which we are employing ,
the @xmath8 asymmetry in @xmath87 is @xmath88 @xcite , so the recent experimental result @xmath89\cdot 10^{-7}$ ] @xcite gives a bound @xmath90 , which is the order of magnitude expected by nda . a first lattice qcd calculation at a pion mass @xmath91 mev gives , in our convention for @xmath86 , @xmath92\cdot 10^{-7}$ ] @xcite .
@xmath41 interactions are expected to be due , mostly , to the dimension - four qcd @xmath37 term , parameterized by @xmath93 , and the dimension - six operators that result from integrating out physics at the scale @xmath40 and the heavy degrees of freedom in the sm .
the complete set of @xmath41 dimension - six operators at the electro - weak scale has been given in ref .
@xcite , and the relevant operators at the hadronic scale have been summarized in ref
they are the isoscalar and isovector quark edm ( qedm ) and quark chromo - edm ( qcedm ) , the weinberg operator , which gives rise to a gluon chromo - edm ( gcedm ) , and four @xmath41 four - quark operators .
two of these four - quark operators are invariant under the sm gauge group and can be generated directly at the electroweak scale . their effect in the chiral eft at low energy can not be separated from the gcedm and we refer to these collectively as chiral - invariant sources ( @xmath94iss ) .
the other two four - quark operators break isospin and result from integrating out the weak gauge bosons and running to low energy . because they mix left- and right - handed quarks we denote these as fqlr .
the various @xmath41 sources are further discussed in app .
[ app : dimsix ] .
the dimension - four and six @xmath41 operators have different transformation properties under the chiral group @xmath95 , which has consequences for the @xmath41 couplings in chiral eft @xcite .
the interactions relevant to the rest of the paper are @xmath96 , \label{g0lag}\end{aligned}\ ] ] where @xmath97 ( @xmath98 ) is the isoscalar ( isovector ) @xmath41 pion - nucleon coupling , @xmath99 ( @xmath100 ) a short - range contribution to the isoscalar ( isovector ) nucleon edm , and @xmath101 a short - range @xmath41 two - nucleon interaction .
the term proportional to @xmath102 is a recoil correction and depends on the sum of the incoming and outgoing nucleon momenta .
other @xmath41 interactions , some expected to be of comparable size , will not be needed below because of the quantum numbers of the deuteron .
the relative importance of the operators in eq . depends on the chiral properties of the @xmath41 source at the quark - gluon level . as described in app .
[ app : ordmag ] , the dimensionless one - nucleon couplings @xmath103 and @xmath104 are given by the dimensionless strengths of the underlying @xmath41 interactions , times factors of @xmath105 that depend on the chiral transformation properties of the source .
for the qcd @xmath37 term , the qcedm , and the fqlr , which violate chiral symmetry , non - derivative pion - nucleon couplings like @xmath97 can appear in the chiral lagrangian at lo . in this case @xmath106 and pion effects tend to dominate because of the low mass .
in contrast , @xmath94iss can generate pion - nucleon non - derivative couplings only through insertion of the quark mass , which costs two powers of @xmath107 , so that , for example , @xmath108 .
the @xmath97 term still appears in the lo lagrangian , but it is accompanied by the equally important two - nucleon and electromagnetic operators , whose construction does not require any insertion of the quark mass . finally , the presence of a photon field causes the qedm to contribute mainly to the photon - nucleon sector , purely hadronic operators being suppressed by powers of @xmath109 . the interactions in eqs .
and can be used to compute the @xmath8 , @xmath41 and @xmath10 form factors of nuclei .
the nucleon does not possess a @xmath31 form factor .
we summarize here the results for the nucleon tdff and electric dipole form factor ( edff ) , which are needed for the calculation of the deuteron tqff in sec .
[ calculation ] .
the @xmath8 and @xmath41 currents are written as , respectively , @xmath110 \label{currenttdff}\ ] ] and @xmath111 , \label{currentedff}\ ] ] where @xmath112 denotes the four - momentum of the photon and @xmath113 is the sum of the nucleon momenta . we write @xmath114 and @xmath115 where @xmath116 and @xmath117 ( @xmath118 and @xmath119 ) are the nucleon isoscalar and isovector anapole ( electric dipole ) moments ,
@xmath120=1 , and @xmath121 is finite . at lo ,
the nucleon tdffs come entirely from pion loops , in which one vertex is the @xmath8 pion - nucleon coupling @xmath86 . by nda one
expects @xmath122 .
the calculation of ref .
@xcite shows that the nucleon anapole form factor is , at lo , isoscalar and finite , @xmath123 , \qquad a_1^{({\rm lo } ) } = 0 .
\label{leadinganapolescalar}\ ] ] the isovector anapole form factor appears only at nlo , where short - range contributions to the moments also are present . neglecting @xmath124 numbers ,
the result for @xmath125 is a factor of @xmath126 larger than the nda estimate , as often happens in baryon chpt .
the nucleon edff was computed in ref .
@xcite to nlo for all @xmath41 sources of dimension up to six . for the qcd @xmath37 term , the qcedm , and the fqlr , the isovector nucleon edm receives a one - loop contribution from @xmath97 at lo . at the same order
there are also short - range isoscalar ( @xmath99 ) and isovector ( @xmath100 ) contributions , the latter being required by renormalization - group invariance .
the isoscalar and isovector nucleon edms are given by @xcite @xmath127 where we used dimensional regularization in @xmath26 spacetime dimensions , with @xmath128 , and @xmath129 the renormalization scale .
in this case there is no @xmath126 enhancement , and the nucleon edm is suppressed by the loop factor @xmath130 with respect to the pion nucleon coupling @xmath97 .
the momentum dependence of the edff is purely isovector in lo and governed by the scale @xmath131 , as is the case for the isoscalar tdff , but it is not needed in the following . for the qedm and the @xmath94iss , @xmath132 is at most as large as the short - range coupling @xmath133 , and the loop suppression makes its contribution negligible .
the edff is then momentum independent at lo and completely determined by the low - energy constants @xmath134 , @xmath135 in this case the momentum dependence appears in higher order and is determined by short - range physics .
the @xmath41 couplings @xmath97 , @xmath119 , and @xmath136 are not known and , in order to estimate the magnitude of the tqff they induce , we will need to make some reasonable assumptions .
first , we assume that there are no cancellations between @xmath137 and @xmath138 , so that , for @xmath41 violation from the qedm and @xmath94iss , the bound on the neutron edm @xmath139 can be directly translated into the bound @xmath140 fm .
second , as pointed out in ref .
@xcite , we should not expect any cancellation in eq . between pieces that are non - analytic and analytic in @xmath141 . with the reasonable value @xmath142 ,
the same bound applies for @xmath143 in the case of @xmath37 term , qcedm and fqlr .
moreover , since the long - range contributions give the estimate @xmath144 fm , the existing experimental bound on the neutron edm yields an approximate bound on the @xmath41 pion - nucleon coupling , @xmath145 .
with the interactions described in sec .
[ interactions ] , we can calculate the long - range contributions to the deuteron tqff , using the techniques of refs .
@xcite . as usual in such a calculation , the orders of magnitude of the various contributions can be found by combining the power counting rules of chpt based on nda with the rules for two - nucleon states as summarized , for example , in ref .
a pion propagator scales as @xmath146 .
a loop involving a single nucleon contributes a factor @xmath147 from the integration and a factor @xmath148 from the nucleon propagator .
the infrared enhancement of a loop involving two nucleons gives a factor @xmath149 from the integration and a factor @xmath150 from each nucleon propagator .
the deuteron wavefunction contributes an overall normalization factor @xmath151 .
the deuteron itself is built out of the two - nucleon contact interaction with coefficient @xmath152 and the nucleon kinetic terms in eqs . and .
pion exchange originating from the pion kinetic terms and pion - nucleon coupling in eq . contributes to the deuteron structure at relative @xmath153 , together with a two - derivative contact interaction that accounts for short - range energy dependence in the on - shell two - nucleon amplitude @xcite . since we calculate the tqff to lo only
, @xmath154 is the sole @xmath6 two - nucleon input needed .
the @xmath31 tqff is an intrinsically two - nucleon observable , which requires at least one symmetry - violating interaction between the two nucleons .
we argue in app .
[ app : ordmag ] that @xmath10 interactions are much smaller than contributions from separate @xmath8 and @xmath41 interactions .
the lowest - order diagrams involving the @xmath8 vertex @xmath86 and one of the @xmath41 couplings are shown in figs .
[ fig1][fig3 ] . in these figures ,
only one possible ordering is shown .
circles , triangles , and squares denote the leading @xmath6 , @xmath8 , and @xmath41 interactions in eqs . , , and , respectively ; a circled circle , the @xmath6 magnetic photon - nucleon interactions in eq . ; a twice circled triangle , the @xmath8 anapole moment of the nucleon in eq . .
the hatched circles denote deuteron states obtained from the iteration of the leading two - nucleon interaction , which brings in dependence on the binding momentum @xmath154 .
the natural scale for momentum dependence of the tqff is @xmath155 , so we express our results in terms of @xmath156 .
we also define the ratio @xmath157 of low - momentum scales . .
pions and photons are represented by solid , dashed and wavy lines , respectively .
lo @xmath6 , @xmath41 , and @xmath8 interactions are denoted by circles , squares , and triangles , respectively .
an nlo @xmath6 interaction is denoted by a circled circle .
deuteron states obtained from the iteration of the leading @xmath6 two - nucleon interaction are represented by hatched circles.,width=566 ] ) contributions to the deuteron tqff , @xmath158 .
the notation is as in fig .
[ fig1].,width=566 ] .
the twice - circled triangle stands for the anapole form factor .
the other notation is as in fig .
[ fig1].,width=566 ] let us first consider a photon which interacts without breaking @xmath3 and @xmath4 .
in this case the photon couples to the nucleon via the magnetic couplings in eq . or to a pion via interactions obtained by gauging the derivatives in the pion kinetic energy and pion - nucleon axial coupling in eq . .
diagrams with only one pion exchange and @xmath97 and @xmath86 vertices on each end vanish .
this can be understood from the fact that such diagrams do not have enough powers of momentum in the vertices to generate a form factor of the form of eq . , and it agrees with the more general analysis of the @xmath31 two - nucleon interaction @xcite .
this leaves three - loop diagrams , containing either two pion exchanges ( tpe ) or one pion exchange and a short - range @xmath41 two - nucleon interaction ( 4n ) , figs .
[ fig1 ] and [ fig2 ] respectively . using the power - counting rules outlined above , the sizes of the diagrams in figs .
[ fig1 ] and [ fig2 ] are @xmath159 whether the diagrams in figs .
[ fig1 ] or [ fig2 ] are more important depends on the @xmath41 source . for the @xmath160 term ,
the qedm , the qcedm , and the fqlr operator the contributions from the short - range interaction @xmath136 are always suppressed , in this case by @xmath47 , with respect to tpe , because for these sources @xmath161 , see app .
[ app : ordmag ] . for @xmath94iss ,
the opposite is true because of the extra @xmath162 suppression of @xmath163 , which makes the short - range contributions larger by a factor of @xmath164 .
the diagrams in fig .
[ fig1 ] are formally the leading contributions for the qcd @xmath165 term , the qcedm , and the fqlr , while those in fig .
[ fig2 ] are leading for the @xmath94iss .
note that for the isovector qcedm and the fqlr , one should consider not only the isoscalar pion - nucleon coupling @xmath97 but also the isovector pion - nucleon coupling @xmath98 , but such diagrams vanish .
alternatively , the photon can interact with the nucleon with a @xmath8 or @xmath41 interaction , in which case a single two - nucleon interaction , @xmath41 or @xmath8 respectively , is sufficient to produce a tqff see fig .
[ fig3 ] . in diagrams [ fig3](a , b ) one of the nucleons couples to the magnetic field via its anapole moment , with @xmath41 coming either from pion exchange or from a two - nucleon interaction . here
the anapole `` vertex '' stands for a one - loop diagram , which produces the result . in diagram
[ fig3](c ) the photon couples to the nucleon through a recoil correction to the @xmath41 edm , with @xmath8 coming from pion exchange . by power
counting , the contributions of diagrams [ fig3](a , b ) to the tqff are @xmath166 were we used the nda expectation for the anapole moment .
diagrams [ fig3](a , b ) are thus suppressed by one power of @xmath167 compared to the contributions of figs .
[ fig1 ] and [ fig2 ] .
however , @xmath116 in eq . is a factor of @xmath126 larger than the nda estimate , making the corresponding contributions to the tqff competitive with lo .
again , of other possible @xmath41 couplings only @xmath97 and @xmath136 contribute .
diagram 3(c ) represents contributions to the tqff coming from the nucleon edff .
it scales as @xmath168 and there is no contribution from @xmath99 at this order . for @xmath37 term , qcedm , and fqlr , @xmath169 and this contribution
is suppressed by @xmath46 ( a factor coming from the recoil ) compared to the analogous anapole diagram [ fig3](a ) . for @xmath94iss ,
this contribution is comparable to fig .
[ fig2 ] , while for the qedm it is the sole leading contribution , since numerically @xmath170 . to summarize these power - counting arguments
, we expect the tqff induced by the @xmath37 term , the qcedm , and the fqlr to be dominated by the tpe diagrams in fig .
[ fig1 ] , with possibly large corrections from the nucleon tdff , diagram [ fig3](a ) . for @xmath94iss ,
the dominant contribution should come from the diagrams involving the @xmath41 short - range two - nucleon interaction in fig .
[ fig2 ] and from the nucleon edff , diagram [ fig3](c ) , with a sizable correction from the nucleon tdff , diagram [ fig3](b ) . for the qedm
only the nucleon edff contribution [ fig3](c ) should be important .
we now proceed to the evaluation of diagrams in figs .
[ fig1 ] , [ fig2 ] , and [ fig3 ] .
we find that only the isovector magnetic moment gives a non - vanishing contribution to diagram [ fig1](a ) and [ fig2](a ) . diagrams with other photon - nucleon interactions ( the isoscalar magnetic moment and the minimal coupling through the covariant derivative in the nucleon kinetic term ) vanish .
similar diagrams where pion exchanges and contact interaction occur both before or after the insertion of the photon coupling also vanish .
diagrams [ fig1](b , c ) and [ fig1](d , e ) differ only by an isospin factor . the diagrams in fig .
[ fig1 ] are finite in four ( and three ) dimensions .
we express the result for their contribution to the tqff as @xmath171,\ ] ] in terms of two three - loop integrals @xmath172 .
likewise , the results for the tqff in fig . [ fig2 ] are expressed in terms of two two - loop functions @xmath173 as @xmath174 , \label{i2ab}\ ] ] where we have used power - divergence subtraction @xcite .
the @xmath129 dependence is absorbed in @xmath136 itself , since here it appears in the same combination as in the magnetic quadrupole form factor @xcite .
the expansions to order @xmath175 of the integrals @xmath176 are given in app .
[ app : ffs ] .
the resulting contributions to the tqm are @xmath177 \cdot 10^{-2 } \ ; \frac{\bar g_0 h_1}{f_{\pi}^2 } \ ; e \ , \textrm{fm}^3 .
\label{valuelo}\ ] ] and @xmath178 \cdot 10^{-2 } \ , m_n \frac{\mu - \gamma}{4\pi } \bar c_0 \ , h_1 \ , e\ , \textrm{fm}^3.\ ] ] finally , we consider diagrams [ fig3](a , b ) and ( c ) . for isoscalar @xmath41 , only the isoscalar tdff @xmath179 contributes in diagrams [ fig3](a , b ) .
isospin - breaking @xmath41 , for example from insertions of @xmath98 , would contribute to diagram [ fig3](a ) together with the isovector tdff @xmath180 .
however , the isovector tdff is suppressed by @xmath46 ( and by a factor of @xmath126 ) with respect to the isoscalar tdff . therefore , even for sources that generate @xmath97 and @xmath98 at the same level , @xmath98 contributions to the tqff are subleading . diagram [ fig3](c ) is leading only for the qedm and @xmath94iss , for which the isovector edff is momentum independent and coincides with the edm .
diagrams [ fig3](a , b ) and ( c ) result in contributions to the tqff given by @xmath181 \label{ff1}\end{aligned}\ ] ] and @xmath182 respectively .
the expression for the one - loop integral @xmath183 along with the expansions to order @xmath175 of two - loop integrals @xmath184 can be found in app .
[ app : ffs ] .
numerically this gives @xmath185 \cdot 10^{-2 } \ ; h_1 \ ; e \
, \textrm{fm}^3 \label{valuelo'}\ ] ] and @xmath186 the result shows that the tdff contribution , though expected to be suppressed by the factor @xmath187 , is comparable to the lo values in eqs . and , in line with the @xmath126 enhancement in the tdm .
for the qcd @xmath37 term , the qcedm , and the fqlr , we can assess the importance of the edm contribution to the tqff by substituting in eq .
the estimate for the nucleon edm in terms of @xmath97 , @xmath188 fm .
we find @xmath189 @xmath24 , which is numerically small compared to eqs . and , as expected by power counting .
we can now combine the results found so far . for the qcd @xmath37 term , the qcedm , and the fqlr , the tpe contributions from the diagrams in fig .
[ fig1 ] and the tdff contributions in fig .
[ fig3](a ) have comparable size , giving @xmath190 this number is within a factor @xmath191 of the power counting estimate in eq . , indicating that the power counting works well ( apart from the @xmath126 in the anapole moment ) . for @xmath94iss , by power
counting the leading contributions are expected to come from the diagrams in fig .
[ fig2 ] , with insertions of the four - nucleon coupling @xmath136 , and from the edm in diagram [ fig3](c ) .
also in this case , the contribution of the tdff in diagram [ fig3](b ) is numerically important .
we find @xmath192\cdot 10^{-2 } \ ; h_1 \;e \,\textrm{fm}^3 . \label{tdchiiss}\ ] ] if @xmath136 and @xmath100 have their nda values , their respective contributions are numerically comparable . in the case of the qedm ,
the tqm is dominated by the contribution from the nucleon edm , and we find @xmath193 because of dimensionless numerical factors this value is about an order of magnitude smaller than expected by the power - counting estimates based on nda .
it is interesting to compare our results for the deuteron tqm in eqs .
, , and with the largest @xmath41 moment of the deuteron , edm or mqm , for the respective @xmath41 sources . in ref .
@xcite power - counting estimates and lo results were given for the deuteron edm , @xmath194 , and mqm , @xmath195 , in chiral eft with perturbative pion exchange . for sources that break chiral symmetry and generate non - derivative @xmath41 pion - nucleon couplings in lo , @xmath194 and
@xmath195 are expected to be dominated by two - body effects and be enhanced with respect to the nucleon edm . in the case of the qcd @xmath37 term , @xmath195 is expected to be the largest moment ( in natural units ) , because at lo @xmath97 does not contribute to @xmath194 ( except through the nucleon edm , eq . ) . on the other hand ,
the qcedm and the fqlr , which generate also the isovector coupling @xmath98 in lo , induce @xmath194 and @xmath195 of the same size . for @xmath94iss , @xmath194 and @xmath195
are also expected to be of the same size , and of similar size as the nucleon edm .
the deuteron edm is in fact expected to be well approximated by ( twice ) the isoscalar nucleon edm , while the deuteron mqm , in the perturbative - pion counting , receives the largest contribution from the four - nucleon coupling @xmath136 .
for all these sources , we compare the deuteron tqm to its mqm , given by @xcite @xmath196 \ , \frac{1+\xi}{(1 + 2\xi)^2 } + e ( 1+\kappa_0 ) \frac{\mu-\gamma}{2\pi}\bar c_0.\ ] ] we consider the dimensionless ratio @xmath197 , which , by power counting , is expected to be of order @xmath198 . for the @xmath37 term , qcedm , and fqlr , @xmath136 is subleading in eq .
, so that from eq .
@xmath199 for the @xmath37 term , one can neglect @xmath98 and the @xmath41 couplings drop out of the ratio , which is approximately @xmath200 , as expected by power counting . for the qcedm
the ratio in eq . depends on @xmath201 , which by nda is expected to be order one . for the fqlr , as discussed in ref .
@xcite , @xmath97 is somewhat suppressed with respect to @xmath98 , further suppressing the deuteron tqm with respect to the mqm . in the case of @xmath94iss , @xmath136
is expected to be the leading term in eq . ; if we neglect the contribution of the nucleon edm in eq . , we get @xmath202 which is also in good agreement with the nda expectation
. for the remaining dimension - six source , the qedm , @xmath194 is also well approximated by the isoscalar nucleon edm , @xmath203 while @xmath195 is suppressed by one power of @xmath47 with respect to the edm @xcite .
therefore , for the qedm we compare the deuteron tqm with its edm using the dimensionless ratio @xmath204 . from eq . ,
@xmath205 which is a bit smaller than naively expected .
equations , and make it explicit that the deuteron tqff , in natural units , is suppressed roughly by a factor of @xmath206 with respect to the largest @xmath41 moment .
the lack of any significant numerical enhancement thus leads to a very small tqff .
the bounds on @xmath97 , @xmath100 , and @xmath86 inferred in sec .
[ interactions ] allow us to estimate the size of the tqm .
for the qcd @xmath37 term , the qcedm , and the fqlr we find @xmath207 while for the qedm we find the even smaller value @xmath208 for @xmath94iss , one expects a similar value , but to be more precise a bound on @xmath136 is needed .
these estimates have been obtained in chiral eft with perturbative pions .
iterating pions one can extend the regime of validity of the theory beyond @xmath50 at the cost of much more complicated renormalization @xcite .
because the binding momentum of nucleons in the deuteron is @xmath209 , we do not expect drastic changes in the quantities calculated here . in the case of our comparison @xmath41 moments , this expectation has been checked @xcite and shown to be reasonable .
we conclude that the value of the deuteron tqm from parity violation in the sm and parity- and time - reversal violation due to the sm @xmath210 term or dimension - six operators originating beyond the sm is , not surprisingly , tiny .
evidence for a nonzero value for the deuteron tqm that is larger than the `` background '' value @xmath211 would likely be due to new @xmath31 interactions .
u. van kolck acknowledges the hospitality of the kvi groningen on many occasions .
this research was supported by the dutch stichting fom under programs 104 and 114 ( jdv , rget ) and in part by the dfg and the nsfc through funds provided to the sino - german crc 110 `` symmetries and the emergence of structure in qcd '' ( jdv ) , by the us doe under contract de - ac02 - 05ch11231 with the director , office of science , office of high energy physics ( em ) , and under grant de - fg02 - 04er41338 ( uvk ) , and by the universit paris sud under the program attractivit 2013 ( uvk ) .
the @xmath41 operators of dimension four and six , after integrating out physics at the scale @xmath40 and the heavy degrees of freedom in the sm , are given by @xcite @xmath212 the first operator is the qcd @xmath37 term , where @xmath213 denotes the reduced light - quark mass @xmath214 .
we assume that @xmath93 . in the second and third operators , @xmath118
( @xmath215 ) and @xmath216 ( @xmath217 ) are the isoscalar ( isovector ) components of the quark edm ( qedm ) and quark chromo - edm ( qcedm ) , respectively . in the fourth term
, @xmath218 represents the gluon chromo - edm ( gcedm ) .
the remainder consists of @xmath41 four - quark operators .
the ones with coefficients @xmath219 are invariant under the sm gauge group , and can be generated directly at the electroweak scale .
the isospin - breaking four - quark operators ( fqlr ) with coefficients @xmath220 , on the other hand , are generated by integrating out the weak gauge bosons and running to low energy .
the importance of the dimension - six @xmath41 operators depends on the high energy scale @xmath40 , on the detailed mechanism of @xmath3 and @xmath4 breaking in new physics , and on the running to the qcd scale ( for the latter , see ref . @xcite and references therein ) .
we hide all the model dependence by introducing the dimensionless parameters @xmath221 , @xmath222 , @xmath223 , @xmath224 , and @xmath225 for ( isoscalar and isovector ) qedm and qcedm , gcedm , and isospin - breaking and invariant four - quark operators , respectively .
we write @xcite @xmath226 naively , one expects @xmath221 , @xmath227 , @xmath223 , @xmath224 , and @xmath225 to be @xmath228 , @xmath229 , @xmath230 , @xmath231 , and @xmath231 , respectively , but they could be significantly smaller or larger , depending on the new - physics model . as in the case of electroweak interactions , the relative strength of @xmath41 interactions at low energies
is expressed by the ratio of a low - energy scale and the characteristic scale where @xmath3 and @xmath4 violation arise .
the dimension - four @xmath37 term is not suppressed by any high energy scale , and its reduced coupling is @xmath232 , where the pion mass is a reminder of the intimate relationship between the @xmath37 term and the quark masses . in the case of dimension - six operators ,
the qedm and qcedm require an insertion of the higgs vacuum expectation value , which can be traded for the quark mass .
a dimensionless measure of their importance is then @xmath233 . for the remaining dimension - six @xmath41 operators
the relevant low energy scale is @xmath76 , and the reduced couplings are @xmath234 .
in terms of the dimensionless parameters defined in app . [ app : dimsix ] , we can estimate @xcite the size of the couplings @xmath97 and @xmath100 in eq . using nda as @xmath235 here @xmath236
is related to the light quark mass difference by @xmath237 . in the case of the qcd @xmath37 term , chiral symmetry relates @xmath97 to the strong contribution to the neutron - proton mass difference by @xcite @xmath238 mev , using results from a lattice qcd calculation @xcite . for the dimension - six operators , going beyond nda
requires additional input from lattice qcd or other non - perturbative techniques .
the contributions to @xmath97 from the isovector component of the qcedm and the fqlr , which break isospin , are suppressed by the quark - mass difference .
these two sources , as well as the @xmath94iss , generate at leading order also the isovector pion - nucleon coupling @xmath98 . however , as we discuss in the main text , such a coupling does not contribute to the tqff at lo .
for all sources there is also an isoscalar edm component @xmath99 , but this term is not relevant here either .
the estimate of the lec @xmath136 associated with the four - nucleon operator in eq . requires more care @xcite . in the eft where pion exchange is treated perturbatively , pions with momentum above @xmath50 are integrated out , which induces contributions to multi - nucleon interactions . in particular ,
@xmath136 can be generated by a high - energy pion exchange between two nucleons with one vertex originating in @xmath239 and the other in @xmath97 .
for @xmath94iss a larger contribution to @xmath136 is generated if one uses , instead of @xmath97 , @xmath240 , where @xmath241 @xcite .
application of the nda rules at the scale @xmath50 then gives @xmath242 or @xmath243 .
however , this naive scaling is altered because the operator connects to an @xmath29 wave . as discussed in ref .
@xcite , this enhances the scaling by a factor @xmath244 due to non - perturbative renormalization by the leading - order @xmath6 two - nucleon interaction .
one finds @xmath245 combining the scaling of these low - energy constants with the power - counting rules outlined in sec .
[ calculation ] , we find for the size of the diagrams in figs .
[ fig1 ] , [ fig2 ] , and [ fig3 ] : @xmath246 various statements made in the text about relative magnitudes of the @xmath41 sources follow straghtforwardly from these relations .
the same nda technique can be used to estimate the size of other @xmath41 contributions .
for example , the contributions of short - range @xmath41 nucleon - nucleon - photon interactions to the @xmath31 form factor of the deuteron are , for all sources , subleading with respect to figs .
[ fig1 ] , [ fig2 ] , and [ fig3 ] . likewise , short - range @xmath8 nucleon - nucleon interactions contribute to the @xmath8 two - nucleon potential at higher order than one - pion exchange diagrams involving @xmath86 @xcite . as a consequence ,
diagrams involving these interactions are also suppressed compared to the diagrams discussed here .
the same holds for diagrams with @xmath8 nucleon - nucleon - photon ( @xmath247 ) interactions @xcite .
in addition , one can understand why effective @xmath31 interactions contribute at higher orders .
one - pion exchange with a @xmath31 coupling does not contribute to the @xmath31 potential in the two - nucleon sector @xcite .
the latter is dominated by exchanges of heavier mesons , like the @xmath248 and the @xmath117 , which , at low energy , appear as two - nucleon contact interactions with at least two derivatives ( see , _ e.g. _ , ref .
the size of these two - nucleon operators can be again estimated in nda by multiplying the reduced couplings for the @xmath8 and @xmath41 interactions .
for example , for the @xmath160 term a @xmath31 two - nucleon coupling would scale as @xmath249 . inserting this coupling into a two - loop diagram with the photon interacting on a nucleon line via the nucleon magnetic moment
gives rise to a tqm of order @xmath250 .
such a contribution is much smaller than the long - range component that results from the pion - nucleon couplings @xmath86 and @xmath97 in fig .
[ fig1 ] , as can be seen from comparison with eq . .
similar power - counting estimates indicate that contributions from diagrams with short - range @xmath31 @xmath247 vertices are smaller than the contributions in eqs .
( [ fig1scale])-([fig3cscale ] ) by at least a factor @xmath251 . for dimension - six sources
, we can neglect @xmath31 operators in the eft as well . for non - electromagnetic sources such as the qcedm , the fqlr operators , and
the @xmath94iss the argument is analogous to the one used for the @xmath160 term . for
the qedm one might think that @xmath31 pion - nucleon - photon interactions could be relevant .
indeed , a @xmath31 operator of the form @xmath252 gives a nonzero contribution to the deuteron tqm via a two - loop diagram .
however , by power counting we find this diagram to be smaller by a factor @xmath253 compared to the diagram in fig .
[ fig3 ] , which provides the dominant contribution for the qedm .
in sec . [ calculation ] the results for the deuteron tqff @xmath254 were expressed in terms of @xmath255 loop integrals @xmath256 , where @xmath257 and @xmath258 .
we list here the form and expansions of these integrals to terms of order @xmath259 , where @xmath260 . the integrals appearing in the two- and three - loop diagrams are more complicated .
we express them in terms of the dimensionless variables @xmath262 , obtained by rescaling the loop momenta , @xmath263 .
they can be conveniently calculated in coordinate space @xcite . the two - loop functions in eqs . , ,
and depend on three two - loop integrals , @xmath264 expanding in @xmath265 and retaining the first two terms in the expansion , we find @xmath266 the two - loop functions @xmath267 are obtained from @xmath268 by multiplying them by prefactors that take into account spin and isospin factors , and the symmetry factor of each diagram : @xmath269 similarly , the three - loop functions that enter in eq .
( [ i3ab ] ) are defined in terms of two three - loop tensor integrals , @xmath270 which are finite in three and four dimensions . again retaining terms up to order @xmath271 , @xmath272 + \mathcal o ( x^{\ , 4 } ) \label{threel}\end{aligned}\ ] ] and @xmath273 \nonumber \\ & & + x^2 \frac{\xi^3}{90 ( 1 + 2\xi)^5 } \left[\frac{5 + 50 \xi + 116 \xi^2 + 72 \xi^3}{3 } \log \left ( \frac{2(1+\xi)}{1 + 2\xi } \right ) \right .
\nonumber\\ & & \left .
+ \frac{205 + 2494 \xi + 12290 \xi^2 + 29472 \xi^3 + 37540 \xi^4 + 25672 \xi^5 + 8720 \xi^6 + 1120 \xi^7 } { 280 ( 1+\xi)^4 } \right ] \nonumber\\ & & + \mathcal o(x^4),\end{aligned}\ ] ] where @xmath274 is the dilogarithm function .
the function @xmath275 vanishes in the limit of vanishing binding momentum , @xmath276 .
@xmath277 consists of the sum of three terms , each of them divergent for @xmath276 .
because of cancellations between these terms , @xmath277 has a finite limit for vanishing @xmath224 , @xmath278 .
however , this limit is not a good approximation to the value of the three - loop function at the physical value of @xmath224 , @xmath279 .
one needs to keep at least terms up to @xmath280 to obtain an approximation of @xmath281 at the @xmath282 accuracy .
dubovik and a.a .
cheshkov , sov .
jetp * 24 * ( 1967 ) 111 ; v.m .
dubovik , e.p .
likhtman , and a.a .
cheshkov , sov .
jetp * 25 * ( 1967 ) 464 ; d. schildknecht , z. phys .
* 201 * ( 1967 ) 99 ; r. prepost , r.m .
simonds , and b.h .
wiik , phys .
* 21 * ( 1968 ) 1271 ; k.y .
lin , nucl .
b * 18 * ( 1970 ) 162 ; h. arenhvel and s.k .
singh , eur .
j. a * 10 * ( 2001 ) 183 .
eversheim , b. lorentz , and yu .
valdau ( spokespersons ) , test of time - reversal invariance in proton - deuteron scattering at cosy , http://apps.fz-juelich.de/pax/paxwiki/index.php/test_of_time-reversal_invariance_at_cosy_(tric ) ; p.d .
_ , hyperfine interact .
* 193 * ( 2009 ) 335 ; yu . valdau , pos * stori11 * 013 ; m. beyer , nucl .
a * 560 * ( 1993 ) 895 .
j. engel , p.h .
frampton , and r.p .
springer , phys .
d * 53 * ( 1996 ) 5112 ; m.j .
ramsey - musolf , phys .
* 83 * ( 1999 ) 3997 , _ ibid . _
* 84 * ( 2000 ) 5681(e ) ; a. kurylov , g.c .
mclaughlin , and m.j .
ramsey - musolf , phys .
d * 63 * ( 2001 ) 076007 .
hockings and u. van kolck , phys .
b * 605 * ( 2005 ) 273 ; k. ottnad , b. kubis , u .- g .
meiner , and f .- k .
guo , phys .
b * 687 * ( 2010 ) 42 ; j. de vries , e. mereghetti , r.g.e .
timmermans , and u. van kolck , phys .
b * 695 * ( 2011 ) 268 ; e. mereghetti , j. de vries , w.h . hockings , c.m .
maekawa , and u. van kolck , phys .
b * 696 * ( 2011 ) 97 .
beane , p.f .
bedaque , m.j .
savage , and u. van kolck , nucl .
phys . a * 700 * ( 2002 ) 377 ; a. nogga , r.g.e .
timmermans , and u. van kolck , phys .
c * 72 * ( 2005 ) 054006 ; m. birse , phys . rev .
c * 74 * ( 2006 ) 014003 ; m. pavn valderrama , phys .
c * 83 * ( 2011 ) 024003 ; * 84 * ( 2011 ) 064002 ; b. long and c.j .
yang , phys .
c * 84 * ( 2011 ) 057001 ; * 85 * ( 2012 ) 034002 ; * 86 * ( 2012 ) 024001 .
j. de vries , r. higa , c .- p .
liu , e. mereghetti , i. stetcu , r.g.e .
timmermans , and u. van kolck , phys .
c * 84 * ( 2011 ) 065501 ; c .- p .
liu , j. de vries , e. mereghetti , r.g.e .
timmermans , and u. van kolck , phys .
b * 713 * ( 2012 ) 447 . | we calculate the toroidal quadrupole moment and form factor of the deuteron , which violate time - reversal symmetry but conserve parity , at leading order in two - flavor chiral effective field theory with perturbative pion exchange .
we take into account time - reversal and parity violation due to the qcd vacuum angle combined with parity violation resulting from the weak interaction in the standard model .
we also consider time - reversal and parity violation that at the quark - gluon level results from effective dimension - six operators originating from physics beyond the standard model .
* toroidal quadrupole form factor of the deuteron * and * u. van kolck@xmath0 * @xmath1 _ ernest orlando lawrence berkeley national laboratory , university of california , + berkeley , ca 94720 , usa _ |
as reviewed in the recent report @xcite , very soon after the dirac equation for a relativistic electron was discovered @xcite , gordon @xcite ( for all @xmath4 ) and darwin @xcite ( for @xmath5 ) found its solution in the point - like coulomb potential @xmath6 , they obtained the well - known sommerfeld s formula @xcite for energy - spectrum , @xmath7^{-1/2 } , \label{dirac}\ ] ] where the fine - structure constant @xmath8 , the principle quantum number @xmath9 and @xmath10 @xmath11 is the orbital angular momentum corresponding to the upper component of dirac bi - spinor , @xmath12 is the total angular momentum .
the integer values @xmath13 and @xmath12 label bound states whose energies are @xmath14 . for the example , in the case of the lowest energy states , one has @xmath15 for all states of the discrete spectrum , the binding energy @xmath16 increases as the nuclear charge @xmath17 increases .
no regular solution with @xmath18 and @xmath19 ( the @xmath20 ground state ) is found for @xmath21 , this was first noticed by gordon in his pioneer paper @xcite .
this is the problem so - called `` @xmath22 catastrophe '' .
the problem was solved @xcite by considering the fact that the nucleus is not point - like and has an extended charge distribution , and the potential @xmath23 is not divergent when @xmath24 .
the @xmath22 catastrophe disappears and the energy - levels @xmath25 of the bound states @xmath26 , @xmath27 and @xmath28 , @xmath29 smoothly continue to drop toward the negative energy continuum ( @xmath30 ) , as @xmath17 increases to values larger than @xmath31 .
the critical values @xmath32 for @xmath33 were found @xcite : @xmath34 is a critical value at which the lowest energy - level of the bound state @xmath20 encounters the negative energy continuum , while other bound states @xmath35 encounter the negative energy continuum at @xmath36 , thus energy - level - crossings and productions of electron and positron pair takes place , provided these bound states are unoccupied .
we refer the readers to @xcite for mathematical and numerical details .
the energetics of this phenomenon can be understood as follow .
the energy - level of the bound state @xmath20 can be estimated as follow , @xmath37 where @xmath38 is the average radius of the @xmath20 state s orbit , and the binding energy of this state @xmath39 .
if this bound state is unoccupied , the bare nucleus gains a binding energy @xmath40 larger than @xmath41 , and becomes unstable against the production of an electron - positron pair .
assuming this pair - production occur around the radius @xmath42 , we have energies of electron ( @xmath43 ) and positron ( @xmath44 ) : @xmath45 where @xmath46 are electron and positron momenta , and @xmath47 . the total energy required for a pair production
is , @xmath48 which is independent of the potential @xmath49 .
the potential energies @xmath50 of electron and positron cancel each other and do not contribute to the total energy ( [ totaleofep ] ) required for pair production .
this energy ( [ totaleofep ] ) is acquired from the binding energy ( @xmath39 ) by the electron filling into the bound state @xmath20 .
a part of the binding energy becomes the kinetic energy of positron that goes out .
this is analogous to the familiar case that a proton ( @xmath5 ) catches an electron into the ground state @xmath20 , and a photon is emitted with the energy not less than 13.6 ev . in this article
, we study classical and semi - classical states of electrons , electron - positron pair production in an electric potential of macroscopic cores with charge @xmath51 , mass @xmath52 and macroscopic radius @xmath1 .
setting the origin of spherical coordinates @xmath53 at the center of such cores , we write the vectorial potential @xmath54 , where @xmath55 and @xmath56 is the coulomb potential .
the motion of a relativistic electron with mass @xmath57 and charge @xmath58 is described by its radial momentum @xmath59 , total angular momenta @xmath60 and the hamiltonian , @xmath61 where the potential energy @xmath62 , and @xmath63 corresponds for positive and negative energies .
the states corresponding to negative energy solutions are fully occupied .
the total angular momentum @xmath60 is conserved , for the potential @xmath23 is spherically symmetric . for a given angular momentum @xmath64 , where @xmath65 is the transverse velocity ,
the effective potential energy for electron s radial motion is @xmath66 where @xmath63 indicates positive and negative effective energies , outside the core ( @xmath67 ) , the coulomb potential energy @xmath23 is given by @xmath68 inside the core ( @xmath69 ) , the coulomb potential energy is given by @xmath70 , \label{poteninside}\ ] ] where we postulate the charged core has a uniform charge distribution with constant charge density @xmath71 , and the core volume @xmath72 .
coulomb potential energies outside the core ( [ potenoutside ] ) and inside the core ( [ poteninside ] ) are continuous at @xmath73 . the electric field on the surface of the core , @xmath74 where the electron compton wavelength @xmath75 , the critical electric field @xmath76 and the parameter @xmath77 is the electric potential - energy on the surface of the core in unit of the electron mass - energy
. given different values of total angular momenta @xmath60 , the stable circulating orbits @xmath78 ( states ) are determined by the minimum of the effective potential @xmath79 ( [ tote ] ) ( see fig .
[ fig1 ] ) , at which @xmath80 .
we obtain stable orbits locate at the radii @xmath78 outside the core , @xmath81 for different @xmath60-values . substituting eq .
( [ stable ] ) into eq .
( [ tote ] ) , we find the energy of electron at each stable orbit , @xmath82 for the condition @xmath83 , we have @xmath84 , \label{ocon}\ ] ] where the semi - equality holds for the last stable orbits outside the core @xmath85 . in the point - like case @xmath86 ,
the last stable orbits are @xmath87 eq . ( [ mtote ] )
shows that there are only positive or null energy solutions ( states ) in the case of a point - like charge , which corresponds to the energy - spectra eqs .
( [ dirac - k1],[dirac - k2],[dirac - k3 ] ) in quantum mechanic scenario .
while for @xmath88 , radii of stable orbits @xmath89 and energies @xmath90 , classical electrons in these orbits are critically bound for their banding energy goes to zero .
we conclude that the energies ( [ mtote ] ) of stable orbits outside the core must be smaller than @xmath91 , but larger than zero , @xmath92 . therefore , no energy - level crossing with the negative energy spectrum occurs .
we turn to the stable orbits of electrons inside the core .
analogously , using eqs .
( [ tote],[poteninside ] ) and @xmath80 , we obtain the stable orbit radius @xmath93 in the unit of @xmath1 , obeying the following equation , @xmath94 and corresponding to the minimal energy ( binding energy ) of these states @xmath95 . \label{mtote2}\ ] ] there are 8 solutions to this polynomial equation ( [ inorbite ] ) , only one is physical , the solution @xmath78 that has to be real , positive and smaller than one .
as example , the numerical solution to eq .
( [ inorbite ] ) is @xmath96 for @xmath97 and @xmath98 . in following , we respectively adopt non - relativistic and ultra - relativistic approximations to obtain analytical solutions .
first considering the non - relativistic case for those stable orbit states whose kinetic energy term characterized by angular momentum term @xmath60 , see eq .
( [ tote ] ) , is much smaller than the rest mass term @xmath91 , we obtain the following approximate equation , @xmath99 and the solutions for stable orbit radii are , @xmath100 and energies , @xmath101 the consistent conditions for this solution are @xmath102 for @xmath103 , and @xmath104 for non - relativistic limit @xmath105 , where the transverse velocity @xmath106 . as a result ,
the binding energies ( [ mtote3n ] ) of these states are @xmath107 , are never less than zero .
these in fact correspond to the stable states which have large radii closing to the radius @xmath1 of cores and @xmath108 .
second considering the ultra - relativistic case for those stable orbit states whose the kinetic energy term characterized by angular momentum term @xmath60 , see eq .
( [ tote ] ) , is much larger than the rest mass term @xmath91 , we obtain the following approximate equation , @xmath109 and the solutions for stable orbit radii are , @xmath110 which gives @xmath111 for the same values of parameters @xmath77 and @xmath112 in above .
the consistent condition for this solution is @xmath113 for @xmath103 .
the energy levels of these ultra - relativistic states are , @xmath114mc^2 , \label{mtote3}\ ] ] and @xmath115 . the particular solutions @xmath116 and @xmath117 are respectively given by @xmath118 these in fact correspond to the stable states which have small radii closing to the center of cores and @xmath119 . to have the energy - level crossing to the negative energy continuum , we are interested in the values @xmath120 for which the energy - levels ( [ mtote3 ] ) of stable orbit states are equal to or less than @xmath2 , @xmath121mc^2\leq -mc^2 .
\label{mtote4}\ ] ] as example , with @xmath122 and @xmath123 , @xmath124 , @xmath125 .
the lowest energy - level of electron state is @xmath126 with the binding energy , @xmath127 locating at @xmath128 , the bottom of the potential energy @xmath129 ( [ poteninside ] ) .
in order to have further understanding , we consider the semi - classical scenario .
introducing the planck constant @xmath130 , we adopt the semi - classical bohr - sommerfeld quantization rule @xmath131 which are discrete values selected from continuous total angular momentum @xmath60 in the classical scenario .
the variation of total angular momentum @xmath132 in th unit of the planck constant @xmath133 , we make substitution @xmath134 in classical solutions that we obtained in section ( [ class ] ) . 1 . the radii and energies of stable states outside the core ( [ stable ] ) and ( [ mtote ] ) become : @xmath135 where the electron compton length @xmath136 .
the radii and energies of non - relativistic stable states inside the core ( [ inorbite4 ] ) and ( [ mtote3n ] ) become : @xmath137 3 .
the radii and energies of ultra - relativistic stable states inside the core ( [ inorbite2 ] ) and ( [ mtote3 ] ) become : @xmath138mc^2 .
\label{qmtote3}\end{aligned}\ ] ] note that radii @xmath78 in the second and third cases are in unit of @xmath1 .
when these semi - classical states are not occupied as required by the pauli principle , the transition from one state to another with different discrete values of total angular momentum @xmath139 ( @xmath140 and @xmath141 ) undergoes by emission or absorption of a spin-1 ( @xmath133 ) photon . following the energy and angular - momentum conservations , photon emitted or absorbed in the transition
have angular momenta @xmath142 and energy @xmath143 . in this transition of stable states ,
the variation of radius is @xmath144 .
we first consider the stability of semi - classical states against such transition in the case of point - like charge , i.e. , eqs .
( [ qstable],[qmtote ] ) with @xmath11 .
as required by the heisenberg indeterminacy principle @xmath145 , the absolute ground state for minimal energy and angular momentum is given by the @xmath146 state , @xmath147 , @xmath148 and @xmath149 , which corresponds to the last stable orbit ( [ ocon1 ] ) in the classical scenario .
thus the stability of all semi - classical states @xmath150 is guaranteed by the pauli principle .
this is only case for @xmath151 . while for @xmath152 , there is not an absolute ground state in the semi - classical scenario
this can be understood by examining how the lowest energy states are selected by the quantization rule in the semi - classical scenario out of the last stable orbits ( [ ocon1 ] ) in the classical scenario . for the case of @xmath153 , equating @xmath60 in eq .
( [ ocon1 ] ) to @xmath154 ( [ angq ] ) , we find the selected state @xmath146 is only possible solution so that the ground state @xmath146 in the semi - classical scenario corresponds to the last stable orbits ( [ ocon1 ] ) in the classical scenario . while for the case of @xmath152 , equating @xmath60 in eq .
( [ ocon1 ] ) to @xmath154 ( [ angq ] ) , we find the selected semi - classical state @xmath155 in the semi - classical scenario corresponds to the last stable orbits ( [ ocon1 ] ) in the classical scenario .
this state @xmath156 is not protected by the heisenberg indeterminacy principle from quantum - mechanically decaying in @xmath133-steps to the states with lower angular momenta and energies ( correspondingly smaller radius @xmath78 ( [ qstable ] ) ) via photon emissions .
this clearly shows that the `` @xmath22-catastrophe '' corresponds to @xmath157 , falling to the center of the coulomb potential and all semi - classical states ( @xmath139 ) are unstable .
then we consider the stability of semi - classical states against such transition in the case of charged cores @xmath158 .
substituting @xmath60 in eq .
( [ angq ] ) into eq .
( [ ocon ] ) , we obtain the selected semi - classical state @xmath159 corresponding to the last classical stable orbit outside the core , @xmath160^{-1/2}\approx ( z\alpha)^{1/4}\left(\frac{r_c}{\lambda}\right)^{3/4 } > 0 .
\label{is1}\ ] ] analogously to eq .
( [ is0 ] ) , the same argument concludes the instability of this semi - classical state , which must quantum - mechanically decay to states with angular momentum @xmath161 inside the core , provided these semi - classical states are not occupied .
this conclusion is independent of @xmath162-value .
we go on to examine the stability of semi - classical states inside the core . in the non - relativistic case @xmath163
, the last classical stable orbits locate at @xmath157 and @xmath164 given by eqs .
( [ inorbite4],[mtote3n ] ) , corresponding to the lowest semi - classical state ( [ qinorbite4],[qmtote3n ] ) with @xmath146 and energy @xmath107 . in the ultra - relativistic case @xmath165 ,
the last classical stable orbits locate at @xmath157 and @xmath164 given by eqs .
( [ inorbite2],[mtote3 ] ) , corresponding to the lowest semi - classical state ( [ qinorbite2],[qmtote3 ] ) with @xmath146 and minimal energy , @xmath166mc^2\approx - \frac{3}{2}\beta mc^2 . \label{qmin}\end{aligned}\ ] ] this concludes that the @xmath146 semi - classical state inside the core is an absolute ground state in both non- and ultra - relativistic cases .
the pauli principle assures that all semi - classical states @xmath150 are stable , provided all these states accommodate electrons .
the electrons can be either present inside the core or produced from the vacuum polarization , later will be discussed in details .
we are particular interested in the ultra - relativistic case @xmath167 , i.e. , @xmath168 , the energy - levels of semi - classical states can be more profound than @xmath2 ( @xmath169 ) , energy - level crossings and pair - productions occur if these states are unoccupied , as discussed in introductory section .
when the energy - levels of semi - classical ( bound ) states @xmath170 ( [ mtote4 ] ) , energy - level crossings between these energy - levels ( [ mtote3 ] ) and negative energy continuum ( [ tote ] ) for @xmath171 , as shown in fig . [ fig2 ] .
the energy - level - crossing indicates that @xmath172 ( [ mtote3 ] ) and @xmath173 ( [ tote ] ) are equal , @xmath174 where angular momenta @xmath60 in @xmath172 ( [ qmtote3 ] ) and @xmath173 ( [ tote ] ) are the same for angular - momentum conservation .
the production of electron - positron pairs must takes place , provided these semi - classical ( bound ) states are unoccupied .
the phenomenon of pair production can be understood as a quantum - mechanical tunneling process of relativistic electrons .
the energy - levels @xmath172 of semi - classical ( bound ) states are given by eq .
( [ qmtote3 ] ) or ( [ mtote4 ] ) .
the probability amplitude for this process can be calculated by a semi - classical wkb method @xcite : @xmath175 where @xmath176 is transverse momenta and the radial momentum , @xmath177 ^ 2}. \label{px}\end{aligned}\ ] ] the energy potential @xmath23 is either given by @xmath178 ( [ potenoutside ] ) for @xmath179 , or @xmath180 ( [ poteninside ] ) for @xmath181 .
the limits of integration ( [ tprobability1 ] ) : @xmath182 ( [ inorbite2 ] ) or ( [ qinorbite2 ] ) indicating the location of the classical orbit ( classical turning point ) of semi - classical ( bound ) state ; while another classical turning point @xmath183 is determined by setting @xmath184 in eq .
( [ px ] ) .
there are two cases : @xmath185 and @xmath186 , depending on @xmath77 and @xmath112 values . to obtain a maximal wkb - probability amplitude ( [ tprobability1 ] ) of pair production
, we only consider the case that the charge core is bare and * the lowest energy - levels of semi - classical ( bound ) states : @xmath187 , the location of classical orbit ( [ inorbite2 ] ) @xmath188 and energy ( [ mtote3 ] ) @xmath189 ( [ mtote5 ] ) ; * another classical turning point @xmath190 , since the probability is exponentially suppressed by a large tunneling length @xmath191 . in this case
( @xmath190 ) , eq . ( [ px ] ) becomes @xmath192}\left(\frac{r}{r_c}\right)^4 } , \label{px1}\end{aligned}\ ] ] and @xmath171 leads to @xmath193^{1/4}. \label{rn1}\end{aligned}\ ] ] using eqs .
( [ tprobability1],[px1],[rn1 ] ) , we have @xmath194^{3/4}r_c}{c\hbar(mc^2\beta)^{1/2}}}\int_{0}^1 \sqrt{1-x^4}dx\right\}\nonumber\\ & = & \exp\left\{-0.87{\frac{2^{3/2}[(c|{\bf p}_\perp |)^2 + m^2c^4]^{3/4}r_c}{c\hbar(mc^2\beta)^{1/2}}}\right\}. \label{tprobability3}\end{aligned}\ ] ] dividing this probability amplitude by the tunneling length @xmath195 and time interval @xmath196 in which the quantum tunneling occurs , and integrating over two spin states and the transverse phase - space @xmath197 , we approximately obtain the rate of pair - production per the unit of time and volume , @xmath198 where @xmath199 is the electric field on the surface of the core and the compton time @xmath200 . to have the size of this pair - production rate , we consider a macroscopic core of mass @xmath201 and radius @xmath202 km , and the electric field on the core surface @xmath203 ( [ mtote6 ] ) is about the critical field ( @xmath204 ) . in this case ,
@xmath205 , @xmath206 , and the rate ( [ rate2 ] ) becomes @xmath207 which is exponentially small for @xmath208 . in this case , the charge - mass radio @xmath209 , where @xmath210 is the newton constant and proton s charge - mass radio @xmath211 .
it is interesting to compare this rate of electron - positron pair - production with the rate given by the hawking effect .
we take @xmath212 and the charge - mass radio @xmath213 for a naive balance between gravitational and electric forces . in this case
@xmath214 , the rate ( [ rate2 ] ) becomes , @xmath215 where the notation @xmath216 .
this is much larger than the rate of electron - positron emission by the hawking effect @xcite , @xmath217 since the exponential factor @xmath218 is much larger than @xmath219 , where @xmath220 .
in this letter , analogously to the study in atomic physics with large atomic number @xmath17 , we study the classical and semi - classical ( bound ) states of electrons in the electric potential of a massive and charged core , which has a uniform charge distribution and macroscopic radius . we have found negative energy states of electrons inside the core , whose energies can be smaller than @xmath2 , and the appearance of energy - level crossing to the negative energy spectrum . as a result
, quantum tunneling takes place , leading to electron - positron pairs production , electrons then occupy these semi - classical ( bound ) states and positrons are repelled to infinity . assuming that massive charged cores are bare and non of these semi - classical ( bound ) states are occupied
, we analytically obtain the maximal rate of electron - positron pair production in terms of the core radius , charge and mass .
we find that this rate is much larger than the rate of electron - positron pair - production by the hawking effect , even for very small charge - mass radio of the core given by the naive balance between gravitational and electric forces .
( [ tote ] ) , @xmath221 and @xmath222 , to illustrate the radial location @xmath78 ( [ stable ] ) of stable orbits where @xmath223 has a minimum ( [ mtote ] ) .
all stable orbits are described by @xmath224 .
the last stable orbits are given by @xmath225 , whose radial location @xmath157 and energy @xmath226 .
there is no any stable orbit with energy @xmath227 and the energy - level crossing with the negative energy spectrum @xmath173 is impossible . ] and @xmath228 , we plot the positive and negative effective potentials @xmath229 ( [ tote ] ) , in order to illustrate the radial location ( [ inorbite2 ] ) @xmath230 of stable orbit , where @xmath223 s minimum ( [ mtote3 ] ) @xmath231 is .
all stable orbits inside the core are described by @xmath232 .
the last stable orbit is given by @xmath233 , whose radial location @xmath157 and energy @xmath234 ( [ mtote5 ] ) .
we indicate that the energy - level crossing between bound state ( stable orbit ) energy at @xmath235 and negative energy spectrum @xmath173 ( [ mtote3 ] ) at the turning point @xmath183 . ]
y. b. zeldovich and v. s. popov , sov .
uspekhi 14 ( 1972 ) 673 . m. s. marinov and v. s. popov , pisma v zhetf 17 ( 1973 ) 511 [ jetp lett .
17 , ( 1973 ) 368 ] ; g. siff , b. mller and j. rafelski , zeits .
naturforsch 292 ( 1974 ) 1267 .
review article : w. greiner , b. mller and j. rafelski , `` quantum electrodynamics of strong fields '' monograph in physics , isbn 3 - 540 - 13404 - 2 springer - verlag berlin heidelberg ( 1985 ) ( references there in ) . | classical and semi - classical energy states of relativistic electrons bounded by a massive and charged core with the charge - mass - radio @xmath0 and macroscopic radius @xmath1 are discussed .
we show that the energies of semi - classical ( bound ) states can be much smaller than the negative electron mass - energy ( @xmath2 ) , and energy - level crossing to negative energy continuum occurs .
electron - positron pair production takes place by quantum tunneling , if these bound states are not occupied .
electrons fill into these bound states and positrons go to infinity .
we explicitly calculate the rate of pair - production , and compare it with the rates of electron - positron production by the sauter - euler - heisenberg - schwinger in a constant electric field .
in addition , the pair - production rate for the electro - gravitational balance ratio @xmath3 is much larger than the pair - production rate due to the hawking processes . |
ionic microgel and nanogel particles formed by cross - linked polyelectrolyte networks displays many remarkable properties which make them suitable for applications in drug - delivery systems , where molecules can be encapsulated and then released at specific targets @xcite .
this is possible through a swelling ( or de - swelling ) process , where solvent molecules flow into ( or leave ) the cross - linked network .
experimentally @xcite it is well - known that the properties of ionic microgels and nanogels particles can be significantly affected by the changes in temperature , solvent quality , salt concentration , ionic strength , and degree of cross - linking .
although these effects have been extensively studied in ionic macrogels @xcite , little is known about ionic nanogels formed by finite - size cross - linked polyelectrolyte networks @xcite .
theoretically , the swelling processes in polyelectrolyte networks have been studied @xcite using flory s theory @xcite , which combines the mixing term , electrostatic , and elastic contributions into the total free energy of the system .
an important aspect of the theory is the relationship between the ( effective ) number of chains of the network @xmath3 and the elastic free energy @xcite , _ i.e. _ @xmath4 .
clearly , the unambiguous interpretation of this relationship in terms of the network topology and its validation is relevant to the development of predictive theories , _ e.g. _ @xcite . in principle
, such relationship can be verified by computer simulations considering explicit network structures , but to our knowledge current studies have so far only exploited polyelectrolyte networks with diamond - like structures ( _ i.e. _ with tetrafunctional cross - links ) @xcite . here
we explore this issue by considering three types of nanogel particles generated by polyelectrolyte networks of different topologies characterized by the functionality ( also known as coordination number or connectivity ) @xmath2 of the cross - links .
in particular , we will investigate how the equilibrium properties of the nanosized gel particles are affected by the fraction @xmath1 of the ionizable groups and salt concentration @xmath0 in solution .
these important effects have also been explored by experiments@xcite and theory @xcite .
as illustrated in fig .
[ systemdef : zcoord](a ) , the nanogel particle is modeled as a polyelectrolyte network with @xmath5 monomers inside a spherical wigner - seitz ( ws ) cell of volume @xmath6 , where @xmath7 is the number density of nanogel particles in solution . the radius of the ws cell is then @xmath8 .
some monomers of the network are ionizable ( anionic ) and will dissociate releasing a counterion into solution . the fraction of dissociation @xmath1 determines the number of negatively charged monomers @xmath9 in the network , and charge neutrality requires @xmath10 counterions in solution .
the counterions are allowed to diffuse everywhere inside the ws cell . for simplicity
, we will assume that all the ions are spherical while the solvent , water , is modeled as a dielectric continuum . in order to investigate the effect of topology on the swelling of the nanogel particle , we study three different networks that are arbitrarly generated from a regular structures determined by the functionality @xmath2 of the cross - links .
as illustrated in fig .
[ systemdef : zcoord](b)-(d ) , the networks are built considering cross - links with @xmath11 , @xmath12 , and @xmath13 , that are connected to each other through chains with @xmath14 monomers .
the final networks are obtained by cropping the large regular template structures ( with the number of monomers larger than @xmath5 ) , so that only the @xmath5 monomers inside a spherical volume comprise the nanogel particle .
importantly , we have selected such spherical volume in a way that all the different networks are formed by approximately the same number of monomers @xmath5 , which leads to different number of cross - links @xmath15 with functionality @xmath2 , as shown in table [ parameters ] .
note that this procedure leads to networks with higher effective cross - link line density @xmath16 as @xmath2 is lowered . also , the cropping procedure inevitably results in @xmath17 dangling chains ( which could have less than @xmath14 monomers ) near the surface of the polyelectrolyte network , as illustrated in fig .
[ systemdef : zcoord](a ) . .
the dissociation fraction @xmath1 determines the number of charged monomers with ionizable ( anionic ) groups ( shown in dark blue ) , and the number @xmath18 counterions in solution ( shown in green ) ; neutral monomers are displayed as light blue spheres .
inner circle represents the radius of nanogel particle defined as the radius of gyration @xmath19 of the polyelectrolyte network .
( b - d ) schematic representation of cross - links with functionality @xmath2 that connect chains with @xmath14 monomers and are used to generate the network with different topologies . ]
ccccc + @xmath2 & @xmath5 & @xmath15 & @xmath17 & @xmath20 + 4 & 2808 & 152 & 104 & 0.054 + 5 & 2524 & 108 & 108 & 0.043 + 6 & 2715 & 93 & 78 & 0.035 + + [ parameters ] next we introduce the interaction potentials between the components of the system .
to describe a nanogel particle immersed in an implicit athermal solvent , where there is a slight preference of polyelectrolytes to be surrounded by solvent molecules , our model assumes that all monomers in the network interact via a non - bonded , shifted and truncated repulsive lennard - jones potential , _
i.e. _ the weeks - chandler - andersen @xcite ( wca ) potential commonly used in simulations of polyelectrolytes ( see _ e.g. _
@xcite ) , which is given by @xmath21~,\ ] ] if the distance @xmath22 between @xmath23 and @xmath24 monomers is less than a cutoff radius @xmath25 , or zero otherwise ; here @xmath26 and @xmath27 are parameters that determine the energy and length scales , respectively ; and @xmath28 , where @xmath29 is the temperature and @xmath30 is the boltzmann s constant . in addition , adjacent monomers in the network ( which are defined _ a priori _ by construction ) interact via a finitely extensible nonlinear elastic ( fene ) potential @xcite , @xmath31~~ , \label{fenepotential}\ ] ] where @xmath32 is the spring constant , @xmath33 defines the minimum of the potential ( _ i.e. _ its equilibrium distance ) , and @xmath34 is the maximum extension allowed .
all charged particles ( _ i.e. _ monomers and ions ) interact via an electrostatic potential which can be written as @xmath35 where @xmath36 is the bjerrum length and @xmath37 or 0 , depending of the charges of the interacting particles .
we also include a hard core ( excluded volume ) potential between all particles ( _ i.e. _ monomer - monomer , ion - monomer and ion - ion ) which is @xmath38 if @xmath39 or 0 , otherwise .
this potential precludes any two particles to be closer than a distance @xmath40 . in order to obtain the equilibrium properties of the system we perform
monte carlo ( mc ) simulations with standard metropolis acceptance criteria @xcite , where the transition probability is given by @xmath41}\right)$ ] , with @xmath42 and @xmath43 being the total energetic contribution ( _ i.e. _ wca , fene , electrostatic , and hard core potentials ) evaluated for the new and old configurations of the system , respectively .
new configurations are proposed by attempting to move a single particle and a mc step is defined after all particles in the system have attempted a move .
only diffusive - like random movements are considered , which means that a particle could move in any direction within a maximum displacement with modulus @xmath44 .
the value of @xmath45 is considered here since it provided the optimal choice from previous studies @xcite .
next we present results for a salt - free solution , _ i.e. _ when @xmath46 and the only dissociated ions in the system are those from the polyelectrolyte network .
for all topologies consider , the networks are generated with chains including @xmath47 monomers between the cross - links . in all simulations
the nanogel concentration is set at @xmath48@xmath49 ( approximately @xmath50 m ) , and the parameters of the potentials are given by @xmath51 , @xmath52 , @xmath53 , @xmath54 , and @xmath55 . for each value of dissociation
fraction @xmath1 , the mean values and error bars of the quantities presented below were evaluated using configurations taken from 50 independent simulations ( different seeds ) with runs amounting @xmath56 mc steps . to avoid correlations
, we skip @xmath57 mc steps between samples in the same run , so that a measurement for a given fraction @xmath1 corresponds to at least 2500 different configurations . .
( a ) relative nanogel radius @xmath58 ( inset : swelling ratio @xmath59 ) ; and ( b ) effective nanogel charge @xmath60 ( inset : relative number of positive charges inside a nanogel particle @xmath61 ) . ]
first , we investigate the effects of dissociation fraction @xmath1 on the size of a nanogel particle , which is defined in terms of the radius of gyration @xmath62 , where the average @xmath63 is over all @xmath64 pairs in the polylectrolyte network @xcite .
for instance , when the dissociation fraction is zero , the radius of gyration gives @xmath65 for @xmath11 , @xmath66 for @xmath67 , and @xmath68 for @xmath69 .
we note that these values yields similar results for the volume cross - link densities , @xmath70 with @xmath71 , for all networks considered , which contrasts with the line densities @xmath72 presented in table [ parameters ] .
figure [ radius_and_zeff](a ) shows that the relative nanogel radius @xmath73 increases as the dissociation fraction @xmath1 increases , which is in agreement with what is theoretically predicted in refs .
the increasing behavior of @xmath19 is similar for the different networks , but one can observe that it gets more pronounced for those with low connectivity . another useful measurement to quantify this effect is the swelling ratio @xcite , which can be defined as @xmath74 ) , where @xmath75 and @xmath76 are the volume of the nanogel in two solvency conditions ( _ e.g. _ the swelled and de - swelled states , respectively ) .
hence for a given dissociation fraction @xmath1 the swelling ratio can be simply written as @xmath77 , with @xmath78 being the nanogel radius for @xmath79 . as is shown in the inset of fig .
[ radius_and_zeff](b ) , the highest swelling ratios @xmath59 are achieved by the networks formed by cross - links of low connectivity ( @xmath11 ) .
the dissociated counterions can diffuse around the polyelectrolyte network , so that the nanogel particles acquire an effective net charge @xcite , which is smaller than the bare charge @xmath10 of the polyelectrolyte backbone .
the effective charge is given by , @xmath80 , where @xmath81 and @xmath82 are the number of positive and negative ions inside the nanogel volume , respectively .
here we take @xmath83 because there are no negative ions in the solution when salt concentration is zero ( @xmath46 ) .
figure [ radius_and_zeff](b ) shows that @xmath60 increases with the dissociation fraction @xmath1 , which is consistent with the theoretical predictions of ref .
@xcite . for larger values of @xmath1 the network with functionality @xmath11 presents slightly lower values of @xmath60 in comparison to the other networks .
this occurs mainly because the network with functionality @xmath11 allows for more counterions to become trapped between the polyelectrolytes chains , as shown by the relative number of positive charges inside a nanogel @xmath61 ( inset of fig .
[ radius_and_zeff](b ) ) .
these results corroborate the equilibrium behaviour observed in previous simulations @xcite . however , in contrast to the previous theoretical studies @xcite that predicted a scaling @xmath84 , our results suggest a more linear behavior , which might be related to the relatively small size of the polyelectrolyte networks considered in our simulations . on the size of nanogel particles formed by networks with cross - links of functionality @xmath2 for different values of the dissociation fraction @xmath1 .
( a ) @xmath11 , ( b ) @xmath67 , and ( c ) @xmath69 . ] to better characterize this picture we evaluate the monomer @xmath85 and the ionic @xmath86 density profiles . figure [ density_profiles ] shows @xmath85 and @xmath87 for three different networks at different values of dissociation fraction @xmath1 .
as can be seen in the upper panels , the heterogeneity due to the specific topology of the networks is reflected in monomer density profiles and becomes evident as the dissociation fraction @xmath1 increases and the nanogel swells .
such heterogeneities in density profiles have been observed in simulations and were related both to solvent quality conditions @xcite and to monomer - ion steric repulsion @xcite .
interestingly , it has been speculated that the void created at the center of the nanogel particle might have potential to act as a drug carrier @xcite , but our results suggest that the density profiles are inherent to the structure of the network considered , corroborating recent experimental evidences @xcite . for instance , the monomer density profile in fig . [
density_profiles](c ) of the swelled network for @xmath88 displays a peak when @xmath89 approaches the center of the nanogel ( @xmath90 ) .
this happens because a cross - link with functionality @xmath69 is placed ( by construction ) in the center of that network , which differs from the network with @xmath67 in fig .
[ density_profiles](b ) , where no cross - link is placed at the center of the initially generated network . despite these heterogeneities in @xmath85 , the ionic density profiles for the three networks present similar behaviors , _
i.e. _ showing higher values of @xmath87 ( both inside and outside the nanogel particle ) as the fraction @xmath1 increases .
now we turn our attention to the effect of 1:1 salt on the swelling properties of nanogel particles . in this case , additionally to the @xmath10 positively charged ions due to dissociation of the ionizable groups of the polyelectrolyte network , the dissociation of salt at a given concentration @xmath91 leads to @xmath92 coions ( anions ) and @xmath92 counterions ( cations ) in solution . for simplicity ,
such additional ions are considered to be hard spheres which interact with all charged particles in the system via coulomb ( eq . [ coulombpot ] ) and hard core potentials . as shown in fig .
[ salteffect ] , the effect of the presence of salt on the de - swelling behavior of the network is weaker for the networks with high connectivity .
particularly , for the network with @xmath11 , one can see that while the nanogel radius @xmath19 remains constant for low values of the dissociation fraction , it displays a decreasing behavior for @xmath88 , which is in agreement with the behavior predicted by the theory of ref .
@xcite , and also observed in experiments @xciteand in a previous simulation study @xcite .
we note that even though excluded volume effects are taken into account in our model , we are not able to observe a reentrant ( or re - swelling ) behavior expected for high salt concentrations @xcite .
in summary , we presented an analysis of the swelling behavior of ionic nanogels defined by explicit polyelectrolyte networks formed by cross - links of different topologies .
our results indicate that despite the higher effective cross - link line density @xmath72 , the polyelectrolyte network formed by cross - links with low connectivity ( @xmath11 ) display the highest volume changes both in the presence and in the absence of monovalent salt in the system . in particular , we have verified that both nanogel size @xmath19 and its effective net charge @xmath60 increase as the fraction @xmath1 of ionizable groups in the network increases .
also , there is a clear de - swelling behavior for a increasing concentration of salt in solution for the networks with functionality @xmath11 at high values of @xmath1 .
we found that these effects are dramatically reduced for the networks with functionality @xmath69 , indicating that a higher elastic energy contribution is hampering the change in the size of the nanogel particle , which is confirmed in the appendix section . by considering that the effective number of chains in the network @xmath3 is proportional to the functionality @xmath2 of its cross - links , our results corroborates flory s assumption that @xmath4 , although in his theory the elastic free energy has a purely entropic origin @xcite .
finally , we note that our results are in a qualitative agreement with the swelling behavior theoretically predicted in ref .
a quantitative comparison with the theory may be hampered by the assumptions of implicit chain ( modeled by an elastic energy contribution ) and a flat monomer density inside the nanogel volume , as these approximations might not be appropriate for small - sized nanogel particles studied in the present simulations .
the authors acknowledge alexandre p. dos santos and thiago e. colla for the useful discussions , and the financial support from cnpq ( grant no 165153/2014 - 8 ) , inct - fcx , and by the us - afosr under the grant fa9550 - 12 - 1 - 0438 .
here we present an analysis of the elastic energy obtained from simulations of a salt - free solution where different values of the dissociation fraction @xmath1 yields networks with different sizes , which are described by the ratio @xmath93 ( see fig . [
radius_and_zeff]a ) .
estimates for the changes in elastic energy are obtained as the difference in the equilibrium values of the fene potential ( see eq .
[ fenepotential ] for the pairwise definition ) evaluated for the whole network at two dissociation fraction conditions , _ i.e. _ @xmath94 . as shown in figure [ comparisonflory ]
, the elastic energy contribution is higher for the networks with high connectivity .
47ifxundefined [ 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 * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | it is well - known that the swelling behavior of ionic nanogels depends on their cross - link density , however it is unclear how different topologies should affect the response of the polyelectrolyte network .
here we perform monte carlo simulations to obtain the equilibrium properties of ionic nanogels as a function of salt concentration @xmath0 and the fraction @xmath1 of ionizable groups in a polyelectrolyte network formed by cross - links of functionality @xmath2 .
our results indicate that the network with cross - links of low connectivity result in nanogel particles with higher swelling ratios . we also confirm a de - swelling effect of salt on nanogel particles . |
a major motivation for carrying out n - body simulations of large - scale structure formation is to test theories for the origin of structure and to estimate cosmological parameters , such as the density @xmath10 , or the cosmological constant @xmath7 .
physically accurate identification of the positions , velocities and intrinsic properties of galaxies is necessary if such simulations are to provide estimates of statistics such as the spatial and velocity correlations of galaxies , which can be reliably compared with observations .
dissipationless simulations of gravitational clustering using tens of millions of particles are now carried out routinely on parallel supercomputers .
such simulations are able to resolve the formation and evolution of the dark matter halos of typical galaxies over cosmologically significant volumes .
however , once these galaxy - sized halos merge to form larger structures , such as groups or clusters , they are quickly disrupted and are no longer distinguishable as separate entities within the more massive systems .
this is commonly referred to as the `` overmerging problem '' .
several recent papers have demonstrated that with sufficient force and mass resolution ( @xmath11 kpc and @xmath12 ) the central cores of many simulated galaxy halos do `` survive '' in groups and clusters , but even with this resolution , substructure is still erased in the dense central regions ( tormen , diaferio & syer 1998 ; ghigna et al 1998 ; klypin et al .
1998 ) .
dissipationless simulations do not address the fact that on scales of a few kiloparsecs , gas - dynamical processes are known to play a key role in determining the structure of galaxies . numerical simulations that treat the physics of the baryonic component can be used to overcome this deficiency .
indeed , it has been demonstrated that when radiative cooling is included , many individual dense knots of cold gas do survive within a present - day cluster ( e.g. carlberg , couchman & thomas 1990 ; katz , hernquist & weinberg 1992 ; navarro & white 1994 ; evrard , summers & davis 1994 ; frenk et al .
such simulations require a heavy investment of cpu time and with present - day computer technology , it is difficult to follow the formation of galaxies in volumes that are large enough to study large - scale structure .
in addition , the galaxies themselves are very poorly resolved and processes such as star formation and supernova feedback still have to be put into the simulations `` by hand '' because they are strongly influenced by structure much smaller than the resolution limit . in this paper ,
we introduce a technique for following the formation and evolution of galaxies in cosmological n - body simulations .
dissipationless simulations are used to track the formation and merging of dark matter halos as a function of redshift .
the most bound particle in each halo is identified as the site where cold gas condenses and forms stars .
later , when two or more halos merge in the simulation , these particles maintain their identities as separate galaxies except they can merge with the new central object on a timescale set by dynamical friction .
simple prescriptions are adopted for gas cooling rates , star formation and supernova feedback .
these are based on simplified models of the physical processes or on simulation results and are are taken directly from earlier _ semi - analytic _ studies of galaxy formation . these used analytic models rather than n - body simulations to specify the merging histories of dark matter halos for a given set of cosmological initial conditions .
such semi - analytic models were able to account for many aspects of the present - day galaxy population , for example the luminosities , colours and morphologies of galaxies and the observed correlation of these properties with environment ( white & frenk 1991 ; lacey & silk 1991 ; kauffmann , white & guiderdoni 1993 ; cole et al 1994 ; baugh , cole & frenk 1996b ; somerville & primack 1998 ) .
in addition , semi - analytic models have been used to study the evolution of the galaxy population to high redshift and to make predictions for faint galaxy counts , for the redshift distributions of faint galaxies , for the evolution of the morphology - density relation and for the abundance and properties of galaxies and damped ly@xmath13 absorbers at high redshifts ( lacey et al .
1993 ; kauffmann , guiderdoni & white 1994 ; heyl et al 1995 ; kauffmann 1995a , b ; baugh , cole & frenk 1996a ; kauffmann 1996a , b ; kauffmann & charlot 1998 ; baugh et al 1998 ; mo , mao & white 1998 ) . the combination of semi - analytic galaxy formation models with cosmological n - body simulations results in an important advance .
it becomes possible to track the positions and velocities of galaxies as a function of time , in addition to properties such as stellar mass , luminosity , colour and morphology .
one can then compute the standard statistical measures used to quantify the clustering of galaxies , such as the two - point correlation function or the pairwise peculiar velocities , and investigate how the selection of galaxies by luminosity , colour or type influences the results .
one can extract mock catalogues from the simulations in order to study the clustering properties of galaxies in redshift space .
one can investigate whether the galaxy distribution is _ biased _ relative to that of the dark matter and how this bias affects attempts to measure the density parameter @xmath10 from galaxy surveys .
finally , one can also study how the clustering of galaxies evolves with redshift .
some , but not all of these issues have been explored in previous papers .
white et al ( 1987 ) assumed that a galaxy formed at the centre of a halo during its initial collapse and then tracked the positions of the central particles as a function of time .
the circular velocities of galaxies were given by the the circular velocities of the halos in which they formed .
galaxies were also able to merge if they approached within a certain critical radius of each other .
white et al ( 1987 ) demonstrated that at the present day , galaxies with high circular velocities cluster more strongly than galaxies with low circular velocities .
van kampen & katgert ( 1997 ) used a similar approach to follow the formation of galaxies within n - body simulations of cluster formation . in their scheme ,
each group of bound particles in virial equilibrium was replaced by a single massive particle with a softening corresponding to the radius of the group .
the studies of white et al and van kampen & katgert did not include any scheme for star formation and no attempt was made to compute the luminosities or colours of the galaxies in the simulation .
modelling of cooling , star formation , feedback and stellar evolution is included in the cosmological eulerian grid calculations of cen & ostriker ( 1992,1993 ) . even though these calculations have insufficient resolution to follow the collapse and dynamics of galaxy halos , and so
can not reliably identify individual galaxies , cen & ostriker were able to address some of the issues we discuss in the present paper ( e.g. bias and the dependence of clustering on galaxy age ) .
a number of their conclusions prefigure our own .
the results of semi - analytic models have been used to place galaxies within individual outputs of n - body simulations with higher resolution .
analytic methods are used to compute the luminosity function of galaxies within dark matter halos of given mass @xmath14 and galaxies are then assigned randomly to particles in each simulated halo of this mass .
kauffmann , nusser & steinmetz ( 1997 ) used this method to study present - day bias as a function of galaxy luminosity , colour and type , while governato et al .
( 1998 ) used the same technique to analyze the clustering properties of lyman break galaxies at redshifts @xmath15 .
the disadvantage is that the properties of the galaxies within each simulated halo do not depend on the merging history of that particular halo .
correct results are only obtained if an average is taken over many halos .
furthermore there is no detailed correspondence between the galaxy populations assigned to the simulations at different times so that many questions about the evolution of galaxy clustering can not be directly addressed .
finally , the method is subject to all the inaccuracies of the press - schechter theory and its extensions ( see section 5 below ) .
it is therefore advantageous to compute the merging histories of dark halos directly from the simulations .
this has been done by roukema at al .
( 1997 ) who studied the evolution of both halos and galaxies in simulations with scale - free initial conditions .
their method included prescriptions for star formation and they were able to demonstrate that different assumptions could strongly affect the predicted number density of low - luminosity galaxies .
this paper is the first in a series studying the properties of galaxies identified in n - body simulations with cold dark matter ( cdm ) initial conditions .
these simulations are much larger than those analyzed in previous work and are able to resolve the detailed merging history of the dark matter halo of an @xmath9 galaxy within a volume of @xmath16 mpc@xmath17 .
we first describe the simulations and the methods used to construct halo merging trees .
we outline the prescriptions adopted to model cooling , star formation , supernova feedback , galaxy - galaxy merging and the evolution of the stellar populations within galaxies .
we show how the luminosity functions of galaxies in the simulations compare with those derived from the analytic merging trees of kauffmann & white ( 1993 ) .
we then study the properties of the galaxy distribution at @xmath0 .
these include the i - band tully - fisher relation , b and k - band luminosity functions , the two - point spatial correlation function @xmath18 , two - point velocity correlations , cluster @xmath19 ratios , the @xmath1 colour distributions of galaxies of different types , and the predicted @xmath20 luminosity function .
we focus on two variants of a cold dark matter ( cdm ) cosmology : a high - density ( @xmath2 ) model with shape - parameter @xmath3 ( @xmath4cdm ) , and a low - density model with @xmath5 and @xmath6 ( @xmath7cdm ) .
the normalization @xmath21 is chosen to match the observed abundance of rich clusters in the universe at @xmath0 .
the parameters controlling star formation and supernova feedback are set so that galaxies with circular velocities of @xmath22 km s@xmath8 have an i - band magnitude consistent with the tully - fisher relation of giovanelli et al ( 1997 ) .
we show that different ways of treating star formation and feedback in the models have a strong influence both on the galaxy luminosity function and on the slope and amplitude of the two - point correlation function . in order for the high - density @xmath4cdm model to come close to matching the observations
, we have to suppress the formation of galaxies in low - mass halos in the field .
this is achieved by assuming that supernova feedback is so efficient that it can eject a substantial fraction of the baryons from the potential wells of low - mass dark matter halos .
even so , the model still produces an excess of very bright galaxies in groups and clusters .
the low - density @xmath7cdm model produces _ too few _ star - forming field galaxies , even if feedback is weak and gas never escapes from dark halos .
we have also investigated the effect of dust extinction on predictions for the colour distributions and clustering properties of galaxies .
given the uncertainties in modelling some of the critical physical processes , we conclude that we can not reliably constrain the values of cosmological parameters using the properties of the galaxy distribution at @xmath0 .
the simulations we use were run as a part of the gif project , a joint effort of astrophysicists from germany and israel . its primary goal is to study the formation and evolution of galaxies in a cosmological context using semi - analytical galaxy formation models embedded in large high - resolution @xmath23-body simulations .
the code used for the gif simulations is called hydra .
it is a parallel adaptive particle - particle particle - mesh ( ap@xmath24 m ) code ( for details on the code , see couchman , thomas , & pearce 1995 ; pearce & couchman 1997 ) .
the current version was developed as part of the virgo supercomputing project and was kindly made available by them for the gif project .
the simulations were started on the cray t3d at the computer centre of the max - planck society in garching ( rzg ) on 128 processors . once the clustering strength required a finer base mesh and so an even larger amount of total memory , they were transferred to the t3d at the edinburgh parallel computer centre ( epcc ) and finished on 256 processors .
a set of four simulations with @xmath25 and with different cosmological parameters was run ( table 1 ) .
all the simulations are `` cluster normalized '' .
white , efstathiou , & frenk ( 1993 ) introduced this way of fixing the amplitude by determining @xmath21 , the square root of the variance of the density field smoothed over @xmath26mpc spheres , such that the observed abundance of high - mass clusters is matched .
more recent determinations of @xmath21 using the observed cluster x - ray temperature function ( eke , cole , & frenk 1996 ; viana & liddle 1996 ) yield similar results . for the low - density gif simulations , the result by eke et al .
( 1996 ) was taken . for the @xmath2 simulations , slightly larger values than suggested by eke et al .
( 1996 ) were adopted , according to the earlier result by white et al .
( 1993 ) .
the parameters shown in table 1 were chosen not only to fulfil cosmological constraints , but also to allow a detailed study of the clustering properties at very early redshifts .
the masses of individual particles are @xmath27 and @xmath28 for the high- and low-@xmath10 models , respectively .
the gravitational softening was taken to be @xmath29 .
gravitational lensing by clusters in this set of four models has been studied by bartelmann et al ( 1998 ) . in this paper
we focus on two of the models : @xmath4cdm and @xmath7cdm .
both have shape parameter @xmath3 . a value of @xmath3 is usually preferred by analyses of galaxy clustering , for example peacock & dodds ( 1994 ) .
this is achieved in the @xmath4cdm model , despite @xmath2 and @xmath30 , by assuming that a massive neutrino ( usually taken to be the @xmath4 neutrino ) was present during the very early evolution of the universe and came to dominate the energy density for a short period .
it then decayed into lighter neutrinos which are still relativistic , thus delaying the epoch when matter again started to dominate over radiation .
the neutrino mass and lifetime are chosen such that @xmath3 . for a detailed description of such a model see white , gelmini , & silk ( 1995 ) .
in order to follow the merging history of dark matter halos in the simulations , we store particle positions and velocities at 50 different output times , spaced in equal logarithmic intervals in redshift from @xmath31 to @xmath0 .
the construction of merger trees from these simulation outputs involves the following steps .
a friends - of - friends group - finding program is used to locate virialized halos .
we adopt a linking length which is 0.2 times the mean interparticle separation .
only halos containing at least 10 particles are included in our halo `` catalogues '' .
tests show that 10-particle halos are stable systems .
more than 95 % of 10 particle halos identified at one output time are still located within groups of 10 particles or more at subsequent times . by located " , we mean that more than 80% of the particles in one halo are present in the _ same halo _ at the later time . halos with masses below about 7 particles do not survive over many output times according to this criterion .
the lowest luminosity galaxy that we are able to resolve in these simulations thus corresponds to a galaxy in a halo with 10 particles or @xmath32 . in these models , the milky way halo has a mass of @xmath33 , so the faintest resolved galaxies are roughly a tenth as bright ( i.e. comparable in luminosity to the large magellanic cloud ) .
we then compute a set of physical quantities for all the halos in the catalogues .
these are : 1 . _ the central particle index .
_ this is the index of the most - bound particle in the halo .
this particle has particular significance as it marks the position of the _ central galaxy _ of the halo , ie the galaxy onto which gas cools and where it forms stars .
@xmath34 _ , the virial radius , defined as the distance from the central particle within which the overdensity of dark matter is 200 times the background density .
@xmath35 , _ the virial mass .
this is the mass of dark matter contained within @xmath34 .
@xmath36 , _ the circular velocity ( @xmath37 ) .
we begin with the first simulation output that has at least one halo with 10 particles or more .
the central particles of these halos mark the locations of the first galaxies identified in the simulation .
we then go to the next output time and loop through all the halos in the catalogue , searching for _ progenitor halos _ at the previous time .
a halo at redshift @xmath38 is defined to be a progenitor of a halo at redshift @xmath39 , if 1 .
more than half its particles are included in the halo at redshift @xmath40 ; and 2 .
its central particle is also included in that halo .
the most massive progenitor of a halo has a special status the properties of its central galaxy
are transferred to the central galaxy of the new halo .
we thus _ reposition _ the central galaxy at each output time . the particle index corresponding to the central galaxy changes , but
its associated mass and luminosity evolve in a smooth fashion .
this procedure ensures that cold gas always settles at the centre of a halo .
the central galaxies of less massive progenitors become _ satellites_. the particle index of a satellite galaxy then remains fixed until the present day .
a satellite is said to `` belong '' to a halo if its particle index is among those linked together by the groupfinder .
this procedure is repeated at every output time until @xmath0 . at each timestep , and for all the halos in the associated catalogue , we store the following information : 1 .
the index of the central particle of the largest progenitor of the halo ( zero if the halo had no progenitors ) .
the indices of all satellite galaxies contained within the halo ( particles that were central galaxies of smaller halos at earlier times , and that are now incorporated within the present halo ) this is sufficient to follow the evolution and merging of all galaxies within the simulation volume using the scheme outlined in the next section .
occasionally a galaxy identified at one time is not included within any halo at a subsequent time .
this may happen for the following reasons : 1 .
a halo that was above the 10-particle resolution limit at one time , may fall below this limit at a later time . in this case
, there is a central galaxy that does not belong to any halo in the next simulation output .
satellite galaxies are occasionally `` ejected '' out of halos , particular during mergers . in most cases ,
the satellite will fall back into the halo at a later time .
we keep an index list of these `` lost '' galaxies , checking at each subsequent output time to see whether they have been re - incorporated into a halo .
if the recovered particle was previously a central galaxy , its properties are transferred to the central galaxy of the new halo if the difference in mass between the old and new halos is small ( less than a factor 2 ) .
otherwise , the recovered particle becomes a satellite galaxy within the new halo .
recovered satellite galaxies simply join the satellite population of the new halo .
note that lost galaxies comprise only a few percent of the total galaxy population at all times .
moreover , most of them are faint .
our treatment of the physical processes governing galaxy formation is very similar to that described in kauffmann , white & guiderdoni ( 1993 , kwg ) and in white & frenk ( 1991 ) .
the reader is referred to these papers for more detailed discussion and for derivations of some of the equations .
some of our prescriptions , in particular those for feedback and for galaxy - galaxy merging , have changed since these papers were published .
we have also incorporated a number of extra features , such as dust extinction and starbursts during galaxy - galaxy mergers . for completeness
, we now present a summary of our prescriptions for the various physical processes .
a detailed discussion of how these are implemented is given in section 4.8 .
we adopt the simple model for cooling first introduced by white & frenk ( 1991 ) . for simplicity
, dark halos are modelled as isothermal spheres truncated at their virial radius @xmath34 .
we assume that the hot gas _ always _ has a distribution that exactly parallels that of the dark matter .
the total mass of hot gas in the halos is given by @xmath41 where @xmath42 is the baryon density of the universe , @xmath35 is the virial mass of the halo , @xmath43 is the total mass of stars that have formed in the halo , @xmath44 is the total cold gas contained in the halo and @xmath45 is the mass of gas ejected out of the halo by supernova feedback ( see section 4.3 ) . the gas temperature can then be derived from the circular velocity of the halo using the equation of hydrostatic equilibrium : @xmath46 at each radius in the halo , we define a local cooling time through the ratio of the specific energy content to the cooling rate @xmath47 .
@xmath48 where @xmath49 is the gas density , @xmath50 is the electron density , @xmath51 is the proton mass and @xmath52 is the molecular weight of the gas . in this paper , we do not include chemical enrichment .
the gas is assumed to have solar metallicity at all times and we use the solar metallicity cooling curve in figure 9.9 of binney & tremaine ( 1987 ) .
( see kauffmann 1996b and kauffmann & charlot 1998 for a description of models including chemical evolution ) . at a given redshift @xmath53
, a cooling radius @xmath54 can be defined as the radius within the halo where the cooling time is equal to the age of the universe . for the case of an einstein - de sitter cosmology , @xmath55 at high redshifts , for small halos , and for high gas fractions , the cooling radius is larger than the virial radius of the halo .
we assume that in the absence of supernovae , all the hot gas in the halo would settle to the centre on a timescale given by the halo dynamical time ( @xmath56 ) .
we thus write the instantaneous cooling rate as @xmath57 .
the gas content of the halo is further affected by feedback ( see next section ) and by the infall of new material , which we determine directly from the merging trees . at later times , for larger halos , and for low gas fractions , the cooling radius lies inside the virial radius and the rate at which gas cools is calculated using the equation @xmath58 as has been shown in kwg and in aragon - salamanca , baugh & kauffmann(1998 ) , for massive halos the cooling rates given by equation 6 lead to the formation of central cluster galaxies that are too bright and too blue to be consistent with observation if the cooling gas is assumed to form stars with a standard initial mass function .
it should be noted that cooling flows of hundreds of solar masses per year are _ observed _ in a number of clusters ( see for example fabian , nulsen & canizares 1991 ; allen & fabian 1997 ) , but that the fate of the cooling gas remains a mystery .
one hypothesis is that it may condense into cold clouds instead of stars ( ferland , fabian & johnstone 1994 ) . in our models ,
we assume the gas cooling in halos with @xmath59 km s@xmath8 does not form visible stars .
note that this is the lowest circular velocity at which we can suppress star formation in cooling flows without destroying our fit to the tully - fisher relation ( see section 6.2 ) . as in kwg
, we adopt a simple star formation law of the form @xmath60 where @xmath13 is a free parameter and @xmath61 is the dynamical time of the galaxy . for a central galaxy in a halo , the dynamical time is given by @xmath62 for disk galaxies ,
this is motivated by noting that , if gas collapses to a centrifugally supported state within an isothermal halo while conserving angular momentum , the contraction factor is @xmath63 , where @xmath64 is the spin parameter of the gas , assumed to be the same as for the halo .
n - body simulations find that @xmath64 scatters around a value of @xmath65 . for a satellite galaxy ,
@xmath61 is held fixed at the value when the galaxy was last a central galaxy .
it should be noted that according to the simple spherical collapse model , the virial radius of a dark matter halo scales with circular velocity and with redshift as @xmath66 , reflecting the fact that halos are smaller and denser at earlier epochs .
this means that @xmath61 is independent of the circular velocity of the halo , but will decrease at higher redshift , so star formation rates are higher in halos of the same cold gas content at high @xmath53 .
the star formation law in equation 7 has received considerable _ empirical _ support from a recent study of the star formation rates and gas masses in 61 nearby spiral galaxies and 36 `` starburst '' systems by kennicutt ( 1997 ) , who finds that such a law can fit the data over several orders of magnitude in star formation rate and gas density . in normal spirals , about 10 percent of the available gas is turned into stars per orbital time .
the effect of energy ejected by supernova explosions into the interstellar medium of a galaxy has profound implications for the observed properties of galaxies .
as shown by cole et al ( 1994 ) and somerville & primack ( 1998 ) , strong feedback in low - mass galaxies is required to fit the flat ( @xmath67 ) faint - end slope of the galaxy luminosity function .
kauffmann & charlot ( 1998 ) have demonstrated that substantial feedback in _ massive _ ( @xmath68 ) galaxies is also needed to fit the observed slope of the colour - magnitude relation of elliptical galaxies .
unfortunately , both theoretical and observational understanding of how feedback operates in different types of galaxies is extremely limited at present , so we have no option other than to experiment with a variety of different prescriptions in order to see what difference they make to our results .
using basic energy - conservation arguments , it is possible to estimate how much cold gas could be reheated to the virial temperature of the halo for a given mass of stars formed in the galaxy .
for the scalo ( 1986 ) initial mass function we use to model the evolution of the stellar populations in our galaxies ( see section 4.6 ) , the number of supernovae expected per solar mass of stars formed is @xmath69 .
the kinetic energy of the ejecta from each supernova , @xmath70 , is about @xmath71 erg .
if a fraction @xmath72 of this energy is used to reheat cold gas to the virial temperature of the halo , the amount of cold gas lost in time @xmath73 can be estimated as @xmath74 in the models , @xmath72 is treated as a free parameter . one major uncertainty is whether the gas heated by supernova explosions will leave the halo . in our previous work ,
reheated gas was always retained within the halo .
another possibility is that reheated gas will be ejected out of halo . in the models of cole
et al ( 1994 ) , reheated gas was removed until the halo grew in mass by a factor of two or more , whereupon it was added once again to the hot gas component .
the re - incorporation of expelled gas at a later time ensured that the total baryonic mass in halos was conserved and that the baryon fraction in clusters was close to the global value . in this paper , we will experiment with both feedback prescriptions .
we call the model in which reheated gas is always trapped within the halo the `` retention '' model . the model in which reheated gas is expelled from the halo will be called the `` ejection '' model . according to our star formation prescription ,
stars form efficiently in low mass halos at high redshifts . in the ejection model ,
the energy injected into the ism by the first generation of star formation is sufficient to expel most of the gas from the halo .
the star formation rate then drops and the galaxy fades until the gas is re - incorporated on the next collapse .
stars in low - mass galaxies are thus formed in a series of `` bursts '' associated with each factor 2 doubling in halo mass .
this bursting behaviour is much less pronounced for massive galaxies , since their potential wells are deeper and much less gas will be expelled ( equation 9 ) . the effect of the two feedback prescriptions on the typical star formation history of a @xmath9 galaxy (
more specifically , the central galaxy of a halo with @xmath75 km s@xmath8 ) is shown in figure 1 . as can be seen , the ejection prescription results in both higher and more irregular rates of star formation at high redshift when the galaxies reside within low@xmath36 halos . at low redshifts , the halo potential wells are deeper and so more effective at retaining the gas heated during supernova explosions .
the star formation rates for for the two prescriptions thus do not differ very much at the present epoch .
n - body plus smoothed particle hydrodynamic ( sph ) simulations of the assembly of galaxies in a hierarchical universe show that as dark matter halos coalesce , the embedded disk galaxies merge on a timescale that is consistent with dynamical friction estimates based on their total ( gas + surrounding dark matter ) mass ( navarro , frenk & white 1995 ) .
the dynamical friction timescale , in the form given by binney & tremaine ( 1987 ) is : @xmath76 where @xmath36 is the circular velocity of a singular isothermal sphere representing the primary halo , @xmath77 is the mass of the orbiting satellite , @xmath78 is the coulomb logarithm which we approximate as @xmath79 .
the function @xmath80 allows for the angular momentum of the satellite s orbit , expressed in terms of the circularity parameter @xmath81 , the ratio of the angular momentum of the satellite to that of a circular orbit with the same energy .
lacey & cole ( 1993 ) show that @xmath82 for @xmath83 . finally , @xmath84 is the radius of the circular orbit with the same energy as the satellite orbit .
navarro , frenk & white show that equation 10 provides , on average , a fair estimate of the merger timescale of a satellite galaxy , provided @xmath77 is taken to be the total gaseous + dark matter mass of the satellite . in our models , we use equation 10 to compute the timescale for an accreted satellite to reach the centre of the halo and merge with the central galaxy .
the orbital eccentricity parameter @xmath72 is drawn randomly from a uniform distribution from 0 to 1 .
the radius @xmath85 is set equal to @xmath34 , the virial radius of the primary halo .
@xmath77 is taken as the baryonic mass of the satellite plus the mass of its surrounding halo at the time it was last a central galaxy .
note that it is assumed that satellite galaxies merge only with the central galaxy of the primary halo and only after time @xmath86 .
if the primary halo is later accreted by a larger system , new orbital parameters for all the remaining unmerged satellites are drawn , the dynamical friction timescales are recomputed and the merger clock is reset to zero . in practice ,
substructure in halos will not be erased immediately and for a while a satellite galaxy may still be able to merge with its old central object .
if two galaxies merge and the mass ratio between the satellite and the central object is greater than 0.3 , we add the stars of both objects together and create a bulge component .
if @xmath87 , we add the stars and cold gas of the satellite to the disk component of the central galaxy .
the value of 0.3 is motivated by a series n - body simulations of merging disk galaxies of unequal mass ( barnes , private communication ) . when a bulge is formed by a merger , all cold gas present in the two galaxies is transformed into stars in a `` starburst '' with a timescale of @xmath88 years .
further cooling of gas in the halo may lead to the formation of a new disk .
the morphological classification of galaxies is made according their b - band disk - to - bulge ratios ( simien & de vaucouleurs 1986 ) . if @xmath89 mag , then the galaxy is classified as early - type ( elliptical or s0 ) .
it should be noted that although we track the formation of galaxies in halos as small as 10 particles , we do not accurately predict the morphologies of galaxies contained in such halos , simply because their merging histories are not resolved .
accurate morphologies are only obtained for central galaxies in halos with @xmath90 100 particles , i.e. for galaxies with luminosities @xmath68 .
we use the new stellar population synthesis models of bruzual & charlot ( in preparation ) , which include updated stellar evolutionary tracks and new spectral libraries .
in this paper , all stars are assumed to have solar metallicity .
the star formation history of any galaxy can be approximated by a series of delta - function `` bursts '' of different masses .
the stellar population models are used to generate lookup tables of the luminosity of a burst of fixed mass as a function of age in each photometric band .
the magnitude of the galaxy at @xmath0 is calculated by summing the mass - weighted luminosities of each burst .
we have adopted a scalo ( 1986 ) initial mass function with upper and lower mass cutoffs of 100 @xmath91 and 0.1 @xmath91 .
as shown in kauffmann & charlot ( 1998 ) , the stellar mass - to - light ratios of an old ( 8 - 10 gyr ) stellar population are then in good agreement with the observed mass - to - light ratios of elliptical galaxies within an effective radius . in the models there is a strong distinction between central and satellite galaxies .
gas cools continuously onto central galaxies and they form stars at a roughly constant rate .
satellite galaxies lose their supply of new gas and , as a result , have exponentially declining star formation rates , leading to redder colours and lower gas fractions than central galaxies .
a central galaxy is thus considerably brighter than a satellite companion of the same stellar mass at ultraviolet and optical wavelengths .
as we will show in section 6.4 , this influences the amplitude of the correlation function of galaxies selected according to b - band magnitude .
it is a well - known but oft - ignored observational fact that gas - rich , star - forming galaxies contain dust , which absorbs a substantial fraction of the light emitted in the short wavelength part of the spectrum and re - radiates it in the far - infrared .
we include a simple , empirically - motivated recipe for dust extinction in our models in order to investigate how much this effect can influence our estimates of clustering in the models .
wang & heckman ( 1996 ) have studied the correlation of the optical depth of dust in galactic disks with the total luminosity of the galaxy .
they studied 150 normal late - type galaxies with measured far - ultraviolet ( uv , @xmath92 ) fluxes and compiled the corresponding far - infrared ( fir , @xmath93 ) fluxes measured by the iras satellite .
they then modelled the absorption and emission of radiation by dust with a simple model of a uniform plane - parallel slab in which the dust that radiates in the iras band is heated exclusively by uv light from nearby hot stars .
they find that their observed uv - to - fir ratios can be explained by the face - on extinction optical depth @xmath4 varying with the intrinsic uv luminosity of the galaxy as @xmath94 with @xmath95 the same scaling law was also able to account for the @xmath96 ratios measured for a subset of the galaxies .
re - expressed in the blue - band , wang & heckman derive a total extinction optical depth of @xmath97 at the fiducial observed blue luminosity of a schechter @xmath9 galaxy ( @xmath98 ) .
we use the galactic extinction curve of cardelli , clayton & mathis ( 1989 ) to derive @xmath99 . although the extinction curves of smaller , less - metal rich galaxies such as the lmc and smc are quite different at ultraviolet wavelengths , at optical wavelengths they are all fairly similar .
this leads to face - on extinction for an @xmath9 disk galaxy of 0.9 , 0.8,0.4 and 0.08 mag in the b , v , i and k bands respectively .
finally , one must also take into account the inclination of the galaxy to the line - of - sight when making an extinction correction . for a thin disk where dust and stars
are uniformly mixed , the total extinction in magnitudes is @xmath100 in our dust - corrected models , we apply an extinction correction to every galaxy forming stars at a rate greater than 0.5 @xmath101 .
we use @xmath102 in equation 11 , and scale @xmath103 using the _ uncorrected _ b - magnitudes predicted by the model .
we then pick a random inclination for each galaxy and apply the final correction using equation 12 . in the simulation ,
each galaxy carries a number of `` labels '' corresponding to different physical properties .
these are : 1 .
@xmath104 , total stellar mass 2 .
@xmath105 , stellar mass of the bulge component 3 .
@xmath44 , cold gas mass 4 .
@xmath106 , total present - day luminosity in a given waveband 5 .
@xmath107 , the present - day bulge luminosity satellite galaxies carry two additional labels : 1 .
@xmath108 , the time until the satellite galaxy should merge with the central object .
@xmath109 , the index of the halo in which the satellite resided at the previous timestep . in the ejection feedback scheme
, each _ halo _ carries an array @xmath110 , which is the amount of gas ejected by galaxies in progenitor halos of mass @xmath111 . at each output time
, we loop over all the halos in the catalogue and follow a series of steps : 1 .
we re - incorporate gas that was ejected out of progenitor halos more than a factor of two less massive than the present halo 2 .
we calculate the mass of hot gas in the halo at the start of the timestep .
this is given by @xmath112 where @xmath113 is the sum over all galaxies in the halo .
note that this assumes that when halos merge with each other , all gas that is not already cooled is shock heated to the virial temperature of the new halo .
the sum over @xmath111 extends only over those progenitors more massive than @xmath114 .
we loop through the satellite galaxies in the halo and update their merging timescales .
if the satellite was not in the largest progenitor of the halo at the previous timestep , its merging clock is reset .
if the galaxy is predicted to merge during the timestep , it is flagged .
the time between the beginning of the timestep and the merging event is noted .
we then solve a set of coupled differential equations for the time evolution of the cold gas and stars in each of the galaxies in the halo .
we adopt a small timestep for this calculation ( there are typically 100 timesteps between each pair of simulation outputs ) . for central galaxies ,
the change of the cold gas component over time @xmath73 is given by : @xmath115 where @xmath116 , @xmath117 and @xmath118 are given by equations 5,6 and 7(or 9 ) respectively .
for satellite galaxies , the equation is the same except there is no source term from gas cooling in the halo .
infall of cold gas onto satellites is assumed to be _ disrupted _ as soon as they are accreted so that star formation within them continues only until their existing cold gas reservoirs are exhausted . +
the stars formed during time @xmath73 are added to @xmath104 .
we look up the present - day luminosity of a burst of mass @xmath119 and age @xmath120 in our population synthesis tables and update the luminosities of the galaxies . finally , the hot gas mass of the halo at the end of the timestep is calculated using @xmath121 note that this is valid for retention feedback . for ejection feedback ,
reheated gas is added instead to @xmath122 , where @xmath35 is the virial mass of the halo .
if a satellite is predicted to merge during the timestep @xmath73 , its stars and gas are added to the central galaxy .
if the mass of the satellite is greater than one - third the mass of the central object , we set the stellar mass of the bulge component equal to the mass of the central galaxy plus the mass of the merged satellite .
all the cold gas in the merger remnant is converted into stars at a constant rate over @xmath88 years .
stars that form in such a burst do not reheat or eject cold gas .
in addition , we assume that the the burst is decoupled from the evolution of rest of the galaxy , i.e. we set @xmath123 in the above equations immediately after the merger takes place . both assumptions are motivated by n - body simulations , which show that gas in merging disk galaxies quickly loses angular momentum and ends up in a dense knot at the very centre of the remnant , and that feedback has rather little effect ( barnes & hernquist 1996 ) . for a given set of cosmological initial conditions ( including the baryon density @xmath42 )
, there are two `` free parameters '' in our models the star formation efficiency @xmath13 and the feedback efficiency @xmath72 .
as in kwg , we tune these two parameters to match the luminosity and cold gas mass of a fiducial reference galaxy , which we take to be the central galaxy in a halo with circular velocity @xmath124 km s@xmath8 . in most previous work
we have normalized the models using standard parameters for the milky way , but here , following the suggestions of somerville & primack ( 1998 ) , we adopt a normalization that is consistent with the velocity - based zero - point of the i - band tully - fisher relation as determined by giovanelli et al ( 1997 ) . matching to i- band rather than
b - band data ought to be more robust , since the i - band is considerably less affected by dust extinction and star formation . moreover , a very large amount of data on the i - band tully - fisher relation is now available .
giovanelli et al obtain the following fit , based on 555 galaxies in 24 clusters : @xmath125 to convert between the measured @xmath126 line - widths @xmath127 and the model circular velocities we assume @xmath128 . in practice , there is substantial uncertainty in the relation between the circular velocity of the halo and that of the disk ( e.g. mo , mao & white 1998 ) , since the transformation between the two quantities depends on the detailed density profile of the halo and whether or not gas loses angular momentum before settling onto the disk .
we set the parameters @xmath13 and @xmath72 so that the central galaxy in a halo with @xmath129 km s@xmath8 has @xmath130 and a cold gas mass @xmath131 .
it should be noted that this normalization is considerably brighter than that adopted by kwg , who took the observed b - band magnitude of the milky way to be @xmath132 . for a hubble constant of 50 km s@xmath8 mpc@xmath8 and assuming @xmath133 for a spiral galaxy ( de jong 1996 )
, the new normalization means that the central galaxy in a halo with @xmath124 km s@xmath8 has a b - band magnitude of -21.8 more than a factor 3 more luminous than the observed value .
consistency of this tully - fisher normalization with the estimated magnitudes of our own galaxy and of m31 is only obtained for substantially higher values of the hubble constant , which do not give acceptable ages for a universe with @xmath2 .
the fact that there are only two free parameters in the model should not be interpreted as an indication that the physical processes governing galaxy formation are well - specified . on the contrary
, we will show that the way in which we choose to implement feedback and whether or not dust is included in the model can have a very large effect both on the luminosity function and on the amplitude of the correlation function on scales below a few mpc .
we will attempt to clarify how different parametrizations of these processes affect our results and which processes appear to be most critical if one is to succeed in matching the observational data .
if halo merging trees derived using methods based on the extended press - schechter formalism ( cole 1991 ; kauffmann & white 1993 ; rodrigues & thomas 1996 somerville & kolatt 1998 ) were statistically equivalent to the trees found in the n - body simulations , the same galaxy luminosity functions would be obtained using the two approaches , provided the same recipes for cooling , star formation and feedback were employed , and the same halo mass resolution limit was adopted . in practice , the halo mass function derived by the press - schechter argument does not fit very accurately the mass function in n - body simulations . for halo masses between @xmath134 and @xmath135 @xmath91 ,
the press - schechter theory predicts roughly twice as many halos as are actually found in the gif simulations for both the @xmath4cdm and the @xmath7cdm cosmologies .
most of the missing mass in the simulations is in the form of `` unresolved '' material
single particles or groups with less than 10 members .
this discrepancy between the press - schechter and simulation mass functions propagates to the halo progenitor distributions and to the monte - carlo merging trees ( somerville , lemson & kolatt , in preparation ) .
here we compare the the predictions of the two approaches for the luminosity functions of galaxies within halos of given mass and for the `` field '' luminosity function . in figure 2
, we plot the number of galaxies brighter than a given v - magnitude in halos with circular velocity @xmath36 .
solid squares show mean results derived using analytic trees constructed as described in kauffmann & white ( 1993 ) .
solid circles show results derived from the simulation .
the error bars on the points show the rms scatter in the number of galaxies in halos of given circular velocity .
we perform the comparison using the @xmath4cdm model and we use the _ same _ prescriptions ( from kwg ) for cooling , star formation , feedback , merging and stellar population synthesis in both cases .
we truncate the analytic trees at a halo mass of @xmath136 so that the resolution is the same as in the simulations .
the four panels in figure 2 show the number of galaxies per halo brighter than @xmath137 and @xmath138 .
as can be seen , the mean number of lower - luminosity galaxies per halo ( @xmath139 ) agrees remarkably well .
the analytic approach appears to overpredict the number of bright galaxies in high @xmath36 halos by a factor of @xmath140 , but since there are not many such halos in the simulation and the number of bright galaxies within them is small , the discrepancy may not be statistically significant .
figure 3 compares the field luminosity functions derived using the semi - analytic trees and the simulations .
the top panel shows the difference between the abundance of halos predicted by press - schechter theory , and that found in the simulation .
the press - schechter overprediction of halo abundance at intermediate masses in quite evident .
the middle panel compares the magnitude of the _ central _ galaxy as a function of halo circular velocity .
there is good agreement between the two approaches .
note that the dip in magnitude at @xmath141 km s@xmath8 arises because we assume stars do not form in the cooling flows of halos with circular velocities greater than this value .
the bottom panel compares the v - band `` field '' luminosity functions .
except perhaps near the `` knee '' , the agreement between the two approaches is excellent .
the comparison in this section demonstrates that although merging trees derived using the extended press - schechter theory differ in detail from those found in the simulations , the two methods give very similar results both for the mean number and luminosity of galaxies within halos , and for the scatter in these quantities .
in figure 4 , we show slices of thickness 8 @xmath142 mpc from the @xmath4cdm simulation . the distribution of dark matter is shown in the top left panel . in the top right panel , all the galaxies in the slice with magnitudes brighter than @xmath143 are plotted as solid white circles . galaxies with magnitudes brighter than @xmath144 are shown in the bottom panels .
we divide these galaxies into two equal subsamples at the median star formation rate per unit stellar mass .
the bottom left panel shows the distribution of galaxies with low star formation rates .
the bottom right panel shows galaxies with high star formation rates .
galaxies of all luminosities trace the filaments and knots visible in the dark matter distribution .
galaxies with high star formation rates are less clustered than galaxies with low star formation rates .
star - forming galaxies tend to avoid dense clusters and groups and occur more frequently in voids .
figure 5 shows a similar set of slices from the @xmath7cdm simulation .
the area and thickness of these slices have been chosen to have the same dimensions in _ redshift space _ as those in figure 4 ( @xmath145 mpc @xmath146 mpc @xmath147 mpc ) .
the same systematic effects are visible in the two figures and the most obvious difference between them is the difference in galaxy abundance .
we return to this in section 6.3 .
as described in section 4.9 , all models are normalized so that the average i - band magnitude of the central galaxy in a halo of circular velocity @xmath129 km s@xmath8 is @xmath148 , in accordance with the zero point of the i - band tully - fisher relation derived by giovanelli et al .
figure 6 shows the tully - fisher relation for spiral galaxies in the simulations .
we have selected central galaxies with @xmath149 ( appropriate for sb / sc type galaxies ) .
the solid line is the fit to the data ( equation 16 ) .
the upper two panels show the @xmath4cdm model for the two different feedback prescriptions discussed in section 4.3 .
although the slopes are similar , the ejection prescription results in considerably more scatter , particularly at low velocity widths , where there is a marked tail of low - luminosity galaxies .
the lower panel shows the @xmath7cdm model with retention feedback . we do not show a @xmath7cdm model with ejection feedback as this model would result in too low a total luminosity density , as explained in the next section .
it is interesting that the scatter obtained for both the @xmath7cdm and @xmath4cdm relations , even for retention feedback , is not much smaller than the observed scatter .
the model scatter arises because of differences in star formation history and stellar mass between central galaxies in different halos of the same circular velocity .
as pointed out by mo , mao & white ( 1998 ) , scatter in the spin parameter of dark matter halos leads to a spread in contraction factors for disks forming in halos of a given circular velocity .
halos with high values of @xmath64 produce disks with rotation curves that are slowly rising , whereas halos with low values of @xmath64 produce disks with rotation curves that rise steeply and then decline .
this leads to intrinsic scatter in the predicted tully - fisher relation even if disks in halos of given circular velocity are assumed to have fixed mass and mass - to - light ratio .
this must be added to the scatter already visible in figure 6 .
more complex modelling is necessary to investigate these issues in detail . in figure 7 , we show b - band luminosity functions for the @xmath4cdm simulation .
a sequence of different models is shown in order to illustrate the effect of including different prescriptions for feedback and dust extinction .
the results are compared with the b - band luminosity functions derived from several recent redshift surveys : the apm - stromlo survey ( loveday et al 1992 ) , the las campanas redshift survey ( lcrs ) ( lin et al 1997 ) and the eso - slice survey ( zucca et al 1997 ) .
note that we only plot the luminosity functions down to magnitudes where our simulations are `` complete '' ( i.e. the luminosity of the central galaxy in a halo with 10 particles ) . in
the top left panel , we plot the luminosity function for a model with retention feedback and no dust extinction . in this case
, the model luminosity function exceeds the observed one by more than a factor of 10 at all luminosities .
the inclusion of dust brings the model more in line with the observations at luminosities around @xmath9 , but there are still far too many galaxies at both fainter and brighter magnitudes .
the model with ejection feedback agrees much better with the observations at @xmath9 and below .
this is because ejection feedback suppresses the formation of bright galaxies in low mass halos .
the first generation of stars formed when the halo collapses generates enough energy to empty the halo of gas and halt further star formation .
this process is only efficient in low - circular velocity systems , so the excess of galaxies brighter than @xmath9 still remains a problem . including dust in the ejection model ( lower right panel ) helps somewhat , but the predicted abundance of bright galaxies is still more than a factor 10 above the schechter function representation of the observations . in figure 8
, we compare the model k - band luminosity functions with the schechter fits derived by gardner et al ( 1997 ) and szokoly et al ( 1998 ) for two independent surveys . unlike b - band magnitudes , the k - band luminosities of galaxies depend on their total stellar masses , rather than their present - day star formation rates .
moreover , k - magnitudes are only weakly affected by dust extinction .
the fact that the model k - band luminosity functions also disagree with the data at the bright end shows that there are too many _ massive _ galaxies in the simulations .
b - band and k - band luminosity functions for the @xmath7cdm model are shown in figure 9 .
retention feedback is assumed .
even so , this model produces a factor @xmath150 too few galaxies at magnitudes around @xmath9 .
there is again an excess of very bright galaxies , but the problem is less severe than in @xmath4cdm .
the same trends are apparent in the k - band . in summary ,
the luminosity functions for both models have the wrong shape , with or without dust extinction . instead of declining exponentially at bright magnitudes , they exhibit a gentler turn down . in previous work , kwg and cole et al ( 1994 ) obtained much better fits to the bright end of the luminosity function ( see also figure 3 ) . as explained in section 4.9 ,
this is because they adopted a much fainter normalization .
if one is to obtain an exponential cutoff at high luminosities , bright galaxies must reside primarily in very massive halos on the exponentially declining part of the mass function .
the shape problem is considerably worse for the @xmath4cdm model than for the @xmath7cdm model .
although both models are normalized to reproduce the observed abundance of rich clusters , the @xmath4cdm model has a higher abundance of intermediate mass halos than the @xmath7cdm model .
it is the luminous galaxies in these halos that create the excess at magnitudes brighter than @xmath9 .
the @xmath4cdm model produces too many galaxies below @xmath9 unless gas is efficiently ejected out of low mass halos .
the @xmath7cdm model , on the other hand , produces too _ few _ @xmath9 galaxies even if feedback is inefficient and if gas never escapes these halos . in figure 10
, we show the galaxy two - point correlation function @xmath18 for the @xmath4cdm simulation .
we select galaxies with b - band luminosities @xmath151 . a sequence of models is again shown to illustrate different prescriptions for feedback and dust extinction .
the results are compared with the real space correlation function measured from the apm galaxy survey ( baugh 1996 ) . for the model with retention feedback
, @xmath18 follows a power law of slope @xmath152 down to scales @xmath153 mpc , below which it flattens and then turns over .
this is in marked contrast to the observed @xmath18 , which follows the same power law down to scales below 100 kpc .
the inclusion of dust extinction lowers the luminosity of bright star - forming field galaxies and raises the amplitude of the correlation function on small scales , but @xmath18 still flattens and is a factor 6 below the observed value at @xmath154 kpc . as described in the previous section , ejection feedback suppresses bright star - forming galaxies in low circular velocity halos and brings the faint end of the luminosity function into better agreement with the observations . as seen in the bottom left panel of figure 10 , ejection feedback also prevents @xmath18 from flattening below 1 mpc .
the results shown in the bottom right panel of figure 10 , for the model with ejection feedback and dust extinction , agree rather well with the observations .
figure 11 shows the galaxy two - point correlation function for the @xmath7cdm simulation .
retention feedback is assumed .
@xmath18 does not turn over on small scales and the model with dust is actually _ too steep _ compared to observations .
= 9.3 cm what determines the slope and amplitude of the galaxy correlation functions in our models ?
in a recent paper , mo & white ( 1996 ) developed a simple analytic model for the gravitational clustering of dark matter halos .
the positions and formation times of halos were determined from the statistics of the initial linear density field and modifications caused by gravitationally - induced motions were treated using a spherical collapse approximation .
mo & white showed that on large scales , the halo autocorrelation function is simply proportional to the dark matter correlation function , with the constant of proportionality dependent on the masses of the selected halos .
because halos are spatially exclusive and have a finite radius , the halo - halo correlation function drops below the mass correlation function on scales comparable to the diameter of typical halos in the sample . from these results
, it follows that if there is only one bright galaxy per halo , the galaxy correlation function will simply follow the form of the halo correlation function , with a flattening and turnover on scales less than @xmath153 mpc . in order for the galaxy correlation to continue as a power law to small scales ,
a substantial fraction of bright galaxies must exist as groups _ within the same halo_. this is illustrated quantitatively in figure 12 , where we plot the fraction of galaxies in the simulation with @xmath155 that occur in halos of a given circular velocity .
the @xmath4cdm model with retention feedback has a much higher fraction of galaxies in low @xmath36 halos than the @xmath4cdm model with ejection feedback .
low @xmath36 halos typically contain only one bright galaxy ( see figure 2 ) , whereas high @xmath36 halos contain many bright galaxies that contribute to clustering amplitude on sub - megaparsec scales .
figure 12 shows that there is a relatively small fraction of bright galaxies in low @xmath36 halos in the @xmath7cdm model , even with retention feedback .
recall that both @xmath4cdm and @xmath7cdm are normalized to fit the observed number density of rich clusters at the present day . because the mass density in the @xmath7cdm model is lower , the number density of low mass halos is also smaller .
@xmath7cdm thus `` naturally '' has more galaxies in clusters relative to the field and as a result , @xmath18 is steep on small scales .
figure 13 shows the first ( @xmath156 ) and the second moment ( @xmath157 ) of the one - dimensional radial pairwise velocity distribution of galaxies and dark matter particles as a function of pair separation for our best - fit models .
bold lines are the moments for the dark matter particles .
solid and dotted lines are the moments for galaxies brighter than @xmath158 and @xmath159 , respectively .
brighter cutoffs yield similar results .
galaxies do not show any significant velocity bias on any non - linear scale regardless of their luminosity .
this result is not trivial because galaxies are not a random subsample of dark matter particles .
this result is consistent with the weak density bias and the lack of luminosity segregation described in sect .
the pairwise velocity dispersion profile @xmath160 is sensitive to the presence of clusters within the sample and therefore to the total volume of the sample .
our simulation box has a volume of the order of @xmath161 mpc@xmath24 which is comparable to the volume @xmath162 mpc@xmath24 within @xmath163 km s@xmath8 of the northern slice of the center for astrophysics redshift surveys ( hereafter cfa2n , de lapparent et al 1991 ; geller & huchra 1989 ; huchra et al 1990 ; huchra et al 1995 ) . filled squares in figure 13 show @xmath160 for the cfa2n slice .
both @xmath4cdm and @xmath7cdm models are in remarkably good agreement with the cfa2n slice . lacking the full three - dimensional information , we compute the cfa2n values by fitting the redshift space correlation function @xmath164 with the convolution model ( fisher 1995 ) : we model @xmath164 as the convolution of the real space correlation function with an exponential pairwise velocity distribution ( davis & peebles 1983 ; details of this procedure
are given in marzke et al 1995 ) . in order to apply this procedure , we need to assume a model for the mean streaming velocity @xmath165 .
we adopt the similarity solution suggested by davis & peebles ( 1983 ) @xmath166 $ ] where @xmath167 mpc is the cfa2n galaxy - galaxy correlation length .
assuming @xmath168 usually yields lower values of @xmath157 ( see e.g. marzke et al 1995 ; somerville , primack & nolthenius 1997 ) .
note that our simulations show @xmath169 km s@xmath8 at relative separations @xmath170 mpc and do not support the assumption of a null @xmath165 . in future papers ,
we use mock redshift surveys to show that this procedure yields pairwise velocity dispersion profiles @xmath160 in reasonably good agreement with the real profile .
it is therefore encouraging that the @xmath160 profiles of our best - fit models match the cfa2n result .
to11.cm ' '' '' an important issue we wish to address is that of galaxy `` bias '' , i.e. whether the clustering amplitude of the galaxies in our models differs from that of the underlying dark matter . in figure 14 , we compare @xmath18 for galaxies of different luminosities , morphological types and colours with @xmath18 for the dark matter . for @xmath4cdm , we show the model with ejection feedback and dust extinction and for @xmath7cdm , we show the model with retention feedback and dust extinction .
these are the `` best fit '' models in both cases . . on large scales , galaxies of all luminosities
are unbiased in both models . on small scales ,
the @xmath4cdm galaxies are slightly more clustered than the dark matter , whereas the @xmath7cdm galaxies are a factor @xmath150 less clustered .
neither model displays any luminosity segregation ; galaxies of all magnitudes cluster in the same way . as discussed by kauffmann , nusser & steinmetz ( 1997 )
, positive bias is obtained for galaxies in halos more massive than @xmath171 , the mass of the typical collapsed object at the present day . for cluster - normalized cdm models , @xmath104 is large ( @xmath172 ) .
our bright normalization also causes luminous galaxies to be placed in relatively low mass halos .
this is why we do not see the luminosity - dependent bias discussed by kauffmann , nusser & steinmetz ( 1997 ) . on the other hand , red galaxies and to a lesser extent
early - type galaxies are more clustered than the underlying population , particularly on small scales .
these galaxies occur preferentially in the cores of groups and clusters ( see figures 4 and 5 ) .
conversely , star - forming galaxies are less clustered on small scales because they are predominantly central halo galaxies and are thus spatially exclusive .
a similar result can be seen in the earlier work of cen & ostriker ( 1992 ) .
the relative bias between different types of galaxies in the southern sky redshift survey has recently been analyzed in a paper by willmer , da costa & pellegrini ( 1998 ) .
these authors find that galaxies with magnitudes below @xmath68 do not exhibit any luminosity - dependent bias , but at magnitudes brighter than @xmath9 , galaxies are significantly more strongly clustered ( @xmath173 ) .
as discussed above , our models do not exhibit this luminosity - dependent bias , perhaps indicating that our adopted normalization is too bright .
willmer , da costa & pellegrini also find that both early - type galaxies and red galaxies cluster more strongly than the galaxy population as a whole .
the bias is stronger on small scales ( @xmath174 @xmath142 mpc ) than on larger scales .
in addition , galaxies with _ colours _ characteristic of old stellar populations exhibit more bias than galaxies with early - type morphologies .
these results are in good qualitative agreement with the results shown in figure 14 and the earlier results of kauffmann , nusser & steinmetz ( 1997 ) .
clusters of galaxies are the largest collapsed objects in the universe and have often been used as a means of estimating the cosmological density parameter @xmath10 .
the standard procedure is as follows .
the mass of a cluster within some radius is measured using techniques such as galaxy kinematics , x - ray profiles or gravitational lensing .
this is compared to the luminosity within the same radius to calculate a mass - to - light ratio .
the mean mass density of the universe is then estimated by multiplying its mean luminosity density by this number .
the validity of this technique depends on two questionable assumptions : 1 .
luminosity is `` conserved '' when field galaxies fall into a cluster . in practice ,
we know that cluster galaxies are predominantly ellipticals with no ongoing star formation , whereas field galaxies are predominantly spirals in which stars are forming at several solar masses per year
. the stellar mass - to - light ratios of field and cluster galaxies thus differ strongly at optical and uv wavelengths
the efficiency of galaxy formation is the same in all environments .
most studies of cluster mass - to - light ratios give values corresponding to @xmath175 ( see for example carlberg et al 1996 ) .
if a high - density cosmology is to be consistent with these observations , we require the mass - to - light ratios of clusters to be significantly _ lower _ than the mass - to - light ratio of the universe as a whole .
this is often also referred to as `` bias '' in the literature , though this kind of bias is quite different to what we measure when we compare the correlation functions of galaxies and dark matter .
figure 15 shows the mass - to - light ratios of our simulated halos divided by the mass - to - light ratio of the simulation as a whole .
the results are shown as a function of the virial mass of the halo and for both b and i - bands .
it is interesting that in the @xmath2 @xmath4cdm model , the mass - to - light ratios of clusters _ are _ significantly lower than the global value .
the factor ( @xmath140 ) that we obtain is somewhat too small , however , to bring the model into good agreement with the observations .
rich clusters in the @xmath7cdm model have mass - to - light ratios only 10 - 20 % smaller than the global value .
it is also interesting that groups of galaxies are predicted to have lower mass - to - light ratios than clusters in both models .
this agrees with the observed trends .
we now compare the @xmath1 colour distributions of our model galaxies with colours of bright nearby galaxies in the rc3 catalogue ( de vaucouleurs et al 1991 ) .
we have selected 1580 galaxies with @xmath176 , which have both morphological classifications and magnitudes in the johnson @xmath177 and @xmath178-bands . of these galaxies ,
653 are classified as early - type .
note that the fraction of early - type galaxies is rather high , which may reflect the fact that the sample is dominated by galaxies in the virgo cluster . in our simulation @xmath179 of galaxies to this magnitude limit have @xmath89 mag and so are classified as early types . in figure 16
, we compare the colour distributions of galaxies in the @xmath4cdm model with that of the rc3 sample .
the left panel shows the model with ejection feedback , but without any dust extinction .
the right panel shows the same model with dust extinction . without dust extinction , the model colour distribution peaks at @xmath180 in contrast to the observed distribution , which is relatively flat from @xmath181 to @xmath182 .
in particular , early - type galaxies appear to be significantly too blue in this model . with the inclusion of dust ,
the color distribution broadens significantly .
many of the field ellipticals on our models have a lower - luminosity , star - forming disk component .
the inclusion of dust in the models reddens these objects and brings the colour distribution of early - type galaxies into better agreement with the observations .
the colour distributions of galaxies in the @xmath7cdm models are shown in figure 17 . because galaxies form earlier in this cosmology , the colours tend to be redder on average , but the effect is small .
again , the effect of dust extinction is to broaden the colour distributions and to bring them into better agreement with the observations .
in addition to the luminosity function and the colour distributions of galaxies , observational estimates of the number density of galaxies as a function of star formation rate provide an important constraint on galaxy formation theories .
gallego et al ( 1995 ) have analyzed a complete sample of emission - line galaxies and have computed an h@xmath13 luminosity function . assuming a scalo ( 1986 ) imf and using case b recombination theory to predict the luminosity of the h@xmath13 emission line for a given rate of star formation , they derive an integrated star formation rate density of the local universe of @xmath183 mpc @xmath184 ( for @xmath185 km s@xmath8 mpc@xmath8 ) . in figure 18 , we compare the h@xmath13 luminosity functions of our models with the results of gallego et al .
we have used the same transformation as these authors to convert from star formation rate to h@xmath13 luminosity , @xmath186 the results of the @xmath4cdm model agree better with the observations than those of the @xmath7cdm model . for @xmath4cdm
, we obtain an integrated sfr density of 0.017 @xmath91 yr@xmath8 mpc@xmath184 . for @xmath7cdm
the present - day sfr density is a factor @xmath187 too low . from figure 18
, we see that there there is a lack of galaxies forming stars at rates between 1 and 10 @xmath91 yr@xmath8 .
these are the same galaxies that are missing at the `` knee '' of the b - band luminsity function in figure 9 .
none of the models analyzed in this paper provides a completely satisfactory fit to all the observed properties of galaxies at @xmath0 .
if supernova feedback is weak , the @xmath2 @xmath4cdm model produces too high a total luminosity density .
most of the excess luminosity is in the form of star - forming `` field '' galaxies . as a result
, the galaxy correlation function turns over on scales below a few mpc . the observed two - point correlation function , on the other hand , is well represented by a single power law of slope @xmath188 down to scales approaching 10 kpc ( baugh 1996 ) .
if one assumes instead that supernova feedback is able to eject gas out of a halo , so that it becomes unavailable for further cooling and star formation until much later times , the number of star - forming galaxies in low - mass field halos is significantly lowered and this improves the fit to the observed luminosity and correlation functions .
even so , the model still produces an excess of very bright galaxies in groups and clusters . the @xmath5 @xmath7cdm model has the opposite problem
: it produces too few @xmath9 galaxies , even if feedback is weak and no gas leaves the halo .
the two point correlation function is also too steep on scales below a few mpc .
one obvious inference we can draw from these results is that a cdm model with density parameter somewhere between @xmath189 and @xmath190 may well do better in matching the observations .
indeed , in recent work , somerville & primack ( 1998 ) find that a model with @xmath191 provides the best fit .
care must be exercised , however , before drawing strong conclusions about cosmological parameters .
many of our difficulties in fitting the observed shape of the galaxy luminosity function come about because we have chosen a very bright normalization for the models .
this is because we _ force _ the models to fit the zero - point of the giovanelli et al i - band tully - fisher relation at a circular velocity of @xmath192 km s@xmath8 .
we have also assumed that the circular velocities of central disk galaxies are the same as those of their surrounding halos . in practice , the relation between the circular velocities of disk and halo is not straightforward and requires a detailed model for the internal structure of the halo and of the galaxy within it .
more detailed calculations of disk formation in cdm halos predict that the circular velocity of the disk is generally _ higher _ than that of the surrounding halo ( mo , mao & white 1998 ) .
this results in a higher total luminosity density for the same tully - fisher normalization , and would improve the fit of the @xmath7cdm model while making the @xmath4cdm model even more discrepant .
one might also question whether normalizing the models to match the tully - fisher relation is sensible , since the physical origin of this relation is still uncertain and depends on structural properties of the galaxies which we are not attempting to model .
we have argued that the distribution of k - band luminosities should provide a robust test of galaxy formation theory , since the k - band luminosity of a galaxy depends primarily on its stellar mass and not on its instantaneous star formation rate or its dust content . with the advent of wide - field k - band galaxy surveys such as the two micron
all sky survey ( 2mass , kleinmann et al 1994 ) , normalizing the models to match the observed k - band luminosity function may be the best choice for the future .
another effect we have investigated is that of dust extinction in star - forming galaxies .
if dust is ignored , the @xmath2 @xmath4cdm model produces too many blue galaxies compared to the observations .
galaxy formation occurs rather late in this model , so the stellar populations of galaxies are younger than in the @xmath7cdm model .
however , the inclusion of a simple , empirically - motivated dust extinction scheme makes this problem go away , and the colour distributions of galaxies in the @xmath4cdm and @xmath7cdm models are then virtually indistinguishable .
since dust extinction only affects star - forming field galaxies , models with dust produce somewhat steeper and higher amplitude correlation functions .
finally , dust lowers the predicted number of galaxies at the bright end of the b - band luminosity function .
our models tend to produce a power - law `` tail '' of high - luminosity galaxies , rather than a sharp exponential cutoff .
although dust can reduce this tail in the b - band , the problem persists in the k - band . finally ,
although we have tried to illustrate the effect of different schemes for star formation , feedback and dust , we can not be sure that we have taken into account all the key physical processes that may seriously affect our results .
for example , we have proposed that very efficient feedback can stop galaxies forming in low - mass halos , but alternative schemes may be possible .
for example , jimenez et al ( 1997 ) have proposed that the gaseous disks that form in low @xmath36 halos are not subject to the instabilities that cause star formation and are thus not seen in the optical or uv , although their emission at 21 cm wavelengths should be detectable .
the only way to make progress is through observational studies of how star formation rates , cold gas and dust contents , and the shells and bubbles seen around star - forming regions , vary with galactic mass , metallicity , morphology and environment .
this is a massive and long - term undertaking , but it seems necessary if we are to build a picture of galaxy formation that is more than just schematic . we conclude that we can not reliably constrain the values of cosmological parameters using the properties of the galaxy distribution at @xmath0 because our results are strongly influenced by our adopted normalization and the way in which we choose to parametrize star formation and supernova feedback .
the main effects of changing the cosmological parameters in our set of cluster - normalized cdm models are the adopted normalization and the recipes for feedback and star formation affect the efficiency with which stars form as a function of redshift and galaxy mass .
it is thus not surprising that there is considerable freedom to `` tune ''
our results to fit the observations .
nevertheless it is interesting that very different star formation and feedback schemes are required to come close to reproducing the observations in the two cosmologies we have explored . for @xmath2
, we require that feedback be very efficient at removing gas both from galaxies and from their surrounding halos .
many lower - mass halos are `` dark '' in this model , containing a faint , possibly low - surface brightness object . in a low - density universe
, feedback must be inefficient and every halo must contain a star - forming galaxy if we are to match the observations .
this gives us confidence that if the cosmological parameters of the universe can be determined using techniques such as the analysis of small - scale fluctuations in the microwave background ( e.g. bond , efstathiou & tegmark 1997 ) or the light - curves of high - redshift supernovae ( e.g. perlmutter et al 1997 ) , the methodology introduced in this paper will become valuable for understanding and constraining the detailed physical processes operating within galaxies .
it is also encouraging that our best - fit models come reasonably close to fitting many different observed measures of the galaxy distribution _ at the same time_. the models that give the best fits to the b - band luminosity function , also give the best fits to the k - band luminosity function , the correlation function , the color distribution and the star formation rate function . in paper
ii we demonstrate that these same models provide a reasonably good match to the observed properties of galaxy groups .
this gives us confidence that the fundamental theoretical framework , which determines how galaxy properties scale with halo mass , may indeed be correct .
+ the simulations in this paper were carried out at the computer center of the max - planck society in garching and at the eppc in edinburgh .
codes were kindly made available by the virgo consortium .
we especially thank adrian jenkins and frazer pearce for help in carrying them out .
we are also grateful to margaret geller , john huchra and ron marzke for enabling us to compare the cfa2n data directly with the simulations .
is a marie curie fellow and holds grant erbfmbict-960695 of the training and mobility of researchers program financed by the ec .
l*7c model & @xmath193 & @xmath194 & @xmath195 & @xmath21 & @xmath196 & box size + & & & & & & [ @xmath197 + scdm & 1.0 & 0.0 & 0.5 & 0.60 & 0.50 & 85 + @xmath4cdm & 1.0 & 0.0 & 0.5 & 0.60 & 0.21 & 85 + @xmath7cdm & 0.3 & 0.7 & 0.7 & 0.90 & 0.21 & 141 + ocdm & 0.3 & 0.0 & 0.7 & 0.85 & 0.21 & 141 + | we introduce a new technique for following the formation and evolution of galaxies in cosmological n - body simulations .
dissipationless simulations are used to track the formation and merging of dark matter halos as a function of redshift .
simple prescriptions , taken directly from semi - analytic models of galaxy formation , are adopted for gas cooling , star formation , supernova feedback and the merging of galaxies within the halos .
this scheme enables us to explore the clustering properties of galaxies and to investigate how selection by luminosity , colour or type influences the results . in this paper
, we study the properties of the galaxy distribution at @xmath0 . these include b and k - band luminosity functions , two - point correlation functions , pairwise peculiar velocities , cluster mass - to - light ratios , @xmath1 colours and star formation rates .
we focus on two variants of a cold dark matter ( cdm ) cosmology : a high - density ( @xmath2 ) model with shape - parameter @xmath3 ( @xmath4cdm ) , and a low - density model with @xmath5 and @xmath6 ( @xmath7cdm ) .
both models are normalized to reproduce the i - band tully - fisher relation of giovanelli et al .
( 1997 ) near a circular velocity of 220 km s@xmath8 .
our results depend _ strongly _ both on this normalization and on the adopted prescriptions for star formation and feedback .
very different assumptions are required to obtain an acceptable model in the two cases . for @xmath4cdm
, efficient feedback is required to suppress the growth of galaxies , particularly in low - mass field haloes . without it , there are too many galaxies and the correlation function exhibits a strong turnover on scales below 1 mpc . for @xmath7cdm ,
feedback must be weaker , otherwise , too few @xmath9 galaxies are produced and the correlation function is too steep . although neither model is perfect both come close to reproducing most of the data . given the uncertainties in modelling some of the critical physical processes ,
we conclude that it is not yet possible to draw firm conclusions about the values of cosmological parameters from studies of this kind .
further observational work on global star formation and feedback effects is required to narrow the range of possibilities .
-0.55 cm -0.55 cm -2.0truecm 16.75 cm clustering of galaxies in a hierarchical universe : + i. methods and results at @xmath0 + guinevere kauffmann , jrg m. colberg , antonaldo diaferio & simon d.m .
white + + keywords : galaxies : formation ; galaxies : halos ; cosmology : large - scale structure ; cosmology : dark matter |
consider the task of solving the linear system : @xmath0 where @xmath1sign@xmath2 is a unitary matrix , the identity matrix , @xmath3 the hermitian wilson operator , @xmath4 and @xmath5 the bare fermion mass .
the overlap operator @xmath6 is non - hermitian . for such operators gmres ( generalised minimal residual ) and fom ( full orthogonalisation method )
are known to be the fastest .
it is shown that when the norm - minimising process of gmres is converging rapidly , the residual norms in the corresponding galerkin process of fom exhibit similar behaviour @xcite . but they are based on long recurrences and thus require to store a large number of vectors of the size of matrix columns . however , exploiting the fact that the overlap operator is a shifted unitary matrix one can construct a gmres type algorithm with short recurrences @xcite . similarly , a short recurrences algorithm can be obtained from fom .
the method is based on an observation of rutishauser @xcite that for upper hessenberg unitary matrices one can write @xmath7 , where @xmath8 and @xmath9 are lower and upper bidiagonal matrices . applying this decomposition for the arnoldi iteration : @xmath10 one obtains an algorithm which constructs arnoldi vectors @xmath11 by short recurrences @xcite : @xmath12 projecting the linear system ( [ lin_sys ] ) onto the krylov subspace one gets : @xmath13 which can be equivalently written as : @xmath14 note that the matrix on the left hand side is tridiagonal .
it can be shown that one can solve this system and therefore the original system using short recurrences @xcite .
the resulting algorithm is called the shifted unitary orthogonal method ( suom ) and is given below : @xmath15 @xmath16 @xmath17 @xmath18 @xmath19 @xmath20 @xmath21 @xmath22 @xmath23 @xmath24 @xmath25 @xmath26 @xmath27 @xmath28 @xmath29 @xmath30 stop if @xmath31 tol @xmath32 note that in an actual implementation one can store @xmath33 and @xmath34 as separate vectors , which can be used in the subsequent iteration to compute @xmath35 .
therefore only one multiplication by @xmath36 is needed at each step .
a straightforward application of multigrid algorithms is hopeless in the presence of non - smooth gauge fields .
however , the situation is different for the 5-dimensional formulation of chiral fermions where there are no gauge connections along the fifth dimension .
here , i will limit my discussion in the easiest case which consists of two grids : the `` fine '' grid , which is the continuum along the fifth coordinate and a coarse grid , which is the lattice discretisation of the `` fine '' grid .
i define chiral fermions on the coarse grid using truncated overlap fermions @xcite .
the corresponding 5-dimensional matrix @xmath37 in blocked form is given by : @xmath38 where @xmath39 .
let @xmath40 be the above matrix but with bare quark mass @xmath41 and @xmath42 the permutation matrix : @xmath43 it can be shown that the following result hold @xcite : [ proposition ] let @xmath44 be the linear system defined on the 5-dimensional lattice with @xmath45 and @xmath46 .
then @xmath47 is the solution of the linear system @xmath48 , where @xmath49 as @xmath50 .
this result lends itself to a special two - grid algorithm @xcite .
indeed , @xmath51 is the ( fifth euclidean ) coordinate of interest since it contains the information about the 4-dimensional physics .
@xmath52 let @xmath53 let @xmath54 solve @xmath55 until @xmath56 @xmath57 @xmath58 stop if @xmath59 tol @xmath60 one way of exploiting this is to use `` decimation '' over the fifth coordinate in order to get the 5d - vector @xmath61 . using proposition [ proposition ] one can evaluate directly the first 4d - component of @xmath61 by @xmath62 , @xmath63 being an approximate solution .
the rest can be padded with zero 4d - vectors .
the second step is to solve the problem on the coarse grid .
finally , one can extract the 4d - solution @xmath47 on this grid and correct the `` fine '' grid solution by @xmath64 . in the second cycle one
has to repeat the same decimation method , since the `` fine '' 5d - operator is not available .
hence , the whole scheme is a restarted two - grid algorithm , which is given here as algorithm [ two_grid_algor ] .
in fig . 1 is shown the convergence of variuos algorithms as a function of wilson matrix - vector multiplication number on a fixed gauge background on a @xmath65 lattice at @xmath66 .
the convergence is measured using the norm of the residual error . for the overlap matrix - vector multiplication
is used the double pass lanczos algorithm ( without small eigenspace projection of @xmath3 ) as described in @xcite . together with the algorithms described in the previous sections fig .
1 shows the performance of conjugate residuals ( cr ) , conjugate gradients on normal equations ( cgne ) and cg - chi .
the latter is the cgne which solves simultaneously the decoupled chiral systems appearing in the matrix @xmath67 .
one can observe a gain over cgne which may be explained due to the reduced number of eigenvalues at each chiral sector .
however , this gain is no more than @xmath68 . on the other hand suom and cr preform
rather similar with suom being slightly faster in this scale .
the gain over cgne is about a factor two .
the two - grid algorithm performs the best with a gain of at least a factor 6 over suom and more than an order of magnitude over cgne .
this situation repeats itself for a different gauge configuration which is not shown here for the lack of space .
however , if the projection of small eigenvalues is used the gain over suom / cr should be smaller since the two - grid algorithm is much less intensive in the application of the overlap operator .
it is exactly the purpose of this comparison to make clear this feature of the two - grid algorithm .
finally , it is ( not ) surprising that suom and cr perform similarly : cr can be shown to be an efficient method for normal matrices .
since it is easier to implement cr is more appealing than other krylov solvers . | in this paper i describe a new optimal krylov subspace solver for shifted unitary matrices called the shifted unitary orthogonal method ( suom ) . this algorithm is used as a benchmark against any improvement like the two - grid algorithm .
i use the latter to show that the overlap operator can be inverted by successive inversions of the truncated overlap operator .
this strategy results in large gains compared to suom .
it is well - known that overlap fermions @xcite lead to much more expensive computations than standard fermions , i.e. wilson or kogut - sussking fermions .
this is obvious since for every application of the overlap operator an extra linear system solving is needed . for the time being , it seems that to get chiral symmetry at finite lattice spacing one should wait for a petaflops computer being built .
however , algorithmic research is far from exhausted . in this paper
i give an example that this is the case if one uses the two - grid algorithm @xcite . before i do this ,
i introduce briefly an optimal krylov subspace solver for shifted unitary matrices . |
recent experimental success @xcite in bose - einstein condensation ( bec ) at ultralow temperature in dilute trapped alkali metal and hydrogen atoms has intensified theoretical studies of the inhomogeneous , weakly interacting dilute bose gas and its condensation @xcite . at zero temperature
such a system is supposed to be fully condensed and described by the mean - field gross - pitaevskii ( gp ) equation @xcite .
two most interesting phenomena in bec are the collapse for attractive interaction and the formation of coupled atomic condensate . for attractive atomic interaction @xcite , the condensate is stable for a maximum critical number of atoms . in the presence of an external source of atoms when the number increases beyond the critical number , due to interatomic attraction the radius of bec tends to zero and the central density tends to infinity .
consequently , the condensate collapses emitting atoms until the number of atoms is reduced below the critical number and a stable configuration is reached .
then the condensate can grow again and a series of collapses can take place , which was observed experimentally in the bec of @xmath0li with attractive interaction @xcite
. there has also been experimental realization of bec involving atoms in two different quantum states @xcite . in one experiment @xmath1rb atoms formed in the @xmath2 , @xmath3 and @xmath4 ,
@xmath5 states by the use of a laser served as two different species , where @xmath6 and @xmath7 are the total angular momentum and its projection @xcite . in another experiment
a coupled bec was formed with the @xmath1rb atoms in the @xmath8 and @xmath9 states @xcite .
it is possible to use the same magnetic trap to confine atoms in two quantum states and this makes these experiments technically simpler compared to a realization of bec with two different types of atoms requiring two independent trapping mechanisms .
here we study theoretically the dynamics of a coupled bec composed of two quantum states 1 and 2 of an atom using the coupled time - dependent gp equation .
we include an interaction term in the hamiltonian that allows for a coherent boson - number - conserving nondissipative transition of atoms from one quantum state to another .
this interaction simulates the action of a laser which can been used experimentally to transform one species of atoms to another .
this model exhibits rich and novel phenomenology which we study using the numerical solution of the coupled gp equation , which can generate oscillation in the number of atoms in a specific state in the presence of the interaction that transforms one spices of atoms to another . also , depending on the nature of the atomic interactions one can have three possibilities of bec : ( i ) stable condensate of both components for all repulsive interactions , ( ii ) collapse of both components for all attractive interactions , and ( iii ) a stable condensate for one component and collapse in the other .
the collapse for attractive interaction(s ) in a coupled condensate is studied by introducing absorptive contact interactions responsible for a growth in the atomic numbers from external source in addition to imaginary three - body quartic interactions leading to recombination loss @xcite . in the presence of these imaginary interactions
the gp equation does not conserve the overall number of atoms .
if the strengths ( and signs ) of these imaginary interactions are properly chosen , one can have collapse in one or both components of the condensate .
the solution of the gp equation could produce a growth of the condensate with time for atom numbers less than the critical numbers for collapse .
once they increase past the critical numbers , the three - body interaction takes control and the numbers suddenly drop below the critical level by recombination loss signaling a collapse @xcite .
then the absorptive term takes over and the individual numbers start to increase .
this continues indefinitely showing an infinite sequence of collapses for one or both components .
we motivate this study by considering two possible atomic states 1 and 2 of @xmath0li ( attractive interaction ) and of @xmath1rb ( repulsive interaction ) whenever possible . in the case of @xmath0li
the interaction in state 1 is taken to be attractive which is responsible for collapse .
we study the collapse with different possibilities of attraction and repulsion between atoms in states 1 and 2 of @xmath0li .
we consider the following hamiltonian to describe the coupled bec allowing for the possibility of transforming atoms from one state to another @xcite @xmath10 where the kinetic and potential energies are @xmath11\hat \psi_j(\vec r),\ ] ] @xmath12 @xmath13 , \ ] ] where @xmath14 and @xmath15 are the field operators for annihilation and creation of the bosonic atom of mass @xmath7 in state @xmath16 , @xmath17 is external trapping potential , @xmath18 is the usual nonlinear potential between the atoms interacting via a contact interaction of strength @xmath19 between two atoms in states @xmath20 and @xmath21 , @xmath22 is the scattering length of atoms in states @xmath20 and @xmath21 , and @xmath23 is the potential which allows for the laser - induced transition of an atom from state @xmath24 to @xmath25 and vice versa via a contact interaction of strength @xmath26 .
coupled mean - field gp equations are obtained by replacing the field operators @xmath14 by the wave functions @xmath27 , which are the expectation values of the field operators , in the heisenberg equation @xcite : @xmath28,\ ] ] where the time ( @xmath29 ) dependence is explicitly shown and the result is @xcite @xmath30 \psi_j(r,\tau)+\sum_{l=1}^2\chi ( 1-\delta_{jl } ) \psi_l(r,\tau)=0,\end{aligned}\ ] ] where @xmath31 is the kronecker delta and @xmath32 is the number of condensed atoms in state @xmath20 . here
we have specialized to the radially symmetric case and have used the harmonic oscillator trapping potential @xmath33 where @xmath34 is the trap frequency and @xmath35 is a parameter introduced to vary independently the trap potential in each quantum state .
the normalization condition for @xmath36 is given by @xmath37 as in refs .
@xcite it is convenient to use dimensionless variables defined by @xmath38 , and @xmath39 where @xmath40 , and @xmath41 . in terms of these variables
( [ cc ] ) becomes @xcite @xmath42\phi_j ( { x},t)+\sum_{l=1}^2\eta ( 1-\delta_{jl } ) \phi_l ( { x},t)=0 , \end{aligned}\ ] ] where @xmath43 and @xmath44 could be negative ( positive ) when the corresponding interaction is attractive ( repulsive ) . in eqs .
( [ e ] ) we have introduced a diagonal absorptive i@xmath45 and a quartic three - body term @xmath46 appropriate to study collapse @xcite .
a nonzero @xmath45 allows the possibility of the absorption of atoms from an external source into the condensate , whereas a nonzero @xmath47 allows for the possibility of ejection of atoms from the condensate due to three - body recombination . the total number of atoms in the condensate is not conserved for nonzero @xmath45 and @xmath47 .
the fluctuation of the number of atoms in the two states of the condensate is best studied via the quantities @xmath48 in the presence of a general absorptive interaction @xmath49 and @xmath50 the reduced number for the two components of the condensate are given by @xmath51 and @xmath52 , the quantities @xmath53 and @xmath54 carry the information about time evolution of the number of the two components .
the actual number of atoms in the two components are @xmath55 , @xmath16 . for @xmath56 ,
the number of atoms in the two components are conserved separately .
this condition leads to @xmath57 . for @xmath58 and @xmath59 ,
the numbers of atoms in the two components are not conserved separately @xmath60 ; but the total number is conserved . in the presence of absorptive interaction @xmath61 and/or @xmath62 , there is no conservation of even the total number of atoms .
to solve eqs .
( [ e ] ) numerically the proper boundary conditions as @xmath63 and @xmath64 are needed . for a confined condensate the asymptotic form of the physically acceptable solution
is given by @xmath65 . at @xmath66
one has the regularity condition @xmath67 .
next we discretize eqs .
( [ e ] ) in both space and time by using a space step @xmath68 and time step @xmath69 with a finite difference scheme using the unknown @xmath70 which are approximation to the exact solution @xmath71 where @xmath72 and @xmath73 .
after discretization we obtain a set of algebraic equations which could then be solved by using the known asymptotic boundary conditions @xcite .
the time derivative in these equations involves the wave functions at times @xmath74 and @xmath75 .
to discretize eqs .
( [ e ] ) we express the wave functions and their derivatives in terms of the average over times @xmath74 and @xmath75 @xcite and the result is the following crank - nicholson - type scheme @xcite @xmath76\nonumber \\ & + & \frac{1}{2}\left[\frac{c_jx_p^2}{4}+\sum_{l=1}^2 n_{jl}\frac{|(\phi_l)_p^{k}|^2}{x_p^2}-i\xi_j \frac{|(\phi_j)_p^{k}|^4}{x_p^4}+i\gamma_j \right]\nonumber \\ & \times & \left[(\phi_j)_p^{k+1}+(\phi_j)_p^k\right]+\sum_{l=1}^2\eta(1-\delta_{lj } ) { ( \phi_l)_p^k } , \label{fx } \end{aligned}\ ] ] where @xmath16 .
considering that the wave function components @xmath77 are known at time @xmath74 , eqs .
( [ fx ] ) involve the unknowns @xmath78 @xmath79 and @xmath80 at time @xmath75 . in a lattice of @xmath81 points eqs .
( [ fx ] ) represent a tridiagonal set for @xmath82 for a specific component @xmath77 .
this set has a unique solution if the wave functions at the two end points @xmath83 and @xmath84 are known @xcite .
these values at the end points are provided by the known asymptotic boundary conditions .
to solve eqs .
( [ fx ] ) we employ space step @xmath85 0.0001 with @xmath86 and time step @xmath87 = 0.05 . after some experimentation we find that these values of the steps give good convergence .
the iteration is started with the known normalized ( harmonic oscillator ) solution of eqs .
( [ e ] ) obtained with @xmath88 .
the nonlinear parameters @xmath89 are increased by equal amounts over 1000 time iterations starting from zero at @xmath90 until the desired final value is reached at @xmath91 . during the iteration we keep @xmath92 and attribute a small numerical value to @xmath93 ( @xmath94 ) .
the resulting solution is the ground state of the condensate corresponding to the specific nonlinearity and built up with this particular value of @xmath93 .
then to study the dynamics we allow the system to evolve by continuing the iteration with @xmath95 and @xmath96 for positive @xmath97 , but maintaining @xmath93 and @xmath98 at the constant final values attained at @xmath91 . to study the dynamics for positive @xmath97 for attractive atomic interaction , the absorptive contact ( @xmath45 ) and the three - body interactions ( @xmath47 )
are switched on at @xmath91 .
this will clearly show the possible collapse of the system .
first we consider the stationary solution of eqs .
( [ e ] ) with @xmath58 in the repulsive case .
we now choose the following two sets of of parameters before the actual calculation : @xmath99 @xmath100 @xmath101 , @xmath102 and ( b ) @xmath103 .
this will show the effect of the feedback between the two atomic states . in this case
all interactions are repulsive corresponding to positive sign of all @xmath104 .
although the parameters above are in dimensionless units , it is easy to associate them with a physical problem of experimental interest in the case @xmath105 .
for example , for the mixture of @xmath106 and @xmath107 states of @xmath1rb , the ratio of scattering lengths @xmath108 @xcite .
if we label state @xmath109 by 1 and @xmath110 by 2 , and consider @xmath111 , then @xmath112 corresponds to @xmath113 for @xmath105 .
hence these models can simulate the actual experimental situation composed of two states of @xmath1rb . in fig .
1 we plot the wave functions @xmath114 for the two components calculated with two nonzero values of @xmath93 above . in figs .
2 ( a ) and ( b ) we plot the quantities @xmath53 and @xmath54 for these models . when @xmath105 , @xmath115 . as a nonzero value of @xmath93
is taken , the quantities @xmath116 oscillate with time as can be seen from figs .
2 ( a ) and ( b ) .
when @xmath53 increases , @xmath54 decreases and this denotes a net transformation of atoms from state 2 to 1 and vice versa .
these transformations lead to interesting oscillation in @xmath116 and consequently , in the actual number of atoms in the two states . for a small @xmath93
these oscillations are small and as @xmath93 increases these oscillations increase in amplitude .
these oscillations clearly show the continued dynamical conversion of one species of atoms to another in the otherwise stable condensate .
now we study the simplest case of collapse by taking only the interaction between the atoms in state 1 to be attractive corresponding to a negative @xmath117 .
all other scattering lengths @xmath78 @xmath118 and @xmath119 (= @xmath120 ) @xmath78 are taken to be positive .
quite expectedly , here the first component of the condensate could experience collapse .
we also consider the collapse in both components when all scattering lengths are negative .
the calculation is performed in these cases with the following two sets of parameters ( a ) @xmath121 @xmath122 @xmath123 , @xmath124 , @xmath125 and ( b ) @xmath126 @xmath127 @xmath128 , @xmath129 , @xmath130 .
the wave function components at @xmath91 in these cases are plotted in fig .
3 . the above parameters are in dimensionless units and one can associate them with a problem of experimental interest for @xmath105 in terms of two states of @xmath0li . in case of model ( a )
above we consider the state 1 to be the ground state of @xmath0li with attractive interaction as in the actual collapse experiment with @xmath131 @xcite . as @xmath132
this number is larger than the maximum number of atoms permitted in the bec of a single component @xmath0li which is about 1400 @xcite . the state 2 could be one of the excited states of @xmath0li with repulsive atomic interaction .
the presence of the second component with repulsive interaction allows for the formation of a stable bec with more @xmath0li atoms in quantum state 1 than allowed in the single - component bec .
we find from fig .
3 that @xmath133 ( full line ) is very much centrally peaked compared to @xmath134 ( dashed line ) in this case . this corresponds to a small radius and large central density for the first component denoting an approximation to collapse . if the number @xmath135 is slightly increased beyond 2200 the first component of the condensate wave function becomes singular at the origin and could experience collapse . in case of model ( b )
we take the atomic interaction in both states of @xmath0li to be attractive corresponding to negative scattering lengths . for @xmath136 as in the actual experiment @xcite
, one has @xmath137 and @xmath138 for @xmath105 .
the total number of particles in this case is roughly 1460 , which is close to the critical number 1400 found in the actual experiment of collapse in @xmath0li .
both wave - function components could become singular in this case as all possible interactions are attractive .
the small extension of the wave - function components in space in this case as in fig .
3 denote an approximation to collapse . although the collapse of the coupled condensates could be inferred from the shape of the stationary wave functions of fig .
3 ( sharply peaked centrally with small radii ) , we also study the dynamics of collapse from a time evolution of the full gp equation ( [ e ] ) in the presence of an absorption and three - body recombination , e.g. , for @xmath139 and @xmath140 as in the uncoupled case @xcite . the general nature of time evolution
is independent of the actual values of @xmath45 and @xmath47 employed provided that a small value for @xmath47 and a relatively larger one for @xmath141 are chosen @xcite .
first we consider model ( a ) . for a dynamic description of this problem
we take for @xmath142 @xmath143 , @xmath144 and @xmath145 and allow the solution of the gp equation to evolve in time using eq .
( [ fx ] ) . in the presence of these absorptive interactions
the total number of atoms is not a constant of motion .
the time evolution of @xmath53 and @xmath54 is shown in fig .
4 ( a ) . there is small oscillation in @xmath116 for @xmath146 signifying a transfer of one type of atoms to the other . for @xmath142 these oscillations
are almost undetectable . because of the attractive interaction , the component 1 could experience a succession of collapse .
this is manifested in the successive growth and decay of @xmath53 .
the condensate corresponding to the second component controlled by the repulsive interaction keeps on growing , manifested by the continuous growth of @xmath54 with time .
next we consider the time evolution of the gp equation for model ( b ) with the following absorptive parameters : @xmath147 . in this case as all interactions are attractive , both components can experience collapse . in fig 4 ( b )
we plot the evolution of @xmath53 and @xmath54 .
the oscillation of @xmath53 and @xmath54 is clearly visible for @xmath146 . for @xmath142 ,
the absorptive interactions take control and the total number of atoms start to increase .
consequently , the dynamics of collapse is more visible than the oscillation in @xmath53 and @xmath54 .
both @xmath53 and @xmath54 is found to exhibit a succession of growth and decay corresponding to collapse .
hence we find that one can have a dynamical collapse of one or both components of the bec in the case of a coupled condensate .
to conclude , we studied the collapse in a trapped bec of atoms in states 1 and 2 using the gp equation when some of the atomic interactions are attractive .
we include an interaction in the hamiltonian which allow for a spontaneous transformation of atoms from state 1 to 2 and vice versa .
experimentally , this can be achieved by a laser and makes the present study richer .
in case of all repulsive atomic interactions one can have a stable condensate of both components .
when some of the atomic interactions are attractive one or both components of the condensate could experience collapse . the collapse could be predicted from a stationary solution of the gp equation .
the time evolution of collapse is studied via the time - dependent gp equation with absorption and three - body recombination .
the number of particles of the component(s ) of bec experiencing collapse alternately grows and decays with time . with the possibility of a detailed study of a coupled bec
, the results of this study could be verified experimentally in the future .
j. r. ensher , d. s. jin , m. r. matthews , c. e. wieman , e. a. cornell , phys .
77 ( 1996 ) 4984 .
k. b. davis , m. o. mewes , m. r. andrews , n. j. van druten , d. s. durfee , d. m. kurn , w. ketterle , phys .
lett . 75 ( 1995 ) 3969 .
d. g. fried , t. c. killian , l. willmann , d. landhuis , s. c. moss , d. kleppner , t. j. greytak , phys .
81 ( 1998 ) 3811 .
s. giorgini , l. p. pitaevskii , s. stringari , phys .
a 54 ( 1996 ) r4633 .
m. edwards , p. a. ruprecht , k. burnett , r. j. dodd , c. w. clark , phys
77 ( 1996 ) 1671 .
s. k. adhikari , a. gammal , physica a 286 ( 2000 ) 299 .
m. r. matthews , b. p. anderson , p. c. haljan , d. s. hall , c. e. wieman , e. a. cornell , phys .
83 ( 1999 ) 2498 .
m. r. matthews , d. s. hall , d. s. jin , j. r. ensher , c. e. wieman , e. a. cornell , f. dalfovo , c. minniti , s. stringari , phys . rev . lett .
81 ( 1998 ) 243 .
a. sinatra , p. o. fedichev , y. castin , j. dalibard , g. v. shlyapnikov , phys .
82 ( 1999 ) 251
. b. d. esry , c. h. greene , j. p. burke , jr .
, j. l. bohn , phys .
78 ( 1997 ) 3594 .
. ho , v. b. shenoy , phys . rev .
77 ( 1996 ) 3276 .
wave function components @xmath148 ( full line ) and @xmath149 ( dashed line ) vs. @xmath150 for two coupled gp equations with ( a ) @xmath121 @xmath122 @xmath123 , @xmath124 , @xmath125 and ( b ) @xmath126 @xmath127 @xmath128 , @xmath129 , @xmath130 . | the dynamics of a coupled bose - einstein condensate involving trapped atoms in two quantum states is studied using the time - dependent gross - pitaevskii equation including an interaction which can transform atoms from one state to the other .
we find interesting oscillation of the number of atoms in each of the states .
for all repulsive interactions , stable condensates are formed .
when some of the atomic interactions are attractive , the possibility of collapse is studied by including an absorptive contact interaction and a quartic three - body recombination term .
one or both components of the condensate may undergo collapse when one or more of the nonlinear terms are attractive in nature . *
pacs number(s ) : 03.75.fi * |
00 r. f. sawyer , hep - ph/0408265 .
r. f. sawyer , phys .
d * 72 * , 045003 ( 2005 ) [ hep - ph/0503013 ] .
h. duan , g. m. fuller and y. z. qian , phys .
d * 74 * , 123004 ( 2006 ) [ astro - ph/0511275 ] .
h. duan , g. m. fuller , j. carlson and y. z. qian , phys .
d * 74 * , 105014 ( 2006 ) [ astro - ph/0606616 ] .
s. hannestad , g. g. raffelt , g. sigl and y. y. y. wong , phys .
d * 74 * , 105010 ( 2006 ) [ astro - ph/0608695 ] .
h. duan , g. m. fuller , j. carlson and y. z. qian , phys .
d * 75 * , 125005 ( 2007 ) [ astro - ph/0703776 ] .
g. g. raffelt and g. sigl , phys .
d * 75 * , 083002 ( 2007 ) [ hep - ph/0701182 ] .
a. esteban - pretel , s. pastor , r. toms , g. g. raffelt and g. sigl , phys .
d , in press ( 2007 ) [ astro - ph/0706.2498 ] . g. g. raffelt and a. yu .
smirnov , phys .
d * 76 * , 081301 ( 2007 ) [ arxiv:0705.1830 ] .
g. g. raffelt and a. yu . smirnov , phys .
d * 76 * , 125008 ( 2007 ) [ arxiv:0709.4641 ] .
h. duan , g. m. fuller and y. z. qian , phys . rev .
d * 76 * , 085013 ( 2007 ) [ arxiv:0706.4293 ] .
h. duan , g. m. fuller , j. carlson and y. z. qian , phys .
, in press [ arxiv:0707.0290 ] . g. l. fogli , e. lisi , a. marrone and a. mirizzi , jcap , in press [ hep - ph/0707.1998 ] . h. duan , g. m. fuller , j. carlson and y. z. qian , arxiv:0710.1271 .
a. esteban - pretel , s. pastor , r. toms , g. g. raffelt and g. sigl , arxiv:0712.1137 .
e. lisi , `` collective neutrino flavor transitions in supernovae : analytical and numerical aspects '' , talk at taup 2007 , 1115 sept .
2007 , sendai , japan .
g. sigl and g. raffelt , nucl .
b * 406 * , 423 ( 1993 ) .
f. j. botella , c. s. lim and w. j. marciano , phys .
d * 35 * , 896 ( 1987 ) .
e. k. akhmedov , c. lunardini and a. yu .
smirnov , nucl .
b * 643 * , 339 ( 2002 ) [ hep - ph/0204091 ] . | we study two - flavor collective neutrino oscillations in the dense - neutrino region above the neutrino sphere in a supernova ( sn ) .
the angular dependence of the neutrino - neutrino interaction potential causes `` multi - angle '' effects that can lead either to complete kinematical decoherence in flavor space or only to small differences between different trajectories .
this nonlinear system switches abruptly between `` self - maintained coherence '' and `` self - induced decoherence '' among the angular modes , depending on the strength of the deleptonization flux .
for a realistic sn the quasi single - angle behavior is probably typical , simplifying the numerical treatment and probably allowing for the survival of observational features of flavor oscillations .
the neutrinos streaming off a collapsed supernova ( sn ) core are so dense near the neutrino sphere that they produce a significant refractive effect for each other , leading to collective oscillation effects .
the practical importance of these nonlinear phenomena was only recently recognized and studied in a series of papers @xcite .
an introduction was given at this conference by eligio lisi @xcite .
the sn neutrino fluxes are thought to obey the hierarchy @xmath0 where @xmath1 stands for any of @xmath2 , @xmath3 , @xmath4 , and @xmath5 . in other words
, there is an excess of @xmath6 pairs relative to @xmath7 . in an inverted hierarchy situation ,
this pair excess converts collectively into @xmath7 pairs , a process that does not violate flavor - lepton number and thus does not require mixing : it could also proceed as an ordinary pair annihilation process .
neutrino refraction causes this pair process to proceed collectively and very fast , almost independently of the mixing angle , i.e. , we have to do with a collective `` speed - up effect '' @xcite . the unpaired @xmath8 excess flux from deleptonization is conserved @xcite . in the adiabatic limit , it is the low - energy part of the @xmath8 spectrum that survives , leading to a step - like feature in the @xmath8 spectrum ( a `` spectral split '' @xcite caused by a `` step - wise spectral swapping '' @xcite ) .
all of these effects happen in the dense - neutrino region that typically extends from the neutrino sphere out to a few hundred kilometers .
if the ordinary msw resonances occur at larger radii , the collective phenomena and subsequent msw transformations are independent , the former producing the initial condition for the latter . for an inverted hierarchy
the only effect of matter in the collective - transformation region is a decrease of the effective mixing angle @xcite .
conversely , if the matter density profile is very shallow , the ordinary msw effect can prepare the initial condition for the collective effects @xcite .
mixed neutrinos are described by matrices of density @xmath9 and @xmath10 for each ( anti)neutrino mode .
the diagonal entries are the usual occupation numbers whereas the off - diagonal terms encode phase information .
the equations of motion are @xmath11 $ ] , where the hamiltonian is @xcite @xmath12 @xmath13 being the velocity . in the mass basis ,
the matrix of vacuum oscillation frequencies is @xmath14 .
the matter effect is represented , in the weak interaction basis , by @xmath15 .
for antineutrinos the only difference is @xmath16 .
the effective tau - lepton density @xmath17 arises from radiative corrections @xcite and can be important in a genuine three - flavor treatment of ordinary @xcite or collective @xcite sn neutrino oscillations . for the latter , the influence of @xmath18 can be rather sensitive to deviations from maximal 23-mixing . the angular factor @xmath19 in eq .
( [ eq : hamiltonian ] ) derives from the current - current nature of the weak - interaction hamiltonian . in an isotropic ensemble it averages to unity , whereas the neutrinos streaming off a sn core are strongly non - isotropic so that different angular modes experience a different strength of the neutrino - neutrino interaction potential
. one may expect that this effect leads to kinematical decoherence among angular modes and thus to flavor equilibrium @xcite . in a symmetric ensemble with equal densities of neutrinos and antineutrinos , this is indeed the case . such a system is highly unstable in that an infinitesimal deviation from exact isotropy is enough to trigger an exponential run - away towards flavor equilibrium @xcite . on the other hand
, a numerical simulation of the flavor evolution of sn neutrinos revealed that multi - angle effects were small : all angular modes evolved nearly collectively , very similar to an isotropic ensemble @xcite .
likewise , in a multi - energy system every energy mode feels a different hamiltonian , yet the evolution is collective , i.e. , multi - energy effects do not lead to kinematical decoherence .
we have performed a numerical exploration of multi - angle effects in a spherically symmetric system where the neutrinos are emitted from a `` neutrino sphere '' @xcite .
we have identified the asymmetry between the @xmath8 and @xmath20 flux as the crucial parameter .
realistic deleptonization fluxes in sne seem sufficient to suppress multi - angle decoherence .
therefore , in practice multi - angle effects seem to be a subdominant feature of collective sn transformations . a better understanding of collective oscillations can be developed in the two - flavor case in terms of the usual flavor polarization vectors @xmath21 to express the matrices @xmath9 .
the equations of motion are @xmath22 . ignoring the ordinary matter effect , the `` hamiltonian '' is @xmath23 where @xmath24 is a unit vector in flavor space in the `` mass direction '' and @xmath25 the vacuum oscillation frequency . in an isotropic situation where the velocity terms average to zero ,
the second term is of the form @xmath26 where the vector @xmath27 is the difference between the total neutrino and antineutrino polarization vectors and @xmath28 the neutrino - neutrino interaction strength .
the individual @xmath29 all lie in a single plane . in the adiabatic limit
the polarization vectors follow the hamiltonians so that they also lie in this co - rotating plane .
this observation explains the collective nature of the multi - energy evolution ( all polarization vectors stay in the same plane that rotates around @xmath30 with a certain frequency @xmath31 ) and explains the spectral splits in that the final @xmath32 in the co - rotating plane are either aligned or anti - aligned with @xmath30 @xcite .
this picture also illustrates a crucial difference to the symmetric system where initially @xmath33 and the overall evolution is purely pendular @xcite there is no co - rotating plane . in a spherically symmetric situation
every mode is characterized by its vacuum oscillation frequency @xmath34 and its angle @xmath35 relative to the radial direction , providing @xmath36 as the radial velocity .
now the individual hamiltonians are @xmath37 where @xmath38 is the flux term of the neutrino ensemble , i.e. , the same as @xmath27 , but every mode weighted with @xmath39 . if all polarization vectors lie in a single plane , then also all @xmath32 are in that plane so that an evolution in a co - rotating plane is self - consistently possible .
our numerical simulations show that those cases with little kinematical decoherence correspond to the polarization vectors essentially staying in a co - rotating plane with some zenith - angle spread . on the other hand
, the vectors @xmath24 , @xmath27 and @xmath38 do not have to stay in a single plane and in fact , this appears to be an unstable arrangement .
kinematical multi - angle decoherence corresponds to strong deviations from this coplanar situation . in a spherically symmetric situation ,
the neutrino - neutrino interaction strength decreases with @xmath40 . in a toy model with an artificially slow decrease of @xmath41 one
can make the evolution arbitrarily slow and adiabatic . in this case
multi - angle decoherence appears to be unavoidable .
it appears that in a realistic sn the collective evolution is slow enough to be essentially adiabatic with regard to the development of a spectral split , but fast enough that the co - planar arrangement of the polarization vectors survives .
this picture may provide the key to an analytic understanding of the conditions for multi - angle kinematical decoherence .
the assumption of spherical symmetry of the overall system severely restricts possible solutions .
the evolution is a one - dimensional problem along the radial direction .
it remains to be studied if this symmetry has an important impact on kinematical decoherence , i.e. , if a system with fewer symmetries would decohere more easily .
this work was partly supported by the deutsche forschungsgemeinschaft ( tr-27 `` neutrinos and beyond '' ) , by the cluster of excellence `` origin and structure of the universe '' ( garching and munich ) , by the european union ( contracts no.rii3-ct-2004-506222 and mrtn - ct-2004 - 503369 ) , and by the spanish grants fpa2005 - 01269 ( mec ) and acomp07 - 270 ( generalitat valenciana ) .
ae was supported by an fpu grant from the spanish government .
sp and rt were supported by mec contracts ( _ ramn y cajal _ and _ juan de la cierva _ , respectively ) . |
the 3d 3-state potts model plays a role in our understanding of the properties of the deconfining phase transition and the structure of qcd .
the high - temperature vacuum of qcd is characterized by ordered polyakov loops which behave similarly as spins in the low temperature phase of the 3d 3-state potts model .
the polyakov loop serves as an order parameter of su(3 ) pure gauge theory with its symmetry determined by the @xmath6 center of the gauge group , and the phase transition is weakly first order .
this maps naturally onto the order parameter ( magnetization ) of the 3d 3-state potts model @xcite . for this and other reasons
a number of numerical studies of the model were performed @xcite . in this paper
we use the multicanonical technique @xcite to calculate the properly normalized entropy all the way from @xmath7 to across the phase transition . besides this new result
a number of additional estimates ( see the abstract ) , are obtained and compared with the literature .
our paper is organized as follows : in section [ sec_notation ] we introduce basic notation and observables of interest , in section [ sec_sim_res ] we present our simulations and conclusions are given in the final section [ sec_sum ] .
we simulate the @xmath8 potts model with the energy function @xmath9 where the sum is over the nearest neighbors of a 3d cubic lattice of size @xmath0 .
the spins @xmath10 of the system take the values @xmath11 .
the factor of two and the term @xmath12 are introduced to match for @xmath13 with ising model conventions @xcite .
our simulations are carried out in a multicanonical ensemble @xcite , covering an inverse temperature range from @xmath14 ( infinite temperature ) to @xmath15 below the phase transition temperature . instead of relying on a recursion ( see , e.g. , @xcite ) ,
the multicanonical parameters were extracted by finite size ( fs ) extrapolations from smaller to larger system , which is efficient when the fs behavior is controllable .
frequent excursions into the disordered phase all the way to @xmath7 secure equilibration of the configurations around the transition and in the ordered phase . by the fluctuation - dissipation theorem
the specific heat is @xmath16 we put @xmath17 in parenthesis , because the number 2 is later used as a superscript of @xmath17 .
the energy density per site is defined by @xmath18 so that the latent heat is @xmath19 where @xmath20 and @xmath21 are the energy endpoints of the high ( @xmath22 ) and low ( @xmath23 ) temperature phase at the transition temperature @xmath24 .
the entropy density is @xmath25 where @xmath26 is the free energy density .
as the free energy is continuous at the phase transition the entropy gap across the phase transition is @xmath27 to determine the entropy and energy gaps from data on finite lattices we follow ref .
@xcite and study the scaling of the specific heat maxima @xmath28 .
the leading order coefficient @xmath29 of the fit @xmath30 is related to the latent heat by @xmath31 and using ( [ sgap ] ) gives @xmath32 the entropy densities in the disordered ( @xmath22 ) and ordered ( @xmath23 ) phase at the transition temperature are defined by @xmath33 calculation of the endpoints from monte carlo ( mc ) data faces some technical difficulties , which are overcome in section [ subsec_entr ] . in a mc simulation of the gibbs canonical ensemble the entropy of the system
is only determined up to an additive constant , whereas in our multicanonical simulation this constant is determined by the known normalization at @xmath7 : @xmath34
in table [ tab_3dp3q ] we list the lattice sizes used , number of sweeps ( sequential updates of the lattice for which each spin is touched once ) performed with the multicanonical metropolis algorithm and the number of cycles @xmath35 the markov process performed during the production run , where @xmath36 is the effective energy - dependent @xmath17 of the multicanonical procedure .
data are energy histograms , which are collected over the statistics given by the second number in the production statistics column ( sweeps ) of table [ tab_3dp3q ] .
this is repeated for 32 bins , the first number in this column .
the data analysis relies on converting the bins into jackknife bins , which are then reweighted to canonical ensembles using the logarithmic coding of @xcite .
error bars are obtained by performing analysis calculations for each jackknife bin .
.statistics of multicanonical simulations of the 3-state potts model on @xmath0 lattices . [ cols="^,^,^",options="header " , ] from ( [ epmsym ] ) and ( [ spmsym ] ) we find finite lattice estimators @xmath37 and @xmath38 , which are collected in table [ tab_3dp3q_pm ] . fits with @xmath39 corrections give : @xmath40 the values of @xmath41 and @xmath42 are shown in figs .
[ fig_entropy0 ] and [ fig_entropygap ] as horizontal lines .
the vertical line in fig .
[ fig_entropygap ] represents the inverse transition temperature ( [ beta2c ] ) .
the main result of this work is the entropy density with proper normalization , shown in fig .
[ fig_entropy0 ] , where @xmath43 . the transition region
is enlarged in fig .
[ fig_entropygap ] .
the values of the entropy density @xmath41 ( [ sp ] ) on the disordered and @xmath42 ( [ sm ] ) on the ordered side of the transition in % of @xmath44 are , respectively , 82% and 73% .
thus , there are 3 states per site at infinite temperature and effectively 2.45 states on the high- and 2.24 states on the low - temperature side of the phase transition .
our corresponding @xmath45 estimates are given in eqs .
( [ ep ] ) and ( [ em ] ) .
other results are @xmath46 ( [ betac ] ) , the free energy density at the critical point ( [ f0fit ] ) , the entropy gap ( [ sgap ] ) , the latent heat ( [ de_cmax ] ) and the interface tension ( [ sigma_od ] ) .
it is notable that the inclusion of capillary waves enhances the estimate of the interface tension by more than 10% . | using a multicanonical metropolis algorithm we have performed monte carlo simulations of the 3d 3-state potts model on @xmath0 lattices with @xmath1 , 30 , 40 , 50 . covering a range of inverse temperatures from @xmath2 to @xmath3 we calculated the infinite volume limit of the entropy density @xmath4 with its normalization obtained from @xmath5 . at the transition temperature the entropy and energy endpoints in the ordered and disordered phase
are estimated employing a novel reweighting procedure .
we also evaluate the transition temperature and the order - disorder interface tension .
the latter estimate increases when capillary waves are taken into account . |
figure 1 . : : surface incompressibility coefficient versus the nuclear matter incompressibility modulus for the parameter sets of table 1 . figure 2 .
: : @xmath107 versus @xmath92 computed for several uncharged and symmetric nuclei from @xmath108 to @xmath109 for the nl1 , nl3 and nl - sh parameter sets .
: : @xmath110 versus @xmath92 for several uncharged nuclei from @xmath108 to @xmath111 with a neutron excess @xmath112 for the nl1 , nl3 and nl - sh parameter sets .
.values of @xmath26 ( in mev@xmath113 ) and @xmath14 ( in mev ) calculated with the rtf and retf approaches and the scaling method for several parameter sets .
the nuclear matter incompressibility modulus @xmath2 ( in mev ) and the @xmath114 ratio are also listed .
[ cols="^,^,^,^,^,^,^,^ " , ] | by using the scaling method and the thomas - fermi and extended thomas - fermi approaches to relativistic mean field theory the surface contribution to the leptodermous expansion of the finite nuclei incompressibility @xmath0 has been self - consistently computed .
the validity of the simplest expansion , which contains volume , volume - symmetry , surface and coulomb terms , is examined by comparing it with self - consistent results of @xmath0 for some currently used non - linear @xmath1 parameter sets .
a numerical estimate of higher - order contributions to the leptodermous expansion , namely the curvature and surface - symmetry terms , is made .
the curvature of the nuclear matter equation of state ( eos ) , i.e. , the nuclear matter incompressibility @xmath2 is a key quantity in nuclear physics because it is related to many properties of nuclei ( such as radii , masses and giant resonances ) , heavy - ion collisions , neutron stars and supernova collapses .
one important source of information on @xmath2 is provided by the study of the isoscalar giant monopole resonance ( gmr ) ( breathing mode ) in finite nuclei . in the non - relativistic frame , theoretical microscopic calculations based on the random - phase approximation ( rpa ) @xcite and approximations to it such as the scaling method @xcite or constrained calculations @xcite using skyrme @xcite and gogny @xcite effective forces lead to a nuclear matter incompressibility coefficient @xmath2 of @xmath3 mev @xcite . a similar analysis carried out within the relativistic mean - field ( rmf ) theory with non - linear @xmath1 effective lagrangians
gives a value of @xmath2 slightly higher : @xmath4 mev @xcite .
the nuclear matter incompressibilty @xmath2 is not a directly measurable quantity , what is measured is , actually , the energy @xmath5 of the gmr of finite nuclei .
it is convenient to write this energy in terms of the incompressibility @xmath0 for a finite nucleus of mass number @xmath6 as : @xmath7 where @xmath8 is the rms matter radius and @xmath9 the nucleon mass .
the finite nucleus incompressibilty @xmath0 can be parametrized by means of a leptodermous expansion @xcite which is similar to the liquid drop mass formula : @xmath10 where @xmath11 is the neutron excess .
eq .
( [ eqfn2 ] ) suggests that it is possible to fit the coefficients of the expansion to the experimental data in a model independent way .
although some effort along these lines has been made in the past @xcite , the fact that a fit of the parameters of eq .
( [ eqfn2 ] ) to experimental data does not lead to a unique determination of the parameters is well established @xcite .
rather , the nuclear matter incompressibility has to be determined from effective forces which reproduce , in a microscopic calculation , the experimental values of the gmr excitation energy in heavy nuclei @xcite .
it is also possible to fit @xmath0 calculated microscopically within the scaling model for a given effective interaction to the leptodermous expansion eq .
( [ eqfn2 ] ) .
this has been done , for example , in the non - relativistic frame using skyrme forces @xcite . in this case
the coefficients entering eq .
( [ eqfn2 ] ) can be expressed through infinite and semi - infinite nuclear matter properties calculated with the hartree - fock approximation for each considered interaction . in particular , the volume - symmetry ( @xmath12 ) and coulomb ( @xmath13 ) coefficients depend on some parameters of the liquid droplet model @xcite computed only using nuclear matter properties @xcite .
the surface coefficient @xmath14 , also derived in @xcite , can be written as @xcite : @xmath15 . \label{eqfn3}\ ] ] the surface tension @xmath16 is calculated in symmetric semi - infinite nuclear matter and is defined as : @xmath17 where @xmath18 is the density profile whose central value is given by @xmath19 , @xmath20 is the energy density and @xmath21 is the energy per particle in nuclear matter at density @xmath22 . in eq .
( [ eqfn3 ] ) dots indicate the derivatives with respect to the central density and all the quantities are evaluated at a central density equal to the nuclear matter saturation density @xmath23 , which is related to the radius constant @xmath24 through @xmath25 .
the key quantity entering eq .
( [ eqfn3 ] ) is @xmath26 which is the second derivative of @xmath27 with respect to @xmath22 calculated at @xmath28 .
the determination of @xmath26 also requires knowledge of how the density profile @xmath18 is modified during compression @xcite . in the study of the breathing mode a scaling transformation of the densities is assumed .
actually , the coefficients entering the parametrization ( [ eqfn2 ] ) can be derived under this hypothesis @xcite .
the scaling transformation means that the density changes according to the transformation @xmath29 and consequently @xmath30 thus , in the scaling approach : @xmath31_{\rho_0 } = \frac{1}{9\rho_0 ^ 2}\left[\frac{d^2\sigma}{d\lambda^2}\right]_{\lambda=1}. \label{eqfn6}\ ] ] to obtain the surface incompressibility coefficient @xmath14 for a given effective interaction , it is necessary , first of all , to calculate the scaled surface tension @xmath32 by replacing the densities by the scaled densities given by eq .
( [ eqfn5 ] ) in eq .
( [ eqfn4 ] ) . in the non - relativistic frame
this can be easily done within the hartree - fock scheme using zero - range skyrme forces and a simple analytical expression for @xmath32 is obtained @xcite .
the self - consistent calculation of @xmath14 within the rmf approximation using the @xmath1 model is more involved due to the problem of the change in the meson fields induced by the scaled nuclear densities @xcite . to our knowledge ,
only approximate calculations of @xmath14 have been developed in the past for the relativistic model .
this is the case of the relativistic thomas - fermi ( rtf ) calculations of refs .
@xcite where a local density approximation of the meson fields was used .
another approach is related with the study of nuclei under an external pressure .
starting from a schematic energy density functional and adding a density - dependent constraint which simulates the pressure , analytical expressions for the surface tension @xmath16 as a function of the bulk density @xmath22 can be derived for a wide class of compression modes , in particular , for the scaling mode @xcite . this way one obtains the following formula for @xmath33 in the scaling mode : @xmath34 where @xmath35 is the surface diffuseness parameter of a symmetric fermi density .
this pocket formula has been employed to estimate @xmath14 in the rmf model for several non - linear @xmath1 parameter sets @xcite .
a symmetric fermi function that reproduces in the best way the density profile obtained from a hartree calculation of semi - infinite nuclear matter has been used in ref .
@xcite to determine the @xmath35 parameter of eq .
( [ eqfn7 ] ) .
very recently , the scaling method applied to the rmf theory in the rtf and relativistic extended thomas - fermi ( retf ) approaches has been used to self - consistently obtain the excitation energy of the gmr of finite nuclei @xcite .
our aim in the present paper is , firstly to obtain the surface coefficient @xmath14 self - consistently in the rtf and retf approaches developed in @xcite for some linear and non - linear @xmath1 parameter sets .
on the other hand , we want to check whether the leptodermous expansion of the finite nucleus incompressibility eq .
( [ eqfn2 ] ) can reproduce the corresponding fully self - consistent value obtained in the retf approach @xcite with some selected non - linear @xmath1 parameter sets .
the key point of our semiclassical approach is that the local fermi momentum @xmath36 and the effective mass @xmath37 scale as @xcite : @xmath38 where @xmath39 is still a function of @xmath40 . with the help of eq .
( [ eqfn8 ] ) , the nuclear part of the energy and the scalar density including @xmath41 corrections , which are functionals of @xmath36 and @xmath37 , scale as : @xmath42 again @xmath43 and @xmath44 are functions of the collective coordinate @xmath40 because of their dependence on @xmath45 . thus the scaled surface tension can be written as @xcite : @xmath46 \frac{d(\lambda z)}{\lambda } \nonumber \\[3 mm ] & & \mbox { } = \int d(\lambda z)\left \{\lambda^3\tilde{\cal e } + \lambda^2g_{v}v_{\lambda}\rho-\frac{1}{2}\lambda \left [ ( \mbox{\boldmath$\nabla$ } v_{\lambda})^2+\frac{m_{v}^2}{\lambda^2}v_{\lambda}^2\right ] + \frac{1}{2}\lambda \left [ ( \mbox{\boldmath$\nabla$ } \phi_{\lambda})^2 + \frac{m_{s}^2}{\lambda^2}\phi_{\lambda}^2\right ] \right .
\nonumber \\[3 mm ] & & \mbox { } \left .
+ \frac{b\phi_{\lambda}^3}{3\lambda } + \frac{c\phi_{\lambda}^4}{4\lambda } -\lambda^2e_{\infty}(\lambda^3\rho_0 ) \rho\right \ } , \label{eqfn10}\end{aligned}\ ] ] where all densities and fields depend on the variable @xmath47 . with the help of the klein - gordon equations for the scaled vector and scalar fields derived from ( [ eqfn10 ] )
, the scaled surface tension can be recast as : @xmath48 using the explicit rtf or retf expressions for the nuclear part of the energy and for the scalar density @xcite together with the klein - gordon equations for @xmath49 , @xmath50 , @xmath51 and @xmath52 derived from eq .
( [ eqfn10 ] ) , after some algebra the first and second derivatives of the scaled surface tension @xmath32 with respect to @xmath40 at @xmath53 read ( see refs .
@xcite for more details ) : @xmath54 & & \mbox { } \left . -\frac{1}{2}b\phi^3-\frac{1}{4}c\phi^4 + \frac{1}{2}g_{v}\rho v+m_{v}^2 v^2 \right \ } = 0 \label{eqfn12}\end{aligned}\ ] ] and @xmath55 & & \mbox { } \left .
+ 2m_{v}^2v\frac{\partial v_{\lambda}}{\partial \lambda}|_{\lambda=1 } -3m_{v}^2v^2+m\frac{\delta\rho_{s}}{\delta m^ * } \left(m^*+g_{s}\frac{\partial \phi_{\lambda}}{\partial\lambda}|_{\lambda=1 } \right ) -k_{\infty}\rho\right \}. \label{eqfn13}\end{aligned}\ ] ] the first derivative of @xmath56 at @xmath53 is just three times the pressure calculated at saturation density and thus it vanishes , while the second derivative gives @xmath57 @xcite . on the other hand , since in the self - consistent rtf and retf calculations the inputs for computing eqs .
( [ eqfn12])-([eqfn13 ] ) are quantities obtained from the solution of the variational equations associated with the surface tension ( [ eqfn10 ] ) at @xmath53 , the so - called `` sigma dot '' theorem is rigorously fulfilled @xcite .
the method therefore allows @xmath33 and consequently @xmath14 to be computed on top of a self - consistent rtf or retf calculation of the surface tension in symmetric semi - infinite nuclear matter .
this is similar to what happens in the non - relativistic frame with skyrme forces @xcite , although in the relativistic case additional klein - gordon equations for @xmath58 and @xmath59 at @xmath53 have to be solved .
now we shall discuss the results obtained from the self - consistent rtf and retf methods in the scaling approximation .
table 1 collects @xmath2 , @xmath33 and @xmath14 for the non - linear nl - z2 @xcite , nl1 @xcite , nl3 @xcite , nl - ra1 @xcite , nl - sh @xcite and nl2 @xcite and the linear hs @xcite and l1 @xcite parameter sets .
one observes that in both the rtf and retf calculations @xmath33 and @xmath14 decrease ( become more negative ) with increasing bulk incompressibility @xmath2 .
the rtf and retf values of @xmath33 and @xmath14 for a given parameter set are , in general , rather different from one another , which means that the precise value of these quantities is model dependent .
this is known to happen also with other quantities related with the nuclear surface .
for example , such is the case of the surface energy coefficient of the leptodermous expansion of the binding energy of a nucleus , which is calculated as @xmath60 .
the quality of the rtf and retf approximations for semi - infinite nuclear matter and finite nuclei with respect to the rmf hartree approach , and its dependence on the effective interaction , was investigated in refs .
@xcite by analyzing the results obtained with many different parameter sets .
it was noticed that the rtf results fluctuate around the corresponding hartree results . due to
this fact there exist parametrizations for which the rtf approximation agrees by chance with the hartree result .
the behaviour of retf results in comparison with the hartree solutions was found to be less dependent on the parameters of the force than in the rtf case , and it turned out that on the average the retf results are in better agreement with the hartree ones @xcite . the first contribution to @xmath14 in eq .
( [ eqfn3 ] ) comes from the surface tension , let us call it @xmath61 .
the deviation found in the value of the surface tension from rtf calculations with respect to the corresponding rmf hartree calculations is strongly correlated with the value of the effective mass in nuclear matter @xmath62 @xcite . for small values of @xmath62 the rtf surface tension is larger than the hartree one , it practically agrees with the hartree result for @xmath63 , and it becomes smaller than the hartree result for larger @xmath62 . on the other hand ,
the retf result for the surface tension is consistently lower than the hartree result and much less dependent on the specific value of @xmath62 .
( a similar situation is found for the total energy of finite nuclei @xcite . ) these trends , of course , are also reflected in @xmath61 .
for example , for nl1 ( @xmath64 ) we have @xmath65 , 377 and 429.3 mev in the hartree , retf and rtf approaches , respectively . for nl2 ( @xmath66 )
it is @xmath67 , 439.1 and 465.7 mev in the hartree ,
retf and rtf calculation , respectively .
the second contribution to @xmath14 in eq .
( [ eqfn3 ] ) is due to the second derivative of the surface tension .
the results for @xmath33 in the rtf approach decrease with @xmath68 faster than in the retf calculation . at small values of @xmath68
the rtf value of @xmath33 is less negative than that computed in the retf approach , while the opposite happens for higher values of @xmath68 .
both approaches predict the same value of @xmath33 for an incompressibility around that of nl1 ( 211 mev ) .
a similar behaviour is displayed by the self - consistent values of @xmath14 , although the crossing point between the rtf and retf predictions is shifted to a larger value of @xmath68 ( around 280 mev ) due to the fact that the contribution proportional to @xmath16 ( @xmath61 ) is larger in the rtf approach than in the retf approach for the parameter sets considered here .
the @xmath33 values obtained from the pocket formula eq .
( [ eqfn7 ] ) using the surface diffuseness of the rtf or retf semi - infinite nuclear matter density profiles also decrease with @xmath2 , though the estimate provided by eq .
( [ eqfn7 ] ) does not reproduce in general the self - consistent values very precisely .
the approximate @xmath33 is always smaller when calculated from the rtf approach than from the retf approach for the parametrizations of table 1 . using eq .
( [ eqfn7 ] ) to estimate the value of @xmath14 in first approximation , one finds that the retf result is closer to the hartree value than the rtf result for the sets nl - z2 , nl1 , nl3 and nl - ra1 . for nl2 and l1
it is the rtf estimate which lies closer to the hartree estimate . for nl - sh and hs
the approximate hartree prediction lies roughly in between of the approximate rtf and retf values . to the extent that eq .
( [ eqfn7 ] ) is applicable , it provides a hint of where the unknown exact hartree value for @xmath14 should lie with respect to the self - consistent rtf and retf results presented in table 1 .
another different approach to computing @xmath14 was proposed in refs .
@xcite .
it is based on the scaling method together with a local density approximation for the meson fields within the rtf approach . in ref .
@xcite a @xmath14 of approximately @xmath69 mev was reported for a linear set with @xmath70 mev ( similar to the hs set ) .
this result is in good agreement with that of our self - consistent rtf calculation for the hs set , though it is clearly larger in absolute value than the retf result for @xmath14 obtained with the hs set . on the other hand ,
our self - consistent rtf and retf surface incompressibilities differ considerably from the estimate of ref .
@xcite where approximate @xmath14 values of @xmath71 and @xmath72 mev were reported for the nl1 and nl - sh parametrizations calculated with the method used in @xcite .
it should also be pointed out that in our self - consistent semiclassical calculations we find that the ratio between the surface and bulk incompressibilities increases with @xmath2 ( in agreement with the results of @xcite ) . in the retf case
this ratio is close to one , as happens for the non - relativistic skyrme forces @xcite , provided that the bulk incompressibility @xmath2 of the interaction is not excessively high . in the rtf case the ratio between the surface and bulk incompressibilities increases much faster with @xmath2 than in the retf calculations , and it considerably differs from unity for parametrizations with either a very low or a very high bulk incompressibility . in figure 1
we plot @xmath73 as a function of @xmath2 for the parameter sets considered in table 1 . as in the non - relativistic case @xcite
, @xmath14 varies roughly linearly with @xmath2
. a linear fit of all the points gives @xmath74 in the retf model and @xmath75 in the rtf model .
if only the non - linear parametrizations are included in the fit one obtains @xmath76 and @xmath77 in the retf and rtf cases , respectively .
the surface incompressibility coefficient is both large and negative , thus its contribution considerably reduces the finite nucleus incompressibility @xmath0 with respect to the nuclear matter limit @xmath2 .
this result , although obtained in the scaling model , illustrates the physical effect that the compression of the surface provides a considerable reduction of @xmath0 , which is also found in more fundamental rpa calculations @xcite . in ref .
@xcite we have self - consistently computed the finite nucleus incompressibility @xmath0 using the retf approach and the scaling method which we have employed in the present work to obtain @xmath14 .
thus we can now precisely check the ability of the leptodermous expansion eq .
( [ eqfn2 ] ) in reproducing the full calculation of @xmath0 carried out in ref .
@xcite for various finite nuclei .
the coefficients @xmath12 and @xmath13 entering eq .
( [ eqfn2 ] ) are computed using nuclear matter properties only .
explicit expressions for these coefficients in the non - linear @xmath1 model are reported in @xcite . in our analysis
we will use the nl1 , nl3 and nl - sh parameter sets for which the numerical values of these coefficients are given in @xcite .
the surface incompressibility coefficient is the self - consistent value taken from table 1 .
table 2 collects @xmath0 obtained from the full self - consistent retf calculation @xcite as well as the value @xmath78 given by eq .
( [ eqfn2 ] ) for @xmath79ca , @xmath80ca , @xmath81ni , @xmath82zr , @xmath83sn , @xmath84sn , @xmath85sm and @xmath86pb . from this table
it can be seen that the leptodermous expansion with the terms given in eq .
( [ eqfn2 ] ) fails to describe small nuclei and also very asymmetric nuclei such as @xmath84sn or @xmath86pb .
in addition , some words of caution should be said about the coulomb term in eq .
( [ eqfn2 ] ) . in the self - consistent scaling calculation of the finite nucleus incompressibility ,
the coulomb energy does not participate directly if the scaling eq .
( [ eqfn5 ] ) for the density is assumed to apply @xcite .
thus , the coulomb term in eq .
( [ eqfn2 ] ) should be related to the change in @xmath87 when the coulomb interaction is switched off in the self - consistent calculation .
the coulomb term in eq .
( [ eqfn2 ] ) overestimates this change by approximately 6 mev for nl1 , 3 mev for nl3 and 1 mev for nl - sh .
now we would like to analyze whether the addition of some higher order terms in the leptodermous expansion eq .
( [ eqfn2 ] ) improves the agreement with the @xmath0 results calculated self - consistently .
in particular , we will focus our attention on the curvature @xmath88 and the surface - symmetry @xmath89 terms .
although these terms could be derived by enlarging the leptodermous expansion of blaizot @xcite , as has been done in the non - relativistic case @xcite , it becomes more complicated in the relativistic case .
thus , for a fast estimate of the curvature and surface - symmetry terms , we perform a numerical fit . to do this , we follow the same strategy as in ref .
@xcite .
first we consider symmetric nuclei with the coulomb force switched off . in this case
the leptodermous expansion eq.([eqfn2 ] ) ( adding the curvature term ) reduces to : @xmath90 in figure 2 we plot @xmath91a^{1/3}$ ] versus @xmath92 for the three parameters sets used in this analysis . here
@xmath2 is the nuclear matter incompressibility given in table 1 and @xmath87 are the self - consistent incompressibilities calculated for @xmath6 ranging from 300 up to 300000 . in the linear fit of these curves the @xmath93axis
intercept gives @xmath14 of the corresponding force , while the slope gives @xmath94 .
the surface terms obtained in this way are @xmath95246.1 , @xmath95328.4 and @xmath95435.8 mev for the nl1 , nl3 and nl - sh parameter sets , which are very close to the corresponding self - consistent values ( see table 1 ) .
the estimates of the curvature term in the leptodermous expansion of the finite nucleus incompressibility obtained with nl1 , nl3 and nl - sh are @xmath95317.2 , @xmath95229.8 and @xmath95185.6 mev respectively . to obtain the surface - symmetry contribution ,
we have found it convenient to parametrize the difference between the self - consistent incompressibilities @xmath0 of a given nucleus with neutron excess @xmath96 and the corresponding symmetric nucleus as : @xmath97 where again uncharged nuclei have been considered .
for each parameter set and according to ( [ eqfn15 ] ) , if @xmath98i^{-2}$ ] is plotted versus @xmath92 a unique curve should be found which is independent of the value of i. however , one obtains a family of almost parallel lines whose slope is @xmath99 .
the splitting of these lines gives us information on the higher order symmetry contributions missed in the parametrization ( [ eqfn15 ] ) .
thus we will estimate the surface - symmetry term from a linear fit of the curve corresponding to @xmath100 , which roughly corresponds to an average asymmetry along the periodic table .
this curve is plotted in figure 3 for @xmath6 ranging from 250 to 200000 for each considered parameter set .
the corresponding y - axis intercepts agree very well with the @xmath12 values calculated in nuclear matter ( @xmath95676.1 , @xmath95698.9 and @xmath95794.5 mev for nl1 , nl3 and nl - sh respectively @xcite ) .
our estimate of the surface - symmetry contribution to @xmath0 corresponds to the slopes of these linear fits , which are 1951.4 , 1754.0 and 1716.5 mev for nl1 , nl3 and nl - sh respectively .
table 3 collects the self - consistent finite nuclei incompressibility @xmath87 ( without coulomb ) compared with the macroscopic parametrizations @xmath78 ( eq . ( [ eqfn2 ] ) ) and @xmath101 which contains the curvature and surface - symmetry contributions obtained from the previously discussed fits .
again , the self - consistent incompressibilities corresponding to the lightest nuclei and the very asymmetric nuclei are not well reproduced by the simplest expansion eq .
( [ eqfn2 ] ) .
if the curvature and surface - symmetry corrections are included , the improved macroscopic formula @xmath101 reproduces the self - consistent incompressibilities with an error , on average , smaller than @xmath102 , @xmath103 and @xmath104 for the nl1 , nl3 and nl - sh parameter sets . in order to gain some insight into the accuracy of our estimate of the curvature and surface - symmetry contributions
, we fit the self - consistent results for the finite nuclei considered in table 3 to a leptodermous expansion including curvature and surface - symmetry terms . the volume , surface and volume symmetry coefficients are taken from self - consistent infinite and semi - infinite nuclear matter calculations .
the results of this calculation show that the difference of the curvature contribution obtained from the fit in the asymptotic region and from finite nuclei is always less than @xmath105 , whereas the difference in the surface - symmetry contribution lies below @xmath106 .
we have applied the scaling method in the thomas - fermi and extended thomas - fermi approximations to the relativistic mean field theory to self - consistently calculate the surface coefficient @xmath14 of the leptodermous expansion of the finite nucleus incompressibility derived within the blaizot model .
the ratio between the surface and bulk incompressibilities obtained in our semiclassical calculation increases with the nuclear matter incompressibility , more strongly in the rtf than in the retf case . in the retf calculations this ratio is close to one , as in the case of non - relativistic skyrme forces , for the non - linear parameter sets which have a nuclear matter incompressibility not larger than roughly 300 mev . for the analyzed @xmath1 parameter sets , the leptodermous expansion eq .
( [ eqfn2 ] ) is not able to reproduce very precisely the finite nuclei incompressibilities obtained self - consistently .
in particular the macroscopic contribution of the coulomb force can differ from the self - consistent contribution up to 6 mev .
we have numerically estimated higher order contributions to the leptodermous expansion , namely curvature and surface - symmetry terms , in the asymptotic region ( i.e. , for very large uncharged systems ) .
we have found that the finite nuclei incompressibilities are reasonably well reproduced by an extended leptodermous expansion which includes curvature and surface - symmetry contributions .
useful discussions with e. vives , m. farine and j.n .
de are acknowledged .
the authors would like to acknowledge support from the dgicyt ( spain ) under grant pb98 - 1247 and from dgr ( catalonia ) under grant 2000sgr-00024 .
s.k.p .
thanks the spanish education ministry grant sb97-ol174874 for financial support and the departament destructura i constituents de la matria of the university of barcelona for kind hospitality .
= -1.5 mm p. ring and p. schuck , _ `` the nuclear many - body problem '' _ , springer - verlag , berlin ( 1980 ) .
j.p .
blaizot , phys . rep .
* 64 * ( 1980 ) 171 .
o. bohigas , a. lane and j. martorell , phys . rep .
* 51 * ( 1979 ) 267 .
b.k .
jennings and a.d .
jackson , phys .
rep .
* 66 * ( 1980 ) 141 .
j. treiner , h. krivine , o. bohigas and j. martorell , nucl . phys .
* a371 * ( 1981 ) 253 .
j.p .
blaizot , j.f .
berger , j. decharg , and m. girod , nucl .
phys .
* a591 * ( 1995 ) 435 .
m. farine , j.m .
pearson and f. tondeur nucl .
phys . *
a615 * ( 1997 ) 135 .
zhongyu ma , nguyen van giai , a. wandelt , d. vretenar , and p. ring , nucl .
phys .
* a686 * ( 2001 ) 173 .
m.m .
sharma , w.t.a .
borghols , s. brandeburg , s. crona , a. van der woude and m.n .
harakeh , phys . rev .
* c38 * ( 1988 ) 2562 ; m.m .
sharma , w. stocker , p. gleissl and m. brack , nucl .
phys . *
a504 * ( 1989 ) 337 .
j.m .
pearson , phys .
lett . *
b271 * ( 1991 ) 1 . s. shlomo and d.h .
youngblood , phys .
rev .
* c47 * ( 1993 ) 529 .
r.c .
nayak , j.m .
pearson , m. farine , p. gleissl and m. brack , nucl .
phys .
* a516 * ( 1990 ) 62 .
w.d .
myers and w.j .
swiatecki , ann .
phys .
( n.y . ) * 55 * ( 1969 ) 395 .
j.m .
pearson , m. farine and j. ct , phys . rev .
* c26 * ( 1982 ) 267 . m. brack and w. stocker , nucl
. phys .
* a388 * ( 1982 ) 320 ; nucl . phys .
* a406 * ( 1983 ) 413 .
w. stocker and m.m .
sharma , z. phys .
* a339 * ( 1991 ) 147 .
m.m .
sharma , m.a .
nagarajan and p. ring , ann . of phys .
( ny ) * 231 * ( 1994 ) 110 .
d. von - eiff , j. m. pearson , w. stocker and m.k .
weigel phys . rev .
* c50 * ( 1994 ) 831 .
t. v. chossy and w. stocker , phys . rev .
* c56 * ( 1997 ) 2518 .
s.k .
patra , m. centelles , x. vias and m. del estal , phys . lett .
* b523 * ( 2001 ) 67 .
s.k .
patra , x. vias , m. centelles and m. del estal , nucl . phys . *
a * ( 2002 ) , in press .
m. centelles , x. vias , m. barranco and p. schuck , ann . of phys .
( n.y . ) * 221 * ( 1993 ) 165 ; m. centelles and x. vias , nucl
. phys . *
a563 * ( 1993 ) 173 ; m. del estal , m. centelles and x. vias , phys . rev .
* c56 * ( 1997 ) 1774 .
m. centelles , m. del estal and x. vias , nucl .
phys .
* a635 * ( 1998 ) 193 .
m. centelles , x. vias , m. barranco , s. marcos and r.j .
lombard , nucl . phys .
* a537 * ( 1992 ) 486 .
c. speicher , e. engel and r.m .
dreizler , nucl . phys .
* a562 * ( 1993 ) 569 .
s. nishizaki , h. kurasawa and t. suzuki , nucl . phys .
* a462 * ( 1987 ) 689 .
m. farine and w. stocker , nucl . phys .
* a454 * ( 1986 ) 71 .
m. bender , k. rutz , p .- g .
reinhard , j.a .
maruhn and w. greiner , phys .
rev .
* c60 * ( 1999 ) 034304 .
p .-
g .
reinhard , m. rufa , j. maruhn , w. greiner and j. friedrich , z. phys .
* a323 * ( 1986 ) 13 .
g.a .
lalazissis , j. knig and p. ring , phys . rev .
* c55 * ( 1997 ) 540 .
m. rashdan , phys . rev .
* c63 * ( 2001 ) 044303 .
m.m .
sharma , m.a .
nagarajan and p. ring , phys . lett . *
b312 * ( 1993 ) 377 .
s.j .
lee , j. fink , a.b .
balantekin , m.r .
strayer , a.s .
umar , p .-
g .
reinhard , j.a .
maruhn and w. greiner , phys .
rev .
lett .
* 57 * ( 1986 ) 2916 .
c.j . horowitz and b.d .
serot , nucl . phys .
* a368 * ( 1981 ) 503 .
m.v .
stoitsov , m.l .
cescato , p. ring and m.m .
sharma , j. phys .
* g20 * ( 1994 ) l149 . |
the study of close binary stars is of interest for several reasons .
for example , the understanding the structure and evolution of stars is a basic goal of stellar astronomy , and is required in most other branches of astronomy .
critical tests of evolution theory ( i.e. predicting the radius and luminosity of a star as a function of its mass and age ) for stars other than the sun are practical only for a small set of eclipsing binary stars ( see , e.g. , pols et al .
@xcite ; schrder et al .
@xcite ; lacy et al .
in addition , our knowledge of the masses of stellar black holes ( and many neutron stars and white dwarfs as well ) depends on our ability to interpret ellipsoidal and usually non - eclipsing light curves ( e.g. avni & bahcall @xcite ; avni @xcite ; mcclintock & remillard @xcite ; haswell et al .
@xcite ; shahbaz et al .
@xcite ; sanwal et al .
@xcite ; orosz & bailyn @xcite ) .
finally , considerable observational effort has been put forth recently to use detached eclipsing binaries as extragalactic distance indicators ( mochejska et al .
@xcite ; ribas et al .
@xcite ) . since close binary stars are of such importance , it is crucial that we have the ability to construct accurate synthetic light curves for a variety of close binaries . the light curve expected from a particular close binary depends on the system geometry ( i.e. the figures of the stars , their relative sizes , separation , viewing angle , etc . ) and on the radiative properties of the stars , which are set mainly by their effective temperatures , surface gravities and chemical compositions .
the equations describing the basic system geometry for a close binary are reasonably simple and have been known for a long time ( e.g. see the text by kopal @xcite ) . in practice , however , the direct computation of light curves requires a computer , and codes to compute light curves have been in use since the late 1960s ( e.g.lucy @xcite ; hill & hutchings @xcite ; wilson & devinney @xcite ; mochnacki & doughty @xcite ; avni & bahcall @xcite ; see also the review by wilson @xcite ) . on the other hand , the detailed computation of stellar atmosphere models which describe the specific intensity of radiation emitted by the stellar surfaces is quite involved . as a result
, approximations are frequently used in the computations of light curves .
the planck function is usually used to compute the normal monochromatic intensity of a surface element with an effective temperature @xmath3 : @xmath4^{-1}\ ] ] where @xmath5 is the effective wavelength of the observation and @xmath6 , @xmath7 , and @xmath8 are the usual physical constants .
after @xmath9 is found , the specific intensities for other emergent angles are computed using a simple parameterisation ( the `` limb darkening law '' ) : @xmath10 .
\label{linear}\ ] ] the coefficient @xmath11 depends on the temperature , gravity , and chemical composition of the star being modelled .
there are also many two - parameter limb darkening laws including a `` quadratic '' law : @xmath12 , \label{quad}\ ] ] the `` square root '' law ( daz - cordovs & gimnez @xcite ) : @xmath13 , \label{sqrt}\ ] ] or the `` logarithmic '' law ( klinglesmith & sobieski @xcite ) : @xmath14 . \label{log}\ ] ] there are many tabulations of the coefficients for these various laws for a wide variety of temperatures , gravities , and chemical compositions ( e.g. al - naimiy @xcite ; wade & rucinski @xcite ; van hamme @xcite ; claret @xcite ) .
we discuss in this paper our technique for computing light curves using model atmosphere specific intensities directly , thereby eliminating the need for the black body approximation and the one- or two - parameter limb darkening laws .
although our emphasis here is on stars with low effective temperatures and surface gravities ( @xmath0 k and @xmath15 ) , the technique is quite general . in the next section
we discuss some previous work in this area and give details of our method .
we then discuss some of the basic results and their implications , and give a summary of this work .
we have also included an appendix to the paper which gives some of the details of our light curve code not directly related to the stellar atmospheres .
the use of model atmosphere computations inside a light curve synthesis code is by no means new .
the widely used wilson - devinney ( @xcite ; hereafter w - d ) code has routine which applies a correction to the normal ( black body ) intensity : @xmath16 where @xmath17 is the ratio of the filter - integrated stellar atmosphere model characterised by @xmath3 and @xmath18 to the filter - integrated blackbody intensity .
the specific intensities for other angles are then computed from @xmath19 using the limb darkening law .
the correction routine supplied with the w - d code is based on the carbon & gingerich ( @xcite ) models .
r. e. wilson ( priv .
comm . ) informs us he is in the process of updating the correction routine .
milone et al .
( @xcite ) have independently written a correction routine for the w - d code based on the kurucz ( @xcite ) models .
linnell & hubeny ( @xcite , @xcite ) have written a series of codes to compute synthetic spectra and light curves for binary stars , including ones with disks .
they use hubeny s general spectrum synthesis code synspec ( hubeny et al .
@xcite ) to generate the model spectra , using as input kurucz ( @xcite ) atmosphere models for cool stars ( @xmath20 k ) and tlusty ( hubeny @xcite ) and tlusdisk ( hubeny @xcite ) models for hotter stars and disks , respectively .
they have applied their model with some success to ( linnell et al .
@xcite ) and ( linnell et al .
tjemkes et al .
( @xcite ) have used kurucz ( @xcite ) models in their light curve synthesis code by tabulating filter - integrated normal intensities for a grid of effective temperatures and surface gravities .
specific intensities for other emergent angles are computed from a limb darkening law tabulated from kurucz models .
this code has been applied successfully to x - ray binaries such as ( heemskerk & van paradijs @xcite ) and ( van der hooft et al .
@xcite ) and to the millisecond pulsar ( callanan et al .
shahbaz et al .
( @xcite ) mention of the use of kurucz ( @xcite ) fluxes in a code used at oxford , but this paper does not specifically describe how the model atmosphere fluxes are incorporated into the light curve synthesis code . similarly , sanwal et al .
( @xcite ) mention the use of bell & gustafsson ( @xcite ) model atmosphere fluxes in the light - curve synthesis code developed at the university of texas at austin ( zhang et al .
@xcite ; haswell et al .
@xcite ) without giving specific details .
we use a technique suggested to us by marten van kerkwijk to incorporate model atmosphere intensities into our light curve synthesis code . in general
, a detailed model atmosphere computation yields the specific intensity @xmath21 at a given wavelength @xmath5 and emergent angle @xmath22 , where @xmath23 at the centre of the apparent stellar disk and @xmath24 at the limb .
the total disk - integrated intensity @xmath25 observed at the wavelength @xmath5 is then @xmath26 normally , the light curves of a binary star are observed in a broad bandpass .
the observed intensity @xmath27 in a given bandpass is @xmath28 w_{\rm filt}(\lambda ) d\lambda , \label{eq1}\end{aligned}\ ] ] where @xmath29 is the wavelength - dependent transmission of the filter bandpass in question .
we can reverse the order of the integrals in eq .
( [ eq1 ] ) to give @xmath30\mu d\mu \cr & = & \int^1_0i_{\rm filt}(\mu)\mu d\mu . \label{eq2}\end{aligned}\ ] ] the quantity in square brackets in eq .
( [ eq2 ] ) , namely @xmath31 is independent of any geometry ( for non - irradiated atmospheres ) , and as such can be computed in advance of the light curve synthesis computations . in our current implementation
we compute for each model characterised by a temperature @xmath3 and gravity @xmath18 a table of eight filter - integrated intensities for each specific angle @xmath32 .
the eight filters currently are the johnson @xmath33 system where we have used the optical filter response curves given in bessell ( @xcite ) and the infrared filter response curves given in bessell & brett ( @xcite ) .
using all of the models we then generate a table of the form [ cols= " < " , ] the first two numbers are the temperature and gravity of the first model .
the next line gives the number of specific intensities that follow .
the next @xmath34 lines give the value of @xmath32 followed by the eight filter - integrated intensities .
the format is then repeated for the additional models .
the table is sorted in order of increasing temperature , and for each temperature , the table is sorted in order of increasing gravity . during the course of computing a synthetic light curve the specific intensity of each surface element must be specified .
these values of @xmath35 are interpolated from the table using a simple linear interpolation procedure .
first , we locate the 4 nearest models ( @xmath36 ) , ( @xmath37 ) , ( @xmath38 ) , and ( @xmath39 ) , where @xmath40 and @xmath41 .
next , within each of the 4 nearest models we find the filter - integrated intensities appropriate for @xmath42 , using linear interpolation . finally , the desired values of + @xmath35 are found by linear interpolation first in the @xmath18 direction and then the @xmath3 direction . returning the filter - integrated intensities for eight different filters at once
is advantageous when modelling observations in several bandpasses since one has to search the intensity table only once per surface element per phase .
= 8.2 cm = 8.2 cm = 8.2 cm as we noted above , the technique we just outlined is quite general .
we now turn our attention to the recent grid of nextgen models for cool giants computed by phh using the parallelised version of his general purpose code phoenix ( hauschildt et al .
the computational techniques and the input physics used to compute the nextgen models for cool giants have been discussed elsewhere ( hauschildt et al .
@xcite ) and will not be repeated here .
these spherical models require an input mass to compute the sphericity factor , and we use @xmath43 for all of the models .
we did examine a few models with masses of @xmath44 and @xmath45 and found no discernible differences compared to the @xmath43 models . for the moment we restrict ourselves to the models with solar metallicity .
for each model the specific intensities are computed for 64 different angles over a wavelength range of 3000 - 24,998 in 2 steps .
the distribution of the emergent angles @xmath32 is chosen by the phoenix code based on the structure of the atmosphere under consideration .
hence , in general , the 64 values of @xmath32 are different from model to model . fig .
[ fig1 ] shows the quantity @xmath46 for two different nextgen models and filter combinations .
the top panel shows the integrated intensities in the johnson @xmath47 and @xmath48 filters for the model with @xmath49 k and @xmath50 and the bottom panel shows the integrated intensities in the johnson @xmath51 and @xmath52 filters for the model with @xmath53 k and @xmath54 . for comparison , we also show the square root limb darkening laws ( eq . ( [ sqrt ] ) )
computed using coefficients taken from the tabulation of van hamme ( @xcite ) .
these four curves essentially show the limb darkening behaviour predicted by the kurucz models to within a few per cent , although is important to note that the intensities for @xmath55 are different between the kurucz models and the nextgen models ( see hauschildt et al .
( @xcite ) for further discussion of this point ) .
this figure is quite striking .
it shows the radically different limb darkening behaviour predicted by the spherical nextgen models and the plane - parallel kurucz models . the nextgen model predicts a sharp decrease in the intensity at relatively large @xmath32 values ( as large as @xmath56 ) , whereas the limb darkening parameterisations predict substantial intensity all the way to @xmath24
this sudden decrease in the intensity is a consequence of the spherical geometry : for sufficiently low gravities ( @xmath57 for most temperatures ) there is much less material near the limbs and hence much less radiation .
the radiation that would have come out near the limb in the redder bandpasses comes out at much shorter wavelengths ( see hauschildt et al .
@xcite ) . the value of @xmath32 where the sudden fall - off in the intensities occurs depends on the effective temperature and gravity of the stellar model , and on the bandpass .
we define the slope of the intensities at @xmath58 to be @xmath59 the location @xmath60 of the edge is defined to be the point where @xmath61 is the largest . fig .
[ fig3 ] shows contour plots of @xmath60 for three different bands ( johnson @xmath62 , @xmath48 , and @xmath52 ) .
in general , for a fixed effective temperature , @xmath60 gets larger as the gravity of the model decreases .
also , for a fixed @xmath18 , the hotter models tend to have a slightly larger values of @xmath60 .
there is some irregular behaviour seen in the contours of @xmath60 at temperatures between about 3000 k and 4000 k. the reason for this seems to be that for models in this temperature range , the fall - off in the intensity is generally less sudden , especially in the bluer filters .
thus @xmath60 is less clearly defined in these cases .
it is quite obvious that the quantity @xmath27 ( eq.([eq2 ] ) ) evaluated for the nextgen intensities will be less than the corresponding quantity evaluated for a one- or two - parameter limb darkening law . for each model and filter
, we computed the `` lost light '' in the following way .
we first `` linearised '' our table of nextgen intensities : we fitted a line to the intensities for the ten specific angles nearest to @xmath55 .
then the specific intensities for the remaining 54 angles were replaced by the intensities determined from the extrapolation of the fitted line . in most cases ,
the extrapolated intensities at the limb were above zero .
there were a few cases such as the @xmath53 , @xmath54 @xmath51-band model shown in fig .
[ fig1 ] where the extrapolated intensities at the limb would have been negative . in such cases
we replaced all negative intensities with zero .
( [ eq2 ] ) was evaluated for the linearised table and the regular table , and we define the lost light as @xmath63 @xmath64 @xmath65 @xmath66 @xmath67 . fig .
[ figlost ] shows contour plots of + @xmath63 for the @xmath62 ( top ) , @xmath48 ( middle ) , and @xmath52 ( bottom ) filters . by our definition
the lost light in many of the @xmath62-band models is zero .
otherwise , for the @xmath62 band , the lost light is typically a few per cent and about 3 or 4 per cent for the hottest models .
on the other hand , the lost light in the @xmath48 and @xmath52 bands can be quite substantial , up to 17 per cent for the models with lowest gravities and the hottest temperatures .
we can easily make an image of the disk a model star as it would appear in the sky .
the intensity maps displayed in fig . [ fig2 ] compare the nextgen intensities for @xmath53 k and @xmath54 to monochromatic black body intensities ( limb darkened using the square root law ( eq . ( [ sqrt ] ) ) for three different filters ( johnson @xmath62 , @xmath47 , and @xmath48 ) .
the star with the nextgen intensities appears smaller on the sky .
the difference in the limb darkening closer to the disk centre is also apparent from the figure .
finally , for comparison purposes , we generated an intensity table from kurucz models .
we retrieved the file `` ip00k2.pck19 '' from kurucz s public world wide web pages .
this file contains the specific intensities for 17 different angles for a grid of solar abundance models .
the coolest models have @xmath68 k. all of these models are plane - parallel and lte .
we have written a new light curve synthesis code which is based on the work of avni ( avni & bahcall @xcite ; avni @xcite ) .
this code , named elc , can model any semi - detached or detached binary with a circular orbit ( we will generalise the code to include overcontact binaries and binaries with eccentric orbits within the next year ) .
the second star can be surrounded by an accretion disk .
we will defer giving specific details of the code to the appendix and now turn our attention to the results related to the use of the nextgen intensities . for various test cases we computed light curves in four different ways : using our elc code in the black body mode , using the w - d code in the black body mode
( we have added a phase shift of 0.5 to the w - d models , see the appendix ) , using our elc code with the kurucz intensity table , and using our elc code with the nextgen intensity table .
[ fig4 ] shows four different normalised light curves ( in @xmath47 , @xmath48 , and @xmath52 ) for a single roche lobe filling star in synchronous rotation with @xmath49 k and @xmath50 .
the mass ratio is 5 , meaning the unseen second star is five times more massive than the cool giant .
the inclination is 75 degrees . for the black body models we used the square root limb darkening law with coefficients from van hamme ( @xcite ) .
the gravity darkening exponent was @xmath69 , appropriate for a star with a convective envelope .
( we note that the black body light curves were computed using a single limb darkening law for the entire star , whereas the light curves computed using model atmosphere intensities ( nextgen or kurucz ) will have location - specific limb darkening `` built in '' . ) in all three cases , the two black body curves are nearly identical ( the largest deviations between the two are on the order of 2 millimags ) . typically the nextgen curves have deeper minima than the black body curves . depending on the filter , the nextgen light curves are different from the the light curves computed with kurucz intensities , although for the example shown in fig .
[ fig4 ] the two @xmath48-band curves have only minor differences .
= 8.2 cm = 8.2 cm in general , the light curves computed using the next - gen intensity table are different from light curves computed using the same geometry but with black body intensities .
this comes as no surprise , given the potentially large differences in the limb darkening between the next - gen models and the one- or two - parameter limb darkening laws .
it is usually the case ( especially in the @xmath48 , @xmath70 , @xmath71 , and @xmath52 filters ) that the nextgen light curves have larger amplitudes ( in normalised intensity or in magnitudes ) than the corresponding black body light curves .
[ fig5 ] shows why this is so .
the top curve is the @xmath48-band light curve of the single star described in the fig .
[ fig4 ] caption made with the linearised intensity table .
since the square root limb darkening law for this model ( @xmath49 , @xmath50 ) and bandpass closely matches the actual nextgen intensities near disk centre ( see fig .
[ fig1 ] ) , the corresponding black body light curve is nearly identical to the `` linearised '' light curve .
the middle curve in fig .
[ fig5 ] is the light curve for the same star made with the regular nextgen intensity table , plotted on the same intensity scale .
the average light level in this light curve is about 5 per cent lower than the linearised one , which is roughly what we would expect from fig .
[ figlost ] .
the bottom curve is the difference between the two top light curves .
this lower curve is basically the light lost near the limb , and it is essentially constant with phase .
the two top curves have roughly the same amplitude in these intensities units .
that is , @xmath72 is the same for both curves .
however , the _ relative _ amplitude @xmath73 for the regular nextgen light curve will be larger since its mean light level is lower . hence , the amplitude of this particular nextgen light curve ( and most others ) in _ normalised intensity or in magnitudes _ is larger than the amplitude of the corresponding black body curve .
there are some situations such as the @xmath51 model shown in the bottom of fig .
[ fig1 ] where the square root limb darkening law does not match the slope of the nextgen intensities near @xmath23 . for cases like this
, the black body light curve can have a larger amplitude than the nextgen light curve .
[ fig4 ] nicely shows the difference in the light curves between black body intensities and model atmosphere intensities ( e.g. either nextgen or kurucz ) .
however , in the example shown , the nextgen light curves and the kurucz light curves had only minor differences . to illustrate how large the difference between a nextgen light curve and a kurucz light curve can be we consider a binary similar to .
this semidetached binary consists of a b3ib star which rotates near its centrifugal limit and a roche lobe - filling a2 ?
star ( e.g. olson & etzel @xcite ) .
the mean gravities of the two stars are @xmath74 and @xmath75 , so sphericity effects should be important here . as a result of its rapid rotation near the critical limit ( see van hamme & wilson @xcite for a discussion of `` critical rotation '' ) , the b - star is significantly flattened .
furthermore , the surface gravity near its equator is relatively low , giving rise to a relatively large range in temperatures owing to the increased gravity darkening .
we currently can not model the real binary using the nextgen intensities because the mean temperature of the b - star is well outside our current model grid .
indeed , there are parts on the b - star where the temperature and gravity combination fall outside the kurucz grid , so we can not model the real binary with the kurucz grid either . for this example we therefore modified the temperatures of the two stars and `` slowed down '' the mass gaining primary slightly so that the temperature and gravity combination of each point on each star is contained in both the nextgen and kurucz grids .
the adopted model parameters are summarised in the caption of fig .
[ rzscuti ] . fig .
[ rzscuti ] itself shows the difference between the nextgen light curve and the kurucz light curve _ in magnitudes _ for three filters .
there are large differences between the two models , and the size of the difference depends on the filter bandpass .
the difference between @xmath52-band light curves is as large as 0.05 mag near phase 0 , when the cooler star is eclipsed by the hotter star .
one can also note from fig .
[ rzscuti ] that the @xmath51 and @xmath47 difference curves have means which are less than zero . in other words
the binary is _ bluer _ in @xmath76 and @xmath77 when nextgen intensities are used compared to when kurucz intensities are used .
= 8.2 cm = 8.2 cm to quantify how much the difference in the light curve amplitudes between the nextgen models and the black body models might matter when fitting the light curves of a real binary star , we consider a model binary similar to .
is an s - type symbiotic binary where an m4 giant orbits an optically faint hot companion with a period of about 227.6 days ( kraft @xcite ; kenyon & garcia @xcite ; fekel et al .
there are no eclipses observed in the uv ( selvelli et al .
@xcite ) . nevertheless , the large amplitude of the ellipsoidal light curves suggests the m giant fills its roche lobe and is viewed at a large inclination ( bailey @xcite ; lines et al .
@xcite ; yudin & munari @xcite ) . since the density of a roche lobe filling star is a function only of the orbital period to a good approximation ( faulkner et al .
@xcite ; eggleton @xcite ) , the surface gravity of the m giant is only a weak function of its assumed mass . in the case of ,
@xmath78 , so we expect the sphericity effects to be important here .
the mass ratio of the binary is not well known .
kraft ( @xcite ) measured hydrogen emission line radial velocities on seven plates and found @xmath79 .
however , this mass ratio implies a rotational velocity of @xmath80 km s@xmath81 , which is much larger than the upper limit of @xmath82 km s@xmath81 measured by kenyon & garcia ( @xcite ) .
we discuss below some of the potential problems associated with measuring the rotational velocity of a tidally distorted cool giant , and in view of this discussion the rotational velocity upper limit of kenyon & garcia ( @xcite ) should be treated with caution .
the @xmath70 and @xmath52 band light curves of t crb collected between 1987 august and 1995 june are stable and representative of the true ellipsoidal modulation ( yudin & munari @xcite ; shahbaz et al .
the full amplitude of the @xmath70 light curve is about 0.2 magnitudes . on the other hand ,
the @xmath47 light curve shows additional sources of variability not associated with the underlying ellipsoidal modulation ( e.g. lines et al .
nevertheless , between about 1989 and the beginning of 1996 , the @xmath47 light curve seemed to be reasonably stable .
hric et al .
( @xcite ) refer to this period as the `` quiet stage '' .
the full amplitude of the @xmath47 light curve during this quiet stage is about 0.4 magnitudes . for this discussion
we will assume this light curve represents the true ellipsoidal light curve .
both shahbaz et al .
( @xcite ) and belczyski & mikolajewska ( @xcite ) had difficulty fitting the optical ( @xmath47 or @xmath48 ) light curves simultaneously with the infrared ( @xmath70 or @xmath52 ) light curves .
if the amplitude of the @xmath70 band light curve was matched , then the model @xmath47 and @xmath48 light curves ( computed using black body intensities ) had amplitudes that were about a factor of two too small .
a thorough analysis of existing data is beyond the scope of the present paper and will be deferred to a future paper .
for now it will suffice to discuss simulated t crb - like light curves . for these light curves
we will use the system parameters adopted by belczyski & mikolajewska ( @xcite ) , namely @xmath83 , which in our notation is @xmath84 , @xmath85 , @xmath86 k , and a gravity darkening exponent of @xmath69 , appropriate for a star with a convective outer envelope .
we also used the orbital period and @xmath52-velocity of the m giant given in kenyon & garcia ( @xcite ) . using the nextgen intensity table we computed light curves for the johnson @xmath47 and @xmath70 filters
the model light curves were converted to magnitudes for compatibility with our optimising routines .
we used elc in its black body mode to fit various combinations of the simulated light curves , and the results are shown in fig .
[ fitfig ] .
the top panel shows the results of fitting the @xmath70 light curve only , using the inclination as the sole free parameter .
the fitted inclination is @xmath87 , and the @xmath88 residuals for the @xmath70 band are reasonably small ( @xmath89 mag ) .
however , if we take the predicted @xmath47 light curve for this geometry and compare it to the simulated @xmath47 light curve , we find that the amplitude of the black body @xmath47 light curve is much _ smaller _ than the amplitude of the simulated @xmath47 ( nextgen ) light curve .
thus we have reproduced the same problem that shahbaz et al .
( @xcite ) and belczyski & mikolajewska ( @xcite ) had when they attempted simultaneous optical and infrared fits .
not surprisingly , if we attempt to fit both our simulated @xmath47 and @xmath70 light curves simultaneously using the inclination as the only free parameter , the fits to the @xmath70 light curve get worse ( middle panel ) .
finally , we fit both light curves simultaneously using both the inclination and mass ratio as free parameters .
the @xmath88 residuals for both filters are not excessively large and might be comparable to the observational errors in the real binary
. however , the fitted parameters are quite different from the input ones : @xmath90 and @xmath91 , compared to input values of @xmath92 and @xmath93 . naturally
, the derived component masses from the fit are significantly different from the input masses : @xmath94 and @xmath95 , compared to input values of @xmath96 and @xmath97 . measurements of the rotational velocities of the stars in close binaries provide powerful constraints on the light curve solution . indeed , any light curve solution that specifies the mass ratio , inclination , and the angular velocity ratios predicts specific values for the observed values of @xmath98 , provided of course that at least one radial velocity curve is available . in cases where a star fills its critical lobe exactly and is in synchronous rotation ( generally safe assumptions in cataclysmic variables and low mass x - ray binaries ) ,
a measurement of its rotational velocity constrains the mass ratio of the binary ( e.g. wade & horne @xcite ) .
if spectra with high resolution and high signal - to - noise are available , then one can use fourier techniques to measure the rotational velocity ( e.g. gray @xcite and cited references ) .
when only spectra of lower quality are available , then often one measures @xmath98 by comparing the spectrum of interest with a spectrum of a slowly rotating template star ( observed with similar instrumentation ) that has been convolved with a broadening kernel @xmath99 ) .
various trial values of the width of the broadening kernel are tried until the optimal match is found .
gray ( @xcite , pp . 370 - 374 ) gives a clear description of how to compute the broadening kernel @xmath100 analytically for the case where the intrinsic line profile @xmath101 has the same shape over the entire disk .
essentially , one can place an @xmath11,@xmath102 coordinate system on the apparent disk of the star on the sky where the @xmath102-axis is parallel to the axis of rotation . the disk of the star
can then be divided up into a number of strips parallel to the @xmath102-axis , each having its own doppler shift according to its @xmath11-coordinate .
the broadening kernel is evaluated at a particular doppler shift by integrating the intensity over the appropriate strip .
normally one assumes a linear limb darkening law with a coefficient of @xmath103 when computing @xmath100 . finally , the broadened line profile is the convolution of the intrinsic line profile with the kernel : @xmath104 .
= 8.2 cm = 8.2 cm the above discussion applies to spherical stars .
a star that fills a substantial fraction of its roche lobe departs markedly from spherical symmetry , and as such will have distorted line profiles ( e.g. kopal @xcite ; marsh et al .
@xcite ; shahbaz @xcite ) .
furthermore , the degree of the line profile distortion depends on the roche geometry ( i.e. mass ratio and inclination ) and the orbital phase ( shahbaz @xcite ) . the intensity maps that the elc program can write ( e.g. fig .
[ fig2 ] ) can be used to numerically evaluate @xmath99 ) , thereby accounting for both the nonspherical shape of the star and the nonlinear limb darkening .
[ fig6 ] shows rotational broadening kernels @xmath100 for the star described in the caption to fig .
[ fig4 ] for phases 0.0 , @xmath105 , and @xmath106 , when the giant is behind its invisible companion .
the situation is complicated .
near phase 0.0 , the projected surface of the giant in the plane of the sky is not too different from a circle ( it is flattened slightly in the @xmath102-direction )
. thus the observed profile will be nearly symmetric .
however , the kernel , computed using the nextgen intensities , is _ narrower _ than the analytic kernel and the kernel computed using black body intensities and linear limb darkening since the star with the nextgen intensities appears smaller on the sky ( fig .
[ fig2 ] ) . at phase @xmath105 ( quadrature ) ,
the star is elongated in the @xmath11-direction . as a result ,
the broadening kernel computed from the black body intensities is _ wider _ than the analytic kernel , and it is asymmetric . as a result of the sharp cut - off in the limb darkening , the kernel computed using the nextgen intensities in narrower than the black body kernel .
however , for this particular situation , the nextgen kernel has roughly the same width as the analytic kernel . finally , at phase @xmath106 , when the giant is behind its unseen companion , all three kernels have more or less the same full width at half maximum .
however , the kernel computed using the black body intensities is flat - topped ( shahbaz ( @xcite ) also noted absorption line profiles with flat bottoms near this phase ( his fig . 1 ) ) , and the kernel computed with the nextgen intensities has a local minimum at zero velocity .
the reason the kernels have flat tops or even central depressions near phase @xmath106 is quite straightforward .
the @xmath107 point is in full view near this phase , and for high inclinations there are no points on the apparent stellar disk with @xmath23 . for the example shown in fig .
[ fig6 ] , @xmath108 everywhere on the star at phase @xmath106 .
hence the central parts of the apparent stellar disk are fainter ( owing to limb darkening ) than they would have been in the spherical case and there is less contribution to the kernel near zero velocity .
if the intensities are computed using the nextgen table and the gravity is sufficiently low , then the central part of the star can be even darker . in some cases such as the one shown in fig .
[ fig6 ] , the @xmath107 point will be so dark that the broadening kernel will actually have a central depression near zero velocity . given the variety of distortions in the broadening kernel as shown in fig .
[ fig6 ] , it would be prudent to compute phase - specific broadening kernels when extracting @xmath109 measurements from spectra ( see also shahbaz @xcite ) . as a specific example , we consider the eclipsing binary , which consists of a k giant in a 96.7 day orbit about an essentially invisible ( in the optical ) companion ( e.g. eaton et al .
there is good evidence for an accretion disk around the unseen companion ( zola et al.@xcite ) , so we will assume that the k giant fills its roche lobe exactly and that it is in synchronous rotation .
eaton et al .
( @xcite ) obtained a series of high resolution spectra and noted a change in the doppler broadening of certain metallic absorption lines as a function of phase .
this behaviour is consistent with expectations ( shahbaz @xcite ) . using a linear limb darkening law with @xmath110 , eaton et al .
( @xcite ) determined a rotational velocity of @xmath111 km s@xmath81 for a spectrum near a quadrature phase and derived a mass ratio of @xmath112 ( using elc , we derive a mass ratio @xmath113 from @xmath111 , where we have adopted a mass function for the unseen companion of @xmath114 ( popper @xcite ) ) .
we plot in fig .
[ fig7 ] the analytic broadening kernel for @xmath110 and a mean rotational velocity of @xmath115 km s@xmath81 , where @xmath116 is the sphere - equivalent roche lobe radius .
we also show a broadening kernel computed using nextgen intensities , assuming @xmath117 and @xmath118 km s@xmath81 . we have adopted an inclination of @xmath119 and an effective temperature of 3900 k for the k giant ( zola et al .
the two broadening kernels have nearly the same full width at half maximum . however , the implied component masses for the two cases are quite different . assuming @xmath120 , @xmath113 ( the value from eaton et al .
@xcite ) implies @xmath121 , @xmath122 , and @xmath123 , whereas our value of @xmath117 gives @xmath124 , @xmath125 , and @xmath126 .
zola et al .
( @xcite ) find a photometric mass ratio from their light curve solutions of @xmath127 , which in our notation is @xmath128 .
the case of is perhaps an extreme example , but it serves to illustrate the importance of using broadening kernels which account for deviations from linear limb darkening and spherical geometry .
= 8.2 cm = 8.2 cm = 8.2 cm given the relatively large differences in the predicted light curves and broadening kernels that we predict , it is worthwhile to discuss the numerical accuracy of the output light curves .
since the computation of the nextgen models is somewhat time consuming , we can not tabulate models for all effective temperatures and gravities .
we have done extensive testing on the integration and interpolation procedures , and we believe the present table ( steps of 0.5 dex in @xmath18 and 200 k in @xmath3 ) is a reasonable compromise between cpu time and sampling accuracy .
there are various ways to see how `` smooth '' the filter - integrated intensities are as a function of the effective temperature and gravity
[ fig8 ] and [ fig9 ] show two such representations . in fig.[fig8 ] we show as a function of the gravity the integrated @xmath48-band intensity for four effective temperatures ( 4000 , 4400 , 5000 , and 5600 k ) at six different emergent angles ( @xmath129 , 0.10 , 0.15 , 0.20 , 0.25 , and 0.30 ) . in general , these curves are smooth , and a linear interpolation of the intensity in the @xmath18 direction gives reasonable accuracy .
[ fig9 ] shows a similar plot , but with the intensities as a function of the effective temperature . as before ,
the curves are quite smooth and a linear interpolation in @xmath3 gives reasonable results .
there is some irregular behaviour in the curves for @xmath129 and @xmath130 . in practice , however , the contribution to the overall integrated light curve from angles less than 0.1 is small since the intensities are weighted by the value of @xmath32 .
another way to test the accuracy of the interpolation is to leave specific models out of the table , compute a light curve , and compare it to the light curve computed using the full intensity table .
[ fig10 ] shows results of this exercise for the star described in the caption of fig .
we computed `` regular '' light curves where we used all models in the intensity table and `` cut '' light curves where the @xmath49 , @xmath50 model was excluded from the table .
thus , at @xmath49 , the local step size in @xmath18 was 1 dex , rather than 0.5 dex . to compare the light curves
, we define the `` difference curve '' as @xmath131 , where @xmath132 is the integrated light computed using the full intensity table and @xmath133 is the integrated light computed using the intensity table with the @xmath49 , @xmath50 model removed .
we see that the systematic difference between the two curves is usually about 0.3 per cent and is at most 0.75 per cent in @xmath47 .
the corresponding values for @xmath48 and @xmath52 are even smaller .
we also computed light curves using two different grids of surface elements : the `` standard '' grid with @xmath134 and @xmath135 and the `` fine '' grid with @xmath136 and @xmath137 . using the fine grid has little effect on
the difference curves : the curves are smoother but the overall shapes are the same .
the lowermost curve in fig . [ fig10 ] compares the regular @xmath47 light curve computed using the standard and fine grids .
the two curves are the same to within 0.05 per cent ( i.e. less than about 0.5 millimag ) .
we conclude that for most situations our regular intensity table and the standard grid are adequate .
there is nothing in our integration and interpolation techniques that limits us to @xmath138 per cent accuracy .
since our interpolation scheme does not have a fixed step size in the temperature or gravity we can easily add a few models with the appropriate temperatures and gravities for special cases where the data demand the highest accuracy .
other light curve codes also suffer systematic errors on the few per cent level .
for example , most versions of the w - d code use a single limb darkening law for the entire star .
since the temperature and gravity are not constant over the surface of a tidally distorted star , small systematic errors can be present when a limb darkening law for a single temperature and gravity are used .
van hamme & wilson ( @xcite ) have discussed this point in more detail .
in this paper we have presented a way to include specific intensities from detailed model atmosphere computations into a light curve synthesis code .
we have shown that using the model atmosphere intensities directly is almost required for cool giants since the limb darkening behaviour for a cool low - gravity atmosphere is radically different than simple one- or two - parameter limb darkening laws ( fig .
[ fig1 ] ) .
this departure from the near - linear limb darkening law is a consequence of the spherical geometry used in the computation of the nextgen models .
other workers have computed spherical model atmospheres for cool giants before ( e.g.scholz & tsuji @xcite ; scholz & takeda @xcite ; plez et al .
@xcite ; jrgensen et al .
@xcite ) , and the strongly nonlinear limb darkening behaviour has previously been noted by scholz & takeda ( @xcite ) , and more recently by hofmann & scholz ( @xcite ) .
we have shown that this strongly nonlinear limb darkening has a large effect on the light curve amplitude .
in general , the light curves computed using the nextgen intensities have larger amplitudes than the light curves computed using the same geometry but with black body intensities and a one- or two - parameter limb darkening law .
also , the strongly nonlinear limb darkening has an effect on rotational broadening kernels in that for spherical stars the kernel computed using the nextgen intensities will be narrower than a numerical or an analytic kernel that has a linear limb darkening law .
if the star fills a substantial fraction of its roche lobe , then the broadening kernel will also be different than the analytic one owing to the nonspherical shape of the star .
thus we basically have two different types of models with which to fit close binary light curves : our elc code with the nextgen intensities , and codes that use either black body intensities or plane - parallel model atmosphere intensities and near - linear limb darkening ( e.g. w - d , elc in black body mode , elc with the kurucz table , etc . ) .
it should be possible to test which class of models provides the better description of the observational data available for binaries with cool giants . in this regard , we have shown that observations in several bandpasses are useful in discriminating between the models .
for example , figs .
[ fig4 ] and [ rzscuti ] show how the difference in the depths of the minima between the nextgen and kurucz light curves depends on the bandpass , while fig .
[ fitfig ] shows that black body light curves provide a poor simultaneous fit to nextgen @xmath47 and @xmath70 light curves of a binary like .
there is no shortage of potentially good binary stars to study .
for example , , as we have already noted , has @xmath47 , @xmath48 , @xmath70 , and @xmath52 light curves available ( although the light curves are perhaps somewhat noisy ) .
is an eclipsing binary consisting of a hot @xmath139b star and a k supergiant .
the orbital period is 262 days .
reasonably good light curves in the bluer filters exist ( e.g. knee et al .
@xcite ; olson @xcite ) , and additional observations in @xmath70 , @xmath71 , and @xmath52 would be extremely valuable .
kruszewski & semeniuk ( @xcite ) present a catalog of poorly studied eclipsing binaries with good parallaxes measured by the _ hipparcos _ mission .
many of these binaries have long periods ( more than 10 days ) and contain evolved cool components .
needless to say , we encourage observers to systematically study these binaries in multiple colours . with a binary of known distance , one has the added challenge of obtaining the correct integrated flux .
our interest should not be confined to binaries with a supergiant component .
for example , there appears to be a small class of `` long period '' algol type binaries ( periods between about 10 and 20 days ) where the mass losing cool star appears to be slightly underfilling its roche lobe . some examples are ( olson & etzel @xcite ) , ( olson & etzel @xcite ) , and ( etzel & olson @xcite ) .
the cool stars in these three systems have surface gravities near @xmath140 , so the sphericity effects are not as pronounced as they are in systems like . on the other hand ,
all three of these binaries have excellent five - colour ( @xmath48(kron)@xmath141 ) photometry , and and are totally eclipsing and are double - lined .
olson and etzel report that in each case , the cool star seems to underfill its roche lobe by about 10 per cent .
it would be worthwhile to re - evaluate the light curve solutions for these three stars using elc and the nextgen intensities to see if the sphericity effects can account for the apparent slight underfilling of the roche lobes by the cool stars .
we thank marten van kerkwijk , frank verbunt , and robert e. wilson for helpful discussions .
jao gratefully acknowledges the hospitality of the institute of astronomy in cambridge .
phh s work was supported in part by the cnrs , insu and by nsf grant ast-9720704 , nasa atp grant nag 5 - 8425 and ltsa grant nag 5 - 3619 , as well as nasa / jpl grant 961582 to the university of georgia .
this work was supported in part by the ple scientifique de modlisation numrique at ens - lyon .
some of the calculations presented in this paper were performed on the ibm sp and sgi origin 2000 of the uga ucns , on the ibm `` blue horizon '' of the san diego supercomputer center ( sdsc ) , with support from the national science foundation , and on the cray t3e and ibm sp of the nersc with support from the doe .
we thank all these institutions for a generous allocation of computer time .
yorham avni was interested in , among other things , mass determinations of the compact objects in high mass x - ray binaries ( e.g. , , etc . ) . during the course of his research he wrote a fortran code to compute the ellipsoidal light curve of a single ( usually ) roche - lobe filling star ( avni & bahcall @xcite ; avni @xcite ) .
avni passed on this code to jeff mcclintock and ron remillard shortly before his death in march of 1988 ( see mcclintock & remillard @xcite ) .
the code was passed on to jao from ron remillard sometime in 1994 when jao was a graduate student at yale . during the spring of 1995
it became clear that the black hole binary was an eclipsing system and that the avni code in its original form was not adequate .
the code was substantially modified by jao in the summer of 1995 to include light from an accretion disk and to account for eclipses ( orosz & bailyn @xcite ) .
the code described in orosz & bailyn ( @xcite ) is somewhat unwieldy and it was becoming more and more difficult to read and modify .
therefore a more general and more modular code was developed by jao .
although we have used much of avni s notation , and we have followed his basic method of setting up the roche geometry and integrating the observable flux , most of the code is new . owing to space limitations
, we can not give below each equation used .
rather , we give below some details of the parts of the code which have been substantially revised with respect to the earlier versions ( there are numerous other papers and texts which go into varying amounts of detail , wilson & devinney @xcite ; wilson @xcite ; avni & bahcall @xcite ; avni @xcite ; linnell @xcite ; wilson @xcite ; orosz & bailyn @xcite ) .
we assume surface of the star is an equipotential surface of the following potential , which includes the gravitational , centrifugal , and coriolis forces ( avni & bahcall @xcite ) : @xmath142 \label{pot}\ ] ] where @xmath143 is the mass of the star under consideration , @xmath144 is separation between two stars , @xmath145 is the mass ratio , @xmath146 is the star s rotational angular velocity , @xmath147 is the orbital angular velocity , @xmath148 and @xmath149 are the distances to the stellar centres in units of @xmath144 , and @xmath11 and @xmath102 are normalised coordinates centred at star 1 . for a binary with a given mass ratio @xmath150 , rotational angular velocity @xmath146 and orbital angular velocity @xmath147 , there is a critical value of the potential @xmath151 where the star exactly fills its limiting lobe .
stars that are smaller then their limiting lobes will have @xmath152 . to fully define the surface of the star ,
the user specifies @xmath150 , @xmath153 , and the `` filling factor '' @xmath154 , where @xmath155 is the @xmath11-coordinate of the `` nose '' of the star and @xmath156 is the @xmath11-coordinate of the @xmath107 point . in our case ,
the filling factor @xmath157 is exactly 1 for roche lobe - filling stars and less than 1 for detached stars .
situations with @xmath158 ( the contact binaries ) are currently not allowed . when @xmath159 , the program computes @xmath155 from @xmath160 and then computes @xmath161 , which is the adopted potential for the star .
once the surface of the star is defined , a grid of surface elements is made using a polar coordinate system with @xmath162 latitude rows equally spaced in the angle @xmath163 and @xmath164 longitude points per latitude row equally spaced in the angle @xmath165 .
( earlier versions of the code used a cylindrical coordinate system with @xmath162 rings equally spaced along the line of centres , running from the @xmath107 point to the back of the star and @xmath164 surface points per ring equally spaced in angle . )
it is convenient to use an internal rectangular coordinate system centred on star 1 where the @xmath11 axis points to the centre of the other object and the @xmath166 axis is parallel to the direction of @xmath146 .
the value of @xmath167 and its derivatives are computed for each element , and from these quantities the local gravities @xmath168 and the surface normal vectors follow .
the temperature of the secondary star was defined in the avni code by its polar temperature @xmath169 . given @xmath169 , the temperatures of the other surface elements followed from the well - known von zeipel relation : @xmath170^{\beta}.\ ] ] the exponent @xmath171 is 0.25 for stars with a radiative atmosphere ( von zeipel @xcite ) and 0.08 for stars with convective envelopes ( lucy @xcite ) .
wilson ( @xcite ) has pointed out that the polar temperature and the _ mean _ temperature of a distorted star will be different . in most cases , of course , one measures the mean temperature via the spectral type or colour index .
therefore , following wilson ( @xcite ) , we now have as input the mean ( or equivalently effective ) temperature of the star , denoted by @xmath3 .
the effective temperature is computed from the bolometric luminosity @xmath172 where @xmath173 is the surface area .
@xmath169 is then given by ( wilson 1979 ) @xmath174^{1/4}\ ] ] where @xmath175 and @xmath176 is an element of surface area ( eqs .
( a8 ) , ( a9 ) , and ( a10 ) of orosz & bailyn @xcite ) .
adding a second star to the code is relatively simple .
we `` flip '' the mass ratio ( define @xmath177 ) , and solve for the potential and its gradients using the same subroutines as for the first star .
when integrating the observed flux , we add @xmath106 to the phase and use the same subroutines as for the first star .
one complication occurs when the second star is not in synchronous rotation ( wilson @xcite ) . in this case
, the @xmath11 derivative of the potential that is used in the detailed reflection routine ( see below ) has a different form .
thus the subroutine that returns the potential gradients returns two sets of @xmath11 derivatives and the appropriate one is used for the detailed reflection .
stars in a close binary can heat each other , and this mutual heating leads to easily observed consequences .
wilson ( @xcite ) divides the `` reflection '' theory into four main parts : the geometric aspect , the bolometric energy exchange , the intensity from an irradiated stellar atmosphere , and the effect on the envelope structure . in this paper wilson gives a complete description of his `` detailed reflection '' scheme which treats the first two parts of the theory essentially exactly .
we have fully implemented wilson s treatment of the reflection effect , which is a big improvement over the scheme used by orosz & bailyn ( @xcite ) .
there are two points about this scheme that are noteworthy : ( i ) wilson s scheme makes use of _ bolometric _ limb darkening approximations .
we have seen that the limb darkening in cool giants is not well parameterised by the common limb darkening laws .
fortunately , in most practical situations , a relatively hot star with a high surface gravity ( @xmath178 ) irradiates a much cooler star . the irradiation of the hot star by the cool star can be neglected ( in which case we do nt care about the details of the cool star limb darkening ) , and the well - known limb darkening laws apply nicely to the hot , high - gravity star .
( ii ) wilson introduced the use of the @xmath179-function in the reflection scheme to allow for multiple reflection . for each element
, the @xmath179-function is defined by @xmath180 where @xmath181 is the ratio of the total irradiating flux from star 2 seen at the local surface element on star 1 and @xmath182 is the reverse .
the new temperatures of the irradiated surface elements are then @xmath183 it is usually assumed that the specific intensity of an irradiated surface element is the same as the intensity of an unheated surface element with the same temperature . at some point the irradiation become intense enough that this assumption must break down .
alencar & vaz ( @xcite ) have computed some irradiated atmosphere models and presented limb darkening coefficients for use in light curve synthesis codes .
the use of these coefficients is perhaps somewhat limited as the widely available versions of the popular w - d code have only a single limb darkening law for the entire star ( which of course includes the unheated back hemisphere ) .
one of us ( phh ) is in the process of computing irradiated atmospheres with the phoenix code .
it should be relatively simple to include the irradiated atmospheres into the elc code , at least for specific binaries .
there exist many eclipsing close binaries where a hot subdwarf o / b star irradiates its g - m main sequence companion ( e.g. hilditch et al .
such systems can provide strong tests of irradiated atmospheres .
the accretion disk used in the orosz & bailyn ( @xcite ) code was a flattened cylinder . to make the disk perhaps more realistic ,
we have modified the disk so that its thickness as a function of the radius is proportional to the radius .
the disk , if present , is always around star 2 .
star 2 does not necessarily have to be present ( as in an x - ray binary ) .
the disk is described by five basic parameters : @xmath184 , the radius of the outer edge of the disk in terms of the effective roche lobe radius of star 2 ; @xmath185 , the inner radius of the disk in the same units as the filling factor of star 2 ( @xmath186 ) ; @xmath187 , the opening angle of the disk rim above the plane ; @xmath188 , the temperature of the _ inner _ disk ( in the orosz & bailyn ( @xcite ) code the temperature of the _ outer _ rim was specified ) ; and @xmath189 , the power - law exponent on the temperature profile of the disk : @xmath190 for a steady - state accretion disk , @xmath191 ( pringle @xcite ) . for a disk heated by a central source
, the exponent @xmath189 can take on a range of values ( @xmath192 ( e.g.friedjung @xcite ; vrtilek et al .
@xcite ; bell @xcite ) ) .
since the surface of the disk is flared , each element on the face of the disk will have different `` projection factors '' @xmath193 to the line of sight . here
@xmath193 is equivalent to the angle @xmath32 discussed above .
we define a polar coordinate system centred at the centre of the disk .
the angle @xmath163 is measured from the @xmath11-axis in the direction of the positive @xmath102-axis . for a given the orbital phase @xmath165 , @xmath194
is given by @xmath195 if @xmath196 , the point is not visible .
there are several geometrical details which we must account for when the disk is flared and/or when star 2 has a relatively large radius .
first , for cases when the inclination @xmath197 is within @xmath171 degrees of 90 , parts of the disk face that have @xmath198 will be below the rim as seen by a distant observer . to account for these hidden points
, we define the `` horizon '' of the top rim of the disk . in this discussion
the horizon of an object is the outline of the object in sky coordinates .
a point on the disk face is visible if its sky coordinates are _ inside _ the top horizon .
star 2 can block parts of the disk if its radius is relatively large . since we currently require the inner radius of the disk to be equal to the radius of star 2 ( if present ) , some or all of the `` bottom '' part of star 2 may be hidden by the disk . again
, the shadowing of the disk by star 2 and the lower part of star 2 by the disk is easily accounted for by defining the suitable horizons .
finally , a disk with a large radius can inhibit the mutual irradiation of the two stars since the `` top '' of star 1 can not `` see '' the `` bottom '' of star 2 , and vice - versa .
if a disk is present then inside the detailed reflection subroutine each line of sight between points on star 1 and star 2 is checked to see if it passes through the disk .
currently we assume that each surface element on the disk has a specific intensity that is the same as a normal stellar atmosphere . following pringle et al .
( @xcite ) we use the model with the largest gravity for each effective temperature .
of course , much more detailed model atmospheres specific to accretion disks are available ( for example the grid of models for accretion disks in cataclysmic variables presented by wade & hubeny ( @xcite ) ) , and in principle a separate intensity table for accretion disks can almost trivially be added to elc .
indeed , linnell & hubeny ( @xcite ) create light curves for binaries with disks by first computing detailed spectra for the disk using the hubeny codes .
however , the hubeny code tlusdisk is best suited for atmospheres hotter than about 10,000 k , and as such can not really be used to model the cool outer parts of many disks . as it is now ,
our treatment of the accretion disk is perhaps the most appropriate for systems where the disk is optically faint and its main effect on the light curves is geometrical ( i.e.it eclipses the bright mass - losing star ) .
examples of such systems are ( zola @xcite ) , ( orosz & bailyn @xcite ) , and ( benson et al .
@xcite ) . in many algol type binaries
there is good evidence for a fainter third star that is gravitationally bound .
for example , the @xmath88 residuals of the eclipse timings of are periodic and can be explained by the presence of a third body in a 5.8 year eccentric orbit about the inner binary ( ogloza et al .
@xcite ) . in such triple systems ,
the `` third light '' dilutes the observed amplitudes of the light curves from the binary , provided of course that the third star is sufficiently bright .
we have a trivial way to self - consistently add third light to light curves in different bandpasses . in elc
, the user specifies three parameters for the third light : the temperature of the third star , its gravity , and its surface area relative to the surface area of star 1 .
the code interpolates the filter - integrated intensities for the third star from the table , scales appropriately based on the surface area ratio , and adds the light to each light curve .
the integration of the observed flux from a single star is straightforward and is sufficiently accurate for a reasonably small number of surface elements .
however , quantisation errors can become noticeable in the light curve of a star going through an eclipse .
in many cases ( e.g. grazing eclipses , no reflection effect ) the number of surface elements can be modestly increased so that a smooth light curve can be obtained without a large increase in the cpu time required to compute the model . in other cases ( e.g. deep
eclipses and several iterations of detailed reflection ) the number of surface elements needed to get smooth light curves becomes so large that the required cpu time becomes excessive .
thus elc has two features which prove to be quite effective in greatly reducing the numerical noise of light curves though eclipse .
the horizon of the eclipsing star ( the one `` in front '' ) is defined to be a collection of points on the star which have @xmath24 . in previous versions of the code ,
the program would step though the surface grid in the `` @xmath199 '' direction and record which surface elements were last visible ( i.e. the last point with @xmath200 ) .
the resulting collection of sky coordinates of these surface elements would then define the horizon of the star .
there is a systematic error introduced when the number of surface elements is small since the numerical horizon of the star will be slightly smaller than the actual horizon . in the current version of the code ,
the program steps along each latitude row and records the @xmath165-coordinate of the last visible point @xmath201 and the @xmath165-coordinate of the first point hidden below the horizon @xmath202 .
a simple bisection procedure is used to find the @xmath165-coordinate ( with a given latitude @xmath203 ) where @xmath24 .
fifteen iterations of this bisection procedure are enough to find @xmath204 to better than @xmath205 radians when @xmath206 ( i.e. 24 longitude points ) .
the corresponding angles @xmath32 are all @xmath207 .
the @xmath11 and @xmath102 sky coordinates of the point with the surface coordinates @xmath208 are then determined .
a similar procedure is done where the program steps through the @xmath163 angle for each longitude @xmath209 and @xmath210 is found from @xmath211 and @xmath212 using bisection . after the list of points on the star with @xmath24 is generated , the @xmath11 and @xmath102 rectangular coordinates of each point on the sky are converted to a @xmath179,@xmath213 polar coordinate system and sorted in the polar angle @xmath213 .
the sorted array forms a convex polygon on the sky .
if the number of surface elements on the star is relatively small ( @xmath214 and @xmath215 ) , then the actual horizon of the star can have some curvature between adjacent points on the polygon .
since the radius @xmath179 as a function of @xmath213 is always a very smooth function , we use spline interpolation to resample @xmath179 for every @xmath216 in @xmath213 .
the new resampled polygon with 360 points always has enough points so that the horizon of the star essentially has no curvature between adjacent points .
we have done numerous tests and found that the horizons derived using a small number of surface elements with the new routine are always the same as the horizons found by the old routine with a very large number of surface elements ( @xmath217 and @xmath218 ) . to compute the observed flux in a given bandpass from a star at a given phase
, we numerically evaluate the integral in eq .
[ eq2 ] using all of the surface elements with @xmath200 : @xmath219 where @xmath220 and @xmath221 are the angular spacings of the elements in latitude and longitude , respectively , and where @xmath171 is the angle between the surface normal and the radius from the centre of the star . in other words , we perform a simple numerical quadrature along each latitude row where the points are equally spaced in the angle @xmath165 . if there is an eclipsing body in front of the star
whose flux is being evaluated , then each point on the star in back is projected onto the sky and a simple routine is used to see if this point is inside the polygon representing the horizon of the body in front .
if the point in question is eclipsed , its flux contribution is simply left out of the summation .
in general , the horizon of the star in front will not pass exactly between two adjacent points on a given latitude row on the star in back . as a result there will be `` visible '' points
that are actually centred in partially eclipsed surface elements and `` eclipsed '' points that are centred in partially visible surface elements .
if the number of surface elements on the star in back is large enough , the contribution from the fractionally eclipsed surface elements will tend to cancel out .
however , it is much more computationally efficient to use a smaller number of surface elements and make a correction for the fractionally eclipsed pixels . at each latitude row @xmath203 on the star in back ,
the program determines the @xmath165-coordinate of the last point visible before the horizon of the eclipsing body @xmath201 and the @xmath165-coordinate of the first point hidden behind the horizon of the eclipsing body @xmath202 .
another bisection procedure is used to determine the @xmath165-coordinate on the star in back where the horizon of the eclipsing body intersects the @xmath203 latitude row @xmath204 . if @xmath222 , then the last point visible before the horizon is centred in a partially eclipsed surface element and a negative correction is added to the flux summation : @xmath223f_{\rm vis}\ ] ] where @xmath224 is flux from the last visible point , i.e.@xmath225 likewise , if @xmath226 , then the first point hidden behind the horizon is centred in a partially visible surface element and a positive correction is added to the flux summation : @xmath227f_{\rm hid}\ ] ] where @xmath228 is the flux the eclipsed point .
if there is an annular eclipse , then this procedure for finding @xmath204 may have to be done twice for a given latitude row .
we have tested this simple interpolation procedure quite extensively and have found it to be quite effective .
smooth light curves through eclipse can be obtained for grid sizes as small as @xmath229 and @xmath206 .
we have done extensive testing of the elc code in its black body mode against the w - d code .
unfortunately , there is some confusion over the notation of some of the input parameters between the two codes . in particular , @xmath230 in elc is the ratio of the star s angular velocity to the orbital velocity , whereas @xmath230 in w - d refers to the potential .
the w - d @xmath230-potentials essentially define the shapes of the stars .
perhaps to conserve parity , @xmath157 in elc is the `` filling factor '' which has the same function as the @xmath230-potential in w - d ( i.e. it defines the shape of the star ) , whereas in w - d , @xmath231 is the ratio of the star s angular velocity to the orbital velocity ( elc s @xmath230 ) .
the phase convention between the two codes is different . in elc ,
star 1 is in front of star 2 at phase 0.0 , whereas in w - d star 1 is _ behind _ star 2 at phase 0.0 . to facilitate comparisons between elc and w - d
, we do two things .
first , since the internal form of the potential is the same for both codes , elc prints out the program values of the potentials which then are the input @xmath230 potentials for w - d .
second , we add a phase shift of 0.5 to the w - d light curves ( the pshift input parameter ) .
we computed various model binary light curves using the two codes and compared them by normalising the light curves at phase 0.25 ( quadrature ) .
we of course used exactly the same effective wavelengths , limb darkening laws and coefficients , and reflection schemes for both codes . in most cases
, the light curves agreed to better than 0.1 per cent ( better than 1 millimag ) .
if we compared the light curves by adjusting the normalisation of one of the curves to match the other , then the largest deviations became even smaller .
finally , we compared computed geometric quantities between the two codes ( i.e. the `` polar '' radius , `` point '' radius , etc . ) .
the agreement was essentially exact .
we also computed sphere - equivalent roche lobe radii and compared our results with eggleton s ( @xcite ) results and likewise found nearly exact agreement .
alencar s. h. p. , vaz l. p. r. , 1999 , a&as 135 , 555 al - naimiy h. m. , 1978 , ap&ss 53 , 181 avni y. , 1978 , in giacconi r. , ruffini r. , eds . ,
physics and astrophysics of neutron stars and black holes .
north - holland , amsterdam , p. 42 avni y. , bahcall j. n. , 1975 , apj 197 , 675 bailey j. , 1975 , j. brit .
85 , 217 belczyski k. , mikolajewska j. , 1998 , mnras 296 , 77 bell k. r. , 1999 , apj 526 , 411 bell r. a. , gustafsson b. , 1989 , mnras 236 , 653 benson p. , dullighan a. , bonanos a. , mcleod k. k. , kenyon s. j. , 2000 , aj 119 , 890 bessell m. s. , 1990 , pasp 102 , 1181 bessell m. s. , brett j. m. , 1988 , pasp 100 , 1134 callanan p. j. , van paradijs j. , rengelink r. , 1995 , apj 439 , 928 carbon d. f. , gingerich o. j. , 1969 , in gingerich o. j. , ed . , theory and observations of normal stellar atmospheres . mit press , cambridge ( mass . ) , p 377 claret a. , 1998 , a&a 335 , 647 daz - cordovs j. , gimnez a. , 1992 , a&a 259 , 227 eaton j. a. , hall d. s. , honeycutt r. k. , 1991 , apj 376 , 289 eggleton p. p. , 1983 , apj 268 , 368 etzel p. b. , olson e. c. , 1995 , aj 110 , 1809 faulkner j. , flannery b. p. , warner b. , 1972 , apj 175 , l79 fekel f. c. , joyce r. r. , hinkle k. h. , skrutskie m. f. , 2000 , aj 119 , 1375 friedjung m. , 1985 , a&a 146 , 366 gray d. f. , 1992 , observations and analysis of stellar photospheres , cambridge university press , cambridge haswell c. a. , robinson e. l. , horne k. , stiening r. f. , abbott t. m. c. , 1993 , apj 411 , 802 hauschildt p. h. , baron e. , allard f. , 1997 , apj 483 , 390 hauschildt p. h. , allard f. , ferguson j. , baron e. , alexander d. r. , 1999 , apj 525 871 heemskerk m. h. m. , van paradijs j. , 1989 , a&a 223 , 154 hilditch r. w. , harries t. j. , hill g. , 1996 , mnras 279 , 1380 hill g. , hutchings j. b. , 1970 , apj 162 , 265 hofmann k .- h . , scholz m. , 1998 , a&a 335 , 637 hric l. , petrk k. , urban z. , niarchos p. , anupama g. c. , 1998 , a&a 339 , 449 hubeny i. , 1988 , comput .
52 , 103 hubeny i. , 1991 , in bertout c. , et al .
, eds . , structure and emission properties of accretion disks .
fong & sons , singapore , p. 227 hubeny i. , lanz t. , jeffrey c. s. , 1994 , in jeffrey c. s. , ed . ,
newsletter on analysis of astronomical spectra .
20 , st andrews univ .
jrgensen u. g. , johnson h. r. , nordlund a. , 1992 , a&a 261 , 263 kenyon s. j. , garcia m. r. , 1986 , aj 91 , 125 klinglesmith d. a. , sobieski s. , 1970 , aj 75 , 175 knee l. b. g. , scarfe c. d. , mayor m. , baldwin b. w. , meatheringha s. j. , 1986 , a&a 168 , 72 kopal z. , 1959 , close binary systems .
john wiley & sons , new york kraft r. p. , 1958 , apj 127 , 620 kruszewski a. , semeniuk i. , 1999 , acta ast .
49 , 561 kurucz r. l. , 1979 , apjs 40 , 1 lacy c. h. s. , torres g. , claret a. , et al . ,
2000 , apj 119 , 1389 lines h. c. , lines r. d. , mcfaul t. g. , 1988 , aj 95 , 1505 linnell a. p. , 1984 , apjs 54 , 17 linnell a. p. , hubeny i. , 1994 , apj 434 , 738 linnell a. p. , hubeny i. , 1996 , apj 471 , 958 linnell a. p. , hubeny i. , harmanec p. , 1998a
, apj 509 , 379 linnell a. p. , etzel p. b. , hubeny i. , olson e. c. , 1998b , apj 494 , 773 lucy l. b. , 1967 , zs . f. ap . 65 , 89 marsh t. r. , robinson e. l. , wood j. h. , 1994 , mnras 266 137 mcclintock j. e. , remillard r. a. , 1990 , apj 350 , 386 milone e. f. , stagg c. r. , kurucz r. l. , 1992 , apjs 79 , 123 mochejska b. j. , kaluzny j. , stanek k. z. , krockenberger m. , sasselov d. d. , 1998 , aj 118 , 2211 mochnacki s. w. , doughty n. a. , 1972 , mnras 156 , 51 ogloza w. , zola s. , tremko j. , kreiner j. m. , 1998 , a&a 340 , 81 olson e. c. , 1993 , aj 106 , 754 olson e. c. , etzel p. b. , 1993a , aj 106 , 759 olson e. c. , etzel p. b. , 1993b , aj 106 , 1162 olson e. c. , etzel p. b. , 1994 , aj 108 , 262 orosz j. a. , bailyn c. d. , 1997 , apj 477 , 876 ( erratum apj 482 , 1086 ) plez b. , brett j. m. , nordlund a. , 1992 , a&a 256 , 551 pols o. r. , tout c. a. , schrder k .- p . , eggleton p. p. , manners j. , 1997 , mnras 289 , 869 popper d. m. , 1977 , pasp 89 , 315 pringle j. e. , 1981 , ara&a 19 , 137 pringle j. e. , verbunt f. , wade r. a. , 1986 , mnras 221 , 169 ribas i. , guinan e. f. , fitzpatrick e. l. , et al . , 2000 ,
apj 528 , 692 sanwal d. , robinson e. l. , zhang e .- h . , et al . , 1996 , apj 460 , 437 scholz m. , takeda y. , 1987 , a&a 186 , 200 scholz m. , tsuji t. , 1984 , a&a 130 , 11 schrder k .-
p . , pols o. r. , eggleton p. p. , 1997
, mnras 285 696 selvelli p. l. , cassatella a. , gilmozzi r. , 1992 , apj 393 , 289 shahbaz t. , 1998 , mnras 298 , 153 shahbaz t. , ringwald f. a. , bunn j. c. , et al . , 1994 , mnras 271 , l10 shahbaz t. , somers m. , yudin b. , naylor t. , 1997 , mnras 288 , 1027 tjemkes s. a. , zuiderwijk e. j. , van paradijs j. , 1986 , a&a 154 , 77 van der hooft f. , heemskerk m. h. m. , alberts f. , van paradijs j. , 1998 , a&a 329 , 538 van hamme w. , 1993 , aj 106 , 2096 van hamme w. , wilson r. e. , 1990 , aj 100 , 1981 van hamme w. , wilson r. e. , 1994 , bull .
american ast .
185 , 85.14 vrtilek s. , raymond j. c. , garcia m. r. , et al . , 1990 ,
a&a 235 , 162 von zeipel h. , 1924 , mnras 84 , 665 wade r. a. , rucinski s. m. , 1985 , a&as 60 , 471 wade r. a. , horne k. , 1988 , apj 324 , 411 wade r. a. , hubeny i. , 1998 , apj 509 , 350 wilson r. e. , 1979 , apj 234 , 1054 wilson r. e. , 1990 , apj 356 , 613 wilson r. e. , 1994 , pasp 106 , 921 wilson r. e. , devinney e. j. , 1971 , apj 166 , 605 wilson r. e. , sofia s. , 1976 , apj 203 , 182 yudin b. , munari u. , 1993 , a&a 270 , 165 zhang e. h. , robinson e. l. , nather r. e. , 1986 , apj 305 , 740 zola s. , 1996 , a&a 308 , 785 zola s. , hall d. s. , henry g. w. , 1994 , a&a 285 , 531 | we have written a light curve synthesis code that makes direct use of model atmosphere specific intensities , in particular the nextgen model atmosphere grid for cool giants ( @xmath0 k and @xmath1 , hauschildt et al .
@xcite ) .
we point out that these models ( computed using spherical geometry ) predict a limb darkening behaviour that deviates significantly from a simple linear or two - parameter law ( there is less intensity at the limb of the star ) .
the presence of a significantly nonlinear limb darkening law has two main consequences .
first , the ellipsoidal light curve computed for a tidally distorted giant using the nextgen intensities is in general different from the light curve computed using the same geometry but with the black body approximation and a one- or two - parameter limb darkening law . in most cases the light curves computed with the nextgen intensities have deeper minima than their black body counterparts .
thus the light curve solutions for binaries with a giant component obtained with models with near linear limb darkening ( either black body or plane - parallel model atmosphere intensities ) are biased .
observations over a wide wavelength range ( i.e. both the optical and infrared ) are particularly useful in discriminating between models with nearly linear limb darkening and the nextgen models .
second , we show that rotational broadening kernels for roche lobe filling ( or nearly filling ) giants can be significantly different from analytic kernels due to a combination of the nonspherical shape of the star and the radical departure from a simple limb darkening law . as a result , geometrical information inferred from @xmath2 measurements of cool giants in binary systems are likewise biased . |
recently , there has been a resurgence of interest in extreme value questions for random walks and its variants , particularly in the physics literature @xcite . this is motivated to a large extent by the observation that extreme value statistics ( evs ) plays a crucial role in the physics of complex and disordered systems @xcite . although the evs of independent and identically distributed ( i.i.d .
) random variables is very well understood , much less is known for strongly correlated variables , which is the case frequently encountered in physics . from that point of view , random walks offer a very useful framework where the effects of correlations on evs can be characterized analytically .
the statistics of the maximal displacement of a random walk ( rw ) after a given number of steps @xmath14 is now quite well understood , thanks for instance to the pollaczeck - spitzer formula @xcite , from which precise asymptotic estimates in the large @xmath14 limit can be obtained @xcite .
the full order statistics , i.e. the statistics of the @xmath15 maximum for any integer @xmath16 , was investigated in @xcite in the case of ordinary rws with a narrow jump distribution ( characterized by a lvy exponent @xmath17 ) .
of particular interest is the _ gap _ between two consecutive maxima , which makes it possible to characterize the `` crowding '' of the walker positions near their maximum . along this line
we recently obtained a complete description of the joint distribution of the gap and time interval between the first two maxima of rws not only for narrow jump distributions ( @xmath18 ) , but also for lvy flights characterized by a lvy exponent @xmath19 @xcite .
since the characteristic displacement of the walker after @xmath14 time steps grows like @xmath20 , the results of @xcite are of interest for diffusive ( @xmath18 ) and super - diffusive ( @xmath21 ) processes only , leaving out the important case of _ sub - diffusive _ processes . extending the analysis of @xcite to sub - diffusive rws has been one of the motivations for this work .
a common model to describe sub - diffusive processes is the so called continuous time random walk ( ctrw ) . in this model ,
the walker performs a usual random walk where two consecutive jumps are separated by a certain trapping time @xmath1 .
the trapping times are i.i.d .
random variables drawn from a common distribution @xmath4 which has a power law tail @xmath22 with @xmath23 .
this model , initially suggested by scher and montroll in the setting of non - gaussian transport in disordered electronic systems @xcite , has been widely used to describe anomalous sub - diffusive dynamics in various complex systems phenomenologically @xcite .
indeed , for @xmath24 the average trapping time between two successive jumps is infinite and , consequently , ctrws are characterized by an anomalous growth of displacement @xmath25 which is indeed sub - diffusive if @xmath7 ( and super - diffusive for @xmath8 ) .
the case of a finite average trapping time between two successive jumps corresponds to standard , non sub - diffusive , rws .
the fluctuations of the maximal displacement of ctrws for @xmath17 and arbitrary @xmath26 was studied in ref .
@xcite using real space renormalization techniques ( see also @xcite for a feynman - kac approach )
. a precise estimate of the expected value of the maximum of the ctrw was obtained in ref .
@xcite by combining the results of ref .
@xcite with the so called subordination property. however , nothing is known about the order statistics , beyond the first maximum , for such ctrws . in this paper , we provide a complete analytic description of the distribution of the gap and the time interval between the first two maxima of such ctrws for any values of the parameters @xmath3 and @xmath27 , as well as for a finite average trapping time between two successive jumps .
we consider a ctrw starting at the origin , @xmath28 , and evolving according to @xmath29 where @xmath30 denotes the walker position between the @xmath31-th and the @xmath32-th jumps .
the jumps @xmath33 are i.i.d .
random variables distributed following a symmetric , bounded and piecewise continuous distribution @xmath2 the fourier transform of which , @xmath34 , has the small @xmath16 behavior @xmath35 where @xmath3 is the lvy index and @xmath36 is the characteristic length scale of the jumps . for @xmath17 , the variance of the jump distribution @xmath37
is well defined and @xmath38 . on the other hand ,
for @xmath39 , @xmath2 does not possess a well defined second moment because of its heavy tails , @xmath40 ( @xmath41 ) , and the rw ( [ def_rw ] ) is a lvy flight of index @xmath42 .
the time intervals @xmath43 between the @xmath44-th and the @xmath31-th jumps ( @xmath45 ) are i.i.d .
random variables , independent of @xmath33 , with pdf @xmath4 the laplace transform of which , @xmath46 , has the small @xmath47 behavior @xmath48 where @xmath49 and @xmath50 is the characteristic time scale of the jumps . for @xmath51 , the mean time between two successive jumps , @xmath52 , exists and @xmath53 .
for @xmath27 , the mean time between two successive jumps does not exist . as a function of time ,
the walker position is thus given by @xmath54 with @xmath55 and @xmath14 the total number of jumps in the walk .
the solution to ( [ def_ctrw ] ) is readily found to be given by @xmath56 with @xmath30 given by ( [ def_rw ] ) .
the goal of the present paper is to provide an exhaustive discussion of the different behaviors of the joint pdf of the gap @xmath9 and time interval @xmath10 between the first two maxima of the walk that may arise in the large @xmath14 limit , depending on the large argument behavior of @xmath2 and @xmath4 . before entering the details of the calculations ,
it is useful to summarize our main results .
we first show that the joint pdf @xmath57 has a well defined limiting pdf as @xmath58 : @xmath59 where the laplace transform of @xmath60 with respect to ( w.r.t . )
@xmath10 is given in eqs .
( [ eq2.8 ] ) and ( [ eq2.5 ] ) .
then , we perform a detailed analysis of @xmath11 in the plane @xmath61 for @xmath42 and @xmath62 in the whole ranges @xmath3 and @xmath49 , and for the three different main classes identified in @xcite : ( i ) slow , ( ii ) exponentially , and ( iii ) fast decreasing @xmath2 at large @xmath63 . one of the most remarkable result of this study may be the existence of a scaling form for the joint pdf @xmath11 in the case of a fast decreasing jump distribution ( @xmath18 ) , such that @xmath64 for @xmath41 [ see eq .
( [ eq5.14 ] ) below for a more precise definition ] , with @xmath27 .
this class includes the case of a gaussian jump distribution for which @xmath65 .
this is a new result , characteristic of ctrws , without any counterpart in the discrete time rw setting considered in @xcite .
more specifically , we show that there is a scaling regime @xmath66 with fixed @xmath67 , in which @xmath11 takes the scaling form @xmath68 with the asymptotic behaviors @xmath69 where the amplitudes @xmath70 and @xmath71 are given in eqs .
( [ eq5.26 ] ) and ( [ eq5.27 ] ) , respectively .
physically , the switch from the first to the second behavior ( [ scaling1b ] ) around @xmath72 corresponds to the cross - over from a ` concentration ' or ` one - step ' regime ( for @xmath73 ) where the walker get stuck for a long time @xmath10 at the second maximum and then jumps directly to the first maximum , to a ` many - steps ' regime ( for @xmath74 ) where she / he travels a long walk of total duration @xmath10 ( with many steps ) between the second and the first maxima . for @xmath51 , there is no scaling form and ( [ scaling1a ] ) is replaced with the uniform expression ( [ eq5.30 ] ) .
an other important result is how the scaling form derived in @xcite for a discrete time lvy flight ( see sec .
4.1 and appendix d in @xcite ) is affected when one switches to a ctrw .
lvy flights have type ( i ) jump distributions since , for @xmath39 , @xmath2 necessarily has an algebraic , slow decreasing , tail . in this case
we show that in the scaling regime @xmath66 with fixed @xmath75 , the joint pdf @xmath11 takes the scaling form @xmath76 with the asymptotic behaviors @xmath77 and @xmath78 where @xmath79 in the first line of ( [ scaling2b ] ) denotes the integer part of @xmath62 and the last two lines are for @xmath27 only .
it can be checked that for @xmath51 , eqs .
( [ scaling2a ] ) , ( [ scaling2b ] ) , and ( [ scaling2c ] ) reduce to eqs .
( 87 ) , ( 88 ) , and ( 90 ) of @xcite in which the number of jumps between the first two maxima , @xmath80 , is merely replaced with @xmath81 , as expected from simple law of large number arguments giving @xmath82 as @xmath83 .
our results provide thus a non trivial extension of the ones in @xcite to the case @xmath27 in which the mean time between two successive jumps does not exist and law of large number arguments can not be used .
finally , by integration of the joint pdf @xmath11 over @xmath9 , one finds that the marginal distribution @xmath84 displays a power law tail with logarithmic corrections and an exponent depending only on @xmath42 and @xmath62 .
namely , for @xmath51 one finds @xmath85 where the amplitudes @xmath86 , @xmath87 and @xmath88 are given in eq .
( [ eq7.1a ] ) , and for @xmath27 one has @xmath89 where the amplitudes @xmath90 , @xmath91 and @xmath92 are given in eq .
( [ eq7.6a ] ) .
again , it can be checked that eq .
( [ eq1:result_p_of_t ] ) is nothing but eq .
( 112 ) in @xcite with @xmath93 , as expected ( _ cf . _
the end of the preceding paragraph ) . note that for @xmath94 and @xmath51 , one has @xmath95 whatever the distributions @xmath2 and @xmath4 possessing a second and a first moment , respectively .
the third lines of ( [ eq1:result_p_of_t ] ) and ( [ eq2:result_p_of_t ] ) reveal an interesting freezing phenomenon , as a function of @xmath42 , of the large @xmath10 behavior of @xmath84 as @xmath42 decreases past the value @xmath96 .
it follows in particular from ( [ eq1:result_p_of_t ] ) and ( [ eq2:result_p_of_t ] ) that the first moment of @xmath84 is never defined .
this means that , although the typical size of @xmath10 is @xmath97 , its average diverges with the total duration , @xmath0 , of the random walk . for
@xmath49 one can estimate from ( [ eq1:result_p_of_t ] ) and ( [ eq2:result_p_of_t ] ) that @xmath98 for @xmath99 , while @xmath100 for @xmath101 and @xmath102 for @xmath103 .
table [ tab1 ] summarizes the functional dependence of the different asymptotic behaviors of @xmath84 at large @xmath10 in the @xmath104 plane . .
functional dependence of the different asymptotic behaviors of @xmath84 at large @xmath10 in the @xmath104 plane .
[ cols="^,^,^,^",options="header " , ] note that the marginal distribution of the gap between the first two maxima , @xmath105 , obtained by integration of the joint pdf @xmath11 over @xmath10 , does not depend on @xmath4 ( hence on @xmath62 ) and is the same as the one already studied in sec . 3 of ref .
@xcite in the discrete time rw setting .
the outline of the paper is as follows . in section [ sec2 ]
we give the laplace transform of @xmath11 with respect to @xmath10 for a free - end random walk and a random bridge in which the walker is conditioned to return to @xmath106 at the end of the walk .
section [ sec3 ] deals with the asymptotic behavior of @xmath11 for large @xmath10 at fixed @xmath9 , from which we derive , in section [ sec4 ] , the asymptotic behavior of @xmath11 for a lvy flight at large @xmath10 then large @xmath9 .
the asymptotic behavior of @xmath11 for large @xmath9 is investigated in section [ sec5 ] for three classes of jump distributions encompassing a wide range of jumps of practical interest .
it is shown in section [ sec5.3 ] that when the jump distribution has a fast decreasing tail , the behavior of @xmath60 when both @xmath9 and @xmath80 are large takes on a scaling form capturing the existence of two complementary regimes , one dominated by walks with only one step between the first two maxima , and the other dominated by walks with many steps between the first two maxima . in section [ sec6 ] we give the scaling form of the asymptotic behavior of @xmath11 for a lvy flight at large @xmath10 and @xmath9 . in section [ sec7 ] we use the results obtained in sections [ sec3 ] and [ sec6 ] to determine the large @xmath10 behavior of @xmath84 and show its freezing when the random walk is a lvy flight with index @xmath107 .
finally , section [ sec8 ] is devoted to the comparison of our analytical results with numerical simulations .
let @xmath57 denote the joint pdf of the gap , @xmath9 , and the time interval , @xmath10 , between the first two maxima of a continuous time random walk after @xmath14 steps .
since @xmath110 we will restrict ourselves to @xmath111 without loss of generality .
let @xmath80 denote the number of jumps between the first to maxima . for @xmath111 one
has @xmath112 where @xmath113 has the laplace transform @xmath114 taking the laplace transform of ( [ eq2.1 ] ) with respect to @xmath10 and using ( [ eq2.2 ] ) one gets @xmath115 where @xmath116 is the generating function of @xmath117 with respect to @xmath118 . the large @xmath14 limit of ( [ eq2.3 ] )
can then be readily obtained from the main result of @xcite according to which @xmath119 exists and is given by @xmath120 with @xmath121 where the functions @xmath122 and @xmath123 are the inverse laplace transforms of @xmath124 and @xmath125 , respectively : @xmath126 with @xmath127 injecting this result into the large @xmath14 limit of ( [ eq2.3 ] ) yields @xmath128 thus , @xmath57 converges to a limiting distribution @xmath11 as @xmath129 whose laplace transform with respect to @xmath10 , @xmath130 , is given by ( [ eq2.8 ] ) .
it was proved in @xcite that the expression ( [ eq2.4 ] ) of @xmath131 actually holds for both free - end random walks and random bridges in which the walker is conditioned to return to @xmath106 at the end of the walk . since , for a given @xmath4
, @xmath132 is entirely determined by @xmath131 through ( [ eq2.8 ] ) , it follows immediately that the expression of @xmath132 , hence the one of the limiting joint pdf @xmath11 , is exactly the same for free - end walks and bridges . therefore , all the results obtained in the following apply to free - end random walks as well as to random bridges without any modification .
sections [ sec3 ] to [ sec6 ] are devoted to the asymptotic behavior of the joint pdf @xmath11 when either @xmath9 or @xmath10 ( or both ) is large . from the results obtained for the large @xmath10 behavior of @xmath11 it will then be possible to derive the large @xmath10 behavior of the marginal distribution @xmath134 , which will be the subject of sec . [ sec7 ] .
first , we consider the limit @xmath83 at fixed @xmath9 .
as we will see , the behavior of @xmath11 in this limit depends on the value of @xmath62 , leading to new , non trivial , asymptotic expressions if @xmath27 . in this case the mean time between two successive jumps , @xmath136 , exists and by a simple law of large number argument , the large @xmath10 behavior of @xmath11 can be expected to be given by @xmath137 , where @xmath93 is large .
let us check this result explicitly by comparing the large @xmath10 behavior of @xmath11 and the large @xmath80 behavior of @xmath60 .
the behavior of @xmath11 at large @xmath10 is determined by the one of @xmath132 at small @xmath47 .
thus , from ( [ eq2.8 ] ) and ( [ def_gamma ] ) with @xmath51 , one has @xmath138 where we have made the change of variable @xmath139 .
similarly , for large @xmath80 the behavior of @xmath60 is determined by the one of @xmath140 near @xmath141 . making the change of variable @xmath142 one gets @xmath143 comparing eqs .
( [ eq3.1 ] ) and ( [ eq3.2 ] ) , one obtains the expected expression @xmath144 finally , eq . ( 65 ) of @xcite together with eq .
( [ eq3.3 ] ) above yield , for all @xmath145 , @xmath146 if @xmath27 the mean time between two successive jumps , @xmath148 , does not exist and the large @xmath10 behavior of @xmath11 is no longer related to @xmath60 by the simple expression ( [ eq3.3 ] ) .
following the same line as for eq .
( [ eq3.1 ] ) with @xmath27 , one now has @xmath149 where we have made the change of variable @xmath139 .
it remains to compute the inverse laplace transform on the right - hand side of ( [ eq3.5 ] ) .
one finds three different regimes depending on the value of @xmath42 .
for @xmath151 the behavior of @xmath140 near @xmath141 is given by , ( see eq . ( 63 ) in @xcite ) , @xmath152 with @xmath153 .
injecting ( [ eq3.6 ] ) into ( [ eq3.5 ] ) yields @xmath154 performing then the inverse laplace transform in ( [ eq3.7 ] ) and using the reflection formula @xmath155 one obtains , for all @xmath156 , @xmath157 note that the exponent of @xmath10 in ( [ eq3.8 ] ) depends on @xmath42 and ( [ eq3.8 ] ) tends to ( [ eq3.4 ] ) as @xmath158 .
for @xmath107 the behavior of @xmath140 near @xmath141 reads , ( see eqs . ( 63 ) and ( c9 ) in @xcite ) , @xmath160 with @xmath161 where @xmath162 is defined by its laplace transform @xmath163 with @xmath164 .
equations ( [ eq3.5 ] ) and ( [ eq3.9 ] ) now yield @xmath165 which replaces eq .
( [ eq3.8 ] ) . performing the inverse laplace transform in ( [ eq3.12 ] ) and using the reflection formula @xmath155 one obtains , for all @xmath166 , @xmath167 note that the exponent of @xmath10 in ( [ eq3.13 ] ) does _ not _ depend on @xmath42 , unlike its counterpart in eq .
( [ eq3.8 ] ) .
it follows in particular that ( [ eq3.13 ] ) does not tend to ( [ eq3.4 ] ) as @xmath158 .
this exponent ` freezing ' from @xmath168 to @xmath169 as @xmath42 decreases past the critical value @xmath170 ( at fixed @xmath27 ) is due to a switch of leading terms in the asymptotic expansion of @xmath171 near @xmath141 from @xmath172 for @xmath151 , to @xmath173 for @xmath107 .
( see eqs .
( 63 ) and ( c9 ) in @xcite ) .
although a term @xmath173 , as regular , does not contribute to the large @xmath80 behavior of @xmath60 , it gives a term @xmath174 in the asymptotic expansion of @xmath175 near @xmath176 in ( [ eq3.5 ] ) which is singular if @xmath27 and thus contributes to ( and dominates ) the large @xmath10 behavior of @xmath11 for @xmath107 .
finally , at the critical value @xmath178 , the equation ( c9 ) in @xcite gives @xmath179 which , together with eq .
( [ eq3.5 ] ) , yields @xmath180 using then @xmath181 and the reflection formula @xmath155 , one obtains @xmath182 defining @xmath183 and @xmath184 we can gather the equations ( [ eq3.4 ] ) , ( [ eq3.8 ] ) , ( [ eq3.13 ] ) , and ( [ eq3.16 ] ) in a more concise form @xmath185 which summarizes the large @xmath10 behavior of @xmath11 at fixed @xmath9 and @xmath49 .
the term @xmath79 in the first line of ( [ eq3.17 ] ) denotes the integer part of @xmath62 and the last two lines are for @xmath27 only .
the random walk is a lvy flight if @xmath21 , in which case the variance of @xmath2 does not exist and @xmath186 at large @xmath63 . in this section
we investigate the asymptotic behavior of @xmath11 for a lvy flight in the limit @xmath83 then @xmath187 . in the cases
@xmath51 with @xmath21 and @xmath27 with @xmath188 , this asymptotic behavior follows readily from eqs .
( [ eq3.4 ] ) , ( [ eq3.8 ] ) , ( [ eq3.16 ] ) , and the large @xmath9 behavior of @xmath189 : @xmath190 ( see eq . ( 48 ) in @xcite )
. one finds , @xmath191 for @xmath51 with @xmath21 and @xmath27 with @xmath192 , and @xmath193 for @xmath27 with @xmath103 .
it remains the case @xmath27 with @xmath107 . injecting the asymptotics @xmath194 ( see eq .
( 41 ) in @xcite ) , and @xmath195 into the right - hand side of ( [ eq3.11 ] ) , one gets @xmath196 the inverse laplace transform of which gives @xmath197 to determine the large @xmath9 behavior of @xmath198 we make the change of variable @xmath199 in eq .
( [ eq3.10 ] ) , let @xmath187 , and use ( [ eq4.7 ] ) and the large @xmath200 behavior of @xmath201 : @xmath202 ( see eq . ( 46 ) in @xcite ) .
one finds @xmath203 where the integral over @xmath204 in the second line is equal to @xmath205 .
thus , from ( [ eq3.13 ] ) , ( [ eq4.1 ] ) , and ( [ eq4.9 ] ) one obtains @xmath206 finally , writing @xmath207 , @xmath208 , and @xmath209 , one has @xmath210 where @xmath79 in the first line denotes the integer part of @xmath62 and the last two lines are for @xmath27 only .
now , we consider the asymptotic behavior of @xmath11 in the large @xmath9 limit . if the support of the jump distribution @xmath2 is bounded , the first gap can not be larger than the ( finite ) diameter of this support and one trivially has @xmath211 for every @xmath10 and @xmath212 with @xmath213 large enough . if the support of @xmath2 is not bounded the result depends on the tail of @xmath2 .
we have singled out three classes of tails which encompass a wide range of jumps of practical interest .
slow decreasing jump distributions are defined by @xmath215 for any fixed @xmath200 . jump distributions with an algebraic tail are slow decreasing .
the large @xmath9 behavior of @xmath216 is then readily obtained from ( [ eq2.5 ] ) , ( [ eq2.6 ] ) , and ( [ eq5.1 ] ) .
one gets @xmath217 where we have used @xmath218 as given by eq .
( [ eq2.7 ] ) in the limit @xmath219 , ( see also e.g. @xcite ) .
the large @xmath9 behavior of @xmath189 depends on @xmath2 and can not be specified further on at this point .
thus , eqs ( [ eq2.4 ] ) and ( [ eq5.2 ] ) yield @xmath220 and from eqs .
( [ eq2.8 ] ) and ( [ eq5.3 ] ) one finds @xmath221 the large @xmath10 behavior of ( [ eq5.4 ] ) is then obtained from the small @xmath47 behavior of @xmath222 , eq .
( [ def_gamma ] ) .
one has @xmath223 which together with eq .
( [ eq5.4 ] ) gives @xmath224 for a lvy flight of index @xmath21 , the large @xmath9 behaviors of @xmath225 and @xmath189 are respectively given by ( [ eq4.8 ] ) and ( [ eq4.1 ] ) , and eqs .
( [ eq5.4 ] ) and ( [ eq5.6 ] ) read @xmath226 and @xmath227 respectively , with @xmath228 and @xmath229 exponentially decreasing jump distributions are defined by @xmath230 for some @xmath231 and any fixed @xmath200 . substituting ( [ eq5.7 ] ) for @xmath232 on the right - hand side of ( [ eq2.5 ] ) and using ( [ eq2.6 ] ) ,
one gets @xmath233 and @xmath234 which , together with eq .
( [ eq2.4 ] ) , yield @xmath235 and from eqs .
( [ eq2.8 ] ) and ( [ eq5.10 ] ) one finds @xmath236 the large @xmath10 behavior of ( [ eq5.11 ] ) is obtained from the behavior of ( [ eq5.10 ] ) near its dominant singularity at @xmath141 and the small @xmath47 behavior of @xmath222 , eq .
( [ def_gamma ] ) .
the former is given by , ( see eq . ( 60 ) in @xcite with @xmath18 and @xmath237 ) , @xmath238 injecting then ( [ eq5.12 ] ) into the right - hand side of ( [ eq5.11 ] ) and performing the inverse laplace transform with @xmath222 given by ( [ def_gamma ] ) , one finds @xmath239 where we have used the reflection formula @xmath155 . fast decreasing jump distributions are defined by @xmath240 for some @xmath241 , any fixed @xmath200 , and where @xmath242 is such that @xmath243 .
super - exponentially distributed jumps are fast decreasing . substituting ( [ eq5.14 ] ) for @xmath232 on the right - hand side of ( [ eq2.5 ] ) and using ( [ eq2.6 ] ) ,
one gets , ( see sec . 4.3 in @xcite for details ) , @xmath244 and @xmath245 eqs .
( [ eq5.15 ] ) and ( [ eq5.16 ] ) , together with eq .
( [ eq2.4 ] ) , yield @xmath246 from which one readily obtains @xmath60 for large @xmath9 , by expanding the logarithm in power series of @xmath247 , ( see eq . ( 103 ) in @xcite ) .
one finds in particular , @xmath248 and @xmath249 for all @xmath250 , where @xmath251 depends on @xmath80 only .
these results reveal a concentration of @xmath60 onto @xmath252 for fast decreasing jump distributions as @xmath187 . from this concentration and the @xmath129 limit of ( [ eq2.1 ] ) , @xmath253 , one has @xmath254 which reads , using @xmath255 and eq .
( [ eq5.18 ] ) , @xmath256 for @xmath257 , @xmath258 has an algebraic tail at large @xmath10 , @xmath259 , which follows from the small @xmath47 behavior ( [ def_gamma ] ) of @xmath222 , and one gets the large @xmath10 behavior of ( [ eq5.21 ] ) as @xmath260 where we have used the reflection formula @xmath155 . on the other hand , eq .
( [ eq3.8 ] ) with @xmath18 and @xmath189 given by eq .
( [ eq5.16 ] ) lead to @xmath261 equations ( [ eq5.22 ] ) and ( [ eq5.23 ] ) suggest the scaling form @xmath262 with @xmath263 where @xmath264 and @xmath265 note that @xmath72 corresponds to the cross - over from a ` concentration ' or ` one - step ' regime ( for @xmath73 ) where the walker get stuck for a long time @xmath10 at the second maximum and then jumps directly to the first maximum , to a ` many - steps ' regime ( for @xmath74 ) where she / he travels a long walk of total duration @xmath10 ( with many steps ) between the second and the first maxima .
it remains to check whether or not the scaling form ( [ eq5.24 ] ) really exists .
rewriting ( [ eq2.7 ] ) as @xmath266 in the first line of ( [ eq5.15 ] ) , with @xmath267 , one finds after some straightforward algebra , @xmath268 where we have made the change of variable @xmath269 in the integral over @xmath16 in eq .
( [ eq5.28 ] ) and use the fact that the integral over @xmath270 is equal to @xmath271 .
inverse laplace transforming ( [ eq5.29 ] ) w.r.t .
@xmath47 and using ( [ eq5.27 ] ) yields @xmath272 equation ( [ eq5.30 ] ) gives a uniform expression of @xmath11 for a fast decreasing @xmath2 when both @xmath9 and @xmath10 are large . for @xmath257 , one has the large @xmath10 behavior @xmath273 from which it is readily seen that ( [ eq5.30 ] ) reduces to the scaling form ( [ eq5.24 ] ) with the scaling function @xmath274 which fulfills the large and small argument behaviors ( [ eq5.25 ] ) , as it should be . for @xmath51 ,
there is no scaling form such as ( [ eq5.24 ] ) but the uniform expression ( [ eq5.30 ] ) makes it possible to determine the domains in the @xmath275 plane corresponding to the ` concentration ' or ` one - step ' regime and to the ` many - steps ' regime , respectively .
taking for instance @xmath276 and comparing the two terms on the right - hand side of ( [ eq5.30 ] ) , one finds that the ` one - step ' regime corresponds to the domain @xmath277 and the ` many - steps ' regime to the complementary domain ( i.e. with @xmath278 replaced with @xmath279 ) .
figure [ fig1 ] summarizes the different behaviors of @xmath11 in the plane @xmath280 for a fast decreasing jump distribution and @xmath27 .
the ` one - step ' ( resp . ` many - steps ' )
regime is on the left ( resp .
right ) of the diagonal . for fast decreasing jump distributions and @xmath27 (
amplitudes are given in the text ) .
for @xmath281 and @xmath282 , @xmath11 is given by eq .
( [ eq3.8 ] ) with @xmath18 . for both @xmath281 and @xmath283 ,
one has the scaling form ( [ eq5.24 ] ) with @xmath284 respectively given by the first line of ( [ eq5.25 ] ) if @xmath285 and by the second line of ( [ eq5.25 ] ) if @xmath286 . for @xmath287 and @xmath283 , @xmath11
is given by eq .
( [ eq5.21]).___,width=340 ]
in the preceding section we have found that for a fast decreasing jump distribution and @xmath27 , the asymptotic behavior of @xmath11 at large @xmath10 and @xmath9 takes the scaling form ( [ eq5.24 ] ) .
here we consider a lvy flight of index @xmath21 corresponding to a slow decreasing jump distribution .
we will see that , in this case too , @xmath11 at large @xmath10 and @xmath9 can be put into a scaling form for all @xmath49 .
consider first the case @xmath51 . by eq .
( [ eq3.3 ] ) and the scaling form ( 87 ) in @xcite , @xmath11 depends on @xmath10 through the combination @xmath288 only as both @xmath10 and @xmath9 are large .
now , for @xmath27 , the comparison of eqs .
( [ eq3.2 ] ) and ( [ eq3.5 ] ) indicates that @xmath10 should be replaced with @xmath289 .
thus , for @xmath49 , @xmath11 is expected to depend on @xmath10 through the combination @xmath75 only as both @xmath10 and @xmath9 are large .
the simplest scaling ansatz for @xmath11 consistent with this argument and the asymptotic behaviors ( [ eq4.11 ] ) and ( [ eq5.6a ] ) reads @xmath290 with the following large and small argument behaviors for the scaling function : @xmath291 and @xmath292 the term @xmath79 in the first line of ( [ eq6.2 ] ) denotes the integer part of @xmath62 and the last two lines are for @xmath27 only . the scaling form ( [ eq6.1 ] ) holds in the regime where both @xmath9 and @xmath10 are large , with @xmath75 fixed .
figures [ fig2a ] , [ fig2b ] , and [ fig2c ] summarize the different behaviors of @xmath11 in the plane @xmath293 for a a lvy flight of index @xmath21 .
they correspond respectively to the cases @xmath294 with @xmath49 , @xmath103 with @xmath27 and @xmath107 with @xmath27 . for a lvy flight of index @xmath294 and @xmath49 (
amplitudes are given in the text ) .
for @xmath281 and @xmath282 , @xmath11 is given by the first line of eq .
( [ eq3.17 ] ) . for both @xmath281 and @xmath283 ,
one has the scaling form ( [ eq6.1 ] ) with @xmath295 given by the first line of ( [ eq6.2 ] ) if @xmath296 and by ( [ eq6.3 ] ) if @xmath297 . for @xmath287 and @xmath283 , @xmath11
is given by eq .
( [ eq5.4a]).___,width=340 ] for a lvy flight of index @xmath103 and @xmath27 ( amplitudes are given in the text ) .
for @xmath281 and @xmath282 , @xmath11 is given by the second line of eq .
( [ eq3.17 ] ) . for both @xmath281 and @xmath283 ,
one has the scaling form ( [ eq6.1 ] ) with @xmath295 given by the second line of ( [ eq6.2 ] ) if @xmath298 and by ( [ eq6.3 ] ) if @xmath299 . for @xmath287 and @xmath283 , @xmath11
is given by eq .
( [ eq5.4a ] ) with @xmath103.___,width=340 ] for a lvy flight of index @xmath300 and @xmath27 ( amplitudes are given in the text ) .
for @xmath281 and @xmath282 , @xmath11 is given by the third line of eq .
( [ eq3.17 ] ) . for both @xmath281 and @xmath283 ,
one has the scaling form ( [ eq6.1 ] ) with @xmath295 given by the third line of ( [ eq6.2 ] ) if @xmath296 and by ( [ eq6.3 ] ) if @xmath297 . for @xmath287 and @xmath283 , @xmath11
is given by eq .
( [ eq5.4a]).___,width=340 ]
we can now determine the large @xmath10 behavior of the marginal distribution @xmath84 of the time between the first two maxima .
the marginal distribution @xmath105 of the gap between the first two maxima does not depend on @xmath62 and is the same as the one already studied in sec .
3 of ref .
@xcite in the discrete time rw setting . for @xmath51 ,
the large @xmath10 behavior of @xmath84 is readily obtained by integrating ( [ eq3.3 ] ) over @xmath9 and by using the large @xmath80 asymptotics ( 112 ) in @xcite .
one finds @xmath302 with @xmath303 to determine the large @xmath10 behavior of @xmath84 for @xmath27 , fix @xmath304 , @xmath305 , and for @xmath306 write @xmath307 for @xmath151 and @xmath83 , ( [ eq7.2 ] ) is dominated by the first two integrals in which one can use the large @xmath10 expression ( [ eq3.8 ] ) of @xmath11 .
one finds @xmath308 for @xmath107 and @xmath83 , ( [ eq7.2 ] ) is dominated by the contribution of large , @xmath309 , values of @xmath9 , i.e. by the last two integrals in which one can use the large @xmath10 and @xmath9 scaling form ( [ eq6.1 ] ) of @xmath11 .
one gets @xmath310 where we have made the change of variable @xmath311 and used the third line of ( [ eq6.2 ] ) .
finally , for @xmath103 and @xmath83 we use the second line of ( [ eq6.2 ] ) in the second line of ( [ eq7.4 ] ) with @xmath103 , which gives @xmath312 to summarize , for @xmath27 the large @xmath10 behavior of the marginal distribution @xmath84 is given by @xmath313 with @xmath314 where @xmath207 , @xmath208 , and @xmath209 , the amplitudes @xmath315 and @xmath316 being respectively given by ( [ eq3.17a ] ) and ( [ eq3.17b ] ) .
it can be seen that the smaller @xmath62 is , the slower @xmath84 decreases as @xmath10 increases .
the third lines of eqs .
( [ eq7.1 ] ) and ( [ eq7.6 ] ) reveal an interesting freezing , as a function of @xmath42 , of the large @xmath10 behavior of @xmath84 as @xmath42 decreases past the critical value @xmath170 .
it is the continuous time counterpart of the freezing phenomenon pointed out in sec .
5 of @xcite for discrete time random walks . in particular , it follows immediately from ( [ eq7.1 ] ) and ( [ eq7.6 ] ) that the first moment of @xmath84 is never defined .
this means that although the typical size of @xmath10 is @xmath317 , its average always diverges with the total duration of the walk .
writing @xmath0 the total duration of the walk , eqs .
( [ eq7.1 ] ) and ( [ eq7.6 ] ) yield the estimate @xmath318 for @xmath49 .
in this section , we check our main analytical results by comparing them with the ones of numerical simulations of the corresponding ctrw . in the simulations ,
the random jump variables @xmath33 s and the random time intervals between two successive jumps @xmath43 s were drawn from the following distributions . _ jump variables . _ for @xmath94 , the jumps were drawn either from a gaussian distribution @xmath319 when we investigate the joint pdf of @xmath9 and @xmath10 for fast decreasing jump distribution or from a symmetric exponential distribution @xmath320 when we study the marginal distribution @xmath84 in which case the amplitudes @xmath86 and @xmath321 in ( [ eq1:result_p_of_t ] ) and ( [ eq2:result_p_of_t ] ) can be evaluated exactly . on the other hand , for @xmath322 , we chose for @xmath2 a symmetric pareto distribution of index @xmath42 of the form @xmath323 for which it is straightforward to compute the parameter @xmath324 in eq .
( [ def_mu ] ) .
one finds , @xmath325^{1/\mu } \;.\end{aligned}\ ] ] the reason for such a choice is twofold : ( i ) random numbers drawn from ( [ eq : pareto_jump ] ) are easy and fast to generate numerically ( see for instance ref .
@xcite ) and ( ii ) @xmath2 is a power law for all @xmath326 ( instead of @xmath327 ) which allows us to access the asymptotic regime characterizing lvy flights rather quickly in numerical simulations
. _ time intervals . _ similarly , to generate a ctrw numerically , it is also convenient to draw the time variables @xmath43 s from a distribution @xmath4 which is a ( one - sided ) pareto distribution of index @xmath26 of the form @xmath328 for such a distribution the characteristic time @xmath329 is given by @xmath330^{1/\gamma } \;.\end{aligned}\ ] ] computing a joint pdf numerically is notoriously a quite hard task .
so , instead of @xmath11 we consider the cumulative distribution @xmath331 which yields better statistics at a lesser cost .
we first present the result for a fast decreasing jump distribution @xmath2 in the sense of sec .
[ sec5.3 ] , i.e. @xmath332 when @xmath41 [ see eq .
( [ eq5.14 ] ) ] . in this case , one expects from eq .
( [ scaling1a ] ) that in the limit @xmath333 and @xmath334 keeping @xmath335 fixed , @xmath336 takes the following scaling form : @xmath337 where the scaling function @xmath338 is related to the scaling function @xmath339 through the relation @xmath340 from the asymptotic behaviors of @xmath339 in eq .
( [ scaling1b ] ) , one deduces the asymptotic behaviors of @xmath338 as follows @xmath341 as discussed above , the limit @xmath342 corresponds to the ` single - step ' regime while the limit @xmath343 corresponds to the ` many - steps ' regime . in fig .
[ fig : fast ] we show a plot of @xmath344 ^ 2 $ ] as a function of @xmath345 for a fast decreasing jump distribution , namely a gaussian jump distribution , corresponding to @xmath65 , and with @xmath346 , for different values of @xmath347 and @xmath348 , while the total duration of the ctrw is @xmath349 .
the data have been obtained by averaging over @xmath350 independent realizations of the ctrw .
we see on this plot that , for sufficiently large values of @xmath9 , the curves for different times @xmath10 tend to collapse on a master curve , as predicted by eq .
( [ cumul_fast ] ) , which indeed holds only in the limit @xmath334 , @xmath333 with @xmath335 fixed . for smaller values of @xmath9
we see a clear deviation from this scaling form .
we emphasize on the fact that , in our simulations , we have been able to access the regime where the scaling variable @xmath351 is small ( corresponding to the ` many - steps ' regime ) , where @xmath338 is expected to behave as @xmath352 .
it is extremely difficult to reach the complementary ` single - step ' regime corresponding to @xmath353 . indeed , for such high values of @xmath354 the condition @xmath353 with @xmath65 and @xmath346 yields @xmath355 , which is already large enough to prevent any access to reliable statistics ( to understand the difficulty , just realize that for the gaussian jump distribution we consider in the ` single - step ' regime , the typical gaps are greater than @xmath356 less often than one time out of the avogadro s number ) . consequently , the clear numerical observation of the crossover behavior between the ` many - steps ' and the ` single - step ' regimes described by the scaling function in eq .
( [ cumul_fast ] ) goes beyond the scope of the present paper .
notice that , for @xmath357 and @xmath358 , the ` single - step regime ' was clearly demonstrated in ref .
@xcite . ^2 $ ] as a function of @xmath345 for a fast decreasing jump distribution , namely a gaussian jump distribution , corresponding to @xmath65 , and with @xmath346 .
the different curves correspond to different values of @xmath347 and @xmath348 for a ctrw of duration @xmath349 .
the ( relatively ) good collapse of the different curves for large values of @xmath9 corroborate the scaling form predicted in eq .
( [ cumul_fast ] ) .
the solid line is a guide to eyes , corresponding to the small argument behavior of the scaling function @xmath338 in eq .
( [ asympth ] ) with @xmath65 .
* inset : * same data as in the main panel but without the rescaling . ]
we now present numerical results for lvy flights ( @xmath21 ) , for which @xmath2 is slow decreasing in the sense of sec .
[ sec5.1 ] . in this case
, one expects from eq .
( [ scaling2a ] ) that in the limit @xmath333 and @xmath334 at fixed @xmath359 , @xmath336 takes the following scaling form : @xmath360 where the function @xmath361 is related to the scaling function @xmath362 through the relation @xmath363 the small and large @xmath364 behaviors of @xmath365 are readily obtained from the asymptotic behaviors of @xmath362 given in eqs .
( [ scaling2b ] ) and ( [ scaling2c ] ) .
one gets @xmath366 for large @xmath364 , where @xmath367 , and @xmath368 for small @xmath364 , where the last two lines are for @xmath27 only . in fig .
[ fig_joint_combined ] we show a scaled plot of @xmath369 as a function of @xmath370 with @xmath371 for different values of @xmath372 and @xmath373 , for ctrws of total duration @xmath374 .
the data have been obtained by averaging over @xmath350 independent realizations of the ctrw .
this plot demonstrates a good agreement with our predictions for the scaling form of the joint pdf in eq .
( [ scaling_cumul ] ) as well as with the asymptotic behaviors predicted in eqs .
( [ g : large ] ) and ( [ g : small ] ) . an interesting feature of our results for the distribution @xmath84 of the time @xmath10 between the first two maxima is the presence of logarithmic corrections for @xmath375 and @xmath27 [ see eq . ( [ eq2:result_p_of_t ] ) ] . to demonstrate these logarithmic corrections numerically
, it is convenient to study the distribution @xmath376 of the logarithm of the time , @xmath377 instead of the time distribution itself . from @xmath378 and eq .
( [ eq2:result_p_of_t ] ) it follows that the asymptotic behavior of @xmath376 at large @xmath379 and for @xmath27 is given by @xmath380 with the amplitudes @xmath381 where the value of @xmath382 corresponds to a symmetric exponential jump distribution ( see the beginning of this section ) . in fig .
[ fig_time1 ] , we show a plot of @xmath376 as a function of @xmath379 for three different values of @xmath42 : @xmath94 ( with exponentially distributed jumps ) , @xmath383 , and @xmath384 , that are representative of the three different regimes in eq .
( [ eq : asympt_log ] ) , for a fixed value of @xmath385 .
this plot shows that the large @xmath379 behavior of @xmath376 does depend on @xmath42 as predicted in eq .
( [ eq : asympt_log ] ) . to make a precise comparison between our numerical results and our analytical predictions ( and in particular to demonstrate the expected logarithmic corrections ) , we now analyze these data , for different values of @xmath42 , more carefully .
[ fig_precise_exp ] corresponds to the case of a sub - diffusive walk with @xmath18 ( with exponentially distributed jumps ) and @xmath386 .
we show a plot of @xmath387 for @xmath385 as a function of @xmath388 for ctrws of increasing length , namely @xmath374 , @xmath389 and @xmath349 . here
, the data have again been obtained by averaging over @xmath350 independent ctrws .
it can be seen that in the time interval @xmath390 where our prediction actually holds , @xmath387 approaches a constant which gets closer to the expected value @xmath382 as the total duration @xmath0 of the walk increases .
the observed small upward discrepancy , as well as the steep upward bending for larger @xmath379 , are finite size effects .
note the very stretched vertical scale . for the longest walks with @xmath349 , numerical and analytical results are in agreement to within an accuracy of less than 5@xmath391 .
next we consider the case @xmath384 , which corresponds to the third regime in eq .
( [ eq : asympt_log ] ) . in fig .
[ fig_precise_iii ] we show a plot of @xmath392 , for @xmath384 and @xmath386 as a function of @xmath379 and for different durations of the ctrw , @xmath374 , @xmath393 , and @xmath349 . as before ,
the data have been obtained by averaging over @xmath350 independent ctrws .
the linear behavior , in a good agreement with our analytical prediction @xmath394 , is a clear signature of the logarithmic correction of @xmath84 as expected from our analytical result in eq .
( [ eq2:result_p_of_t ] ) . finally , fig .
[ fig_precise_ii ] shows a plot of @xmath395 as a function of @xmath379 , for @xmath103 , @xmath385 , and the same walk durations as before ( averaging procedure is also the same ) .
it can be seen that our numerical data are compatible with the expected quadratic behavior ( [ eq : asympt_log ] ) , @xmath396 .
recalling that @xmath388 , this quadratic behavior confirms our analytical predictions for @xmath84 in eq .
( [ eq2:result_p_of_t ] ) .
in this paper , we have performed a detailed analytical study of the statistics of the gap @xmath9 and the time interval @xmath10 between the first two maxima of a ctrw in the limit where the number of steps in the walk goes to infinity .
the results we have obtained are quite general as we have addressed the question for the wide class of a symmetric , bounded , and ( piecewise ) continuous jump distribution , including the case of lvy flights of index @xmath42 , with @xmath21 .
the average trapping time @xmath397 between two successive jumps can be either finite , @xmath398 , or infinite . in the latter case
, the trapping time distribution @xmath4 is assumed to have a power law tail @xmath22 , with @xmath23 . for a finite average trapping time between two successive jumps , our results coincide with the ones in @xcite , devoted to discrete time rws , in which the number of jumps between the first two maxima , @xmath80 , is merely replaced with @xmath81 , as expected from simple law of large number arguments .
on the other hand , our study provides a non trivial extension of the one in @xcite to the case where the average time between two successive jumps does not exist and law of large number arguments can not be used . in particular , by taking @xmath399 it makes it possible to address the question for sub - diffusive rws , out of reach of the discrete time analysis of @xcite . we have first shown that the joint pdf @xmath57 converges to a stationary pdf , @xmath400 , as @xmath58 , where @xmath14 is the total number of steps in the walk .
we have then obtained an explicit expression of the laplace transform of @xmath11 w.r.t .
@xmath10 , given in equations ( [ eq2.8 ] ) and ( [ eq2.5 ] ) , from which we have performed a detailed analysis of @xmath11 in the plane @xmath61 for @xmath42 and @xmath62 in the whole ranges @xmath3 and @xmath49 , and for the three different main classes identified in @xcite : ( i ) slow , ( ii ) exponentially , and ( iii ) fast decreasing @xmath2 at large @xmath63 . in the case ( iii ) of a fast decreasing jump distribution ( @xmath18 ) with @xmath27 and in the scaling regime @xmath66 with fixed @xmath67 , we have shown that @xmath11 takes a scaling form , eqs .
( [ scaling1a ] ) and ( [ scaling1b ] ) .
the switch from the first to the second behavior ( [ scaling1b ] ) corresponds to the cross - over from a ` concentration ' or ` one - step ' regime where the walker get stuck for a long time @xmath10 at the second maximum and then jumps directly to the first maximum , to a ` many - steps ' regime where she / he travels a long walk of total duration @xmath10 ( with many steps ) between the second and the first maxima .
we have summarized these results in figure [ fig1 ] .
the numerical observation of the cross - over and ` one - step ' regime is still out of reach as it requires a prohibitively large value of @xmath9 .
note , however , that we expect a similar scaling form for @xmath11 if @xmath2 has a bounded support @xmath401 and @xmath187 is replaced with @xmath402 , the largest possible ( bounded ) value of the gap .
this should make a numerical observation of the cross - over and ` one - step ' regime easier in this case .
the study of @xmath11 for jump distributions with a bounded support ( in both discrete and continuous time settings ) will be the subject of a future work . for type ( i ) lvy flights of index @xmath42 , with @xmath39 , and in the scaling regime @xmath66 with fixed @xmath75
, we have shown that @xmath11 takes the scaling form ( [ scaling2a ] ) with the asymptotic behaviors ( [ scaling2b ] ) and ( [ scaling2c ] ) .
these results , summarized in figures [ fig2a ] to [ fig2c ] , provide a non trivial extension of the ones in @xcite to the case @xmath27 .
finally , from the stationary joint pdf @xmath11 we have computed the stationary marginal distribution , @xmath84 , of the time between the first to maxima of the walk .
its asymptotic behavior , which depends on @xmath62 and the lvy index @xmath42 , is summarized in equations ( [ eq1:result_p_of_t ] ) , ( [ eq2:result_p_of_t ] ) , and in table [ tab1 ] .
the third lines of ( [ eq1:result_p_of_t ] ) and ( [ eq2:result_p_of_t ] ) reveal that the freezing phenomenon of the large @xmath10 behavior of @xmath84 , as a function of @xmath42 , as @xmath42 decreases past the value @xmath96 , which was first put in evidence for discrete time rws in @xcite , also exists for ctrws . as a consequence , the first moment of @xmath84 is never defined .
our study raises interesting open questions .
for instance , this work focuses on the first gap and it would be interesting to study higher order gaps . in ref .
@xcite , the study of the @xmath16-th gap of long random walks , for @xmath18 revealed a rich and universal behavior ( i.e. , independent of the jump distribution ) of the stationary pdf of the @xmath16-th gap in the large @xmath16 limit .
the extension of these results , which also hold for ctrw with @xmath18 and any value of @xmath403 , to lvy flights , i.e. @xmath39 , remains a challenging open question .
similarly , it would very interesting to study the time between the @xmath16-th and @xmath404-th maximum for ctrw ( both for @xmath405 and @xmath257 ) , for which one also expects universality for @xmath406 .
finally , the generalization of our results to higher dimensions would also be very interesting . | we consider a one - dimensional continuous time random walk ( ctrw ) on a fixed time interval @xmath0 where at each time step the walker waits a random time @xmath1 , before performing a jump drawn from a symmetric continuous probability distribution function ( pdf ) @xmath2 , of lvy index @xmath3 .
our study includes the case where the waiting time pdf @xmath4 has a power law tail , @xmath5 , with @xmath6 , such that the average time between two consecutive jumps is infinite .
the random motion is sub - diffusive if @xmath7 ( and super - diffusive if @xmath8 ) .
we investigate the joint pdf of the gap @xmath9 between the first two highest positions of the ctrw and the time @xmath10 separating these two maxima .
we show that this pdf reaches a stationary limiting joint distribution @xmath11 in the limit of long ctrw , @xmath12 .
our exact analytical results show a very rich behavior of this joint pdf in the @xmath13 plane , which we study in great detail .
our main results are verified by numerical simulations .
this work provides a non trivial extension to ctrws of the recent study in the discrete time setting by majumdar _
et al . _
( _ j .
stat .
mech .
_ p09013 , 2014 ) . |
the advent of a new generation of large area radio surveys which sample the radio sky down to mjy levels ( e.g. first , becker et al . , 1995 ; nvss , condon et al . , 1998 ;
sumss , bock et al . , 1999 ) has introduced great advantages for both radio astronomy and cosmology . from the cosmological point of view
, the high surface density of sources allows studies of the large - scale structure of the universe and its evolution up to high redshifts ( @xmath15 ) and large physical scales ( see e.g. cress et al . , 1996 ; magliocchetti et al . , 1998 , 1999 ) .
furthermore , due to the low flux limits reached , these surveys sample populations of sources , such as star - forming galaxies , where the radio signal is produced by phenomena other than agn activity .
a complete investigation of the nature of radio sources and their properties can only be achieved with multi - wavelength follow - up . in particular optical identifications
enable the acquisition of optical spectra , which can be used to derive their spectral type and redshift distribution ( see e.g. magliocchetti et al . , 2000 - ma2000 hereafter ) .
great effort has recently been made to determine the photometric and spectroscopic properties of radio sources at mjy levels and fainter ( gruppioni et al .
, 1998 ; georgakakis et al . , 1999 ; sadler et al . , 1999 ; ma2000 ; masci et al . ,
these studies have proven extremely useful in characterizing the populations of radio sources , but there are still significant uncertainties .
for instance it has been established that there is starbursting population of radio sources , but there is still some debate about the overall fraction ( see for instance the results from sadler et al . , 1999
as compared with those derived in gruppioni et al . , 1998 ;
georgakakis et al .
, 1999 ; ma2000 ) .
these follow - up studies however suffer from two great limitations : either they survey small areas of the sky and therefore generate relatively small samples of objects , or they lack completeness .
the small number of sources leads to a low confidence level for any result drawn from statistical studies , while radio or optical incompleteness leads to biases in the determination of the redshift distribution of radio sources at such low - flux levels .
in this paper we present an analysis of the properties and angular clustering of @xmath16 4100 radio sources brighter than 1 mjy with optical counterparts brighter @xmath17 .
this sample was obtained by matching together objects from the first and apm surveys over @xmath18 square degrees near the celestial equator . despite not having redshift measurements for these sources ,
their joint radio , photometric and morphological properties can be used to infer extremely useful information on their nature .
furthermore , for @xmath19 , we obtain a catalogue of sources which is @xmath20 complete in the optical band ( ma2000 ) and 80% complete in the radio band ( with a completeness level rising to 100% for radio fluxes @xmath21 mjy , becker et al . , 1995 ) .
the photometric and morphological properties of the optical identifications are then used to divide the whole sample of radio sources into two well defined populations .
the first population consists of low - redshift radio galaxies and the second one of high - z qso and objects with no counterpart on the ukst plates .
the homogeneity and completeness of the low - z sample allows us to study the clustering properties of these sources .
lastly , the area chosen for our analysis coincides with some of the fields observed in the 2df galaxy redshift survey ( maddox 1998 , colless , 1999 ) .
this sample of optical identifications of @xmath16 4100 radio sources therefore provides an excellent starting point for further wide - area spectroscopic follow - up .
the layout of the paper is as follows : section 2 briefly describes the two surveys and the data coming from them .
section 3 presents the procedure we adopted to match radio and optical sources together , while section 4 is devoted to the analysis of the photometric properties of the optical identifications .
section 5 examines the clustering properties of the sample and section 6 summarizes our conclusions . throughout the paper
we will assume @xmath22 , @xmath23 , @xmath24 .
the first ( faint images of the radio sky at twenty centimetres ) survey ( becker , white and helfand , 1995 ) began in the spring of 1993 and will eventually cover some 10,000 square degrees of the sky in the north galactic cap and equatorial zones . the beam - size is 5.4 arcsec at 1.4 ghz , with an rms sensitivity of typically 0.15 mjy / beam .
a map is produced for each field and sources are detected using an elliptical gaussian fitting procedure ( white et al . , 1997 ) ; the @xmath25 source detection limit is @xmath26 mjy .
the astrometric reference frame of the maps is accurate to 0.05 arcsec , and individual sources have 90 per cent confidence error circles of radius @xmath27 0.5 arcsec at the 3 mjy level , and 1 arcsec at the survey threshold .
the surface density of objects in the catalogue is @xmath28 per square degree , though this is reduced to @xmath29 per square degree if we combine multi - component sources ( magliocchetti et al . , 1998 ) .
the depth , uniformity and angular extent of the survey are excellent attributes for investigating , amongst others , the clustering properties of faint sources .
the catalogue derived from this survey has been estimated to be 95 per cent complete at 2 mjy and 80 per cent complete at 1 mjy ( becker et al . , 1995 ) .
we used the 5 july 2000 version of the catalogue which contains approximately 722,354 sources from the north and south galactic caps .
this is derived from the 1993 through 2000 observations that cover nearly 7988 square degrees of sky , and includes most of the area @xmath30 , @xmath31 and @xmath32 , @xmath33 .
the apm galaxy survey has been extensively described in maddox et al . , ( 1990a , 1990b , 1996 ) . briefly , it is based on apm scans of @xmath34 ukst plates , covering about 7000 square degrees in the regions dec @xmath35 in the southern galactic cap and @xmath36 dec @xmath37 , @xmath38 dec @xmath39 respectively in the equatorial sgp and ngp . a scan of a typical plate records about 300,000 images , with a limiting magnitude for image detection @xmath40 .
for each image , the measurements show a positional accuracy of @xmath26 arcsec and an isophotal magnitude accuracy of @xmath41 mag .
the photometry from each plate is corrected so that it is consistent with the neighbouring plates , in the way described in maddox et al .
( 1990b ) .
the limit for uniform image detection over the whole survey is set by the shallowest plate , and is found to be @xmath42 . [
cols="<,<,<,<,<",options="header " , ] columns 2 and 3 of table [ table_identifications ] respectively give the total number of radio sources and number of sources with optical identifications ( with @xmath40 ) found in each field .
the total number of sources in the area @xmath43 , @xmath44 with an identified optical counterpart within 2 arcsecs is 4075 , 16.7 per cent of the whole radio sample .
this percentage drops to 13 per cent ( corresponding to 3176 identifications , column 4 of table [ table_identifications ] ) if we include the further constraint @xmath45 , to obtain a uniform sample from the apm survey . in order to quantify how many matches are likely to be random coincidences
, we ran the matching code once again , after shifting all the radio positions 1 arcmin north .
the results are shown in figure [ fig : allres2 ] .
the positional residuals are uniformly distributed on the @xmath46 @xmath47 plane as expected for random coincidences .
furthermore , we find only 225 sources with offsets within our chosen matching radius of 2 arcsecs , corresponding to @xmath48 per cent of the true matches .
contamination from spurious coincidences at this level is not a problem in our analysis .
note that if we had chosen a matching radius as large as 5 arcsecs , the percentage of random matches would have been 1200/5866 @xmath49 per cent , which would have required a more sophisticated treatment of contamination .
these numbers are in good agreement with the results obtained by evaluating the expected number of chance coincidences : for 68.7 radio sources per square degree , one in fact respectively obtains 182 and 1139 random matches by considering radii of 2 and 5 arcsecs over the 350 square degrees of analysed area .
we divided the radio sources with optical identifications into galaxies and stellar objects based on the optical classifications described in section 2 .
figure [ fig : hist_identifications ] shows the number of galaxies ( dotted line ) and stellar sources ( solid line ) in the radio subsample with optical identifications ( hereafter called the id - sample ) as a function of @xmath50 magnitude . as expected from the analysis performed in section 2 , the number of sources identified as galaxies rises with increasing magnitude up to @xmath51 , where the optical images become too faint to allow the detection of any extended structure . correspondingly , the number of stellar objects increases beyond its real value for @xmath52 .
again @xmath53 is seen as the faintest magnitude for which we can discriminate between extended images ( radio galaxies ) and stellar images ( mainly qso ) .
the stellar identifications also include a very small number of galactic stars - less than 0.1 per cent of the objects in the first survey are expected to be associated with stars , see helfand et al . , 1997 ) .
columns 5 and 6 of table [ table_identifications ] indicate the number of non - stellar and stellar sources found on each of the ukst plates in our analysis . a joint analysis of the radio and photometric properties of the id - sample can provide us with extremely useful insights on the nature of the radio sources under examination .
the @xmath50 and @xmath9 magnitudes are plotted against radio fluxes in figures [ fig : fb ] , [ fig : fr ] , where black dots indicate galaxies and red dots indicate stellar objects .
note that beyond @xmath51 ( corresponding to @xmath9 @xmath54 ) the distinction between galaxies and stars becomes very difficult , therefore all the sources appear in the catalogue as stellar .
the following analysis will therefore only apply to magnitudes @xmath3 .
the radio - to - optical ratio is defined as r@xmath55 @xmath56 , where s is the radio flux in mjy and m is the apparent magnitude ( either @xmath9 or @xmath50 depending on the specific case .
we recall here that a source is considered radio - loud if r@xmath57 @xmath58 ( see e.g. urry & padovani 1995 ) .
it has been shown ( see for instance magliocchetti , celotti & danese , 2001 ) that for star - bursting galaxies , where the star - formation activity is so intense that it hides the signatures of any central bh , r@xmath57 @xmath59 .
thus the small number of sources below the r@xmath57=100 line in figure [ fig : fb ] shows that there are @xmath60 star - forming galaxies in the first sample down to 1 mjy .
this is further confirmation of the result that star - bursting sources show up only at radio fluxes below the mjy level ( see for instance , gruppioni et al . 1998 , georgakakis et al . 1999 and ma2000 ) .
the region of the s-@xmath50 plane between r @xmath61 and r @xmath62 is mainly dominated by sources identified as ( radio ) galaxies , and this population constitutes the majority of the id - sample . for r @xmath63 and @xmath64 stellar identifications , most likely to be high - z qso , dominate the sample .
fainter than @xmath65 this effect is mainly an artefact of the data because the optical images are too faint to resolve galaxies .
figure [ fig : fr ] illustrates the flux - magnitude diagram in the @xmath9 band .
apart from the features already seen in figure [ fig : fb ] , we note an interesting shift of the population of radio galaxies towards brighter red magnitudes .
this shift is especially significant for s @xmath58 mjy and is not observed for the stellar identifications , highlighting the fact that radio - galaxies are typically red , early - type objects .
this effect can be seen more clearly in figure [ fig : frb ] , which plots the @xmath4 colour of sources in the id - sample for objects with @xmath50 @xmath66 .
the majority of the id - sample consists of galaxies with very red colours , up to @xmath67 .
this conclusion is actually made even stronger as in our analysis we might have missed very red sources which had @xmath50 magnitudes too faint to be included in the original catalogue .
the id - sample was drawn from a catalogue selected on @xmath50 magnitude ; the red magnitudes were added later as explained in section 2 .
note the clear distinction in figure [ fig : frb ] between radio galaxies , which appear mostly with @xmath68 and qso , which dominate the region @xmath69 .
this division becomes striking for radio fluxes s @xmath70 10 , where the two distributions take the shape of a tuning fork .
again we can see that star - forming galaxies showing blue ( i.e. for @xmath71 ) colours and low ( s @xmath72 ) radio - fluxes constitute only a few tens of objects in the lower part of the plot
. also when we consider the number of sources per unit of radio flux ( see figure [ fig : hist_flux ] ) we see that stellar and non - stellar ids show different behaviours .
the number of galaxies ( the dashed line ) , shows a steeper slope than the number of qso ( the solid line ) .
this difference reflects the fact that galaxies tend to have lower radio fluxes , while qso dominate the region @xmath73 mjy .
the acquisition of optical identifications for radio sources is not only useful in providing optical information , it also gives the opportunity to make a distinction between low - redshift sources and more distant ones .
passive radio galaxies are known to be reliable standard candles ( see e.g. hine & longair , 1979 ; rixon , wall & benn , 1991 ) . the mean absolute magnitude in the red band
is @xmath74 ( see e.g. ma2000 ) , and there is little scatter about the mean ( ma2000 estimate @xmath75m@xmath76 at the 1@xmath77 level ) .
thus measurements of the apparent magnitude @xmath9 for these sources can provide us with an estimate of their redshifts according to the relation @xmath78 where @xmath79 the luminosity distance .
the id - sample is complete to @xmath80 , and from this formula we expect that objects brighter than @xmath80 should be found in the redshift range @xmath81 .
+ as we saw in the previous section , the optically extended id - sample is mostly made up of early - type galaxies , with star - bursting objects contributing only a small fraction .
so we can apply equation ( [ eq : r ] ) to the whole subsample of radio galaxies , and deduce that it contains radio - sources that are mostly closer than @xmath14 .
the spread in the @xmath82 relation means that the upper redshift limit will be a gradual cutoff ; see also ma2000 .
the other sources that appear in the id - sample are mainly qso with a very small contamination from nearby stars .
unfortunately , we have no simple way to distinguish between these two classes of objects , or to estimate the qso redshift . however , apart from rare exceptions , we expect qso to be all placed at high redshifts , at least beyond z @xmath83 .
+ the optical distinction between point - like and extended sources as discussed in the previous section , allows us to divide the id - sample in a low - z subsample - comprising mostly of early - type galaxies and roughly complete to @xmath14 - and a sample of qso mostly found at redshifts well beyond 0.3 .
figure [ fig : distrib ] shows the angular distribution of these sources projected on the sky .
the bottom panel represents the distribution of all the objects in the id - sample with @xmath84 , limit for completeness and uniformity of the optical catalogue , while the middle one includes radio sources identified as galaxies and the top one is for point - like sources ( both in the case of @xmath85 ) .
we can quantify the level of clustering in each sample by means of the two - point correlation function .
ideally one would like to obtain the spatial correlation function @xmath86 but , since we do not have measured redshifts for any sources in the id - sample , we have to deal with its angular counterpart @xmath12 .
we recall here that the angular two - point correlation function measures the excess probability , with respect to a random poisson distribution , of finding two sources in the solid angles @xmath87 @xmath88 separated by an angle @xmath89 .
it is defined as @xmath90 \label{eqn : wthetadef}\end{aligned}\ ] ] where @xmath91 is the mean number density of objects in the catalogue under consideration .
we calculated @xmath12 using the estimator ( hamilton , 1993 ) @xmath92 where the number of data - data , random - random and data - random pairs separated by an angle @xmath89 are denoted by dd , rr and dr respectively .
+ random catalogues were generated with a spatial distribution modulated by both the apm and first coverage maps , so that the instrumental window functions do not affect the measured clustering .
we estimate the errors on the @xmath12 measurements by assuming the the pair counts follow poisson statistics .
this underestimates the uncertainties , but we do not expect this bias to be a large factor .
however , the errors in @xmath12 at each point are correlated , so it is easy to overestimate the significance of any features apparent in @xmath93 . in figure
[ fig : wall ] we show the results for @xmath12 of the whole id - sample with @xmath84 .
as already stated , error - bars are given by poisson estimates for the catalogue under consideration .
the clustering signal is significantly greater than zero at all angular scales ; if we assume a power - law form for @xmath12 , @xmath94 , we find , via a @xmath95 fit to the data , @xmath96 and @xmath97 .
+ however , as we have already seen , the id - sample is a hybrid mixture of low- and high - z sources , which means that this measurement is not particularly relevant .
more interesting information can be derived from the analysis of the clustering signal produced by the two different classes of objects , namely stellar and extended sources .
we then estimated @xmath12 for the stellar - like sources ( 535 objects with @xmath3 ) , and the resulting @xmath98 is shown in figure [ fig : wstars ] .
the clustering signal is dominated by large errors due to the small number of objects .
nevertheless the signal is positive at almost all angular scales up to @xmath99 .
the most interesting result is represented by the angular correlation function for the 1494 radio galaxies in the id - sample .
as illustrated by figure [ fig : wgals ] , in this case the signal is quite strong , showing once again that radio sources are more clustered than normal galaxies ( see also peacock & nicholson , 1991 ; cress et al . , 1996 ; loan , wall & lahav , 1997 ; magliocchetti et al . ,
1998 : magliocchetti et al . , 1999 ) .
in fact , if we parameterize @xmath12 as a power - law : @xmath100 , we find - via a @xmath95 fit to the data - @xmath101 and @xmath102 , with a value for the slope close to those obtained for samples of early - type galaxies only , known to be more strongly clustered than other galaxy populations ( see e.g. maddox et al . , 1990c and loveday et al . , 1995 ) .
note that this result also agrees with the observational evidence for agn - fuelled radio sources to be hosted by elliptical / early - type galaxies ( see e.g. mclure et al . , 1999 ) .
+ since this last measurement has been obtained for a homogeneous sample of sources with a fairly well determined redshift distribution , we can predict the clustering for a range of models to be directly compared with the data .
the first step in our model is to calculate the mass - mass correlation function @xmath103 .
we use the analytical form derived using the peacock & dodds ( 1996 ) approach , extended to z @xmath104 by matarrese et al .
this starts with a primordial cdm power spectrum ( cobe normalized in our case ; bunn & white , 1997 ) and writes the non - linear part of the power - spectrum @xmath105 as a function of its linear part @xmath106 , where @xmath107 and @xmath108 are the linear and non - linear wave - number respectively ( see magliocchetti et al .
( 2000a ) for a complete derivation of the various quantities ) .
the predicted angular correlation function @xmath109 can then be obtained from @xmath103 using the relativistic limber equation ( peebles , 1980 ) @xmath110 ^ 2 } , \label{eq : limber}\end{aligned}\ ] ] where @xmath111 is the comoving coordinate , @xmath112 is related to the spatial distance @xmath113 between two sources separated by an angle @xmath89 ( in the small angle approximation ) via @xmath114 ( @xmath115 is the curvature correction ) , @xmath116 is the redshift distribution for the sources under consideration , @xmath117 is the maximum redshift of the sample and @xmath118 is the bias function which determines the way radio sources trace the mass distribution .
note that , by imposing a @xmath117 we have assumed that the sample completeness is a step function , which immediately drops to zero beyond the redshift where the sample starts to become incomplete ( 0.3 in our case ) .
we use the @xmath116 obtained by ma2000 ( fig .
this is based on redshift measurements for 76 radio sources selected from both the first survey and the phoenix survey ( hopkins et al .
, 1998 ; georgakakis et al . , 1999 )
these samples were selected at a limit of 1 mjy and include redshift measurements up to z @xmath119 .
a little more attention has to be devoted to the issue of bias and its evolution .
of the various models presented in literature , the so - called _ merging model _
( mo & white , 1996 ; matarrese et al . , 1997 ; moscardini et al . , 1998 ; martini & weinberg , 2001 ) has the most appealing physical motivation : dark matter halos simply undergo dissipative collapse ( i.e. merging ) .
however this model has been shown to not fit the clustering data at low - redshifts ( see e.g. baugh et al . , 1999 ; magliocchetti et al .
, 1999 ; magliocchetti et al . , 2000a to mention just a few ) .
the reason for the discrepancy may be that the merging rate of objects is very high for z @xmath120 but becomes negligible for z @xmath121 .
+ a better bias model to describe the data in this situation is the so - called _ galaxy conserving model _ ( fry , 1996 ) .
this model assumes that galaxies form at some epoch and then the clustering pattern simply evolves as the galaxies follow the continuity equation without losing their identity .
it can be shown ( nusser & davis , 1994 ; fry , 1996 ) that bias for such galaxies evolves as @xmath122 where @xmath123 is the linear growth rate for clustering , and @xmath124 is the bias at the present epoch , @xmath125 , where @xmath126 is the rms fluctuation amplitude in a sphere of radius 8 h@xmath127mpc .
we calculate @xmath128 assuming a correlation radius @xmath129 h@xmath127mpc for radio galaxies ( peacock & nicholson , 1991 ) .
we calculate @xmath130 using a cobe - normalized @xmath131cdm model with @xmath22 , @xmath23 , @xmath24 .
the resulting prediction for @xmath109 is presented in figure [ fig : wgals ] .
the match with the data is extremely good , especially when one considers that there are no free parameters in the model .
this gives us confidence on our understanding of the processes regulating clustering in general and in particular the one of radio sources , at least in the nearby universe .
we have presented an analysis of the properties of optical counterparts of radio sources at the 1 mjy level .
the optical identifications have been obtained by matching together objects from the first and apm surveys over the region of the sky @xmath0 , @xmath1 .
we found 4075 radio objects with fluxes greater than 1 mjy to be identified in the apm catalogue for @xmath40 within a matching radius of 2 arcsecs ( id - sample ) .
this corresponds to 16.7 per cent of the original radio sample .
a cut for @xmath50 = 21.5 has subsequently been applied to ensure a uniform optical sample : 3176 identifications are found brighter than this magnitude limit .
such a wide area sample allows detailed statistical studies on the photometry of radio sources .
furthermore , with the application of the methods introduced in maddox et al . , ( 1990a ) , we managed to divide radio sources with optical identifications and @xmath3 into radio galaxies and stellar objects ( mainly qso with some contamination from nearby stars ) .
we find the population of radio galaxies to be almost exclusively dominated by early - type galaxies , with radio - to - optical ratios in the blue band between @xmath132 and @xmath133 and very red colours ( up to @xmath134 ) .
starbursting objects make up only a negligible fraction of the sample .
qso tend to have @xmath135 and preferentially lie in the region @xmath136 of the colour - radio flux diagram at all radio fluxes .
since passive radio galaxies are known to be reliable standard candles , measurements of the apparent magnitude in the red band allowed us to divide the id - sample into two low - redshift and high - redshift catalogues , the first one - roughly complete to @xmath14 - made by radio galaxies and the second one including high - z qso . the completeness and homogeneity and defined redshift cut of the galaxy sample allowed a direct comparison between models for the angular correlation function and the data .
we find an extremely good match between observational results and models if we treat radio galaxies as highly biased tracers of the underlying dark matter distribution and if we assume a bias evolution with look - back time of the form introduced by fry ( 1996 ) .
this is in agreement with previous results obtained by magliocchetti et al .
( 1999 ) for the whole sample of first radio sources .
note that , as well as the results presented in this paper , such a large sample of radio sources with optical identifications forms an excellent basis for spectroscopic follow - up observations , which could , amongst others , fully determine the form of the redshift distribution of these objects at the mjy level out @xmath137 . we will tackle this issue in a future paper where we will present results from the 2df survey .
we would like to thank mike irwin for the help provided during the analysis of the apm data and gianfranco de zotti and annalisa celotti for very useful discussions and comments on the manuscript .
baugh c.m . ,
benson a.j . , cole s. , frenk c.s . ,
lacey c.g . , 1999 ,
mnras , 305 , 21 becker r.h . ,
white r.l . ,
helfand d.j . , 1995 ,
apj , 450 , 559 bock d. c - j . ,
large m.i . , sadler , e.m .
, 1999 , aj , 117 , 1578 bunn e.f . , white m. , 1997 , apj , 480 , 6 colless m. , 1999 , phil trans r soc lond .
a , 357 , 105 condon j.j .
, cotton w.d . , greisen e.w . ,
yin q.f . , perley r.a . , taylor g.b . , broderick j.j . , 1998 , aj , 115 , 1693 cress c.m . , helfand d.j . , becker r.h
. , gregg m.d . , white r.l . , 1996 , apj , 473 , 7 fry j.n . , 1996 ,
apj , 461 , l65 georgakakis a. , mobasher b. , cram l. , hopkins a. , lidman c , rowan - robinson m. , 1999 , mnras , 306 , 708 gruppioni c. , mignoli m. , zamorani g. , 1998 , mnras , 304 , 199 hamilton a.j.s .
, 1993 , apj , 417 , 19 helfand d.j . at al . , 1997 ,
aas , 190 , 4304 hine r.g . ,
longair m.s .
, 1979 , mnras , 188 , 111 hopkins a. , mobasher b. , cram l. , rowan - robinson m. 1998 , mnras , 296 , 839 loan a.j . , wall j.v . , lahav o. , 1997 , mnras , 286 , 994 loveday j. , maddox s.j . , efstathiou g. , peterson b.a . , 1995 ,
apj , 442,457 lu n.y .
, lyle hoffman g.l . ,
salpeter e.e .
, houck j.r . , 1996 , apjs , 103 , 331 maddox s.j . , efstathiou g. , sutherland w.j . , loveday j. , 1990c , mnras , 242 , 43p maddox s.j . , efstathiou g. , sutherland w.j . , loveday j. , 1990a , mnras , 243 , 692 maddox s.j . , efstathiou g. , sutherland w.j . , 1990b , mnras , 246 , 433 maddox s.j .
, efstathiou g. , sutherland w.j .
, 1996 , mnras , 283 , 1227 magliocchetti m. , maddox s.j . , lahav o. , wall j.v . , 1998 , mnras , 300 , 257 magliocchetti m. , maddox s.j .
, lahav o. , wall j.v . , 1999 , mnras , 309 , 943 magliocchetti m. , bagla j. , maddox s.j . , lahav o. , 2000a , mnras , 314 , 546 magliocchetti m. , maddox s.j . , wall j.v . ,
, cotter g. , 2000 , mnras , 318 , 1047 [ ma2000 ] magliocchetti m. , celotti a. , danese l. , 2001 , mnras , submitted martini p. , weinberg d.h . , 2001 , apj , 547 , 12 masci f.j .
, condon j.j . , barlow t.a . , lonsdale c.j . , xu c. , shupe d.l . , pevunova o. , cutri r. , 2001 , pasp , 113 , 10 matarrese s. , coles p. , lucchin f. , moscardini l. , 1997 , mnras , 286 , 115 mclure r.j . ,
kukula m.j .
, dunlop j.s . ,
baum s.a . ,
odea c.p . , hughes d.h . , 1999 ,
mnras , 308 , 377 mo h. , white s.d.m . , 1996 , mnras , 282 , 347 moscardini l. , coles p. , lucchin f. , matarrese s. , 1998 , mnras , 299 , 95 nusser a. , davis m. , 1994 , apj , 421 , l1 peacock j.a . , nicholson d. , 1991 , mnras , 253 , 307 peacock j.a . ,
dodds s.j . , 1996 , mnras , 267 , 1020 peebles p.j.e . , 1980 , the large - scale structure of the universe , princeton university press rixon g.t . , wall j.v
, benn c.r . , 1991 , mnras , 251 , 243 sadler e.m . , mcintyre v.j . ,
jackson c.a . , cannon r.d . , 1999 ,
pasa , 16 , 247 urry c.m . , padovani p. , 1995 ,
pasp , 107 , 803 white r.l . ,
becker r.h .
, helfand d.j . ,
gregg m.d . , 1997 ,
apj , 475 , 479 | we present an analysis of the properties of optical counterparts of radio sources down to 1 mjy .
optical identifications have been obtained by matching together objects from the apm and first surveys over the region @xmath0 , @xmath1 . selecting radio sources down to 1 mjy , and adopting a uniform optical limit of @xmath2 we find 3176 have a counterpart in the apm catalogue , corresponding to 13 per cent of the radio sample . for @xmath3 we can divide radio sources into resolved radio galaxies and stellar - like objects ( principally qso ) .
we find the population of radio galaxies to be mainly made by early - type galaxies with very red colours ( @xmath4 up to @xmath5 ) and a radio - to - optical ratio @xmath6 .
the contribution of starbursting objects is negligible . in general qso show @xmath7 and @xmath8 . on the basis of the @xmath9 magnitudes
, we estimate the sample of radio galaxies to be complete up to @xmath10
. we can therefore divide the whole sample of radio sources into a low - z and high - z population .
the low - z one includes the objects identified as galaxies in the apm survey , the high - z one includes sources either identified as qso or with no optical counterpart for @xmath11 .
we find that radio galaxies are strongly clustered and highly biased tracers of the underlying mass distribution .
models for the angular correlation function @xmath12 show good agreement with the observations if we assume a bias factor @xmath13 at @xmath14 .
galaxies : active - galaxies : starburst - cosmology observations -radio continuum galaxies |
in his seminal paper on rational homotopy theory @xcite , sullivan isolated an important topological property of spaces .
a connected , finite - type cw - complex @xmath1 is said to be _ formal _ if its minimal model is quasi - isomorphic to @xmath2 ; in particular , the rational homotopy type of @xmath1 is determined by its rational cohomology ring .
examples of formal spaces include compact khler manifolds @xcite and complements of complex hyperplane arrangements @xcite .
there is a related property of groups .
a finitely generated group @xmath3 is said to be _
@xmath0-formal _ if its malcev completion ( in the sense of quillen @xcite ) is a quadratic complete lie algebra ; see for instance @xcite for more details . if the first betti number of @xmath3 is @xmath4 or @xmath0 , then @xmath3 is @xmath0-formal
. other examples of @xmath0-formal groups include finitely generated artin groups @xcite , finitely presented bestvina - brady groups @xcite , as well as pure welded braid groups @xcite . as noted in @xcite , fundamental groups of formal spaces are @xmath0-formal ; thus , khler groups and arrangement groups belong to this class . in general ,
smooth complex quasi - projective varieties are not formal , but fundamental groups of complements of complex projective hypersurfaces are always @xmath0-formal @xcite .
a well - known obstruction to formality is provided by the massey products .
if @xmath1 is a formal space , then all such products ( of order @xmath5 and higher ) in @xmath6 vanish , up to indeterminacy .
similarly , if @xmath3 is a @xmath0-formal group , all higher - order massey products in @xmath7 vanish . in this note
, we develop a new formality criterion , based on the interplay between three ingredients .
the first ingredient is a close connection between the algebraic monodromy action on the homology of an infinite cyclic galois cover , and the aomoto
betti numbers associated to the cohomology ring of the base .
this connection , established in @xcite and adapted to our context in section [ sect : mono ] , requires no formality assumption .
a second general fact , noted by dwyer and fried in @xcite and further refined in @xcite , is that the homological finiteness properties ( over @xmath8 ) of a free abelian galois cover are controlled by the jump loci for the twisted betti numbers of the base , with rank @xmath0 complex coefficients .
finally , @xmath0-formality comes into play , by allowing an intimate relation between the aomoto
betti numbers and the twisted betti numbers in degree one .
this relation is based on a key result from our joint work with a. dimca , @xcite . for a group @xmath3 ,
denote by @xmath9 its betti numbers .
our main result ( proved in section [ sect : proof thm ] ) reads as follows .
[ thm : main ] suppose given two group extensions , @xmath10 & n\ar[r ] & g \ar^{\nu}[r ] & { \mathbb{z}}\ar[r]&1 } , \tag{$\xi_{\nu}$ } \label{xinu}\\ & \xymatrix{1\ar[r ] & k\ar[r ] & \pi \ar^{\chi}[r]&g \ar[r ] & 1 } , \tag{$\xi_{\chi}$ } \label{xichi}\end{aligned}\ ] ] such that [ hyp1 ] @xmath11 and @xmath12 ; [ hyp2 ] @xmath13 is finitely presented and @xmath0-formal .
then , if the monodromy action of @xmath14 on @xmath15 has eigenvalue @xmath0 , the corresponding jordan blocks are all of size @xmath0 .
the @xmath0-formality hypothesis is crucial here , even in the particular case ( treated in proposition [ prop = special ] ) when @xmath16 and @xmath17 .
for example , the heisenberg group of integral , unipotent @xmath18 matrices is not @xmath0-formal , yet it arises as an extension of @xmath19 by @xmath20 , with monodromy @xmath21 .
we illustrate theorem [ thm : main ] with a variety of geometric applications . to start with ,
we derive in section [ sect : bundles ] the following consequence .
[ thm : fgm intro ] let @xmath22 be a closed , connected , smooth manifold , and let @xmath23 be a diffeomorphism , with mapping torus @xmath24 .
let @xmath25 be a locally trivial smooth fibration , with closed , connected fiber @xmath26 .
if the operator @xmath27 has eigenvalue @xmath0 , with a jordan block of size greater than @xmath0 , then @xmath28 is not a formal space .
this theorem extends a result of fernndez , gray , and morgan @xcite , valid for bundles with fiber @xmath29 .
it is unclear to us whether their approach which relies on massey products as an obstruction to formality still works for a general fiber @xmath26 .
one of the inspirations for this work was the following result of eisenbud and neumann @xcite : if @xmath30 is a fibered graph multilink , then its algebraic monodromy has only @xmath31 jordan blocks for the eigenvalue @xmath0 .
as noted by these authors , though , there do exist fibered links in @xmath32 for which larger jordan blocks do occur . necessarily , then , the complement of such a link is not formal .
when combined with a result from @xcite on the bieri - neumann - strebel invariant , our approach yields a formality criterion for closed , orientable @xmath5-manifolds which fiber over the circle , see corollary [ cor : jordan blocks ] .
given a reduced polynomial function @xmath33 with @xmath34 and connected generic fiber , there are two standard fibrations associated with it .
one is the milnor fibration , @xmath35 , where @xmath36 is the link at the origin , for @xmath37 small . as shown by durfee and hain @xcite
, the link complement is formal .
theorem [ thm : main ] allows us then to recover the basic fact that the algebraic monodromy has no jordan blocks of size greater than @xmath0 for the eigenvalue @xmath38 .
the other fibration , @xmath39 , is obtained by base localization . "
as noted in @xcite , the algebraic monodromy of this fibration can have larger jordan blocks for @xmath38 , see example [ ex : dimca ] .
it follows that the space @xmath40 may be non - formal .
this observation is rather surprising , in view of the well - known fact that , upon localizing near @xmath41 the complement of the curve @xmath42 , one obtains a space homotopy equivalent to the formal space @xmath43 .
fix a field @xmath45 .
the group - ring @xmath44 may be identified with the ring of laurent polynomials in one variable , @xmath46 $ ] .
since @xmath45 is a field , the ring @xmath47 is a principal ideal domain .
thus , every finitely - generated @xmath47-module @xmath28 decomposes into a finite direct sum of copies of @xmath47 ( the free part ) , with torsion modules of the form @xmath48 , where @xmath49 is an irreducible polynomial in @xmath47 , and @xmath50 is the size of the corresponding jordan block .
define @xmath51 to be the _
@xmath49-primary part _ of @xmath28 . clearly , the module @xmath52 consists of those elements of @xmath28 annihilated by some power of @xmath49 .
hence , the following naturality property holds : every @xmath47-linear surjection , @xmath53 , between finitely generated torsion @xmath47-modules restricts to a @xmath47-linear surjection , @xmath54 , for all irreducible polynomials @xmath55 .
let @xmath1 be a connected , finite - type cw - complex .
without loss of generality , we may assume @xmath1 has a single @xmath4-cell , say @xmath56 .
let @xmath57 be the fundamental group .
given an epimorphism @xmath58 , let @xmath59 be the galois cover corresponding to @xmath60
. denote by @xmath61 the group ring of @xmath19 , viewed as a ( right ) module over @xmath62 via extension of scalars : @xmath63 , for @xmath64 and @xmath65 . by shapiro s lemma , @xmath66 , as modules over @xmath67
$ ] . since @xmath1 is of finite - type , @xmath68 is a finitely generated @xmath47-module , for each @xmath69 .
denote by @xmath70 the class in @xmath71 determined by @xmath72 .
let @xmath73 be the induced homomorphism .
the corresponding cohomology class , @xmath74 , is the image of @xmath70 under the coefficient homomorphism @xmath75 .
we then have @xmath76 indeed , by obstruction theory , there is a map @xmath77 and a class @xmath78 such that @xmath79 .
hence , @xmath80 , and the claim follows by naturality of cup products with respect to coefficient homomorphisms . as a consequence , left - multiplication by @xmath81 turns the cohomology ring @xmath82 into a cochain complex , @xmath83 & h^1(x,{\bbbk } ) \ar^{\nu_{{\bbbk}}}[r ] & h^2(x,{\bbbk } ) \ar[r ] & \cdots } \ ] ] which we call the _ aomoto complex _ of @xmath84 , with respect to @xmath81 .
the _ aomoto
betti numbers _ of @xmath1 , with respect to the cohomology class @xmath74 , are defined as @xmath85 in @xcite , we defined a spectral sequence , starting at @xmath86 , and converging to @xmath87 . in the process , we investigated the monodromy action of @xmath19 on @xmath88 , relating the triviality of the action to the exactness of the aomoto complex defined by @xmath74 , as follows .
[ thm : trivial action ] for each @xmath89 , the following are equivalent : 1 .
[ m1 ] for each @xmath90 , the @xmath44-module @xmath68 is finite - dimensional over @xmath45 , and contains no @xmath91-primary jordan blocks of size greater than @xmath0 .
2 . [ m2 ] @xmath92 .
similar considerations apply to groups .
let @xmath3 be a group , and let @xmath58 be an epimorphism with kernel @xmath93 .
let @xmath94 be a classifying space for @xmath3 .
then the monodromy action of @xmath46 $ ] on @xmath95 is induced by conjugation : the generator @xmath96 acts on @xmath97 by @xmath98 , where @xmath64 is such that @xmath99 and @xmath100 is the corresponding inner automorphism .
clearly , the aomoto complex depends in this case only on @xmath3 and @xmath81 ; thus , we may denote @xmath101 by @xmath102 . in the sequel
, we will need the following group - theoretic version of theorem [ thm : trivial action ] .
[ cor = gmonotest ] let @xmath103 be a group extension , with @xmath3 finitely presented .
then , the following are equivalent : 1 .
[ mono1 ] the @xmath44-module @xmath97 is finite - dimensional over @xmath45 , and contains no @xmath91-primary jordan block of size greater than @xmath0 .
[ mono2 ] @xmath104 .
let @xmath1 be a finite presentation @xmath105-complex for @xmath3 .
note that @xmath106 applying theorem [ thm : trivial action ] to @xmath1 , for @xmath107 , we find that @xmath108 is finite - dimensional over @xmath45 , with trivial monodromy action on the @xmath91-primary part , if and only if @xmath109 . but
@xmath110 and @xmath111 , and we are done .
this section is devoted to proving theorem [ thm : main ] from the introduction .
we first prove the special case in which @xmath16 and @xmath17 , and then reduce the general case to the special one , by means of two lemmas .
we start by recalling a result ( corollary 6.7 from @xcite ) , relating the homological finiteness properties of certain normal subgroups of a @xmath0-formal group to the vanishing of the corresponding aomoto
betti numbers .
[ thm : b1ker ] let @xmath3 be a finitely generated , @xmath0-formal group .
suppose @xmath58 is an epimorphism , with kernel @xmath93 .
then @xmath112 if and only if @xmath113 .
we are now ready to prove an important particular case of theorem [ thm : main ] .
[ prop = special ] let @xmath3 be a finitely presented , @xmath0-formal group , and let @xmath114 be an epimorphism , with kernel @xmath93 . if @xmath11 , then the eigenvalue @xmath0 of the monodromy action of @xmath14 on @xmath15 has only @xmath31 jordan blocks . since @xmath3 is @xmath0-formal and @xmath115 , theorem [ thm : b1ker ] insures that @xmath113 .
the conclusion follows from corollary [ cor = gmonotest ] , for @xmath116 . before proceeding ,
recall the extensions and from section [ subsec : intro2 ] .
set @xmath117 .
we then have an extension @xmath118 & \gamma \ar[r ] & \pi \ar^{\nu\circ \chi}[r ] & { \mathbb{z}}\ar[r]&1}. \tag{$\xi_{\nu\chi}$ } \label{xinuchi}\ ] ] pulling back to @xmath93 gives rise to another extension , @xmath118 & k \ar[r ] & \gamma \ar^{\eta}[r ] & n \ar[r]&1}. \tag{$\xi_{\eta}$ } \label{xieta}\ ] ] the first lemma is a standard application of the hochschild serre spectral sequence of a group extension , as recorded for instance in brown @xcite . for the sake of completeness
, we include a proof .
[ lem = hs ] if both @xmath119 and @xmath120 are finite , so is @xmath121 .
consider the spectral sequence of extension , with @xmath122-trivial coefficients @xmath8 : @xmath123 clearly , @xmath124 is a quotient of @xmath125 , and @xmath126 .
thus , both @xmath127 and @xmath128 are finite - dimensional over @xmath8 .
hence , @xmath129 is also finite - dimensional .
the second lemma shows that the jordan blocks of size greater than @xmath0 for the eigenvalue @xmath0 propagate from to . as usual , write @xmath130 $ ] .
[ lem = propa ] if the @xmath47-module @xmath131 is trivial , then so is the module @xmath132 .
pick @xmath64 such that @xmath99 , and then @xmath133 such that @xmath134 .
clearly , @xmath135 .
it follows that the surjection @xmath136 is @xmath47-linear with respect to the corresponding monodromy actions .
now , since both @xmath137 and @xmath15 are finitely generated , torsion @xmath47-modules , we infer from the discussion in section [ subsec : kz mod ] that @xmath138 induces a @xmath47-linear surjection on @xmath91-primary parts , @xmath139 the claim follows . the finite - dimensionality assumption , together with lemma [ lem = hs ]
imply that @xmath140 .
suppose that the monodromy action on @xmath15 coming from extension has eigenvalue @xmath0 , with a jordan block of size greater than @xmath0 .
by lemma [ lem = propa ] , the same holds for the monodromy action on @xmath137 , coming from extension .
finally , we use the formality assumption . from proposition
[ prop = special ] , applied to extension , we deduce that all jordan blocks of the monodromy action on @xmath137 have size @xmath0 .
but this is a contradiction , and the proof is finished .
let @xmath22 be a closed , connected , smooth manifold , and let @xmath23 be a diffeomorphism , with mapping torus @xmath24 .
let @xmath25 be a locally trivial smooth fibration , with closed , connected fiber @xmath26 .
theorem [ thm : fgm intro ] follows from the next proposition . [
prop : fgm1 ] if the operator @xmath27 has eigenvalue @xmath0 , with a jordan block of size greater than @xmath0 , then @xmath141 is not a @xmath0-formal group . the homotopy exact sequence of the mapping torus fibration , @xmath142 , yields an extension of fundamental groups , @xmath143 .
clearly , @xmath144 , and the monodromy action of @xmath14 on @xmath145 coincides with @xmath146 .
similarly , the fibration @xmath147 yields an epimorphism on fundamental groups , @xmath148 .
the group @xmath149 is a quotient of @xmath150 ; hence , @xmath151 .
plainly , the group @xmath152 is finitely presented . given the assumption on @xmath146 , theorem [ thm : main ] implies that @xmath13 is not @xmath0-formal . for a class @xmath153 ,
denote by @xmath154 its image in @xmath155 .
proposition [ prop : fgm1 ] gives a substantial extension of the following result of fernndez , gray , and morgan , contained in proposition 9 and note 10 from @xcite .
[ thm : fgm ] let @xmath156 be an orientation - preserving diffeomorphism of a closed , connected , oriented manifold @xmath22 .
suppose there is an element @xmath157 with @xmath158 , such that @xmath159 . if @xmath160 is a circle bundle , then @xmath161 is not formal .
[ rem : fgm ] when @xmath22 is a symplectic manifold , and @xmath156 is a symplectic diffeomorphism , the circle bundle @xmath160 from above can be chosen so that @xmath161 carries a symplectic structure ( * ? ? ?
* theorem 4 ) . yet , by theorem [ thm : fgm ] and the main result of @xcite , the manifold @xmath161 admits no khler metric . in a similar vein , the particular case of theorem [ thm : main ] proved in section [ sect : proof thm ] has the following geometric counterpart .
[ prop:1jordan ] let @xmath162 be a smooth fibration , with monodromy @xmath163
. assume the fiber @xmath26 is connected and has the homotopy type of a cw - complex with finite @xmath105-skeleton .
if the algebraic monodromy , @xmath164 , has eigenvalue @xmath0 , with a jordan block of size greater than @xmath0 , then the group @xmath165 is not @xmath0-formal .
let @xmath143 be the extension of fundamental groups determined by the fibration . from the hypothesis on the fiber @xmath26 , the group @xmath3 is finitely presented , and @xmath115 .
the conclusion follows from proposition [ prop = special ] .
the question of deciding whether a link complement is formal has a long history .
in fact , the higher - order products were defined by w. massey precisely for detecting higher - order linking phenomena for links such as the borromean rings .
the following example shows how the non - formality of a fibered link complement can be detected by the algebraic monodromy .
it also shows that the @xmath0-formality hypothesis is essential in theorem [ thm : main ] .
-component link , width=172,height=172 ] [ ex : en ] let @xmath30 be the @xmath5-component link in @xmath32 depicted in figure [ fig : en link ] . as explained in ( * ? ? ?
* example 11.6 ) , the link exterior , @xmath166 , fibers over @xmath167 , with fiber @xmath26 a connected , compact surface , and with algebraic monodromy having two jordan blocks of type @xmath168 . by corollary [ prop:1jordan ] ,
the group @xmath169 is not @xmath0-formal .
turning now to closed @xmath5-manifolds , let us first recall the following result from @xcite , based on the interplay between the bieri
strebel invariant , @xmath170 , and the ( first ) resonance variety , @xmath171 , of a @xmath0-formal group @xmath3 . [ prop:3mfd ]
let @xmath28 be a closed , orientable @xmath5-manifold which fibers over the circle . if @xmath172 is even , then @xmath173 is not @xmath0-formal . combining this result with corollary [ prop:1jordan ] puts strong restrictions on the algebraic monodromy of a fibration @xmath174 .
[ cor : jordan blocks ] let @xmath28 be a closed , orientable @xmath5-manifold with @xmath0-formal fundamental group .
suppose @xmath28 fibers over the circle , and the algebraic monodromy has @xmath0 as an eigenvalue . then , there are an even number of @xmath31 jordan blocks for this eigenvalue , and no higher size jordan blocks . by corollary [ prop:1jordan ] ,
the algebraic monodromy has only @xmath31 jordan blocks for the eigenvalue @xmath0 .
let @xmath175 be the number of such blocks . from the wang sequence of the fibration
, we deduce that @xmath176 . by proposition [ prop:3mfd ]
, @xmath175 must be even .
let @xmath177 be a reduced polynomial map .
the link of @xmath49 at @xmath178 , denoted @xmath36 , is the intersection of the hypersurface @xmath179 with the sphere @xmath180 , for sufficiently small @xmath37 .
as shown by milnor in @xcite , the map @xmath181 is a smooth fibration .
moreover , the milnor fiber @xmath182 has the homotopy type of an @xmath183-dimensional finite cw - complex .
let @xmath184 be the monodromy map .
as is well known , the monodromy operator , @xmath185 , has only roots of unity as eigenvalues , with corresponding jordan blocks of size at most @xmath186 , see for instance @xcite . in the case
when @xmath49 has an isolated singularity at @xmath178 , more can be said .
first of all , as shown by milnor , the fiber @xmath26 is @xmath187-connected .
moreover , @xmath188 has jordan blocks of size at most @xmath183 for the eigenvalue @xmath0 , see for instance @xcite .
particularly tractable is the case of a reduced polynomial map @xmath189 with @xmath34 , when @xmath178 is necessarily an isolated singularity . as shown by durfee and hain in (
* theorem 4.2 ) , the link complement , @xmath190 , is a formal space .
corollary [ prop:1jordan ] , then , explains from a purely topological point of view why the jordan blocks of @xmath191 corresponding to the eigenvalue @xmath0 must all have size @xmath0 .
let @xmath189 be a polynomial map with connected generic fiber .
for @xmath37 , there is a fibration @xmath192 induced by @xmath49 , where @xmath193 ; the topological type of this fibration is independent of small @xmath194 .
now let @xmath26 be the generic fiber of @xmath49 , and let @xmath184 be the monodromy of the fibration .
note that @xmath26 is a finite cw - complex , up to homotopy ; see for instance @xcite .
it was shown in @xcite that @xmath191 can have a jordan block of size @xmath105 for the eigenvalue @xmath0 .
the following concrete examples were communicated to us by alex dimca .
[ ex : dimca ] consider the polynomial @xmath195 .
in this case , the central fiber @xmath196 has just a node singularity . using theorem 1 or corollary 6 from @xcite
, we see that @xmath191 has a jordan block of type @xmath168 .
other examples are suitable translates of the polynomials @xmath197 and @xmath198 introduced in ( * ? ? ?
* example 4 ) , such that the central fiber corresponds to the kodaira types @xmath199 or @xmath200 . in this case ,
the central fibers are smooth , but the non - triviality of the monodromy is due to singularities at infinity .
the polynomial @xmath197 ( known as the brianon polynomial ) is also considered in example 5 from @xcite . for
a non - constant polynomial map , @xmath202 with @xmath34 , let @xmath203 be the corresponding plane algebraic curve , and @xmath204 its complement .
in recent work , cogolludo agustin and matei @xcite , and macinic @xcite , proved , by different methods , that @xmath205 is a formal space . due to the local conic structure of affine varieties
( see for instance @xcite ) , for small enough @xmath206 , the intersection @xmath207 has the same homotopy type as the link complement @xmath208 . in conclusion ,
the formality property of @xmath205 is preserved by localization near @xmath209 , but it may be destroyed by base localization , that is , when replacing the complement @xmath205 by @xmath210 e. artal bartolo , p. cassou - nogus , a. dimca , _ sur la topologie des polynmes complexes _ , in : singularities ( oberwolfach , 1996 ) , pp .
317343 , progress in math .
162 , birkhuser , basel , 1998 .
http://www.ams.org/mathscinet-getitem?mr=1652480[mr1652480 ] b. berceanu , s. papadima , _ universal representations of braid and braid - permutation groups _ ,
j. knot theory ramifications * 18 * ( 2009 ) , no . 7 , 9991019 .
e. brieskorn , _ sur les groupes de tresses _ , in : sminaire bourbaki , 1971/72 , lect .
notes in math .
* 317 * , springer - verlag , 1973 , pp .
http://www.ams.org/mathscinet-getitem?mr=0422674[mr0422674 ] k. s. brown , _ cohomology of groups _ , grad .
texts in math .
87 , springer - verlag , new york - berlin , 1982 .
http://www.ams.org/mathscinet-getitem?mr=0672956[mr0672956 ] j. i. cogolludo - agustin , d. matei , _ cohomology algebra of plane curves , weak combinatorial type , and formality _ , preprint ` arxiv:0711.1951 ` .
p. deligne , p. griffiths , j. morgan , d. sullivan , _ real homotopy theory of khler manifolds _ , invent . math . * 29 * ( 1975 ) , no . 3 , 245274 .
a. dimca , _ singularities and topology of hypersurfaces _ , universitext , springer - verlag , new york , 1992 . http://www.ams.org/mathscinet-getitem?mr=1194180[mr1194180 ] a. dimca , _ monodromy at infinity for polynomials of two variables _ , j. algebraic geom . * 7 * ( 1998 ) , no . 4 , 771779 . http://www.ams.org/mathscinet-getitem?mr=1642749[mr1642749 ]
a. dimca , s. papadima , a. suciu , _ topology and geometry of cohomology jump loci _ , duke math .
journal * 148 * ( 2009 ) , no . 3 , 405457 . a. durfee , r. hain , _
mixed hodge structures on the homotopy of links _
ann . * 280 * ( 1988 ) , no . 1 , 6983 ; http://www.ams.org/mathscinet-getitem?mr=0928298[mr0928298 ] w. g. dwyer , d. fried , _ homology of free abelian covers .
london math .
* 19 * ( 1987 ) , no . 4 , 350352 .
http://www.ams.org/mathscinet-getitem?mr=0887774[mr0887774 ] w. ebeling , _ monodromy _ , in : singularities and computer algebra , pp . 129155 , london math .
lecture note ser .
324 , cambridge univ .
press , cambridge , 2006 .
d. eisenbud , w. neumann , _ three - dimensional link theory and invariants of plane curve singularities _ , annals of math .
studies , vol .
110 , princeton university press , princeton , nj , 1985 .
http://www.ams.org/mathscinet-getitem?mr=0817982[mr0817982 ] m. fernndez , a. gray , j. morgan , _ compact symplectic manifolds with free circle actions , and massey products _ , michigan math .
j. * 38 * ( 1991 ) , no . 2 , 271283 .
http://www.ams.org/mathscinet-getitem?mr=1098863[mr1098863 ] m. kapovich , j. millson , _ on representation varieties of artin groups , projective arrangements and the fundamental groups of smooth complex algebraic varieties _ , inst .
hautes tudes sci .
publ . math .
* 88 * ( 1998 ) , 595 .
http://www.ams.org/mathscinet-getitem?mr=1733326[mr1733326 ] t. kohno , _ on the holonomy lie algebra and the nilpotent completion of the fundamental group of the complement of hypersurfaces _ , nagoya math .
j. * 92 * ( 1983 ) , 2137 .
a. macinic , _ cohomology rings and formality properties of nilpotent groups _ ,
preprint ` arxiv:0801.4847 ` .
f. michel , c. weber , _ on the monodromies of a polynomial map from @xmath211 to @xmath8 _ , topology * 40 * ( 2001 ) , no . 6 , 12171240 .
http://www.ams.org/mathscinet-getitem?mr=1867243[mr1867243 ] j. w. milnor , _ singular points of complex hypersurfaces _ , ann . of math .
studies , no .
61 princeton univ . press ,
princeton , n.j . , 1968 .
http://www.ams.org/mathscinet-getitem?mr=0239612[mr0239612 ] s. papadima , a. suciu , _
chen lie algebras _ , international math .
research notices * 2004 * ( 2004 ) , no . 21 , 10571086 .
http://www.ams.org/mathscinet-getitem?mr=2037049[mr2037049 ] s. papadima , a. suciu , _ algebraic invariants for bestvina - brady groups _
, j. london math .
society , * 76 * ( 2007 ) , no . 2 , 273292 .
http://www.ams.org/mathscinet-getitem?mr=2363416[mr2363416 ] s. papadima , a. suciu , _ the spectral sequence of an equivariant chain complex and homology with local coefficients _ , preprint ` arxiv:0708.4262 ` , to appear in trans .
d. quillen , _ rational homotopy theory _ , ann . of math .
* 90 * ( 1969 ) , no . 2 , 205295 .
d. sullivan , _ infinitesimal computations in topology _ , inst .
hautes tudes sci .
publ . math . *
47 * ( 1977 ) , 269331 . | given a fibration over the circle , we relate the eigenspace decomposition of the algebraic monodromy , the homological finiteness properties of the fiber , and the formality properties of the total space . in the process , we prove a more general result about iterated group extensions . as an application
, we obtain new criteria for formality of spaces , and @xmath0-formality of groups , illustrated by bundle constructions and various examples from low - dimensional topology and singularity theory .
partially supported by cncsis grant id-1189/2009 - 2011 of the romanian ministry of education and research ] partially supported by nsa grant h98230 - 09 - 1 - 0012 , and an enhance grant from northeastern university ] |
the scaling law ( [ n_to_s ] ) is available when no domains are embedded inside the circular domain .
figure [ fig_s_out ] shows a schematic of a @xmath14-domain with area @xmath145 in which several domains are embedded inside the outer wall ( @xmath146 ) . here
, the contributions from the embedded domains are shown to be statistically negligible .
this claim is fulfilled when the following approximation is valid : @xmath147 here , @xmath148 is the number of vortices that exist inside the outer domain ( @xmath145 ) .
the total vorticity @xmath149 along the outer wall ( @xmath146 ) is written as @xmath150 with @xmath151 .
similarly , the vorticity @xmath152 on the wall of @xmath153 reduces to @xmath154 with the vorticity @xmath155 on the wall of @xmath156 , and so forth . as a result , @xmath157 , where @xmath158 is the sum of the circulation number @xmath159 of all @xmath160-domains inside the outer wall . in this way , eq .
( [ gamma_out ] ) reduces to @xmath161 the approximation ( [ s_gamma_nv ] ) is valid for @xmath162 , consider the total area @xmath163 , which is the area enclosed by the outer wall @xmath146 .
we write @xmath164 , where @xmath165 is the sum of the areas of all embedded domains .
the area @xmath166 of each embedded domain should be smaller than or at most on the order @xmath145 , @xmath167 ; otherwise , the outer domain violates the scaling relation ( [ tau_d ] ) with @xmath102 .
according to this argument , we have @xmath168 .
then , we obtain @xmath169 with @xmath11 and @xmath170 . therefore , we may write @xmath171 similarly , we can evaluate @xmath172 with @xmath173 . by using eq .
( [ n_to_s ] ) and eq .
( [ stotal_sout ] ) , we obtain @xmath174 this is equivalent to the inequality ( [ nvsout_nsigma ] ) for @xmath175 . for example , consider the case of @xmath176 . then , the inequality ( [ nvsout_nsigma ] ) is applicable to the regime @xmath177 .
we may demand this inequality for the entire range of the microscopic regime ( @xmath178 ) in the context of our dynamic scaling hypothesis , where domain structures should be statistically similar in the regime with a unique scaling behavior . | the domain - area distribution in the phase transition dynamics of @xmath0 symmetry breaking is studied for quasi - two - dimensional multi - component superfluids .
the distribution is divided into microscopic and macroscopic regimes with distinct power - law exponents .
the macroscopic regime universally exhibits fischer s law in percolation theory , while the microscopic regime depends on the microscopic dynamics of the system .
numerical experiments of segregating superfluids reveal that the exponent in the microscopic regime reaches its theoretical upper limit in the presence of the circular vortex sheet described by quantum hydrodynamics .
an analogy to quantum turbulence and the applicability to different muti - component superfluids are also discussed . in the phase transition dynamics of spontaneous symmetry breaking ( ssb ) , a number of topological defects are nucleated , forming a complicated network or texture in order parameter fields . this type of phenomenon can occur universally in systems ranging from condensed matter to cosmology and high energy physics @xcite .
the dynamic - scaling law in phase ordering kinetics hypothesizes that the growth of order parameter fields preserves the statistical similarity of the spatial patterns @xcite .
this hypothesis has been accepted empirically by observing that the structure factors or correlation functions of the fields fit a universal function after rescaling length by the characteristic size @xmath1 , which obeys a power law @xmath2 with the dynamic exponent @xmath3 .
the application of percolation theory @xcite to ssb development has recently attracted attention@xcite and has become an important problem in statistical physics @xcite . since percolation theory reveals different statistical aspects of the phase ordering kinetics , such studies could lead to a greater understanding of the highly non - equilibrium physics of ssb development .
for example , the most fundamental problem of phase ordering kinetics @xcite is two - dimensional @xmath0 symmetry breaking in which the order parameter is a real scalar field and the defect is a linear object called a domain wall .
one significant prediction of percolation theory is that the time development of the number distribution @xmath4 of domains of area @xmath5 obeys a universal power law @xmath6 with the fisher exponent @xmath7 of two - dimensional percolation @xcite .
there has been growing interest in the phase ordering kinetics of atomic bose - einstein condensates ( becs ) , and the dynamic - scaling law in superfluid systems has been investigated theoretically in different situations @xcite .
recently , percolation theory has been applied to the segregation of binary becs in quasi - two - dimensions , and the dynamic finite - size - scaling analysis revealed that the domain structures preserve the percolation criticality during @xmath0 symmetry breaking with the percolation threshold @xmath8 @xcite .
recently , the domain - area distribution in segregating superfluids was shown to obey fisher s power law for large - scale structures@xcite .
interestingly , the distribution of domains smaller than the characteristic domain size @xmath1 in the superfluid systems shows anomalous behavior different from that observed in conventional coarsening systems of nonconserved and conserved fields @xcite .
these results suggest that the dynamic - scaling hypothesis holds even when the quantum effect becomes important in superfluid systems since the small - scale structure strongly depends on the `` microscopic '' dynamics of the domain wall in the system under consideration .
this anomalous behavior of smaller domains indicates the existence of a different dynamic - scaling regime that reflects the quantum hydrodynamics , namely , `` microscopic '' dynamics in quantum fluids .
can such a scaling regime firmly coexist with the universal scaling regime of percolation theory in superfluid coarsening systems ?
if so , the critical exponent and the effect of the microscopic nature on the scaling behavior should be important for deeper understanding of the physics of ssb development . to answer these questions
, we demonstrate the dynamic - scaling analysis according to the combined theory of phase ordering kinetics and percolation , called _ phase ordering percolation_. the theoretical conjecture presented here is summarized in fig .
[ fig_loglog ] .
assuming an asymptotic form of the domain - area distribution , this theoretical analysis is universally applicable to different coarsening systems .
the domain - area distribution has two distinct scaling regimes , namely the microscopic and macroscopic regimes , which are characterized by the exponents @xmath9 and @xmath10 , respectively .
numerical experiments of binary becs reveal that the microscopic exponent @xmath9 of this system takes the theoretical upper limit @xmath11 , which is quite different from those of conventional coarsening systems .
this anomalous behavior is connected to the quantum - fluid effect of a circular vortex sheet by finding a characteristic scaling law derived from the quantum hydrodynamics . finally , an analogy to hierarchy in quantum turbulence and the applicability of this scaling theory to different multi - component superfluids will be discussed .
we first formulate a generalized model for evaluating the scaling behavior of phase ordering percolation before discussing superfluid systems . consider the time evolution of domain - area distribution @xmath12 in a coarsening system that undergoes @xmath0 symmetry breaking in two dimensions .
there are two kinds of domains , namely @xmath13- and @xmath14-domains .
we write the number of @xmath13- or @xmath14-domains , which have areas between @xmath5 and @xmath15 , divided by the system area @xmath16 at time @xmath17 as @xmath18 . according to the theory of phase ordering percolation ,
@xmath4 is described by the non - dimensional universal function @xmath19 of @xmath20 with characteristic domain area @xmath21 and mean domain size @xmath1 : @xmath22 this is another expression of the dynamic - scaling hypothesis in the sense that the hypothesis is conventionally described by the structure factor or the correlation function @xcite .
the dynamic - scaling law ( [ rhos ] ) has been experimentally @xcite and numerically @xcite confirmed in different systems . to analytically evaluate the rescaled distribution function ( [ rhos ] ) ,
it is convenient to introduce the normalization conditions for the domain - area distribution .
the condition or sum rule for the total domain area was demonstrated in ref .
@xcite . here
, we formulate the condition for the total domain - wall length @xmath23 that yields useful input for the analysis of microscopic exponent @xmath9 .
because domain walls exist between @xmath13- and @xmath14-domains , the length @xmath24 is calculated by integrating the total length of the walls surrounding all @xmath13- or @xmath14-domains .
thus , by using the length @xmath25 of the wall surrounding a domain of area @xmath5 , we have the normalization condition @xmath26 with @xmath27 and @xmath28 the lower cutoff @xmath29 is determined by the microscopic length of the system , that is , by the thickness @xmath30 of the domain wall : @xmath31 .
the upper cutoff comes from percolation theory .
the area @xmath32 of the largest domain is connected with the percolation probability @xmath33 at percolation threshold @xmath34 , @xmath35 , with the exponents @xmath36 and @xmath37 for two - dimensional percolation @xcite . here , @xmath8 is assumed for the coarsening dynamics of conventional systems @xcite , which has been numerically confirmed for segregating binary becs @xcite .
we assume an asymptotic form for @xmath38 and @xmath39 ( see fig .
[ fig_loglog ] ) , @xmath40 with different exponents @xmath41 and @xmath42 in the two dynamic - scaling regimes : the microscopic regime ( @xmath43 ) with @xmath44 and the macroscopic regime ( @xmath45 ) with @xmath46 .
the exponent @xmath47 characterizes the fractal behavior of the domain walls . the trivial exponent , @xmath48 , is realized for circular domains with @xmath49 , for example .
the upper bound , @xmath50 , comes from the spatial dimension of our system . under our assumption
, the length normalization condition ( [ normall ] ) is reduced to @xmath51 with @xmath52 and @xmath53 . here , the contribution from the domain walls surrounding the largest ( percolation ) domain is extracted explicitly as @xmath54 in eq .
( [ normall_reduce ] ) .
a theoretical restriction is obtained for @xmath9 and @xmath55 by evaluating the reduced condition ( [ normall_reduce ] ) in the limit of @xmath56 and @xmath57 .
all terms on the left - hand side of eq .
( [ normall_reduce ] ) must be on the order of unity or zero . from eqs .
( [ s_min_s_max ] ) and ( [ tau_d ] ) , we obtain @xmath58 ; thus , @xmath59 in the limit . substituting eq .
( [ tau_d ] ) into @xmath60 , we have the restriction for @xmath9 , @xmath61 similarly , the restriction for @xmath55 is derived as @xmath62 .
a similar evaluation is available for the normalization condition of area , @xmath63 @xcite .
this condition yields @xmath64 and @xmath65 the restriction ( [ rest_mac ] ) has been obtained in ref .
@xcite and is consistent with the theory of phase ordering percolation : @xmath66 .
since we are primarily interested in the microscopic regime , we do not discuss the macroscopic regime further . to validate our theory , the obtained restriction ( [ rest_mic ] ) in the microscopic regime
is compared with results in conventional coarsening systems of nonconserved and conserved order parameters at zero temperature @xcite . for nonconserved coarsening in the two - dimensional ising model ( 2dim ) @xcite ,
the distribution function @xmath12 becomes flat for @xmath67 with @xmath68 . on the other hand , 2dim simulations for the conserved case @xcite indicate @xmath69 .
smaller domains are close to or indeed circular with @xmath70 and @xmath71 for the former and latter cases , respectively @xcite ; the restriction ( [ rest_mic ] ) is safely satisfied .
these results suggest the scaling of exponents in the microscopic regime varies for different types of coarsening systems depending on the `` microscopic '' dynamics of the domain wall while the exponents in the macroscopic regime take the universal values @xmath10 and @xmath72 for conventional systems @xcite .
our main goal is to study the realizability of the microscopic regime in superfluid systems and identify the scaling behavior .
fortunately , the numerical experiment of segregating binary becs @xcite can be used to achieve this goal .
the segregation dynamics were simulated by solving the coupled gross
pitaevskii ( gp ) equations @xcite derived from the gp lagrangian in a quasi - two - dimensional system , @xmath73 here , we used @xmath74 with atomic mass @xmath75 , complex order parameter @xmath76 , and superfluid velocity @xmath77 of the @xmath78-component .
the @xmath0 symmetry breaking is caused by the dynamic instability starting from the initial state with a constant density @xmath79 and zero velocity @xmath80 for the coupling constants @xmath81 .
for strong segregation in our simulation with @xmath82 @xcite , the domain - wall thickness is on the order of the healing length @xmath83 , and we set @xmath84 .
the dynamic instability , triggered by adding small random seeds to the initial state , develops into complicated domain patterns with characteristic length @xmath1 .
we investigated the domain - area distribution after the time when domain patterns emerge clearly at the initial stage of the ssb development .
the characteristic length @xmath85 of the initial domain patterns at that time is of order @xmath86 as determined from the bogoliubov theory @xcite .
the size @xmath87 of the numerical grid must be smaller than @xmath86 to precisely simulate the dynamics of the instability and the quantized vortices , which will appear at a later stage .
we require a computational system whose size is much larger to satisfy the conditions @xmath38 and @xmath39 for attaining the coexistence of the microscopic and macroscopic regimes . in our simulation , which utilized the crank nicolson method with periodic boundary conditions , the size and number of the numerical grids are set to meet these requirements during the time development as @xmath88 and @xmath89 , respectively .
the numerical results are obtained by averaging @xmath90 samples of the time evolution @xcite .
figure [ fig_s_rho ] shows rescaled plots of the time evolution of the domain - area distribution ( [ rhos ] ) . because we have @xmath91 ( @xmath92 ) in the initial domain patterns , there are fewer domains for @xmath93 .
thus , the microscopic regime is ill - established , and the rescaled plots of different times do not coincide for @xmath93 .
however , the macroscopic regime is well - established in the early stages ( @xmath94 ) . in late stages ( @xmath95 ) , the microscopic regime becomes well - established by exhibiting the asymptotic form of fig .
[ fig_loglog ] with distinct exponents @xmath96 and @xmath97 .
the dynamic - scaling plot for @xmath98 is also successful in the late stages ( see fig .
[ fig_s_rho ] inset ) .
we have exponents @xmath99 and @xmath100 , which is similar to the results of conventional systems @xcite .
the exponent @xmath101 takes the upper limit of the restriction ( [ rest_mic ] ) with @xmath102 ; that is , the number of domains in the microscopic regime is maximized in the superfluid system .
this anomalous behavior is in contrast to the behavior of the above - mentioned conventional systems , while the macroscopic regime shows the universal behavior of percolation criticality . generally speaking ,
the dynamic scaling behavior in the microscopic regime of the domain - area distribution reflects the `` microscopic '' dynamics of domain walls in the system under consideration .
an effective theory that describes the dynamics in our system is quantum hydrodynamics for multi - component superfluids .
it is interesting how the hydrodynamics are connected to the anomalous behavior of the microscopic regime in the quantum fluids .
the anomaly is suggested to occur in the presence of circular vortex sheets described by the quantum hydrodynamics . to provide quantitative evidence of the above suggestion
, we introduce a hydrodynamic theory for a vortex sheet between two domains in a multi - component superfluid .
consider a core - less vortex of circulation @xmath103 with the circulation quantum @xmath104 in the superfluid .
the vortex core is occupied by a @xmath14-domain surrounded by a sufficiently large @xmath13-domain , where a vortex sheet is formed by the relative velocity between the domains . according to a theory of quantum fluid dynamics @xcite , such a state
is stabilized by centrifugal force due to the rotational superflow by the formation of a circular vortex sheet ; however , a flat vortex sheet is dynamically unstable without external forces owing to kelvin
helmholtz instability @xcite .
the radius @xmath24 of the circular wall is estimated by the equation of pressure equilibrium @xmath105 with the tension coefficient @xmath106 of the wall and the hydrostatic pressure @xmath107 ( @xmath108 ) along the wall in the @xmath78-domain .
the theory gives the dynamic - scaling law between @xmath109 and @xmath24 in the microscopic regime .
the pressure @xmath110 in the bulk , where the superfluid is at rest , is the sum of @xmath107 and the hydrodynamic pressure @xmath111 ; @xmath112 . here ,
@xmath113 and @xmath114 are the mass density and superfluid velocity along the domain wall in the @xmath78-domain , respectively .
the @xmath0 symmetry of the hamiltonian for the multi - component superfluid corresponds to @xmath115 . using this result and substituting @xmath116 and @xmath117 into eq .
( [ epre ] ) , we obtain the dynamic - scaling relation with @xmath118 : @xmath119 here , the value of @xmath120 characterizes a length scale related to the domain wall , and thus , it should be on the order of the wall thickness @xmath30 . in the above discussion
, we neglected the existence of domains embedded inside the circular domain wall .
however , it is straightforward to show the contribution from such domains is negligible @xcite .
a circular domain ( @xmath102 ) shrinks and finally collapses in bulk owing to dissipation or evaporation in the conventional coarsening systems .
in contrast , a circular domain with vortices ( @xmath121 ) is stable in the presence of rotational superflow .
this is why the number of domains in the microscopic regime is statistically enhanced with the upper limit @xmath11 in the superfluid system .
the existence of the scaling law ( [ n_to_s ] ) is a direct evidence that the microscopic regime is described by the quantum hydrodynamics .
we examined this law for the numerical experiments of binary becs .
figure [ fig_s_nv ] shows the dynamic scaling plot of the ensemble average of @xmath122 for domains with @xmath123 obtained numerically by integrating the superfluid velocity along domain walls .
the scaling law ( [ n_to_s ] ) is observed after the power - law behavior is well - established in the microscopic regime .
the upturn of the plot at @xmath124 reflects the core structure of a vortex whose radius is on the order @xmath30 , which was neglected in the derivation of eq .
( [ n_to_s ] ) @xcite .
the above statistical discussion is meaningful when domains with vortices are dominant over vortex - free domains ( @xmath125 ) in the microscopic regime .
the inset of fig .
[ fig_s_nv ] shows the plot of the probability @xmath126 in which a domain of @xmath127 has @xmath121 .
the probability is negligible in the initial domain pattern because the bogoliubov modes excited by dynamic instability do not causes momentum exchange , leading to a relative flow between two superfluids @xcite .
however , superflow is locally induced by complex motions of domain walls , and quantized vortices are nucleated owing to the kelvin helmholtz instability . as a result
, the probability @xmath128 grows with time , becoming substantially large at a late stage , where almost domains in the microscopic regime contain vortices .
this result clearly shows that the influence of superflow is essential to the understanding of later - stage dynamics .
interestingly , the dynamic - scaling plot for circulation number @xmath109 yields a different scaling behavior with @xmath129 in the macroscopic regime .
this suggests that distribution of velocity or vorticity may obey a different scaling law or power law for length scales larger than the average distance @xmath1 between domain walls .
this situation is similar to the hierarchy in quantum turbulence , where the kolmogorov or semi - classical power law is realized over length scales larger than the average distance between vortex lines and a different power law is expected for smaller length scales @xcite . in this sense , investigation into the aspects of turbulence is fruitful future work of our system . to reinforce our hypothesis on phase ordering dynamics
, the theory can be applied to different multi - component superfluids , where vortex sheets are stabilized .
an important application is quasi - two - dimensional @xmath130he - a confined in a slab system , where chiral domain walls were recently visualized @xcite .
the order parameter is the vector field @xmath131 that represents the direction of orbital angular momentum of the cooper pair in the dipole - locked @xmath130he - a . a one - to - one correspondence exists between the @xmath131 field of @xmath130he - a and the ( pseudo-)spin field of spinor ( binary ) becs , according to the mermin ho relation that represents vorticity distribution due to the vector - field texture @xcite . in this sense
, these superfluid systems can show a similar effect even if the `` microscopic '' dynamics differs between the @xmath130he - a and spinor ( binary ) becs . by quenching the @xmath130he - a system to the superfluid phase from the normal fluid phase ,
domain growth should occur .
then , the characteristic domain - area distribution , as is illustrated in fig .
[ fig_loglog ] , will be observed when the characteristic domain size is much larger than thickness of domain wall and much smaller than the system size .
we are grateful to l. cugliandolo , l. a. williamson , p. b. blakie , o. ishikawa , and k. kasamatsu for useful discussion and comments on this work .
this work was supported by kakenhi from the japan society for the promotion of science ( grant no .
26870500 ) .
99 a. vilenkin and e. p. s. shellard , _ cosmic strings and other topological defects _ ( cambridge university press , cambridge , 1995 ) . , edited by y. m. bunkov and h. godfrin , vol .
549 of nato advance science institute , series c : mathematical and physical sciences ( kluwer , dordrecht , 2000 ) .
a. onuki , _ phase transition dynamics _ ( cambridge university press , cambridge , 2002 ) .
t. vachaspati , _ kinks and domain walls : an introduction to classical and quantum solitons _ ( cambridge university press , cambridge , 2006 ) .
a. j. bray , adv .
phys . * 43 * , 357 ( 1994 ) .
d. stauffer and a. aharony , _ an introduction to percolation _
, 2nd ed .
( taylor and francis , london , 1994 ) .
j. j. arenzon , a. j. bray , l. f. cugliandolo , and a. sicilia , phys .
rev .
lett . * 98 * , 145701 ( 2007 ) ; a. sicilia , j. j. arenzon , a. j. bray , and l. f. cugliandolo , phys .
rev .
e * 76 * , 061116 ( 2007 ) .
a. sicilia , j. j. arenzon , i. dierking , a. j. bray , l. f. cugliandolo , j. martnez - perdiguero , i. alonso , and i. c. pintre , phys .
rev .
lett .
* 101 * , 197801 ( 2008 ) .
j. olejarz , p. l. krapivsky , and s. redner , phys .
rev .
lett . * 109 * , 195702 ( 2012 ) .
a. tartaglia , l. f. cugliandolo , and m. picco , europhys .
lett . * 1116 * , 26001 ( 2016 ) ; l. cugliandolo , _ phase ordering kinetics , aggregation and percolation in two dimensions _ , plenary lecture in statphys26 , lyon , france ( 2016 ) .
k. damle , s. majumdar , and s. sachdev , phys .
rev .
a * 54 * , 5037 ( 1996 ) .
s. mukerjee , c. xu , and j. e. moore , phys .
rev .
b * 76 * , 104519 ( 2007 ) .
h. takeuchi , k. kasamatsu , m. tsubota , and m. nitta , phys .
rev .
lett . * 109 * , 245301 ( 2012 ) .
k. kudo and y. kawaguchi , phys .
rev .
a * 88 * , 013630 ( 2013 ) .
m. karl , b. nowak , and t. gasenzer , phys .
rev .
a * 88 * , 063615 ( 2013 ) .
k. kudo and y. kawaguchi , phys .
rev . a * 91 * , 053609 ( 2015 ) .
j. hofmann , s. s. natu , and s. das sarma , phys .
rev .
lett . * 113 * , 095702 ( 2014 ) .
l. a. williamson and p. b. blakie , phys .
rev .
lett . * 116 * , 025301 ( 2016 ) ; phys .
rev .
a * 94 * , 023608 ( 2016 ) .
h. takeuchi , y. mizuno , and k. dehara , phys .
rev .
a * 92 * , 043608 ( 2015 ) .
h. takeuchi , j. low temp .
phys . * 183 * , 169 ( 2016 ) .
a. bourges and p.
b. blakie , phys .
rev .
a * 95 * , 023616 ( 2017 ) .
a. sicilia , y. sarrazin , j. j. arenzon , a. j. bray , and l. f. cugliandolo , phys .
rev .
e * 80 * , 031121 ( 2009 ) .
the integral represents the total area occupied by @xmath13- or @xmath14-domains , divided by the system area @xmath16 , yielding @xmath132 for the case of the @xmath0 symmetry breaking . c. j. pethick and h. smith , _ bose einstein condensation in dilute gases _ , 2nd ed .
( cambridge university press , cambridge , 2008 ) .
the coupling constants are controllable by changing the atomic interaction through the feshbach resonance and the oblateness of the trapped condensates .
strong segregation can be achieved , for example , s. b. papp , j. m. pino , and c. e. wieman , phys .
rev .
lett .
* 101 * , 040402 ( 2008 ) . since spinor becs can be described effectively by the same hydrodynamics , the system is also available for examining the scaling behavior .
see also the references in _
e.g. _ , ref .
@xcite .
the ensemble averages , except for the average of @xmath1 , were taken by considering the periodic boundary conditions ; averaged quantities are calculated by averaging 64 samples after the secondary average over the @xmath133 pseudo - samples obtained by shifting the field data of a single simulation in the @xmath134 and @xmath135 directions by @xmath136 and @xmath137 with integers @xmath138 and @xmath78 ( @xmath139 ) , respectively .
the length @xmath140 is much larger than the characteristic length @xmath1 of the field , so the secondary average improves the statistical analysis substantially .
s. hayashi , m. tsubota , and h. takeuchi , phys .
rev .
a , * 87 * , 063628 ( 2013 ) . h. takeuchi , n. suzuki , k. kasamatsu , h. saito , and m. tsubota , phys .
rev .
b * 81 * , 094517 ( 2010 ) .
the contribution of embedded domains to the circulation number @xmath109 is statistically negligible .
the contributions to the circulation from the considered domain and the domains embedded within it are calculated analytically and compared statistically according to the dynamic scaling hypothesis in the supplemental material .
the upturn is understood qualitatively by considering the density variation owing to rotational superflow in the static solution . neglecting the quantum pressure term @xmath141 in the lagrangian @xmath142
, we obtain @xmath143 $ ] , where an upturn occur around @xmath144 .
s. ishino , m. tsubota , h. takeuch , phys .
rev .
a * 83 * , 063602 ( 2011 ) . c. f. barenghi , l. skrbek , and k. r. sreenivasan , proc .
natl .
acad .
sci .
usa * 111 * , 4647 ( 2014 ) .
y. sasaki , _ visualizing textural domain walls in superfluid @xmath130he by magnetic resonance imaging _
, invited talk in the international conference on quantum fluids and solids 2016 ( qfs2016 ) , prague , czech republic ( 2016 ) .
g. e. volovik , _ the universe in a helium droplet _ ( clarendon press , oxford , 2003 ) .
k. kasamatsu , m. tsubota , and m. ueda , phys .
rev .
a * 71 * , 043611 ( 2005 ) .
g. e. volovik and m. krusius , physics * 5 * , 130 ( 2012 ) . |
in recent years there have been many infrared surveys of the chamaeleon i dark cloud @xcite .
these surveys provide many new candidate pre - main sequence stars based on photometric criteria .
these observations aim to detect the faintest and thus the least massive pre - main sequence stars and are , in most cases , sensitive enough to detect young stars with masses close or even below the h burning limit .
a complete census of the stellar population in the cloud would provide better estimates of the mass distribution and the star - formation history of chamaeleon i. in particular , these surveys open the door to study the imf down to the substellar regime .
the chamaeleon i cloud is especially adequate for these purposes since it is located nearby ( d @xmath0 160 pc , * ? ? ?
* ) , while also extending over moderate angular scales on the sky ( @xmath0 3 deg@xmath8 , * ? ? ?
* ; * ? ? ?
@xcite determined the imf of the cloud down to @xmath00.3 @xmath9 based on @xmath0 80 optically visible known members and found a good agreement with the miller - scalo model . @xcite
intensively observed a small region ( 300 arcmin@xmath4 ) in the center of chamaeleon i dark cloud .
this area contains 22 young stars , including 13 very low mass members with k @xmath0 1113.5 identified in h@xmath10 .
they found that the imf for the central area of the chamaeleon i cloud was roughly flat ( in logarithmic mass units ) from @xmath0 1 @xmath2 down to @xmath0 0.03 @xmath2 . in order to study the extended low - mass imf in the whole chamaeleon i , it is desirable to determine characteristics of the new candidate young stars such as spectral types .
the near - infrared spectra combined with published pre - main sequence evolutionary tracks can also be used to derive stellar masses and ages . in this contribution we present near - infrared spectra of 46 of the brightest ( _ typically _ @xmath1 12 ) candidate young stellar objects selected from the surveys cited above .
we chose the brightest objects in order to have reasonable completeness and to achieve good s / n ratios in 4-m class telescopes .
in addition , we obtained near - infrared spectra of 63 previously known members of this cloud as well as other nearby southern star - forming regions such as lupus .
we used these known young stellar objects as a spectral type calibration group .
these known objects and the @xcite atlas allow us to identify spectroscopic features usually present in young stellar objects that support the pre - main sequence status of the new objects .
we combine our data with published photometry and spectral types of other sources in the cloud to compile a list of 145 pre - main sequence stars in the chamaeleon i dark cloud and study the mass and age distributions .
in 2 we describe the observations and data reduction . in
3 we present our analysis and results .
we derive spectral types from the water vapor indexes which combined with published photometry allows us to place the stars in the hr diagram .
we conclude with a brief summary in 4 .
we carried out these observations on april 10 12 , 1999 with the eso ntt near - infrared spectrograph / imaging camera sofi ( son of isaac ) and on may 6 9 , 1999 with osiris ( ohio state infrared imager and spectrometer ) on the ctio blanco 4-m telescope .
a few additional spectra were obtained on feb 27 28 , 2002 with this telescope and infrared camera .
tables 1 and 2 list the targets observed .
sofi has two low resolution grisms ( red and blue ) that roughly cover the @xmath11 bands on a hawaii hgcdte 1024@xmath51024 detector with a plate scale of 0.292@xmath12/pix .
the blue grism covers the spectral region between 0.951.63 @xmath13 m and the red grism the region between 1.532.52 @xmath13 m .
the corresponding spectral resolutions ( r @xmath14 @xmath15/@xmath16@xmath15 ) are 930 and 980 for a 0.6@xmath12 slit .
we used a 1@xmath12 wide and 290@xmath12 long slit .
osiris was used in the x - disp ( multi - order / cross - dispersed ) mode with the f/2.8 camera and a slit 30@xmath12 long and 1.2@xmath12 wide .
this configuration gives r ( @xmath15/@xmath16@xmath15 ) @xmath0 1200 ( 2 pixels ) , while covering the @xmath17 , @xmath18 , and @xmath19 bands simultaneously on a 1024 @xmath5 1024 hgcdte detector at a plate scale of 0.403@xmath12/pix .
in particular , the x - disp mode uses a cross - dispersing grism that gives the @xmath11 spectral region in three adjacent orders , from 5th to 3rd . the integration time and the number of exposures per target
were chosen based on the brightness of the source and background contribution each night . in most cases we obtained two sets of 9 blue and red spectra of the same integration time with sofi .
we repeated this sequence for objects fainter than @xmath20 13 .
common integration times were 90180 sec for both grisms . the telescope was nodded 3060@xmath12 along the 290@xmath12 long slit between consecutive positions following the usual abba pattern .
this procedure corresponds to the `` nod throw along slit '' scheme as described in the sofi users manual @xcite .
with osiris we _ typically _ obtained seven images of identical integration time , with the telescope dithered by 7@xmath12 along the 30@xmath12 long slit .
the individual integration times varied between 30 and 180 sec .
for faint objects ( i.e. , @xmath21 12 ) we obtained two 180 sec sequences of seven spectra .
in addition to the program sources we observed several atmospheric standards .
we selected two groups of standards with similar airmass to our candidate sources and observed them ( every @xmath0 2 hours ) each night .
these telluric stars comprise both late ( g3 - 5 ) and early ( o8 ) spectral types and were chosen from the lists available at both observing sites .
we obtained multiple flat field images , with a dome screen , using incandescent lamps on and off .
a xenon lamp , also taken on and off , each night provided the wavelength calibration for our sofi data . to determine the wavelength scale of the osiris spectra we used a henear lamp . to reduce the data we used iraf .
we subtracted one image from another ( using pairs of nodded observations ) to eliminate the background and sky contribution in first approximation .
this subtraction automatically took care of dark current and bias level .
we flat - fielded our data dividing by a normalized dome flat .
we used the _ twodspec _ task apall to trace and extract the spectra along a 6 pixel wide aperture in the sofi data and 10 pixel in the osiris co - added images .
in particular for osiris data we treated @xmath11 x - disp adjacent orders using the _ echelle _ package .
a further sky subtraction was done by fitting a polynomial to the regions on either side of the aperture .
a non linear low order fit to the lines in the xenon and henear lamps were used to wavelength calibrate the spectra .
we removed telluric features from our data dividing each program star by the atmospheric standard spectrum .
for the osiris data we used a g3 - 5 standard and both g3 - 5 and o8 stars for the sofi spectra .
the telluric star differed from the science target by @xmath22 0.2 in airmass .
we used the _ onedspec _ task splot to remove h i ( paschen and brackett ) lines from the spectra of the g3 - 5 and o8 stars , interpolating across each line .
the division by the standard canceled out telluric features in the science spectrum reasonably well but introduced the inverse of the atmospheric standard in the corrected spectrum .
we recovered the true spectral shape multiplying the resultant spectrum by a planck function at the temperature corresponding to the atmospheric star .
we finally combined the @xmath11 x - disp orders for our osiris spectra and the blue and the red grism regions for each object observed with sofi and eliminated regions of deep atmospheric absorptions .
we also trimmed out wavelengths 0.950.97 @xmath13 m from our sofi spectra due to edge effects .
the useful spectral range provided by osiris is 12000 - 23500 while sofi covers from 9700 to 25300 . for the sofi spectra , as mentioned before , we obtained two sets of spectra corresponding to the g3 - 5 and o8 telluric stars that were essentially identical . in section 3.2
we show the set of science spectra obtained applying the o8 standard as telluric corrector .
we observed two groups of stars : a ) 46 candidate young stellar objects detected by several recent surveys of the cloud ; and b ) 63 previously known young stars ( 36 belonging to the chamaeleon i dark cloud and 27 located in other star - forming regions , such as lupus ) .
three objects in the second group ( vw cha , sz 119 , and sz 124 ) were observed with both sofi and osiris .
eight of the stars in the second group already had near - infrared spectra .
@xcite observed haro1 - 1 in the @xmath23 ophiuchi cloud and @xcite cce98 49 ( iso - chai 256 ) in chamaeleon i. @xcite published combined @xmath24 band spectra for cha h@xmath10 16 .
@xcite obtained individual @xmath11 band spectra for cha h@xmath10 1 and cha h@xmath10 2 . in this paper
we report individual @xmath11 data for each of the cha h@xmath10 16 stars .
the 46 new objects were selected , in general , among the brightest ( _ typically _ k @xmath22 12 ) near - infrared objects detected by several recent surveys in the cloud @xcite .
these lists include about @xmath0 150 additional potential members of the cloud with k @xmath3 13 . in this contribution
we analyze data for the brightest objects for which it is feasible to obtain near - infrared spectra with a 3 - 4 meter class telescope in about 20 - 40 minutes total integration time .
the second set of 63 previously known objects were selected among the southern hemisphere pre - main sequence stars to have a group of reference from which to derive common near - infrared spectroscopic features among young stellar objects .
objects with previously known spectral types within this group allowed us to calibrate the water vapor index @xmath25 based classification ( see section 3.4 ) .
tables 3 and 4 compile photometric and spectroscopic data ( when available ) for the observed stars . figure 1
shows the position of new ( stars ) and previously known ( dots ) objects in the near - infrared color - color diagram .
superposed on this plot are the loci of unreddened main sequence dwarfs ( solid line , * ? ? ?
* ) and ctts classical t tauri stars ( long - dash line , * ? ? ?
the reddening band ( dotted line , * ? ? ? * ) and the length of the reddening vector @xcite are also indicated .
@xmath14 5 mag @xcite .
typical photometric errors are displayed in the upper left corner.,width=642 ] although some of the candidate objects have significant near - infrared excesses indicating the presence of circumstellar material , most of the previously known and new sources are located within or close to the reddening band .
the group of known objects includes many bona - fide ctts in the cloud ( e.g. , * ? ? ?
* ; * ? ? ?
* ) without measurable near - infrared excesses and thus the near - infrared color - color diagram may miss a significant number of young members of the cloud .
most of the candidate sources located within the reddening band in figure 1 were proposed by @xcite based on the @xmath28 vs @xmath29 diagram .
near - infrared spectra provide additional information to support the pre - main sequence nature of these candidate sources .
figure 2 shows the @xmath11 band spectra of all candidate objects reported in this paper and figure 3 a selected sub - sample of 12 of the previously known young stellar objects we observed .
prominent atomic and molecular features in the observed spectral range for late type stars ( i.e. , spectral types later than g0 ) are labeled @xcite . both groups ( new objects figure 2 and previously known young stars figure 3 ) have a wide variety of spectral shapes .
some sources have shapes decreasing with @xmath15 , others display a turnover around 1.61.8 @xmath13 m or the spectral shape is roughly constant .
other objects have rising or steeply rising shapes towards 2 @xmath13 m .
considered as groups , new young stellar objects and previously known stars show similar near - infrared spectroscopic characteristics .
band spectra of candidate young stellar objects in the chamaeleon i dark .
prominent atomic and molecular features in the observed spectral range are labeled.,title="fig:",width=672 ] -0.3 in band spectra of 12 of the previously known young stellar objects for which we obtained near - infrared data .
prominent atomic and molecular features in the observed spectral range are labeled.,title="fig:",width=672 ] -0.3 in @xcite presented an atlas of low resolution spectra of @xmath0 100 young stellar objects in different star - forming regions , including taurus , @xmath23 ophiuchi , r cra , ic 5146 , scorpio , and lupus .
these authors describe near - infrared spectroscopic characteristics of class i , ii , and iii objects as well as flat - spectrum and fu orionis objects .
they found that , in general , the spectral shapes correlate with the spectral energy distribution class and that equivalent widths of absorption features decrease from the class iii to the class i sources , as a consequence of the increasing amount of circumstellar material .
in addition many young stellar objects ( usually class ii and i ) have atomic hydrogen lines in emission .
for the previously known young stellar objects observed we find a similar correlation between the spectral shape and the classical ( class ii ) or weak emission ( class iii ) t tauri type .
in general , class ii objects have rising or flat spectral shapes .
class iii stars have decreasing or flat spectral shapes .
we find candidate sources have rising , flat , and decreasing shapes .
the first and second groups probably correspond to class i - ii members of the cloud and the rest to potential class iii young stellar objects .
tables 5 and 6 list equivalent widths corresponding to the strongest lines detected in the spectra of the observed targets . for the sofi data we estimate an uncertainty of @xmath0 1 - 2 in the pa@xmath30 , br@xmath31
, na i doublet ( 2.206 and 2.209 @xmath13 m ) , and ca i triplet ( 2.261 , 2,263 , and 2.266 @xmath13 m ) equivalent widths . the combined co @xmath32 @xmath14 2 - 0 and 3 - 1 bands have a slightly worse precision of @xmath0 3 .
the osiris spectra in general have poorer s / n ratio in relation to sofi data .
we estimate an error of 2 - 3 in our measurements of pa@xmath30 , br@xmath31 , na i doublet , and ca i triplet . the osiris spectra only cover co @xmath32 @xmath14 2 - 0 , and 3 - 1 bands .
we estimate an uncertainty of @xmath0 4 - 5 in the two co bands covered by this instrument .
the three stars observed with both instruments ( vw cha , sz 119 , and sz 124 ) have equivalent width measurements that agree within the estimated uncertainties .
we notice the sofi spectra cover the four co @xmath32 @xmath14 2 - 0 , 3 - 1 , 4 - 3 , and 5 - 3 bands . in tables 5 and 6
we provide the combine co equivalent width corresponding to the @xmath32 @xmath14 2 - 0 and 3 - 1 bands to properly compare these measurements with those obtained for the osiris data .
stars in tables 5 and 6 have , on average , similar or slightly smaller equivalent widths than m type standards @xcite , with exception of few sources ( such as pmk99 ir cha ina1 , ced 110 irs 4 , ced 110 irs 6 , and glass i ) with no absorption lines present , at the spectral resolution used .
@xcite also noticed that many young stellar objects in their sample ( about 100 stars belonging to @xmath23 oph , taurus , and other star - forming regions ) have atomic ( na i and ca i ) and co equivalent widths in the same range as late - type mk standards .
figure 4 shows the ca i vs na i equivalent widths for the groups of previously known class ii ( ctts ) and class iii ( wtts ) objects observed . for the ca i line equivalent width we obtain an average of 3.30 @xmath33 1.39 and 4.88 @xmath33 1.59 for the class ii and class iii , respectively .
the na i line equivalent width gives similar results : 3.48 @xmath33 1.59 for the ctts and 3.87 @xmath33 1.32 for the wtts .
although the uncertainties are large , on average , the class ii sources have slightly smaller equivalent widths compared to the class iii group .
smaller equivalent widths may indicate the presence of circumstellar material partially filling - in these absorption lines .
a similar plot for the new sources gives only a marginal correlation between the equivalent widths and the spectral shapes .
we note that the poor s / n ratio in the spectra of several candidate objects rendered uncertain the detection of spectral line features .
better s / n data are required to confirm this initial correlation .
several stars ( such as cce98 24 , vw cha , hjc c 1 - 6 , and byb 43 ) also show h atomic lines in emission ; the strongest are pa@xmath30 and br@xmath31 .
we notice an apparent tendency of a higher frequency of h i emission lines among previously known young stellar objects than among the new sources .
twenty of the 63 previously known young stars have pa@xmath30 or br@xmath31 in emission , whereas only one of the 29 measured candidate young stars shows these features
. however the modest s / n ratio of some of the new stars spectra may explain at least in part this tendency .
higher s / n data are necessary to properly consider this issue .
figure 5 shows an enlarged spectrum of the source pmk99 ir cha ina1 , that has molecular hydrogen in emission .
in particular h@xmath34 1 - 0 q ( 18 ) as well as h@xmath34 1 - 0 s ( 10 ) are clearly detected .
pmk99 ir cha ina1 is a class i source in the cloud identified by @xcite roughly located at the center of the high velocity bipolar outflow found by @xcite .
based on the physical properties of pmk99 ir cha ina1 ( steep spectral energy distribution and association with a small near - infrared nebula ) @xcite proposed this star rather than hm 23 ( a class ii source lying in the region ) as the exciting source of the bipolar outflow .
hm 23 had been originally associated with the co flow as it was the nearest young star known in the vicinity of the flow .
molecular hydrogen lines in the spectrum of pmk99 ir cha ina1 are probably originated in the shock regions where bipolar outflow impacts on the surrounding molecular gas @xcite .
band spectrum of pmk99 ir cha ina1 .
molecular hydrogen lines , and are in emission.,width=642 ] to derive spectral types for the candidate sources we applied the stellar water vapor indexes proposed by @xcite and @xcite .
these indexes are reddening - free and compare the intensity of the absorption bands near 2 and 1.9 @xmath13 m , respectively .
the strength of these bands are extremely sensitive to the spectral types for m dwarfs .
@xcite s index @xmath35 is defined by the average values of relative flux density calculated in three narrow bands , f1 ( 2.072.13 @xmath13 m ) , f2(2.2672.285 @xmath13 m ) , and f3(2.402.50 @xmath13 m ) , as @xmath36 this index can only be applied to the sofi spectra .
osiris data have poor s / n ratio at wavelengths longer than 23500 . the index @xmath25 @xcite measures the relative fluxes in four narrow band filters ( 0.05 @xmath13 m width ) centered at 1.675 @xmath13 m , 2.075 @xmath13 m , and 2.25 @xmath13 m . denoting the fluxes at each narrow band filter by f1 , f2 , f3 , and f4 , respectively ,
the index @xmath25 is expressed by : @xmath37 the @xmath25 index allows us to derive spectral types for our candidate sources observed with both instruments as , in general , involves spectral regions of good s / n ratios .
we notice that the @xmath35 and @xmath25 indexes yield values which are independent of the extinction .
they provide an appropriate tool to estimate spectral types for young stellar objects usually heavily obscured by the parent cloud dusty material .
@xcite ( see also * ? ? ?
* ) obtained a linear relation to determine the m sub - type as function of @xmath35 using m dwarf standards with optically known spectral types ( see equations ( 2 ) in * ? ? ?
* ; * ? ? ?
these relations have been derived from standards whose spectra were telluric corrected using a0 v stars but whose continuum shapes were not restored ( i.e. , without multiplying by the planck function spectrum at the temperature of the corresponding standard ) . as explained in section 2 we used both g3 - 5 and o8 telluric standards for our sofi data and thus our atmospheric telluric divided spectra ( without correction for the slope of the standard spectrum ) are not entirely comparable to those of @xcite and @xcite .
nevertheless , @xcite noticed little dependence on the spectral type of the telluric standard in the analysis of isocam sources of the chamaeleon i dark cloud . to check the applicability of these calibrations to our data we plot , in figure 6 , the optical spectral types vs the spectral types derived from the index @xmath35 for previously known objects . in this case
we show the plot corresponding to the spectra telluric corrected using the o8 standards .
the results are basically the same when we use the g5 stars .
our spectra of sz 23 and gk-1 have relatively poor s / n ratio in the 2.4 2.5 @xmath13 m region and the fluxes in the band f3 are very uncertain .
these two objects were not included in figure 6 .
we fit a linear relation to the data on this plot with a correlation coefficient r @xmath14 0.95 .
all spectral types derived from the index @xmath35 agree within 1.5 sub - type with the optical types .
this difference is comparable to the accuracy expected by @xcite and @xcite for their calibrations .
vs @xmath35-index - derived m@xmath38 for previously known objects observed with sofi .
the dash line indicates a linear fit to the data.,width=642 ] we derive the index @xmath25 for previously known objects with optical spectral type observed with both instruments and obtain the following linear least - squares fit to the data : @xmath39 with a correlation coefficient r @xmath14 0.94 .
we estimate a maximum uncertainty of about 1.7 sub - classes in this calibration , considering _ typical _ errors in our measurements of the index @xmath25 and the correlation coefficient of equation [ tres ] .
@xcite derived a similar relation of considerably lower precision based only on six previously known observed objects .
we also determined equation ( 3 ) using only sofi and osiris data separately .
figure 7 shows the optical spectral types vs the spectral types derived from the index @xmath25 for previously known objects .
both calibrations agree within the 1.7 sub - classes uncertainty .
poor telluric corrections rendered unreliable the fluxes in one or more bands used by this index for several objects such as sz 119 , chxr 78c , uy cha , ho lup , and sz 77 and thus these stars were omitted in the derivation of equation [ tres ] . vs @xmath40-index - derived m@xmath38 for previously known objects observed with sofi and osiris .
the dash line indicates a linear fit to the data.,width=642 ] in table 7 we give spectral types derived from the application of these indexes for our candidate sources .
several objects in our sample such as cce98 20 , cce98 35 , cce98 39 , and cce98 46 have very poor s / n spectra and thus spectral type derivations would be very uncertain .
other two near - infrared selected objects ( cce98 9 and gk2001 15 ) have better s / n spectra in relation to the ones mentioned above but still rather poor to attempt a reliable spectral type determination . finally two objects with relatively good s / n spectra cce98 40 and cce98 44 have spectral types earlier than m0 justing from the significant br@xmath31 equivalent width absorption lines ( 7 and 8 , respectively ) and the absence of na i , ca i and co features in their spectra ( see table 5 ) .
table 8 lists a group of previously known objects without optical spectral types in the literature for which we applied the water vapor indexes .
poor telluric corrections around 2.05 @xmath13 m compromised the fluxes in the f3 and f3 bands ( @xmath35 and @xmath25 indexes , respectively ) and prevented us from obtaining spectral types for ced 110 irs 4 and ced 110 irs 6 .
hjm c 2 - 3 has a significant br@xmath31 equivalent width ( @xmath0 12 ) and no additional features indicative of a late spectral type .
this object is probably earlier than m0 . for objects with sofi spectra available we derive spectral type from both indexes and obtain a good agreement within 1.7 sub - class .
however , it was not always possible to accurately measure the flux in the f3 band corresponding to the index @xmath35 due to poor s / n ratio in this spectral region .
in these cases we adopted the spectral type given by the index @xmath25 . to derive effective temperatures , intrinsic colors , and bolometric corrections for our targets from the estimated spectral types we adopted the corresponding calibrations obtained by @xcite .
these relations comprise spectral types from m2 to m9 .
we supplemented these determinations with @xcite relations in order to include m0-m1 stars .
the thermal emission from warm dust grains in circumstellar disks around young stellar objects may affect the observed water vapor bands and thus the spectral type estimates for our targets .
this infrared excess emission can be characterized by @xmath41 , where @xmath42 is the measured flux , including the excess ( circumstellar ) emission , and @xmath43 is the flux due to the underlying stellar photosphere .
@xcite analyzed the near - infrared properties of t tauri stars located in the tarus - auriga molecular cloud and derived median values of the j- , h- , and k - band excesses for the group of the ctts and for the wtts .
they obtained @xmath44 @xmath14 0.0 , @xmath45 @xmath14 0.2 , and @xmath46 @xmath14 0.6 for the ctts and @xmath44 @xmath14 @xmath45 @xmath14 @xmath46 @xmath14 0.0 for the wtts . in tables 7 , 8 , and 9 we give r@xmath47 , the @xmath19-band excess , estimated for our target objects using the following relation , @xmath48 } \label{cuatro}\ ] ] ( see * ? ? ?
. in this expression @xmath49 is the observed color ( see tables 3 and 4 ) and @xmath50 corresponds to the intrinsic or photospheric color . the third term in equation ( [ cuatro ] )
is related to the excess in color due only to the extinction to each individual target . to derive @xmath51 we applied the @xcite reddening law , @xmath52 , with @xmath53 where , @xmath54 , is the color excess , and @xmath55 , the intrinsic color .
tables 7 , 8 , and 9 list @xmath56 for the observed targets as well as the corresponding values of r@xmath57 .
this estimate provides only a lower limit to the veiling as equation ( [ cuatro ] ) assumes r@xmath58 @xmath14 r@xmath59 @xmath14 0 ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however we expect the k - band excess , r@xmath57 , provides the main contribution to the total infrared excess , @xmath60 , in view of the median values for the ctts in taurus derived by @xcite .
@xcite have simulated the spectra expected from young stars surrounded by circumstellar matter , adding the disk contribution to the spectra of known standard stars .
these authors have then compared the index @xmath35 calculated for stars of known spectral types _ with _ and _ without _ the disk emission .
the spectral types obtained from star+disk systems that have veiling effects of r@xmath61 0.2 are about 1 sub - class earlier than those derived for objects without the disk contribution .
this effect increases as r@xmath57 increases and can account for @xmath0 2 - 3 sub - classes earlier for r@xmath61 0.6 .
most objects in tables 7 , 8 , and 9 have r@xmath47 @xmath22 0.20.3 with the exception of pmk99 ir cha ina1 ( r@xmath57 @xmath14 0.7 ) .
we expect our spectral type determinations using the water vapor indexes to be little affected by the effects of veiling except for pmk99 ir cha ina1 that may actually have a spectral type @xmath0 3 sub - types later than that given in tables 7 according to its large k - band excess .
it is also likely that in this case at least the h - band excess contribution be not negligible and then increases the total effect . in the table 7 ( in brackets )
we adopt a minimum veiling correction of 3 sub - types later and indicate the corresponding derived parameters .
we include this correction on the estimated mass and age for this object in section 3.6 .
we notice that any additional veiling correction would shift the object practically horizontally on the hr diagram corresponding to decreasing values of mass and age .
for example an additional increase of the spectral type in 0.5 sub - class ( i.e. , m7 spectral type ) would give a mass of 0.05 @xmath2 and an age of 2 @xmath5 10@xmath6 yr for this object .
pre - main sequence stars are known to have surface gravities intermediate between those of dwarfs ( luminosity class v ) and giants ( luminosity class iii ) ( e.g. , * ? ? ?
* ; * ? ? ?
. however we have determined spectral types for our targets neglecting this effect , in particular when applying the @xmath35 index versus the m spectral type relation derived by @xcite and @xcite .
these authors used m main sequence standards to obtain this calibration . to derive equation ( 3 ) , the @xmath40 index calibration , we only used pre - main sequence stars with optically known spectral types and thus this relation may be more appropriate , in terms of the gravity effect , to obtain spectral types for our candidate young stellar objects .
several authors ( see , for example , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have analyzed this effect on spectral type determinations and in general conclude that the surface gravity effect introduces a _
typical _ uncertainty of about 12 sub - classes .
this estimate is appropriate for our water vapor index based spectral classification .
we used the @xmath17 magnitudes to estimate bolometric luminosities .
this spectral band is less affected by contamination from circumstellar infrared excess emission than @xmath19 or @xmath18 .
the bolometric luminosities are calculated from the following expressions : @xmath62 @xmath63 where , dm @xmath14 6.0 is the distance modulus and @xmath64 is the @xmath17 band bolometric correction .
@xmath65 magnitude measurements we used the @xmath19 band data in combination with the @xcite reddening law ( @xmath66 , with @xmath67 ) to estimate the bolometric luminosities . ] and @xmath68 are the extinction and the apparent magnitude , respectively .
luminosities for the observed targets are given in tables 7 , 8 , and 9 .
our determinations agree within a factor of @xmath0 2 with previous derivations for the already known objects in the chamaeleon i and lupus clouds ( cf . * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we chose pre - main sequence evolutionary tracks and isocrones from dantona & mazzitelli ( 1994,1998 ) to derive masses and ages for the observed sources .
this model provides consistent and plausible results for our group of new targets , in reasonable agreement with the higher mass members of the cloud .
in addition , dantona & mazzitelli s ( 1994,1998 ) calculations include stars in a wide range of masses , from 3 @xmath2 to objects well below the h burning limit ( 0.017 @xmath2 ) and thus allows us an uniform and consistent derivation of masses and ages for our targets .
tables 7 , 8 , and 9 list these parameters .
figure 8 shows the location of new and previously known chamaeleon i stars spectroscopically observed on the hr diagram . , the h burning limit .
the dashed lines correspond to the isocrones calculated also by the same authors.,width=642 ] we derived a mass of 0.19 @xmath2 and an age of 4 @xmath5 10@xmath69 yr for the source pmk99 ir cha ina1 using the spectral type m3.5 given in table 7 .
however as discussed in section 3.5 this object may actually have a spectral at least type 3 subtype later due to the high veiling ( r@xmath47 @xmath14 0.7 ) . to correct for the veiling effect we adopt a m6.5 spectral type and recalculate the corresponding bolometric luminosity ( @xmath701.36 ) .
this shifts the star to the right practically horizontally on the hr diagram , corresponding to decreasing values of mass and age ( 0.06 @xmath2 and 4 @xmath5 10@xmath6 yr , respectively ) . in table 7
we indicate in brackets the veiling corrected parameters . to our knowledge
pmk99 ir cha ina1 is the first substellar object associated with a bipolar outflow .
the candidate chamaeleon i young stars ( figure 8) have masses between @xmath0 0.7 and 0.04 @xmath2 .
the previously known members of the cloud in figure 8 , roughly show the same range of masses and with the exception of lkh@xmath10 332 - 17 ( @xmath0 2.4 @xmath2 ) , one of the highest mass members of the cloud ( see also * ? ? ?
gk2001 18 with a mass of 0.04 @xmath2 and pmk99 ir cha ina1 with a mass of 0.06 @xmath2 are the only two substellar objects detected among the observed chamaeleon i candidates .
the new objects have ages between @xmath3 1 @xmath5 10@xmath6 and @xmath0 2 @xmath5 10@xmath7 yr , roughly in agreement with the known members of the cloud @xcite .
our mass determinations for the known members agree within a factor of @xmath0 2 with those derived by @xcite and @xcite .
we find that the age determinations agree within a _ typical _ factor of 4 - 5 , except for hm 15 , hm 16 , glass i and cha h@xmath10 1 6 , for which we find a larger ( about a factor of 10 ) difference in age . in the case
cha h@xmath10 1 6 this difference is a least in part due to the use of different isocrones .
@xcite based their determination of ages on @xcite and @xcite models whether we applied dantona & mazzitelli s isocrones . the lupus previously known young stars span a range of ages of @xmath0 3 @xmath5 10@xmath6 and 1 @xmath5 10@xmath69 yr and masses between @xmath0 0.5 and 0.14 @xmath2 .
the only exception is sz 68 , with a mass of 1.5 @xmath2 .
our determinations of mass and age agree within a factor of roughly 2 and 4 , respectively with those derived by @xcite .
@xcite derived the mass and age distributions for the chamaeleon i dark cloud based on @xmath0 80 previously known members of the cloud , adopting dantona & mazzitelli s ( 1994 ) and @xcite models .
the mass distribution rises from @xmath0 2 @xmath9 , peaks at 0.5 - 0.6 @xmath2 and then falls towards lower masses .
they obtained a median stellar mass of 0.55 @xmath2 as representative of their sample .
the age distribution peaks at ages @xmath22 5 @xmath5 10@xmath69 yr indicating a roughly constant rate of star formation during the last 5 @xmath5 10@xmath69 yr and a rapidly declines towards older ages .
we combined our targets with those from @xcite , @xcite , @xcite and @xcite to determine the mass and ages distribution for 145 objects belonging to the chamaeleon i dark cloud . to analyze an homogeneous as possible data sample we estimated bolometric luminosities and effective temperatures for all the cloud members in the same manner as for the objects in tables 7 and 8 ( see sections 3.4 and 3.5 ) .
we then used dantona & mazzitelli s ( 1994,1998 ) evolutionary tracks and isocrones to derive masses and ages .
figure 9 indicates the mass and age distributions .
the median mass is 0.30 @xmath2 . for the age distribution we obtained a median of 5 @xmath5 10@xmath6 yr . .
the median age is , width=642 ] the mass histogram rises from about 2.5 @xmath2 down to 0.4 @xmath9 , then falls off .
we note a substantial decrease in the median mass of the stars in the chamaeleon i dark cloud , from @xmath0 0.55 @xmath2 @xcite to about @xmath0 0.30 @xmath2 .
this is mainly due to the increasing number of mostly low mass and fainter members of the cloud detected by recent deep infrared survey of the cloud .
the distribution of ages indicates an active star - formation episode within the last @xmath0 5 @xmath5 10@xmath6 yr and a decreasing rate at older ages ( few @xmath5 10@xmath7 yr ) .
@xcite intensively surveyed the central 300 arcmin@xmath4 of the chamaeleon i dark cloud , at different wavelengths combining both photometric and spectroscopic information .
they detected 13 new members of the cloud and used @xcite and @xcite models to identify 4 bona fide brown dwarfs , 6 transition ( stellar / substellar ) objects and 3 very low mass objects . in total
this area contains 22 young stars with masses @xmath22 1 @xmath2 . assuming the pre - main sequence population in this region is essentially complete they measured the imf and obtained a roughly flat behavior in logarithmic mass units within @xmath0 1 and @xmath0 0.03 @xmath2 .
they also found that the majority of identified members in the region ( 19 out of 22 ) have ages near 2 @xmath5 10@xmath69 yr and only 3 objects are significantly older , with ages of @xmath0 2 @xmath5 10@xmath7 yr .
the imf of different star - forming regions ( such as @xmath23 oph , trapezium , ic 348 , l1495e and taurus ) is approximately flat from the substellar regimen to @xmath0 0.60.4 @xmath2 ( see * ? ? ? * and the references therein ) .
we interpret the fall off of the current imf for the whole chamaeleon i dark cloud as population selection effect . assuming that the _ true _ shape of the function is roughly flat in logarithmic mass bins from the subsolar to the substellar regimen and that the census of young stars in the cloud for masses @xmath3 0.4 @xmath2 is basically complete , we estimate that @xmath0 100 stars with masses between 0.4 and 0.04 @xmath2 remain to be identified in the cloud .
we present near - infrared low resolution ( r @xmath0 500 ) spectra of 46 candidate young stellar objects in the chamaeleon i cloud with _ typically _ @xmath71 .
we also observed 63 previously known young stellar objects belonging to this star - forming region and other southern hemisphere clouds .
both groups have similar spectroscopic characteristics .
we used the water vapor indexes to derive spectral types for the new sources and adopted dantona & mazzitelli ( 1998 ) evolutionary tracks and isocrones to estimate masses and ages .
the majority of the new objects have masses between 0.7 and 0.15 @xmath2 ; with only two sources with masses in the substellar regimen
. one of these stars ( pmk99 ir cha ina1 , with a mass of @xmath0 0.06 @xmath2 ) is associated with the bipolar outflow detected by @xcite in the northern part of the cloud .
the other object , gk2001 18 , has an estimated mass of 0.04 @xmath2 . the mass distribution rises from about 2.5 @xmath2 , has a peak around 0.4 @xmath9 and then falls towards lower masses .
the median mass for the 145 compiled stars in the chamaeleon i dark cloud is @xmath0 0.30 @xmath2 , representing a significant decrease with respect to the @xmath0 0.55 @xmath2 derived by @xcite from @xmath0 80 known members .
the scarcity of very low mass objects is interpreted as an observational selection effect toward the fainter ( and thus least massive ) members of the cloud .
we assume the _ true _ imf of the whole cloud is roughly flat in logarithmic mass bins in the interval 0.4
0.04 @xmath2 as found by @xcite in the central 300 arcmin@xmath4 region .
if the current pre - main sequence population with masses @xmath3 0.4 @xmath2 is complete , we estimate that @xmath0 100 stars with masses within these limits remain to be identified in the cloud .
the age distribution indicates an active star - formation episode within the last @xmath0 5 @xmath5 10@xmath6 yr and a decreasing rate at older ages ( few @xmath5 10@xmath7 yr ) .
the new 46 candidates analyzed in this paper are , on average , roughly 1.5 mag fainter in the @xmath19 band than those included in @xcite sample .
these new objects include only two substellar objects ; the rest of the stars have sub - solar masses down to @xmath0 0.15 @xmath2
. however , @xmath0 150 fainter candidate members of the chamaeleon i cloud ( _ typically _ @xmath21 13 ) remain to be spectroscopically characterized @xcite .
these objects may contain at least part of the missing population within the interval 0.4 0.04 @xmath2 .
near - infrared spectra , in combination with pre - main sequence evolutionary tracks and isocrones , would be useful to determine the masses and ages of these objects .
this information would provide a precise determination the behavior of the imf , particular towards and into the substellar regimen for the whole cloud .
we thank leonardo vanzi ( eso ) and robert blum ( ctio ) for technical support with sofi and osiris , respectively .
an anonymous referee provided useful suggestions that improved the content and presentation of this paper .
this research has made use of the simbad database , operated at cds , strasbourg , france .
dm gratefully acknowledges support from universidad de chile fondap 15010003 and did project i021 - 98/2 .
lidman , c. , cuby , j - g . ,
vanzi , l. , & hainaut , o. 2000 , sofi user s manual , doc . no .
lso - man - eso-40100 - 0003 , issue 1.3 ( available at : http://www.ls.eso.org/lasilla/telescopes/newntt/sofi/manual/sofimanual.html ) persi , p. , marenzi , a. r. , olofsson , g. , kaas , a. a. , nordh , l. , huldtgren , m. , abergel , a. , andr , p. , bontemps , s. , boulanger , f. , burgdorf , m. , casali , m. m. , cesarsky , c. j. , copet , e. , davies , j. , falgarone , e. , montmerle , t. , perault , m. , prusti , t. , puget , j. l. , & sibille , f. 2000 , , 357 , 219 cce98 1 & 10 58 22.9 & @xmath7278 29 49 & osiris & 1999 may 08 & + cce98 2 & 10 59 12.3 & @xmath7278 26 40 & osiris & 1999 may 08 & + cce98 5 & 11 01 32.1 & @xmath7277 42 10 & osiris & 1999 may 07 & + cce98 8 & 11 02 47.1 & @xmath7277 38 10 & osiris & 1999 may 08 & + cce98 9 & 11 03 11.5 & @xmath7277 36 36 & osiris & 1999 may 08 & iso - chai 16 + cce98 12 & 11 04 11.2 & @xmath7277 50 13 & sofi & 1999 apr 12 & iso - chai 40 + gk2001 3 & 11 04 49.2 & @xmath7278 04 46 & osiris & 2002 feb 28 & + cce98 14 & 11 05 19.2 & @xmath7277 57 39 & sofi & 1999 apr 12 & + cce98 16 & 11 05 24.1 & @xmath7276 07 44 & osiris & 1999 may 07 & + cce98 17 & 11 05 54.4 & @xmath7277 38 43 & sofi & 1999 apr 10 & iso - chai 69 + cce98 18 & 11 05 55.2 & @xmath7277 35 13 & sofi & 1999 apr 12 & iso - chai 70 + cce98 19 & 11 06 15.9 & @xmath7277 37 51 & sofi & 1999 apr 10/12 & iso - chai 76 + cce98 20 & 11 06 18.9 & @xmath7277 35 18 & sofi & 1999 apr 12 & + gk2001 15 & 11 07 06.5 & @xmath7276 37 17 & osiris & 2002 feb 28 & + cce98 21 & 11 07 09.6 & @xmath7277 18 47 & sofi & 1999 apr 12 & iso - chai 91 + cce98 23 & 11 07 16.5 & @xmath7277 23 08 & sofi & 1999 apr 10/11 & iso - chai 97 + cce98 24 & 11 07 21.7 & @xmath7277 22 12 & sofi & 1999 apr 10/11 & byb 35,iso - chai 101 + cce98 25 & 11 07 23.7 & @xmath7277 41 25 & osiris & 1999 may 07 & iso - chai 102 + cce98 26 & 11 07 36.9 & @xmath7277 35 19 & sofi & 1999 apr 11 & iso - chai 107 + cce98 27 & 11 07 37.5 & @xmath7277 33 09 & sofi & 1999 apr 11 & iso - chai 109 + gk2001 18 & 11 07 46.9 & @xmath7276 15 17 & osiris & 2002 feb 28 & + cce98 32 & 11 08 04.2 & @xmath7277 38 43 & osiris & 1999 may 07 & iso - chai 126 + cce98 33 & 11 08 12.1 & @xmath7277 18 54 & sofi & 1999 apr 11 & iso - chai 130 + cce98 34 & 11 08 12.9 & @xmath7277 19 13 & sofi & 1999 apr 11 & iso - chai 131 + gk2001 21 & 11 08 22.6 & @xmath7276 49 19 & osiris & 2002 feb 28 & + gk2001 24 & 11 08 44.5 & @xmath7276 13 29 & osiris & 2002 feb 28 & + cce98 35 & 11 08 57.2 & @xmath7277 43 28 & sofi & 1999 apr 10 & + iso - chai 177 & 11 09 08.5 & @xmath7276 49 13 & osiris & 2002 feb 27 & + cce98 36 & 11 09 11.4 & @xmath7276 32 50 & sofi & 1999 apr 12 & ots99 2 + cce98 37 & 11 09 12.0 & @xmath7277 39 06 & osiris & 1999 may 07 & iso - chai 179 + cce98 39 & 11 09 23.3 & @xmath7276 31 14 & sofi & 1999 apr 12 & + cce98 40 & 11 09 26.9 & @xmath7276 33 33 & sofi & 1999 apr 12 & ots99 13,hjm c 1 - 3 + pmk99 ir cha ina1 & 11 09 29.4 & @xmath7276 33 28 & sofi & 1999 apr 10/12 & iso - chai 192,cce98 41,ots99 15 + pmk99
ir cha ina2 & 11 09 36.6 & @xmath7276 33 39 & sofi & 1999 apr 10 & ots99 18 + cce98 42 & 11 09 38.0 & @xmath7277 10 41 & osiris & 1999 may 07 & iso - chai 196 + cce98 43 & 11 09 43.5 & @xmath7276 33 31 & sofi & 1999 apr 10 & hjm c 1 - 21,ots99 21 + cce98 44 & 11 09 47.7 & @xmath7276 34 06 & sofi & 1999 apr 10 & hjm c 1 - 22,ots99 25 + cce98 46 & 11 09 53.4 & @xmath7277 17 16 & sofi & 1999 apr 11 & + cce98 47 & 11 09 56.7 & @xmath7277 18 26 & sofi & 1999 apr 11 & iso - chai 227 + pmk99 ir cha ina3 & 11 10 01.5 & @xmath7276 32 50 & sofi & 1999 apr 10 & ots99 30 + pmk99
ir cha ina4 & 11 10 04.5 & @xmath7276 33 09 & sofi & 1999 apr 10 & ots99 32 + cce98 48 & 11 10 11.9 & @xmath7276 35 31 & sofi & 1999 apr 11 & iso - chai 237,hjm c 1 - 8,ots99 45 + cce98 49 & 11 10 54.0 & @xmath7277 25 03 & sofi & 1999 apr 11 & iso - chai 256 + gk2001 42 & 11 11 21.1 & @xmath7278 05 20 & osiris & 2002 feb 28 & + cce98 50 & 11 11 29.5 & @xmath7276 09 29 & osiris & 1999 may 07 & + cce98 51 & 11 11 32.4 & @xmath7277 28 12 & osiris & 1999 may 07 & + lllll & & & & + + & & & & + hn 1 & 11 02 33.1 & @xmath7277 29 26 & osiris & 1999 may 08 + byb 18 & 11 04 43.0 & @xmath7277 41 57 & osiris & 1999 may 06 + chxr 73 & 11 06 26.5 & @xmath7277 37 38 & sofi & 1999 april 10 + ced 110 irs 4 & 11 06 48.1 & @xmath7277 22 29 & sofi & 1999 apr 11 + chxr 74 & 11 06 54.0 & @xmath7277 42 09 & sofi & 1999 apr 10 + uy cha & 11 06 59.5 & @xmath7277 18 54 & sofi & 1999 apr 12 + ced 110 irs 6 & 11 07 10.5 & @xmath7277 23 10 & sofi & 1999 apr 11 + hn 6 & 11 07 13.0 & @xmath7277 46 39 & osiris & 1999 may 08 + cha h@xmath10 1 & 11 07 16 & @xmath7277 35 54 & sofi & 1999 apr 10 + lkh@xmath10 332 - 17 & 11 07 20.7 & @xmath7277 38 07 & osiris & 1999 may 06 + byb 34 & 11 07 35.3 & @xmath7277 34 05 & sofi & 1999 apr 10 + chxr 26 & 11 07 37.0 & @xmath7277 33 33 & osiris & 1999 may 06 + cha h@xmath10 2 & 11 07 42.7 & @xmath7277 34 00 & sofi & 1999 apr 10 + hm 15 & 11 07 44.5 & @xmath7277 39 43 & sofi & 1999 apr 10 + cha h@xmath10 3 & 11 07 58.2 & @xmath7277 37 21 & sofi & 1999 apr 10 + sz 23 & 11 07 58.4 & @xmath7277 42 42 & sofi & 1999 apr 10 + hm 16 & 11 07 59.1 & @xmath7277 38 46 & sofi & 1999 apr 10 + vw cha & 11 08 01.8 & @xmath7277 42 29 & sofi / osiris & 1999 apr 10/1999 may 06 + glass i & 11 08 15.4 & @xmath7277 33 53 & osiris & 1999 may 06 + hm 19 & 11 08 16.9 & @xmath7277 44 37 & osiris & 1999 may 06 + cha h@xmath10 4 & 11 08 20.6 & @xmath7277 39 19 & sofi & 1999 apr 10 + cha h@xmath10 5 & 11 08 24.7 & @xmath7277 41 47 & sofi & 1999 apr 10 + cha ir nebula & 11 08 38.7 & @xmath7277 43 51 & sofi & 1999 apr 10 + cha h@xmath10 6 & 11 08 39.6 & @xmath7277 34 17 & sofi & 1999 apr 10 + chxr 78c & 11 08 54.2 & @xmath7277 32 12 & osiris & 1999 may 06 + hjm c 1 - 6 & 11 09 23.5 & @xmath7276 34 32 & sofi & 1999 apr 12 + hjm c 7 - 2 & 11 09 25.9 & @xmath7277 26 25 & sofi & 1999 apr 12 + hjm c 1 - 25 & 11 09 41.8 & @xmath7276 34 59 & sofi & 1999 apr 11 + hjm c 2 - 3 & 11 09 46.8 & @xmath7276 43 54 & sofi & 1999 apr 12 + hn 10 & 11 09 47.5 & @xmath7276 34 45 & sofi & 1999 apr 11 + byb 43 & 11 09 47.9 & @xmath7277 26 30 & sofi & 1999 apr 12 + hjm c 1 - 2 & 11 09 55.0 & @xmath7276 32 40 & sofi & 1999 apr 11 + hjm c 7 - 3 & 11 10 00.0 & @xmath7277 26 33 & sofi & 1999 apr 12 + hn 11 & 11 10 04.9 & @xmath7276 33 28 & sofi & 1999 apr 11 + gk-1 & 11 10 04.9 & @xmath7276 35 47 & sofi & 1999 apr 11 + byb 53 & 11 14 50.8 & @xmath7277 33 39 & osiris & 1999 may 07 + & & & & + + & & & & + sz 65 & 15 39 27.7 & @xmath7234 46 17 & osiris & 1999 may 09 + sz 68 & 15 45 12.9 & @xmath7234 17 31 & osiris & 1999 may 06 + sz 77 & 15 51 47 & @xmath7235 56 43 & osiris & 1999 may 08 + sz 82 & 15 56 09.3 & @xmath7237 56 06 & osiris & 1999 may 08 + sz 128 & 15 58 07.4 & @xmath7241 51 48 & sofi & 1999 apr 10 + ho lup & 16 07 00.6 & @xmath7239 02 19 & osiris & 1999 may 09 + sz 90 & 16 07 10.1 & @xmath7239 11 03 & osiris & 1999 may 07 + sz 94 & 16 07 49.6 & @xmath7239 04 29 & sofi & 1999 apr 12 + sz 96 & 16 08 12.6 & @xmath7239 08 33 & osiris & 1999 may 06
+ hk lup & 16 08 22.5 & @xmath7239 04 46 & osiris & 1999 may 06 + sz 103 & 16 08 30.3 & @xmath7239 06 11 & osiris & 1999 may 09 + eggen 2 & 16 08 42.7 & @xmath7239 06 18 & osiris & 1999 may 06 + sz 114 & 16 09 01.8 & @xmath7239 05 12 & osiris & 1999 may 08 + sz 117 & 16 09 44.3 & @xmath7239 13 30 & osiris & 1999 may 08 + sz 123 & 16 10 51.5 & @xmath7238 53 14 & osiris & 1999 may 08 + sz 119 & 16 09 57.1 & @xmath7238 59 48 & sofi / osiris & 1999 apr 10/1999 may 07 + sz 121 & 16 10 12.2 & @xmath7239 21 01 & sofi & 1999 apr 10 + sz 122 & 16 10 16.4 & @xmath7239 08 00 & osiris & 1999 may 06 + sz 124 & 16 11 53.3 & @xmath7239 02 16 & sofi / osiris & 1999 apr 10/1999 may 08 + & & & & + + & & & & + rx j0842.4 - 8345 & 08 42 22.4 & @xmath7283 45 24 & sofi & 1999 apr 11 + rx j0844.8 - 7846 & 08 44 50.8 & @xmath7278 46 55 & sofi & 1999 apr 11 + rx j0915.5 - 7608 & 09 15 30 & @xmath7276 08 00 & sofi & 1999 apr 11 + rx j0951.9 - 7901 & 09 51 50.7 & @xmath7279 01 38 & sofi & 1999 apr 11 + haro 1 - 1 & 16 21 34.4 & @xmath7226 12 25 & osiris & 1999 may 08 + rox 13 & 16 26 29 & @xmath7224 19 42 & osiris & 1999 may 07 + lkh@xmath10 104 & 18 02 54.3 & @xmath7224 20 56 & osiris & 1999 may 07 + rno 93 & 17 16 13.8 & @xmath7220 57 46 & osiris & 1999 may 07 + cce98 1 & 11.09 & 9.08 & & 7.50 & ( 1 ) + cce98 2 & 10.65 & 8.93 & & 7.47 & ( 1 ) + cce98 5 & 13.57 & 10.73 & & 8.68 & ( 1 ) + cce98 8 & & 13.75 & & 10.50 & ( 1 ) + cce98 9 & & 13.40 & & 10.01 & ( 1 ) + cce98 12 & 18.48 & 13.77 & 11.52 & 10.55 & ( 1 ) , ( 2 ) + gk2001 3 & & 14.26 & 13.23 & 12.48 & ( 2 ) + cce98 14 & 18.13 & 13.65 & 11.46 & 10.57 & ( 1 ) , ( 2 ) + cce98 16 & 11.99 & 10.00 & & 8.41 & ( 1 ) + cce98 17 & 15.52 & 11.46 & 9.77 & 9.04 & ( 1 ) , ( 2 ) + cce98 18 & 17.98 & 13.32 & 10.96 & 9.92 & ( 1 ) , ( 2 ) + cce98 19 & 16.55 & 12.59 & 10.93 & 10.26 & ( 1 ) , ( 2 ) + cce98 20 & & 16.47 & 14.19 & 13.22 & ( 1 ) + gk2001 15 & & 12.65 & 11.72 & 10.22 & ( 2 ) + cce98 21 & & 14.66 & 12.46 & 11.27 & ( 2 ) + cce98 23 & & 17.55 & 13.60 & 10.97 & ( 2 ) + cce98 24 & & 15.20 & 12.42 & 10.83 & ( 2 ) + cce98 25 & 12.32 & 9.46 & 8.09 & 7.63 & ( 1 ) , ( 2 ) + cce98 26 & 16.31 & 12.44 & 10.75 & 9.91 & ( 1 ) , ( 2 ) + cce98 27 & 18.03 & 13.60 & 11.73 & 10.91 & ( 1 ) , ( 2 ) + gk2001 18 & & 13.87 & 12.93 & 12.22 & ( 2 ) + cce98 32 & 14.39 & 11.51 & 9.67 & 8.30 & ( 1 ) , ( 2 ) + cce98 33 & & 14.28 & 11.72 & 10.47 & ( 2 ) + cce98 34 & & 12.79 & 10.28 & 8.90 & ( 2 ) + gk2001 21 & & 14.26 & 13.44 & 12.32 & ( 2 ) + gk2001 24 & & 14.27 & 13.71 & 12.92 & ( 2 ) + cce98 35 & & 15.52 & 13.86 & 12.56 & ( 2 ) + iso - chai 177 & & 13.02 & 12.20 & 11.56 & ( 2 ) + cce98 36 & & 14.42 & & 11.68 & ( 1 ) + cce98 37 & 16.14 & 12.86 & 11.17 & 10.47 & ( 1 ) , ( 2 ) + cce98 39 & 18.20 & 14.36 & 12.62 & 11.82 & ( 1 ) , ( 2 ) + cce98 40 & 17.77 & 13.15 & 11.19 & 10.17 & ( 1 ) , ( 2 ) + pmk99 ir cha ina1 & & 18.24 & 15.28 & 12.75 & ( 3 ) + pmk99 ir cha ina2 & & 18.25 & 16.72 & 15.97 & ( 4 ) + cce98 42 & 15.28 & 11.43 & 9.81 & 9.11 & ( 1 ) , ( 2 ) + cce98 43 & & 15.56 & 13.26 & 12.29 & ( 2 ) + cce98 44 & & 15.36 & 13.09 & 11.86 & ( 2 ) + cce98 46 & & 15.66 & 13.38 & 12.42 & ( 2 ) + cce98 47 & 16.58 & 12.36 & 10.62 & 9.86 & ( 1 ) , ( 2 ) + pmk99 ir cha ina3 & & 17.79 & 16.31 & 14.98 & ( 3 ) + pmk99 ir cha ina4 & & 17.47 & 15.92 & 14.36 & ( 3 ) + cce98 48 & 13.90 & 10.75 & 9.34 & 8.55 & ( 1 ) , ( 2 ) + cce98 49 & 17.51 & 13.33 & 11.71 & 10.83 & ( 1 ) , ( 2 ) + gk2001 42 & & 14.58 & 13.94 & 13.25 & ( 2 ) + cce98 50 & 12.78 & 9.88 & & 7.99 & ( 1 ) , ( 2 ) + cce98 51 & 14.90 & 11.72 & 10.17 & 9.61 & ( 1 ) , ( 2 ) + & & & & & & & + + & & & & & & & + hn 1 & 12.97 & 11.32 & 10.40 & 10.08 & w & m2 & ( 1 ) , ( 2 ) , ( 13 ) + byb 18 & 13.63 & 11.80 & 11.01 & 10.64 & w & & ( 1 ) , ( 2 ) , ( 3 ) + chxr 73 & & 12.60 & 11.24 & 10.70 & w & k3-m2 & ( 2 ) , ( 3 ) + ced 110 irs 4 & & 16.53 & 13.88 & 12.57 & & & ( 3 ) + chxr 74 & & 11.55 & 10.55 & 10.25 & & k7 & ( 2),(3),(4 ) + uy cha & 12.92 & 11.10 & 10.33 & 9.91 & c & m1.5 & ( 1),(3),(5 ) + ced 110 irs 6 & & & 14.61 & 10.81 & & & ( 3 ) + hn 6 & 13.34 & 11.11 & 10.07 & 9.72 & w & k7 & ( 1 ) , ( 2 ) , ( 3 ) , ( 13 ) + cha h@xmath10 1 & 16.4 & 13.42 & 12.76 & 12.31 & c & m7.5 & ( 3 ) , ( 6 ) + lkh@xmath10 332 - 17 & 9.31 & 7.99 & 7.21 & 6.77 & c & g2 & ( 1 ) , ( 3 ) , ( 5 ) + byb 34 & 14.43 & 12.19 & 11.30 & 10.95 & w & m5 & ( 1 ) , ( 2 ) , ( 3 ) , ( 4 ) + chxr 26 & 15.16 & 11.41 & 9.88 & 9.21 & & k7 & ( 1 ) , ( 2 ) , ( 3 ) + cha h@xmath10 2 & 15.3 & 12.21 & 11.22 & 10.65 & c & m6.5 & ( 3 ) , ( 6 ) + hm 15 & 12.56 & 10.22 & 9.05 & 8.42 & c & m0.5 & ( 1 ) , ( 3 ) , ( 5 ) + cha h@xmath10 3 & 15.0 & 12.35 & 11.69 & 11.27 & w & m7 & ( 3 ) , ( 6 ) + sz 23 & 14.58 & 11.85 & 10.62 & 9.81 & c & m2.5 & ( 1 ) , ( 3 ) , ( 4 ) + hm 16 & 12.76 & 9.71 & 8.19 & 7.16 & c & k7 & ( 1 ) , ( 3 ) , ( 4 ) , ( 5 ) + vw cha & 10.56 & 8.63 & 7.72 & 7.17 & c & k2-k5 & ( 1 ) , ( 3 ) , ( 5 ) , ( 14 ) + glass i & 10.54 & 8.58 & 7.40 & 6.22 & w & k4 & ( 1 ) , ( 3 ) , ( 5 ) + hm 19 & 13.04 & 11.29 & 10.41 & 10.12 & w & m3.5 & ( 1 ) , ( 3 ) + cha h@xmath10 4 & 14.4 & 12.13 & 11.43 & 11.08 & w & m6 & ( 3 ) , ( 6 ) + cha h@xmath10 5 & 14.7 & 12.13 & 11.21 & 10.78 & w & m6 & ( 3 ) , ( 6 ) + cha ir nebula & 15.86 & 11.12 & 9.30 & 8.07 & & & ( 1 ) , ( 3 ) + cha h@xmath10 6 & 15.1 & 12.37 & 11.53 & 11.04 & c & m7 & ( 3 ) , ( 6 ) + chxr 78c & 14.75 & 12.36 & 11.54 & 11.17 & w & m5.5 & ( 1 ) , ( 2 ) , ( 3 ) , ( 4 ) + hjm c 1 - 6 & 17.99 & 12.36 & 10.13 & 8.52 & & & ( 1 ) , ( 3 ) + hjm c 7 - 2 & & 11.64 & 10.05 & 9.34 & & & ( 7 ) + hjm c 1 - 25 & & 13.48 & 11.15 & 9.63 & & & ( 3 ) + hjm c 2 - 3 & 14.50 & 11.73 & 10.72 & 10.14 & & & ( 1 ) , ( 3 ) + hn 10 & 14.81 & 11.94 & 10.68 & 9.96 & & & ( 1 ) , ( 3 ) , ( 13 ) + byb 43 & 16.02 & 12.61 & 11.12 & 10.12 & & & ( 1 ) , ( 3 ) + hjm c 1 - 2 & & 13.42 & 10.98 & 9.25 & & & ( 3 ) + hjm c 7 - 3 & & 12.11 & 10.51 & 9.92 & & & ( 7 ) + hn 11 & 14.46 & 11.53 & 10.05 & 9.20 & c & & ( 1 ) , ( 3 ) , ( 13 ) + gk-1 & 12.86 & 10.39 & 9.48 & 9.10 & w & m0 & ( 3 ) , ( 5 ) , ( 14 ) + byb 53 & 11.89 & 10.49 & 9.76 & 9.53 & w & m1 & ( 1 ) , ( 2 ) , ( 3 ) + & & & & & & + + & & & & & & + sz 65 & 10.51 & 9.19 & 8.39 & 7.99 & c & m0-m0.5 & ( 8) , ( 14 ) + sz 68 & 8.86 & 7.65 & 6.94 & 6.55 & w & k2 & ( 8) + sz 77 & 10.83 & 9.53 & 8.72 & 8.34 & c & m0 & ( 8) , ( 14 ) + sz 82 & 10.20 & 8.87 & 8.10 & 7.75 & w & m0 & ( 8) , ( 14 ) + sz 128 & 12.70 & 11.19 & 10.29 & 9.71 & w & m1.5 & ( 8) + ho lup & 11.42 & 10.04 & 9.14 & 8.60 & c & m1 & ( 8) + sz 90 & 13.28 & 10.46 & 9.44 & 8.82 & c & k7-m0 & ( 8) , ( 14 ) + sz 94 & 12.90 & 11.50 & 10.89 & 10.63 & w & m4 & ( 8) + sz 96 & 11.84 & 10.21 & 9.35 & 8.90 & c & m1.5 & ( 8) , ( 14 ) + hk lup & 10.92 & 9.55 & 8.64 & 7.99 & c & k7-m0 & ( 8) , ( 14 ) + sz 103 & 13.37 & 11.78 & 10.90 & 10.34 & c & m4 & ( 8) , ( 14 ) + eggen 2 & 11.16 & 9.72 & 8.98 & 8.75 & w & m1 & ( 8) + sz 114 & 12.12 & 10.42 & 9.62 & 9.18 & c & m4 & ( 8) + sz 117 & 12.27 & 11.02 & 10.11 & 9.66 & c & m2 & ( 8) + sz 123 & 12.78 & 11.20 & 10.36 & 9.91 & c & m1 & ( 8) + sz 119 & 11.96 & 10.36 & 9.68 & 9.45 & w & m4 & ( 8) + sz 121 & 11.84 & 10.27 & 9.49 & 9.13 & w & m3 & ( 8) + sz 122 & 12.12 & 11.03 & 10.31 & 10.07 & w & m2 & ( 8) + sz 124 & 11.46 & 10.38 & 9.79 & 9.48 & w & k7-m0 & ( 8) + & & & & & & & + + & & & & & & & + rx j0842.4 - 8345 & 10.53 & 9.44 & 8.82 & 8.65 & w & k4-k6 & ( 9 ) , ( 10 ) + rx j0844.8 - 7846 & 10.70 & 9.61 & 8.93 & 8.72 & w & k6-m0 & ( 9 ) , ( 10 ) + rx j0915.5 - 7608 & 10.35 & 9.37 & 8.71 & 8.52 & w & k6-k7 & ( 9 ) , ( 10 ) + rx j0951.9 - 7901 & 9.29 & 8.66 & 8.20 & 8.09 & w & g7-g8 & ( 9 ) , ( 10 ) + haro 1 - 1 & & & & & c & k5-k7 & ( 11 ) + rox 13 & & & & & & & + lkh@xmath10 104 & & & & & c & k5 & ( 12 ) + rno 93 & & & & & c & k5 & ( 12 ) + cce98 1 & & 2 & 6 & 4 & 46 + cce98 2 & & & 2 & 3 & 35 + cce98 5 & & 2 & 3 & & 28 + cce98 8 & & 2 & 3 & 2 & 18 + cce98 9 & & & 3 & & 8 + cce98 12 & & 1 & 6 & 4 & 34 + cce98 14 & & & 3 & 3 & 28 + cce98 16 & & 2 & 4 & 3 & 39 + cce98 17 & & 2 & & 1 & 15 + cce98 18 & & 2 & 2 & 2 & 22 + cce98 19 & & & 5 & 5 & 20 + cce98 21 & & 1 & 4 & 5 & 19 + cce98 23 & & & 3 & 3 & 11 + cce98 24 & @xmath728 & @xmath724 & 3 & 4 & 12 + cce98 25 & & 3 & 6 & & 65 + cce98 32 & & & 5 & & + cce98 33 & & 2 & 1 & 1 & 24 + cce98 34 & & & 4 & 2 & 23 + cce98 36 & & 1 & & & 19 + cce98 37 & & & 5 & & 26 + cce98 40 & & 7 & & & + pmk99 ir cha ina1 & & & & & + cce98 42 & & & 7 & 3 & 23 + cce98 43 & & & 4 & & 21 + cce98 44 & & 8 & & & + cce98 47 & & 2 & & & 17 + cce98 48 & & & 3 & 3 & 15 + cce98 50 & & & 5 & 2 & 35 + cce98 51 & & 2 & 3 & & + & & & & & + + & & & & & + hn 1 & & & 3 & 2 & 8 & + byb 18 & & & 7 & 4 & 33 & + chxr 73 & & 2 & & & 16 & + ced 110 irs 4 & & & & & & + chxr 74 & & & 6 & 5 & 13 & + uy cha & @xmath725 & @xmath722 & 4 & 4 & 15 & + ced 110 irs 6 & & @xmath727 & & & & + hn 6 & & & 5 & 4 & 10 & + cha h@xmath10 1 & & & & & 13 & + lkh@xmath10 332 - 17 & @xmath725 & & & & & + byb 34 & & & 5 & 6 & 12 & + chxr 26 & & & 9 & & & + cha h@xmath10 2 & & & 5 & 6 & 12 & + hm 15 & @xmath725 & & 4 & 2 & 8 & + cha h@xmath10 3 & & & 3 & 2 & 14 & + sz 23 & @xmath726 & @xmath722 & 6 & 5 & 9 & + hm 16 & @xmath728 & & & & & + vw cha & @xmath728 & @xmath723 & 4 & 2 & 7 & sofi data + vw cha & @xmath726 & @xmath722 & 5 & & 6 & osiris data + glass i & & & & & & + hm 19 & & & 5 & 2 & 21 & + cha h@xmath10 4 & & & 3 & 2 & 13 & + cha h@xmath10 5 & & & 4 & 2 & 13 & + cha ir nebula & & & 2 & 2 & 16 & + cha h@xmath10 6 & & & 3 & 2 & 10 & + chxr 78c & & 3 & 3 & & & + hjm c 1 - 6 & @xmath7210 & @xmath722 & 1 & 2 & & + hjm c 7 - 2 & & 2 & 1 & 1 & 16 & + hjm c 1 - 25 & & & 3 & 3 & 7 & + hjm c 2 - 3 & 14 & 12 & & & & + hn 10 & @xmath723 & 3 & 4 & 4 & 9 & + byb 43 & @xmath7217 & @xmath724 & 3 & 2 & 9 & + hjm c 1 - 2 & & & 2 & 2 & 5 & + hjm c 7 - 3 & & 2 & 3 & 2 & 19 & + hn 11 & @xmath729 & @xmath724 & 1 & 3 & 7 & + gk-1 & @xmath722 & 1 & 6 & 4 & 23 & + byb 53 & & 2 & 3 & 3 & 8 & + & & & & & + + & & & & & + sz 65 & & & 2 & & 6 & + sz 68 & & 2 & 5 & 3 & 16 & + sz 77 & & 2 & 2 & 3 & 11 & + sz 82 & & & 3 & 2 & 4 & + sz 128 & @xmath722 & & 5 & 5 & 12 & + ho lup & @xmath7212 & @xmath722 & & & 4 & + sz 90 & @xmath728 & & & 2 & & + sz 94 & & @xmath726 & 10 & 4 & 17 & + sz 96 & @xmath722 & & 3 & & 1 & + hk lup & & & 3 & & 7 & + sz 103 & & 3 & 4 & 7 & 19 & + eggen 2 & & 4 & 4 & 6 & & + sz 114 & & & 4 & & 3 & + sz 117 & & & 3 & & 3 & + sz 123 & @xmath728 & & 2 & 3 & 2 & + sz 119 & & & 6 & 4 & 15 & sofi data + sz 119 & & & 4 & 7 & 13 & osiris data + sz 121 & & & 6 & 6 & 15 & + sz 122 & & & 4 & 5 & 5 & + sz 124 & & & 6 & 6 & 12 & sofi data + sz 124 & & & 5 & 6 & 8 & osiris data + & & & & & + + & & & & & + rx j0842.4 - 8345 & & & 6 & 5 & 8 & + rx j0844.8 - 7846 & & & 5 & 5 & 11 & + rx j0915.5 - 7608 & & & 5 & 5 & 11 & + rx j0951.9 - 7901 & & & 3 & 3 & 5 & + haro 1 - 1 & @xmath7216 & @xmath7211 & 2 & 2 & & + rox 13 & & 2 & 5 & 6 & 11 & + lkh@xmath10 104 & & 2 & & & & + rno 93 & & & 4 & 3 & 11 & + cce98 1 & & m0 & 1.2 & & @xmath730.52 & 0.28 & @xmath22 1 @xmath5 10@xmath6 + cce98 2 & & m0 & 1.0 & & @xmath730.50 & 0.28 & @xmath22 1 @xmath5 10@xmath6 + cce98 5 & & m1 & 2.0 & & @xmath730.15 & 0.275 & 1 @xmath5 10@xmath6 + cce98 8 & & m1 & 3.9 & & @xmath720.26 & 0.38 & 8 @xmath5 10@xmath6 + cce98 12 & & m0 & 4.1 & @xmath720.26 & @xmath720.19 & 0.40 & 8 @xmath5 10@xmath6 + gk2001 3 & & & & & & & + cce98 14 & & m0.5 & 4.0 & @xmath730.33 & @xmath720.20 & 0.38 & 7 @xmath5 10@xmath6 + cce98 16 & & m1 & 1.2 & & @xmath730.14 & 0.275 & 1 @xmath5 10@xmath6 + cce98 17 & m0 & m0.5 & 2.7 & @xmath720.15 & @xmath730.15 & 0.295 & 1 @xmath5 10@xmath6 + cce98 18 & m1.5 & m0.5 & 4.4 & @xmath720.28 & @xmath730.11 & 0.275 & 1 @xmath5 10@xmath6 + cce98 19 & m3 & m2 & 2.9 & @xmath720.27 & @xmath720.31 & 0.17 & 1 @xmath5 10@xmath6 + cce98 21 & m3 & m2.5 & 4.3 & @xmath720.18 & @xmath720.60 & 0.18 & 7.5 @xmath5 10@xmath6 + cce98 23 & m0.5 & m1 & 8.6 & @xmath730.1 & @xmath730.10 & 0.275 & 1 @xmath5 10@xmath6 + cce98 24 & m2 & m2 & 5.9 & @xmath720.2 & @xmath720.15 & 0.22 & 1 @xmath5 10@xmath6 + cce98 25 & & m0.5 & 1.8 & @xmath720.2 & @xmath730.61 & 0.25 & @xmath22 1 @xmath5 10@xmath6 + cce98 26 & & & & & & & + cce98 27 & & & & & & & + gk2001 18 & & m6.5 & 0.8 & @xmath730.1 & @xmath721.74 & 0.04 & 5 @xmath5 10@xmath6 + cce98 32 & & m2 & 3.4 & @xmath730.3 & @xmath730.34 & 0.17 & @xmath22 1 @xmath5 10@xmath6 + cce98 33 & & m0.5 & 4.9 & @xmath720.1 & @xmath720.06 & 0.34 & 4 @xmath5 10@xmath6 + cce98 34 & & m1 & 4.8 & @xmath720.1 & @xmath730.48 & 0.235 & @xmath22 1 @xmath5 10@xmath6 + gk2001 21 & & & & & & & + gk2001 24 & & & & & & & + iso - chai 177 & & m3 & 0.7 & @xmath730.2 & @xmath721.40 & 0.225 & 8 @xmath5 10@xmath69 + cce98 36 & & m0.5 & 3.1 & & @xmath720.86 & 0.60 & 1.5 @xmath5 10@xmath7 + cce98 37 & & m1 & 2.7 & @xmath720.2 & @xmath720.41 & 0.41 & 1.5 @xmath5 10@xmath69 + pmk99 ir cha ina1@xmath74 & & m3.5(m6.5 ) & 6.3(6.15 ) & 0.7(0.52 ) & @xmath721.25(@xmath721.36 ) & 0.19(0.06 ) & 4 @xmath5 10@xmath69(4 @xmath5 10@xmath6 ) + pmk99 ir cha ina2 & & & & & & & + cce98 42 & & m2 & 2.9 & @xmath720.2 & @xmath730.16 & 0.185 & @xmath22 1 @xmath5 10@xmath6 + cce98 43 & m2 & m0.5 & 4.3 & @xmath720.3 & @xmath720.85 & 0.37 & 4 @xmath5 10@xmath69 + cce98 47 & m0 & m0 & 2.8 & @xmath720.1 & @xmath720.16 & 0.40 & 7 @xmath5 10@xmath6 + pmk99 ir cha ina3 & & & & & & & + pmk99 ir cha ina4 & & & & & & & + cce98 48 & m0.5 & m1.5 & 1.9 & @xmath730.1 & 0.125 & 0.275 & 1 @xmath5 10@xmath6 + cce98 49 & m4 & m5 & 2.7 & @xmath720.1 & @xmath720.77 & 0.12 & 2.5 @xmath5 10@xmath6 + gk2001 42 & & & & & & & + cce98 50 & & m2.5 & 1.9 & & @xmath730.35 & 0.155 & @xmath22 1 @xmath5 10@xmath6 + cce98 51 & & m0.5 & 2.3 & @xmath720.2 & @xmath720.10 & 0.35 & 5 @xmath5 10@xmath6 + llllllll byb 18 & & m4 & 0.5 & 0.0 & @xmath720.99 & 0.155 & 2 @xmath5 10@xmath69 + cha ir nebula & m4.5 & m5.5 & 3.2 & 0.0 & @xmath730.32 & 0.11 & @xmath22 1 @xmath5 10@xmath6 + hjm c 1 - 6 & & m1 & 4.1 & @xmath730.3 & @xmath730.36 & 0.205 & @xmath22 1 @xmath5 10@xmath6 + hjm c 7 - 2 & m0.5 & m0.5 & 2.4 & @xmath720.1 & @xmath720.03 & 0.385 & 7 @xmath5 10@xmath6 + hjm c 1 - 25 & m2 & m1.5 & 4.4 & @xmath730.1 & @xmath730.015 & 0.19 & @xmath22 1 @xmath5 10@xmath6 + hn 10 & m3 & m3 & 1.8 & 0.0 & @xmath720.50 & 0.17 & 3 @xmath5 10@xmath6 + byb 43 & m1.5 & m3 & 2.5 & @xmath730.1 & @xmath720.50 & 0.17 & 3 @xmath5 10@xmath6 + hjm c 1 - 2 & & m1.5 & 4.7 & @xmath730.3 & @xmath730.15 & 0.225 & @xmath22 1 @xmath5 10@xmath6 + hjm c 7 - 3 & m0 & m0.5 & 2.4 & @xmath720.2 & @xmath720.21 & 0.345 & 4 @xmath5 10@xmath6 + hn 11 & & m1.5 & 2.1 & @xmath730.1 & @xmath720.10 & 0.265 & 2 @xmath5 10@xmath6 + rox 13 & & m1.5 & & & & & + lrllll & & & & & + + & & & & & + hn 1 & 1.0 & @xmath720.1 & @xmath720.55 & 0.25 & 1 @xmath5 10@xmath69 + chxr 73 & 1.8 & @xmath720.1 & @xmath720.59 & 0.80 & 1.5 @xmath5 10@xmath7 + chxr 74 & 0.9 & @xmath720.1 & @xmath720.55 & 0.72 & 7 @xmath5 10@xmath69 + uy cha & 1.4 & @xmath720.1 & @xmath720.22 & 0.275 & 5 @xmath5 10@xmath6 + hn 6 & 1.0 & @xmath720.05 & @xmath720.33 & 0.60 & 2.5 @xmath5 10@xmath69 + cha h@xmath10 1 & 0.1 & 0.0 & @xmath721.88 & 0.021 & @xmath22 1 @xmath5 10@xmath6 + lkh@xmath10 332 - 17 & 1.2 & 0.1 & @xmath731.22 & 2.4 & 2 @xmath5 10@xmath69 + byb 34 & 0.8 & @xmath720.15 & @xmath721.08 & 0.125 & 1.5 @xmath5 10@xmath69 + chxr 26 & 2.3 & @xmath720.1 & @xmath730.06 & 0.45 & 4.5 @xmath5 10@xmath6 + cha h@xmath10 2 & 1.0 & 0.0 & @xmath721.02 & 0.08 & 2.5 @xmath5 10@xmath6 + hm 15 & 1.3 & @xmath730.1 & @xmath730.10 & 0.30 & 2 @xmath5 10@xmath6 + cha h@xmath10 3 & 0.1 & 0.0 & @xmath721.44 & 0.045 & 2 @xmath5 10@xmath6 + sz 23 & 1.8 & 0.1 & @xmath720.47 & 0.20 & 3 @xmath5 10@xmath6 + hm 16 & 2.3 & @xmath730.3 & @xmath730.73 & 0.35 & @xmath22 1 @xmath5 10@xmath6 + vw cha & 0.8 & @xmath730.2 & @xmath730.61 & 0.70 & 2.5 @xmath5 10@xmath6 + glass i & 1.5 & @xmath730.8 & @xmath730.93 & 0.80 & 1 @xmath5 10@xmath6 + hm 19 & 0.8 & @xmath720.2 & @xmath720.67 & 0.16 & 7 @xmath5 10@xmath6 + cha h@xmath10 4 & 0.2 & @xmath720.1 & @xmath721.29 & 0.08 & 7.5 @xmath5 10@xmath6 + cha h@xmath10 5 & 0.8 & @xmath720.1 & @xmath721.06 & 0.09 & 4 @xmath5 10@xmath6 + cha h@xmath10 6 & 0.6 & @xmath720.1 & @xmath721.26 & 0.05 & 2 @xmath5 10@xmath6 + chxr 78c & 0.6 & @xmath720.1 & @xmath721.24 & 0.10 & 2 @xmath5 10@xmath69 + gk-1 & 0.6 & @xmath730.1 & @xmath720.25 & 0.43 & 9 @xmath5 10@xmath6 + byb 53 & 0.2 & 0.0 & @xmath720.47 & 0.45 & 2 @xmath5 10@xmath69 + & & & & & + + & & & & & + sz 65 & 0.3 & @xmath730.1 & @xmath730.12 & 0.32 & 2 @xmath5 10@xmath6 + sz 68 & 0.6 & @xmath730.15 & @xmath730.97 & 1.5 & 2 @xmath5 10@xmath6 + sz 77 & 0.3 & @xmath730.1 & @xmath720.01 & 0.335 & 4 @xmath5 10@xmath6 + sz 82 & 0.2 & @xmath730.1 & @xmath730.21 & 0.31 & 1 @xmath5 10@xmath6 + sz 128 & 0.6 & @xmath730.2 & @xmath720.58 & 0.35 & 2 @xmath5 10@xmath69 + ho lup & 0.6 & @xmath730.2 & @xmath720.11 & 0.33 & 5 @xmath5 10@xmath6 + sz 90 & 0.9 & @xmath730.2 & @xmath720.16 & 0.40 & 8 @xmath5 10@xmath6 + sz 94 & 0.1 & @xmath730.0 & @xmath721.06 & 0.16 & 2 @xmath5 10@xmath69 + sz 96 & 0.5 & @xmath730.1 & @xmath720.23 & 0.28 & 6.5 @xmath5 10@xmath6 + hk lup & 0.6 & @xmath730.3 & @xmath730.09 & 0.33 & 3 @xmath5 10@xmath6 + sz 103 & 0.8 & @xmath730.15 & @xmath720.89 & 0.15 & 1.5 @xmath5 10@xmath69 + eggen 2 & 0.2 & @xmath730.0 & @xmath720.15 & 0.33 & 7 @xmath5 10@xmath6 + sz 114 & 0.6 & @xmath730.0 & @xmath720.43 & 0.135 & 1 @xmath5 10@xmath6 + sz 117 & 1.0 & @xmath730.0 & @xmath720.44 & 0.24 & 8 @xmath5 10@xmath6 + sz 123 & 0.45 & @xmath730.1 & @xmath720.64 & 0.50 & 4 @xmath5 10@xmath69 + sz 119 & 0.3 & @xmath720.1 & @xmath720.53 & 0.135 & 2 @xmath5 10@xmath6 + sz 121 & 0.6 & 0.0 & @xmath720.34 & 0.17 & 1.5 @xmath5 10@xmath6 + sz 122 & 0.5 & @xmath720.1 & @xmath720.65 & 0.25 & 1.5 @xmath5 10@xmath69 + sz 124 & 0.0 & @xmath730.1 & @xmath720.48 & 0.51 & 2.5 @xmath5 10@xmath69 + & & & & & + + & & & & & + rx j0842.4 - 8345 & 0.1 & 0.0 & @xmath720.02 & 0.44 & 1.5 @xmath5 10@xmath69 + rx j0844.8 - 7846 & 0.1 & @xmath730.1 & @xmath730.11 & 0.48 & 4 @xmath5 10@xmath6 + rx
j0915.5 - 7608 & 0.0 & @xmath730.1 & @xmath720.04 & 0.68 & 7.5 @xmath5 10@xmath6 + rx j0951.9 - 7901 & 0.1 & 0.0 & @xmath730.43 & 1.5 & 7 @xmath5 10@xmath69 + | we present low resolution ( r @xmath0 500 ) near - infrared spectra of 46 candidate young stellar objects in the chamaeleon i star - forming region recently detected in several deep photometric surveys of the cloud .
most of these stars have @xmath1 12 .
in addition , we present spectra of 63 previously known southern hemisphere young stars mainly belonging to the chamaeleon i and lupus dark clouds .
we describe near - infrared spectroscopic characteristics of these stars and use the water vapor indexes to derive spectral types for the new objects .
photometric data from the literature are used to estimate the bolometric luminosities of all sources .
we apply dantona & mazzitelli ( 1998 ) pre - main sequence evolutionary tracks and isocrones to derive masses and ages .
we detect two objects with mass below the h burning limit among the 46 new candidates .
one of this object ( pmk99 ir cha ina1 ) is the likely driving source of a bipolar outflow in the northern region of the cloud . combining our targets with previously known members of the cloud we analyze the mass and age distributions for 145 stars in the chamaeleon i dark could .
the mass histogram rises from about 2.5 @xmath2 up to 0.4 @xmath2 and then falls off .
the median mass is 0.30 @xmath2 .
the current population with masses @xmath3 0.4 @xmath2 is essentially complete .
the scarcity of very low mass members is interpreted as population bias towards the least massive and fainter objects . if we assume the _ true _
chamaeleon i imf is flat ( in logarithmic mass bins ) in the interval 0.4
0.04 @xmath2 as found by @xcite in the central 300 arcmin@xmath4 region , then we estimate that @xmath0 100 stars remain to be found in that mass range .
the distribution of ages indicates an active star - formation episode within the last @xmath0 5 @xmath5 10@xmath6 yr and a decreasing rate at older ages ( few @xmath5 10@xmath7 yr ) . |
1915 was discovered by the _ watch _ instrument on - board _ granat _ in 1992 ( castro - tirado et al .
it is the first galactic source observed to have apparent superluminal motion in radio ( mirabel & rodrguez , 1994 ) , corresponding to the ejection of plasma at a speed of @xmath14 of the speed of light .
it distance is estimated to @xmath15 kpc ( chapuis & corbel 2004 ) , and the mass of the compact object in 1915 is estimated to be 14.0 @xmath16 m@xmath17 ( harlaftis & greiner 2004 ) . +
systematic monitoring in the x - rays ( mainly with the _ rossi x - ray timing explorer , rxte _ ) , revealed a rich pattern of variability on all time scales .
1915 is a source of low and high frequency qpos ( lfqpo , hfqpo , morgan , remillard & greiner 1997 ) , whose properties ( frequency , rms amplitude ) are tightly correlated to the spectral parameters ( morgan et al .
1997 , muno et al .
1999 , markwardt et al .
1999 , rodriguez et al .
2002a , b , vignarca et al . 2002 ) . when analyzing data of black hole binaries , the frequency of lfqpos
have been shown to be best correlated with the slope of the high energy tail of the energy spectra ( vignarca et al .
it should be noted that the lfqpo frequency is usually correlated with the soft x - ray flux , thought to originate from the accretion disk .
+ belloni et al .
( 2000 , hereafter b00 ) analyzing 163 _ rxte _ observations , have shown that , though complex , the behavior of 1915 could be understood as spectral transitions between 3 basic states a , b , c. they identified 12 recurrent classes of variability on a timescale of @xmath18 3000 s. 1915 spends most of the time in the so - called @xmath1 class of variability that corresponds to a steady state in the x - rays , lying in a rather hard part of the color - color diagram ( state c or hard state ) .
based on the x - ray ( spectral end temporal ) and radio properties of 1915 , 4 subclasses ( @xmath19 , @xmath20 , @xmath21 , @xmath22 ) can be distinguished .
two of them have a high level of radio emission with a flat spectrum , lfqpos , and a high energy tail ( b00 , trudolyubov 2001 , muno et al .
2001 , klein - wolt et al .
+ we monitored 1915 with the _ international gamma - ray astrophysics laboratory ( integral ) _ during its first ao , for a total of 6@xmath23100 ks ( hannikainen et al .
2003 , 2004 ) , and obtained 120 ks divided in 6 sequences of simultaneous observations with _ rxte_. one of our _ rxte _ observations was planned during an _ integral _ target of opportunity on 1915 , and allowed a wide band simultaneous spectral and temporal coverage to be performed ( fuchs et al .
2003 , hereafter f03 ) .
the global analysis of the whole campaign is dedicated to a future publication . here
we focus on the timing analysis of the 3 steady c - state _ rxte _ observations .
the data reduction methods are described in sec .
2 , while the results are presented in sec . 3 and discussed in sec . 4
the log of the observations analyzed in this paper can be found in table [ tab : log ] .
each observation covers several satellite orbits .
the analysis was first performed on each single revolution , and when no noticeable ( spectral or temporal ) evolution was found , the different orbits were further averaged to increase the statistics .
we extracted light curves from the _ proportional counter array ( pca ) _ following the standard way described in the cook book and abc of _ rxte _ , with the _
lheasoft v5.3 _ package .
good time intervals ( gti ) are defined as follows : satellite elevation over the earth limb @xmath24 , offset pointing @xmath25 , and pcu 0 and 2 turned on .
light curves were extracted from binned and event data .
we first accumulated a broad band 240 kev ( absolute channels 0 - 94 , epoch 5 ) light curve with the highest time resolution allowed by the ( binned ) data format ( @xmath26 ms ) .
we then extracted light curves in small energy bins , with the highest spectral resolution allowed by the ( binned ) data ( 16 energy bins from 2 to 14.8 kev ) , and over 7 additional spectral bins for the event data ( from 15 to @xmath1840 kev ) .
power density spectra ( pds ) were produced using _
powspec v1.0 _ , and corrected for white noise . in the case of the 240 kev light curves , pds were produced on interval length of 64 s between 15.625 mhz and 64 hz . all intervals were averaged together .
the energy dependent pdss were produced on interval of 160 s length , between 6.25 mhz and 12.8 hz . fig .
[ fig : pds ] shows , as an example , the pds s extracted in three energy bands , for obs.1 we extracted background light curves in all these energy ranges , and used their count rate to obtain the true rms amplitude following berger and van der klis ( 1994 ) .
in addition , to check for short term evolution of the qpo frequency , we produced dynamical power spectra with @xmath27 s resolution between absolute channels 0 - 35 ( @xmath28 kev ) .
the preliminary spectral analysis ( in a multi - wavelength context ) of the first observation is presented in f03 .
the 8 other sequences presented here show similar steady light curves ( see hannikainen et al .
2004 , hereafter h04 , for details on the _ integral / rxte _ campaign ) .
the _ rxte / asm _ light curve , showing the dates of our pointed _ integral / rxte _ observations is represented in fig .
[ fig : asm ] .
while the long term evolution shows a slow decay , a double flare occurs between obs . 1 and obs . 2 - 7 .
this x - ray flare is associated with a radio flare ( f03 ) , probably indicative of a discrete ejection .
we identify the class of variability of our observations as class @xmath1 of b00 .
the high level of radio emission detected during each of these observations ( f03 , h04 ) allows us to further classify the observations as class @xmath29 also known as the radio loud hard state ( muno et al .
2001 ) or type ii hard state ( trudolyubov 2001 )
. it should be noted , however , such a long term decay with the source mostly in class @xmath1 is rather peculiar , and had never been seen previously .
a preliminary spectral analysis ( f03 ; h04 ) shows that the common model of black hole x - ray binaries , i.e. a multicolor disk black body and a power - law , represents the data well .
as mentioned for such classes ( muno et al .
2001 ) , however , the disk temperature returned from the fit is too high ( 3 - 4 kev ) , and the inner radius far too small .
alternative models of broken power - law or broken power - law plus disk component fit the data well and lead to parameters closer to what is seen in other systems ( h04 ) .
we also successfully fitted the _ rxte / pca _ 3 - 25 kev spectra with a cut off power - law with a high energy cut - off of about 20 - 25 kev ( rodriguez et al .
2004 , fig [ fig : spec ] ) .
when adding higher energy spectra , such as those extracted with _ rxte / hexte _ , a large deviation to the spectrum is seen at high energy , indicating the need of an additional spectral component to the model , e.g. an extra power law ( zdziarski et al .
2001 ; hannikainen et al . in prep . ) .
this is illustrated in fig .
[ fig : spec ] ( left panel ) with the particular example of obs .
1 . furthermore ,
rodriguez et al . (
2004 ) have shown that the 20 - 400 kev combined _ rxte / hexte _ and _ integral / ibis _ and _ spi _ spectra could be fitted with a power law of photon index @xmath30 .
note that similar results were found from the osse spectral analysis of zdziarski et al .
( 2001 ) . for all sequences ,
the 240 kev pds were fitted between @xmath1815 mhz and 10 hz with a sum of 2 or 3 lorentzians ( depending on the energy range ) , to account for the wide band variability ( belloni , psaltis & van der klis 2002 ) .
a strong lfqpo is detected in all the pdss , and is modeled with an additional lorentzian ( harmonics are also detected , especially during intervals with the longest exposures ) . a first analysis of obs .
8 showed a rather broad qpo , with parameters poorly constrained .
as the dynamical power spectrum showed two distinct features , we separated this observation in sub - intervals and averaged those showing the qpo at the same frequency .
this resulted in two distinct sets of data , for which we identified two different qpos .
the lfqpo parameters are reported in table [ tab : qpos ] .
+ at first glance , there is apparently no obvious correlation between the qpo frequency and the pca 260 kev count rate ( tables [ tab : log ] & [ tab : qpos ] ) . to further verify this
, we fitted each of the @xmath31 kev 16 s pds used to construct our dynamical power spectra with a lorentzian around the average qpo frequency ( table [ tab : qpos ] ) , and could therefore obtain the variations of the qpo frequency with a time resolution of 16 s. no correlation is found between the qpo frequency and the pca @xmath32 kev count rate from obs . 1 through 7 , whereas we do find a correlation in obs .
+ we further averaged sequences showing the qpo at a similar frequency ( obs .
7 , showing the qpo at 1.06 hz is averaged with obs . 2 and obs .
4 , whereas obs . 3 , 5 and 6 are averaged together ) , and produced pds in the 22 energy bins described in sec .
these energy dependent pdss were fitted between 6.25 mhz and 10 hz . the energy dependences of the amplitude of the four distinct features are reported in fig .
[ fig : qpospec ] . + a clear difference in the shape of the energy dependence of the amplitude of the qpo appears in fig .
[ fig : qpospec ] . a clear turn - over in the amplitude vs. energy relation is visible for the @xmath33 hz qpo detected on mjd 52731 , and another one is visible for the @xmath34 hz qpo from the observation of mjds 52738 - 52739 , although it is not as clear as for the first qpo .
for the three other features we do not see any clear turn - over ( fig .
[ fig : qpospec ] ) , although a flattening is obvious at energies above 10 kev .
this may suggest that the turn - over energy evolves from one observation to the other , and is above the upper energy limit of our qpo spectra . to further test this hypothesis , we fitted the qpo spectra in _ xspec v11.3.0_. for all qpos but the 1.04 hz one , the spectra are well fitted by a cut - off power - law ( cutoffpl in _ xspec _ ) .
the fit parameters are reported in table [ tab : fit ] , while the right panel of fig .
[ fig : spec ] shows the qpo spectrum of obs .
1 with the best fit model superimposed .
it should be added here that a single power law gives a rather good representation of the 1.878 hz qpo detected in obs .
8 , with a reduced chi square of 1.89 ( 20 dof ) .
a cut - off power law model improve the fit ( table [ tab : fit ] ) , although the cut - off energy is poorly constrained ( 4.7-@xmath35 significance on this parameter ) .
it is interesting to note that the break energy seems anti - correlated with the qpo frequency , i.e. the lowest break energy is observed for the highest qpo frequency ( table [ tab : fit ] ) .
caution has to be expressed , however , since the statistical uncertainties on the break energies does not allow us to draw firm conclusion .
the presence of lfqpo in grs 1915 + 105 during class @xmath1 ( as well as during other classes ) is a known fact ( e.g. muno et al .
1999 , rodriguez et al .
2002a , b ) .
it is also known that qpo parameters depend on spectral parameters in black hole binaries ( bhb ) in general . here
we present observations taken during the same state , with few differences between the spectral parameters returned from the spectral fits .
the parameters of the qpo change dramatically from one observation to another . except in obs .
8 , the frequency of the qpo does not seem to correlate to the _ pca _ 215 kev flux or the _ asm _ 1.212 kev flux either ( although the highest frequency is observed when the _ asm _ flux is the highest , fig .
[ fig : asm ] , and table [ tab : qpos ] ) , contrary to what is usually claimed / observed .
this could indicate some definite peculiarities in obs .
1 - 7 that are taken just before and after the x - ray / radio flare ( fig .
[ fig : asm ] , f03 ) . on the other hand , obs .
8 occurs later , after grs 1915 + 105 apparently went off the linear decay phase , after the _ asm _ light curve showed some variability again . + the most striking behavior appears when studying the energy dependence of the qpo amplitude .
it is expected and a known fact that it presents a turn - over at some point ( e.g. tomsick & kaaret 2001 , rodriguez et al .
we report here , for the first time , a clear evolution of the turn - over energy between states that are spectrally similar , and have similar _ pca _ fluxes .
this `` cut - off '' energy has an origin that is unclear .
it could represent , for example , some specific temperature at which the qpo is produced , either through oscillations of a shocked boundary layer between the accretion disk , and a hot inner flow ( e.g. chakrabarti & titarchuk 1995 ) , or by a hot spot orbiting at some specific radius in the disk ( e.g. rodriguez et al .
2002a ; tagger et al .
in these two cases , however , we would expect the frequency of the qpo to scale with the inner radius of the accretion disk , and thus the soft x - ray flux ( molteni , sponholz & chakrabarti 1996 ; tagger & pellat 1999 ) , unless the soft x - rays are not uniquely produced by the accretion disk , but by another physical medium , as e.g. a compact jet ( see markoff & nowak 2004 ) .
the variations of the soft x - rays flux could be due to variations of the compact jet flux ( with a steady thermal flux from the accretion disk ) , as we discuss below .
+ the spectral approach presented in h04 , and rodriguez et al .
( 2004 ) , the systematic analysis of type ii states ( trudolyubov 2001 ) , and the detection of a hard tail up to ( at least ) 600 kev with osse ( zdziarski et al .
2001 ) raise the challenging question of the origin of the third spectral component needed to fit the high energy spectra well .
models of jet emission ( e.g. markoff et al . 2003 , markoff & nowak 2004 ) propose a jet model in which the x - ray spectrum of an xrb would represent the sum of thermal emission from the accretion disk , direct synchrotron from the jet , inverse comptonization ( either through synchrotron self - compton from the jet , or comptonization on the basis of the jet , the `` corona '' ) , reflection of these radiations on the accretion disk .
it should be noted this proposition has found an echo with the radio flux / x - ray flux correlation found in several bhbs when in the low hard state ( when the compact jet is present , e.g. corbel et al .
2003 , gallo et al .
2003 ) , but also in the case of radio loud agn ( e.g. merloni , heinz & di matteo , 2003 ; falcke , krding & markoff , 2004 ) .
our rxte observation of mjd 52731 occurred at a time when the radio flux was high and indicative of the presence of the compact jet ( f03 ) .
the level of radio emission as measured by the ryle telescope at 15 ghz is about 130 - 150 mjy during this observation , with a spectrum extending up to the near infra - red range ( f03 ) .
unfortunately we do not have such a nice coverage for the following observations , but the observation of mjds 52738 - 52739 indicates a level of 15 ghz emission higher ( @xmath36 mjy , f03 ) , that is dropping rapidly .
we remark that this observation occurred just after a radio flare indicative of a discrete ejection .
it is thus very likely that the radio emission this day partly originates from the discrete ejection ( with a different spectrum ) . during the last observation ,
the radio flux is very low compared to the two previous dates , with a level dropping from 107 mjy on mjd 52767 to 44 mjy on mjd 52769 ( h04 ) .
both our spectral analysis ( h04 , rodriguez et al .
2004 ) and the properties of the qpos ( present work ) can be understood easily , if the x - ray emission in grs 1915 + 105 during radio loud / type ii / class @xmath29 observations originate ( as proposed by markoff & nowak 2004 ) from two different physical processes ( beside the thermal emission of the accretion disk ) : comptonization and synchrotron radiation . the high energy spectrum of a source with a compact jet represents thus the sum of these different emission processes . as a result
the spectrum will therefore strongly depend on the relative contribution of each of these emission processes .
the break in the energy spectrum could be representative of the energy at which the relative contributions of these components cross each other . above the break
the contribution of the synchrotron radiation would be the dominant process to the spectrum . then the higher the relative contribution of the synchrotron component ( to the overall spectrum ) the lower the break energy is . in this case , to understand the energy dependence of the qpo amplitude , one has to assume that the qpo is contained in the comptonized flux , and not in the synchrotron flux .
then the position of the cut - off in the energy dependence of the qpo amplitude would be linked to the synchrotron flux emitted by the jet .
we find this interpretation at least qualitatively in good agreement with several observational facts : * the compact jet model has successfully been used in the fitting of different bhbs ( e.g. markoff et al .
* type ii states show a 230 kev level of variability lower than that of type i ( radio quiet ) states ( trudolyubov 2001 ) .
* a compact jet is detected during the observation showing the clear and well constrained cut - off in the energy dependence of the qpo amplitude ( obs . 1 , fig .
[ fig : qpospec ] ) * a high level of radio emission is detected during the observation taken on mjds 52738 - 52739 , a turn - over in the energy dependence of the @xmath37 hz qpo is detected ( fig .
[ fig : qpospec ] , table [ tab : fit ] , although it is absent in the spectrum of the @xmath38 hz qpo ) , while for the last observation ( mjd 52768 ) the radio flux is much lower , a single power law can fit the first qpo spectrum , and the turn over is poorly constrained ( fig .
[ fig : qpospec ] , table [ tab : fit ] ) .
we should add that the optical / uv / x - ray variability ( and presence of lfqpo in those bands ) seen in xte j1118 + 480 ( hynes et al .
2003 ) , a black hole x - ray transient in which the compact jet model has been shown to fit the broad band spectra well ( markoff et al . 2001 ) , is also compatible with our interpretation .
hynes et al .
( 2003 ) pointed out that the variability could not originate from the disk itself , but involved another non - thermal source of photons .
+ the lack of complete simultaneity , between the radio and x - ray observations , prevents us from drawing any firmer conclusions .
in addition , a cut - off in the spectrum of the compact jet is expected in the near infra - red domain .
knowing its exact position would allow us to estimate the flux expected from the jet in the hard x - rays accurately , and thus test our hypothesis .
we hope to obtain such simultaneous coverages in the near future , with _ integral _
, _ rxte _ for the high energies , but also the ryle telescope and the vla , in the radio domain , and _ spitzer _ , and ground based telescopes in the infra red domain .
the authors would like to thank g. henri for useful discussions and g. pooley for kindly providing the ryle data to our group . j.r . acknowledges financial support from the french space agency ( cnes ) .
dch acknowledges the finnish academy .
belloni t. , klein - wolt m. , mndez m. , et al .
2000 , a&a , 355 , 271 , * b00*. belloni t. , psaltis d. , van der klis m. 2002 , apj , 572 , 392 .
berger m. & van der klis m. , 1994 , a&a , 292 , 175 .
castro - tirado a.j . ,
brandt s. , lund n. 1992 , iauc 5590 .
chapuis c. & corbel s. 2004 , a&a , 414 , 659 .
chakrabarti s.k .
, titarchuk l. g. , 1995 , apj , 455 , 623 .
corbel s. , nowak m. , fender r.p .
, tzioumis a. k. , markoff s. 2003 , a&a , 400 , 1007 .
falcke h. , krding e. , markoff s. , 2004 , a&a , in press , astro - ph/0305335 fuchs y. , rodriguez j. , mirabel i.f .
2003 , a&a , 409 , l35 , * f03*. gallo e. , fender r.p . , pooley g.g . 2003 , mnras , 344 , 60 .
hannikainen d.c . , vilhu o. , rodriguez j. , et al .
2003 , a&a 411 , l415 .
hannikainen d.c . , vilhu o. , rodriguez j. , et al . 2004 , proceedings of the 5@xmath39 integral workshop , * h04*. harlaftis e. t. & greiner j. 2004 , a&a , 414 , l13 .
hynes r. i. , haswell c. a. , cui w. , et al .
2003 , mnras , 345 , 292 klein - wolt m. , fender r.p .
pooley g.g .
2001 , mnras , 331 , 745 .
markoff s. , falcke h. , fender r.p .
2001 , a&a , 372 , l25 .
markoff s. , nowak m. , corbel s. , fender r. , falcke h. 2003 , a&a , 397 , 645 .
markoff s. & nowak m. 2004 , apj in press , astro - ph 0403468 .
markwardt c.b . ,
swank j.h . , taam r.e . , 1999 ,
apj , 513 , 37 .
merloni a. , heinz s. , di matteo t. , 2003 , mnras , 345 , 1057 mirabel i.f . , rodrguez l.f .
1994 , nature , 371 , 46 .
molteni d. , sponholz h. , chakrabarti s.k .
, 1996 , apj , 457 , 805 .
morgan e.h .
, remillard r.a .
, greiner j. 1997 , apj , 482 , 993 .
muno m. p. , morgan e. h. , and remillard r. a. 1999 , apj , 527 , 321 .
muno m.p . , remillard , r.a . ,
morgan , e.h .
2001 , apj , 556 , 515 .
rodriguez j. , durouchoux p. , mirabel f. , ueda y. , tagger m. , yamaoka k. 2002a , a&a , 386 , 271 .
rodriguez j. , varnire p. , tagger m. , durouchoux p. 2002b
, a&a 387 , 487 .
rodriguez j. , corbel s. , tomsick j.a .
2003 , apj 595 , 1032 .
rodriguez j. , fuchs y. , hannikainen et al .
2004 , proceedings of the 5@xmath40 integral workshop , astro - ph 0403030 .
tagger m. & pellat r. , 1999 , a & a , 349 , 1003 .
tagger m. varnire p. , rodriguez j. , pellat r. , 2004 , apj , in press , astro - ph 0401539 .
tomsick j.a . , kaaret p. , 2001 ,
apj , 548 , 401 .
trudolyubov s. 2001 , apj , 558 , 276 vignarca f. , migliari s. , belloni t. , psaltis d. , van der klis m. 2003 , a&a , 397 , 729 .
zdziarski a.a . ,
grove e.j . , poutanen j. , rao a.r . , vadawale s.v .
2001 , apj , 554 , l45 .
.log of the rxte observations used in the present analysis .
the first of these observations was performed simultaneously with the multi - wavelength campaign discussed in fuchs et al .
observations are time ordered .
@xmath41 cts / s in the top layer of pcu # 2 , both anodes . [ cols="^,^,^,^,^,^ " , ] | we report on the timing analysis of _ rxte _ observations of the galactic micro - quasar 1915 performed in 2003 . out of a total of six times @xmath0 ks , we focus here only on the three observations during which 1915 is found in a steady c - state ( referred to as class @xmath1 ) resulting in a total of @xmath2 ks . during these observations ,
we detect low frequency quasi - periodic oscillations with high ( @xmath3 ) rms amplitude in the 240 kev energy range .
contrary to what is usually observed in grs 1915 + 105 , in most of our observations the qpo frequency present no correlation with the _ rxte / pca _ count rate , nor with the _ rxte / asm _ count rate .
we present , for the first time , high resolution ( 22 spectral channels ) 2 - 40 kev spectral fits of the energy dependence of the qpo amplitude ( `` qpo spectra '' ) .
the qpo spectra are well modeled with a cut - off power law except on one occasion where a single power law gives a satisfactory fit ( with no cut - off at least up to @xmath4 kev ) .
the cut - off energy evolves significantly from one observation to the other , from a value of @xmath5 kev to @xmath6 kev in the other observations where it is detected .
we discuss the possible origin of this behavior and suggest that the compact jet detected in the radio contributes to the hard x - ray ( @xmath7 kev ) mostly through synchrotron emission , whereas the x - ray emitted below 20 kev would originate through inverse compton scattering .
the dependence of the qpo amplitude on the energy can be understood if the modulation of the x - ray flux is contained in the comptonized photons and not in the synchrotron ones .
1915grs @xmath8+@xmath9 1550xte j@xmath10@xmath11 1655gro j@xmath12@xmath13 |
in the last 25 years more than 40 accretion - powered x - ray pulsars have been detected .
about half of these are transient , of which 12 have known be star companions .
neutron stars with be companions accrete material from the slow , dense , stellar outflow thought to be confined to the equatorial plane of the be star .
recent long term studies by the burst and transient source experiment ( batse ) on the _ compton gamma - ray observatory ( cgro ) _ revealed that be / x - ray binaries exhibit series of often periodic outbursts .
these outbursts are sometimes associated with `` giant '' outbursts accompanied by high spin - up rates and luminosities .
batse observed four additional accreting x - ray pulsars , without identified companions , which are believed to also be be / x - ray binaries because their temporal behavior closely resembles that of systems with be companions ( @xcite ) . this paper reports the discovery and temporal behavior of the fourth member of this group , gro j2058 + 42 .
a 198 second periodic signal was observed in the batse data starting on 1995 september 14 . at the same time
a new source was also detected by earth occultation measurements , which measure phase - averaged ( total ) flux .
a location was determined from both the pulsed data and the non - pulsed data with a 95% confidence error box of about 4 @xmath2 1 ( @xcite ) .
a _ cgro _ target of opportunity was declared and the spacecraft was reoriented to allow scans of the region by the oriented scintillation spectroscopy experiment ( osse ) , resulting in an improved 30 @xmath2 60 ( 95% confidence ) position ( @xcite ) .
the total flux peaked at about 300 mcrab ( 20 - 100 kev ) and the pulsed flux ( rms deviation from mean ) peaked at 140 mcrab ( 20 - 50 kev ) on 1995
september 27 .
this bright outburst continued until 1995 october 30 . a search of archival batse data from 1991 april until this bright outburst showed no previous outbursts .
an analysis of batse data following the bright outburst initially revealed three much weaker outbursts each lasting about two weeks with pulsed flux peaking at 15 - 20 mcrab ( 20 - 50 kev ) .
these outbursts were spaced by about 110 days which allowed the peak of the next outburst to be predicted .
a target of opportunity scan of the osse / batse error box was performed on 1996 november 28 with the _ rxte _ proportional counter array ( pca ) yielding a 90% confidence 4 radius error circle centered on r.a .
= 20@xmath0 [email protected] , decl .
= + 4143(@xcite ) .
batse also detected the source from 1996 november 23- december 1 .
another outburst was detected by batse about 110 days later ( 1997 march 16 - 20 . ) three shorter outbursts with peak pulsed fluxes of about 8 mcrab were detected by batse halfway between the first four 15 - 20 mcrab outbursts .
rxte _ all - sky monitor detected all 8 weak outbursts with approximately equal durations and intensities .
an archival search of _ rosat _ data found no sources within the error circle ( j. greiner , 1997 , private communication . )
no optical counterpart has been found to date .
in this paper we present the batse and _ rxte _ observations of gro j2058 + 42 .
our observations with batse include histories of pulse frequency and pulsed flux from 1995 september - 1997 march , a history of phase - averaged flux for the giant " outburst ( 1995 september - october ) , and pulse profile variations dependent upon energy and outburst phase of the giant " outburst .
an _ rxte _
pca observation of a weak outburst on 1996 november 28 includes a fit to the scan data used to better locate gro j2058 + 42 and pulse profiles .
we compare batse and _ rxte _ asm observations of the 8 weak outbursts from 1995 december - 1997 march .
we then discuss the implications of our results .
batse consists of eight identical uncollimated detector modules positioned on the corners of the _ cgro _ spacecraft such that the normal vectors of the detectors are perpendicular to the faces of a regular octahedron , providing all - sky coverage .
the batse data presented here are taken with the large - area detectors ( lads ) , which are nai(tl ) scintillation crystals with a geometric area of 2025 @xmath3 and a thickness of 1.27 cm .
the lads are sensitive to photons from 20 to 1800 kev .
two batse data types were used in this analysis , the cont ( 2.048 s , 16 energy channel ) data and the discla ( 1.024 s , 4 energy channel ) data .
a more complete description of the instrument and data types can be found in @xcite .
figure [ fig : freqhis ] shows histories of spin frequency and pulsed intensity , determined by fits to 4-day intervals of the batse 20 - 50 kev discla ( 1.024 second ) data .
short intervals of data ( 200 s ) were fit with a background model plus a sixth order fourier expansion in the pulsed phase model to generate pulse profiles at the model frequency .
( see @xcite for a detailed description of pulsed flux and pulsed frequency estimation techniques . )
the pulsed phase model initially consisted of a constant frequency obtained from daily power spectra of the batse discla data .
the data were combined into 600 second intervals with continuous background models . for each four day span , the pulse profiles from the 600 s intervals were shifted in phase according to a range of trial frequency offsets from the pulse phase model ( @xmath4 5 cycles day@xmath5 ) . then the profiles were summed for each trial frequency offset .
the best fit frequency was determined by the z@xmath6 test ( @xcite ) which measured the significance of the first 6 fourier amplitudes .
the frequency estimates were iteratively improved by generating a new phase model from the best fit frequencies and repeating the process . for gro j2058 + 42 , root - mean - square ( rms )
pulsed fluxes were estimated from the best fit pulse profile as @xmath7^{1/2}\ ] ] where @xmath8 is the pulse profile at phase @xmath9 , @xmath10 , and @xmath11 is the average flux .
the pulsed fluxes in the bottom panel of figure [ fig : freqhis ] were determined at 4-day intervals , assuming an exponential energy spectrum , @xmath12 with a normalization @xmath13 and e - folding energy @xmath14 = 20 kev . for pulse periods
longer than about 100 seconds , this frequency estimation method had an added complication if the pulse period was close to a harmonic of the spacecraft orbital period ( @xmath15 93 min ) .
the pulse period of gro j2058 + 42 was close the 28th spacecraft orbital harmonic .
relatively bright sources produced intensity steps in the count rate in source - facing detectors upon entering or exiting earth occultation .
when the data from these detectors were fit for a range of frequencies that included a spacecraft orbital harmonic , the accumulated occultation steps produced a periodic signal which was sometimes more significant than the signal from gro j2058 + 42 .
this effect was removed by discarding an interval of about 30 seconds of discla data which was centered on the occultation step of each interfering source and by allowing a corresponding discontinuity in the background model .
occultations for the crab and cyg x-1 were always removed from the data .
also , occultations for grs 1915 + 105 , sco x-1 , and gro j1744 - 28 were removed when they were bright . from figure
[ fig : freqhis ] , which shows the determined pulse frequency and pulsed flux history of gro j2058 + 42 from batse data , it is apparent that gro j2058 + 42 experienced a large spin up during its initial 46 day outburst .
the period changed by about 2 seconds across this outburst , corresponding to an average @xmath16 12 years .
measurements of the spin - up rate yielded a peak value of @xmath17 hz s@xmath5 ( where we have neglected corrections for the pulsar s orbital motion ) , which is comparable to peak spin - up rates seen by batse in `` giant '' outbursts from known be transients ( @xcite ) . in the bottom panel of figure [ fig : freqhis ] ,
the peak intensity of the initial outburst is about 10 times the peak intensity of later outbursts .
the initial outburst begins at approximately the same orbital phase as later outbursts , but peaks at a later phase .
the large spin - up and brightness of the initial outburst compared to later outbursts leads us to classify it as a `` giant '' outburst ( @xcite ) .
the total ( pulsed+unpulsed ) flux from gro j2058 + 42 was determined using earth occultation measurements .
this technique measures the intensity of a known source by calculating the difference in total count rate in source - facing detectors just before and just after occultation by the earth ( @xcite ) . for sources with long pulse periods and large pulsed fractions , flux measurements from individual occultation edges
can be biased by favoring a particular pulse phase . however ,
this bias was easily eliminated by averaging many occultation measurements to uniformly sample the pulse .
both the earth occultation measurements and the pulsed flux measurements were well fit by an exponential spectrum ( see eq .
[ eqn : spec ] ) with @xmath14 values of 15 - 22 kev .
no dependence of spectral shape on intensity was detected .
the total 20 - 70 kev flux history for gro j2058 + 42 was determined at 2-day intervals by fitting the earth occultation count rates with equation ( [ eqn : spec ] ) and assuming @xmath18 kev .
this flux history is shown in figure [ fig : occhis ] .
the weak outbursts were undetectable in earth occultation data , with an upper limit of @xmath19 mcrab ( 3@xmath20 ) .
pulse profiles for 4 day intervals were generated by epoch - folding batse cont ( 2.048 s ) data into 64 phase bins using a quadratic phase model ( again neglecting the pulsar s unknown orbit ) .
phase models were estimated at four day intervals by a grid search in frequency and frequency derivative , as described above .
each epoch - fold contained the usable data from a single spacecraft orbit .
data were excluded when the gro j2058 + 42 was occulted and during intervals containing occultation steps for crab and cyg x-1 .
gamma - ray bursts , solar flares , south atlantic anomaly passages , and electron precipitation events in the data stream were also excluded .
we searched for pulse shape variations using a statistical comparison of pulse profiles formulated in the frequency domain .
wilson et al . ( 1997 ) show that the harmonic coefficients of a pulse profile are uncorrelated . for each single spacecraft orbit
epoch - folded profile , these harmonic coefficients were calculated as @xmath21 where @xmath22 is the harmonic number , @xmath23 , @xmath24 , the number of phase bins in the pulse profile , and @xmath25 is the count rate in the @xmath26th bin of the pulse profile .
we calculated the mean value @xmath27 from the measurements of @xmath28 from the epoch - folded profiles contained within each 4-day interval corresponding to the phase model .
we calculated the error on the mean @xmath29 from the sample variance of the measurements of @xmath28 . by comparing signal power to noise power @xmath30
, we determined that only the first 6 harmonics were significant .
the average noise level was higher than expected for poisson counting noise because cyg x-1 was nearly always present in the field of view and was in an active state during the giant outburst of gro j2058 + 42 , introducing spurious noise power over a broad range of frequencies .
the observed noise power spectrum exhibited a turnover frequency consistent with cyg x-1 ( @xcite ) .
noise from cyg x-1 also affected the sensitivity of the frequency search .
we attempted to remove cyg x-1 in the time domain by simultaneously fitting the two sources using different assumed spectra . for each phase bin , fluxes were determined for gro j2058 + 42 and cyg x-1 by minimizing @xmath31 where @xmath32 is the mean subtracted count rate for the @xmath33th energy channel in a 4-day mean profile , @xmath34 is the gro j2058 + 42 flux at 30 kev , and @xmath35 is the cyg x-1 flux at 30 kev .
the parameter @xmath36 is the assumed energy spectrum for gro j2058 + 42 , @xmath37 kev)@xmath38 kev@xmath39 , convolved with the detector response to yield the expected count rate in energy channel @xmath33 .
the parameter @xmath40 was calculated in the same way using a photon power law spectrum with photon index -1.85 for cyg x-1 .
the energy channel range used corresponded to 20 - 125 kev .
gro 2058 + 42 was detectable up to about 70 kev while cyg x-1 emission was present up to about 400 kev .
this method worked well for only the brightest profiles , due to statistical limitations .
figure [ fig : cyg ] shows the fit to a profile from 1995
september 23 - 27 near the peak of the giant outburst .
noise from cyg x-1 appears to be causing high frequency features in the profile .
since noise from cyg x-1 could not be directly removed , only the first 6 harmonics of the pulse profiles were retained for pulse shape analysis , thus avoiding most of the noise effects .
groups of harmonic limited 4-day profiles were phase - aligned and averaged to improve statistical significance .
mean profiles from different epochs were compared by minimizing @xmath41 where @xmath42 and @xmath43 were the harmonic coefficients of the first and second mean profiles respectively and @xmath44 and @xmath45 were the corresponding errors .
the second mean profile was scaled by a factor @xmath46 and shifted in phase by @xmath47 to obtain the best fit .
phase aligned pulse profiles were thus constructed from batse 20 - 70 kev cont data from four intervals : ( a)1995 september 11 - 18,(b ) september 23 - 30 , ( c ) october 5 - 17 , and ( d ) october 17 - 28 .
the intensity intervals used to construct these profiles are denoted by dotted lines in figure [ fig : occhis ] and are labeled accordingly .
the profiles themselves are shown in figure [ fig : profile ] .
average count rates measured by earth occultation for these time intervals were added to the profiles shown to demonstrate the relative intensities of the profiles .
errors on these count rates were not included in the errors plotted because the profile comparisons involved only the mean subtracted profiles .
the peak - to - peak pulse fraction , @xmath48 , for these profiles was consistent with a constant value of 0.46@xmath4 0.02 .
the profiles near the peak of the outburst ( figure [ fig : profile]b ) were significantly different from profiles at lower intensities ( @xmath49 100 mcrab ) .
the main pulse appeared to evolve from a faster rise than fall to a slower rise than fall , although the primary differences in the profiles were occurring in the region between the large pulses ( phase 0.5 - 1 ) .
profile ( a ) was relatively flat in the interpulse region .
the deep dip near phase 0.6 persisted in higher channels and was most likely due to cyg x-1 .
profile ( b ) slowly declined in intensity to a minimum near phase 0.7 , after which it slowly increased in intensity .
profiles ( c ) and ( d ) both had a feature from phase 0.8 - 1.0 that did not persist at higher energies , so it was most likely intrinsic to the source .
the profile in ( a ) was significantly different from the profile in ( d ) , although the intensities were similar .
however , profiles ( c ) and ( d ) , both from the fall of the outburst were consistent with one another , although their intensities differed .
profiles during the rise of the outburst differed from those in the fall of the outburst which implies that the shape of the profile was dependent upon more than simply the mass accretion rate .
next we compared profiles for the energy intervals 20 - 40 kev and 40 - 70 kev .
we detected profile variations with energy as shown in figure [ fig : hardness ] .
the profiles shown are from the brightest interval 1995 september 23 - 30
. comparisons from fainter intervals were also suggestive of energy dependent pulse shape variations . in figure
[ fig : hardness ] , average count rates and errors from earth occultation measurements were added to the profiles .
the bottom panel is the hardness ratio given by @xmath50 where @xmath51 is the average count rate measured by earth occultation for 40 - 70 kev , @xmath52 is the mean subtracted pulse profile count rate for 40 - 70 kev , @xmath53 is the average count rate for 20 - 40 kev , and @xmath54 is the mean subtracted profile for 20 - 40 kev .
the errors shown are the errors on the difference @xmath55 , since we are interested in where the hardness ratio deviated from the mean value , 0.74 @xmath4 0.06 .
the hardness ratio demonstrated that the rise of the pulse ( phase 0.0 - 0.25 ) had a harder spectrum than later pulse phases . the proportional counter array ( pca ) on _ rxte _ consists of five proportional counters sensitive to photons from 2 - 60 kev and has a total collecting area of 6500 @xmath3 .
the pca is a collimated instrument with an approximately circular collimator efficiency with a fwhm of about 1 ( @xcite ) .
pca observations of gro j2058 + 42 were performed on 1996 november 28 .
a series of scanning observations was followed by a short ( @xmath56 ) dwell on the source .
a frequency consistent with the batse barycentered measurement , @xmath57 hz at mjd 10416 , and three higher harmonics were clearly detected during the pointed observation .
figure [ fig : xte_profiles ] shows the pulse profiles as a function of energy from the pca data .
the pulse shape appears to be changing with energy .
evidence of variations can be seen by comparing the feature near phase 0.6 in different energy ranges .
this feature is quite prominent at lower energies but is essentially nonexistent above 20 kev .
to estimate the position of gro j2058 + 42 from the pca scanning data , we analyzed the two largest scan peaks simultaneously .
the relatively constant background during this time greatly simplified the analysis .
we used the spacecraft pointing data combined with a model of the collimator response to calculate the contribution to the total 2 - 60 kev count rate from a constant ( in time ) background @xmath58 and a model source of assumed constant intensity @xmath34 ( counts s@xmath5 ) located at celestial coordinates @xmath59 , and @xmath60 .
we minimized the four parameter function @xmath61 here @xmath62 , @xmath63 are the observed and predicted count rates for time interval @xmath33 , and @xmath64 is the variance on the predicted count rate .
the best - fit source position was that which had the minimum value for @xmath65 .
we determined a statistical confidence region from the appropriate value of @xmath66 for the 3 interesting parameters ( @xmath67 . ) to the extent that the background was flat the background parameter was not relevant to the determination of the confidence region ( see @xcite ) .
we show in figure [ fig : pca_mod ] the data and best - fit model .
the resulting best - fit position was @xmath68 , @xmath60 = 4143 ( @xcite ) .
the total uncertainty in the position was not dominated by the statistical error , rather , time variability of the source as well as uncertainties in the collimator model contributed to the error budget as well .
our current understanding of these effects yielded a final circular confidence region ( 90 % ) of radius 4 centered at the above position .
the _ rxte _ all - sky monitor ( asm ) consists of three wide - angle shadow cameras equipped with xenon proportional counters with a total collecting area of 90 @xmath3 .
the asm provides 90 second snapshots of most of the sky every 96 minutes in three energy channels from 2 - 12 kev ( @xcite ) .
archival and new image data at the position of gro j2058 + 42 determined by the pca were searched by the _ rxte _ asm team and weak outbursts were detected .
single dwell ( 90 s time resolution ) data during these outbursts were epoch - folded using phase models determined from batse data , but no pulsations were detected .
figure [ fig : xte ] is a comparison of the batse pulsed flux ( 20 - 50 kev ) and the flux as measured by the _
the batse data indicated significant detections every @xmath15110 days of outbursts peaking at 15 - 20 mcrab ( 20 - 50 kev ) .
halfway between these outbursts batse detected shorter weaker outbursts with peak fluxes @xmath69 15 mcrab .
the asm data show outbursts every 54 days ( @xcite ) that were all similar in intensity and duration .
figure [ fig : orb_profile ] shows batse pulsed flux and asm flux from the first 7 weak outbursts folded at a period of 110 days with epoch julian date 2450302 .
the folding epoch was chosen near the peak of a stronger outburst in the batse data .
the outbursts near phase 0.0 were much brighter in the batse data than those near phase 0.5 , while both outbursts were similar in the asm data .
no known systematic effects would remove or weaken outbursts in the batse data at 55 day intervals .
a closer analysis of the individual _ rxte _ asm energy channels did not show a significant difference in the spectral hardness of the two sets of outbursts within the 2 - 12 kev band .
the optically identified companions for transient pulsars with pulse periods longer than one second are all be or oe stars .
long term studies of these be / x - ray binaries have demonstrated that giant outbursts followed by or interspersed with a series of periodic normal outbursts appear to be typical behavior for be / x - ray binaries ( @xcite ) .
gro j2058 + 42 is most likely a be / x - ray pulsar because it exhibits transient outbursts recurring with a presumed orbital period of 110 days and it has both giant and normal outbursts .
an orbital period of 110 days places gro j2058 + 42 along the orbital period spin period correlation for be / x - ray binaries ( @xcite , @xcite . )
the giant outburst of gro j2058 + 42 showed enough dynamic range that we can use the relationship between torque and observed flux to test accretion theory .
simple accretion theory assumes that material from the companion star is flowing onto a rotating neutron star with a magnetic field .
the magnetic field is so strong that it determines the motion of the material in a region of space surrounding the neutron star called the magnetosphere .
the size of this region , the magnetospheric radius @xmath70 , is defined to be the distance from the neutron star at which all magnetic field lines are just closed loops .
another important length scale is the corotation radius , @xmath71 , the distance from the neutron star where centrifugal forces just balance local gravity .
if @xmath72 , accretion is not expected to occur .
accretion theory predicts @xmath73 for @xmath74 for for disk or wind accretion .
the accretion torque @xmath35 is given by @xmath75 .
we can assume the bolometric flux @xmath76 is related to the mass accretion rate @xmath77 by @xmath78 .
therefore , simple accretion theory predicts @xmath79 , where @xmath80 is the spin - up rate of the neutron star ( @xcite and references therein ) . figure [ fig : flvsfd ] shows the observed 20 - 100 kev flux as measured by earth occultation plotted versus the measured frequency derivative @xmath80 during the giant outburst .
the frequency derivatives were generated by an search performed on the batse discla 20 - 50 kev data over a grid of trial frequencies and frequency derivatives .
this technique was identical to that described in section [ sec : giant ] except that the pulse profiles were shifted according to a grid of frequency offsets and frequency derivatives , rather than only frequency offsets .
the search yielded maximum spin - up rates of 5 @xmath81 hz s@xmath5 in the weak outbursts .
if we assume that the orbital contribution was small compared to the intrinsic @xmath80 , we can treat the measured @xmath80 as equal to the intrinsic @xmath80 and compare it to the total flux in the batse energy range . clearly , @xmath80 and the total flux were correlated , hence an accretion disk was likely to be present .
the power law @xmath82 with @xmath83 , which is predicted by accretion torque theory , is shown in figure [ fig : flvsfd ] .
also shown is the best fit curve with @xmath84 .
orbital contributions to the torque and the fact that batse does not measure the bolometric flux may explain the discrepancy between the measured slope and accretion theory .
interestingly , a similar fit to exosat data from the 1985 giant outburst of the be / x - ray binary exo 2030 + 375 also gave @xmath84 ( @xcite ) .
in addition to testing accretion theory , we can use the maximum @xmath80 ( @xmath85 hz s@xmath5 ) from the giant outburst to estimate a distance to gro j2058 + 42 .
the luminosity is related to @xmath80 by @xmath86 where the notation is @xmath87 g @xmath88 , @xmath89 , @xmath90 cm , @xmath91 g @xmath3 , and @xmath92 ergs s@xmath5 .
the neutron star has magnetic moment @xmath93 , radius r , mass m@xmath94 , moment of inertia i , accretion luminosity l@xmath95 , and
dimensionless torque " @xmath96 .
we assumed the very slow rotator case of @xmath97 ( @xcite and references therein ) .
we used typical values ( m = 1.4 , @xmath98 , @xmath99 ) .
since @xmath100 is less well known , we calculated bolometric luminosities , l@xmath101 ergs s@xmath5 , for the range @xmath102 .
these luminosities were comparable to the eddington limit for neutron stars , l@xmath103 ergs s@xmath5 .
next , to calculate the distance , we can use the 20 - 100 kev flux corresponding to @xmath104 , @xmath105 ergs @xmath106 s@xmath5 .
lastly , we must calculate a bolometric correction @xmath59 , the ratio between the batse flux and the total 2 - 100 kev flux .
simultaneous batse and _ rxte _ measurements from 1996 november yielded a bolometric correction of @xmath107 .
then the distance @xmath108 is given by @xmath109 which gives distances of 7 - 16 kpc for gro j2058 + 42 . for these distances
the estimated bolometric luminosity for the 1996 november outburst is l@xmath110 ergs s@xmath5 .
bolometric luminosities for the intermediate weaker outbursts ( in the batse data ) appear to be similar to the 1996 november outburst , but are more uncertain because complete spectral and pulsed fraction information are unavailable .
the weak outbursts provide interesting , but difficult to interpret , information about this system .
the batse data indicate brighter and longer outbursts every @xmath15110 days with weaker outbursts in between , while the asm data show outbursts every 54 days that are similar in intensity and duration ( fig [ fig : xte ] & @xcite ) .
the intermediate weaker outbursts in the batse data must be of a different character than the stronger outbursts . the fact that batse detects a 110 day periodicity and _ rxte _ detects a periodicity at half
that period implies that either a spectral change or a change in the pulse fraction is occurring every other outburst , indicating a difference in accretion mode .
an analysis of the three _ rxte _ asm energy bands shows that both sets of outbursts have comparable intensity in the 5 - 12 kev band .
this implies that if a spectral change is occurring it is above 12 kev .
the observed periodicity may be interpreted in two ways : ( 1 ) the 110 day cycle observed in the batse data is the orbital period and 2 outbursts are occurring each orbit ; or ( 2 ) the orbital period is 54 days and the spectrum ( or pulse fraction ) alternates every other orbit for at least 9 orbits .
we can find no plausible explanation for an alternating change in spectrum or pulse fraction that persists for at least 9 orbits if the outbursts are produced by periastron passage in a 54 day orbit .
therefore we interpret the observed 110 day periodicity as the orbital period .
we propose that gro j2058 + 42 is undergoing periastron and apastron outbursts in a 110 day orbit .
a neutron star in an inclined orbit combined with concentration of material in the equatorial plane of the be star is a possible explanation for the outburst behavior observed in gro j2058 + 42 .
the same velocity kicks that produce an eccentric orbit could also produce an inclined orbit .
outbursts would occur twice per orbit when the neutron star crosses the equatorial plane of the companion ( @xcite ) . from the batse data
, we measured the separation between the brighter outbursts to be 110 @xmath4 3 days and the separation between brighter and weaker outbursts to be @xmath111 2.3 days .
if we assume the separation between the brighter outbursts is the orbital period , @xmath112 , then @xmath113 falls in the 95% confidence interval @xmath114 . for this interval
the separation between periastron and the line where the neutron star s orbit intersects the equatorial plane of the companion is @xmath115 12.1 for a typical eccentricity , @xmath116 .
hence this mechanism would produce outbursts near periastron and near apastron .
this mechanism does not , however , explain the intensity differences seen only in the higher energies ( 20 - 50 kev ) .
although an inclined orbit has often been proposed as a mechanism for outbursts in be / x - ray binaries , two outbursts per orbit have not been seen in other be / x - ray binaries to date .
however , such behavior has been observed in wind - fed systems such as gx 301 - 2 ( @xcite ) .
another explanation which does not require an inclined orbit suggests that two different accretion mechanisms could be at work in this system .
an accretion disk is likely to be present during giant outbursts .
its presence helps to explain the large and steady spin - up rates seen during giant outbursts .
a peak spin - up rate of 2.5 @xmath117 hz s@xmath5 was observed in gro j2058 + 42 .
bildsten et al .
( 1997 ) suggest that an accretion disk could remain after a giant outburst
. then normal outbursts could be produced by large tidal torques on the accretion disk during periastron passage .
this could explain the series of normal outbursts following the giant outburst .
the apastron outbursts could be produced by accretion from the equatorial outflow from the be star .
a slow dense outflow could produce outbursts at apastron ( @xcite ) .
the two outbursts per orbit would be produced by different accretion mechanisms which could produce differences in spectra or in pulse fraction . clearly , additional observations are needed to fully understand this source .
jahoda , k. , swank , j. h. , giles , a. b. , stark , m. j. , strohmayer , t. e. , zhang , w. & morgan , e. h. 1996 , euv , x - ray , and gamma - ray instrumentation for space astronomy vii , ed . o. h. w. siegmund & m. a. grummin , spie 2808 , 59 | gro j2058 + 42 , a transient 198 second x - ray pulsar , was discovered by the burst and transient source experiment ( batse ) on the _ compton gamma - ray observatory ( cgro ) _ , during a giant " outburst in 1995 september - october .
the total flux peaked at about 300 mcrab ( 20 - 50 kev ) as measured by earth occultation .
the pulse period decreased from 198 s to 196 s during the 46-day outburst .
the pulse shape evolved over the course of the outburst and exhibited energy dependent variations .
batse observed five additional weak outbursts from gro j2058 + 42 , each with two week duration and peak pulsed flux of about 15 mcrab ( 20 - 50 kev ) , that were spaced by about 110 days .
an observation of the 1996 november outburst by the _ rossi x - ray timing explorer ( rxte ) _ proportional counter array ( pca ) localized the source to within a 4 radius error circle ( 90% confidence ) centered on r.a .
= 20@xmath0 [email protected] , decl .
= 4143 ( j2000 )
. additional shorter outbursts with peak pulsed fluxes of about 8 mcrab were detected by batse halfway between the first four 15 mcrab outbursts .
the _
rxte _ all - sky monitor detected all 8 weak outbursts with approximately equal durations and intensities .
gro j2058 + 42 is most likely a be / x - ray binary that appears to outburst at periastron and apastron .
no optical counterpart has been identified to date and no x - ray source was present in the error circle in archival _ rosat _ observations . |
the volume operator in loop quantum gravity ( lqg ) plays a central role in the quantum dynamics of the theory . without it ,
it is not possible to define the hamiltonian constraint operators @xcite , the master constraint operators @xcite , or the physical hamiltonian operators @xcite . the same applies to truncated models of lqg such as loop quantum cosmology ( lqc ) @xcite which is believed to describe well the homogeneous sector of the theory . in lqc
the quantum evolution of operators that correspond to classically - singular dirac observables typically remains finite .
this can be traced back to the techniques used to quantize inverse powers of the local volume that enter into the expression for the triad , co - triad and other types of coupling between geometry and geometry , or geometry and matter @xcite .
specifically , such derived operators arise as commutators between holonomy operators and powers of the volume operator . in view of its importance , it is of considerable interest to verify whether the classical limit of the lqg volume operator coincides with the classical volume . by this
we mean that ( i ) the expectation value of the volume operator with respect to suitable semiclassical states which are peaked at a given point in phase space , coincides with the value of the classical volume at that phase space point , up to small corrections ; and ( ii ) its fluctuations are small .
it should be remarked that there are actually two versions of the volume operator that have been discussed in the literature @xcite .
these come from inequivalent regularisations of the products of operator - valued distributions that appear at intermediate stages .
however , only the operator in @xcite survives the consistency test of @xcite , namely , writing the volume in terms of triads which then are quantised using the commutator mentioned above , gives the same operator up to @xmath0-corrections as that obtained from direct quantisation .
this consistency check is important as otherwise we could not trust the triad and co - triad quantisations that enter the quantum dynamics .
a semiclassical analysis of the volume operator has not yet been carried out , although , in principle , suitable semiclassical ( even coherent ) states for lqg are available @xcite .
this is because the spectral decomposition ( projection - valued measure ) of the volume operator can not be computed exactly in closed form .
however , this is needed for exact , practical calculations .
more precisely , the volume operator is the fourth root of a positive operator , @xmath1 , whose matrix elements can be computed in closed form @xcite but which can not be diagonalised analytically . in more detail , the volume operator has a discrete ( that is , pure - point ) spectrum and it attains a block - diagonal form where the blocks are labelled by the graphs and spin quantum numbers ( labelling the edges of the graph ) of spin - network functions ( snwf ) @xcite .
snwf form a convenient basis of the lqg hilbert space @xcite which carries a unitary representation of the spatial diffeomorphism group and the poisson@xmath2algebra of the elementary flux and holonomy variables @xcite .
the blocks turn out to be finite - dimensional matrices whose matrix elements can be expressed in terms of polynomials of 6j symbols . fortunately , these complicated quantities can be avoided by using a telescopic summation technique @xcite related to the elliot - biedenharn identity @xcite , so that a manageable closed expression results .
however , the size of these matrices grows exponentially with increasing spin quantum numbers and , since the expression for coherent states is a coherent superposition of snwf s with arbitrarily large spin quantum numbers , a numerical computation of the expectation value using the numerical diagonalisation techniques , that are currently being developed @xcite , is still a long way off .
one way forward is to use the semiclassical perturbation theory developed in @xcite , and applied already in @xcite .
the basic idea is quite simple . in practical calculations
one needs the expectation value of @xmath3 where @xmath4 is a rational number in the range @xmath5 . in order to attain
that , one has to introduce the ` perturbation operator ' , @xmath6 , where the expectation value @xmath7 of the positive operator @xmath1 is exactly computable .
notice that @xmath8 is bounded from below by @xmath9 .
then trivially @xmath10^q>$ ] .
now we exploit the existence of positive numbers @xmath11 such that the classical estimates @xmath12 are valid for all @xmath13 .
finer estimates of this form involving arbitrary powers of @xmath14 are also available @xcite . by virtue of the spectral theorem ,
this classical estimate survives at the quantum level and we have @xmath15 where @xmath16 .
it follows that @xmath17 $ ] .
however , @xmath18 and @xmath19/<q>^2 $ ] is proportional to the relative fluctuations of @xmath1 , which are of order of @xmath0 @xcite .
it follows that to zeroth - order in @xmath0 we may replace @xmath20 by @xmath21 which is computable , as are the error estimates in the way shown above .
the exact computation of @xmath22 is feasible , but it is still quite involved . at this point
it is convenient to use the fact that ( see @xcite ) , to zeroth - order in @xmath0 , the computation of these expectation values ( more generally , the expectation values of low - order polynomials in the flux operators ) for @xmath23 spin - network states coincides with the corresponding calculations for @xmath24 spin networks . on the latter ,
the volume operator is even diagonal .
hence we conclude that , as long as we are only interested in the zeroth - order in @xmath0 contribution , we may evaluate the expectation value of the volume operator for a fictitious theory in which the non - abelian group @xmath23 is replaced by the abelian group @xmath24 .
this dramatically simplifies all the calculations .
the coherent states developed so far for lqg have all been constructed using the complexifier method reviewed in @xcite .
this generalises the coherent - state construction for phase spaces that are cotangent bundles over a compact group ; see @xcite .
this construction involves computing the heat - kernel evolution of the delta distribution ( which is the matrix element of the unit operator in the schrdinger ( position ) representation ) with respect to a generalized laplace operator , called the ` complexifier ' , followed by a certain analytic continuation .
now the unit operator in lqg can be written as a resolution of unity in terms of snwf s and , although the heat kernel is a damping operator , since the snwf are not countable ( the lqg hilbert space is not separable ) , the resulting expression is not normalisable .
however , it does give a well - defined distribution ( in the algebraic dual of the finite linear span of snwf s ) which can be conveniently written as a sum over cut off states labelled by finite graphs .
these states , called `` shadows '' in @xcite , _ are _ normalisable , and can be used to perform semiclassical calculations .
of course , one expects that the good semiclassical states will only be those cut
off states that are labelled by graphs which are sufficiently fine with respect to the classical three - metric to be approximated .
thus the input in the semiclassical calculations consists in the choice of a complexifier and the choice of a graph .
one may wonder why the complexifier has to be chosen _ a priori _ rather than being dictated by the dynamics of the theory : that is , why do the coherent states remain coherent under quantum time evolution ?
for example , for the harmonic oscillator , the complexifier is just the laplacian on the real line .
the reason is two - fold . on the one hand
, one could perform a constraint quantisation so that we are working at the level of the kinematical hilbert space on which the quantum constraints have not yet been imposed .
in this situation , all that is needed is to ensure that the volume operator , which is not gauge invariant ( and thus does not preserve the physical hilbert space ) , has a good classical limit on the kinematical hilbert space .
we would certainly not trust a constraint quantization for which even that was not true .
on the other hand , one could also work at the level of the physical hilbert space , as outlined in @xcite .
then , one would expect that the time evolution of any reasonable choice of complexifier coherent states with respect to the physical hamiltonian should remain coherent ( and peaked on the classical trajectory ) for sufficiently short time intervals .
note that for interacting theories , even the simple example of the anharmonic oscillator , globally stable coherent states have yet to be found .
the choice of the complexifier will be guided by practical considerations : namely ( i ) it is diagonal on snwf s ; and ( ii ) it is a damping operator that renders normalisable the heat - kernel evolution of the delta distribution restricted to a graph .
moreover , it should be gauge invariant under the @xmath23 gauge group .
as for the choice of graph , for practical reasons one will use graphs that are topologically regular , that is , have a constant valence for each vertex .
indeed , the semiclassical calculations performed in @xcite were done using graphs of cubic topology , with good results .
the existing literature on such coherent states for lqg can be divided into two classes , depending on a certain structure that defines them .
the first are _ gauge - covariant flux _ coherent states , which depend on collections of surfaces and path systems inside them @xcite .
the second are _ area _ coherent states , which also depend on collections of surfaces but involve area operators rather than flux operators @xcite . in our companion paper @xcite we have reviewed and generalised both constructions . in its most studied incarnations ,
the collection of surfaces involved in @xcite is defined by a polyhedronal partition of the spatial manifold , while the collection used in @xcite uses a plaquette of foliations of the spatial manifold .
the main objective of this series of two papers is to analyse the semiclassical properties of the volume operator with respect to both classes of coherent state . in @xcite
we employed the states of @xcite .
in this paper we employ the states of @xcite .
this exhausts all known coherent states for lqg .
the result of our analysis is that , if we use the states @xcite , _ the correct semiclassical limit is attained with these states for @xmath25 only_. if instead we use the states @xcite , the correct semiclassical limit is attained _ only _ for : 1 ) artificial rescaling of the coherent state label ; and 2 ) particular embeddings of the 4-valent and 6-valent graphs with respect to the set of surfaces on which the complexifier depends .
however , the combinations of euler angles for which such embeddings are attained have measure zero in so(3 ) , and are therefore negligible . thus the _ area
complexifier coherent states are not the correct tools with which to analyse the semiclassical properties of the volume operator_. if one wants to obtain embedding independence , a possible strategy is to sample over graphs ( dirichlet - voronoi sampling @xcite ) as outlined in @xcite .
what this strategy amounts to is that , instead of singling out one particular coherent state @xmath26as defined in terms of a single graph @xmath27one considers an ensemble of coherent states constructed by averaging the one - dimensional projections @xmath28 onto the states @xmath26 over a subset @xmath29 of the set of all allowed graphs . in other words ,
one considers a mixed state ( with an associated density matrix ) rather than a single coherent state .
in such a way , if the subset @xmath29 is big enough , it can be shown ( @xcite ) that it is possible to eliminate the embedding dependence ( the ` staircase problem ' for a surface @xmath30 .
if we compute the expectation value for @xmath31 with respect to a coherent state @xmath26 , such that the surface @xmath30 intersects transversely one and only one edge @xmath32 of @xmath27 , then the expectation value of the area operator coincides with the classical value @xmath33 .
however , if the surface @xmath30 lies transversally to the edges , then we do not obtain the correct classical limit . ] ) .
it is straightforward to deduce that the area complexifier coherent states can not be used to construct embedding - independent , mixed coherent states because of condition 2 ) above .
we thus claim that _ area complexifier coherent states should be ruled out as semiclassical states altogether if one wants to attain embedding independence_. instead one should use the flux coherent states , as was done in @xcite . for such states
we will show that _ the correct semiclassical limit is attained only for @xmath25 _ .
in other words , _ up to now , there are no semiclassical states known other than those with cubic - graph topology ! _
thus the implication of our result for lqg is that the semiclassical sector of the theory is spanned by snwf that are based on cubic graphs .
this has some bearing for spin - foam models @xcite which are supposed to be
but , so far , have not been proved to be the path - integral formulation of lqg .
spin foams are certain state - sum models that are based on simplicial triangulations of four manifolds whose dual graphs are therefore 5-valent .
the intersection of this graph with a boundary three - manifold is 4-valent , and therefore we see that spin - foam models , based on simplicial triangulations , correspond to boundary hilbert spaces spanned by spin - network states based on 4-valent graphs only . however , we have proved that the correct semiclassical states for analysing the semiclassical properties of the volume operator are the _ gauge covariant flux _ states . for such states , only those of cubic topology
give the correct semiclassical value of the volume operator .
even if the mismatches between the 4-valent sector of lqg and the boundary hilbert space of spin foams could be surmounted , the result of our analysis seems to be that _ the boundary hilbert space of current spin - foam models does not contain any semiclassical states !
_ this apparently contradicts recent findings that the graviton propagator derived from spin - foam models is correct @xcite .
however , it is notable that these latter results only show that the propagator has the correct fall - off behaviour : the correct tensorial structure has not yet been verified . one straightforward way of possibly repairing this situation is to generalise spin - foam models to allow for arbitrary in particular , cubic triangulations , as suggested in @xcite . in the present paper we will perform detailed calculations of the expectation value of the volume operator for the 4- , 6- , and 8-valent graphs using the area coherent states reviewed in @xcite .
the main difference between the flux coherent states and the area coherent states is that the former depend on a discrete set of flux operators while the latter depend on a continuous family of area operators .
in particular , the surfaces on which the area operators depend fill all of space .
hence the question arises whether the area coherent states depend less severely on the relative orientation of the chosen cut - off graph with respect to the system of surfaces .
specifically , it would be interesting to know if the calculations of the expectation values are insensitive to euclidean motions of the graph ( translations and rotations ) when the spatial metric to be approximated is the euclidean one .
the structure of the paper is as follows . in section [ s4 ]
we calculate the expectation value for a general graph .
the expectation value depends on the edge metric , already encountered in @xcite , which is close to diagonal when the edge lengths , @xmath34 , are much larger than the lengths , @xmath35 , of the boundaries of the surfaces . in order to compute the off - diagonal metric entries we stick to the graphs defined in our companion paper as
otherwise we encounter severe ` bookkeeping ' problems .
these graphs are dual to tetrahedronal , cubical and octahedronal triangulations , all of which derive from a cubulation of the spatial manifold , as discussed in @xcite .
one can expand the expectation value into a convergent power series in @xmath36 .
we establish that , contrary to the results in @xcite , to zeroth - order in both @xmath0 and @xmath36 the correct expectation value is obtained for @xmath25 and @xmath37 , but _ only _ for specific embeddings of the graph with respect to the surfaces on which the area operators depend .
however , we show that such embeddings have measure _ zero _ in @xmath38 .
moreover , by using the coherent states in @xcite we also find that the graphs ( 4- , 6 , and 8-valent ) are all euclidean invariant , up to sets of measure zero in @xmath38 and up to linear order in @xmath36 .
nonetheless , the embeddings with non - zero measure all fail to give the correct semiclassical value for the volume operator .
this shows that coherent states @xcite are not the correct tools for analysing the semiclassical properties of the volume operators .
in section 3 we perform an explicit calculation of the expectation value of the volume operator for 4- , 6- , and 8-valent graphs , including the higher - order effects in @xmath36 .
this contribution is only euclidean - invariant up to linear order . for each of these valences
we first consider a graph constructed in terms of regular simplicial , cubical and octahedron cell complexes , as done in @xcite .
after that , we study the effects of rotations and translations .
in section 4 we summarise our results and draw appropriate conclusions .
the more technical calculations have been transferred to an appendix in order not to distract from the line of argument in the main text .
in this section we compute the expectation value of the operator @xmath39 for an arbitrary @xmath40-valent vertex , @xmath41 , for the stack family coherent states using the replacement of @xmath23 by @xmath24 .
this uses the calculational tools developed in @xcite , to which we refer for our notation .
we may , therefore , replace the @xmath23 right - invariant vector fields by @xmath24 right - invariant vector fields @xmath42 acting on @xmath43 .
the crucial simplification is that these vector fields mutually commute .
their common eigenfunctions are the spin - network functions which , for @xmath24 , take the explicit form [ 4.0.5 ] t_,n(a)=_ee ( ) _ j=1 ^ 3 [ a^j(e)]^n_j^e we will refer to them as ` charge network states ' because the @xmath44 are integer valued . using the spectral theorem we may immediately write down the eigenvalues of @xmath39 on @xmath45 as [ 4.0.6 ] _
n , v=_p^3 what follows is subdivided into four parts .
we begin by performing the calculation for a general graph .
this leads to the inverse of the edge metric which , for large graphs , is beyond analytical control . in the second part
we restrict the class of graphs , which lets us perform perturbative computations of the inverse of the edge metric .
this gives a good approximation of the actual expectation value . in the third and fourth parts we consider the dependence of our results on the relative orientation of the graph with respect to the family of stacks .
the coherent state associated with an @xmath40-valent graph in which more than one edge intersects a given plaquette @xmath30 , is as follows : [ 4.1 ] _ z , = _ ne^- n^ttn e^z^t n t _ , n the norm of the coherent states is given by [ 4.2 ] ||_z,||^2 = _
n e^-n^t t n e^2 pn where [ 4.2a ] p^j(e)=i_e ( z - a)= : e^e_j the length parameter , @xmath50 , that appears here is generally different from the parameter , @xmath51 , that enters the classicality parameter @xmath52 , as explained in @xcite .
the expectation value for the volume operator is [ 4.3 ] < > _ z,= = where [ 4.4 ] _
, v(n)=_p^3 are the eigenvalues of the volume operator @xmath39 .
we have introduced the notation [ 4.5 ] _
e e e^(n):= ^jkln^e_j
n^e_k n^e^_l the semiclassical limit of the volume operator is obtained from ( [ 4.3 ] ) in the limit of vanishing @xmath53 .
that is , it is the zeroth - order in @xmath53 of the expansion of ( [ 4.3 ] ) in powers of @xmath53 .
since ( [ 4.3 ] ) converges slowly for small values of @xmath53 , we will perform a poisson transform which replaces @xmath53 by @xmath54 which converges quickly . to this end ,
analogous to @xcite , we introduce the following variables [ 4.6 ] t_e:=t_ee , t_e:=,x^e_j:= t_e n^e_j , y^e_j(x):=x^e_j / t_e , c^e_j:=p^e_j / t_e , w^e_j(n):=n^e_j / t_e , a_e
e^jk:= a_e e ^jk:= notice that the diagonal entries of @xmath55 equal unity .
the off - diagonal ones , however , are bounded from above by unity by the schwarz inequality applied to the scalar product defined by @xmath55 and are restricted to the six - dimensional subspace restricted to vectors with non - zero entries for the @xmath56 components only . in order to perform the gaussian integrals in ( [ 4.8 ] )
we notice that , by construction , @xmath55 is a positive - definite , finite - dimensional matrix , so that its square root , @xmath58 , and its inverse are well defined via the spectral theorem .
hence we introduce as new integration variables [ 4.9 ] z^e_j:=_e,k ( ) ^ee_jk x^e_k , y^e_j(z)= x^e_j = _ e k ( [ ] ^-1])^ee_jk z^e_k the jacobian from the change of variables drops out in the fraction , and ( [ 4.8 ] ) becomes [ 4.10 ] < > _ z , = now one would like to shift @xmath59 into the complex domain by @xmath60 and then perform the ensuing gaussian integral .
this is unproblematic for the denominator of ( [ 4.8 ] ) which is analytic in @xmath59 ; however , the numerator is not .
the careful analysis in @xcite shows the existence of branch cuts in @xmath61 of the fourth - root function involved . in turn
, this shows that , in the semiclassical limit , both numerator and denominator are dominated by the @xmath62 term while the remaining terms in the series are of order @xmath63 ( i.e they decay as @xmath64 for some @xmath65 ) .
see @xcite for the details .
the upshot is that to any polynomial order in @xmath0 we may replace ( [ 4.10 ] ) by [ 4.11 ] < > _ z , = which is now defined unambiguously because the argument of the fourth root is the square of a real number .
the denominator of ( [ 4.11 ] ) simply equals @xmath66 .
therefore , the only @xmath0-dependence of ( [ 4.11 ] ) lies in the numerator in the function @xmath67 .
we now note that the eigenvalues of the volume operator come with a factor of @xmath68 , as displayed in ( [ 4.4 ] ) . pulling it under the square root into the modulus , and noticing that the modulus is a third - order polynomial in @xmath69
, we see that [ 4.12 ] _ , v(y(z+^-1 c ) ) = where we have [ 4.13 ] ^e_j & = & _ e,k ^e e_jk ^e_k + & = & _ e,k ( a^2 ^e e_jk z^e_k + ^e e_jk ( ) ^2 e^e_k ) where ( [ 4.2a ] ) has been used in the last line . to extract the leading order in @xmath53 of ( [ 4.11 ] )
is now easy .
first note that the matrix elements of both @xmath55 and @xmath70 are of order unity .
since @xmath51 is some macroscopic length scale , the first term that is proportional to @xmath59 in ( [ 4.13 ] ) is therefore of order @xmath71 , while the second is of zeroth - order in @xmath53 .
therefore @xmath72 is of the form [ 4.14 ] f(z)=4 where @xmath73 is a certain sixth - order polynomial in @xmath74 with no zeroth - order term , while @xmath1 is independent of @xmath59 .
moreover , @xmath75 is non - negative for all @xmath59 because it is the square of a third - order polynomial in @xmath59 . in particular
, this holds at @xmath76 , and therefore @xmath1 is also a non - negative number .
then , provided @xmath77 , we can define [ 4.15 ] f(z):==:4 where @xmath78 is a sixth - order polynomial with no zeroth - order term which is bounded from below by @xmath9 .
now , as in @xcite , we exploit the existence of @xmath79 such that [ 4.16 ] 1 + r - r r^2f1 + r for all @xmath80 . inserting this estimate into ( [ 4.11 ] )
we can bound the integral from above and below because the gaussian is positive .
the integrals over @xmath78 and @xmath81 are finite and are at least of order @xmath53 because odd powers of @xmath59 do not contribute to the gaussian integral .
this is as far as we can go with the calculation for a general graph .
notice that the inverse of the edge metric appears in this expression and , for a general graph , this is beyond analytical control .
therefore we will now make restrictions on the graph so as to analyse ( [ 4.18 ] ) further .
* _ coordinate chart _ + the graph , the region @xmath78 and the families of stacks lie in a common coordinate chart @xmath82 .
this is not a serious restriction because the general situation may be reduced to this one by appropriately restricting attention to the various charts of an atlas that covers @xmath83 . *
_ tame graphs _ + we assume that the graph is _ tame _ with respect to the stacks .
by this we mean that for each direction @xmath84 , and each stack @xmath85 , a given edge , @xmath32 , of the graph enters and leaves that stack at most once .
this means that the graph does not ` wiggle ' too much on the scale of the plaquettes .
analytically , it means that @xmath86 vanishes whenever @xmath87 for any @xmath32 , and that , for given @xmath32 , the number @xmath88 is non - vanishing at most for either @xmath89 or @xmath90 but , not both , and independently of @xmath85 .
finally , it means that the sets @xmath91 are connected . * _ coarse graphs _ + we assume that the graph is much coarser than the plaquettation in the sense that any edge intersects many different stacks in at least one direction @xmath84 . * _ non - aligned graphs _ + we exclude the possibility that distinct edges are ` too aligned ' with each other in the sense that the number of stacks that they commonly traverse is much smaller than the number of stacks that they individually traverse . a consequence of the tameness , coarseness and ` alignedness ' assumption is that @xmath92 for all @xmath56 , as is immediately obvious from the formulae displayed in our companion paper because the number of stacks with @xmath93 will be very much smaller than the number of stacks with @xmath94 or @xmath95 . hence the edge
metric will be almost diagonal .
this is important because we need its inverse , which can only be calculated with good approximation ( that is , for large , semiclassically relevant graphs ) if it is almost diagonal .
the graphs that we will eventually consider are embeddings of subgraphs dual to tetrahedronal , cubical or octahedronal triangulations of @xmath96 .
these correspond to embeddings of regular @xmath97-,@xmath98-,@xmath99-valent lattices , which ensures the non - alignedness property .
thus , without loss of generality , we may choose the stacks and plaquettes as follows . using the availability of the chart @xmath100 we consider the foliations
@xmath101 defined by the leaves @xmath102 where for @xmath103 we set @xmath104 .
the stacks are labelled by @xmath105 .
the corresponding plaquettes are given by @xmath106l);\;u\in [ 0,1)^2\}$ ] where @xmath107 is a positive number .
likewise , using the availability of the chart , we take the edges of the graph to be embeddings of straight lines in @xmath96 ( with respect to the euclidean background metric available there ) , that is , @xmath108 where @xmath109 is a vector in @xmath96 and @xmath110 defines the beginning point of the edge .
after these preparations , we can now analyse ( [ 4.17 ] ) and ( [ 4.18 ] ) further . recall that [ 4.19 ] e^e_j=_,i dt e_j(p^i_t ) ( p^i_t , e ) and [ 4.20 ] t^_ee=_i dt ( p^i_t , e ) ( p^i_t , e ) by the assumption about the graphs made above , the signed intersection number takes at most the numbers @xmath111 and independently of @xmath85 , so that @xmath112 for certain @xmath113 which takes
the value @xmath114 if the orientation of @xmath32 agrees with that of the leaves of the foliation , @xmath9 if it disagrees , and @xmath115 if it lies inside a leaf . if we assume that the electric field @xmath116 is slowly varying at the scale of the graph ( and hence at the scale of the plaquettes as well ) then we may write [ 4.21 ] e^e_j _ i t^_e e ^i_e
e_j(p^i_v ) where @xmath117 and @xmath41 is the vertex at which @xmath32 is adjacent and which is under consideration in @xmath118 .
it follows that ( [ 4.18 ] ) can be written as [ 4.22 ] y^e_j=()^2_e,k ^e e_jk _ i t^_e e ^i_e e_k(p^i_v ) = ( ) ^2_e ^e e_i ^i_e e_j(p^i_v ) where we have used @xmath119 .
now , by construction , @xmath120 with @xmath121 off - diagonal and with small entries [ 4.23 ] b_ee= which are of the order of @xmath36 since two distinct edges will typically only remain in the same stack for a parameter length @xmath35 , while the parameter length of an edge is @xmath34 . now notice that under the assumptions we have made , we have @xmath122 if @xmath56 are not adjacent .
define @xmath123 to be the subset of edges which are adjacent to @xmath32 . then [
4.24 ] ||b x||^2 & = & _ e_e , e^s_e x_e b_e e b_e e^ x_e^ + & & [ _ e , e b_e e^2 ] _
e ^2 + & & ( ) ^2 _ e ^2 + & & ( ) ^2 m _ e _
es_e x_e^2 + & = & ( ) ^2 m _
x_e^2 _ e _
s_e(e ) + & & ( ) ^2 m^2 ||x||^2 here , in the second step we estimated the matrix elements of @xmath121 from above ; in the third step we applied the schwarz inequality ; in the fourth step we estimated @xmath124 , where @xmath125 is the maximal valence of a vertex in @xmath27 ; and in the sixth step we exploited the symmetry [ 4.25 ] _
s_e(e ) = \ { it follows that for @xmath126 , @xmath121 is bounded from above by unity .
therefore , the geometric series @xmath127 converges in norm .
hence we are able to consider the effects of a non - diagonal edge metric up to arbitrary order , @xmath40 , in @xmath36 . here
we will consider @xmath128 only and write @xmath129 .
but before considering corrections from the off - diagonal nature of @xmath55 notice that , to zeroth - order in @xmath36 , equation ( [ 4.22 ] ) becomes simply [ 4.25a ] y^e_j = ( ) ^2 _ i ^i_e e_j(p^i_v ) inserting ( [ 4.25a ] ) into ( [ 4.17 ] ) we find [ 4.26 ] < > _ z , ( ) ^3 where [ 4.27 ] ( p_v)=_ijk^abc _ [ 0,1)^2 d^2u n_a
t_i(v)^_i(v ) i(u ) _ [ 0,1)^2 d^2u n_b t_j(v)^_j(v ) i(u ) _ [ 0,1)^2 d^2u n_c t_k(v)^_k(v ) i(u ) on recalling that @xmath130 with @xmath131 for @xmath103 , we find [ 4.28 ] ( p_v)l^6 _ x(s)=v hence ( [ 4.26 ] ) becomes [ 4.29 ] < > _ z , ( ) ^3 |_x(s)=v| we can draw an important conclusion from expression ( [ 4.29 ] ) .
namely , the first _ three _ factors approximate the classical volume @xmath132 as determined by @xmath133 of an embedded cube with parameter volume @xmath134 . when we sum ( [ 4.29 ] ) over the vertices of @xmath27 , which have a parameter distance , @xmath34 , from each other where @xmath135 by assumption , then the volume expectation value only has a chance to approximate the classical volume when the graph is such that @xmath136 , or @xmath137 .
this could never have been achieved for @xmath138 and explains why we had to rescale the labels of the coherent states by @xmath139 while keeping the classicality parameter at @xmath52 .
see our companion paper for a detailed discussion .
there we also explained why one must have @xmath140 actually equal to @xmath141 and not just of the same order : while one could use this in order to favour other valences of the volume operator , the expectation value of other geometrical operators such as area and flux would be incorrect .
assuming @xmath137 we write ( [ 4.29 ] ) as [ 4.30 ] <
> _ z,=:v_v(e ) g_,v thereby introducing the graph geometry factor @xmath142 .
it does not carry any information about the phase space , only about the embedding of the graph relative to the leaves of the three - foliations . from the fact that ( [ 4.30 ] ) reproduces the volume of a cube up to a factor
, we may already anticipate that the geometry factor will be close to unity for , at most , a cubic graph . whether this holds for an arbitrary orientation of the graph with respect to the stack family will occupy a large part of the analysis which follows .
we start by investigating the behaviour of the graph geometry factor @xmath142 under diffeomorphisms , @xmath143 , of @xmath83 , that is , under @xmath144 while the linearly - independent families of stacks are left untouched .
this will answer the question of how much the geometry factor depends on the relative orientation of the graph with respect to the stacks .
in fact , the orientation factor @xmath145 is invariant under diffeomorphisms of the spatial manifold @xmath83 .
the signature factor [ 4.31 ] _ e , ee^()=_ijk ^i_e ^j_e ^k_e^ is obviously invariant under any diffeomorphism that preserves the foliations @xmath101 , i.e. which map leaves onto leaves , because [ 4.32 ] ^i_e=_e dx^a _ abc _ l_it dy^bdy^c ( x , y ) where @xmath146 is any leaf in @xmath53 which intersects @xmath32 transversely .
since we consider graphs whose edges are embedded lines in @xmath96 with the same embedding that defines the stacks , it follows that the geometry factor is invariant under any embedded global translations in @xmath96 .
next , since global rescaling in @xmath96 preserves the foliations and the topological invariant ( [ 4.32 ] ) , the geometry factor is also invariant under embedded global rescalings of @xmath96 .
finally , any embedded global rotations of @xmath96 that preserves all the orientation factors @xmath147 will leave the geometry factors invariant .
since the orientation factors only take the values @xmath148 ( depending on whether an edge agrees , disagrees with the orientation of the leaves , or lies within a leaf ) , there will be a vast range of euler angles for which this condition is satisfied , if the graph is an embedded , regular lattice of constant valence .
hence , in order to check whether the geometry factor is rotationally invariant under any rotation we need only worry about those rotations which lead to changes in the @xmath147 .
likewise , if we rotate a graph which is dual to a polyhedronal complex , we expect that the expectation value remains invariant as long as the graph remains dual to the complex . fortunately , using the explicit formulae derived for the edges and vertices for @xmath37- , @xmath98- , @xmath99-valent graphs displayed in @xcite we can calculate the @xmath147 for each edge @xmath32 . intuitively , it is clear , that whenever many of the @xmath147 change from @xmath114 to @xmath9 , we can expect a drastic change of the expectation value . however , one has to take into account the combined effect of these changes , and this is what makes rotational invariance possible . as a first step
we determine the action of a rotation on the sign factors . in what follows we discuss the cases that show a drastic change in the value of @xmath149 caused by a change in the values of @xmath147 .
to carry out this calculation we will perform a rotation of each of the three different types of lattice analysed so far : namely , the @xmath97- , @xmath98- and @xmath99-valent lattices .
these rotations will be parameterised by euler angles and will be centred at a particular vertex of the lattice , for example @xmath150 .
the effects of a rotation will depend on the distance of the vertices from the centre of the rotation .
in fact , the position of each vertex in the lattice after rotation will depend on both the distance from the centre of the rotation and the euler angles used in the rotation .
fortunately , the values of the terms @xmath147 will not depend on the former but only on the latter .
this is easy to see since the value of @xmath147 can be either 1 , -1 or 0 depending on whether the edge is outgoing , ingoing , or lies on the plaquette in the direction @xmath84 .
thus it will only depend on the angle the edge makes with the perpendicular to the plaquette in any given direction , i.e. it will depend on the angles the edge makes with respect to a coordinate system centred at the vertex at which the edge is incident .
clearly , only the values of the euler angles of the rotation will affect the angles each edge has with respect to the vertex at which it is incident .
in particular , since the graph we are using is regular , following the rotation , all edges which were parallel to each other will remain such , and thus will have the same angles with respect to the vertex at which they are incident .
this implies that in order to compute the values of the terms @xmath151 , we can consider each vertex separately and apply the same rotation to each vertex individually . on the other hand
, the distance from the centre of the rotation affects the position of each vertex with respect to the plaquette structure , and thereby affects both the values of the terms @xmath152 and the number of them that are different from zero .
these effects can be easily understood with the aid of the two - dimensional diagram ( figure [ fig1 ] ) .
it is clear that , for any two parallel edges , the angle each of them has with respect to the vertex at which they are incident , is independent of the distance of the edge from the centre of rotation . on the other hand , the values of the @xmath152 will depend on both the rotation and the distance of the centre of rotation , since the position of the rotated vertex with respect to the plaquette depends on both these parameters .
therefore , we can tentatively assume that two different geometric factors will be involved in the computation of the volume operator : 1 .
@xmath153 , which indicates how the terms @xmath151 are affected by rotation .
this geometric factor affects all orders of approximation of the expectation value of the volume operator .
2 . @xmath154 , which indicates the effect of rotation on the terms @xmath152 .
this term affects only the first- and higher - order approximations of the expectation value of the volume operator , not the zeroth - order . in
what follows we will analyse the geometric term @xmath155 : i.e. , we will analyse the changes in the values of the @xmath156 due to a rotation applied at each vertex independently .
we will do this for the 4- , 6- and 8-valent graphs separately . the geometric factor @xmath154 will be analysed in subsequent sections .
as we will not see , our calculations show that for all 4- , 6- and 8-valent graphs , the rotations that produce drastic change in the values of @xmath157 have measure zero in @xmath38 since they occur for specific euler angles rather than for a range of them .
let us start with the 6-valent graph ( figure [ fig6 ] ) .
in order to compute the change in the values of the individual @xmath156 , we will divide into eight small sub - cubes the cube formed by the intersection of the plaquettes in the three directions and containing the vertex we are analysing ( @xmath150 ) .
it is then easy to see that for each edge @xmath32 , the corresponding value of @xmath158 depends on the sub - cube in which it lies . in particular
, we have the following table for the values of @xmath147 . where @xmath159 , @xmath160 , @xmath161 , @xmath162 , @xmath163 , @xmath164 , and the quantities @xmath165 , @xmath166 , @xmath167 are the @xmath14- , @xmath69- , @xmath59-coordinates of the vertex @xmath168 . in all the above formulae the term @xmath169 comes from the fact that @xmath170 for all edges @xmath171 . c. rovelli .
loop quantum gravity , _ living rev .
* 1 * ( 1998 ) , 1 .
[ gr - qc/9710008 ] + a. ashtekar and j. lewandowski .
background independent quantum gravity : a status report . _
* 21 * ( 2004 ) , r53 .
[ gr - qc/0404018 ] + t. thiemann .
lectures on loop quantum gravity .
notes phys . _
* 631 * ( 2003 ) , 41 - 135 .
[ gr - qc/0210094 ] t. thiemann .
anomaly - free formulation of non - perturbative , four - dimensional lorentzian quantum gravity .
_ physics letters _ * b380 * ( 1996 ) , 257 - 264 .
[ gr - qc/9606088 ] + t. thiemann .
quantum spin dynamics ( qsd ) .
quantum grav .
_ * 15 * ( 1998 ) , 839 - 73 .
[ gr - qc/9606089 ] + t. thiemann .
quantum spin dynamics ( qsd ) : ii .
the kernel of the wheeler - dewitt constraint operator . _ class .
quantum grav . _ * 15 * ( 1998 ) , 875 - 905 .
[ gr - qc/9606090 ] + t. thiemann .
quantum spin dynamics ( qsd ) : iii . quantum constraint algebra and physical scalar product in quantum general relativity . _ class .
quantum grav . _ * 15 * ( 1998 ) , 1207 - 1247 .
[ gr - qc/9705017 ] + t. thiemann .
quantum spin dynamics ( qsd ) : iv .
2 + 1 euclidean quantum gravity as a model to test 3 + 1 lorentzian quantum gravity .
quantum grav .
_ * 15 * ( 1998 ) , 1249 - 1280 .
[ gr - qc/9705018 ] + t. thiemann .
quantum spin dynamics ( qsd ) : v. quantum gravity as the natural regulator of the hamiltonian constraint of matter quantum field theories . _ class .
quantum grav . _
* 15 * ( 1998 ) , 1281 - 1314 .
[ gr - qc/9705019 ] + t. thiemann .
quantum spin dynamics ( qsd ) : vi . quantum poincar algebra and a quantum positivity of energy theorem for canonical quantum gravity . _ class .
quantum grav . _ * 15 * ( 1998 ) , 1463 - 1485
. [ gr - qc/9705020 ] + t. thiemann .
kinematical hilbert spaces for fermionic and higgs quantum field theories .
quantum grav . _ * 15 * ( 1998 ) , 1487 - 1512 .
[ gr - qc/9705021 ] t. thiemann .
the phoenix project : master constraint programme for loop quantum gravity . _ class .
* 23 * ( 2006 ) , 2211 - 2248 .
[ gr - qc/0305080 ] + t. thiemann .
quantum spin dynamics ( qsd ) : viii . the master constraint . _ class .
* 23 * ( 2006 ) , 2249 - 2266 .
[ gr - qc/0510011 ] k. giesel and t. thiemann .
consistency check on volume and triad operator quantisation in loop quantum gravity .
i. _ class .
* 23 * ( 2006 ) , 5667 - 5691 .
[ gr - qc/0507036 ] + k. giesel and t. thiemann .
consistency check on volume and triad operator quantisation in loop quantum gravity .
* 23 * ( 2006 ) , 5693 - 5771 .
[ gr - qc/0507037 ] t. thiemann and o. winkler .
gauge field theory coherent states ( gcs ) : ii .
peakedness properties .
* 18 * ( 2001 ) , 2561 - 2636 .
[ hep - th/0005237 ] + t. thiemann and o. winkler .
gauge field theory coherent states ( gcs ) : iii .
ehrenfest theorems . _ class .
* 18 * ( 2001 ) , 4629 - 4681 .
[ hep - th/0005234 ] + t. thiemann and o. winkler .
gauge field theory coherent states ( gcs ) : iv .
infinite tensor product and thermodynamic limit .
* 18 * ( 2001 ) , 4997 - 5033 .
[ hep - th/0005235 ] + h. sahlmann , t. thiemann and o. winkler .
coherent states for canonical quantum general relativity and the infinite tensor product extension .
phys . _ * b606 * ( 2001 ) , 401 - 440 .
[ gr - qc/0102038 ] a. ashtekar and c.j .
representations of the holonomy algebras of gravity and non - abelian gauge theories . _ class .
grav . _ * 9 * ( 1992 ) , 1433 .
[ hep - th/9202053 ] + a. ashtekar and j. lewandowski .
representation theory of analytic holonomy @xmath172 algebras . in
_ knots and quantum gravity _ , j. baez ( ed . ) , ( oxford university press , oxford 1994 ) . [ gr - qc/9311010 ] j. lewandowski , a. okolow , h. sahlmann and t. thiemann .
uniqueness of diffeomorphism invariant states on holonomy flux algebras .
* 267 * ( 2006 ) , 703 - 733 .
[ gr - qc/0504147 ] + c. fleischhack .
representations of the weyl algebra in quantum geometry .
[ math - ph/0407006 ] j. brunnemann and d. rideout .
spectral analysis of the volume operator in loop quantum gravity .
[ gr - qc/0612147 ] + j. brunnemann and d. rideout .
properties of the volume operator in loop quantum gravity . i. results . _ class . quant .
* 25 * ( 2008 ) , 065001 .
[ arxiv:0706.0469 [ gr - qc ] ] + j. brunnemann and d. rideout .
properties of the volume operator in loop quantum gravity ii : detailed presentation .
[ arxiv:0706.0382 ] h. sahlmann and t. thiemann . towards the qft on curved spacetime limit of qgr .
1 . a general scheme .
* 23 * ( 2006 ) , 867 - 908 .
[ gr - qc/0207030 ] + h. sahlmann and t. thiemann . towards the qft on curved spacetime limit of qgr .
2 . a concrete implementation . _ class . quant .
* 23 * ( 2006 ) , 909 - 954 .
[ gr - qc/0207031 ] k. giesel and t. thiemann .
algebraic quantum gravity ( aqg ) .
i. conceptual setup . _ class .
* 24 * ( 2007 ) , 2465 - 2498 .
[ gr - qc/0607099 ] + k. giesel and t. thiemann .
algebraic quantum gravity ( aqg ) .
ii . semiclassical analysis . _ class .
* 24 * ( 2007 ) , 2499 - 2564 .
[ gr - qc/0607100 ] m. varadarajan .
fock representations from u(1 ) holonomy algebras .
_ * d61 * ( 2000 ) , 104001 .
[ gr - qc/0001050 ] + m. varadarajan .
photons from quantised electric flux representations .
_ * d64 * ( 2001 ) , 104003 .
[ gr - qc/0104051 ] + m. varadarajan .
gravitons from a loop representation of linearised gravity .
* d66 * ( 2002 ) , 024017 .
[ gr - qc/0204067 ] + m. varadarajan .
the graviton vacuum as a distributional state in kinematic loop quantum gravity . _ class .
* 22 * ( 2005 ) , 1207 - 1238 .
[ gr - qc/0410120 ] e. livine and s. speziale .
a new spinfoam vertex for quantum gravity .
* d76 * ( 2007 ) , 084028 . [ arxiv:0705.0674 [ gr - qc ] ] + j. engle , r. pereira and c. rovelli . the loop - quantum - gravity vertex - amplitude . _ phys .
* 99 * ( 2007 ) , 161301 .
[ arxiv:0705.2388 [ gr - qc ] ] + e. livine and s. speziale .
consistently solving the simplicity constraints for spinfoam quantum gravity .
[ arxiv:0708.1915 [ gr - qc ] ] + j. engle , r. pereira and c. rovelli .
flipped spinfoam vertex and loop gravity .
* b798 * ( 2008 ) , 251 - 290 .
[ arxiv:0708.1236 [ gr - qc ] ] + l. freidel and k. krasnov . a new spin foam model for 4d gravity .
[ e - print : arxiv:0708.1595 [ gr - qc ] ] + j. engle , e. livine , r. pereira and c. rovelli .
lqg vertex with finite immirzi parameter .
phys . _ * b799 * ( 2008 ) , 136 - 149 . [ arxiv:0711.0146 [ gr - qc ] ] j. w. barrett and l. crane .
relativistic spin networks and quantum gravity . _ j. math .
* 39 * ( 1998 ) , 3296 - 3302 .
[ gr - qc/9709028 ] + j. w. barrett and l. crane . a lorentzian signature model for quantum general relativity .
* 17 * ( 2000 ) , 3101 - 3118 .
[ gr - qc/9904025 ] h. sahlmann and t. thiemann .
irreducibility of the ashtekar
lewandowski representation .
* 23 * ( 2006 ) , 4453 - 4472 .
[ gr - qc/0303074 ] + c. fleischhack .
irreducibility of the weyl algebra in loop quantum gravity .
* 97 * ( 2006 ) , 061302 .
e. alesci and c. rovelli .
the complete lqg propagator .
i. difficulties with the barrett - crane vertex .
* d76 * ( 2007 ) , 104012 . [ arxiv:0708.0883 [ gr - qc ] ] + e. alesci and c. rovelli . the complete lqg propagator .
asymptotic behavior of the vertex .
_ phys.rev .
_ * d77 * ( 2008 ) , 044024 .
[ arxiv:0711.1284 [ gr - qc ] ] + v. bonzom , e. livine , m. smerlak and s. speziale + towards the graviton from spinfoams : the complete perturbative expansion of the 3d toy model .
[ arxiv:0802.3983 [ gr - qc ] ] a. ashtekar , j. lewandowski , d. marolf , j. mouro and t. thiemann .
quantisation of diffeomorphism invariant theories of connections with local degrees of freedom .
* 36 * ( 1995 ) , 6456 - 6493 .
[ gr - qc/9504018 ] b. bahr and t. thiemann .
gauge - invariant coherent states for loop quantum gravity . i. abelian gauge groups .
[ arxiv:0709.4619 [ gr - qc ] ] + b. bahr and t. thiemann .
gauge - invariant coherent states for loop quantum gravity .
non - abelian gauge groups .
[ arxiv:0709.4636 [ gr - qc ] ] j. velhinho .
a groupoid approach to spaces of generalised connections .
_ j. geom .
* 41 * ( 2002 ) , 166 - 180 .
[ hep - th/0011200 ] + j. velhinho . on the structure of the space of generalised connections .
phys . _ * 1 * ( 2004 ) , 311 - 334 .
[ math - ph/0402060 ] + b. bahr and t. thiemann , automorphisms in loop quantum gravity .
[ arxiv:0711.0373 [ gr - qc ] ] | we continue the semiclassical analysis of the loop quantum gravity ( lqg ) volume operator that was started in the companion paper @xcite . in the first paper we prepared the technical tools , in particular the use of complexifier coherent states that use squares of flux operators as the complexifier . in this paper ,
the complexifier is chosen for the first time to involve squares of area operators .
both cases use coherent states that depend on a graph . however , the basic difference between the two choices of complexifier is that in the first case the set of surfaces involved is discrete , while , in the second it is continuous . this raises the important question of whether the second set of states has improved invariance properties with respect to relative orientation of the chosen graph in the set of surfaces on which the complexifier depends . in this paper
, we examine this question in detail , including a semiclassical analysis .
the main result is that we obtain the correct semiclassical properties of the volume operator for i ) artificial rescaling of the coherent state label ; and ii ) particular orientations of the 4- and 6-valent graphs that have measure zero in the group so(3 ) . since
such requirements are not present when analysing dual cell complex states , we conclude that coherent states whose complexifiers are squares of area operators are not an appropriate tool with which to analyse the semiclassical properties of the volume operator . moreover ,
if one intends to go further and sample over graphs in order to obtain embedding independence , then the area complexifier coherent states should be ruled out altogether as semiclassical states .
_ i would like to dedicate this paper to my parents elena romani and luciano flori _ |
by studying isotopes with enhanced sensitivity to fundamental symmetries , nuclear physicists can test various aspects of the standard model in ways that are complementary to other sciences .
for example , a possible explanation for the observed asymmetry between matter and anti - matter in the universe could be studied by searching for a permanent electric dipole moment larger than standard model predictions in heavy radioactive nuclei that have permanent octupole shapes .
likewise , the superallowed @xmath0-decays of a handful of rare isotopes with similar numbers of protons and neutrons , in which both the parent and daughter nuclear states have zero angular momentum and positive parity , are the unique laboratory to study the strength of the weak force .
what makes these pure vector - current - mediated ( fermi ) decays so useful for testing the standard model is the hypothesis of the conserved vector current ( cvc ) , that is , independence of the vector current on the nuclear medium .
the consequence of the cvc hypothesis is that the product of the statistical rate function @xmath4 and partial half - life @xmath5 for the superallowed @xmath6 fermi @xmath0-decay should be nucleus independent and equal to : @xmath7 where @xmath8gev@xmath9s is a universal constant ; @xmath10 stands for the vector coupling constant for semi - leptonic weak interaction , and @xmath11 is the nuclear matrix element of the isospin rising or lowering operator @xmath12 .
the relation ( [ ft ] ) does not hold exactly and must be slightly amended by introducing a set of radiative corrections to the @xmath13-values , and a correction to the nuclear matrix element due to isospin - symmetry breaking : @xmath14 with the left - hand side being nucleus independent . in eq .
( [ ftnew ] ) , @xmath15% stands for the nucleus - independent part of the radiative correction @xcite , @xmath16 is a transition - dependent ( @xmath17-dependent ) but nuclear - structure - independent part of the radiative correction @xcite , and @xmath18 denotes the nuclear - structure - dependent part of the radiative correction @xcite . in spite of theoretical uncertainties in the evaluation of the radiative and isospin - symmetry - breaking corrections ,
the superallowed @xmath0-decay is the most precise source of experimental information for determining the vector coupling constant @xmath10 , and provides us with a stringent test of the cvc hypothesis . in turn , it is also the most precise source of the matrix element @xmath19 of the cabibbo - kobayashi - maskawa ( ckm ) three - generation quark mixing matrix @xcite .
this is so because the leptonic coupling constant , @xmath20 gev@xmath21 , is well known from the muon decay @xcite .
the advantage of the superallowed @xmath0-decay strategy results from the fact that , within the cvc hypothesis , @xmath22 can be extracted by averaging over several transitions in different nuclei . for precise tests of the standard model , only these transitions that have @xmath13-values known with a relative precision better than a fraction of a percent are acceptable . currently , 13 canonical " transitions spreading over a wide range of nuclei from @xmath23 to @xmath24 meet this criterion ( have @xmath13-values measured with accuracy of order of 0.3% or better ) and are used to evaluate the values of @xmath10 and @xmath22 @xcite . in this work
we concentrate on the isospin - breaking ( isb ) corrections @xmath25 , which were already computed by various authors , using a diverse set of nuclear models @xcite .
the standard in this field has been set by towner and hardy ( ht ) @xcite who used the nuclear shell - model to account for the configuration mixing effect , and the mean - field ( mf ) approach to account for a radial mismatch of proton and neutron single - particle ( s.p . )
wave functions caused by the coulomb polarization . in this study , which constitutes an extension of our earlier work @xcite , we use the isospin- and angular - momentum - projected density functional theory ( dft ) .
this method can account , in a rigorous quantum - mechanical way , for spontaneous symmetry - breaking ( ssb ) effects , configuration mixing , and long - range coulomb polarization effects .
our paper is organized as follows .
the model is described in sec .
[ model ] .
the results of calculations for isb corrections to the superallowed @xmath26 fermi transitions are summarized in sec .
[ sec03 ] .
the isb corrections to the fermi matrix elements in mirror - symmetric @xmath27 nuclei are discussed in sec .
[ sec05 ] .
section [ sec06 ] studies a particular case of the fermi decay of @xmath28cl .
finally , the summary and perspectives are given in sec .
[ sec07 ] .
the success of the self - consistent dft approach to mesoscopic systems @xcite in general , and specifically to atomic nuclei @xcite , has its roots in the ssb mechanism that incorporates essential short - range ( pairing ) and long - range ( spatial ) correlations within a single deformed slater determinant .
the deformed states provide a basis for the symmetry - projected dft approaches , which aim at including beyond - mean - field correlations through the restoration of broken symmetries by means of projection techniques @xcite .
the building block of the isospin- and angular - momentum - projected dft approach employed in this study is the self - consistent deformed mf state @xmath29 that violates both the rotational and isospin symmetries . while the rotational invariance is of fundamental nature and is broken spontaneously , the isospin symmetry is violated both spontaneously and explicitly by the coulomb interaction between protons .
the strategy is to restore the rotational invariance , remove the spurious isospin mixing caused by the isospin ssb effect , and retain only the physical isospin mixing due to the electrostatic interaction @xcite .
this is achieved by a rediagonalization of the entire hamiltonian , consisting the isospin - invariant kinetic energy and skyrme force and the isospin - non - invariant coulomb force , in a basis that conserves both angular momentum and isospin . to this end
, we first find the self - consistent mf state @xmath29 and then build a normalized angular - momentum- and isospin - conserving basis @xmath30 by using the projection method : @xmath31 where @xmath32 and @xmath33 stand for the standard isospin and angular - momentum projection operators : @xmath34 where , @xmath35 is the rotation operator about the @xmath36-axis in the isospace , @xmath37 is the wigner function , and @xmath38 is the third component of the total isospin @xmath39 . as usual
, @xmath40 is the three - dimensional rotation operator in space , @xmath41 are the euler angles , @xmath42 is the wigner function , and @xmath43 and @xmath44 denote the angular - momentum components along the laboratory and intrinsic @xmath45-axis , respectively @xcite .
note that unpaired mf states @xmath29 conserve the third isospin component @xmath46 ; hence , the one - dimensional isospin projection suffices .
the set of states ( [ itbasis ] ) is , in general , overcomplete because the @xmath44 quantum number is not conserved .
this difficulty is overcome by selecting first the subset of linearly independent states known as _ collective space _
@xcite , which is spanned , for each @xmath47 and @xmath48 , by the so - called _ natural states _
@xmath49 @xcite .
the entire hamiltonian including the isb terms is rediagonalized in the collective space , and the resulting eigenfunctions are : @xmath50 where the index @xmath51 labels the eigenstates in ascending order according to their energies .
the amplitudes @xmath52 define the degree of isospin mixing through the so - called isospin - mixing coefficients ( or isospin impurities ) , determined for a given @xmath51th eigenstate as : @xmath53 where the sum of norms corresponds to the isospin @xmath48 dominating in the wave function @xmath54 .
one of the advantages of the projected dft as compared to the shell - model - based approaches @xcite is that it allows for a rigorous quantum - mechanical evaluation of the fermi matrix element using the bare isospin operators : @xmath55 where @xmath56 denotes the rank - one covariant one - body spherical - tensor operators in the isospace , see the discussion in ref .
indeed , noting that each @xmath57th eigenstate ( [ ktmix ] ) can be uniquely decomposed in terms of the original basis states ( [ itbasis ] ) , @xmath58 with microscopically determined mixing coefficients @xmath59 , the expression for the fermi matrix element between the parent state @xmath60 and daughter state @xmath61 can be written as : @xmath62 where tilde indicates the slater determinant rotated in space : @xmath63 . the matrix element appearing on the right - hand side of eq .
( [ mate ] ) can be expressed through the transition densities that are basic building blocks of the multi - reference dft @xcite . indeed , with the aid of the identity @xmath64 which results from the general transformation rule for spherical tensors under rotations or isorotations , @xmath65 the matrix element entering eq .
( [ mate ] ) can be expressed as : @xmath66 for unpaired slater determinants considered here , the double integral over the isospace euler angles in eq .
( [ mate ] ) can be further reduced to a one - dimensional integral over the angle @xmath67 using the identity @xmath68 which is the one - dimensional version of the transformation rule ( [ trans ] ) valid for rotations around the @xmath69 axis in the isospace . the final expression for the matrix element in eq .
( [ mate1 ] ) reads : @xmath70 where @xmath71 is the isovector transition density , and the double - tilde sign indicates that the right slater determinant used to calculate this density is rotated both in space as well as in isospace : @xmath72 . the symbol @xmath73 denotes the overlap kernel
. since the natural states have good isospin , the states ( [ ktmix ] ) are free from spurious isospin mixing .
moreover , since the isospin projection is applied to self - consistent mf solutions , our model accounts for a subtle balance between the long - range coulomb polarization , which tends to make proton and neutron wave functions different , and the short - range nuclear attraction , which acts in an opposite way .
the long - range polarization affects globally all s.p .
wave functions .
direct inclusion of this effect in open - shell heavy nuclei is possible essentially only within the dft , which is the only _ no - core _ microscopic framework that can be used there .
recent experimental data on the isospin impurity deduced in @xmath74zr from the giant dipole resonance @xmath75-decay studies @xcite agree well with the impurities calculated using isospin - projected dft based on modern skyrme - force parametrizations @xcite .
this further demonstrates that the isospin - projected dft is capable of capturing the essential piece of physics associated with the isospin mixing . as discussed in ref .
@xcite , the isospin projection technique outlined above does not yield singularities in energy kernels ; hence , it can be safely executed with all commonly used energy density functionals ( edfs ) .
however , as demonstrated in ref .
@xcite , the isospin projection alone leads to unphysically large isospin mixing in odd - odd @xmath76 nuclei .
it has thus been concluded that in order to obtain reasonable results
isospin projection must be augmented by angular - momentum projection .
this not only increases the numerical effort , but also brings back the singularities in the energy kernels @xcite and thus prevents one from using the modern parametrizations of the skyrme edfs , which all contain density - dependent terms @xcite .
therefore , the only option @xcite is to use the hamiltonian - driven edfs . for the skyrme - type functionals ,
this leaves us with one choice : the sv parametrization @xcite . in order to better control the time - odd fields ,
the standard sv parametrization must be augmented by the tensor terms , which were neglected in the original work @xcite .
this density - independent parameterization of the skyrme functional has the isoscalar effective mass as low as @xmath77 , which is required to reproduce the actual nuclear saturation properties .
the unusual saturation mechanism of sv has a dramatic impact on the overall spectroscopic quality of this force , impairing such key properties like the symmetry energy @xcite , level density , and level ordering .
these deficiencies also affect the calculated isospin mixing , which is a prerequisite for realistic estimates of @xmath25 .
in particular , in the case of @xmath74zr discussed above , sv yields @xmath78% , which is considerably smaller than the mean value of @xmath79% obtained by averaging over nine popular skyrme edfs including the msk1 , sko , skp , sly4 , sly5 , sly7 , skm@xmath80 , skxc , and siii functionals , see ref .
@xcite for further details .
even though the isb corrections @xmath25 are primarily sensitive to _ differences _ between isospin mixing in isobaric analogue states , the lack of a reasonable hamiltonian - based skyrme edf is probably the most critical deficiency of the current formalism .
the aim of this study is ( i ) to provide the most reliable set of the isb corrections that can be obtained within the current angular - momentum and isospin - projected single - reference dft , and ( ii ) explore the sensitivity of results to edf parameters , choice of particle - hole configurations , and structure of time - odd fields that correlate valence neutron - proton pairs in odd - odd @xmath76 nuclei . in particular , to quantify uncertainties related to the skyrme coupling constants , we have developed a new density - independent variant of the skyrme force dubbed hereafter shz2 , see table [ tab1 ] .
the force was optimized purposefully to properties of light magic nuclei below @xmath81sn .
the coupling constants @xmath82 , @xmath83,@xmath84,@xmath85 of shz2 were found by means of a @xmath86 minimization to experimental @xcite binding energies of five doubly - magic nuclei : @xmath87o , @xmath88ca , @xmath89ca , @xmath90ni , and @xmath81sn .
the procedure reduced the @xmath86 from @xmath91 for sv set to @xmath92 for shz2 .
most of the nuclear matter characteristics calculated for both sets are similar .
it appears , however , that the fit to light nuclei only weakly constrains the symmetry energy .
the bulk symmetry energy of shz2 is @xmath93mev , i.e. , it overestimates the accepted value @xmath94mev by almost 30% . while this property essentially precludes using shz2 in detailed nuclear structure studies , it also creates an interesting opportunity for investigating the quenching of isb effects due to the large isospin - symmetry - restoring components of the force .
crrr ' '' '' param . &
sv & shz2 & change ( % ) + ' '' '' @xmath82 & @xmath951248.290 & @xmath951244.98830 & @xmath950.26 + @xmath83 & 970.560 & 970.01156 & @xmath950.06 + @xmath84 & 107.220 & 99.50197 & @xmath957.20 + @xmath85 & @xmath950.170 & 0.01906 & @xmath95111.21 + @xmath96 & 150 & 150 & @xmath97 all calculations presented below were done by using the code hfodd @xcite , version ( 2.48q ) or higher , which includes both the angular - momentum and isospin projections . in order to obtain converged results for isospin mixing with respect to basis truncation , in our sv calculations we used @xmath98 harmonic oscillator ( ho ) shells for @xmath99 nuclei , 12 shells for @xmath100 nuclei , and 14 shells for @xmath101 nuclei . in shz2 test calculations
, we took @xmath98 shells for @xmath102 and @xmath103 shells for @xmath104 nuclei .
for the numerical integration over the euler angles in space and isospace ( @xmath105 ) we used the gauss - tchebyschev ( over @xmath106 and @xmath75 ) and gauss - legendre ( over @xmath0 and @xmath67 ) quadratures .
we took @xmath107 and @xmath108 ( or 10 ) integration points .
this choice guarantees that the calculated values of @xmath25 are not affected by the numerical integration error .
the @xmath110 fermi @xmath0-decay proceeds between the ground state ( g.s . ) of the even - even nucleus @xmath111 and its isospin - analogue partner in the @xmath76 odd - odd nucleus , @xmath112 .
the corresponding transition matrix element is : @xmath113 the g.s .
state @xmath111 in eq .
( [ fermime ] ) is approximated by a projected state @xmath114 where @xmath115 is the g.s . of the even - even nucleus
obtained in self - consistent mf calculations . the state @xmath115 is unambiguously defined by filling in the pairwise doubly degenerate levels of protons and neutrons up to the fermi level .
the daughter state @xmath112 is approximated by @xmath116 where the self - consistent slater determinant @xmath117 ( or @xmath118 ) represents the so - called anti - aligned configuration , selected by placing the odd neutron and the odd proton in the lowest available time - reversed ( or signature - reversed ) s.p.orbits .
the s.p .
configuration @xmath119 manifestly breaks the isospin symmetry as schematically depicted in fig .
[ fig01 ] . the isospin projection from @xmath29 as expressed by eq .
( [ oddphi ] ) is essentially the only way to reach the @xmath120 states in odd - odd @xmath76 nuclei .
= @xmath17 nucleus , as described by the conventional deformed mf theory .
these degenerate configurations are called aligned ( upper ) and anti - aligned ( lower ) , depending on what levels are occupied by the valence particles .
the right panel shows what happens when the isospin - symmetry is restored .
the aligned configuration is isoscalar ; hence , it is insensitive to the isospin projection .
the anti - aligned configuration represents a mixture of @xmath48=0 and @xmath48=1 states .
the isospin projection removes the degeneracy by lowering the @xmath48=0 level . ]
( @xmath121 ) , @xmath122 ( @xmath123 ) , and @xmath124 ( @xmath17 ) discussed in the text .
the long ( @xmath69 ) , intermediate ( @xmath125 ) , and short ( @xmath126 ) principal axes of the nuclear mass distribution are indicated by thick arrows . the odd - neutron ( @xmath127 ) and odd - proton ( @xmath128 ) angular momentum oriented along the @xmath125 , @xmath126 , or @xmath69 axes
is shown by thin arrows .
note that in each case the total angular - momentum alignment , @xmath129 , is zero . ] at variance with the even - even parent nuclei , the anti - aligned configurations in odd - odd daughter nuclei are not uniquely defined .
one of the reasons , which was not fully appreciated in our previous work @xcite , is related to the relative orientation of the nuclear shapes and currents associated with the valence neutron - proton pairs . in all signature - symmetry - restricted calculations for triaxial nuclei , such as ours , there are three anti - aligned slater determinants with the s.p .
angular momenta ( alignments ) of the valence protons and neutrons pointing , respectively , along the @xmath125 , @xmath126 , or @xmath69 axes of the intrinsic shape defined by means of the long ( @xmath69 ) , intermediate ( @xmath125 ) , and short ( @xmath126 ) principal axes of the nuclear mass distribution .
these solutions , hereafter referred to as @xmath130 , @xmath122 , and @xmath124 , are schematically illustrated in fig .
[ fig02 ] .
their properties can be summarized as follows : * the three solutions are not linearly independent .
their hartree - fock ( hf ) binding energies may typically differ by a few hundred kev .
the differences come almost entirely from the isovector correlations in the time - odd channel , as shown in the lower panel of fig .
[ fig03 ] for a representative example of @xmath131cl .
let us stress that these poorly - known correlations may significantly impact the isb corrections , as shown in the upper panel of fig .
[ fig03 ] . *
the type of the isovector time - odd correlations captured by the hf solutions depends on the relative orientation of the nucleonic currents with respect to the nuclear shapes .
solutions oriented perpendicular to the long axis , @xmath130 and @xmath122 , are usually similar to one another ( they yield identical correlations for axial systems ) and differ from @xmath124 , oriented parallel to the long axis , which captures more correlations due to the current - current time - odd interactions . *
the three @xmath132 states projected from the @xmath133 , @xmath122 , and @xmath124 slater determinants differ in energy by only a few tens of kev , see the lower panel of fig .
[ fig03 ] .
hence , energy - wise , they represent the same physical solution , differing only slightly due to the polarization effects originating from different components of the time - odd isovector fields .
however , since these correlations are completely absent in the even - even parent nuclei , they strongly impact the calculated @xmath25 .
the largest differences in @xmath25 have been obtained for @xmath134 and @xmath24 systems , see fig .
[ fig03 ] and tables [ tab2 ] and [ tab3 ] .
* symmetry - unrestricted calculations always converge to the signature - symmetry - conserving solution @xmath135 which , rather surprisingly , appears to be energetically unfavored ( except for @xmath136f ) . in spite of our persistent efforts ,
no self - consistent tilted - axis solutions have been found . for the @xmath137 superallowed @xmath0-decays @xmath131ar@xmath138cl ( open circles ) and @xmath131cl@xmath138s ( full circles ) determined for the shape - current orientations @xmath121 , @xmath123 , and
@xmath17 depicted schematically in fig .
[ fig02 ] .
lower panel : differences between the energies of the @xmath121 and @xmath123 configurations , and the @xmath17 configuration in @xmath131cl .
full triangles correspond to the total hf energies and open triangles correspond to contributions from the time - odd isovector channel .
full dots show the total energy differences obtained for the angular - momentum and isospin - projected states . ] owing to an increased density of s.p.nilsson levels in the vicinity of the fermi surface for nearly spherical nuclei , there appears another type of ambiguity in choosing the slater determinants representing the anti - aligned configurations . within the set of nuclei
studied in this work , this ambiguity manifests itself particularly strongly in @xmath140sc , where we deal with four possible anti - aligned mf configurations built on the nilsson orbits originating from the spherical @xmath141 and @xmath142 sub - shells .
these configurations can be labeled in terms of the quantum number @xmath44 as @xmath143 with @xmath144 , 3/2 , 5/2 , and 7/2 . in the extreme shell - model picture ,
each of these states contains all the @xmath145 and @xmath47=0 , 2 , 4 , and 6 components . within the projected dft picture ,
owing to configuration - dependent polarizations in time - odd and time - even channels , the situation is more complicated because the slater determinants @xmath146 corresponding to different @xmath44-values are no longer degenerate .
consequently , for each angular momentum @xmath47 , one obtains four different linearly - dependent solutions .
calculations show that in all @xmath47=0 and @xmath147 states of interest , the isospin mixing @xmath148 is essentially independent of the choice of the initial slater determinant .
in contrast , the calculated isb corrections @xmath25 and energies depend on @xmath44 , see fig .
[ fig04 ] .
top : the isb corrections to the @xmath140sc@xmath149@xmath140ca superallowed @xmath0-transition , calculated using sv ( circles ) and shz2 ( dots ) forces by projecting the @xmath146 configurations in @xmath140sc for @xmath144 , 3/2 , 5/2 , and 7/2 . from top to bottom ,
horizontal lines mark ( i ) the average isb correction using sv ( thick solid line ) ; ( ii ) the value of ref .
@xcite ( dotted line ) ; and ( iii ) @xmath25 of ref .
shaded regions mark the related uncertainties .
bottom : projected energies of states @xmath150 in @xmath140sc obtained from the configurations @xmath146 , relative to the projected energy of the @xmath144 state . ] based on the discussion presented in secs . [ sec03a ] and [ sec03b ] , the recommended calculated values of @xmath25 for the superallowed @xmath137 @xmath0-decay are determined by averaging over three relative orientations of shapes and currents . only in the case of @xmath151
, we adopt for @xmath25 an arithmetic mean over the four configurations associated with different @xmath44-orbitals .
to minimize uncertainties in @xmath148 and @xmath25 associated with the truncation of ho basis in hfodd , we used different ho spaces in different mass regions , cf .
[ basis ] . with this choice , the resulting systematic errors due the basis cut - off should not exceed @xmath152% . to illustrate the dependence of @xmath25 on the number of ho shells , fig .
[ fig05 ] shows the case of the superallowed @xmath153v@xmath149@xmath153ti transition obtained by projecting from the @xmath124 solution in @xmath153v . in this case ,
the parent and daughter nuclei are axial , which allows us to reduce the angular - momentum projection to one - dimension and extend the basis size up to @xmath154 ho shells . with increasing @xmath155 , @xmath25 increases , and asymptotically it reaches the value of @xmath156% .
this limiting value is about 6.7% larger than the value of @xmath157% obtained for @xmath103 shells , that is , for a basis used to compute the @xmath158 cases . for @xmath159 nuclei , which were all found to be triaxial , we have used @xmath160 shells .
the further increase of basis size is practically impossible .
nonetheless , as seen in fig .
[ fig05 ] , a rate of increase of @xmath25 slows down exponentially with @xmath155 , which supports our 10% error estimate due to the basis truncation .
( color online ) the convergence of the isb correction @xmath25 to the @xmath153v@xmath149@xmath153ti superallowed @xmath0-decay versus the number of oscillator shells considered . ]
the total error of the calculated value of @xmath25 includes the standard deviation from the averaging , @xmath161 , and the assumed 10% uncertainty due to the basis size : @xmath162 .
the same prescription for @xmath163 was also used in the test calculations with shz2 , even though a slightly smaller ho basis was employed in that case . for @xmath164 nuclei ,
our model predicts the unusually large correction @xmath165% .
the origin of a very different isospin mixing obtained for odd - odd and even - even members of this isobaric triplet is not fully understood .
most likely , it is a consequence of the poor spectroscopic properties of sv .
indeed , as a result of an incorrect balance between the spin - orbit and tensor terms in sv , the @xmath166 subshell is shifted up in energy close to the fermi surface .
this state is more sensitive to time - odd polarizations than other s.p .
states around @xmath88ca core , see table i in ref .
the calculated equilibrium deformations @xmath167 of the @xmath168 and @xmath169 @xmath164 isobaric triplet are very similar , around ( 0.090 , 60@xmath170 ) . in the following , the @xmath171k@xmath149@xmath171ar transition is excluded from the calculation of the @xmath22 matrix element .
lccrrrcrrcrrcrr parent & @xmath13 & @xmath173 & @xmath174 & @xmath175 & @xmath176 & @xmath173 & @xmath177 & @xmath178 & @xmath173 & @xmath179 & @xmath180 & @xmath173 & @xmath181 & @xmath178 + nucleus & ( s ) & & ( % ) & ( % ) & ( % ) & & ( % ) & ( s ) & & ( % ) & & & ( % ) & ( s ) + @xmath182 & & & & & & & & & & & & & & + @xmath183c & 3041.7(43 ) & & 0.559 & 0.559 & 0.823 & & 0.65(14 ) & 3062.1(62 ) & & 0.37(15 ) & 3.7 & & 0.462(65 ) & 3067.8(49 ) + @xmath184o & 3042.3(11 ) & & 0.303 & 0.303 & 0.303 & & 0.303(30 ) & 3072.3(21 ) & & 0.36(06 ) & 0.8 & & 0.480(48 ) & 3066.9(24 ) + @xmath185 mg & 3052.0(70 ) & & 0.243 & 0.243 & 0.417 & & 0.301(87 ) & 3080.5(75 ) & & 0.62(23 ) & 1.9 & & 0.342(49 ) &
3079.2(72 ) + @xmath131ar & 3052.7(82 ) & & 0.865 & 0.997 & 1.475 & & 1.11(29 ) & 3056(12 ) & & 0.63(27 ) & 3.1 & & 1.08(42 ) & 3057(15 ) + @xmath186 & & & & & & & & & & & & & & + @xmath187al & 3036.9(09 ) & & 0.308 & 0.308 & 0.494 & & 0.370(95 ) & 3070.5(31 ) & & 0.37(04 ) & 0.0 & & 0.307(62 ) & 3072.5(23 ) + @xmath131cl & 3049.4(11 ) & & 0.809 & 0.679 & 1.504 & & 1.00(38 ) & 3060(12 ) & & 0.65(05 ) & 48.4 & & 0.83(50 ) & 3065(15 ) + @xmath140sc & 3047.6(12 ) & & & & & & 0.77(27 ) & 3069.2(85 ) & & 0.72(06 ) & 0.5 & & 0.70(32 ) & 3071(10 ) + @xmath153v & 3049.5(08 ) & & 0.486 & 0.486 & 0.759 & & 0.58(14 ) & 3074.6(47 ) & & 0.71(06 ) & 4.5 & & 0.375(96 ) & 3080.9(35 ) + @xmath188mn & 3048.4(07 ) & & 0.460 & 0.460 & 0.740 & & 0.55(14 ) & 3074.1(47 ) & & 0.67(07 ) & 3.1 & & 0.39(13 ) & 3079.2(45 ) + @xmath189co & 3050.8(10 ) & & 0.622 & 0.622 & 0.671 & & 0.638(68 ) & 3074.0(32 ) & & 0.75(08 ) & 2.0 & & 0.51(20 ) & 3078.0(66 ) + @xmath190ga & 3074.1(11 ) & & 0.925 & 0.840 & 0.881 & & 0.882(95 ) & 3090.0(42 ) & & 1.51(09 ) & 44.0 & & 0.49(11 ) & 3102.3(45 ) + @xmath191rb & 3084.9(77 ) & & 2.054 & 1.995 & 1.273 & & 1.77(40 ) & 3073(15 ) & & 1.86(27 ) & 0.1 & & 0.90(22 ) & 3101(11 ) + & & & & & & & @xmath192 & 3073.6(12 ) & & @xmath193 & 112.2 & & @xmath192 & 3075.0(12 ) + & & & & & & & @xmath194 & 0.97397(27 ) & & @xmath195 & 10.2 & & @xmath194 & 0.97374(27 ) + & & & & & & & & 0.99935(67 ) & & & & & & 0.99890(67 ) .[tab3]similar as in table [ tab2 ] , except for the unmeasured transitions .
[ cols="<,^,^,^,>,^,>,^ " , ] the obtained values of the isb corrections to the fermi transitions , @xmath196 are collected in table [ tab5 ] and illustrated in fig .
[ fig13 ] .
full circles : calculated values of the isb corrections to the fermi transitions in @xmath197 mirror nuclei .
open circles with errors : results calculated by severijns _
since the calculations were performed in a relatively large basis , the basis - cut - off - related uncertainty in @xmath198 could be reduced to approximately @xmath199 , cf .
[ sec03c ] . except for one case , theoretical spins and parities of decaying states
were taken equal to those found in experiment : @xmath200(g.s . ) . only for @xmath201 ,
no @xmath202 component was found in the hf wave function , and thus the lowest solution corresponding to @xmath203 was taken instead .
it should be mentioned that , owing to the poor spectroscopic quality of sv , the projected states corresponding to @xmath204(g.s . )
are not always the lowest ones .
this situation occurs for @xmath205 , 25 , and 45 , where the lowest states have @xmath203 , @xmath206 , and @xmath207 , and the corresponding @xmath198 values are 0.308 % , 0.419% , and 0.636% , respectively .
a relatively strong dependence of the calculated isb corrections on spin is worth noting .
the calculations also indicate an appreciable impact of the signature - symmetry constraint on @xmath198 , in particular , in the @xmath208-shell nuclei with @xmath209 , 47 , and 49 .
a similar effect was calculated for the @xmath137 transitions , see @xmath25-values at fixed shape - current orientations in tables [ tab2 ] and [ tab3 ] .
the @xmath22 values extracted by using diverse techniques including @xmath210 nuclear decays , nuclear mirror decays , neutron decay , and pion decay are subject to both experimental and theoretical uncertainties .
the latter pertain to calculations of radiative processes and for nuclear methods to the nuclear isb effect . the uncertainties in radiative and isb corrections affect the overall precision of @xmath22 at the level of a few parts per 10@xmath211 each @xcite .
it should be stressed , however , that the isb contribution to the error bar of @xmath212 was calculated only for a single theoretical model ( sm - ws ) .
other microscopic models , including the sm - hf @xcite , rh - rpa @xcite , and projected dft @xcite , yield @xmath25 corrections that may differ substantially from those obtained in sm - ws calculations .
inclusion of the model dependence in the calculated uncertainties is expected to increase the uncertainty of @xmath22 . according to ref .
@xcite the increase can reach even an order of magnitude . in our opinion , a reasonable assessment of systematic errors ( due to the model dependence ) can not be done at present , as it requires the assumption that all the nuclear structure models considered are either equally reliable or their performance can be graded in an objective way . a good way to verify the reliability of various models is to compare their predictions with empirically determined @xmath25 .
recently , an anomalously large value of @xmath213% has been determined from a precision measurement of the @xmath75 yields following the @xmath0-decay of @xmath214 state in @xmath28cl to its isobaric analogue state ( fermi branch ) in @xmath28s @xcite .
this value offers a stringent test on nuclear - structure models , because it is significantly larger than any value of @xmath25 in the @xmath215 nuclei .
the physical reason for this enhancement can be traced back to a mixing of two close - lying @xmath216 states seen in @xmath28s at the excitation energies of 7002kev and 7190kev , respectively @xcite .
the lower one is the isobaric analogue state having predominantly @xmath145 component while the higher one is primarily of @xmath217 character .
the experimental value @xmath213% is consistent with the sm - ws calculations : @xmath218% . in our projected - dft approach ,
we also see fingerprints of the strong enhancement in @xmath219 value in @xmath28cl as compared to other @xmath215 nuclei .
unfortunately , a static dft approach based on projecting from a single reference state is not sufficient to give a reliable prediction .
this is because , as sketched in fig .
[ fig14 ] , there exist ambiguities in selecting the hf reference state . in the extreme isoscalar s.p .
scenario , by distributing four valence protons and neutrons over the nilsson s.p .
levels in an odd - odd nucleus , one can form two distinctively different s.p .
configurations , see fig .
[ fig14 ] . schematic illustration of several possible mean - field configurations in the odd - odd @xmath220 ( left ) and even - even @xmath76 ( right ) nuclei .
the pairs of proton ( neutron ) s.p .
levels , labeled as @xmath221 and @xmath222 ( @xmath223 and @xmath224 ) , are assumed to be degenerated due to the intrinsic signature symmetry .
the orbits @xmath223 and @xmath221 carry the signature quantum number @xmath225 @xmath226 while @xmath224 and @xmath222 have @xmath227 @xmath228 . ]
the total signature of valence particles determines the total signature of the odd - odd nucleus and , in turn , an approximate angular - momentum distribution in its wave function @xcite ; the total additive signature @xmath229 corresponds then to even ( odd ) spins in the wave function @xcite .
it is immediately seen that the anti - aligned configuration shown in fig .
[ fig14 ] has @xmath230 ; hence , in the first approximation , it can be disregarded . in this sense ,
the reference wave function in @xmath28cl ( or , in general , in any @xmath231 odd - odd nucleus ) corresponds to the uniquely defined aligned state . as seen in fig .
[ fig14 ] , this does not hold for @xmath28s ( or , in general , for any @xmath76 even - even nucleus ) , where one must consider two possible slater determinants having @xmath232 , obtained by a suitable proton or neutron particle - hole excitation .
the above discussion indicates that , contrary to transitions involving the odd - odd @xmath76 nuclei studied in sec .
[ sec03 ] , those involving even - even @xmath76 nuclei can not be directly treated within the present realization of the model . to this end
, the model requires enhancements including the configuration mixing ( multi - reference dft ) .
nevertheless , we have carried out an exploratory study by independently calculating two isb corrections for the two configurations discussed above .
these calculations proceeded in the following way : * we select the appropriate reference configurations which , in the present case , are : @xmath233 \pi [ 5,6,3,3]$ ] in @xmath28cl and @xmath234 : @xmath235 \pi [ 4,6,3,3]$ ] and @xmath236 : @xmath237 \pi [ 5,5,3,3]$ ] in @xmath28s .
the labels denote the numbers of neutrons and protons occupying the lowest nilsson levels in each parity - signature block @xmath238 counting from the bottom of the hf potential well , as defined in ref .
* we determine the lowest @xmath239 and @xmath240 ( @xmath241 ) states by projecting onto subspaces of good angular momentum and isospin , and performing the @xmath44-mixing and coulomb rediagonalization as described in sec .
[ model ] . * finally
, we calculate matrix elements of the fermi operator @xmath242 and extract @xmath25 . the resulting isb corrections are @xmath243% and @xmath244% for the @xmath245 and @xmath236 configurations , respectively . as before
, we assumed a 10% error due to the basis size ( @xmath98 spherical ho shells ) .
projections from the same configurations cranked in space to @xmath246 = 1@xmath247 ( see discussion in ref .
@xcite ) leaves isb corrections almost unaffected : @xmath248% and @xmath249% .
a simple average value would read @xmath250% , which is indeed strongly enhanced as compared to the @xmath215 cases .
the obtained central value is smaller than both the empirical value and the sm - ws result .
it is worth noting , however , that within the stated errors our mean value 3.4(10)% agrees with the sm - ws value 4.6(05)% .
whether or not the configuration - mixing calculations would provide a significant enhancement is an entirely open question .
within the recently - developed unpaired projected - dft approach , we carried out systematic calculations of isospin mixing effects and isb corrections to the superallowed @xmath26 fermi decays in @xmath251 nuclei and @xmath0-transitions between the isobaric analogue states in mirror @xmath27 nuclei with @xmath252 .
our predictions are compared with empirical values and with predictions of other theoretical approaches .
using isospin - breaking corrections computed in our model , we show that the unitarity of the ckm matrix is satisfied with a precision better than 0.1% .
we also provide isb corrections for heavier nuclei with @xmath253 nuclei that can guide future experimental and theoretical studies .
we carefully analyze various model assumptions impacting theoretical uncertainties of our calculations : basis truncation , definition of the intrinsic state , and configuration selection . to assess the robustness of our results with respect to the choice of interaction
, we compared sv results with predictions of the new force shz2 that has been specifically developed for this purpose .
the comparison of sv and shz2 results suggest that isb corrections are sensitive to the interplay between the bulk symmetry energy and time - odd mean - fields .
while the overall agreement with the empirical values offered by the projected - dft approach is very encouraging , and the results are fairly robust , there is a lot of room for systematic improvements .
the main disadvantages of our model in its present formulation include : ( i ) lack of pairing correlations ; ( ii ) lack of ph interaction ( or functional ) of good spectroscopic quality ; ( iii ) the use of a single hf reference state that can not accommodate configuration mixing effects ; ( iv ) ambiguities in establishing the hf reference state in odd and odd - odd nuclei caused by different possible orientations of time - odd currents with respect to total density distribution .
the work on various enhancements of our model , including the inclusion of @xmath217 and @xmath145 pairing within the projected hartree - fock - bogoliubov theory , better treatment of configuration mixing using the multi - reference dft , and development of the spectroscopic - quality edf used in projected calculations , is in progress . this work was supported in part by the academy of finland and university of jyvskyl within the fidipro programme , and by the office of nuclear physics , u.s .
department of energy under contract nos .
de - fg02 - 96er40963 ( university of tennessee ) and de - sc0008499 ( nuclei scidac-3 collaboration ) .
we acknowledge the csc - it center for science ltd , finland , for the allocation of computational resources .
59ifxundefined [ 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 10.1103/physrevlett.94.092502 [ * * , ( ) ] link:\doibase 10.1103/physrevc.71.055501 [ * * , ( ) ] link:\doibase 10.1103/physrevc.77.025501 [ * * , ( ) ] link:\doibase 10.1103/physrevc.79.055502 [ * * , ( ) ] link:\doibase 10.1103/physrevc.82.065501 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.032002 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.10.531 [ * * , ( ) ] @noop * * , ( ) http://stacks.iop.org/0954-3899/37/i=7a/a=075021 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevc.53.2163 [ * * , ( ) ] link:\doibase 10.1103/physrevc.79.064316 [ * * , ( ) ] link:\doibase 10.1103/physrevc.79.035502 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.106.132502 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/revmodphys.73.463 [ * * , ( ) ] @noop * * ( ) @noop * * , ( ) @noop _ _ ( , ) link:\doibase 10.1103/physrevlett.103.012502 [ * * , ( ) ] link:\doibase 10.1103/physrevc.81.054310 [ * * , ( ) ] @noop _ _ ( , ) @noop * * , ( ) link:\doibase 10.1103/physrevc.76.044304 [ * * , ( ) ] link:\doibase 10.1103/physrevc.52.2455 [ * * , ( ) ] link:\doibase 10.1103/physrevc.78.035501 [ * * , ( ) ] link:\doibase 10.1103/physrevc.80.064319 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevc.79.044318 [ * * , ( ) ] link:\doibase 10.1103/physrevc.84.041304 [ * * , ( ) ] link:\doibase 10.1103/physrevc.76.054315 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevc.77.024316 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.181803 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.142302 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.64.491 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevc.71.024305 [ * * , ( ) ] link:\doibase 10.1103/physrevc.74.011301 [ * * , ( ) ] link:\doibase 10.1103/physrevc.53.214 [ * * , ( ) ] link:\doibase 10.1103/physrevc.78.064311 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevc.78.055501 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/revmodphys.78.991 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.107.182301 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) | background : : the superallowed @xmath0-decay rates provide stringent constraints on physics beyond the standard model of particle physics . to extract crucial information about the electroweak force , small isospin - breaking corrections to the fermi matrix element of superallowed transitions must be applied .
purpose : : we perform systematic calculations of isospin - breaking corrections to superallowed @xmath0-decays and estimate theoretical uncertainties related to the basis truncation , time - odd polarization effects related to the intrinsic symmetry of the underlying slater determinants , and to the functional parametrization .
methods : : we use the self - consistent isospin- and angular - momentum - projected nuclear density functional theory employing two density functionals derived from the density independent skyrme interaction .
pairing correlations are ignored .
our framework can simultaneously describe various effects that impact matrix elements of the fermi decay : symmetry breaking , configuration mixing , and long - range coulomb polarization .
results : : the isospin - breaking corrections to the @xmath1 pure fermi transitions are computed for nuclei from @xmath2=10 to @xmath2=98 and , for the first time , to the fermi branch of the @xmath3 transitions in mirror nuclei from @xmath2=11 to @xmath2=49 .
we carefully analyze various model assumptions impacting theoretical uncertainties of our calculations and provide theoretical error bars on our predictions .
conclusions : : the overall agreement with empirical isospin - breaking corrections is very satisfactory . using computed isospin - breaking corrections we show that the unitarity of the ckm matrix is satisfied with a precision better than 0.1% . |
accessible information about a quantum system is restricted by the noncommutability of observables .
the nature of this restriction can be classified essentially into two categories : fluctuations inherent in a quantum system and the error caused by the process of measurement .
these aspects of uncertainty constitute the two distinctive features of quantum mechanics .
the kennard - robertson uncertainty relation such as @xmath0 describes quantum fluctuations that are independent of the measurement process @xcite .
according to bell s theorem @xcite , this type of quantum fluctuations prohibits us from presupposing any `` element of reality '' @xcite behind the probability distributions of observables . the measurement error , on the other hand
, is determined by the process of measurement which is characterized by a positive operator - valued measure ( povm ) @xcite . in the idealized error - free limit
, quantum measurement is described by projection operators which , however , can not always be implemented experimentally .
information about more than one observable can be obtained from a single povm in simultaneous measurement of two noncommuting observables and quantum - state tomography .
it is known , however , that , in simultaneous measurements , at least one of the observables can not be measured without incurring a measurement error @xcite . in this context , various uncertainty relations between the measurement errors of noncommuting observables have been studied @xcite . in this paper , we quantify the measurement accuracy and the measurement error of observables in terms of a given povm @xmath1 by introducing @xmath2 accuracy matrix @xmath3 calculated from the povm . based on this accuracy matrix , we derive trade - off relations between the measurement accuracy of two or three observables , these being stronger trade - off relations than those derived in our previous work @xcite
. they can be interpreted as the uncertainty relations between the measurement errors of noncommuting observables in simultaneous measurements or as the uncertainty relations between the measurement error and back - action of the measurement @xcite .
in addition , a no - cloning inequality @xcite is derived from the trade - off relations . in a rather different context
, the maximum - likelihood estimation @xcite has been investigated as the standard scheme of quantum state tomography for a finite number of samples .
several studies have focused on the efficiency and optimality of the estimation of an unknown quantum state @xcite .
we show that our characterization of the measurement accuracy can be related to the maximum - likelihood estimation and that the accuracy matrix can be interpreted as an average of the fisher information matrix over the state to be measured .
the trade - off relations can also be interpreted as those concerning the accuracy of the estimate of various probability distributions of noncommuting observables .
the constitution of this paper is as follows . in sec .
ii , we formulate the general quantum measurement of a qubit ( spin-1/2 ) system . in sec .
iii , we define the accuracy matrix and investigate its properties . based on this accuracy matrix , we define the accuracy parameter and error parameter in a particular direction of measurement . in sec .
iv , we derive the trade - off relations between the accuracy parameters or the error parameters in two or three directions . in sec .
v , we apply the trade - off relations to specific problems : the uncertainty relations between measurement errors in nonideal joint measurements , the uncertainty relations between the error and back - action , a no - cloning inequality , and quantum state tomography . in sec .
vi , we point out a close connection between the accuracy matrix and the fisher information matrix .
we conclude this paper in sec .
we consider a quantum measurement described by povm @xmath1 ( @xmath4 ) on state @xmath5 of a qubit system , where @xmath6 denotes the outcome of the measurement .
povm @xmath7 satisfies @xmath8 , with @xmath9 being the identity operator , and can be parameterized as @xmath10 where @xmath11 represents the pauli matrices .
the requirements that the sum of @xmath12 s equals the identity operator and that all of them be nonnegative are met if and only if @xmath13 we can also parameterize density operator @xmath5 as @xmath14 where @xmath15 is the bloch vector satisfying @xmath16 .
conversely , for a given @xmath5 , @xmath15 is calculated as @xmath17 . the probability of obtaining the measurement outcome @xmath6 is then given by @xmath18 any observable @xmath19 of the qubit system can be diagonalized as @xmath20 where @xmath21 and @xmath22 are the corresponding eigenvalues , @xmath23 and @xmath24 are projection operators with @xmath25 being a three - dimensional unit vector , and @xmath26 the probability distribution of observable @xmath19
is then given by @xmath27 if we are not interested in eigenvalues of the observables but are only concerned with the directions ( @xmath28 ) of the outcome , we can replace @xmath19 with @xmath29 by setting @xmath30 . in the following analysis
, we identify observable @xmath31 with the observable @xmath29 and refer to the probability distribution in eq .
( [ distribution ] ) as that in the direction of @xmath25 .
we discuss three typical examples .
* example 1 * ( _ projection measurement _ ) .
we can precisely measure @xmath29 by the projection measurement described by the povm @xmath32 . * example 2 * ( _ nonideal measurement _ )
. a more general class of measurements can be described by the povm @xmath7 consisting of two positive operators parametrized as @xmath33 where @xmath25 is a unit vector , @xmath34 , @xmath35 , @xmath36 , and @xmath37 .
this povm corresponds to a nonideal measurement of the observable @xmath29 @xcite .
it can be reduced to a projection measurement @xmath38 if and only if @xmath39 and @xmath40 . on the other hand ,
the povm is trivial ( i.e. , @xmath41 and @xmath42 ) if and only if @xmath43 ; then we can not obtain any information about @xmath5 .
equations ( [ nonideal1 ] ) can be rewritten as @xmath44 where @xmath45 is the @xmath46 transition - probability matrix @xmath47 which satisfies @xmath48 and @xmath49 .
note that @xmath45 describes a binary symmetric channel @xcite if and only if @xmath40 and @xmath50 .
it follows from eq .
( [ projection - povm ] ) that any measurement process described by a povm consisting of two positive operators is formally equivalent to a measurement process in which a classical error is added to the projection measurement .
the physical origin of this error , however , lies in the quantum - mechanical interaction .
* example 3 * ( _ probabilistic measurement _ ) .
suppose that a nonideal measurement of @xmath51 is performed with probability @xmath52 ( @xmath53 ) and that @xmath54 is performed with probability @xmath55 .
the povm corresponding to this probabilistic measurement consists of four operators : @xmath56 as the number of measured samples increases , this measurement asymptotically approaches the measurements on @xmath57 identically prepared samples which are divided into two groups in the ratio @xmath58 , with @xmath59 being measured for the first group and @xmath60 for the second group .
other important examples such as nonideal joint measurements and quantum state tomography are discussed in sec .
we will characterize the accuracy of an arbitrary observable in such a manner that it depends only on the process of measurement and not on the measured state @xmath61 .
we first define the accuracy matrix . * definition 1 * ( _ accuracy matrix _ ) . the @xmath2 accuracy matrix @xmath62 characterizing the measurement accuracy of observables in terms of the povm @xmath7
is defined as @xmath63 where @xmath64 denotes the @xmath65th component of the real vector @xmath66 and @xmath67 shows indices of matrix elements of @xmath62 .
we introduce the notation @xmath68 with @xmath69 as @xmath70 that is , @xmath71 denotes the transposed vector of @xmath72 and @xmath68 denotes the projection matrix onto direction @xmath72 in @xmath73 whose @xmath67 matrix element is given by @xmath74
. we can then rewrite ( [ am1 ] ) in matrix form as @xmath75 note that @xmath62 is positive semidefinite and hermitian , and can therefore be diagonalized by an orthonormal transformation .
the physical meaning and useful properties of the accuracy matrix will be investigated subsequently , and its foundation from an information - theoretic point of view will be established in terms of the maximum - likelihood estimation of the probability distribution of observables in sec . vi .
in fact , the accuracy matrix is closely related to fisher information matrix ( [ fisher1 ] ) or ( [ fisher2 ] ) , although physical quantities such as the measurement error can be directly derived from the accuracy matrix without resort to fisher information .
noting that @xmath76 , we can obtain the following fundamental inequality which forms the basis of trade - off relations to be discussed later .
* theorem 1 .
* three eigenvalues @xmath77 of @xmath62 satisfy @xmath78 or equivalently , @xmath79 where we denote the trace of the @xmath2 matrix as @xmath80 to reserve symbol @xmath81 for the trace of a quantum - mechanical @xmath46 matrix .
the equality @xmath82 , or @xmath83 , holds if and only if @xmath84 for all @xmath6 .
the following corollary follows from the positivity of @xmath62 .
* corollary 1 . *
the accuracy matrix satisfies the following matrix inequality : @xmath85 where @xmath86 is the @xmath2 identity matrix , and @xmath87 means that all eigenvalues of @xmath88 are nonnegative .
the following examples illustrate the physical meaning of the accuracy matrix .
we first consider a nonideal quantum measurement ( see also example 2 in sec .
ii ) . we can rewrite eq.([nonideal1 ] ) as @xmath89 where @xmath90 and @xmath91 .
the accuracy matrix can then be represented by @xmath92 where @xmath93 is the eigenvalue of @xmath94 corresponding to the eigenvector @xmath25 , and is given by @xmath95 we can also write @xmath93 in terms of the transition - probability matrix introduced in eq .
( [ transition ] ) as @xmath96 the accuracy parameter @xmath93 satisfies @xmath97 where @xmath98 holds if and only if @xmath99 and @xmath40 ; that is , @xmath7 describes the projection measurement of observable @xmath29 . note that @xmath100 holds in this case . on the other hand
, @xmath101 holds if and only if @xmath102 . in this case , @xmath103 holds , and we can not obtain any information about @xmath5 .
the nonzero eigenvalue @xmath93 thus characterizes the measurement accuracy of @xmath29 ; the larger @xmath93 , the more information we can extract about @xmath29 from the measurement outcome .
these properties can be generalized for an arbitrary povm as shown below .
another example is the probabilistic measurement of two noncommuting observables ( see example 3 in sec .
we consider the nonideal measurement of @xmath59 whose accuracy matrix is @xmath104 and that of @xmath60 whose accuracy matrix is @xmath105 .
the accuracy matrix of the probabilistic measurement is given by @xmath106 this representation suggests that the measurement accuracy concerning @xmath59 is degraded by a factor of @xmath52 compared with the single nonideal measurement of @xmath59 , because we can not observe @xmath59 with probability @xmath107 .
a similar argument applies to @xmath60 as well .
equation ( [ accuracy1 ] ) shows that @xmath62 is the linear combination of the accuracy matrices of povms measuring @xmath59 and @xmath60 , where the coefficients @xmath52 and @xmath55 give the probabilities of measuring @xmath59 and @xmath60 , respectively .
this can be generalized as follows .
let us consider three povms : @xmath108 , @xmath109 , and @xmath110 with @xmath111 .
the povm @xmath7 describes the probabilistic measurement of @xmath112 with probability @xmath52 and that of @xmath113 with probability @xmath107 . according to the definition of the accuracy matrix
, we have @xmath114 we thus obtain the following theorem .
* theorem 2 * ( _ linearity _ ) : @xmath115 or more symbolically , @xmath116 note that we can take as a scalar measure of the measurement accuracy the largest eigenvalue of the accuracy matrix which we denote as @xmath117 .
it satisfies @xmath118 , where @xmath119 holds if and only if @xmath7 describes the projection measurement of a particular direction and @xmath120 if and only if the povm is trivial : @xmath121 , where @xmath9 is the identity operator and @xmath122 denotes the probability of finding outcome @xmath6 , with @xmath123
. we may alternatively choose the scalar measure to be @xmath124 ; it has the linear property from theorem 2 and satisfies @xmath125 , where @xmath126 if and only if the povm is trivial .
we next parametrize the measurement accuracy of a particular observable .
we denote the support of @xmath127 as @xmath128 ; that is , @xmath128 is the subspace of @xmath129 spanned by all eigenvectors of @xmath62 with nonzero eigenvalues .
* definition 2 * ( _ measurement accuracy _ ) .
the accuracy parameter @xmath130 in direction @xmath131 is defined as @xmath132 where @xmath133 is assumed to act only on subspace @xmath128 .
if @xmath134 , we set @xmath135 .
this definition is closely related to the fisher information concerning a particular direction defined in eq .
( [ fisher3 ] ) . * definition 3 * ( _ measurement error _ ) the error parameter of the measurement in direction @xmath25 is defined as @xmath136 the parameters @xmath130 and @xmath137 satisfy the following inequalities .
* theorem 3 * : @xmath138 @xmath139 the equality @xmath140 , or equivalently @xmath141 , holds if and only if the measurement described by @xmath7 is equivalent to a projection measurement in direction @xmath25 . in this case , the measurement involves no measurement error .
the other limit of @xmath142 , or equivalently @xmath143 , holds if and only if @xmath144 . in this case
, we can not obtain any information about direction @xmath25 from the measurement .
* proof * since @xmath3 commutes with the identity operator @xmath86 , we can show that @xmath145 from inequality ( [ trade - off2 ] ) in corollary 1 .
we thus obtain @xmath146 inequalities ( [ inequality1 ] ) and ( [ inequality2 ] ) are the direct consequences of this inequality . the condition that @xmath147 and @xmath148 hold follows from the definitions of @xmath149 and @xmath150 .
we next show the condition that @xmath151 and @xmath152 hold . if @xmath7 is the projection measurement in direction @xmath25 , then @xmath153 .
conversely , from inequality ( [ trade - off1 ] ) , it can be shown that if @xmath140 and @xmath154 hold , then @xmath25 is the eigenvector corresponding to eigenvalue @xmath155 and that the other two eigenvalues are @xmath156 .
it follows from the condition of equality @xmath157 in theorem 1 that @xmath158 for all @xmath6 .
therefore , without loss of generality , we can write the povm as @xmath159 where @xmath160 , because @xmath161 and @xmath162 hold .
we define two operators as @xmath163 then @xmath164 describes the projection measurement in direction @xmath25. let @xmath165 , @xmath166 , and @xmath167 be the eigenvectors of @xmath62 , and @xmath168 , @xmath169 , and @xmath170 be the corresponding eigenvalues .
it can be shown that @xmath171 according to theorem 1 , we can not simultaneously measure the three directions corresponding to the eigenvectors with the maximum accuracy @xmath172 for all @xmath65 ; the trade - off relation ( [ trade - off1 ] ) or ( [ trade - off1 ] ) is equivalent to @xmath173 this trade - off relation represents the uncertainty relation between the measurement errors in the three directions .
we define that the povm @xmath7 is optimal if and only if @xmath174 ; that is , @xmath7 reaches the upper bound of trade - off relation ( [ trade - off1 ] ) , ( [ trade - off1 ] ) , or ( [ trade - off1 ] ) . on the other hand
, we define that @xmath7 is symmetric if and only if @xmath175 holds for any @xmath25 and @xmath176 . in this case
, @xmath62 is proportional to the @xmath2 identity matrix .
we next introduce the concept of `` reconstructive subspace '' and `` reconstructive direction . ''
the following theorem can be directly shown from the definition of the accuracy matrix .
* theorem 4 * @xmath128 corresponds to the subspace spanned by the set of basis vectors @xmath177 of the accuracy matrix ( [ am2 ] ) .
suppose that we perform the measurement @xmath178 and obtain the probability distribution @xmath179 for each outcome @xmath6 .
can we then reconstruct the premeasurement distribution @xmath180 of the system from @xmath179 ?
the answer is given by the following theorem .
* theorem 5 * ( _ reconstructive subspace and reconstructive direction _ ) we can reconstruct the probability distribution @xmath181 from the measured distribution @xmath179 if and only if @xmath131 .
we thus refer to @xmath128 as a reconstructive subspace and to a unit vector in @xmath128 as a reconstructive direction .
* proof * we can show from eq .
( [ qk ] ) that @xmath182 where @xmath183 is a @xmath184 matrix : @xmath185 let @xmath186 and @xmath187 be the kernel and image of @xmath183 , respectively
. it can easily be shown that @xmath188 .
let us introduce the equivalence relation `` @xmath189 '' as @xmath190 .
we denote the equivalence class of @xmath191 as @xmath192 $ ] , where @xmath193 $ ] is an element of the quotient space @xmath194 . from the homomorphism theorem ,
the quotient map @xmath195 is a linear isomorphism from @xmath194 to @xmath196 . noting that @xmath197
, we obtain @xmath198 = ( m/ \sim ) ^{-1 } \left ( \left ( \begin{array}{c } q_1
\\ \vdots \\
q_m \end{array }
\right ) - \left ( \begin{array}{c } r_1 \\ r_2 \\ \vdots \\ r_m \end{array }
\right ) \right).\ ] ] by taking a representative @xmath199 $ ] , we can reconstruct @xmath200 as @xmath201 which gives @xmath202 through eq .
( [ distribution]). we consider the nonideal measurement of @xmath51 with the povm @xmath203 in eq .
( [ nonideal1 ] ) .
the nonideal measurement is characterized with the accuracy matrix @xmath104 . in this case
, we can show that @xmath204 .
it follows that @xmath205 and @xmath206 for @xmath207 .
we next consider the probabilistic measurement of @xmath208 and @xmath209 in example 3 in sec .
the probabilistic measurement is characterized with the accuracy matrix of the joint povm @xmath7 given in ( [ accuracy1 ] ) , so the reconstructive subspace is two dimensional : @xmath210 .
a straightforward calculation shows that @xmath211 if the classical noise described by a transition - probability matrix is added to the measurement outcomes , the measurement accuracy should deteriorate .
this fact can be expressed as a data processing inequality . *
theorem 6 * ( _ data processing inequality _ ) .
suppose that two povms @xmath212 and @xmath213 are related to each other by @xmath214 where @xmath215 is an @xmath216 transition - probability matrix satisfying @xmath217 .
it then follows that @xmath218 where matrix inequality ( [ data - processing ] ) means that all the eigenvalues of @xmath219 are non - negative . * proof * we can parametrize the povms as @xmath220 @xmath221 where @xmath222 . introducing the function @xmath223 , with arbitrary vector @xmath224 as @xmath225 we can show that @xmath226 @xmath227 the hessian of @xmath228 , which is defined as @xmath229 , becomes @xmath230 so that @xmath228 is a concave function
. therefore @xmath231 holds for any @xmath232 satisfying @xmath233 and @xmath234 for all @xmath6 .
taking @xmath235 , inequality ( [ con2 ] ) becomes @xmath236 or equivalently , @xmath237 noting that @xmath238 , we obtain @xmath239 which implies that @xmath240 since ( [ 52 ] ) holds for arbitrary @xmath241 , we obtain ( [ data - processing]). the following corollary is a direct consequence of the foregoing theorem .
* corollary 6 * suppose that @xmath242 is obtained by a coarse graining of @xmath212 : @xmath243 , and @xmath244 , with @xmath245
. then @xmath246 holds .
inequality ( [ grain ] ) means that the measurement accuracy in any direction is decreased by a coarse graining .
we can also express the data processing inequality in terms of the accuracy parameter in an arbitrary direction .
* theorem 7 * ( _ data processing inequality _ ) .
we consider the povms @xmath7 and @xmath112 satisfying eq .
( [ data - processing2 ] ) .
suppose that @xmath247 .
then , @xmath248 holds , or equivalently , @xmath249 holds for arbitrary @xmath25 .
* proof * let @xmath168 , @xmath169 , and @xmath170 be the eigenvalues of @xmath3 , and @xmath165 , @xmath166 , and @xmath167 be the corresponding eigenvectors .
similarly , let @xmath250 , @xmath251 , and @xmath252 be the eigenvalues of @xmath253 , and @xmath254 , @xmath255 , and @xmath256 be the corresponding eigenvectors .
it follows from the data processing inequality ( [ data - processing ] ) that @xmath257 for @xmath258 , where @xmath259 . applying the concave inequality to @xmath260
, we obtain @xmath261 for arbitrary @xmath25 , we can show that @xmath262 which implies ( [ data - processing3 ] ) and ( [ data - processing4]).
we now derive general trade - off relations between the measurement errors of noncommuting observables , which are the main results of this paper .
let @xmath165 , @xmath166 , and @xmath167 be the respective eigenvectors of @xmath62 corresponding to the eigenvalues @xmath168 , @xmath169 , and @xmath170 , where @xmath263 ( @xmath258 ) .
we define the error parameters as @xmath264 .
inequality ( [ trade - off1 ] ) or ( [ trade - off1 ] ) in theorem 1 can be rewritten in terms of the error parameters as @xmath265 considering two eigenvalues alone ( i.e. , @xmath266 ) , we can simplify the trade - off relation : @xmath267 the trade - off relations ( [ trade - off1 ] ) and ( [ trade - off1 ] ) can be generalized to the case of arbitrary directions .
we first consider the case of two observables .
* theorem 8 * ( _ trade - off relation _ ) .
we consider a simultaneous measurement in two directions @xmath268 and @xmath269 ( @xmath270 ) described by the povm @xmath7 .
we assume @xmath271 and @xmath272 , and define @xmath273 and @xmath274 ( @xmath275 ) . then the trade - off relation @xmath276 or equivalently
, @xmath277 holds . * proof .
* we divide the proof into two steps .
we consider a situation in which both @xmath268 and @xmath269 lie in a plane spanned by two eigenvectors . without loss of generality ,
we choose @xmath165 and @xmath166 as the two eigenvectors , and expand @xmath268 and @xmath269 as @xmath278 @xmath279 where @xmath280 .
it can be shown that @xmath281 @xmath282 applying the cauchy - schwarz inequality and @xmath283 , we obtain @xmath284 the equality @xmath285 holds if and only if @xmath286 and @xmath287 . in the case of @xmath288
( i.e. , the measurement errors are symmetric ) , the equality holds if and only if @xmath289 .
_ we next consider a more general case .
we choose an orthonormal basis @xmath290 such that both @xmath268 and @xmath269 are in the plane spanned by @xmath291 and @xmath292 .
we introduce the notation @xmath293 ( @xmath258 ) .
let @xmath294 be a @xmath2 orthogonal matrix which transforms @xmath295 into @xmath290 .
it can be shown that @xmath296 note that @xmath297 because @xmath294 is an orthogonal matrix , and that the function @xmath298 is concave .
it follows from a concave inequality that @xmath299 combining this with @xmath300 , we obtain @xmath301 therefore @xmath302 this inequality means that @xmath303 , or equivalently , @xmath304 we can derive inequality ( [ trade - off3 ] ) by following the same procedure as in step 1 .
we can directly derive inequality ( [ trade - off4 ] ) from ( [ trade - off3]). we note that the equalities in ( [ trade - off3 ] ) and ( [ trade - off4 ] ) hold in the case that the povm @xmath7 is given by @xmath305 ( @xmath306 ) , where @xmath307 and @xmath308 .
the accessible regime for @xmath309 and @xmath310 is illustrated in fig.1 for the case of @xmath311 , @xmath312 , and @xmath313 .
note that regime q can be reached only through simultaneous measurement for the case of @xmath314 @xcite .
, the union of p and q indicate the regime satisfying inequality ( [ trade - off4 ] ) for the case of @xmath314 , and the union of p , q and r indicate the regime satisfying the inequality for the case of @xmath315 .
we can only access regime q through simultaneous measurement for the case of @xmath314.,width=264 ] the trade - off relation can be interpreted as the uncertainty relation between measurement errors @xcite .
it offers a rigorous representation of bohr s principle of complementarity @xcite which dictates `` the mutual exclusion of any two experimental procedures '' when we measure two noncommuting observables simultaneously .
the trade - off relation between three observables can be formulated as follows . * theorem 9 * we consider a simultaneous measurement in three directions @xmath268 , @xmath269 , and @xmath316 described by the povm @xmath7 .
let us assume that @xmath268 , @xmath269 , and @xmath316 are linearly independent .
we set the notation @xmath317 and @xmath318 , where @xmath319 .
then the inequality @xmath320 holds . the equality in ( [ trade - off5 ] ) holds if and only if @xmath321 and @xmath322 are orthogonal .
* proof * introducing the notation @xmath323 where @xmath324 ( @xmath325 ) , it can be shown that @xmath326 we thus obtain @xmath327 we can show @xmath328 from @xmath329 ; therefore we obtain ( [ trade - off5]).
we have discussed in sec .
iv trade - off relations ( [ trade - off3 ] ) and ( [ trade - off4 ] ) which describe the uncertainty relations in generalized simultaneous measurements . in this section ,
we discuss possible applications of these trade - off relations .
we consider a class of simultaneous measurements called nonideal joint measurements , where two observables @xmath51 and @xmath54 are simultaneously measured . since their eigenvalues are @xmath330 , each measurement should give a pair of outcomes @xmath331 ( @xmath332 ) for observables @xmath59 and @xmath60
. the joint povm @xmath333 can be parametrized as @xmath334 the marginal povms @xmath335 ( @xmath275 ) are defined by @xmath336 and can be parametrized by @xmath337 where @xmath338 this simultaneous measurement can be regarded as a nonideal joint measurement @xcite if and only if the marginal povm @xmath339 corresponds to the nonideal measurement of @xmath340 , that is , @xmath341 we can define the @xmath342 transition - probability matrices of @xmath343 as in eq .
( [ transition ] ) : @xmath344 in this case , we can calculate the accuracy of @xmath59 and @xmath60 by two different methods . one method to calculate the accuracy parameter is based on the joint povm @xmath7 : @xmath345 where @xmath275 .
the other is based on the marginal povm @xmath343 : @xmath346 note that @xmath347 .
these two accuracy parameters are equivalent as shown by the following theorem .
* theorem 10 * : @xmath348 the proof of theorem 10 is given in the appendix .
for @xmath349 , we can show that @xmath25 is not an element of @xmath350 ; therefore , @xmath351 . on the other hand
, @xmath352 holds by definition .
we can thus obtain the following corollary .
* corollary 10 * for arbitrary @xmath25 ,
@xmath353 we next discuss the relationship between the present work and our earlier work @xcite for the case of nonideal joint measurement . in ref .
@xcite , we have introduced the accuracy parameter @xmath354 and error parameter @xmath355 as @xmath356 on the other hand , the accuracy parameter @xmath357 and error parameter @xmath358 in the present paper are given by @xmath359 it can be easily shown that @xmath360 so the trade - off relations derived in the present paper are stronger than our previous ones ( @xmath361 and @xmath362 ) derived in ref .
the latter trade - off relations can thus be derived from those obtained in the present paper .
we have interpreted trade - off relation ( [ trade - off3 ] ) as the uncertainty relation between the measurement errors . in this subsection , we show that it can be interpreted as the uncertainty relation between the measurement error and back - action of the measurement @xcite . let us suppose that @xmath363 is a state immediately after the measurement of @xmath51 for the premeasurement state @xmath5 .
if the measurement of @xmath59 is described by measurement operators @xmath364 , we can write @xmath365 as @xmath366 . for simplicity , we assume that the number of measurement outcomes is @xmath367 : @xmath368 . to identify the disturbance of @xmath54 caused by the measurement of @xmath59 , we consider how much information about @xmath60 for the premeasurement state @xmath5 remains in post - measurement state @xmath363 .
we characterize this by considering how much information on @xmath61 can be obtained by performing the projection measurement of @xmath60 for @xmath363 .
note that we can regard the projection measurement of @xmath60 on @xmath365 described by the povm as the measurement of @xmath5 described by the povm @xmath369 , where @xmath370 .
the joint operation of measurement @xmath59 followed by measurement @xmath60 can be described by a povm @xmath371 , where @xmath372 we can construct the marginal povms as @xmath373 it is possible to interpret @xmath374 as a measure of the back - action of @xmath60 caused by measurement of @xmath59 . defining the measurement error of @xmath59 as @xmath375 and the back - action of the measurement on @xmath60 as @xmath376 , we can obtain the trade - off relation between the error and back - action based on inequality ( [ trade - off3 ] ) .
* theorem 11 * ( _ uncertainty relation between measurement error and back - action _ ) .
@xmath377 we note that a non - selective measurement process for @xmath59 can simulate the decoherence caused by the environment . in this case , the trade - off relation ( [ e - b - trade - off ] ) gives a lower bound on the back - action of @xmath60 in the presence of decoherence characterized by @xmath309 .
another application of the trade - off relation is the derivation of a no - cloning inequality .
we consider a quantum cloning process from qubit system @xmath378 to qubit system @xmath379 described as follows : let @xmath5 be an unknown density operator of system @xmath378 to be cloned , @xmath380 be that of system @xmath379 as a blank reference state , and @xmath381 be that of the environment .
the density operator of the total system is initially given by @xmath382 , and becomes @xmath383 after unitary evolution @xmath384 .
we define @xmath385 and @xmath386 . we can write @xmath387 and @xmath388 in the operator - sum representation as @xmath389 and @xmath390 . the no - cloning theorem @xcite states that there exists no unitary operator @xmath384 that satisfies @xmath391 for arbitrary input state @xmath5
. if @xmath384 is the identity operator , then all information about @xmath5 remains in system @xmath378 , and no information is transferred into system @xmath379 ; @xmath392 and @xmath393 . as another special case , if @xmath384 describes the swapping operation between @xmath378 and @xmath379 ( i.e. , @xmath394 )
, then all information about @xmath5 is transferred into @xmath379 with no information left in @xmath378 .
intermediate cases between the identity operation and the swapping operation can be quantitatively analyzed by the no - cloning inequality @xcite .
we derive here another simple no - cloning inequality based on the trade - off relation .
we first consider how much information about @xmath5 remains in @xmath387 .
we can characterize this by considering how much information about @xmath29 of @xmath5 can be obtained by the measurement of @xmath29 on @xmath387 .
we can regard the measurement of @xmath29 on @xmath387 as the measurement described by the povm @xmath395 on @xmath5 , where @xmath396 .
we can thus characterize the amount of information that remains in @xmath378 by the accuracy parameter @xmath397 .
similarly , we can consider how much information about @xmath5 is transferred into @xmath388 .
we characterize this by considering how much information about @xmath29 of @xmath5 can be obtained by the measurement of @xmath29 on @xmath388 .
we can regard the measurement of @xmath29 on @xmath388 as the measurement described by the povm @xmath398 on @xmath5 .
we thus characterize the amount of information which is transferred from @xmath378 to @xmath379 by the accuracy parameter @xmath399 . for mathematical convenience , we use @xmath400 and @xmath401 , instead of @xmath402 and @xmath403 , to derive our no - cloning inequality .
the amount of information about @xmath5 which remains in @xmath378 is characterized by @xmath404 averaged over all directions , and the amount of information about @xmath5 which is transferred into @xmath379 is characterized by @xmath405 averaged over all directions .
* definition 4 * ( _ cloning parameter _ ) .
we define the cloning parameters @xmath406 and @xmath407 as @xmath408 since @xmath409 and @xmath410 , the cloning parameters satisfy @xmath411 the cloning parameters depend only on @xmath380 , @xmath381 , and @xmath384 , and characterize the performance of the cloning machine @xmath412 .
the smaller @xmath406 is , the more information about @xmath5 remains in system @xmath378 , while the smaller @xmath407 is , the more information about @xmath5 is transferred into system @xmath379 by the cloning machine .
for example , if @xmath384 is the identity operator , then @xmath413 and @xmath414 hold , which implies that all information about @xmath5 is left in system @xmath378 . on the other hand ,
if @xmath384 describes the swapping operation between @xmath378 and @xmath379 , then @xmath415 and @xmath416 hold . for intermediate cases between them ,
the following no - cloning inequality between @xmath406 and @xmath407 can be derived from trade - off relation ( [ trade - off3 ] ) .
* theorem 12 * ( _ no - cloning inequality _ ) : @xmath417 * proof .
* it can be shown that there exists a povm @xmath418 , with @xmath419 , satisfying @xmath420 we can also show that @xmath421 and @xmath422 are its marginal povms . from inequality ( [ data - processing5 ] ) and the trade - off relation ( [ trade - off3 ] ) , we obtain @xmath423 averaging ( [ c ] ) over all directions and using @xmath424 @xmath425 we obtain ( [ no - cloning]). inequality ( [ no - cloning ] ) represents the trade - off relation between the information remaining in the original system @xmath378 and the information transferred to the reference system @xmath379 .
the impossibility of achieving @xmath426 implies the no - cloning theorem . note that if @xmath427 , then @xmath428 , which implies that if a cloning machine transfers all of the information about @xmath5 into system @xmath379 , then no information can be left in system @xmath378 .
we next apply our framework to quantum - state tomography @xcite . as shown in sec .
vi , characterization of the measurement accuracy by the accuracy matrix is closely related to the asymptotic accuracy of the maximum - likelihood estimation which is considered to be the standard scheme for quantum - state tomography .
we first consider the standard strategy to estimate the three components of bloch vector @xmath15 .
we divide @xmath57 identically prepared samples into three groups in the ratio @xmath429 , and measure @xmath430 for the first group , @xmath431 for the second group , and @xmath432 for the third group .
as @xmath57 increases , this scheme becomes asymptotically described by povm consisting of six operators : @xmath433 we can reconstruct the quantum state by quantum - state tomography and hence reconstruct the probability distributions in all directions .
in fact , the accuracy matrix for the standard tomography ( [ tomography1 ] ) is given by @xmath434 which attains the upper bound of the inequality @xmath435 .
this expression manifestly shows that the reconstructive subspace of the standard quantum state tomography is @xmath73 and that the accuracy of the tomography is optimal and symmetric in the sense discussed in sec .
iii b. we next consider the minimal qubit tomography .
_ have shown that the following four measured probabilities are just enough to estimate the bloch vector @xcite : @xmath436 where @xmath437 the minimal qubit tomography is also optimal and symmetric , in the sense that the corresponding accuracy matrix is again given by ( [ tomography - accuracy ] ) .
note that the povm @xmath7 satisfying @xmath438 can be regarded as tomographically complete @xcite .
in this section , we point out a close connection between the accuracy matrix and the fisher information @xcite .
we consider the quantum measurements described by the povm @xmath1 for each of @xmath57 ( @xmath439 ) samples prepared in the same unknown state @xmath5 .
note that @xmath440 . our task is to estimate the bloch vector @xmath15 by maximum - likelihood estimation .
we denote @xmath441 as the maximum - likelihood estimator of @xmath15 from @xmath57 measurement outcomes .
the asymptotic accuracy of maximum - likelihood estimation is characterized by the fisher information . in our situation
, the fisher information takes the matrix form given by @xmath442 or equivalently , @xmath443 note that @xmath444 is a @xmath2 positive and hermitian matrix , and that the support of @xmath444 coincides with that of @xmath445 . focusing on a particular direction @xmath25
, we can reduce the fisher information content to @xmath446 the greater the fisher information , the more information we can extract from the measurement outcome . in the case of @xmath447 , the variance of the estimator @xmath448 diverges , so we can not gain any information about the probability distribution in direction @xmath25 .
this is the case of @xmath25 not being in any reconstructive direction . replacing @xmath122 by @xmath449 in the fisher information ( [ fisher1 ] ) or ( [ fisher2 ] ) ,
we can obtain the accuracy matrix in eq .
( [ am1 ] ) or ( [ am2 ] ) .
note that @xmath449 is the average of @xmath122 over the entire bloch sphere .
the trade - off relations ( [ trade - off3 ] ) , ( [ trade - off4 ] ) , and ( [ trade - off5 ] ) can thus be interpreted as the trade - off relations between the asymptotic accuracy of the maximum - likelihood estimation of the probability distributions of observables .
a finite number of samples only gives us imperfect information about the probability distribution of an observable for an unknown state .
as we have shown @xcite , this imperfection further deteriorates in the case of simultaneous estimation due to the noncommutability of the observables .
figure 2 shows the results of simulations for the value of the maximum - likelihood estimators @xmath450 ( red curves ) and @xmath451 ( blue curves ) in the the case of an optimal nonideal joint povm which satisfies the equality in ( [ trade - off3 ] ) or ( [ trade - off4 ] ) with @xmath452 and @xmath453 ( red curves ) and @xmath451 ( blue curves ) for the case of @xmath314 , @xmath454 , and @xmath455 .
the abscissa indicates sample number @xmath57 and the ordinate indicates the value of the estimators .
the number of simulations is @xmath456.,width=264 ] let us next consider a simple estimation scheme by dividing @xmath57 prepared samples into two groups in the ratio @xmath457 and performing a nonideal measurement of @xmath458 by the povm @xmath459 with accuracy @xmath309 for the former group , and similarly we perform a nonideal measurement of @xmath460 by the povm @xmath461 with accuracy @xmath310 for the latter group ( see also example 3 in sec .
this measurement can asymptotically be described by the povm @xmath462 whose accuracy matrix is @xmath463 from eq.([accuracy2 ] ) in sec .
ii , the accuracy parameters in directions @xmath268 and @xmath269 are given by @xmath464 and thus @xmath465 we can therefore conclude that a simultaneous measurement has the advantage over this simple method in that the former can access the domain @xmath466 for @xmath467 , i.e. , domain q in fig.1 .
projection measurements can not always be implemented experimentally .
this raises the question of how accurately we can obtain information about observables from a given imperfect measurement scheme . to quantitatively characterize such measurement accuracy , we have introduced the @xmath2 accuracy matrix @xmath62 , with @xmath1 being the corresponding povm .
we have considered the accuracy matrix of the most general class of measurements of a qubit system : generalized simultaneous measurements including nonideal joint measurements and quantum - state tomography . from the outcomes of generalized simultaneous measurements ,
we can obtain information about more than one observable . in terms of the accuracy matrix ,
we have defined accuracy parameter @xmath130 and error parameter @xmath137 for a direction of @xmath25 corresponding to the observable @xmath29 .
these parameters satisfy @xmath468 and @xmath469 .
if @xmath470 , or equivalently @xmath471 , the measurement is equivalent to the projection measurement of @xmath29 . on the other hand ,
if @xmath472 , or equivalently @xmath473 , we can not obtain any information about the measured system by this measurement .
the accuracy matrix and accuracy parameters give us information about observables for which we can reconstruct the probability distribution from the measured distribution @xmath179 , where @xmath474 .
in fact , we can reconstruct the probability distribution of observable @xmath29 if and only if @xmath475 , or equivalently @xmath476 . in other words ,
the direction @xmath25 is a reconstructive direction if and only if @xmath131 , where the subspace @xmath128 of @xmath129 is spanned by the eigenvectors of @xmath62 corresponding to nonzero eigenvalues .
the main results of this paper are trade - off relations ( [ trade - off3 ] ) , ( [ trade - off4 ] ) , and ( [ trade - off5 ] ) between the accuracy parameters and the error parameters .
we can interpret them as the uncertainty relations between measurement errors in generalized simultaneous measurements ; the more information we obtain about an observable , the less information we can access about the other noncommuting observable . trade - off relation ( [ trade - off3 ] ) can also be interpreted as the uncertainty relation between the measurement error and back - action of measurement as formulated in inequality ( [ e - b - trade - off ] ) .
the new no - cloning inequality in ( [ no - cloning ] ) is derived from the trade - off relations . to derive this ,
we have introduced the cloning parameters @xmath406 and @xmath407 , where @xmath378 indicates the system to be cloned and @xmath379 indicates the blank reference system .
let @xmath5 be the pre - cloned state of system @xmath378 .
after a cloning operation , all the information about @xmath5 remains in system @xmath378 if and only if @xmath413 , and the information about @xmath5 is completely transferred to system @xmath379 if and only if @xmath477 . the impossibility of attaining @xmath426 implies the no - cloning theorem .
the condition of the equality in our no - cloning inequality ( [ no - cloning ] ) has yet to be understood .
we have also applied the trade - off relations to analyze the efficiency of quantum - state tomography .
the accuracy matrix of the standard qubit - state tomography or the minimal qubit tomography is given by @xmath478 with @xmath86 being the @xmath2 identity matrix , which implies that the efficiency of quantum - state tomography is optimal and symmetric .
we have pointed out a close relationship between the accuracy matrix and the fisher information .
we have also shown that the trade - off relations can be interpreted as being those concerning the accuracy of the maximum - likelihood estimators of the probability distributions of noncommuting observables .
while we focus on the spin-1/2 system in the present paper , many results can be generalized for higher - dimensional systems .
we conclude this paper by outlining such generalization . in the case of a @xmath479-dimensional system ( @xmath480 ) , the parametrization of the hermitian operator @xmath481 is given by @xmath482 where @xmath483 is a real number , @xmath72 is a @xmath484-dimensional real vector , and @xmath485 is the elements of the lie algebra of su(@xmath479 ) satisfying @xmath486 and @xmath487 with @xmath488 being the kronecker delta .
the necessary and sufficient condition for @xmath481 to be a positive operator is given by @xmath489 and @xmath490 ( @xmath491 ) , where @xmath492 is an @xmath493th - degree polynomial for @xmath72 @xcite .
the condition for @xmath494 is given by @xmath495 , which is equivalent to @xmath496 for @xmath497 , @xmath498 is given by @xmath499 where @xmath500 is defined as @xmath501 with @xmath502 .
we note that if @xmath481 is a rank-@xmath155 projection operator , then @xmath503 . however , the hermitian operator @xmath481 with @xmath504 is not necessarily a positive operator .
the accuracy matrix for a @xmath479-dimensional system assumes the same form as eq .
( [ am2 ] ) using parametrization ( [ d - parameterize ] ) . in this case
, @xmath62 is a @xmath484 square matrix .
moreover , we can define the accuracy parameter and the error parameter according to eqs .
( [ a - parameter ] ) and ( [ e - parameter ] ) , respectively . using condition ( [ positivity2 ] )
, we can derive trade - off relations ( [ trade - off3 ] ) and ( [ trade - off4 ] ) for a @xmath479-dimensional system . in this sense ,
the trade - off relations serve as universal uncertainty relations holding true for all finite - dimensional systems .
however , bounds of trade - off relations ( [ trade - off3 ] ) and ( [ trade - off4 ] ) would not necessarily be able to be reached for @xmath480 , because @xmath505 and @xmath506 are not sufficient for positivity of the povm .
moreover , while the accuracy parameter @xmath130 for @xmath507 characterizes the measurement accuracy of spin observables @xmath29 , the accuracy parameter @xmath130 for @xmath480 can not characterize the measurement accuracy of , for example , the spin-@xmath479 observable @xmath508 ; it only characterizes the accuracy of a rank-@xmath155 projection operator .
therefore the results of this paper based on @xmath130 can not be applied straightforwardly for @xmath480 .
a full investigation of this problem is underway .
we prove the case of @xmath509 .
for simplicity of notation , we define that @xmath510 , @xmath511 , @xmath512 , and @xmath513 .
the accuracy matrix is given by @xmath514 and the accuracy parameter in direction @xmath268 is @xmath515 the marginal povm @xmath208 is @xmath516 @xmath517 and the marginal accuracy matrix is @xmath518 where @xmath519 our objective is to show that @xmath520 . for simplicity
, we introduce the notation @xmath521 we can then write @xmath3 as @xmath522 where @xmath523 using eq .
( [ a1 ] ) , we can calculate the determinant of @xmath3 : @xmath524 ^ 2 .
\end{split}\ ] ] on the other hand , the cofactor matrix of @xmath3 is @xmath525 therefore @xmath526 the inverse matrix is given by @xmath527 ^ 2}.\ ] ] noting that @xmath528 ^ 2 \\ & = ( r_1 + r_2 ) \left ( \bigl [ \bm a_1 \cdot ( \bm a_2 \times \bm a_3 ) \bigr]^2 + \bigl [ \bm a_1 \cdot ( \bm a_2 \times \bm a_4 ) \bigr]^2 \right ) \end{split}\ ] ] and @xmath529 we obtain @xmath530 ^ 2 + \bigl [ \bm a_1 \cdot ( \bm a_2 \times \bm a_4 ) \bigr]^2 } { \det \chi ( \textbf{e})}. \end{split}\ ] ] similarly , we can show that @xmath531 ^ 2 + \bigl [ \bm a_3 \cdot ( \bm a_4 \times \bm a_2 ) \bigr]^2 } { \det \chi ( \textbf{e})}. \end{split}\ ] ] let us define @xmath532 ^ 2 + \bigl [ \bm a_1 \cdot ( \bm a_2 \times \bm a_4 ) \bigr]^2 } { \det \chi ( \textbf{e})},\ ] ] @xmath533 ^ 2 + \bigl [ \bm a_3 \cdot ( \bm a_4 \times \bm a_2 ) \bigr]^2 } { \det \chi ( \textbf{e } ) } , \ ] ] @xmath534 , and @xmath535 .
noting that @xmath536 and @xmath537 we obtain @xmath538 we can thus conclude @xmath539 therefore @xmath540 which is our objective .
this work was supported by a grant - in - aid for scientific research ( grant no .
17071005 ) and by a 21st century coe program at tokyo tech , `` nanometer - scale quantum physics '' , from the ministry of education , culture , sports , science and technology of japan . | we formulate the accuracy of a quantum measurement for a qubit ( spin-1/2 ) system in terms of a 3 by 3 matrix .
this matrix , which we refer to as the accuracy matrix , can be calculated from a positive operator - valued measure ( povm ) corresponding to the quantum measurement .
based on the accuracy matrix , we derive trade - off relations between the measurement accuracy of two or three noncommuting observables of a qubit system .
these trade - off relations offer a quantitative information - theoretic representation of bohr s principle of complementarity . they can be interpreted as the uncertainty relations between measurement errors in simultaneous measurements and also as the trade - off relations between the measurement error and back - action of the measurement .
a no - cloning inequality is derived from the trade - off relations .
furthermore , our formulation and the results obtained can be applied to analyze quantum - state tomography .
we also show that the accuracy matrix is closely related to the maximum - likelihood estimation and the fisher information matrix for a finite number of samples ; the accuracy matrix tells us how accurately we can estimate the probability distributions of observables of an unknown state by a finite number of quantum measurements . |
this paper generalizes the discrete memoryless channel ( dmc ) coding theorem for multiple codebooks of @xcite .
similar models have been analyzed recently in the context of random access communication , see @xcite , @xcite , @xcite ( typically for multiple access scenarios , not entered here ) , and unequal error protection @xcite .
it is assumed that the sender has a codebook library consisting of several codebooks .
each codebook consists of codewords of the same length and type . as a new feature compared to @xcite , here not only the type but also the length of the codewords may vary across codebooks , thus a model in between fixed length and variable length coding is addressed .
the sender uses the codebooks alternately , in any order he chooses .
the receiver is not aware of the codebook choices of the sender .
the different codeword lengths cause a certain asynchronism at the receiver , who should also estimate the boundaries of the codewords and avoid error propagation . a maximal mutual information based universal decoder is proposed to meet these challenges .
the main theorem shows that simultaneously for each codebook , the same error exponent can be achieved as the random coding error exponent for this codebook alone provided that it does not exceed the rate of this codebook .
the method of types , more exactly the subtypes technique of @xcite is used along with second order types ( see @xcite for a detailed explanation ) . in a related work @xcite
the problem of transmitting a discrete memoryless source over a dmc with variable - length codes is analyzed . using random coding argument along with maximum likelihood decoder @xcite
shows that variable - length source - channel codes achieve an error exponent equal to the random coding exponent of the channel evaluated at the source entropy .
a channel coding problem as in this work is not explicitly discussed in @xcite . in this paper , to keep the discussion simple , all codeword length ratios are assumed to be in @xmath0 . this assumption could be relaxed replacing @xmath1 by an arbitrary constant , then our main theorem could be used to give an alternative proof of this result of @xcite .
the advantage of this approach would be the universality of the decoder . due to the length - constraint ,
we omit the elaboration of this idea .
note that the topic of this paper is also connected to strong asynchronism ( @xcite,@xcite ) .
the notation @xmath2 denotes a quantity growing subexponentially as @xmath3 , that could be given explicitly . for some subexpontial sequences individual notations
are used and the parameters on which these sequences depend will be indicated in parantheses .
denote the set @xmath4 by @xmath5 $ ] .
notations like @xmath6 ( type class ) , @xmath7 ( family of n - types ) , @xmath8 , @xmath9 ( information measures under distribution v ) follow @xcite , @xcite and @xcite . denote @xmath10 the mutual information @xmath9 with @xmath11 equal to the joint type of @xmath12 .
if @xmath13 is a distribution on @xmath14 then @xmath15 , @xmath16 and @xmath17 denote the associated marginal resp .
conditional distributions .
the concatenation of an @xmath18-type @xmath19 and an @xmath20-type @xmath21 is the @xmath22-type @xmath23 with @xmath24 let @xmath25 , @xmath26 , @xmath27 be @xmath18 , @xmath20 , @xmath28-type respectively and let @xmath29 , @xmath30 , @xmath31 , @xmath32 then @xmath33 note that the nonegative quantities ( [ f - def1 ] ) and ( [ f - def2 ] ) are frequently called jensen - shannon divergence . for a given @xmath34 and @xmath35 , let @xmath36 denote the type for which @xmath37 .
as @xmath38 is always clear from the context it is not indicated in the notation .
finally we define the notion of second order type following @xcite . for a sequence @xmath39
let @xmath40 defined by @xmath41 let @xmath42 denote the second order type class @xmath43 .
we cite from @xcite that @xmath44
a dmc @xmath45 is given .
the transmitter has a codebook library with multiple constant composition codebooks .
the codewords length and type are fixed within codebooks , but can vary from codebook to codebook .
[ constantcomposition ] let positive integers @xmath46 , @xmath47 , @xmath48 with @xmath49 for all @xmath50 $ ] , distributions @xmath51 \}$ ] and rates @xmath52\}$ ] be given parameters .
a codebook library with the above parameters , denoted by @xmath53 , consists of constant composition codebooks @xmath54 such that @xmath55 with @xmath56 , @xmath57 $ ] , @xmath58 , @xmath59 $ ] . in the sequel , @xmath46
will be referred to as length - bound .
the transmitter continuously sends messages to the receiver through channel @xmath60 . before sending a message
, the transmitter arbitrarily chooses one codebook of the library .
this choice is not known to the receiver .
the performance of the following decoder is analyzed .
the output symbols of the channel are denoted by the infinite sequence @xmath61 .
assume that decoding related to symbols @xmath62 is already performed and now the position of @xmath63 is analyzed .
in the first stage of decoding the decoder tries to find indices @xmath64 which uniquely maximize @xmath65 if the decoder successfully finds a unique maximizer @xmath64 , the second stage of decoding starts . in this stage , if for all @xmath66 and @xmath67 the maximum of ( [ dekodolomuk ] ) is strictly larger than @xmath68 the decoder decodes @xmath69 as the codeword sent in the positions of @xmath70 and jumps to the position of @xmath71 , there the same but shifted procedure is performed .
otherwise the decoder goes to the position of @xmath72 without decoding at the position of @xmath63 .
[ stage2 ] . the output of this decoder can be described by a sequence of triplets @xmath73 , where the first coordinate is a codebook index , the second one is a message index related to this codebook , and the third refers to the starting position of the codeword .
this sequence is denoted by @xmath74 . are considered.,width=321 ] as mentioned above , the transmitter arbitrarily chooses codebooks .
his choices are described by an infinite codebook index sequence @xmath75 where @xmath76 $ ] . to each fixed @xmath77
, there corresponds a sequence @xmath78 , @xmath79 , of mutually independent random messages , where @xmath80 is uniformly distributed on @xmath81 $ ] .
let @xmath82 be the starting position of the @xmath83th message .
the average decoding error probability of the @xmath83th message is defined by @xmath84 where the probability is calculated over the random choice of the messages , and the channel transitions .
capital letters are used to indicate randomness . in practice ,
typically also the sequence @xmath77 is random : the messages at any time instant @xmath83 may be one of @xmath47 different kinds , with @xmath85 equiprobable messages of kind @xmath86 .
of course , this scenario is covered by our main theorem , as our bound does not depend on @xmath77 .
[ maintheorem ] for each @xmath46 let codebook library parameters as in definition [ constantcomposition ] be given with length - bound @xmath46 and with @xmath87 as @xmath3 . then there exist a sequence @xmath88 with @xmath89 and for each @xmath46 a codebook library @xmath53 with the given parameters such that for all infinite codebook index sequences @xmath77 and index @xmath83 @xmath90 where @xmath91 is the random coding error exponent function , i.e. , it is equal to @xmath92 to avoid additional technical difficulties it is assumed in the calculation below that all three consecutive messages are different ( see the terms below the sum in ( [ packingt2 ] ) and ( [ packingt3 ] ) ) .
the minimum in ( [ felsobecsles ] ) is present due to this assumption .
the next packing lemma provides the appropriate codebook library for theorem [ maintheorem ] .
note that the constructed codebook library works simultaneously for all infinite codebook index sequences @xmath77 .
the notations are explained on fig .
[ packing ] . , @xmath18 , @xmath20 , @xmath28 , @xmath25 , @xmath26 and @xmath27 .
] [ rc - packing - lemma ] let a sequence of codebook library parameters be given as in theorem [ maintheorem ] . then there exist a sequence @xmath93 with @xmath94 and for each @xmath46 a codebook - library @xmath53 with the given parameters such that the following bounds hold : _ t1 : _ for any @xmath95 $ ] , @xmath96 $ ] with @xmath97 , @xmath98 and for all joint types @xmath99 @xmath100
\triangleq \sum_{{\genfrac{}{}{0pt}{}{a \in [ n^{k_1 } ] , d \in [ n^{\hat{k}}]}{d \ne a \textnormal { if } k_1=\hat{k}}}}\hspace{-11pt}\mathds 1_{t1,q}^{v_{1}}({\mathbf x}_a^{k_1 } , { \mathbf x}_{d}^{\hat{k } } ) \label{packingt1}\\ & \le \delta^{'}_n \cdot 2^{-n_1\operatorname{i}_{v_1}(x \wedge \hat x)+l^{k_1}r^{k_1}+l^{\hat{k}}r^{\hat{k } } } \notag\end{aligned}\ ] ] here @xmath101 , the indicator function @xmath102 equals @xmath103 if filling the pattern t1 of fig .
[ packing ] by @xmath104 and @xmath105 the joint type of the indicated subsequences equals @xmath25 .
furthermore , @xmath106 denotes the set of all joint type pairs that may occur in this way .
_ t2 : _ for any @xmath95 $ ] , @xmath107 $ ] , @xmath96 $ ] , @xmath108 $ ] with @xmath109 and for all @xmath110 : @xmath111 \label{packingt2 } \\ & \triangleq \sum_{{\genfrac{}{}{0pt}{}{a \in [ n^{k_1}],b \in [ n^{k_2 } ] , d \in [ n^{\hat{k}}]}{a\ne b \text { if } k_1=k_2}}}\mathds 1_{t2,q}^{v_{1},v_{2}}({\mathbf x}_a^{k_1 } , { \mathbf x}_b^{k_2 } , { \mathbf x}_{d}^{\hat{k } } ) \notag\\ & \le \delta^{'}_n \cdot 2^{-\sum\limits_{i=1}^{2 } n_i\operatorname{i}_{v_i}(x \wedge \hat x)-l^{\hat{k}}f(v_1^{\hat{x}},v_2^{\hat{x}})+\sum\limits_{i=1}^{2}l^{k_i}r^{k_i } + l^{\hat{k}}r^{\hat{k } } } \notag\end{aligned}\ ] ] here @xmath112 , @xmath113 , the indicator function @xmath114 equals @xmath103 if filling the pattern t2 of fig
. [ packing ] by @xmath104 , @xmath115 and @xmath105 the joint types of the indicated subsequences equal @xmath25 and @xmath26 respectively .
furthermore , @xmath116 denotes the set of all joint type triples that may occur in this way . _
t3 : _ for any @xmath95 $ ] , @xmath107 $ ] , @xmath117 $ ] , @xmath118 $ ] , @xmath108 $ ] with @xmath119 and for all @xmath120 : @xmath121 \label{packingt3 } \\ & \triangleq \sum_{{\genfrac{}{}{0pt}{}{a \in [ n^{k_1}],b \in [ n^{k_2 } ] , c \in [ n^{k_3 } ] , d \in [ n^{\hat{k}}]}{a\ne b \text { if } k_1=k_2 , a \ne c \text { if } k_1=k_3 , b \ne c \text { if } k_2=k_3}}}\hspace{-4pt}\mathds 1_{t3,q}^{v_{1},v_{2},v_3}({\mathbf x}_a^{k_1 } , { \mathbf x}_b^{k_2},{\mathbf x}_c^{k_3 } , { \mathbf x}_{d}^{\hat{k } } ) \notag\\ & \le \delta^{'}_n \cdot 2^{-\sum\limits_{i=1}^{3 } n_i\operatorname{i}_{v_i}(x \wedge \hat x)-l^{\hat{k}}f(v_1^{\hat{x}},v_2^{\hat{x}},v_3^{\hat{x}})+\sum\limits_{i=1}^{3}l^{k_i}r^{k_i } + l^{\hat{k}}r^{\hat{k } } } \notag\end{aligned}\ ] ] here @xmath112 , @xmath122 , @xmath123 , the indicator function @xmath124 equals @xmath103 if filling the pattern t3 of fig . [ packing ] by @xmath104 , @xmath115 , @xmath125 and @xmath105 the joint types of the indicated subsequences equal @xmath25 , @xmath26 and @xmath27 respectively . furthermore , @xmath126 denotes the set of all joint types that may occur in this way . choose the codebook library @xmath53 at random , i. e. , for all @xmath50 $ ] the codewords of @xmath127 are chosen independently and uniformly from @xmath128 .
denote the exponents in upper - bounds ( [ packingt1 ] ) , ( [ packingt2 ] ) and ( [ packingt3 ] ) by @xmath129 $ ] , @xmath130 $ ] and @xmath131 $ ] respectively .
we first claim that under this random selection the expected value of the expressions @xmath1002^{-e^{k_1,\hat{k}}_{t1,q}[v_{1 } ] } \label{packingt1rendez}\\ & k^{k_1 , k_2 , \hat{k}}_{t2,q}[v_{1},v_{2 } ] 2^{-e^{k_1 , k_2 , \hat{k}}_{t2,q}[v_{1},v_{2 } } ] \label{packingt2rendez}\\ & k^{k_1 , k_2,k_3 , \hat{k}}_{t3,q}[v_{1},v_{2},v_3]2^{-e^{k_1 , k_2,k_3 , \hat{k}}_{t3,q}[v_{1},v_{2},v_3 ] } \label{packingt3rendez}\end{aligned}\ ] ] are bounded above by a subexponential function of @xmath46 that depends only on the alphabet size @xmath132 .
we establish this claim for ( [ packingt2rendez ] ) , the other cases are similar or easier .
assume that @xmath95 $ ] , @xmath107 $ ] , @xmath96 $ ] and @xmath108 $ ] fulfills the requirements of case t2 .
first let @xmath133 be indices with @xmath134 . in this case @xmath135
then with some algebric rearrangement we get : @xmath136 now let @xmath133 be indices with @xmath137 and @xmath138 ( the case @xmath139 can be treated similarly ) .
then @xmath140 in this case upper bounding @xmath141 is a bit technical .
we have to separately investigate two cases : @xmath142 , @xmath143 . for didactic purposes
the second case is divided further into two subcases : @xmath144 , @xmath145 . _
( a ) , @xmath146 , @xmath142 : _ in this case we divide the block corresponding to @xmath26 into consecutive subblocks of length @xmath147 ( the length of the last block , denoted by @xmath148 , is at most @xmath147 ) .
we perform the counting from left to right ( see part ( a ) of fig .
[ caset2i ] ) : @xmath149 here in ( [ markovi1 ] ) the sum is over subtype sequences corresponding to the division in part ( a ) of fig . [ caset2i ] .
these @xmath150 have convex combination @xmath26 . in ( [ markovi2 ] )
we used the concavity of the entropy and the fact that the number of subtype sequences in the sum is subexponential . substituting ( [ markovi2 ] ) into ( [ t2becsles4 ] ) and proceeding similarly as in ( [ t2becslesreszeleje])-([t2becslesreszkozepe ] ) we get ( [ t2becslesvegso ] ) .
_ ( b ) , @xmath146 , @xmath144 : _ in this case we can upper - bound @xmath141 by @xmath151 here the term @xmath152 comes from ( [ markovrabecsles ] ) . substituting ( [ markovi3 ] ) into ( [ t2becsles4 ] ) and proceeding similarly as in ( [ t2becslesreszeleje])-([t2becslesreszkozepe ] ) we get ( [ t2becslesvegso ] ) . _
( c ) , @xmath146 , @xmath153 : _ the length of the block corresponding to @xmath26 is @xmath20 .
we divide this block into @xmath18 `` subblocks '' consisting of non - consecutive elements : the first corresponds to indices @xmath154 , the second corresponds to indices @xmath155 , @xmath156 , the last corresponds to indices @xmath157 ( see part ( c ) of fig .
[ caset2i ] ) .
let @xmath158 , , @xmath148 denote the lengths of these non - consecutive subblocks ( note that @xmath159 for all @xmath86,@xmath83 ) .
then : @xmath160 @xmath161 here in ( [ markovi4 ] ) again ( [ markovrabecsles ] ) is used and the sum is over type sequences corresponding to the division in part ( c ) of fig .
[ caset2i ] whose convex combination equals @xmath26 . in ( [ markovi5 ] )
again the concavity of the entropy and the fact that the number of subtype sequences in the sum is subexponential in @xmath46 are used . substituting ( [ markovi5 ] ) into ( [ t2becsles4 ] ) and proceeding similarly as in ( [ t2becslesreszeleje])-([t2becslesreszkozepe ] ) we get ( [ t2becslesvegso ] ) . as ( [ t2becslesvegso ] )
is fulfilled in all cases , taking into account the definition of @xmath162 , @xmath163 and @xmath164 we proved the claim for ( [ packingt2rendez ] ) .
next , denote by @xmath165 the sum of form ( [ packingt1rendez])-([packingt3rendez ] ) for all possible @xmath166 , @xmath167 , @xmath168 , @xmath169 , @xmath170 and joint types @xmath25 , @xmath26 , @xmath27 .
as @xmath47 grow at most subexponentially and the number of types is polynomial it follows that @xmath171 for suitable @xmath93 with @xmath94 . hence , there exists a realization of the codebook library with @xmath172 . in the proof of lemma [ rc - packing - lemma ] the separate investigation of cases @xmath173 and
@xmath174 ensures that the number of terms of the sums in both ( [ markovi1 ] ) and ( [ markovi4 ] ) is subexponential .
we prove that the codebook library @xmath53 provided by lemma [ rc - packing - lemma ] with the decoder specified in section [ modell ] fulfills ( [ felsobecsles ] ) .
let an infinite codebook sequence @xmath77 and an index @xmath83 be given .
let @xmath175 denote the output symbols affected by the @xmath83-th message .
for @xmath176 $ ] and @xmath177 , let @xmath178 be following event @xmath179 \text { } ( c\ne m_j \text { if } k = h_j \text { and } d = s_j ) , \end{array}\hspace{-3pt}\right\ } \label{errorpattern}\ ] ] and let @xmath180 denote its probability .
then @xmath181 for @xmath182 and @xmath183 by standard argument @xmath184 to prove the theorem it is enough to show that ( [ biz1 ] ) holds for all @xmath176 $ ] and @xmath177 ( the number of such @xmath185 pairs is subexponential ) .
fix such a @xmath185 pair . without any loss in generality
assume that @xmath186 and that both the @xmath187th and @xmath188th messages affect outputs @xmath189 ( see fig .
[ hibabecsleskep ] ) .
the analyzes of other cases are similar .
assume further that @xmath190 .
let @xmath191 .
then @xmath180 can be upper - bounded by : @xmath192 , b \in [ n^{h_{j-1}}]}{c \in [ n^{h_{j } } ] } } } \hspace{-28pt } { \textnormal{pr}}\ { \mathcal{e}_j^{\mathbf{h}}(k , d ) | m_{j-2}=a , m_{j-1}=b , m_j = c\ } \label{teljesval}\ ] ] let @xmath193 , @xmath194 , @xmath195 , @xmath196 and @xmath197 ( lemma [ rc - packing - lemma ] will be used with these choices ) and @xmath198 be equal to @xmath199 let @xmath200 be the length of the analyzed output window ( see again fig . [ hibabecsleskep ] )
. then ( [ teljesval ] ) can be further upper - bounded by @xmath201 , b \in [ n^{h_{j-1}}]}{c \in [ n^{h_{j } } ] } } } \hspace{-1pt } \prod_{i=1}^{4 } \hspace{-1pt } \big ( 2^{-n_i \operatorname{d}(v_i^{y|x}||w|v_i^{x } ) } \notag\\ & \cdot 2^{-n_i \operatorname{h}_{v_i}(y|x ) } \big ) \cdot \big| \{\mathbf{y } \in \mathcal{y}^{l}:\mathds{1}_{f4}^{\mathbf{v } } \ { { \mathbf x}_a^{k_1},{\mathbf x}_b^{k_2},{\mathbf x}_c^{k_3 } , { \mathbf x}_d^{\hat{k}},{\mathbf y}\ } \notag\\ & = 1 \text { for some } d \in [ n^{\hat{k } } ] \big| , \label{hibabindikator}\end{aligned}\ ] ] where the indicator function @xmath202 equals @xmath103 if filling the pattern of fig .
[ hibabecsleskep ] by @xmath104 , @xmath115 , @xmath125 , @xmath203 and @xmath204 the joint types of the indicated subsequences equal @xmath25 , @xmath26 , @xmath27 and @xmath205 respectively .
we can upper - bound the set size in ( [ hibabindikator ] ) two different ways .
the first bound is @xmath206 .
let @xmath207 , @xmath208 .
the second bound is : @xmath209 substituting these bounds into ( [ hibabindikator ] ) and using ( [ packingt3 ] ) we get that @xmath180 can be upper - bounded by : @xmath210 the term inside @xmath211 can be bounded from below by @xmath212 by convexity ( [ lassanuccso ] ) can be further lower - bounded by @xmath213 substituting ( [ lassanuccso2 ] ) into ( [ hibabindikator2 ] ) and using ( [ hibatipus ] ) we get @xmath214 hence , choosing @xmath215 , continuity argument and the convexity of the divergence prove @xmath216 if @xmath217 or @xmath218 then in ( [ teljesval ] ) we have to separately investigate the cases when there is coincidence between @xmath219 , @xmath220 and @xmath221 .
the minimum in ( [ felsobecsles ] ) is present due to this cases .
we would like to thank prof .
imre csiszr for his help and advice .
we also thank the support of the hungarian national foundation for scientific research grant otka k105840 and the mta - bme stochastics research group .
i. csiszr , j. krner , _ information theory , coding theorems for discrete memoryless systems , @xmath222 edition _ , cambridge university press , 2011 .
i. csiszr , `` joint source - channel error exponent , '' _ prob .
contr . & info .
vol . 9 , no . 5 , pp.315323 , 1980 . i. csiszr , `` the method of types , '' _ ieee transactions on information theory _ , vol .
2505 - 2523 , 1998 . l. farkas and t. ki , `` random access and source - channel coding error exponents for multiple access channels , ''
_ ieee transactions on information theory _ , vol .
61 , pp . 3029 - 3040 , jun . 2015 | csiszr s channel coding theorem for multiple codebooks is generalized allowing the codeword lenghts differ across codebooks .
it is shown that simultaneously for each codebook an error exponent can be achieved that equals the random coding exponent for this codebook alone , with possible exception of codebooks with small rates
. multiple codebook , error exponent , variable length , joint source - channel coding , random access |
* shear - induced structures versus flow instabilities * sup .
movies 1 , 2 and 3 corresponding respectively to fig .
2(a - b ) , fig .
3(a - b ) and fig .
3(c - d ) of the paper are available on request from the corresponding author .
see the captions of the corresponding figures for details .
the movies also display some velocity profiles in the bottom left quadrant .
the square symbols correspond to examples of velocity profiles obtained at three different heights @xmath38 , whereas the circles correspond to the velocity profile obtained by averaging along @xmath38 .
the line represents the purely azimuthal couette flow profile expected for no - slip boundary conditions .
c , at @xmath63 s@xmath7 , below the onsets of both tvf and sis formation .
the flow remains perfectly laminar with no noticeable wall slip .
( a ) global shear stress @xmath47 measured by the rheometer ( in black ) and dimensionless slip velocity @xmath48 ( in red ) .
( b ) spatiotemporal diagram ( center ) of the velocity @xmath49 at @xmath106 mm from the rotor .
the dotted lines show the times @xmath80 and @xmath107 s corresponding to the velocity maps @xmath52 shown on the left and right respectively . the color scale is linear and goes from 0 to @xmath53 .
[ supfig],width=302 ] | the taylor - couette flow of a dilute micellar system known to generate shear - induced structures is investigated through simultaneous rheometry and ultrasonic imaging .
we show that flow instabilities must be taken into account since both the reynolds number and the weissenberg number may be large . before nucleation of shear - induced structures ,
the flow can be inertially unstable , but once shear - induced structures are nucleated the kinematics of the flow become chaotic , in a pattern reminiscent of the inertio - elastic turbulence known in dilute polymer solutions .
we outline a general framework for the interplay between flow instabilities and flow - induced structures . beside their tremendous industrial importance in applications such as detergence , oil recovery , or drag reduction ,
micellar systems have long been used as a model system for rheological research @xcite . in particular
some surfactant systems are known to form rodlike micelles , which can grow to become wormlike and entangled when the concentration in surfactant and/or salt increases , or when the temperature decreases @xcite .
the dilute regime of rodlike micelles has been studied in the context of highly elastic shear - induced structures ( sis ) and associated shear - thickening , whereas the so - called `` semi - dilute '' and `` concentrated '' regimes have been studied in the context of shear - banding , which is an extreme form of shear - thinning where the velocity gradient field becomes inhomogeneous even in simple shear @xcite . on the one hand ,
shear - banding micellar systems have a high viscosity , such that inertial flow instabilities can generally be neglected .
nevertheless semi - dilute and concentrated solutions have high normal stresses , such that the weissenberg number is large and purely elastic instabilities can develop @xcite , similar to the case of polymeric boger fluids @xcite .
on the other hand , the possibility for flow instabilities has never been considered thoroughly in dilute , shear - thickening systems even though these systems have both large reynolds ( @xmath0 ) and weissenberg ( @xmath1 ) numbers @xcite .
the aim of the present letter is to fill this gap by showing that both inertial and elastic instabilities are constantly encountered in the taylor - couette flow of a well - known dilute micellar system . vs shear rate @xmath2 at @xmath3 , 25 , and
30@xmath4c ( filled symbols , from top left to bottom right ) for ctat at 0.16% wt . seeded with 1% wt .
polyamide spheres ( see text ) .
open symbols correspond to a ctat sample free of contrast agents at @xmath5c .
the addition of polyamide spheres increases the viscosity by about 15% but does not change the overall behavior .
data points correspond to averages over the last 5 s of shear rate steps of total duration 15 s. vertical dashed lines indicate the onset of shear - thickening @xmath6 , 35 , and 80 s@xmath7 respectively inferred at @xmath3 , 25 , and 30@xmath4c from start - up experiments of typical duration 500 s .
[ fig1],width=264 ] the sample under study is made of 0.16% wt .
hexadecyltrimethylammonium p - toluenesulfonate ( ctat ) in water . for this system , which has become a benchmark example of dilute micellar fluids @xcite
, the overlap concentration is estimated at @xmath8 wt .
% and shear - thickening is found over the range 0.05 - 0.8% wt .
@xcite .
figure [ fig1 ] shows the viscosity @xmath9 of our sample as a function of the applied shear rate @xmath2 for three different temperatures @xmath10 , exhibiting the typical behavior of shear - thickening , dilute surfactant systems @xcite : a zero - shear viscosity close to the viscosity of water ; then a jump in @xmath9 at a characteristic shear rate @xmath11 that increases with @xmath10 ; finally , a shear - thinning viscosity branch at high shear rates . this behavior was first explained by postulating the formation of sis @xcite : above @xmath11 , micelles grow in length and undergo a transition from rodlike to wormlike aggregates .
this microscopic scenario was confirmed through neutron and light scattering experiments @xcite .
the shear - induced state can then be shear - thinning due to the increasing alignment of the worms . in the present experiments the fluid
is sheared in a taylor - couette ( tc ) device adapted to a rheometer ( arg2 , ta instruments ) and with dimensions ( height @xmath12 mm , gap @xmath13 mm , radius of the inner rotating cylinder @xmath14 mm ) that ensure the `` small gap approximation '' , @xmath15 , without any strong end effect at the top and bottom boundaries ( @xmath16 ) , so that the laminar base flow is a simple shear with @xmath17 , where @xmath18 and @xmath19 are the rotor linear and angular velocities respectively .
we recall that in a tc device with inner rotation , taylor showed that a newtonian fluid becomes unstable and develops a secondary flow made of toroidal counter - rotating vortices @xcite .
larson , shaqfeh and muller discovered that non - newtonian fluids can develop a similar vortex flow not driven by inertia , but driven by elasticity @xcite . in both cases
the instability develops when the taylor number @xmath20 exceeds a given threshold @xmath21 . in the purely inertial case @xmath22 , with @xmath23 and @xmath24 @xcite , while in the purely elastic case @xmath25 , with @xmath26 , where @xmath27 is a characteristic polymeric relaxation time @xcite instead of the viscous dissipation time @xmath28 , and @xmath29 kg.m@xmath30 is the fluid density . in the latter case
, the value of @xmath21 depends on the constitutive relation of the non - newtonian fluid , e.g. , linear stability theory predicts @xmath31 for the upper - convected maxwell model @xcite .
more generally the balance between elasticity and inertia can be estimated by the elasticity number @xmath32 .
although predicting the onset of instability for arbitrary elasticity is not yet possible , a range of inertio - elastic instabilities is expected but remains mostly unexplored @xcite . since we expect secondary flows to emerge , our geometry is equipped with a recently developed two - dimensional ultrasonic velocimetry technique that allows the simultaneous measurement of 128 velocity profiles over 30 mm along the vertical direction in the tc geometry @xcite .
we use ultrafast plane wave imaging and cross - correlation of successive images @xcite to infer velocity maps from the echoes backscattered by acoustic contrast agents seeding the fluid , namely 1% wt .
polyamide spheres ( arkema orgasol 2002 es 3 nat 3 , mean diameter 30 @xmath33 m , relative density 1.03 ) , which do not affect significantly the rheological behavior of our solution ( see fig .
[ fig1 ] ) .
this technique yields the component @xmath34 of the velocity vector , @xmath35 in cylindrical coordinates , projected along the acoustic propagation axis @xmath36 as a function of the radial distance @xmath37 to the rotor and of the vertical position @xmath38 with a temporal resolution down to 50 @xmath33s @xcite .
the acoustic axis @xmath36 is horizontal and makes an angle @xmath39 with the normal to the outer cylinder so that @xmath40 .
finally we define the measured velocity as @xmath41 , which coincides with the azimuthal velocity @xmath42 in the case of a purely azimuthal flow @xmath43 .
more generally @xmath44 combines contributions from both azimuthal and radial velocity components .
nevertheless , close to instability onset , secondary flows are usually much weaker than the main flow , such that @xmath45 , as checked recently on an inertially unstable newtonian fluid and on an elastically unstable shear - banding fluid @xcite .
c for @xmath46 s@xmath7 ( see also supp .
movie 1 ) .
the flow is first inertially unstable and then inertio - elastically unstable due to sis formation .
( a ) global shear stress @xmath47 measured by the rheometer ( in black ) and dimensionless slip velocity @xmath48 ( in red ) @xcite .
( b ) spatiotemporal diagram ( center ) of the velocity @xmath49 at @xmath50 mm from the rotor .
the dotted lines show the times @xmath51 and 22.5 s corresponding to the velocity maps @xmath52 shown on the left and right respectively .
the color scale is linear and goes from 0 to @xmath53 for the velocity maps and from @xmath54 to @xmath53 for the diagram . the vertical axis @xmath38 is oriented downwards with @xmath55 being taken at about 6 mm from the top of the tc cell .
[ fig2],width=257 ] figure [ fig2 ] reports the start - up flow of ctat at @xmath56c for @xmath46 s@xmath7 ( see also supp .
movie 1 ) . at very short times
, a laminar boundary layer extends from the inner cylinder to the outer cylinder . a taylor vortex flow ( tvf ) then develops for @xmath57 s , deforming the main flow which becomes periodic along @xmath38 : slow moving fluid is brought inward in regions of centripetal radial flow and fast moving fluid is pushed outward in regions of centrifugal radial flow .
this initial sequence of events would be exactly similar if the fluid was pure water @xcite .
even though such a tvf had never been reported before for dilute micelles , it should not be surprising since , assuming the fluid to be influenced only by inertia in this initial sequence , we have @xmath58 for @xmath46 s@xmath7 . in computing @xmath28
, we have used the dynamic viscosity relevant to the short - time behavior , i.e. the zero - shear viscosity @xmath59 mpa.s . as shown in fig .
[ fig2](a ) the onset of tvf at @xmath60 s corresponds to a slight increase of the shear stress @xmath61 ( after an initial spike due to the feedback with the rheometer inertia ) .
this first stress increase is simply due to the formation of vortices breaking the viscometric assumption @xcite .
in contrast , for @xmath62 s , the stress ( or alternatively the viscosity ) climbs up much more dramatically . since @xmath46 s@xmath7 falls into the shear - thickening range for @xmath56c ( see fig . [ fig1 ] ) , this next sequence of events can clearly be attributed to slow sis formation .
meanwhile the structure of the vortex flow is disrupted .
the formerly well - defined wavelength and amplitude of the main flow shown in fig .
[ fig2](b , left ) are lost and the flow becomes chaotic - like in fig .
[ fig2](b , right ) .
this latter state is reminiscent of the inertio - elastic turbulent state called `` disordered oscillations '' @xcite or `` elastically dominated turbulence '' @xcite . while iso - velocity lines in the initial state can be approximated by harmonic functions , the new secondary flows associated with the sis deform the main flow intermittently .
the state in fig .
[ fig2](b , right ) is representative of the asymptotic turbulent flow .
it continuously generates large fluctuations in the viscosity and stress , which have been reported before but never accounted for in terms of elastic turbulence @xcite .
note also that the turbulent nature of the flow can locally and transiently generate plug flow profiles that may explain some earlier 1d velocity measurements @xcite . at lower shear rates ,
e.g. @xmath63 s@xmath64 and @xmath65 , the flow is below both thresholds for tvf and sis formation and remains purely azimuthal as shown in supp .
fig . 1 .
as it turned out for @xmath56c , the critical shear rate @xmath11 for sis formation and the critical shear rate @xmath66 for the onset of tvf are about the same value @xmath67 s@xmath7 in our tc geometry ( @xmath68 ) . in order to separate the inertial tvf and the turbulence associated with sis more readily , we reproduced similar shear start - up protocols at two other temperatures , @xmath3 and @xmath69c , as shown in fig .
[ fig3 ] .
increasing the temperature slightly lowers the zero - shear viscosity ( see fig .
[ fig1 ] ) so that @xmath70 only decreases from 45 s@xmath7 at @xmath71c to 33 s@xmath7 at @xmath69c .
in contrast , the same temperature change has a much stronger impact on @xmath11 .
as reported extensively in the literature @xcite , lowering the temperature leads to easier sis formation hence shifting @xmath11 to lower values .
figure [ fig1 ] indicates @xmath6 and 80 s@xmath7 at @xmath71c and @xmath69c respectively .
therefore at the highest temperature , we should be able to observe tvf without sis , whereas sis without tvf may be expected at the lowest temperature .
this scenario is fully confirmed in fig .
[ fig3](a , b ) and ( c , d ) where spatiotemporal dynamics are compared for shear rates such that @xmath72 and @xmath73 at @xmath74 and @xmath71c respectively . for ( a , b ) @xmath5c and @xmath75 s@xmath7 showing only inertial instability and ( c , d ) @xmath76c and @xmath63 s@xmath7 where the flow slowly develops sis and the associated inertio - elastic turbulence without initial tvf ( see also supp .
movies 2 and 3 respectively ) . in ( b ) @xmath77 mm and @xmath78 ( left ) and 175 s ( right ) . in ( d )
@xmath79 mm and @xmath80 ( left ) and 225 s ( right ) .
the color scale goes from 0 to @xmath53 for all velocity maps and from @xmath81 ( resp .
@xmath82 ) to @xmath53 for the diagram in ( b ) [ resp .
( d ) ] . [ fig3],width=257 ]
the fact that sis and tvf can occur separately is an indication that these two phenomena are not consequences of one another .
sis do not need tvf to nucleate , which suggests that the out - of - equilibrium growth of the worms is driven by the base shear flow , as usually postulated @xcite .
early velocity measurements at a single height @xmath38 reported that sis first form at the inner wall and generate significant slip on this wall @xcite .
the dimensionless slip velocities @xmath83 @xcite shown in figs .
[ fig2](a ) and [ fig3](c ) confirm that the onset of wall slip is concomitant with sis formation , whereas no noticeable wall slip is reported in the presence of tvf alone [ see fig .
[ fig3](a ) ] .
of course tvf does not need sis since it can occur even in simple molecular fluids .
moreover the spatiotemporal structure of the flow in figs .
[ fig2](b , right ) and [ fig3](d , right ) are very similar so that tvf appears to have negligible impact on the asymptotic turbulent flow after sis nucleation .
it seems that before the nucleation of sis inertia is dominating the instability of the flow , whereas once sis are formed elasticity dominates . in a shear - thickening dilute surfactant solution ,
both @xmath28 and @xmath27 depend on @xmath2 and @xmath84 so that @xmath0 , @xmath1 , and @xmath85 also depend on @xmath2 and @xmath84 . to illustrate this
, we evaluate a bulk - averaged value of @xmath85 in the case of fig .
[ fig2 ] ( @xmath56c and @xmath46 s@xmath7 ) . assuming the fluid density @xmath86 to remain constant during sis formation
, we first estimate the reynolds number by @xmath87 , where @xmath88 is the `` true '' shear rate corrected for wall slip and @xmath89 the corresponding `` true '' viscosity .
this yields a decrease from @xmath90 before sis formation to @xmath91 in the final state .
estimating the characteristic viscoelastic time @xmath27 of the sis is more challenging .
we take the longest time of the stress relaxation after flow cessation either before or after sis formation , which yields @xmath92 before and @xmath93 after sis formation in accordance with the literature @xcite . thus in the early stages of the dynamics
@xmath94 , i.e. inertia dominates , whereas after sis formation @xmath95 , validating the dominance of elasticity . in some sense , the dynamics of fig .
[ fig2](b ) can be seen as the superposition of the dynamics of fig .
[ fig3](b ) and ( d ) @xcite . ) .
( 2 ) inertial instability ( tvf ) without sis ( @xmath96 ) . ( 3 ) inertio - elastic instability of the sis without initial tvf ( @xmath97 ) . ( 4 )
inertio - elastic instability of the sis after initial tvf ( @xmath98 ) .
larger empty symbols correspond to the experiments in figs .
[ fig2 ] and [ fig3 ] .
stars show the transitions between the various regimes .
solid lines are guides to the eye .
[ fig4],width=302 ] the stability diagram of fig .
[ fig4 ] summarizes the interplay between inertial instability ( tvf ) and elastic instability of the sis based on an extensive data set .
we believe that such a diagram should be systematically sought for in complex fluids subject to flow instabilities in order to sort out the influences of flow - induced microscopic and macroscopic phenomena . here
we find that sis are always unstable and systematically lead to elastic - like turbulence . whether or not this is a feature common to all elastic sis stands out as an open issue .
another fundamental question is the possibility to predict the various boundaries in fig .
[ fig4 ] using phenomenological approaches or even microscopic models . deriving an instability criterion based on the elasticity number @xmath85 and accounting for both @xmath0 and @xmath1 thus appears as a first crucial theoretical step .
m .- a.f . and c.p .
contributed equally to this work .
the authors thank s. lerouge for providing the ctat sample and for motivating this study .
this work was funded by the institut universitaire de france and by the european research council under the european union s seventh framework programme ( fp7/2007 - 2013 ) / erc grant agreement no .
258803 .
37 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , * * , ( ) . , in _ _
( , ) , pp . . , * * , ( ) . ,
* * , ( ) . ,
* * , ( ) . , * * , ( ) . , * * , ( ) . , , , * * , ( ) . ,
* * , ( ) . , * * , ( ) . , , , , * * , ( ) . , * * , ( ) . , * * , ( ) . ,
* * , ( ) . , * * , ( ) . , , , * * , ( ) . , * * , ( ) . , , , , , , * * , ( ) . , , , * * , ( ) .
, , , ( ) . , * * , ( ) . , * * , ( ) . ,
* * , ( ) .
, , , * * , ( ) . , , , * * , ( ) .
, , , , * * , ( ) .
linear fits of the velocity profiles @xmath99 over about 300 @xmath33 m in the radial direction @xmath37 are used to estimate the local shear rates close to each wall and extrapolate the fluid velocities @xmath100 and @xmath101 . the dimensionless total slip velocity is then defined as @xmath102 / \omega r_i$ ] where the average is taken over the 128 simultaneous measurements along the vertical direction @xmath38 .
error bars in fig .
[ fig2](a ) show the standard deviation along @xmath38 .
the sis formation time in fig .
[ fig3](d ) is much longer than in fig .
[ fig2](b ) .
this is due to the fact that in the first case @xmath103 and it is well known that the formation time of sis goes like @xmath104 @xcite .
along the same lines apparent wall slip ( @xmath105% ) is visible after onset of tvf in fig .
[ fig2](a ) while it remains negligible in fig .
[ fig3](a ) . this can be attributed to a larger distance to @xmath70 in fig .
[ fig2 ] and to larger effects of radial velocity components on @xmath44 and @xmath83 . |
self - avoiding walks ( saws ) on regular lattices provide a successful description of the universal configurational properties of polymer chains in good solvent @xcite .
in particular , the average square end - to - end distance @xmath0 of saws with @xmath1 steps obeys the scaling law @xmath2 where the universal exponent @xmath3 only depends on the space dimension @xmath4 . for regular lattices ,
its value is well established ( @xmath5 for @xmath6 ) .
new challenges have been raised recently in studies of biopolymers in natural cellular environments , which are usually very crowded by many other biochemical species occupying a large fraction of the total volume @xcite . in a minimalistic description ,
we assume here this `` volume exclusion '' to be random and frozen , i.e. , quenched , and model the available space for the saws by site percolation clusters on hypercubic lattices at the percolation threshold @xmath7 in @xmath8 .
note that a percolation cluster itself is a fractal object with fractal dimension @xmath9 dependent on the space dimension @xmath4 .
the scaling law ( [ scaling ] ) still holds , but with an exponent @xmath10 @xcite .
when studying physical processes on complicated fractal objects , one often encounters the situation of coexistence of a family of singularities , each associated with a set of different fractal dimensions @xcite . in these problems , an infinite set of critical exponents
is needed to characterize the different moments of the distribution of observables , which scale independently .
these peculiarities are usually referred to as multifractality @xcite .
multifractal properties arise in many different contexts , for example in studies of turbulence in chaotic dynamical systems and strange attractors @xcite , human heartbeat dynamics @xcite , anderson localization transition @xcite , etc .
although the behavior of saws on percolative lattices served as a subject of numerous numerical and analytical studies since the early 80th , not enough attention has been paid to clarifying the multifractality of the problem .
it was only recently proven in field - theoretical studies @xcite that the exponent @xmath11 alone is not sufficient to completely describe the peculiarities of saws on percolation clusters .
instead , a whole spectrum @xmath12 of multifractal exponents emerges @xcite : @xmath13 with @xmath14 .
note that putting @xmath15 in ( [ spectr ] ) , we restore an estimate for the dimension @xmath16 of the underlying backbone of percolation clusters via @xmath17 , and @xmath18 gives us the exponent @xmath11 , governing the scaling law for the averaged end - to - end distance of saws on the backbone of percolation clusters . in the present paper , we report a careful computer simulation study of saws on percolation clusters .
we consider site percolation on regular lattices of edge lengths up to @xmath19 in dimensions @xmath20 , respectively .
each site of the lattice is occupied randomly with probability @xmath21 and empty otherwise . to extract the backbone of a percolation cluster ,
we apply the algorithm proposed in ref .
@xcite .
we construct a saw on percolation clusters , applying the pruned - enriched chain - growth algorithm @xcite .
we let a trajectory of saw grow step by step , until it reaches some prescribed distance ( say @xmath22 ) from the starting point .
then , the algorithm is stopped , and a new saw grows from the same starting point .
in such a way , we are interested in constructing different possible trajectories with fixed end - to - end distance . for each lattice size @xmath23 ,
we change @xmath22 up to @xmath24 to avoid finite - size effects , since close to lattice borders the structure of the backbone of percolation clusters is distorted and thus can falsify the saw statistics . let us denote by @xmath25 the total number of constructed saw trajectories between @xmath26 and @xmath22 ( we perform @xmath27 saws for each value of @xmath22 ) .
then , for each site @xmath28 of the backbone we sum up the portion of trajectories , passing through this site . in such a way
, we prescribe a weight @xmath29 to each site @xmath30 of the underlying fractal cluster .
critical exponent @xmath11 of saws on percolation clusters as function of space dimension @xmath4 ; stars : numerical data , obtained by us . left : dotted line : analytical results of janssen and stenull ( eq . ( [ spectr ] ) ) , solid line : analytical results of ref .
right : @xmath31/[2]$ ] pad approximants of the same analytical results.,title="fig:",width=222 ] critical exponent @xmath11 of saws on percolation clusters as function of space dimension @xmath4 ; stars : numerical data , obtained by us . left : dotted line : analytical results of janssen and stenull ( eq . ( [ spectr ] ) ) , solid line : analytical results of ref . @xcite .
right : @xmath31/[2]$ ] pad approximants of the same analytical results.,title="fig:",width=222 ] the multifractal moments @xmath32 are defined as follows : @xmath33 averaged over different configurations of the constructed backbones of percolation clusters , they scale as : @xmath34 with exponents @xmath12 that do not depend on @xmath35 in a linear or affine fashion , implying that saws on percolation clusters are multifractals . to estimate the numerical values of @xmath12 on the basis of data obtained by us ,
linear least - square fits are used .
at @xmath15 we just count the number of sites of the cluster of linear size @xmath22 , and thus @xmath36 corresponds to the fractal dimension of the backbone @xmath16 .
our results give @xmath37 , @xmath38 , @xmath39 . at @xmath40 ,
we restore the value of the exponent @xmath11 , governing the scaling law of the end - to - end distance for saws on the backbone of percolation clusters .
we obtain @xmath41 , @xmath42 , @xmath43 , in perfect agreement with our recent numerical estimates @xcite based on the scaling of the end - to - end distance with the number of saw steps .
comparison of our results for @xmath11 to that of the analytical studies @xcite are presented in fig .
[ nusaw ] .
spectrum of multifractal exponents @xmath12 as function of @xmath35 in @xmath44 .
the dotted lines present @xmath31/[2]$ ] pad approximants to the analytical results of janssen and stenull ( eq . ( [ spectr ] ) ) .
the inset for @xmath45 shows a comparison with results from ref .
@xcite ( bold solid line ) .
, title="fig:",width=188 ] spectrum of multifractal exponents @xmath12 as function of @xmath35 in @xmath44 . the dotted lines present @xmath31/[2]$ ] pad approximants to the analytical results of janssen and stenull ( eq . ( [ spectr ] ) ) . the inset for @xmath45 shows a comparison with results from ref .
@xcite ( bold solid line ) .
, title="fig:",width=192 ] spectrum of multifractal exponents @xmath12 as function of @xmath35 in @xmath44 .
the dotted lines present @xmath31/[2]$ ] pad approximants to the analytical results of janssen and stenull ( eq . ( [ spectr ] ) ) .
the inset for @xmath45 shows a comparison with results from ref .
@xcite ( bold solid line ) .
, title="fig:",width=192 ] the precision of our estimates decreases with increasing @xmath35 .
this problem turns out to be also especially crucial when exploring the moments with negative powers @xmath35 : the sites with small probabilities to be visited , which are determinant in negative moments , are very difficult to probe . our estimates of the exponents @xmath12 for different @xmath35 are presented in fig . [ nuq ] .
these values appear to be in perfect correspondence with analytical estimates down to @xmath45 dimensions , derived by applying pad approximation to the @xmath14-expansion ( [ spectr ] ) , presenting the given series as ratio @xmath46/[n]$ ] of two polynomials of degree @xmath47 and @xmath48 in @xmath49 .
we used the @xmath31/[2]$ ] approximant , because it appears to be most reliable in restoring the known estimates in the limiting case @xmath15 .
a direct use of the expression ( [ spectr ] ) gives worse results , especially for low dimensions @xmath4 where the expansion parameter @xmath14 is large .
to conclude , we have shown numerically that saws residing on the backbone of percolation clusters give rise to a whole spectrum of singularities , thus revealing multifractal properties . to completely describe peculiarities of the model , the multifractal scaling
should be taken into account .
we have found estimates for the exponents , governing different moments of the weight distribution , which scale independently .
work supported by eu marie curie rtn `` enrage '' : _ random geometry and random matrices : from quantum gravity to econophysics _ under grant no .
mrtn - ct-2004 - 005616 , schsische dfg - forschergruppe for877 , dfg graduate school of excellence `` buildmona '' , and research academy leipzig ( ral ) .
v.b . is grateful for support through the
marie curie international incoming fellowship " eu programme and a research fellowship of the alexander von humboldt foundation .
h.g . hentschel and i. procaccia , physica d 8 ( 1983 ) 435
; r. benzi , g. paladin , g. parisi , and a. vulpiani , j. phys . a 17 ( 1984 ) 3521 ; a. brandenburg , i. procaccia , d. segel , and a. vincent , phys . rev . a 46 ( 1992 ) 4819 ; d. queiros - conde , phys . rev .
e 64 ( 2001 ) 015301 ; l. biferale , g. boffetta , a. celani , b.j .
devenish , a. lanotte , and f. toschi , phys .
( 2004 ) 064502 .
m. schreiber and h. grussbach , phys .
67 ( 1991 ) 607 ; h. grussbach and m. schreiber , phys .
b 51 ( 1995 ) 663 ; a. mildenberger , f. evers , and a.d .
mirlin , phys .
b 66 ( 2002 ) 033109 ; a.d .
mirlin , y.v .
fyodorov , a. mildenberger , and f. evers , phys .
97 ( 2006 ) 046803 . | the scaling behavior of linear polymers in disordered media , modelled by self - avoiding walks ( saws ) on the backbone of percolation clusters in two , three and four dimensions is studied by numerical simulations .
we apply the pruned - enriched rosenbluth chain - growth method ( perm ) .
our numerical results yield estimates of critical exponents , governing the scaling laws of disorder averages of the configurational properties of saws , and clearly indicate a multifractal spectrum which emerges when two fractals meet each other . |
pks 1441 + 25 is a known high - energy ( he , @xmath4 ) @xmath5-ray flat spectrum radio quasar ( fsrq ) @xcite located at @xmath6 . in january 2015 it became active from @xmath5 rays to the near - infrared @xcite . in april , the detection of the source with a hard spectral index with the _ fermi gamma - ray space _ large area telescope ( lat ) together with increased multi - wavelength ( mwl ) emission triggered the magic observations .
they resulted in the first detection of this source at very high energy ( vhe , @xmath7 gev ) @xcite , later followed up by veritas @xcite .
this detection makes pks1441 + 25 the 5th fsrq with a firm classification detected at vhe , and the most distant known vhe source , along with qso b0218 + 357 ( @xmath8 , * ? ? ?
* ) . in this letter ,
the mwl observations are discussed in the context of an external compton model for four different states of activity , dubbed periods a ( mjd 57125.057130.0 ) , b ( 57130.057135.5 ) , c ( 57135.557139.5 ) and d ( 57149.057156.0 ) .
upper limits on the extragalactic background light ( ebl ) are obtained in the vhe band .
magic is a stereoscopic system consisting of two 17 m diameter imaging atmospheric cherenkov telescopes located at the observatorio del roque de los muchachos , on the canary island of la palma .
the current sensitivity for low - zenith observations ( @xmath9 ) above 220gev is @xmath10 of the crab nebula s flux in 50 hr @xcite .
the magic telescopes monitored pks 1441 + 25 from 2015 april 18 to 27 ( mjd 5713057139 , for a total of 29.9 hr ) and may 8 - 9 ( mjd 5715057151 , for 1.8 hr ) , the observational gap being due to the full - moon break . the observations were performed in wobble mode with a @xmath11 offset and four symmetric positions with respect to the camera center @xcite .
the data were collected in the zenith angle range of @xmath12 .
the analysis of the data is performed using the standard magic analysis framework _ mars _
@xcite and monte carlo ( mc ) simulations matching the night - sky background levels .
pks 1441 + 25 is detected with a significance of @xmath13 ( @xmath5-ray like excess @xmath14 ) during periods _ _
b__+_c_. no significant emission was found in period _
d_. the differential vhe spectrum is measured from 40 to 250gev and 50 to 160gev in periods _
b _ and _ c _ respectively . a power - law ( pwl ) can describe both observed and ebl - corrected spectra using the model of @xcite : @xmath15 where the normalization constant @xmath16 , the spectral index @xmath17 and the goodness of the fit ( @xmath18 and @xmath19-value ) are : 1 .
period _ b _ : a. observed : @xmath20 @xmath21 and @xmath22 ( @xmath23 , @xmath24 ) b. ebl - corrected : @xmath25 @xmath21 and @xmath26 ( @xmath27 , @xmath28 ) 2 .
period _ c _ : a. observed : @xmath29 @xmath21 and @xmath30 ( @xmath31 , @xmath32 ) b. ebl - corrected : @xmath33 and @xmath34 ( @xmath35 , @xmath36 ) from a likelihood ratio test ( lrt ) , a model with intrinsic curvature such as a log - parabola ( lp ) is preferred at @xmath37 during period _ b_. it is defined as : @xmath38 where @xmath39 @xmath40 , @xmath41 and @xmath42 ( @xmath43 , @xmath44 ) . a full description of the magic systematic uncertainties can be found in @xcite and references therein .
in nominal survey mode the lat monitors the entire @xmath5-ray sky every 3 hr in the energy range from @xmath45 to at least @xmath46 @xcite .
we select pass 8 source class events collected from 2015 april 8 to may 23 ( mjd 5712057165 ) between 100 mev to 500 gev and within a @xmath47 region of interest ( roi ) centered at the location of pks 1441 + 25 . in order to reduce contamination from the earth limb , a zenith angle cut of @xmath48
is applied .
the analysis is performed with the sciencetools software package version v10r0p5 using the ` p8r2_source_v6 ` instrument response functions and the ` gll_iem_v06 ` and ` iso_p8r2_source_v6_v06 ` models for the galactic and isotropic diffuse emission , respectively .
the likelihood fit is performed using ` gtlike ` , including all 3fgl sources @xcite within @xmath49 from pks 1441 + 25 .
the spectral indices and fluxes are left free for sources within @xmath47 , while sources from @xmath47 to @xmath49 have their parameters fixed to the catalog value .
both the flux and the spectral index of pks 1441 + 25 are left free for the light curve calculation , while the parameters for the rest of the sources in the roi are fixed except the diffuse components .
five photons of energies 10 - 50gev were detected with a probability of association with pks 1441 + 25 larger than @xmath50 , calculated with ` gtsrcprob ` .
the spectrum of pks 1441 + 25 is well fit by a pwl ( as in the 3fgl catalog ) and no significant curvature was found . during the flare ( period b+c ) , the spectral index is @xmath51 , harder than the 3fgl value @xmath52 .
( nuclear spectroscopic telescope array ; @xcite ) is a hard x - ray telescope operating in the energy range between @xmath53 and 79kev .
pks 1441 + 25 was observed with _ nustar _ on 2015 april 2526 ( mjd 57137.1113 ) for a total ( on - source ) exposure of 40 ks .
the data are processed using the standard ` nupipeline ` script ( version 1.4.1 ) available in the nustardas software package @xcite .
the source spectrum extends up to @xmath5425 kev , and can be described by a pwl with spectral index @xmath55 ( @xmath56 ) .
no significant variability is detected during the observation .
a _ swift _ @xcite target of opportunity started on 2015 april 15 .
_ swift_-xrt @xcite observed the source in photon - counting mode .
standard filtering and analysis of the data were employed .
the source exhibited a soft x - ray photon index ( from @xmath57 to @xmath58 ) and is described by an absorbed pwl model , with the galactic absorption fixed to @xmath59 @xmath60 @xcite during april may 2015 . for comparison , the observations on 2010 june 12 ( mjd 55359 ) can be fit with a pwl with spectral index @xmath61 the _ swift_-uvot @xcite flux in several bands was estimated using aperture photometry . de - reddening is performed using @xmath62 @xcite and @xmath63 @xcite .
optical r - band observations started on mjd 57130 and were performed using the 35 cm celestron telescope attached to the kva 60 cm telescope ( la palma , canary islands , spain ) and the 50 cm hans - haffner - telescope ( hettstadt , wrzburg , germany ) .
the data are analyzed using differential photometry and corrected for galactic extinction @xcite .
the host galaxy contribution is negligible compared to the flux of the source during these observations .
the optical emission shows a high degree of polarization , reaching a maximum of @xmath64 on mjd 57132 @xcite .
nir observations in the j , h , and k@xmath65 bands started on mjd 57141 with canica , a direct camera at the 2.1 m telescope of the guillermo haro observatory located at cananea , mexico . the flux is estimated by means of differential photometry using the 2mass catalog @xcite .
the observations of pks 1441 + 25 with the metshovi 13.7-m radio telescope started on mjd 57135 .
the measurements were made with a 1 ghz - band dual beam receiver centered at 37 ghz . a detailed description of the observation and analysis methods can be found in @xcite .
the mwl light curve is presented in fig . [
fig : mwl_lc ] . in the vhe band
, the no - variability hypothesis can be discarded as it results in a @xmath66 ( @xmath67 ) for _ _ b__+__c__+_d_. a constant fit is also unlikely for the flare in april ( _ _ b__+_c _ ) with a @xmath68 ( @xmath69 ) .
we gauge the characteristic variability time scale by heuristically fitting the vhe light curve with a gaussian function .
the fit provides a standard deviation @xmath70 ( halving flux time of @xmath71 ) and a peak flux of @xmath72 ( @xmath73 , @xmath74 ) . for x - rays , a halving flux time of @xmath75
was found .
the average flux in _ b _ is larger than in _
c _ by a factor of @xmath76 in vhe .
a similar pattern was found in x - rays ( @xmath77 ) , optical ( @xmath78 ) and a hint in the he ( @xmath79 ) .
no intra - night variability is detected .
cccccccccc _ a _ & 57125.057130.0 & @xmath80 & @xmath81 & @xmath81 & @xmath82 & @xmath83 & @xmath84 & @xmath85 & @xmath86 + _ b _ & 57130.057135.5 & @xmath80 & @xmath81 & @xmath81 & @xmath87 & @xmath83 & @xmath88 & @xmath89 & @xmath90 + _ c _ & 57135.557139.5 & @xmath91 & @xmath92 & @xmath81 & @xmath93 & @xmath94 & @xmath95 & @xmath96 & @xmath97 + _ d _ & 57149.057156.0 & @xmath91 & @xmath98 & @xmath99 & @xmath100 & @xmath101 & @xmath102 & @xmath103 & @xmath104 + archival & - & @xmath105 & @xmath106 & @xmath107 & @xmath108 & @xmath109 & @xmath110 & @xmath111 & @xmath112 + [ tab : model ] the mwl spectral energy distributions ( seds ) shown in fig .
[ fig : mwl_sed ] indicate a shift of both synchrotron and inverse - compton ( ic ) peaks to higher energies during the 2015 observation campaign with respect to the archival data , accompanied by a significant variation of the x - ray and he @xmath5-ray spectral indices .
this behavior resembles the less extreme outburst seen in pmn j23451555 @xcite , and can be explained by a change in the emitting region location : within the broad line region ( blr ) in the quiescent state to beyond the blr during the outburst , where the external photon field is dominated by the optical uv from the blr or the ir thermal emission of a dusty torus , respectively ( conventional framework for @xmath5-loud fsrq , e.g. * ? ? ?
* ; * ? ? ?
the consequences of this scenario are two - fold : ( 1 ) since the radiation energy density of the ir component is much lower than the one associated with the optical
uv photons from the blr , the electron radiative cooling is less effective and the energy reachable by the acceleration process could be higher , accounting for the shift of the sed peak toward higher energies ; ( 2 ) given the much lower energy of the external photons , absorption of @xmath5 rays by pair production occurs only above several hundreds of gev ( e.g. , * ? ? ?
* ) , enabling the detection of fsrqs at vhe . for an emission region
well within the blr , strong absorption features are expected for energies above tens of gev ( see e.g. * ? ? ?
* ; * ? ? ?
* ) , which are not observed in the 2015 mwl seds .
according to this framework , the proposed sed model for the 2015 observations assumes that the emission region is located at a distance @xmath113 from the central compact object . adopting the simple scaling proposed by @xcite
, @xmath114 depends only on the disk luminosity , @xmath115 cm .
the latter can be inferred from the luminosity of the optical broad lines , @xmath116 @xcite , resulting in @xmath117 . in the same way , the size of the dusty torus can be inferred from a similar scaling law , @xmath118 cm .
the resulting emission is calculated using the code described in @xcite .
the emission region is assumed to be spherical ( in the source frame ) with radius @xmath119 , in motion with bulk lorentz factor @xmath17 at angle @xmath120 with respect to the line of sight .
it contains a tangled magnetic field with intensity @xmath121 and a population of relativistic leptons described by a smoothed broken pwl energy distribution between lorentz factors @xmath122 and @xmath123 , with a break at @xmath124 , slopes @xmath125 and @xmath126 and normalization @xmath127 estimated at @xmath5=1 . the external photon field ( dominated by the ir torus emission )
is assumed to follow a black body spectrum with luminosity @xmath128 ( @xmath129 , following * ? ? ?
* ) and temperature @xmath130 , diluted within a region of radius @xmath131 .
rrrrrrrrr pwl & no ebl & - & - & @xmath132 & @xmath133 & & & @xmath134 + pwl & f08 & @xmath135 & @xmath136 & @xmath137 & @xmath138 & & & @xmath139 + pwl & d11 & @xmath140 & @xmath141 & @xmath142 & @xmath138 & & & @xmath143 + pwl & g12 & @xmath144 & @xmath145 & @xmath146 & @xmath147 & & & @xmath148 + pwl & s14 ( max ) & @xmath140 & @xmath141 & @xmath142 & @xmath138 & & & @xmath143 + pwl & s14 ( min ) & @xmath149 & @xmath150 & @xmath146 & @xmath147 & & & @xmath151 + lp & no ebl & - & - & @xmath132 & @xmath152 & @xmath153 & & @xmath154 + lp & f08 & @xmath155 & @xmath156 & @xmath157 & @xmath158 & @xmath159 & & @xmath160 + lp & d11 & @xmath161 & @xmath162 & @xmath157 & @xmath163 & @xmath164 & & @xmath154 + lp & g12 & @xmath165 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & & @xmath160 + lp & s14 ( max ) & @xmath161 & @xmath162 & @xmath157 & @xmath163 & @xmath164 & & @xmath154 + lp & s14 ( min ) & @xmath170 & @xmath171 & @xmath142 & @xmath172 & @xmath173 & & @xmath174 + pwlsc & no ebl & - & - & @xmath175 & @xmath176 & @xmath177 & @xmath178 & @xmath179 + pwlsc & f08 & @xmath180 & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath143 & @xmath179 + pwlsc & d11 & @xmath185 & @xmath162 & @xmath147 & @xmath164 & @xmath186 & @xmath187 & @xmath179 + pwlsc & g12 & @xmath188 & @xmath189 & @xmath190 & @xmath191 & @xmath192 & @xmath179 & @xmath179 + pwlsc & s14 ( max ) & @xmath185 & @xmath162 & @xmath147 & @xmath164 & @xmath186 & @xmath187 & @xmath179 + pwlsc & s14 ( min ) & @xmath193 & @xmath171 & @xmath194 & @xmath154 & @xmath195 & @xmath94 & @xmath196 [ tab : ebl ] the model also includes @xmath5-ray absorption within the ir radiation field of the torus . assuming a temperature @xmath197 for the ir torus , the maximum absorption is reached at @xmath198 in the source frame , with an optical depth @xmath199 @xmath200 .
given the large optical depth and the relatively broad spectrum of the target photons , the absorption is appreciable at few hundreds of gev , i.e. @xmath201 at @xmath202 and @xmath203 at @xmath46 in the observer frame .
note also that an additional softening of the spectrum can be due to the fact that the emission in the vhe band is produced by scattering occurring in the klein
nishina ( kn ) regime ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . to decrease the number of free parameters , we fix the bulk lorentz and doppler factor to @xmath204 and @xmath205 , close to the average obtained for a large sample of @xmath5-loud fsrq @xcite .
this implies a viewing angle of the jet @xmath206 , and the aperture angle is fixed to @xmath207 rad ( @xmath208 ) .
we assume that the emission region is located beyond but not very far from the blr , @xmath209 , implying @xmath210 .
the low - energy slope @xmath125 is fixed to the standard value of 2 .
the remaining parameters are chosen to reproduce the synchrotron and ic components ( see table [ tab : model ] ) . to account for the different flux states , an evolution in both the electron distribution and the magnetic field is required . for comparison , the archival data ( representation of the quiescent state ) were modeled considering the emitting region partially within the blr ( standard framework ) at @xmath211 cm , so that the @xmath212 optical depth is small as indicated by the highest energy point of the 3fgl spectrum .
the ratio between the peak luminosities ( compton dominance , cd ) , are reported in table [ tab : model ] . during the outburst
, @xmath213 lies more than an order of magnitude outside the fsrq parameter space in the cd sequence proposed by @xcite , indicating a shift in the sequence during flares .
the high degree of optical polarization suggests that the emission may come from a compressed region in the jet like an internal shock , which is also an ideal site for electron acceleration / injection .
vhe @xmath5 rays from distant blazars can interact with the optical
uv photons from the ebl via pair production @xcite , resulting in an attenuation of the intrinsic vhe spectrum .
the ebl imprint in the @xmath5-ray spectra from distant blazars can be used to constrain the ebl density .
measurements of the ebl absorption can be derived under some assumptions on the intrinsic spectrum of the source ( see e.g. * ? ? ?
* ; * ? ? ?
* ) . with a redshift of @xmath0 and a strong detection in the vhe band , pks 1441 + 25
allows us to probe ebl models at a distance never explored before in this energy regime with ground - based gamma - ray instruments .
however , kn effects together with an expected intrinsic @xmath5-ray absorption in the vhe band ( see sec . [ broadbandsed ] ) , can mimic the effect of ebl absorption , making it difficult to disentangle the two effects .
a lrt was used to compare a null hypothesis ( no ebl absorption ) with respect to the hypothesis of ebl absorption with a scaled opacity @xmath214 as in @xcite .
predicted opacities from @xcite ( @xmath215 ) , @xcite ( @xmath216 ) , @xcite ( @xmath217 ) and @xcite ( @xmath218 ) are considered , while @xmath219 is a free scaling parameter .
different intrinsic spectral shapes were assumed : pwl @xmath220 , lp @xmath221 and pwlsc @xmath222}$ ] where @xmath223 is measured in @xmath224 and @xmath225 .
the limits are reported in table [ tab : ebl ] and an example is given in fig .
[ fig : ebl_sed ] .
a possible overall systematic error of @xmath226 in the absolute energy scale of the instrument is considered . under the assumption that no curvature is present in the intrinsic vhe spectrum ,
the measured spectrum is compatible with the present generation of ebl models .
the @xmath227 confidence level limit obtained in this work for @xcite is compatible with the one found in @xcite for @xmath228 , @xmath229 , which is obtained from observations with a wide range of redshift values while our ul is calculated for a precise redshift value .
the estimated scaling on the optical depth can be translated into ebl density constraints as shown in @xcite and @xcite .
the observed vhe spectrum allow us to constrain the ebl density between @xmath230 and @xmath231 , where the optical depth with respect to the nominal value of @xcite , @xmath232 , implies in the local universe @xmath233 .
magic has detected for the first time vhe emission from the @xmath0 blazar pks 1441 + 25 during a mwl outburst in 2015 april .
pks 1441 + 25 is , together with qso b0218 + 357 , the most distant vhe source detected so far .
this allow us to study vhe blazars when the universe was only half of its current age .
the evolution of the mwl sed is studied in the framework of an external compton emission model .
the absence of intrinsic absorption features in the he and the vhe regime constrains the localization of the emitting region to be just outside of the blr during this period of high activity , while it is expected to be partially compatible with the blr during the period of low activity .
the sed evolution reveals changes in the electron distribution and the magnetic field .
for the first time , the vhe measurements are used to indirectly probe the ebl at redshifts out to @xmath234 with ground - based gamma - ray instruments .
although an internal cutoff can not be excluded , the measured vhe spectrum is consistent with a steepening due to attenuation caused by the ebl . employing state - of - the - art ebl models , upper limits to the ebl density
are derived .
the upper limits on the opacity calculated under the assumption of an intrinsic spectrum compatible with a pwl function for different ebl models result in @xmath235 , @xmath236 , @xmath237 , @xmath238 and @xmath239 for ebl models from @xcite , @xcite , @xcite and maximum and minimum from @xcite , respectively .
abdo , a. a. , ackermann m. , ajello , m. , et al .
2010a , apjs , 188 , 405 abdo , a. a. , ackermann , m. , ajello , m. , et al .
2010b , apj , 723 , 1082 abramowski , a. , acero , f. , aharonian , f. , et al .
2013 , a&a , 2013 , 550 , a4 acero , f. , ackermann , m. , ajello , m. et al .
2015 , apjs , 218 , 23 ackermann , m. , ajello , m. , albert , a. , et al .
2012 , science , 338 , 1190 ackermann , m. , ajello , m. , allafort , a. , et al .
2013 , apjs , 209 , 34 aleksi , j. , ansoldi , s. , antonelli , l. a. , et al .
2015 , aph , 72 , 76 atwood , w. b. , abdo , a. a. , ackermann , m. , et al .
2009 , apj , 697 , 1071 blumenthal , g. r. & gould , r. j. , 1970 , rev .
42 , 237 burrows , d. n. , hill , j. e. , nousek , j. a. , et al .
2005 , space sci .
120 , 165 - 195 carrasco , l. , miramon , j. , porras a. et al .
2015 , the astronomer@xmath240s telegram # 6895 domnguez , a. , primack , j. r. , rosario , d. j. , et al .
2011 , mnras , 410 , 2556 - 2578 .
donea , a. c. & protheroe , r. j. , 2003 , astroparticle physics , 18 , 377 - 393 finke , j. d. , 2013 , apj , 763 , 134 fomin , v. p. , sepanian , a. a. , lamb , r. , et al .
1994 , astropart .
phys . , 2 , 137 franceschini , a. ; rodighiero , g. ; vaccari , m. 2008 , a&a 487 , 837 - 852 gehrels , n. , chincarini , g. , giommi , p. , et al .
2004 , apj , 611 , 1005 ghisellini , g. & tavecchio , f. , 2009 , mnras , 396 , 105 - 109 ghisellini , g. , tavecchio , f. , foschini , l. , bonnoli , g. & tagliaferri , g. , 2013 , mnras , 432 , 66 - 70 ghisellini , g. & tavecchio , f. , 2015 , mnras , 448 , 1060 - 1077 gilmore , r. c. , somerville , r. s. , primack , j. r. & domnguez , a. , 2012 , mnras , 422 , 3189 - 3207 gould r. j. & schreder g. p. , 1967 ,
phrv , 155 , 1404 harrison , f. a. , craig , w. w. , christensen f. e. , et al .
2013 , apj , 770 , 103 kalberla , p. m. w. , burton , w. b. , hartmann , d. , 2005 , a&a , 440 , 775 liu , h. t. , bai , j. m. , apj , 2006 , 653 , 1089 - 1097 maraschi , l. & tavecchio , f. 2003 , apj , 593 , 667 .
mirzoyan , r. on behalf of the magic collaboration , 2015 , the astronomer@xmath240s telegram # 7416 moderski , r. , sikora , m. , coppi , p. s. , & aharonian , f. , 2005 , mnras , 363 , 954 - 966 mukherjee , r. on behalf of the veritas collaboration , 2015 , the astronomer@xmath240s telegram # 7433 nolan , p. l. , abdo , a. a. , ackermann , m. , et al .
2012 , apjs , 199 , 31 ojha , r. on behalf of the fermi - lat collaboration 2015 , the astronomer@xmath240s telegram # 6878 pacciani , l. 2015 , the astronomer@xmath240s telegram # 7402 perri , m. , puccetti , s. , spagnuolo , n , et al .
2014 , _ the nustar data analysis software guide _ , http://heasarc.gsfc.nasa.gov/docs/nustar/analysis/nustar_swguide.pdf protheroe , r. j. & biermann , p. l. , 1997 , astroparticle physics , 6 , 293 - 300 pursimo , t. & ojha , r. 2015 , the astronomer@xmath240s telegram # 6923 roming , p. w. a. , kennedy , t. e. , mason , k. o. , et al .
2005 , space sci .
120 , 95 - 142 schlafly e. f. & finkbeiner d. p. , 2011
, apj , 737 , 103 , g. v. & wiemer , w. , 1975 , a&a , 43 , 133 - 139 scully , s. t. , malkan , m. a. , stecker , f. w. , 2014 , apj , 784 , 138 shaw , m. s. , romani , r. w. , cotter , g. , et al .
2012 , apj , 748 , 49 sitarek , j. , becerra gonzlez , j. , dominis prester , d. et al .
2015 , pos icrc 2015 , 825 skrutskie , m. f. , cutri , r. m. , stiening , r. , et al .
2006 , aj , 131 , 1163 smith , p. s. & tutar ozdarcan , d. , 2015 , the astronomer@xmath240s telegram # 7417 stecker f. w. , dejager o. c. , salamon m. h. , 1992 , apj , 390 , l49 tavecchio , f. , becerra gonzalez , j. , ghisellini , g. , et al .
2011 , a&a , 534 , id.a86 tersranta h. , tornikoski , m. , mujunen , a. , et al .
1998 , a&as , 132 , 305 zanin , r. , carmona , e. , sitarek , j. , et al .
2013 , proc . of the 33th international cosmic ray conference ( icrc ) , 2 - 9 july 2013 , rio de janeiro , brazil , # 773 zdziarski , a. a. , & krolik , j. h. 1993 , apj , 409 , l33
we would like to thank the instituto de astrofsica de canarias for the excellent working conditions at the observatorio del roque de los muchachos in la palma .
the financial support of the german bmbf and mpg , the italian infn and inaf , the swiss national fund snf , the erdf under the spanish mineco ( fpa2012 - 39502 ) and mecd ( fpu13/00618 ) , and the japanese jsps and mext is gratefully acknowledged .
this work was also supported by the centro de excelencia severo ochoa sev-2012 - 0234 , cpan csd2007 - 00042 , and multidark csd2009 - 00064 projects of the spanish consolider - ingenio 2010 programme , by grant 268740 of the academy of finland , by the croatian science foundation ( hrzz ) project 09/176 and the university of rijeka project 13.12.1.3.02 , by the dfg collaborative research centers sfb823/c4 and sfb876/c3 , and by the polish mniszw grant 745/n - hess - magic/2010/0 . the _ fermi_-lat collaboration acknowledges support for lat development , operation and data analysis from nasa and doe ( united states ) , cea / irfu and in2p3/cnrs ( france ) , asi and infn ( italy ) , mext , kek , and jaxa ( japan ) , and the k.a .
wallenberg foundation , the swedish research council and the national space board ( sweden ) .
science analysis support in the operations phase from inaf ( italy ) and cnes ( france ) is also gratefully acknowledged . | the flat - spectrum radio quasar pks 1441 + 25 at a redshift of @xmath0 is detected between 40 and 250gev with a significance of 25.5@xmath1 using the magic telescopes .
together with the gravitationally lensed blazar qso b0218 + 357 ( @xmath2 ) , pks 1441 + 25 is the most distant very high energy ( vhe ) blazar detected to date .
the observations were triggered by an outburst in 2015 april seen at gev energies with the large area telescope on board _
fermi_. multi - wavelength observations suggest a subdivision of the high state into two distinct flux states . in the band covered by magic
, the variability time scale is estimated to be @xmath3 days .
modeling the broadband spectral energy distribution with an external compton model , the location of the emitting region is understood as originating in the jet outside the broad line region ( blr ) during the period of high activity , while being partially within the blr during the period of low ( typical ) activity .
the observed vhe spectrum during the highest activity is used to probe the extragalactic background light at an unprecedented distance scale for ground - based gamma - ray astronomy .
= 2 = 200 |
[ cols="<,<,<,<,^,^,^,^,^",options="header " , ] in i18151 , s106 , and v645 cyg , the co @xmath12 bandhead emission is detected ( as previously seen in s106 by @xcite and in v645 cyg , as discussed in @xcite ) .
fig.[co_bh ] presents the spectra .
s106 shows a clear double - peaked profile at the bandhead .
the interpretations of the spectral shape of the bandhead have been provided by several authors @xcite .
figure2 in the paper by @xcite is instructive to explain the spectral features of a disc model .
the emission of a single line from an inclined keplerian rotating disc is described with the double - horned rotational broadening function .
the model bandhead is then simply the convolution of this broadening function with the intrinsic co emission .
this leads to a distinct `` blue shoulder '' of the bandhead where there is blue shifted emission relative to the line of sight , which sets a constraint on both the line width and the inclination of the source .
this is clearly seen in s106 .
it should be noted that a double - peaked feature can also be reproduced with a wind model @xcite , when the opening angle of the conical wind is so small that the observer sees the red component only from the far side and the blue component only from the front side .
this is not the case in s106 because this object is seen near to edge - on @xcite .
the spectral features of v645 cyg can also be explained with the same models .
the emission peak is located around 2.293 @xmath13 m , which is between the blue shoulder and red peak in s106 and where the difference of the system velocities is negligible . in the face - on case ,
the rotational broadening function is single - peaked , and the peak stays at the bandhead wavelength . in this object ,
only the blueshifted components are detected in the and outflow emissions @xcite .
this result indicates that the system is seen near face - on and the redshifted part is obscured by the equatorial material , presumably an accretion disc .
a single - peak is also reproduced in a wind model if the opening angle of the wind is large .
however we would then expect the peak of the emission to be shifted redwards compared to the nominal @xmath14 of the source and it clearly is not . for i18151 , the data , although noisy , show no significant emission bluewards of the peak of the bandhead that can not be explained by simple line broadening .
the inclination angle of i18151 is estimated to be @xmath15 @xcite , intermediate between the cases of s106 and v645 cyg , and the appearance of the spectrum is consistent with this if the origin of the emission is a disc .
= 20 bandhead spectra .
the thin and thick lines denote the observations and the model results , respectively .
the bars around @xmath16 m indicate the local standard of rest .
, width=321 ] we have also used a simple model of the gas phase co disc to fit the data .
we analyze the spectra attributed to the co first over - tone bandhead emission based on @xcite to derive the physical parameters of the disc geometry ( see also * ? ? ?
* ; * ? ? ?
* ) . in our model ,
up to @xmath17 rotational levels for the @xmath8 under the local thermodynamic equillibrium are considered .
the co energy levels are calculated with the @xmath18 parameters ( see equation ( 6 ) ; * ? ? ?
* ; * ? ? ?
* ) , and the einstein coefficients are taken from @xcite . the co@xmath19h@xmath3 ratio is assumed to be @xmath20 .
for the disc geometry , we use a temperature gradient model , @xmath21 , and a geometrically thin disc with a surface density @xmath22 , where @xmath23 is the distance from the central star .
the disc has inner and outer radii of @xmath24 and @xmath25 , respectively , and is tilted with an inclination angle , @xmath26 , where @xmath27 indicates face - on .
the co molecules are assumed to have a keplerian rotating motion @xmath28 , where @xmath29 and @xmath30 are the gravitational constant and the stellar mass , respectively .
the disc is divided into 75 radial rings each with 75 azimuthal cells .
the intrinsic line shape of each transition has a gaussian function with a line width of @xmath31 .
the total spectra are obtained by summing the flux from individual cells , in which the wavelength shift due to the rotational velocity of the disc is taken into account .
the outer radius is set by the condition that the gas phase temperature drops below 1000 k , at which point the @xmath8 emission is no longer excited .
in addition , the c - o bonding is dissociated at temperature above 5000 k. hence , the maximum temperature is chosen not to exceed 5000 k. with these physical assumptions the co bandhead spectra are produced and compared to the observed spectra using an idl script .
the best fit model parameters are found using @xmath32 minimization .
we examine the three objects with co emission , i18151 , s106 and v645 cyg .
the distance and the stellar mass are assumed to be 2.0 kpc and 15 @xmath11 for i18151 , 1.7 kpc and 20 @xmath11 for s106 , and 3.6 kpc and 20 @xmath11 for v645 cyg ( see table [ obs_parm ] ) . with these assumptions , we searched for the best fit parameters of @xmath24 , @xmath25 , @xmath33 , @xmath34 , @xmath26 , @xmath35 , @xmath31 , and the column density of co , @xmath36 .
it turned out that the @xmath33 and @xmath34 parameters have particularly large uncertainties .
hence , we set fixed values of @xmath37 , respectively , which are derived for intermediate mass stars @xcite .
the results of other six parameters are presented in table[tab : cobh ] . only @xmath26 and @xmath35 are estimated with good accuracy .
a similar analysis was done for s106 by @xcite . although the estimated inclination angle they used of @xmath38 is somewhat different from ours , their estimate of the size of the co emitting region of 0.8 to 1.5 au is in reasonable agreement . in both s106 and v645 cyg ,
the best fitting inclination angle is consistent with the qualitative expectations discussed previously , as well as available results from the literature as given in appendix [ discussion ] , suggesting that our estimated stellar masses are also reasonable .
d10 analysed the low j absorption line velocity structure of co @xmath8 .
their data show a pair of blueshifted and redshifted components at east and west with respect to the flux peak position in the @xmath1-band continuum in w33a .
they found the classical signature of a keplerian rotating structure in the position - velocity ( pv ) diagram and a central mass of 15 @xmath11 , which includes the masses of the central star and the hot gas inside the cool gas region .
both i18151 and afgl 2136 show absorption features in the low @xmath9 transition of co @xmath8 . from the cube data
, we used 11 absorption lines , which appear between 2.32 @xmath13 m and 2.35 @xmath13 m @xcite .
we first derived the wavelength positions of individual absorption features , which are regarded as the zero - velocity .
the images of all branches are aligned with the relative velocity and are added to increase the signal - to - noise ratio .
then , the differential images between the redshifted ( @xmath39 ) and blueshifted ( @xmath40 ) components , i.e.the subtraction of the blueshifted component image from the redshifted component images , the images of the velocity centroids , and the pv diagrams are made .
the top and bottom panels of fig.[co_vc ] show the results of i18151 and afgl 2136 , respectively .
the velocities range between @xmath41 kms@xmath5 for i18151 and @xmath42 kms@xmath5 for afgl 2136 .
although the standard deviations in velocity are large , we can clearly see a pair of redshifted and blueshifted components in the @xmath43 images ( left panels ) .
the separations and the position angles of these components are obtained by measuring the absorption centroids of the blueshifted and redshifted components to be 0.038 arcsec at @xmath44 for i18151 and 0.054 arcsec at @xmath45 for afgl 2136 with typical errors of one tenth of a pixel on the centroid and 5@xmath46 on the angle . applying a slit along these position angles ,
the pv diagrams are made ( right panels ) and the velocity centroid at each position pixel ( 0.043 arcsec spacing ) is plotted with pink dots , with a typical error on the velocity of better than one tenth of a resolution element , or kms@xmath5 .
the redshifted component is located on the right side of the panel . in the velocity centroid maps , we see a skew , which is often seen in objects with keplerian rotation , though the correspondence with a keplerian model is not perfect even given the estimated errors . if we assume that this is also the case in our results , we can estimate the central masses .
for this purpose , we made a simple model .
the keplerian rotating velocity of the co gas , @xmath47 , is calculated by @xmath48 , where @xmath49 is the central mass .
the inclination angles , @xmath26 , and the distances to these objects are assumed to be @xmath15 and 2.0 kpc for i18151 @xcite and @xmath50 and 2.0 kpc for afgl 2136 @xcite . to allow direct comparison with the observations ,
the resulting velocity curve is obtained by convolving gaussian functions with the psf sizes ( 0.1 arcsec ) . in the pv diagram ,
the modeled velocity curves are plotted with green lines .
the observed and model data are sufficiently similar to conclude that we do see keplerian rotation in the cool absorbing gas in these sources .
the resultant central masses are 30 @xmath11 for i18151 and 20 @xmath11 for afgl 2136 , with error of approximately 10% assuming the keplerian model is a good fit .
the central mass of 30 @xmath11 for i18151 is larger than that of the inferred mass for a b0 star of @xmath51 that was used in modelling the co bandhead emission .
however , the absorbing gas includes the mass of not only the central star but also all the gas located inside the maximum extent of the region of absorbing gas .
although the error on these measurements is large enough to suggest both absorption and emission kinematical masses could be equivalent , it may also suggest a larger reservoir of material still available for accretion by the central star . the large 220@xmath11 mass found for the `` disc - like '' structure by @xcite
is certainly consistent with the picture in which more accretion could take place .
a similar situation was found by d10 for w33a as well .
the orientation of the redshifted and blueshifted components in the co @xmath8 absorption line is nearly northeast - southwest in i18151 , perpendicular to the larger scale outflow .
this is notably different from both the ionised spectroastrometric signature seen for br@xmath2 , and the small scale north - south h@xmath3 jet we see in our data .
it is possible that the outflow is precessing with time , though no intermediate angles are seen in the h@xmath3 jet by d04 .
it is also possible that the jet initially launches in a north - south direction , but is deflected by the larger scale molecular envelope in the observed southeast - northwest direction . in either case
, the structure of the accreting gas in the immediate vicinity of the star appears to lie at different orientations when moving further away from the star .
m on velocity scale . , width=321 ] the orientation of the co absorption in afgl 2136 is also puzzling .
the blue- and red - shifted components of the absorption centroid lie at a position angle of @xmath52 with a separation of @xmath53100 au . as noted in appendix [ discussion ] , the observed outflow axis runs se nw , consistent with the polarisation disc with a position angle of @xmath5345@xmath46 found by @xcite .
it is possible that the distribution of mass on a few thousand au is different from that on hundreds au , i.e. a simple axisymmetric structure may not be applicable in this object , just as appears to be the case for i18151 . nevertheless , by assuming a keplerian rotating motion in the cold co gas
, the velocity structure can be reproduced well with a 20 @xmath11 mass model .
although the stellar parameters of afgl 2136 have not been precisely determined , it corresponds to a late o zero - age main sequence star in terms of the estimated luminosity of 5.3@xmath5410@xmath55 , corresponding to a @xmath56 star . in our sample ,
the @xmath57 ( @xmath58 m ) line is detected in i19065 , s106 , and v645 cyg , and the @xmath59 ( @xmath60 m ) blend is present as a weak absorption feature in s106 .
figure[hei ] presents the line profile on a @xmath14 scale for the @xmath57 in all three objects .
we see emission in i19065 and absorption in s106 and v645 cyg . notably the emission feature seen in i19065 is broader than expected from unresolved emission from the ionised hii region , suggesting that this , like the br@xmath2 emission , arises from the star in this source .
@xcite have obtained @xmath1-band spectra of be stars , and @xcite and @xcite of a more general sample of ob stars of all classes .
the trends with spectral type are relatively clear and can be used to classify the emission as seen in our targets .
hei @xmath57 appears in emission in later type stars in dwarfs , giants and supergiants , and in many of the b0ve and b1ve stars in @xcite , and in oe stars with types as early as o6ve in @xcite .
the lack of any detectable @xmath59 line in i19065 however suggests a relatively narrow range of type of around b0ve or b1ve , which is consistent with the known luminosity of this source from the rms survey @xcite .
@xcite previously detected the @xmath57 line in s106 , mostly in absorption as we also see it .
they argue that this absorption probably arises due to high optical depth in the 584 line and a high electron density ( @xmath61 @xmath62 ) .
the fwhm is 65 kms@xmath5 and is in good agreement with the result of 70 kms@xmath5 by @xcite .
this velocity width is significantly smaller than that of the br@xmath2 emission lines ( @xmath53100 200 kms@xmath5 ) .
@xcite proposed that this line traces the region where the outflow from a disc - wind is still accelerating .
weak hei @xmath59 absorption is seen , but there is no niii 87 emission at @xmath63 m emission as would be seen in earlier spectral types @xcite .
this suggests a type of o9ve b0ve .
we also carried out a spectroastrometric study of the @xmath57 line given its strength .
overall there is only a weak signature near the systemic velocity , in the direction of the bipolar nebula , perhaps consistent with a slight degree of infilling of the absorption line profile by larger scale nebular emission .
our lower resolution spectroscopy taken for the rms survey @xcite shows nebular hei @xmath57 emission consistent with this .
the spectroastrometry does not show evidence for any other feature .
in particular , there is nothing similar to the disc like structure seen in fig.[bg_spcast ] . in v645
cyg , the absorption feature is strongly blueshifted and appears around @xmath64@xmath6 kms@xmath5 .
a similar trend is seen in the spectrum presented by @xcite . however , the spectral feature spreads up to @xmath65 kms@xmath5 in their data , but is detected only up to @xmath66 kms@xmath5 in our spectrum .
v645 cyg changes its spectral appearance regularly , complicating any attempts to derive an exact spectral type .
however , the spectra we have is consistent with approximately o9ve , similar to s106 .
the star can not be hotter , given the lack of niii , and the lack of any hei @xmath59 suggests it can not be much cooler either .
the other metal lines that we see in our spectra , such as nai , mgii or feii are less sensitive to the spectral type . in all three cases
it is plausible that the emission actually comes from the accretion disc rather than the stellar atmosphere @xcite .
overall the hei and other metal lines suggest a common picture in which a stellar wind or disc wind can be present , along with an accretion disc
. it would be interesting to determine whether spectroastrometry could be used on some of these metal lines in future higher signal - to - noise studies to test whether this scenario is correct .
we presented high - resolution @xmath1-band integrated field spectroscopy of six mysos obtained using the ao assisted nifs instrument on the gemini - north telescope .
this observing technique allows us to analyse three - dimensional spectroscopy across emission and absorption lines . in the previous paper ( d10 )
, we successfully detected a bipolar jet like feature in the br@xmath2 emission line and a keplerian rotating motion of the neutral co gas disc around a @xmath5315 @xmath11 mass central star in w33a . in this new work
, we increase the sample , with four out of six being already well studied and two new .
our new data display a number of spectral lines such as br@xmath2 emission lines , co bandhead emissions and absorptions , h@xmath3 @xmath67 s(1 ) , s(2 ) , and .
some clear features can be inferred from our data despite the heterogeneous nature of the sample .
all of our targets show evidence of br@xmath2 emission .
the measured @xmath68 ranges between 100 and 200 kms@xmath5 , which is consistent with stellar wind or disc - wind velocities @xcite . in v645
cyg , the p - cygni profile is detected in a velocity range between @xmath6 kms@xmath5 and @xmath7 kms@xmath5 .
this is consistent with a spherical wind . in the br@xmath2 emission images ,
the appearance is similar to those of the @xmath1-band continuum images in all objects , suggesting that these features are seen in scattering by dust in the circumstellar disc or envelope .
there is a clear spectroastrometric signature in both s106 and i18151 , running perpendicular to the known larger scale outflow axis in both cases . at least for s106 ,
the orientation is consistent with the disc - wind picture previously developed for this source , and we conclude the same should be true for i18151 as well .
this is the opposite of what d10 saw in w33a , where the best explanation for the spectroastrometric signature was an ionised bipolar jet . in the two cases presented here
it would suggest that the signature arises in ionised gas in , or from , a disc around the star . in the case of the well
studied s106 , the orientation of our spectroastrometric disc is also consistent with the morphology of the radio emission observed by @xcite .
the presence of a stellar wind in this object is clear from the strong hei absorption , lending support to the disc - wind picture of @xcite .
the estimates that are derived for the stellar masses from the spectroastrometry assuming keplerian orbits , although clearly imprecise due to the scatter in the data as shown in fig.[bg_spcast ] , and strictly lower limits , are consistent with the known properties of these sources .
the evidence that these two sources contain discs is therefore strong .
the other objects show spectroastrometric data consistent with a null result in most cases , revealing only a random - walk like pattern , though afgl 2136 may show evidence for a weak hii region near the systemic velocity .
i19065 shows a larger scatter in its random - walk pattern than the other non - detections .
this may be due to `` contamination '' of the stellar spectrum due to the extended nebular hii region emission .
the lack of any signature in v645 cyg , given the excellent signal - to - noise for that source , places a strong constraint on what we see there , consistent with the p - cygni wind profile observed in the line .
evidence for a disc from co bandhead emission was present in i18151 , s106 and v645 cyg .
the disc parameters for i18151 and s106 are consistent with those found from our analysis of the spectroastrometry of br@xmath2 .
the edge - on nature of s106 is clearly seen in these data as well .
v645 cyg shows evidence for a more pole - on orientation , which may be why we see no evidence for a disc - wind component in this source in the spectroastrometry . it may also simply be that v645 cyg is relatively more evolved than the other sources . finally , in i18151 and afgl 2136
, a series of co absorption lines are detected .
if we assume that these lines originate from the co gas disc and the velocity structures reflect the keplerian rotating motion , the central masses can be estimated .
our data show a separation of the flux centroids between the blueshifted and redshifted components .
this orientation is nearly perpendicular to the jet direction in i18151 , giving a strong evidence for a rotating motion in our data , whereas that orientation is less correlated with the direction of the outflow or the nir polarization disc in afgl 2136 .
nevertheless , the rotation velocities and the central masses are estimated to be @xmath69 kms@xmath5 and 30 @xmath11 for i18151 and @xmath70 kms@xmath5 20 @xmath11 for afgl 2136 .
our work has demonstrated that three - dimensional spectroastrometry across spectral emission and absorption lines is a powerful diagnostic technique to study the geometrical and kinematic properties of ionized or molecular regions in vicinity of the central star .
the ionized hydrogen traces the mass accretion channel inside the accretion disc and disc winds , which is driven by the central star . on the other hand ,
the molecular gas probes the keplerian rotating discs .
the consistency of the results obtained using different methods strongly supports the hypothesis that massive stars form by mass accretion via circumstellar discs .
based on observations ( proposal i d : gn-2011a - q-56 ) obtained at the gemini observatory , which is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the science and technology facilities council ( united kingdom ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministrio da cincia , tecnologia e inovao ( brazil ) and ministerio de ciencia , tecnologa e innovacin productiva ( argentina ) .
we thank the referee , dr . scott wolk , for his helpful comments .
i18151 ( @xmath71 kpc ) has been studied well , having featured in a large number of survey observations of myso candidates @xcite .
images at the msx 8 @xmath13 m @xcite and 1.2 mm continuum @xcite wavelengths revealed four clumps mm14 spreading across @xmath53@xmath72 scale .
irs 1 is located at mm1 , which we observed .
@xcite performed a radiative transfer modeling of the dust disc of this object and obtained a luminosity of @xmath73 .
however , the spectral energy distribution was not well constrained in the far infrared in their model .
unpublished herschel pacs data at 70@xmath13 m reveals a source with flux @xmath74jy .
@xcite scaled all `` large beam '' flux down by a factor of four to match their mm interferometry .
if we adopt the same correction , their best fit spectral energy distribution lies a factor of @xmath75 too low .
the herschel data is only about 10% smaller than the published large beam iras flux we used to derive a luminosity of @xmath76 for the rms database .
it seems likely the luminosity is actually @xmath77 given these remarks , correspondingq to a @xmath51 star @xcite .
@xcite also found a huge disc of 220 @xmath11 mass and 5000 au radius from the mm interferometry . only the inner 30 au satisfy the toomre criterion , a condition of gravitational stability @xcite .
hence , most of the mass of their modeled disc would include the outer envelope .
our results are consistent with the luminosity and mass for the central star implied by these earlier results .
the disc orientation we find in br@xmath2 is east - west , whereas the large scale h@xmath3 jet traced by d04 lies nw se .
the co absorption agrees with the latter orientation .
this suggests that the inner disc lies at an angle to the larger scale envelope structure .
although a couple of stars or cores are often found in a high - mass star formation region , there is no other source within @xmath78 from afgl 2136 in the mid - infrared and submillimeter wavelength ranges @xcite .
the near - infrared images show three features with extensions of a few arcsec and the central proto - star is located at irs 1 @xcite .
the co emission line observations show a bipolar outflow extending at a position angle of @xmath53135@xmath46 and the se and nw components are blueshifted ( 14 to 20 kms@xmath5 ) and redshifted ( 24 to 30 kms@xmath5 ) respectively .
the orientation of the polarization disc in the nir data is found to be @xmath5345@xmath46 , approximately perpendicular to the co outflow direction @xcite .
these results indicate that the orientation of the equatorial plane lies at a position angle of @xmath5345@xmath46 .
the results of radiative transfer modeling suggest that the viewing angle is about @xmath50 , i.e.seen near edge - on @xcite .
@xcite found that a situation requiring both a gaseous disc and dusty envelope is needed to explain the observed mid - infrared interferometry .
the observed baseline for their data lay along the position angle of the polarization disk .
@xcite detected both water and class ii methanol maser spots about 1 arcsec west with respect to irs 1 .
these masers are usually tracers of outflows , but they do not align with the co outflow presented by @xcite . instead they lie perpendicular to the plane of the co absorption feature we find .
the strongest water maser feature found by @xcite is redshifted with respect to the systemic velocity .
our more recent water maser observations using the gbt @xcite shows that this feature has weakened considerably , with the strongest component now arising much closer to the systemic velocity .
@xcite concluded that this may be an example where the water maser arises in the infalling gas , in the post - shock gas just behind the accretion shock .
it may be that the co absorption disc is evidence for larger accretion flow that is `` twisted '' relative to the small scale gaseous disk suggest by @xcite , giving rise to a stronger shock than would otherwise be the case , and reconciling the position angles of the small scale disc and larger scale absorbing gas .
further observations are clearly required to fully resolve this puzzle .
the inner circumstellar region of s106 has been studied by several authors ( e.g. * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the dark lane is thought to be an edge - on massive disc @xcite .
@xcite found a rotation axis of @xmath53@xmath79 , which is perpendicular to the dark lane .
the radio spectrum at 22 ghz shows a flux , @xmath80 , proportional to @xmath81 , suggesting an ionized stellar wind @xcite .
the wind has a high - mass loss rate of @xmath82yr@xmath5 @xcite .
this is too high for ob type main sequence stars , indicating a young stellar object .
the disc plane is well defined by the radio observations of @xcite to be @xmath83 .
line observations have shown strong p - cygni profiles in hydrogen recombination and @xcite and double peaks in br12 and @xcite .
these features can not be explained with an accretion flow , but suggest emission from stellar@xmath19disc winds or an accelerating outflow .
this can be see as consistent with the fact that the radio emission lies in the disc plane and not perpendicular to it if the source of both is a disc - wind type geometry @xcite .
our own br@xmath2 spectroastrometry is consistent with these previous findings , and the inferred stellar mass agrees with the observed luminosity .
the physical properties of the central star of v645 cyg has been frequently studied since its identification @xcite .
it has a luminosity of about @xmath84 @xcite .
the near - infrared imaging polarimetry shows a centro - symmetric pattern surrounding n0 nebula @xcite .
some h@xmath3o maser spots are detected exactly toward the n0 @xcite .
the spectral type of the central star was estimated to be o7e @xcite .
on the other hand , @xcite and @xcite detected the p - cygni profiles in , , , and , suggesting herbig ae@xmath19be stars .
recently , @xcite assessed the effective temperature from the optical spectroscopy .
they suggest that @xmath85 is about 25000 k , but that the star is probably extended .
given the variability in the spectrum , it is not inconceivable that this result is in agreement with the earlier result of @xcite , who found a temperature nearer 30000k , and a type near o9ve as we do here ( which is also more consistent with the known luminosity ) .
the lack of heii emission as clearly pointed out by @xcite is not a problem with this classification .
the p - cygni profile is detected in br@xmath2 ( this work and * ? ? ?
the blueshifted absorption component ranges up to @xmath6 kms@xmath5 ( cf.@xmath65 kms@xmath5 , * ? ? ?
a similar profile is detected in w33a ( d10 ) .
our spectroastrometry of br@xmath2 shows no significant signal .
we suggest that this is due to the fact that br@xmath2 arises in a spherically outflowing wind in this source , which essentially has a null spectroastrometric signature .
this also naturally explains the strong p - cygni profile .
we previously concluded that w33a has an asymmetric wind by comparison .
this suggests that v645 cyg is actually nearer to its final main sequence configuration .
our fit to the co bandhead also suggests that v645 cyg must be close to pole - on .
these objects are identified as mysos by the rms survey @xcite and this work for the first time reports their individual study .
i19065 is a weak hii region , with a small cluster of stars .
the brightest of those seen at @xmath1 may be the most important ionizing source . in i19065 , the line at 2.058 @xmath13
m is detected in emission , as is br@xmath2 .
the width of the lines suggest an origin partly from the star as well as the hii region .
the estimated bolometric luminosity is 1.3@xmath5410@xmath55 @xcite .
we find good agreement between this luminosity and a possible b1ve classification from the properties of the hei 2.058 @xmath13 m line .
g082 has no special characteristic besides the detections of the br@xmath2 and emission lines in our data .
taking into account the estimated bolometric luminosity of 6.3@xmath5410@xmath86 , this object is probably a mid - b star .
mcgregor , p. j. , hart , j. , conroy , peter g. et al . in iye m. , moorwood a. f. m. , eds , proc .
spie conf .
4841 , instrument design and performance for optical@xmath19infrared ground tested telescopes .
spie , bellingham , 4841 , 1581 | we present medium resolution ( @xmath0 ) @xmath1-band integral field spectroscopy of six massive young stellar objects ( mysos ) .
the targets are selected from the red msx source ( rms ) survey , and we used the altair adaptive optics assisted near - infrared integral field spectrometer ( nifs ) mounted on the gemini north telescope .
the data show various spectral line features including br@xmath2 , co , h@xmath3 , and .
the br@xmath2 line is detected in emission in all objects with @xmath4 200 kms@xmath5 .
v645 cyg shows a high - velocity p - cygni profile between @xmath6 kms@xmath5 and @xmath7 kms@xmath5 .
we performed three - dimensional spectroastrometry to diagnose the circumstellar environment in the vicinity of the central stars using the br@xmath2 line .
we measured the centroids of the velocity components with sub - mas precision .
the centroids allow us to discriminate the blueshifted and redshifted components in a roughly east west direction in both iras 181511208 and s106 in br@xmath2 .
this lies almost perpendicular to observed larger scale outflows .
we conclude , given the widths of the lines and the orientation of the spectroastrometric signature , that our results trace a disc wind in both iras 181511208 and s106 .
the co @xmath8 absorption lines at low @xmath9 transitions are detected in iras 181511208 and afgl 2136 .
we analysed the velocity structure of the neutral gas discs , which we find to have nearly keplerian motions . in iras 181511208 , the absorption centroids of the blueshifted and redshifted components are separated in a direction of north - east to south - west , nearly perpendicular to that of the larger scale @xmath10 jet .
the position - velocity relations of these objects can be reproduced with central masses of 30 @xmath11 for iras 181511208 and 20 @xmath11 for afgl 2136 .
we also detect co @xmath8 bandhead emission in iras 181511208 , s106 and v645 cyg .
the results can be fitted reasonably with a keplerian rotation model , with masses of 15 , 20 and 20 @xmath11 respectively .
these results for a sample of mysos can be explained with disc and outflow models and support the hypothesis of massive star formation via mass accretion through discs as is the case for lower mass counterparts .
[ firstpage ] ism : region ism : individual : iras 181511208 , afgl 2136 , iras 19065 + 0526 , s106 irs4 , g082.5682 + 00.4040 , and v645 cyg . |
quantum mechanics on metric graphs is a subject with a long history which can be traced back to the paper of ruedenberg and scherr @xcite on spectra of aromatic carbohydrate molecules elaborating an idea of l. pauling .
a new impetus came in the eighties from the need to describe semiconductor graph - type structures , cf .
@xcite , and the interest to these problems driven both by mathematical curiosity and practical applications is steadily growing ; we refer to @xcite or the proceedings @xcite for a bibliography to the subject .
since quantum graphs are supposed to model various real graph - like structures with the transverse size which is small but non - zero , one has to ask naturally how close are such system to an `` ideal '' graph in the limit of zero thickness .
this problem is not easy and a reasonable complete answer is known in case of `` fat graphs '' with neumann boundary conditions and similar systems .
a pioneering work in this area was done by freidlin and wentzell @xcite and the papers @xcite and @xcite can be mentioned as important milestones .
we managed to contribute to this problem in a series of papers , @xcite , @xcite and @xcite , in which we improved the approximation using the intrinsic geometry of the manifold only , demonstrating the norm resolvent convergence , and finally extending the approximation also to resonances by means of complex scaling .
while these results provide in our opinion a solid insight into the neumann - type situation , we must acknowledge as the authors that the three papers are long and rather technical , and some may find them not easy to read .
this motivated us to write the present survey in which we intend to describe this family of approximation results without switching in the heavy machinery ; let the reader judge whether we have succeeded . before proceeding let us mention that there is an encouraging recent progress in the more difficult dirichlet case , see @xcite ,
however , we will not discuss it here .
let us briefly describe the contents of the paper . in the next section
we describe the two basic objects of this paper , quantum graphs and graph - like manifolds ( cf .
figure [ fig : graph - mfd ] ) .
section [ sec : disc - spec ] is devoted to convergence of the discrete spectrum summarizing the main results of ref .
an extension to non - compact graphs and a resolvent convergence coming from @xcite is given in section [ sec : res - conv ] .
finally , in section [ sec : reson - conv ] describe the results of ref .
@xcite showing how the resonances on quantum graphs and graph - like manifolds approximate each other .
[ fig : graph - mfd ] ( 0,0 ) with one external edge , five internal edges and four vertices and the associated graph - like manifold @xmath0 , here with cross section manifold @xmath1.,title="fig : " ] ( 5172,3964)(532,-3415 ) ( 5344,-720 ) @xmath2 ( 4621,-991 ) @xmath3 ( 655,-171 ) @xmath4 ( 5704,-2645)@xmath5 ( 4714,-3231)@xmath6 ( 655,-2419 ) @xmath0
suppose that @xmath4 is a connected metric graph given by @xmath7 where @xmath8 is a usual graph , i.e. , @xmath9 denotes the set of vertices , @xmath10 denotes the set of edges , @xmath11 associates to each edge @xmath3 the pair @xmath12 of its initial and terminal point ( and therefore an orientation ) .
the space @xmath4 being a _ metric _ graph means that there is a _ length function _ @xmath13 } $ ] associating to each edge @xmath3 a length @xmath14 .
we often identify the edge @xmath3 with the interval @xmath15 and use @xmath16 as a coordinate .
in addition , we denote @xmath17 the lebesgue measure on @xmath3 . in this way
, @xmath4 becomes a topologically @xmath18-dimensional space with singularities at the vertices .
moreover , @xmath4 carries a natural metric by defining the distance of two points to be the length of the shortest path in @xmath19 joining these points .
we call an edge @xmath3 _ external _ iff @xmath20 and _ internal _ otherwise and we denote the corresponding sets by @xmath21 and @xmath22 . properly speaking ,
the map @xmath23 is only defined on @xmath24 . for external edges
, we do not consider the `` end point '' at infinity as a vertex of @xmath19 , i.e. , @xmath25 contains only the initial vertex @xmath26 .
for each vertex @xmath27 we set @xmath28 i.e. , @xmath29 consists of all edges starting ( @xmath30 ) resp . ending ( @xmath31 ) at @xmath2 and @xmath32
is their disjoint union .
note that the _ disjoint _ union is necessary in order to allow single - edge loops , i.e. , edges having the same initial and terminal point .
we adopt the following uniform bounds on the degree @xmath33 and the length function @xmath34 : [ eq : deg.len ] @xmath35 where @xmath36 and @xmath37 .
needless to say , both the assumptions are fulfilled if @xmath38 and @xmath39 are finite .
we consider the hilbert space defined naturally as the orthogonal sum , @xmath40 a metric graph @xmath4 becomes a _
graph when it is equipped with a self - adjoint operator @xmath41 in the hilbert space @xmath42 .
the latter is assumed to act as @xmath43 on each edge for @xmath44 , so a quantum graph is determined by fixing the _ vertex boundary conditions _ at each vertex , in order to turn the formal laplacian into a self - adjoint operator .
we will use the _ free _
( often called , not quite properly , _ kirchhoff _ ) boundary conditions : a function lies in the operator domain @xmath45 and the relations [ eq : kirchhoff ] @xmath46 are fulfilled for all @xmath27 where @xmath47 defines the _ inward _ derivative of @xmath48 at @xmath2 . under the assumptions ,
the operator @xmath41 is self - adjoint in @xmath49 ( cf .
@xcite ) . the quadratic form associated with the operator
@xmath41 is @xmath50 and can be expressed as @xmath51 let us pass to a model of a quantum network which we will consider , corresponding to the idea that the graph has a small , but non - zero thickness .
let @xmath0 be a @xmath52-dimensional connected manifold with metric @xmath53 ( cf .
figure [ fig : graph - mfd ] ) .
if @xmath0 has boundary , we denote it by @xmath54 ; let us stress that our discussion covers different kind of models , `` full '' fat graphs considered by kuchment and zeng @xcite where the boundary is present , as well as `` hollow '' or `` sleeve - type '' manifolds having no boundary .
we assume that @xmath0 can be decomposed into open sets @xmath6 and @xmath5 , i.e. , here and in the following , the expression @xmath55 means that the @xmath56 s are open ( in @xmath57 ) , mutually disjoint and the interior of @xmath58 equals @xmath57 ; recall that in an @xmath59-theory it is enough to have charts covering a set of full measure . ]
@xmath60 denote the metric on @xmath0 by @xmath61 .
we assume that @xmath6 and @xmath5 are isometric to [ eq : met ] @xmath62 where @xmath63 is a compact @xmath64-dimensional manifold with @xmath65 , and @xmath66 is an @xmath67-independent @xmath52-dimensional manifold shows a slightly different situation where the edge neighborhoods are shortened ( cf . remark [ rem : short.ed ] below ) . ] .
note that @xmath68 has @xmath69-many components isometric to @xmath70 denoted by @xmath71 for @xmath72 .
we sometimes write @xmath73 .
[ rem : collar ] for technical reasons ( cf .
lemma [ lem : cn ] ) we assume that near @xmath74 , the ( unscaled ) manifold @xmath75 has a collar neighborhood @xmath76 . if such a collar neighborhood happens to be too small , we can just change the decomposition in such a way that we add a cylinder of length @xmath77 ( the length taken in the edge coordinates @xmath78 on @xmath79 ) from the edge neighborhood to the vertex neighborhood ( becoming here a cylinder of length @xmath80 in the vertex coordinates on @xmath81 since @xmath82 ) .
the cross section manifold @xmath83 has a boundary or does not have one , depending on the analogous property of @xmath0 . for simplicity
, we suppose that @xmath84 . clearly , we have @xmath85 for the riemannian densities .
we consider the hilbert space @xmath86 and the laplacian @xmath87 ( with neumann boundary conditions if @xmath88 ) defined on @xmath89 the associated quadratic form @xmath90 can be expressed as @xmath91 where @xmath92 is the exterior derivative of @xmath93 .
note that the quadratic form expression contains only the metric @xmath61 , but no derivatives of @xmath61 .
[ rem : short.ed ] we have chosen the full edge length on the edge neighborhood
although this assumption is not valid if @xmath0 is the @xmath94-neighborhood of a metric graph @xmath4 _ embedded _ into a euclidean space from which it inherited its metric . in such a situation , however , the metric @xmath95 differs from the metric @xmath96 only by a small longitudinal error . using the fact that the laplacian on @xmath0 defined via its quadratic form depends only on @xmath61 ( and not on its derivatives ) , it can be shown that a small ( uniform ) perturbation of the product structure has only a small effect on the laplacian , its spectrum , etc .
in addition , we assume that the following uniformity conditions are valid , @xmath97 where @xmath98 denotes the second ( i.e. , first non - zero ) neumann eigenvalue of @xmath66 . roughly speaking , the requirements mean that the region @xmath81 remains small w.r.t .
the vertex index see the discussion in ( * ? ? ?
* rem . 2.7 ) ) for more details .
needless to say , these assumptions are trivially satisfied once the vertex set @xmath9 is finite .
recall also one more domain related to the laplacian for @xmath99 .
we denote by @xmath100 the scale of hilbert spaces associated with the self - adjoint , non - negative operator @xmath101 , together with its natural norm @xmath102 for negative exponents , we set @xmath103 note that @xmath104 can be again viewed as the completion of @xmath105 with respect to the appropriate norm , in this case @xmath106 .
let us start with a simple thing , a convergence result using a variational characterization of discrete eigenvalues .
we assume that @xmath4 is a compact metric graph , in other words , @xmath107 in this case , the graph - like manifold @xmath0 is also compact , and the spectrum of @xmath101 , @xmath99 , is purely discrete .
we denote by @xmath108 the @xmath109-th eigenvalue of @xmath101 , @xmath110 , repeated with respect to the multiplicity .
the min - max variational characterization is then @xmath111 where the infimum is taken over all @xmath109-dimensional subspaces @xmath112 of @xmath113 .
our main result in this section , coming from @xcite , is the following : [ thm : disc ] assume that the metric graph @xmath4 is compact , i.e. , the graph is finite and each edge has a finite length .
then the @xmath109-th ( neumann ) eigenvalue of the laplacian on the graph - like manifold @xmath108 converges to the @xmath109-th eigenvalue of the laplacian on the metric graph @xmath4 with the free boundary conditions . specifically , we have @xmath114 where the error term @xmath115 depends on the index @xmath109 .
the proof is based on an abstract comparison result . in order to compare the respective eigenvalues of the quantum graph and the graph - like manifold , we need identification maps expressed in terms of the quadratic forms , namely @xmath116 we have the following abstract eigenvalue comparison result ( see , e.g. , ( * ? ? ?
2.1 ) ) : assume that there are @xmath117 and @xmath118 such that [ eq : j1.est ] @xmath119 for all @xmath120 and @xmath121 .
then @xmath122 where the upper bound is @xmath123 depending on @xmath124 and the lower bound is @xmath125 depending additionally on @xmath126 . in our concrete example , the @xmath127 identification operator can be chosen as @xmath128 note that the definition makes sense since functions in @xmath129 are continuous .
for the map in the opposite direction , we first introduce the following averaging operators @xmath130 for @xmath131 .
what they yield is nothing else than the fourier coefficient corresponding to the first ( transverse ) eigenfunction @xmath132 on @xmath83 and @xmath133 on @xmath81 , respectively ; note that these eigenfunctions are constant and that @xmath134 .
we set @xmath135 for @xmath136 , where @xmath137 } $ ] is a smooth function with @xmath138 the interpolating contribution related to @xmath139 is needed in order to make @xmath140 continuous at each vertex .
the following lemma ensures that the error coming from this correction remains small : [ lem : cn ] we have @xmath141 for all @xmath121 and @xmath142 , where @xmath143 denotes the @xmath59-norm on @xmath5 .
due to remark [ rem : collar ] , each component @xmath144 of @xmath145 has a collar neighborhood @xmath146 of length @xmath80 ( in the unscaled coordinates of @xmath81 ) . the cauchy - schwarz inequality and the sobolev trace estimate ( see e.g. ( * ? ? ?
8) ) @xmath147 yield ( with @xmath148 in the collar coordinates , and then integration over @xmath149 ) @xmath150 where @xmath151 is considered to be a constant function on @xmath81 and @xmath152 is the @xmath59-norm on @xmath81 .
now @xmath153 is orthogonal to the first ( constant ) eigenfunction of the neumann laplacian on @xmath81 , and the min - max principle ensures that the squared norm of @xmath153 can be estimated by @xmath154 . using the scaling @xmath155 of the metric and , we obtain the desired estimate
. we also have to make sure that eigenfunctions @xmath93 of @xmath101 belonging to eigenvalues _ bounded _ with respect to @xmath67 , can not concentrate on the vertex neighborhoods : [ lem : vx ] we have @xmath156 for @xmath121 , where @xmath3 is any edge adjacent to the vertex @xmath2 .
the constant @xmath157 in this inequality depends only on @xmath158 , @xmath159 and @xmath160 .
we employ the estimate @xmath161 the first summand can be treated as in the previous proof , the second by lemma [ lem : cn ] , and the last one by a sobolev trace estimate on the _ edge _ neighborhood similar to , namely @xmath162 .
it remains to show that the conditions are fulfilled .
we do not give the details here referring to ( * ? ? ?
* sec . 5 ) .
the proof of is simple , and it works even with @xmath163 , hence one obtains a stronger estimate , @xmath164 . for the opposite inequality , we need to verify . to this end
we need lemma [ lem : cn ] in the norm and a quadratic form estimate .
the norm estimate uses in addition lemma [ lem : vx ] , and the estimate @xmath165 which follows from . for the quadratic form estimate we need
the simple cauchy - schwarz bound @xmath166 .
similar results can be obtained for more general situations when the vertex and edge neighborhoods scale at different rates , cf .
@xcite , and in certain situations also for the dirichlet laplacian , cf .
@xcite , where , however , the resulting graph operator is decoupled .
next we would like to go further and prove also results for non - compact graphs , and also convergence of eigenfunctions or resolvents . to do
so we need some more notation .
we write @xmath167 and @xmath168 for the @xmath67-dependent spaces , @xmath169 .
we stress that the parameter @xmath67 enters only through the quantity @xmath170 and one can interpret it as a label for the second hilbert space involved see also the appendix in @xcite for the concept of a `` distance '' between two hilbert spaces and associated non - negative operators . for brevity
, we set @xmath171 [ def : quasi ] we say that an operator @xmath172 is _ @xmath173-quasi - unitary _ with respect to @xmath174 , iff @xmath175 , @xmath176 , and @xmath177 where @xmath178 is the identity on @xmath167 , and @xmath179 is the operator norm of @xmath180 . in our particular situation , we will employ the quasi - unitary operator @xmath181 where @xmath182 is the lowest normalized eigenfunction on @xmath183 and in turn @xmath184 is the zero function on @xmath81 .
the quasi - unitarity is stated in the following lemma : [ lem : j.quasi ] the map @xmath185 defined by ( [ eq : j ] ) is @xmath186-quasi - unitary , where @xmath187 depends only on @xmath158 , @xmath160 and @xmath159 .
a simple calculation shows that @xmath175 , @xmath176 , and that @xmath188 the function @xmath189 is orthogonal to the constant function on @xmath83 , and by the min - max principle we infer that @xmath190 where @xmath191 is the first non - zero ( neumann ) eigenvalue on @xmath83 , and @xmath192 is the derivative with respect to the transverse variable(s ) .
the estimate of the sum over the vertex contributions follows from lemma [ lem : vx ] .
we also need a tool to compare the laplacians on @xmath49 and @xmath167 .
to this end we put : we say that @xmath174 and @xmath193 are _
@xmath173-close _ w.r.t .
the map @xmath172 iff @xmath194 where @xmath195 denotes the operator norm of @xmath196 .
note that a @xmath197-quasi - unitary map is indeed unitary .
furthermore , if @xmath193 and @xmath174 are @xmath197-close with respect to a @xmath197-quasi - unitary map @xmath185 , then @xmath193 and @xmath174 are unitarily equivalent . in this sense
, the concept of quasi - unitarity and closeness provides a quantitative way to measure how far a pair of operators is from being unitarily equivalent . in order to show that the operators @xmath174 and @xmath193 are
@xmath173-close , it is often easier to deal with the respective quadratic form domains as we have already done when demonstrating the convergence of the discrete spectrum .
we thus want to compare the identification operators on the scale of order @xmath18 with the quasi - unitary map @xmath185 : we say that the identification maps are _
@xmath173-compatible _ with the map
@xmath172 iff @xmath198 by means of an adjoint we obtain from @xmath199 a natural map @xmath200 now it is easy to derive the following criterion for @xmath173-closeness : [ lem : j.comm1 ] assume that @xmath201 and @xmath199 are @xmath173-compatible w.r.t .
the map @xmath185 , and that @xmath202 then @xmath174 and @xmath193 are @xmath203-close with respect to @xmath185 .
we first check that the identification maps @xmath201 and @xmath199 are indeed @xmath115-compatible with respect to the map @xmath185 : [ lem : j.comp ] the maps @xmath201 and @xmath199 as defined in and are @xmath186-compatible with @xmath185 , where @xmath187 depends only on @xmath158 , @xmath204 , @xmath160 and @xmath159
. we have @xmath205 by a standard sobolev estimate
see , e.g. , estimate or ( * ? ? ?
2.4 ) ) we can estimate the latter sum by @xmath206 .
for the other identification operator we have @xmath207 using now lemma [ lem : cn ] and reordering the sum , we obtain the additional factor @xmath204 , the maximum degree of a vertex , and the second estimate follows as well .
next , we will now indicate briefly how to prove the closeness of the laplacians : [ lem : j.comm ] the laplacians @xmath174 and @xmath193 are @xmath186-close with respect to the map @xmath185 defined in where @xmath187 depends only on @xmath158 , @xmath204 and @xmath159 .
we check the condition of lemma [ lem : j.comm1 ] which reduces to estimating @xmath208 in terms of @xmath209 ; the claim follows from lemma [ lem : cn ] and cauchy - schwarz .
putting together the previous results , we come to the following conclusion : [ thm : res ] adopt the uniformity conditions and
. then the laplacians @xmath101 and @xmath41 are @xmath115-close with respect to the quasi - unitary map @xmath185 defined in , i.e. @xmath210 where the error term depends only on @xmath158 , @xmath204 , @xmath160 and @xmath159 .
in addition , we have @xmath211 the first estimate follows from lemmata [ lem : j.comm1][lem : j.comm ] ; the second one in turn is a consequence of the first estimate and lemma [ lem : j.quasi ] . one can now develop the standard functional calculus for the laplacians @xmath174 and @xmath193 , and deduce estimates similar to the ones in theorem [ thm : res ] , but with the resolvent replaced by more general functions @xmath212 of the laplacians . specifically , @xmath213 need to be measurable , continuous in a neighborhood of the spectrum of @xmath193 , and the limit at infinity must exist . for example , one can control the heat operators via @xmath214 or the spectral projectors via @xmath215 . a proof of the following result
can be found in the appendices of the paper @xcite , see also @xcite : [ thm : res.spec ] under the assumptions of the previous theorem , we have @xmath216 for the spectral projections provided @xmath217 is a compact interval such that @xmath218 . in particular , if @xmath217 contains a single eigenvalue @xmath219 of @xmath41 with multiplicity one corresponding to an eigenfunction @xmath220 , then there is an eigenvalue @xmath221 and an eigenfunction @xmath222 of @xmath101 such that @xmath223 in addition , the spectra converge uniformly on @xmath224 $ ] , i.e. @xmath225 \to { \sigma_{\mathrm{{}}}({\delta_{{x_0}}^ { { { } } } } ) } \cap [ 0,\lambda]\ ] ] in the sense of hausdorff distance on compact subsets of @xmath224 $ ] .
the same result is true if we consider only the essential or the discrete spectral components .
naturally , the above stated spectral convergence reduces to the claim of theorem [ thm : disc ] in the situation when the spectra are purely discrete .
in the final section we will deal with the convergence of resonances in the present setting .
it is useful to include into the considerations also eigenvalues embedded in the continuous spectrum , because it may happen that resonances of a `` fat graph '' converge to such an eigenvalue , as it can be seen , e.g. , in a simple motivating example of the metric graph consisting of a single loop with a half - line `` lead '' attached @xcite .
a standard and successful method of dealing with resonances is based on the concept of _ complex scaling _
, often an _ exterior _ one .
the method has its roots in the seminal papers @xcite and a lot of work was devoted to it ; we refer to @xcite for a sample bibliography .
the main virtue is that it allows to reformulate treatment of resonances , i.e. poles of the analytically continued resolvent , and embedded eigenvalues , to analysis of discrete eigenvalues of a suitable _ non - selfadjoint _ operator . as we will see below the complex - scaling approach suits perfectly , in particular , to our convergence analysis . in this section
, we assume that the metric graph is finite , but non - compact , i.e. @xmath226 which means , in particular , that the assumptions and are satisfied .
we decompose the metric graph @xmath4 and the graph - like manifold @xmath0 into an _ interior _ and _ exterior _ part @xmath227 and @xmath228 , respectively . for technical reasons ,
it is easier to do the cut not at the initial vertices @xmath25 of an external edge @xmath229 , but at a fixed distance , say one , from @xmath25 along @xmath3 , and similarly for the graph - like manifold .
we therefore consider the _ internal _ metric graph @xmath230 consisting of all vertices , all edges of finite length and the edge parts @xmath231 for each external edge .
the exterior metric graph @xmath232 is just the disjoint union of @xmath233-many copies of a half - line @xmath234 , and we use the corresponding parametrization on an external edge . in other words , we do not regard the _ boundary points _ @xmath235 as vertices . similarly , let @xmath236 be the common boundary of @xmath227 and @xmath237 ; note that @xmath236 is isometric to @xmath233-many copies of @xmath70 .
now we introduce the exterior dilation operator .
for @xmath238 we define by @xmath239 one - parameter unitary groups on @xmath240 and @xmath241 , respectively , acting non - trivially on the _ external _ part only .
we call the operator @xmath242 for @xmath99 the _ dilated _ laplacian on @xmath0 with the domain @xmath243 for _ real _ @xmath244 .
a simple calculation shows that @xmath245 holds for internal edges and that @xmath246 is true for external edges with the domain @xmath247 given by @xmath248 here , @xmath249 are the functions from @xmath250 restricted to @xmath251 , i.e. without any condition at @xmath236 .
in addition , @xmath252 is defined as the longitudinal derivative on the common boundary @xmath236 oriented _ away _ from the internal part .
the expression of @xmath253 now can be generalized to _ complex _
@xmath244 in the strip @xmath254 where @xmath255 ; we call the corresponding @xmath253 the _ complex dilated _ laplacian .
the operators @xmath256 form a family with spectrum contained in the common sector @xmath257 moreover , we can determine the essential spectrum coming from the external part .
note that the branches of the essential spectrum associated with the higher transverse eigenvalues @xmath258 , on the graph - like manifold all vanish as @xmath259 .
[ lem : ess.sp ] let the metric graph be finite and non - compact , i.e. is fulfilled .
then @xmath260 in particular , since @xmath261 , we have @xmath262 for any bounded set @xmath263 provided @xmath169 is small enough .
in addition , we have the following important result : [ thm : analytic ] for @xmath264 not contained in the @xmath265-sector @xmath266 , the resolvents @xmath267 depend analytically on @xmath268 .
this is a highly non - trivial fact since @xmath253 is neither of type a nor of type b , in other words , both the sesquilinear form and the operator domains depend on @xmath244 even for real @xmath244 . to put it differently
, the ( non - smooth ) exterior scaling as defined here is a very singular perturbation of the operator @xmath269 .
the main idea is to compare @xmath270 with the resolvent of a _ decoupled _ operator , where one imposes dirichlet boundary conditions at @xmath236 .
an inspiration for such an idea was used in @xcite ; for a full proof in our situation we refer to @xcite . as a consequence
, we have the following result on the discrete spectrum . see ( * ? ? ?
5.8 ) for more details . ]
[ lem : disc ] the discrete spectrum of @xmath253 is locally constant in @xmath244 . as a consequence ,
discrete ( complex ) eigenvalues are `` revealed '' if @xmath271 is positive and large enough .
the same is true for eigenvalues embedded into the continuous spectrum @xmath234 of the laplacian ; in this case it is sufficient to have @xmath272 .
this is crucial for the above mentioned reformulation .
recall that by the most common definition a _ resonance _ is a pole in a meromorphic continuation of the resolvent over the cut corresponding to the essential spectrum into the `` unphysical sheet '' of the riemann energy surface @xcite .
rotating the essential spectrum one can reveal these singularities ; this allows us to identify a _ resonance _ of @xmath273 with a complex @xmath59-eigenvalue of the dilated operator @xmath253 for @xmath271 large enough .
notice also that such a definition is consistent : it does not depend on where we cut the spaces into an interior and exterior part ; of course , as far as the interior part remains compact . in order to demonstrate convergence properties of resonances ,
we need to introduce a scale of hilbert spaces associated to the non - selfadjoint operator @xmath253 .
in particular , we set @xmath274 with the norm @xmath275 and @xmath276 with the dual norm ; for details we refer to ( * ? ? ?
we say that an operator @xmath172 is _ @xmath173-quasi - unitary _ with respect to @xmath253 , iff @xmath175 , @xmath176 , and @xmath277 where @xmath278 and @xmath279 is the operator norm of @xmath280 .
it is much easier to use the @xmath173-quasi - unitarity with respect to the non - dilated operator ( via its quadratic form ) as in definition [ def : quasi ] . in particular
, we would like to compare the non - dilated scale @xmath281 of order @xmath18 and the dilated scale @xmath282 of order @xmath283 : @xmath174 and @xmath253 are _ compatible _ if there is a family of bounded , invertible operators @xmath284 on @xmath167 such that @xmath285 , @xmath286 and @xmath287 .
given @xmath99 , we say that @xmath174 and @xmath253 are _ uniformly _ compatible with respect to @xmath67 , if there is a constant @xmath288 , _ independent _ of @xmath67 , such that @xmath289 in the situation we consider there is a natural candidate for @xmath290 , namely @xmath291 in contrast to @xmath292 , the operators @xmath290 are defined also for _ complex _ values of @xmath244 . again , as for the analyticity , the proof of uniform compatibility in our example needs some technical preliminaries which we skip here .
in essence , one needs to define the resolvent @xmath293 as a bounded operator from @xmath294 to @xmath295 with @xmath67-independent bound , where @xmath296 is an appropriate space of order @xmath18 for @xmath297 , not necessarily related to @xmath298 .
[ lem : comp.dil ] for a given @xmath169 , the operators @xmath101 and its complex dilated counterparts @xmath297 are ( uniformly ) compatible . in particular , the map @xmath185 as defined in is @xmath115-quasi - unitary with respect to @xmath297 , where @xmath115 depends on @xmath244 . the compatibility is demonstrated in ( * ? ? ?
the second assertion follows from @xmath299 where we employed the fact that @xmath300 and lemma [ lem : j.quasi ] .
now we are in position to state our first convergence result of this section : [ thm : res.dil ] assume that the metric graph @xmath4 is finite and non - compact , cf . .
then the complex dilated laplacians @xmath301 and @xmath297 are @xmath115-close , i.e. @xmath302 where @xmath115 depends on @xmath244 .
the proof is essentially the same as for the non - dilated case .
first we define operators on the scales of order one , specifically @xmath303 in exactly the same way as in and .
recall that we do not consider the boundary points between the internal and external parts as vertices , i.e. the graph - like manifold does not have a vertex neighborhood there . hence @xmath304 and then we have @xmath305 r_0^\theta\ ] ] where @xmath306 .
the last difference at the right - hand side can be estimated by @xmath307 using lemma [ lem : j.comp ] and lemma [ lem : comp.dil ] , and the first one can be treated similarly . for the remaining term
we observe that in order to prove @xmath308 it suffices to show @xmath309 for @xmath310 and @xmath311 .
from the compatibility , we obtain @xmath312 and similarly for @xmath313 ; in particular , we can choose @xmath314 , however , the estimate is almost the same as in the non - dilated case given in lemma [ lem : j.comm ] . as in the non - dilated case , on can develop a functional calculus for the pairs of operators @xmath315 and @xmath253 cf .
since now the operators are not self - ajoint , we only have a _ holomorphic _ functional calculus .
in particular , we can show @xmath316 for the spectral projections , provided @xmath317 is an open disc containing a single discrete eigenvalue @xmath219 of @xmath301
. from here our main result on resonances follows : [ thm : resonances ] assume that the metric graph @xmath4 is finite and non - compact , cf . .
if @xmath219 is a resonance of the laplacian @xmath41 with a multiplicity @xmath318 , then for a sufficiently small @xmath169 there exist @xmath64 resonances @xmath319 of @xmath101 , satisfying @xmath320 and not necessarily mutually different , which all converge to @xmath219 as @xmath259 .
the same is true in the case when @xmath219 is an embedded eigenvalue of @xmath41 , except that then only @xmath321 holds in general .
finally , if the multiplicity of @xmath219 is one with a normalized eigenfunction @xmath322 ( corresponding to a _ resonance _ or _ embedded eigenvalue _ for @xmath41 ) , then there exists a normalized eigenfunction @xmath323 ( related to the respective entity for @xmath101 ) on the graph - like manifold ) such that @xmath324
the first author acknowledges a partial support by gaas and meys of the czech republic under projects a100480501 and lc06002 , the second one by dfg under the grant po-1034/1 - 1 . , _ asymptotics of spectra of neumann laplacians in thin domains _ , advances in differential equations and mathematical physics ( birmingham , al , 2002 ) , contemp . math . , vol .
327 , amer .
soc . , providence , ri , 2003 , pp . | quantum networks are often modelled using schrdinger operators on metric graphs .
to give meaning to such models one has to know how to interpret the boundary conditions which match the wave functions at the graph vertices . in this article
we give a survey , technically not too heavy , of several recent results which serve this purpose .
specifically , we consider approximations by means of `` fat graphs '' in other words , suitable families of shrinking manifolds and discuss convergence of the spectra and resonances in such a setting . |
the moa project * m*icrolensing * o*bservations in * a*strophysics is a collaboration of about 30 astronomers / physicists in new zealand and japan working in about 10 different institutions ( headquarters nagoya and auckland universities ) .
the project was established in 1995 .
the moa project goals are ( a ) the search for dark matter by microlensing in the galactic halo , ( b ) the search for planets by microlensing , in halo and galactic bulge , ( c ) the study of variable stars in the magellanic clouds and bulge and ( d ) the optical identification of @xmath9-ray bursters ( grbs ) the observational program is undertaken on the 61-cm boller & chivens telescope at the university of canterbury s mt john university observatory in the centre of the south island of new zealand . here
there is a dark sky background and typically 13 arcsec seeing ; between a third and a half of the weather is suitable for ccd photometry . at the latitude of @xmath0 s , the magellanic clouds are circumpolar and
can be observed all year .
the telescope is a ritchey - chrtien cassegrain instrument modified to accommodate wide - field f/6.25 optics and a modern computer control system .
three series of moa observations have been undertaken as follows 1 .
jan 1996 dec 1996 f/13.5 cass .
@xmath10 , moacami 2 .
jan 1997 aug 1998 f/6.25 cass .
@xmath11 , moacami 3 .
aug 1998 present f/6.25 cass .
@xmath12 , moacamii in all cases , very broad - band @xmath13 and @xmath14 filters have been used , covering respectively about 395620 nm ( @xmath15 nm ) and 6201050 nm ( @xmath16 nm ) ( see fig .
1 ) . series 1 used the f/13.5 optics and encompassed @xmath17 stars in three lmc fields .
series 2 covered three fields ( nlmc1,2,3 ) in the lmc bar ( @xmath18 stars ) plus two smc fields ( smc1,2 ) ( @xmath19 stars ) . for series 3 , which is currently in progress , the new moacamii is used to observe 16 lmc fields ( all year ) ( @xmath20 stars ) plus 8 smc fields ( all year ) ( @xmath21 stars ) plus 16 galactic bulge fields ( winter only)(@xmath22 stars ) .
each of the moacamii fields covers about 1.25 square degrees .
exposure times are 5 min . in @xmath13 and @xmath14 in the clouds , 3 minutes in the galactic bulge .
the ccd camera moacami comprises a mosaic of nine ti tc215 @xmath23 pixel ccds whose pixel size is 12 @xmath24 m ( @xmath25 at f/13.5 , @xmath26 at f/6.25 ) .
the chips are not butted , so four consecutive interleaved exposures are required to cover a @xmath11 field .
the recently acquired moacamii has now replaced the first camera .
it consists of three butted site @xmath27-pixel thinned , back - illuminated ccds with 15-@xmath24 m pixels ( @xmath28 at f/6.25 ) .
the peak qe is 85% .
the area of the ccd surface is @xmath29 cm , and there are 24 mpixels . both cameras operated at a temperature of @xmath30 c with liquid n@xmath31 .
generally we expose each magellanic cloud field once or sometimes twice per night in both @xmath13 and @xmath14 , and each galactic bulge field about four times per night in @xmath14 and once in @xmath13 , unless a known high magnification microlensing event is in progress , in which case frequent successive exposures are made .
moa observations of ccd images are recorded to 8-mm exabyte tape for sending to japan , while an archival copy on dlt tape is retained at mt john . for series 1 and 2 observations , all the reductions
have been completed using the dophot software package .
one of us ( i.b . ) completed an archived database of all series 1 and 2 frames giving magnitudes in @xmath13 and @xmath14 . in the case of series 2 images ,
there are some 200 data points in each colour and about @xmath32 stars .
series 3 images using moacamii have hitherto not been reduced .
i. bond is currently perfecting implementation of the alard & lupton ( 1997 ) difference imaging technique , which already has shown that a substantial improvement in photometric precision is possible for crowded fields in only moderate seeing .
it is hoped to reduce series 3 images early in 2000 and then to be able to reduce all images on site at the observatory in nearly real time .
a sun enterprise 450 computer with 90 gbyte hard disk drive is used for on - site data analysis at mt john .
mt john @xmath33 moa photometry gives colours which transform linearly to johnson colours , as calibrated by reid , dodd & sullivan ( 1997 ) , as follows : @xmath34 @xmath35 the calibration of moa magnitudes gives the following transformation from the moa instrumental system to that of the hst guide star catalogue .
@xmath36 @xmath37 the typical precision of the series 2 moa magnitudes is @xmath38 mag . at @xmath13 or @xmath39 ; @xmath40 mag . at @xmath13 or @xmath41 15 .
three surveys on variable stars in the moa database have been undertaken or are in progress .
one of these has been by s. noda ( nagoya univ . ,
msc ) and m. takeuti ( tohoku univ . ,
sendai ) , and is a study of variables in series 2 observations .
the stars are selected by several criteria ( see noda et al . in these proceedings ) .
periods have been obtained by the phase dispersion method ( pdm ) .
important aspects of this study include : * study of red variables mainly sr ; 188 variables in the lmc , 35 in the smc * study of eclipsing binaries ; 69 eb in the lmc , 136 in the smc * study of cepheids ; 78 in the lmc , 38 in the smc * study of blue variables ( many detected ) ; may be be stars , at least one is a so - called bumper , resembling a microlensing light curve by using another selection , 282 agb variables were found in the lmc and 69 in the smc ( see takeuti et al . in these proceedings ) .
the periods and the @xmath42-band magnitudes of most of these stars are comparable with those in the literaure .
the relation between the period and the mean colour for the agb variables is reported .
another study has been by n. rattenbury ( auckland univ .
, msc ) , using selected moa fields for series 1 and 2 data .
variable stars were detected by deriving the welch - stetson variability index ( stetson , 1996 ; welch & stetson , 1993 ) . only about 2% of the stars with greatest variability
were selected for further study in this preliminary survey .
typically these stars have a range in the @xmath14 magnitude exceeding @xmath43 .
periods were analysed in several ways , the most successful algorithm being the date compensated discrete fourier transform ( dcdft ) algorithm of ferraz - mello ( 1981 ) .
the following summarizes the preliminary results .
* study of cepheids ( periods 110 d ) : 70 were identified in the lmc and 420 in the smc * study of cepheids or semi - regular variables ( @xmath44 1050 d ) : 433 in the lmc , several more in the smc * eclipsing binaries and rr lyrae stars ( @xmath44 120 h ) : 7 in lmc , 71 in smc * blue variables ( @xmath45 ) : 37 were identified in the smc including 7 of long or irregular period , 2 possible `` bumpers '' , and at least 3 eclipsing binaries . * red variables with @xmath46 and @xmath47 d ( lpv , sr ) :
89 were found in the smc and 5 in the lmc fig .
2 shows @xmath13 and @xmath14 phased light curves for four of the cepheids in the field smc2 ( series 2 ) .
3 shows unphased quasi - periodic light curves for four long period variables from field nlmc2 ( series 2 ) .
finally g. bayne , w. tobin and j. pritchard ( univ .
canterbury , n.z . ) are searching the moa database for eclipsing binaries and analysing the light curves ( series 1 , 2 ) .
this work is still in progress , using the period - searching algorithm of grison ( 1994 ) . | a review of the moa ( microlensing observations in astrophysics ) project is presented .
moa is a collaboration of approximately 30 astronomers from new zealand and japan established with the aim of finding and detecting microlensing events towards the magellanic clouds and the galactic bulge , which may be indicative of either dark matter or of planetary companions .
the observing program commenced in 1995 , using very wide band blue and red filters and a nine - chip mosaic ccd camera .
as a by - product of these observations a large database of ccd photometry for 1.4 million stars towards both lmc and smc has been established .
in one preliminary analysis 576 bright variable stars were confirmed , nearly half of them being cepheids .
another analysis has identified large numbers of blue variables , and 205 eclipsing binaries are included in this sample . in addition 351 red variables ( agb stars ) have been found .
light curves have been obtained for all these stars .
the observations are carried out on a 61-cm f/6.25 telescope at mt john university observatory where a new larger ccd camera was installed in 1998 july . from this latitude ( @xmath0 s )
the magellanic clouds can be monitored throughout the year .
f. abe , b. s. carter , r. j. dodd@xmath1 , m. honda , j. jugaku , s. kabe , + p. m. kilmartin , b. s. koribalski , y. matsubara@xmath2 , k. masuda@xmath2 , + y. muraki@xmath2 , t. nakamura , g. r. nankivell , m. reid , + n. j. rumsey@xmath3 , to .
saito , h. sato@xmath4 , m. sekiguchi@xmath5 , d. j. sullivan@xmath6 , + t. sumi@xmath2 , y. watase@xmath7 , t. yanagisawa@xmath2 , p. c. m. yock@xmath8 and + m. yoshizawa |
much progress has been made in recent decades in understanding the relationship between star formation rate ( sfr ) surface density ( @xmath5 ) and gas surface density ( @xmath6 ) in nearby star - forming spiral galaxies , and this work has been used to infer the physical basis of the star formation law ( e.g. , @xcite ) .
studies relating the two observables with a power law of the form @xmath7 typically find power - law indices _
n _ ranging from 1 to 3 ( e.g. , @xcite ; see @xcite for a review of previous work ) . in a spatially
resolved study of hi and h@xmath4 in star - forming spirals , @xcite find a linear ( @xmath81 ) relation between @xmath5 and molecular - gas surface density ( @xmath9 ) , but little to no correlation between @xmath5 and hi surface density .
in contrast , less is known about the connection between @xmath10 and star formation in early - type galaxies . the good correlation between the morphologies of molecular gas , 24@xmath11 m emission , and radio continuum in local e / s0s hints that there is a relationship between @xmath12 and @xmath5 @xcite . in a single - dish survey of co emission in sauron e / s0s ,
@xcite find that their galaxies follow the @xmath13 disk - averaged power law , characteristic of spirals @xcite .
using multiple star - formation tracers , @xcite find a similar result for a sample of 12 e / s0s , although possibly at lower total - gas star formation efficiencies ( tsfe @xmath14 sfr/@xmath15 ) .
@xcite update the @xcite results with spatially resolved maps , localizing both co emission and star formation in the central regions of the galaxies and finding @xmath5 , @xmath16 , and molecular - gas star formation efficiency ( msfe @xmath14 sfr/@xmath17 ) values similar to those of spirals . however , spatially resolved studies of the h@xmath4-star formation relation at sub - kpc resolution similar to the analyses of @xcite and @xcite have yet to be done for e / s0s .
recent work has identified a local population of star - forming e / s0s that reside alongside spirals on the blue sequence in color vs. stellar mass space ( * ? ? ?
* hereafter kgb ) .
the fraction of e / s0s on the blue sequence increases with decreasing mass , from @xmath18 at stellar mass @xmath19 , up to @xmath20 for @xmath21 ( 2% and 5% , respectively , of all galaxies in these mass ranges ; ) .
high - mass blue - sequence e / s0s are often young major - merger remnants that will fade to the red sequence ( see also @xcite ) .
low - mass blue - sequence e / s0s , in contrast , appear more settled , occupying low - density field environments where gas accretion is likely .
this population may reflect the transformation of `` red and dead '' e / s0s into spirals via inner- and outer - disk regrowth .
argue that many blue - sequence e / s0s occupy a `` sweet spot '' in @xmath22 and stellar concentration index , characterized by abundant gas and bursty , efficient ( when time - averaged over multiple bursts ) star formation , which may enable efficient disk building ( see also [ section.discussion ] ) .
@xcite confirm that blue - sequence e / s0s have fractionally large atomic gas reservoirs , comparable to those of spirals ( 0.11.0 , relative to @xmath22 ) .
the ongoing star formation and large gas reservoirs of these galaxies make them ideal for probing the spatially resolved relationship between @xmath9 and @xmath5 in e / s0s , offering unique insight into whether / how some e / s0s may be actively evolving via bursty , efficient star formation .
our parent sample of 32 galaxies consists of all e / s0s ( 14 blue- , 2 mid- , and 11 red - sequence ) with @xmath23 from the nearby field galaxy survey ( nfgs , @xcite ) , and an additional five blue - sequence e / s0s in the same mass range ( see ) .
the sample was defined for _ spitzer _ and _ galex _ programs go-30406 and gi3 - 046012 ( pi kannappan ) , so all galaxies have new or archival fuv and 24@xmath11 m data for sfr estimation .
of the parent sample , we observed 23 e / s0s in co(10 ) with the combined array for research in millimeter - wave astronomy ( carma ) , detecting 12 of 23 . in this letter , we focus on the 12 e / s0s with carma detections and 7 additional e / s0s with iram 30 m observations ( five detections , two limits ) in co(10 ) .
these 19 e / s0s sample a large range ( table [ table.global ] ) in color and stellar mass ( figure [ fig.sample ] ) , h@xmath4/hi mass ratio ( 0.0063.2 ) , and total gas - to - stellar mass ratio ( 0.073.4 ) .
the carma maps have beam sizes of 2@xmath244@xmath24 , with velocity coverage of 300450 @xmath25 and resolution of @xmath32.5 @xmath25 .
we reduced the data with the miriad package @xcite , using natural weighting .
a more detailed description of the carma data reduction will be provided in l. wei et al .
2010 , in preparation .
iram 30 m co(10 ) observations are from s. kannappan et al .
2010 , in preparation , and d. stark et al .
2010 , in preparation , with additional literature data as noted in table [ table.global ] .
we consider central pointings only , so the 23@xmath24 iram beam probes inner disks ( relative to @xmath30@xmath2661@xmath269 optical diameters ) .
comparison between carma and iram 30 m fluxes suggests that little flux is resolved out by the interferometric observations .
we use the pipeline 24@xmath11 m mosaics from the _ spitzer _ archive , and the background - subtracted pipeline - processed fuv data from the _ galex _ archive .
the carma and _ galex _ data sets were convolved with a kernel that reproduce the mips 24@xmath11 m psf ( @xmath27 @xmath28 ; @xcite ) exactly in the carma/_galex _ images , including the airy ring , which is @xmath29 of the @xmath30 . for our pixel - to - pixel analysis in [ section.pixels ] , we further convolve our data to a resolution of 750pc ( [email protected]@xmath24 ) for comparison with @xcite .
all maps are at least nyquist - sampled .
we estimate @xmath31 and @xmath32 following @xcite : @xmath33 @xmath34 assuming a co - to - h@xmath35 conversion factor ( @xmath36 ) of @xmath37 .
our estimates of @xmath12 do not include helium ; however , helium is included in estimates of molecular gas depletion time .
equation [ eq.sfr ] uses the broken power law initial mass function given by @xcite .
our procedures for calculating @xmath31 and @xmath32 are identical to those of @xcite after their conversion from co(21 ) to co(10 ) ( @xmath38/@xmath39 ) , ensuring a fair comparison between the two data sets .
agn contamination is not an issue for the 750pc resolution analysis , as any agn contribution in the infrared would be contained within the central resolution element .
ngc 4117 and ic 1141 are known agn hosts , and optical line diagnostics from @xcite identify ngc 5173 as a candidate host . however
, irac color - color diagnostics @xcite suggest that ( possible ) agn contributions to the integrated infrared emission in our galaxies are relatively weak , as the removal of the central resolution element does not significantly affect their positions in the infrared color - color diagram .
following @xcite , we estimate possible 24@xmath11 m emission contamination from passively evolving stellar populations to be @xmath408% for our e / s0s .
figure [ fig.ks_pixels ] plots the pixel - to - pixel relationship between @xmath5 and @xmath12 for the eight of our 19 galaxies resolved on 750pc scales .
we also include the 750pc resolution data for the seven spirals observed by @xcite as light blue dots .
the vertical dashed lines demarcate the three different star formation regimes discussed by @xcite : hi - dominated , giant molecular cloud ( gmc)/disk , and starburst .
figure [ fig.ks_pixels ] shows that all but two of the eight galaxies have some regions that fall within the gmc / disk regime , with two blue- and one mid - sequence e / s0s ( ngc 3032 , ugc 6570 , and ugc 7020a ) having the majority of their points in this regime .
we fit the @xmath31@xmath32 relationship with a power - law of the form @xmath41 in log - log space using the ordinary least - squares ( ols ) bisector method ( solid line in figure [ fig.ks_pixels ] ) and list the fit parameters in table [ table.ksfits ] , using coefficient @xmath42 .
note that we set the intercept of our fit at log@xmath43(@xmath44 ) , where @xmath44 is the median @xmath31 for each galaxy , to lessen the effect of the covariance between @xmath45 and @xmath46 . the power - law index _
n _ ranges from 0.62 to 1.92 , with a median of @xmath47 .
it is evident that the majority of e / s0s in figure [ fig.ks_pixels ] exhibit a power - law relation between @xmath31 and @xmath48 , all the way down to the hi - dominated regime .
figure [ fig.ks_pixels ] also plots ( dotted ) lines of constant msfe with @xmath49 , defined as the inverse of the molecular gas depletion time @xmath17/sfr , and illustrates variations in @xmath45 and @xmath50/@xmath51 . for blue- and mid - sequence e / s0s , we find that as msfe increases , @xmath45 seems to steepen and @xmath50/@xmath52 seems to decrease .
we find a wider range of msfes ( 4%70% ) compared to the 3%8% found by @xcite .
we discuss whether msfe truly measures molecular - gas star formation _
efficiency _ in [ section.discussion ] .
in figure [ fig.ks_global ] , we show the 18 kpc aperture - averaged relationship between @xmath53 and @xmath5 for the eight e / s0s of figure [ fig.ks_pixels ] plus 11 more blue - sequence e / s0s with carma detections and/or iram observations , as well as for normal spiral - disk and starburst galaxies @xcite and sauron e / s0s @xcite . with the exception of kennicutt s normal disk points ,
all the other points in figure [ fig.ks_global ] show the surface densities within regions of star formation or molecular gas .
the starbursts were averaged over `` the radius of the starburst region '' determined from co / infrared imaging @xcite , and the sauron e / s0s were averaged over the extent of the star - forming region defined by 8@xmath11 m pah emission @xcite . for our 19 e / s0s ,
we average @xmath50 and sfr over an area with radius ( @xmath54 ) twice the scale length of the 24@xmath11 m emission , where the flux drops by e@xmath55 from the peak .
this area encompasses most of the flux in co , 24@xmath11 m , and fuv , and corresponds well to the visual impression of the extent of co emission for most galaxies .
note that the dots from @xcite plotted in the background for reference are _ local _
( 750pc ) measures of @xmath31 and @xmath32 as in figure [ fig.ks_pixels ] .
in contrast , the @xcite spiral disk points are averaged over @xmath56 , which may dilute the values of @xmath31 and @xmath32 .
we infer this from radial profiles of normal spirals in @xcite , which indicate that h@xmath4 typically extends out to only @xmath57 .
thus the surface densities calculated by @xcite for spiral disks should for consistency move up along lines of constant msfe by @xmath30.62 dex ( grey arrow in figure [ fig.ks_global ] ) .
comparison of these data sets reveals that the aperture - averaged @xmath58 and @xmath32 for our 19 e / s0s overlap the range spanned by spiral disks , with six of the carma - detected e / s0s having sufficiently high surface densities to occupy the same space where @xcite finds the _ centers _ of spirals lie between normal disks and starburst galaxies ( not shown in figure [ fig.ks_global ] , as @xcite tabulates only combined hi+h@xmath4 data for galaxy centers ) .
our e / s0s also appear to span the same range as the typically more massive sauron e / s0s .
figure [ fig.ks_global ] also shows a similarly large spread in aperture - averaged msfes as seen for local msfes in [ section.pixels ] , with over half of the carma - detected and all of the 30@xmath59m - observed e / s0s ( all blue- or mid - sequence ) offset towards apparently higher msfes ( @xmath6010% ) compared to the typical spirals from @xcite .
equivalently , the molecular gas depletion times for our e / s0s range from 2.3 down to 0.1gyr , with a median of 0.5gyr lower than that of the @xcite spirals .
additionally , the offset towards apparently higher msfes is seen in both the carma and iram 30 m galaxies in figure [ fig.ks_global ] , so it is not specific to interferometric data .
we have shown above that the relationship between molecular gas and star formation in low - mass e / s0s ( @xmath61 ) resolved at 750pc is similar to that for spirals , with a roughly linear correlation between @xmath31 and @xmath62 all the way down to the hi - dominated regime .
this suggests that star formation occurs in h@xmath4 and not hi , similar to what @xcite find for spirals .
one intriguing difference is the apparently elevated msfes of our e / s0s compared to the msfes of the bigiel spirals . _ star formation efficiency _
, however , may be a misnomer in some cases , as other factors may contribute to the observed offsets .
one possible cause for apparently elevated msfes is that the co may not trace all of the h@xmath4 in these galaxies , as many are low - mass systems where @xmath36 may be variable ( e.g. , @xcite ) . @xcite
finds a similar scatter in msfes for low - luminosity ( @xmath63 ) disk galaxies , which he attributes to variation in @xmath36 , possibly due to low metallicities .
recent _ herschel _ results support this , finding evidence for excess cold dust that is not well - traced by co in low - mass galaxies @xcite
. however , metallicity measurements ( available for 5/6 carma - detected and 3/7 iram - observed galaxies offset towards higher msfes ) indicate that these galaxies are well within the range ( @xmath641/4@xmath65 ) where much work suggests that @xmath36 is similar to that assumed here ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a second explanation is that some of these galaxies are advanced , h@xmath4-depleted starbursts , where the delay between h@xmath4 exhaustion and fading of star - formation tracers associated with young , massive stars gives the _ appearance _ of elevated msfes .
thus the possible correlation between elevated msfe , lower @xmath66/@xmath51 , and steeper ( higher @xmath45 ) slopes in the schmidt - kennicutt relation found in [ section.pixels ] may reflect the depletion of h@xmath4 in the later stages of star formation , in good agreement with simulations that predict steeper schmidt - kennicutt relations and higher msfes as the molecular gas fraction decreases @xcite .
this is consistent with studies of ngc 1569 , a post - starburst dwarf irregular ( e.g. , @xcite ) that also appears to have an elevated msfe @xcite .
a third possibility is that the observed offsets may reflect truly enhanced msfes .
@xcite , updating @xcite , find that the fractional gas content of galaxies abruptly rises below a gas - richness threshold mass of @xmath67@xmath68 , roughly corresponding to internal velocities of @xmath3120 kms@xmath69 ( see also ) .
this is the same velocity threshold below which the physics of star formation may change due to possibly increased gas accretion , outflow , and metal loss from shallower potentials @xcite . while the _ total _ sfes of dwarf galaxies are low compared to those of normal star - forming spirals ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , recent simulations suggest that gas - rich and/or lower metallicity galaxies deviate from the schmidt - kennicutt relationship towards higher _ molecular _ sfes @xcite .
these predictions are supported by observations of local dwarfs , which find high msfes in ic 10 and m33 compared to nearby spirals @xcite .
this phenomenon could contribute to the scatter in msfes observed by @xcite , as @xmath70 roughly corresponds to the gas - richness threshold mass .
similarly , the two sauron e / s0s in figure [ fig.ks_global ] with the highest msfes are low - luminosity systems with @xmath71 .
unlike previous studies of star formation in blue e / s0s , our sample focuses on galaxies below the gas - richness threshold mass ( 13 of our 19 e / s0s ) . at higher stellar masses , agn and strong starbursts
are observed to dominate the blue e / s0 population ( e.g. , * ? ? ?
* ) , which is not inconsistent with the nature of our six higher - mass e / s0s ( three known / candidate agns and two likely starbursts due to their large distances and the resulting poor resolution of the central region , but their nuclear ew(h@xmath72 ) emission measurements ( 86 and 76 , respectively ) and very blue - centered color gradients suggest recent / ongoing central starbursts . ] ) .
our results suggest that ( possibly milder ) bursts likely play a key role for lower - mass e / s0s as well .
the variation of msfe in this scenario has implications for our understanding ( and theoretical simulations ) of low - mass galaxy evolution .
the dynamical timescales for gas inflow typical for our sample ( 0.060.4 gyr ; @xcite ) are short compared to the molecular gas consumption times we find here ( 0.12.3 gyr ) , which suggests that refueling of h@xmath4 from the hi reservoir to sustain star formation is limited only by the frequency of minor mergers / interactions that trigger gas inflow . if apparently elevated msfes reflect advanced , h@xmath4-depleted bursts , the fact that 15 of our 19 e / s0s have msfe @xmath6010% suggests that low - mass e / s0s may experience frequent small starbursts ( with the caveat that we have sampled the most strongly star - forming examples by favoring blue - sequence e / s0s detected in co ) .
therefore the tsfes of lower - mass galaxies , when time - averaged over many bursts , may be elevated consistent with the finding that the concentration indices in blue - sequence e / s0s are similar to those identified by @xcite as optimal for peak time - averaged tsfe .
our results support the picture of and @xcite that many blue - sequence e / s0s are in a `` sweet spot '' with abundant gas and bursty star formation enabling efficient disk building .
we thank the referee for helpful comments , a. leroy for sharing the things data and for useful discussions , and s. jogee for her role in acquiring the _ spitzer _ data .
we are grateful to a. bolatto , j. gallimore , m. lacy , a. moffett , m. thornley , and s. veilleux for insightful conversations .
this work is based in part on observations made with the _
spitzer space telescope _
, which is operated by jpl , caltech under a contract with nasa .
support for this work was provided by nasa through an award issued by jpl / caltech .
this work uses observations made with the nasa galaxy evolution explorer .
_ galex _ is operated for nasa by caltech under nasa contract nas5 - 98034 .
we acknowledge support from the _ galex _ gi grant nnx07at33 g .
carma development and operations are supported by nsf under a cooperative agreement , and by the carma partner universities . , r. j. , teuben , p. j. , & wright , m. c. h. 1995 , in astronomical society of the pacific conference series , vol .
77 , astronomical data analysis software and systems iv , ed .
r. a. shaw , h. e. payne , & j. j. e. hayes , 433 , k. , et al .
2009 , , 690 , 1672 color vs. stellar mass for galaxies in the nearby field galaxy survey @xcite , plus five additional e / s0s from the literature .
symbols denote morphological types . the red sequence , i.e. the main locus of traditional red e / s0s , lies above the dashed line ( with two borderline `` mid - sequence '' e / s0s boxed , see ) , while the blue sequence ( typically populated by spirals ) lies below .
dark symbols denote the 32 galaxies in the parent sample ; the rest of the nfgs is shown in light grey .
[ fig.sample ] ] vs. @xmath31 at 750pc resolution .
points for the seven spirals from @xcite are plotted in light blue .
black triangles show our eight e / s0s resolved in co(10 ) with carma .
vertical dashed lines demarcate the three different regimes of star formation discussed in @xcite , and dotted lines mark constant msfe , corresponding ( from top to bottom ) to the depletion of 100% , 10% , and 1% of the molecular gas ( including helium ) within @xmath73yr , or equivalently to molecular gas depletion timescales of 0.1 , 1 , and 10 gyr .
solid black lines represent ols bisector fits .
the color sequence ( blue / mid / red ) is noted at the bottom of each panel .
the last panel combines the points for all eight galaxies , color - coded by @xmath74/@xmath51 , illustrating variations in msfe with power - law index @xmath45 and @xmath75/@xmath51 ( see [ section.pixels ] and [ section.discussion ] ) . [ fig.ks_pixels ] ] vs. @xmath76 for normal disk and nuclear starburst galaxies from @xcite , sauron e / s0s from @xcite , and e / s0s from this letter .
the 750pc resolution points from @xcite are also plotted in the background for comparison .
dotted lines mark the same lines of constant msfe as in figure [ fig.ks_pixels ] .
grey arrow in the lower right shows typical shift of spiral disk points from @xcite if averaged over @xmath77 instead of @xmath56 .
[ fig.ks_global ] ] ngc 3011 & b & 0.8 & 25.7 & 9.4 & 8.3 & & @xmath407.0 & 12.1 & 27 & @xmath401.2 & [email protected] & @xmath400.22 + ngc 3032 & b & 1.3 & 25.2 & 9.6 & 8.3 & 8.6 & 8.6 & 10.6 & 26 & [email protected] & [email protected] & [email protected] + ngc 3073 & b & 1.1 & 21.1 & 9.1 & 8.5 & & 7.0 & 10.6 & 31 & [email protected] & [email protected] & [email protected] + ugc 6003 & b & 0.6 & 84.2 & 10.1 & 9.4 & 8.6 & 8.7 & 7.5 & 19 & [email protected] & [email protected] & [email protected] + ic 692 & b & 0.8 & 21.4 & 8.9 & 8.4 & & 6.8 & 12.1 & 42 & [email protected] & [email protected] & [email protected] + ugc 6570 & m & 1.2 & 28.6 & 9.6 & 8.4 & 8.0 & 8.0 & 7.5 & 62 & [email protected] & [email protected] & [email protected] + ugc 6637 & b & 0.9 & 31.5 & 9.2 & 8.6 & & 7.3 & 12.1 & 66 & [email protected] & [email protected] & [email protected] + ngc 3773 & b & 1.1 & 10.5 & 8.6 & 7.9 & 6.6 & & 10.6 & 44 & [email protected] & [email protected] & [email protected] + ngc 3870 & b & 1.0 & 14.5 & 8.8 & 8.4 & 6.9 & 7.2 & 12.1 & 41 & [email protected] & [email protected] & [email protected] + ugc 6805 & b & 0.7 & 20.3 & 8.9 & 7.6 & 7.6 & 7.4 & 9.0 & 45 & [email protected] & [email protected] & [email protected] + ugc 7020a & m & 1.1 & 26.7 & 9.3 & 8.6 & 8.3 & 8.2 & 7.5 & 62 & [email protected] & [email protected] & [email protected] + ngc 4117 & r & 1.8 & 19.0 & 9.7 & 8.3 & 7.7 & & 7.5 & 72 & [email protected] & [email protected] & [email protected] + ngc 5173 & b & 1.2 & 41.2 & 10.3 & 9.3 & 8.2 & & 10.6 & 39 & [email protected] & [email protected] & [email protected] + ngc 5338 & r & 1.9 & 10.3 & 8.9 & 7.3 & 7.3 & 7.4 & 7.5 & 57 & [email protected] & [email protected] & [email protected] + ugc 9562 & b & 0.9 & 25.2 & 8.9 & 9.3 & & 7.0 & 12.1 & 68 & [email protected] & [email protected] & [email protected] + ic 1141 & b & 0.7 & 68.0 & 10.4 & 9.3 & 9.2 & & 7.5 & 26 & [email protected] & [email protected] & [email protected] + ngc 7077 & b & 0.9 & 18.9 & 8.8 & 8.2 & & 7.1 & 10.6 & 40 & [email protected] & [email protected] & [email protected] + ngc 7360 & b & 1.2 & 67.9 & 10.5 & 9.6 & & @xmath407.8 & 12.1 & 67 & @xmath400.5 & [email protected] & @xmath400.55 + ugc 12265n & b & 0.6 & 82.8 & 10.1 & 9.4 & 8.9 & 8.9 & 7.5 & 42 & [email protected] & [email protected] & [email protected] * blue seq . : * & & & & + ngc 3032 & 33.7 & [email protected] & [email protected] & 0.12 + ngc 3773 & 1.7 & [email protected] & [email protected] & 0.20 + ngc 3870 & 2.7 & [email protected] & [email protected] & 0.23 + ugc 6805 & 7.1 & [email protected] & [email protected] & 0.18 + * mid seq .
: * & & & & + ugc 6570 & 11.7 & [email protected] & [email protected] & 0.23 + ugc 7020a & 7.7 & [email protected] & [email protected] & 0.21 + * red seq .
: * & & & & + ngc 4117 & 3.0 & [email protected] & [email protected] & 0.13 + ngc 5338 & 5.2 & [email protected] & [email protected] & 0.09 + median & 7.1 & -1.73 & 1.19 & | we consider the relationship between molecular - gas and star - formation surface densities in 19 morphologically defined e / s0s with stellar mass @xmath0 , paying particular attention to those found on the blue sequence in color vs. stellar mass parameter space , where spiral galaxies typically reside .
while some blue - sequence e / s0s must be young major - merger remnants , many low - mass blue - sequence e / s0s appear much less disturbed , and may be experiencing the milder starbursts associated with inner - disk building as spirals ( re)grow . for a sample of eight e / s0s ( four blue- , two mid- , and two red - sequence ) whose carma co(10 ) , _ spitzer _ mips @xmath1 , and _ galex _
fuv emission distributions are spatially resolved on a 750pc scale , we find roughly linear relationships between molecular - gas and star - formation surface densities within all galaxies , with power law indices @xmath2 0.61.9 ( median 1.2 ) . adding 11 more blue - sequence e / s0s whose co(10 ) emission is not as well resolved , we find that most of our e / s0s have global 18 kpc aperture - averaged molecular - gas surface densities overlapping the range spanned by the disks and centers of spiral galaxies . while many of our e / s0s fall on the same schmidt - kennicutt relation as local spirals , @xmath380% ( predominantly on the blue sequence ) are offset towards apparently higher molecular - gas star formation efficiency ( i.e. , shorter molecular gas depletion time ) .
possible interpretations of the elevated efficiencies include bursty star formation similar to that in local dwarf galaxies , h@xmath4 depletion in advanced starbursts , or simply a failure of the co(10 ) emission to trace all of the molecular gas . |
the study of the intrinsic properties of the @xmath12 meson as well as its production processes in elementary particle and hadron physics is of particular interest for various reasons .
the properties of @xmath12 are largely governed by the dynamics of the qcd @xmath13 axial vector anomaly @xcite .
expressed in the ( pseudoscalar ) quark - flavor basis , the physically observed @xmath14 and @xmath12 mesons may be written as @xmath15 where @xmath16 and @xmath17 describe the strange and nonstrange quark - antiquark states , respectively .
the @xmath13 anomaly mediates @xmath18@xmath19 transitions and therefore plays a central role in understanding the @xmath14@xmath12 mixing @xcite .
the mixing angle @xmath20 is shown to be fairly constant @xcite , and a weighted average value of @xmath21 has been extracted @xcite .
quite recently , the kloe collaboration @xcite has reported a value of @xmath22 . as can be seen from eq .
( [ etaetapmixqs ] ) , such a value of the mixing angle results in a considerable amount of @xmath23 in both the @xmath14 and @xmath12 mesons .
by contrast , the corresponding mixing angle for the vector mesons @xmath24 and @xmath25 is quite small ( @xmath26 @xcite ) , providing an @xmath24 with nearly no @xmath27 and a @xmath25 being almost a pure @xmath23 state .
therefore , instead of using the vector mesons @xmath24 and @xmath25 , production processes involving @xmath12 and @xmath14 offer an alternative way of probing the strangeness content of the nucleon . due to the fact that both @xmath14 and @xmath12 contain a significant amount of @xmath23 , but of opposite phase with respect to the nonstrange @xmath28 component , significant interference effects involving the strange - quark piece of the nucleon wave function are possible @xcite .
in fact , it has been proposed that the relative cross sections for the reactions induced by pseudoscalar mesons , @xmath29 and @xmath30 , provide a sensitive test for the presence of the @xmath23 component in the nucleon wave function @xcite .
due to its nontrivial properties , the qcd vacuum exhibits strong gluonic fluctuations with pseudoscalar quantum numbers to which the @xmath31 states can couple via the @xmath13 axial anomaly .
the nonperturbative gluon dynamics and the axial anomaly @xcite are thought to be responsible for the generation of the much larger mass of @xmath12 as compared to the masses of other members of the @xmath32 pseudoscalar meson nonet known as the goldstone bosons .
the masses of the goldstone bosons are generated by the spontaneous breaking of chiral symmetry @xcite .
the @xmath12 meson is , therefore , thought to couple strongly to gluons via the @xmath13 axial anomaly coupling @xcite .
the unexpectedly large branching ratio measured recently for the inclusive decay of beauty particles , @xmath33 @xcite , has been interpreted as possible experimental evidence in this respect @xcite . to date
, the kloe collaboration has recently found that the gluonium content in the @xmath12 is consistent with a fraction below @xmath34 @xcite . in any case , if there is a strong coupling of @xmath12 meson to gluons , it would be conceivable that short - range reaction processes such as @xmath35 might reveal the gluonic degrees of freedom in the low energy interactions involving nucleons and @xmath12 @xcite .
one of the properties of the @xmath36 meson of extreme importance is its yet poorly known coupling strength to the nucleon .
this has attracted much attention in connection with the so - called `` nucleon - spin crisis '' in polarized deep inelastic lepton scattering @xcite . in the zero - momentum limit ,
the @xmath37 coupling constant @xmath38 is related to the flavor singlet axial charge @xmath39 through the flavor singlet goldberger - treiman relation @xcite ( see also refs.@xcite ) @xmath40 where @xmath41 is a renormalization - group invariant decay constant ; @xmath42 and @xmath43 denote the number of flavors and the pion decay constant , respectively .
@xmath44 describes the coupling of the nucleon to the gluons arising from contributions violating the okubo - zweig - iizuka rule @xcite .
the emc collaboration @xcite has measured an unexpectedly small value of @xmath45@xmath46 . the first term on the right - hand side of the above equation corresponds to the quark contribution to the `` spin '' of the proton , and the second term to the gluon contribution @xcite .
therefore , if @xmath47 is known , eq .
( [ spinfrac ] ) may be used to extract the coupling @xmath48 .
however , unfortunately , there is no direct experimental measurement of @xmath47 so far .
reaction processes where the @xmath12 meson is produced directly off a nucleon , such as @xmath49 and @xmath50 , may thus offer a unique opportunity to extract this coupling constant .
of course , other production mechanisms , such as meson exchange and nucleon resonance currents , must be taken into account before a quantitative determination of @xmath51 is possible .
yet another interesting aspect in studying @xmath12 production processes is that they may provide an alternative tool to extract information on nucleon resonances , @xmath52 .
current knowledge of most of the nucleon resonances is mainly due to the study of @xmath53 scattering and/or pion photoproduction off the nucleon .
reaction processes such as @xmath12 photoproduction provide opportunities to study those resonances that couple only weakly to pions , especially , in the less explored higher @xmath52 mass region of `` missing resonances '' @xcite .
missing resonances are those predicted by quark models but not found in more traditional pion - production reactions @xcite . in the present work ,
we concentrate on the reactions @xmath49 and @xmath35 .
so far there exists only a limited number of studies of the @xmath12 photoproduction both experimentally @xcite and theoretically @xcite .
et al . _
@xcite , in their theoretical investigation using an effective lagrangian approach , have emphasized the role of the @xmath54 resonance in the description of the , then , existing data @xcite , while li @xcite has described those data within a constituent quark model with the off - shell @xmath55 excitation as the dominant contribution .
the authors of ref .
@xcite described their data obtained with much higher statistics than the previous measurements @xcite in the energy region from threshold to 2.6gev under the assumption of resonance dominance .
they considered an @xmath3 and a @xmath4 resonance with extracted masses of 1897 and 1986mev , respectively .
the former resonance was needed to explain the energy dependence of the total cross section which exhibits a steep rise and falloff close to threshold .
the @xmath56 resonance was needed to account for the measured forward rising angular distributions . in a calculation similar to that in ref .
@xcite , zhao @xcite introduced also a @xmath57 and an @xmath58 resonance to describe the saphir data @xcite . in both these quark model calculations ,
no ( @xmath5-channel ) vector meson exchange contribution was considered .
based on a @xmath59 baryon chiral perturbation theory , borasoy @xcite introduced the off - shell @xmath60 and @xmath55 resonances , in addition to the born and vector meson exchange contributions , to describe the data @xcite .
quite recently , chiang _
et al . _
@xcite have put forward a model for @xmath0 photoproduction that considers the @xmath5-channel vector meson exchanges in terms of regge trajectories to comply with high energy behavior . in their calculation , which was applied to the saphir data @xcite ( that cover an energy region @xmath61gev ) , the interference of the regge trajectories with an @xmath3 resonance is the underlying mechanism responsible for reproducing the data and no need of any @xmath4 resonance contribution was found .
in contrast , also in a quite recent calculation , sibirtsev _ et al . _
@xcite have described the saphir data by considering the @xmath5-channel @xmath62- and @xmath24-meson exchange contributions with an exponential form factor at the @xmath63 vertex ( @xmath64 ) .
the observed forward rise of the angular distribution is then largely accounted for by the ( @xmath5-dependent ) exponential form factor .
in addition , the @xmath55 resonance is introduced in order to account for the energy dependence of the total cross section .
et al . _
@xcite have also speculated that the @xmath0 photoproduction at high energies and large @xmath5 may be useful in determining the @xmath65 coupling constant @xmath66 .
new experimental investigations of @xmath12 photoproduction are currently being carried out at jlab by the clas collaboration @xcite and at elsa by the crystal barrel collaboration @xcite .
the @xmath67 reaction has been a subject of increasing attention in the last few years .
experimental data on total cross section exist for excess energies up to @xmath68 mev @xcite , in addition to the total cross section and the angular distribution at @xmath69mev from the disto collaboration @xcite .
the new total cross section data in the excess energy range of @xmath7047mev and an angular distribution at @xmath71mev have been just reported by the cosy-11 collaboration @xcite , filling in partly the gap between the near threshold @xcite and higher energy disto data @xcite .
theoretically , the @xmath67 reaction has been investigated by a number of authors @xcite within meson - exchange approaches of varying degrees of sophistication .
in particular , in ref .
@xcite , we have explored the possible role of the nucleonic , mesonic , and resonance current contributions .
the @xmath72 and @xmath56 resonances as determined by the saphir collaboration @xcite have been considered for the resonance current . due to the scarcity of the then available data ( total cross sections up to @xmath73mev ) , it was not possible to quantitatively constrain each of these currents . with the increase of the data base since then , we are now in a much better position to learn about this reaction than was possible before . diagrams contributing to @xmath74 .
time proceeds from right to left .
the intermediate baryon states are denoted for the nucleon , and for the @xmath3 and @xmath4 resonances .
the intermediate mesons in the @xmath5-channel are @xmath62 and @xmath24 .
the external legs are labeled by the four - momenta of the respective particles and the labels , , and of the hadronic vertices correspond to the off - shell mandelstam variables of the respective intermediate particles .
the three diagrams in the lower part of the diagram are transverse individually ; the three diagrams in the upper part are made gauge - invariant by an appropriate choice ( see text ) of the contact current depicted in the top - right diagram .
the nucleonic current ( nuc ) referred to in the text corresponds to the top line of diagrams ; the meson - exchange current ( mec ) and resonance current contributions correspond , respectively , to the leftmost diagram and the two diagrams on the right of the bottom line of diagrams . ]
the major purpose of the present work is to perform a combined analysis of the @xmath75 and @xmath35 reactions within a relativistic meson - exchange model of hadronic interactions ( see figs . [
fig : etaphotoprod ] and [ fig : ppetaprod ] ) . for the @xmath12 photoproduction , in the @xmath6 and @xmath7 channels , we consider contributions due to the intermediate nucleon and the nucleon resonances and in the @xmath5-channel
, we take into account @xmath62 and @xmath24 meson exchanges . since we employ the physical coupling constants and physical masses for all intermediate particles in all the channels
, the @xmath6-channel diagrams also account for the pole part of the @xmath76 final - state interaction ( fsi ) @xcite . for the nonpole part of the fsi ,
the @xmath7 and @xmath5 channels correspond to the born approximation of the corresponding @xmath77 @xmath78-matrix .
phenomenological form factors are attached to each vertex in all channels .
the total amplitude is constrained to obey gauge invariance following the prescription of refs .
the photoproduction amplitude thus obtained is then used in the construction of the basic @xmath0 production amplitude in @xmath50 by replacing the photons with relevant mesons which , in turn , are attached to the second nucleon ( see fig . [
fig : ppetaprod ] ) .
hereafter , the basic @xmath12 production amplitude is referred to as the @xmath12 production current following the nomenclature employed in ref .
the @xmath35 reaction is then described in a distorted - wave born approximation ( dwba ) which includes both the nucleon - nucleon ( @xmath79 ) final - state interaction and the initial - state interaction ( isi ) .
the present paper is organized as follows . in sec .
ii our model for the @xmath75 and @xmath35 reactions is described briefly .
the numerical results are discussed in sec .
iii , and in sec .
iv we present our summarizing conclusions .
basic production mechanisms for @xmath80 .
time proceeds from right to left . the full amplitude , with additional initial- and final - state contributions , is given by eq .
( [ eq : ppetaprod ] ) . as in fig .
[ fig : etaphotoprod ] , and denote the intermediate nucleon and resonances , respectively , and incorporates all exchanges of mesons @xmath81 , @xmath14 , @xmath62 , @xmath24 , @xmath82 , and @xmath83 ( @xmath84 ) for the nucleon graphs and @xmath81 , @xmath62 , and @xmath24 for the resonance graphs .
external legs are labeled by the four - momenta of the respective particles ; the hadronic vertices , , and here correspond to the same kinematic situations , respectively , as those identified similarly in fig .
[ fig : etaphotoprod ] . the nucleonic , resonance , and meson - exchange contributions referred to in the text correspond , respectively , to the first , second , and third lines of the diagrams on the right - hand side . ]
the dynamical content of our approach is summarized by the graphs of figs . [
fig : etaphotoprod ] and [ fig : ppetaprod ] . for the @xmath0 photoproduction
, we employ the tree graphs of fig .
[ fig : etaphotoprod ] with form factors at the vertices to account for the hadronic structure .
the gauge invariance of this production current is ensured by adding a phenomenological contact current , according to the prescription of refs .
this contact term ( see below ) may also be thought of as mocking up the neglected parts of the final - state interaction .
the hadronic @xmath85 reaction is described according to the model put forward in refs .
the dwba amplitude @xmath86 for this process is given by @xcite @xmath87 where @xmath88 , with @xmath89 , denotes the @xmath79 @xmath78-matrix interaction in the initial ( @xmath90 ) or final ( @xmath91 ) state , and @xmath92 is the corresponding two - nucleon propagator ( which absorbs the factor @xmath90 found in the dwba formula given in ref .
@xmath93 sums up the basic @xmath0 production mechanisms depicted in fig .
[ fig : ppetaprod ] . in the absence of models capable of providing a reliable off - shell @xmath79 isi ,
we consider it only in the on - shell approximation following ref .
this is shown to be a reasonable approximation for calculating cross sections @xcite .
the @xmath79 fsi is generated by using the bonn potential @xcite .
we use the blankenbecler - sugar propagator for the two - nucleon propagator @xmath94 in eq .
( [ eq : ppetaprod ] ) in order to be consistent with the @xmath79 interaction used .
the coulomb force is ignored in the present calculation ; its effect is known to be relevant only in the energy region very close to threshold ( excess energies less than 5 mev ) @xcite . in the present work
we concentrate our attention on the higher excess energy region where the coulomb effect is negligible .
the interaction lagrangian used to construct our model for the basic production amplitudes is given below .
for further convenience , we define the operators @xmath95 the electromagnetic vertices are derived from the following lagrangian densities . _
@xmath96 lagrangian : _
@xmath97 where @xmath98 and @xmath99 stand for the photon and nucleon fields , respectively .
@xmath100 is the nucleon mass , @xmath101 the elementary charge unit , and @xmath102 the anomalous magnetic moment of the proton . _
@xmath103 lagrangian ( resonance @xmath104 ) : _ @xmath105 where @xmath52 stands for the resonance field .
the upper and lower signs on the left refer to the even ( @xmath106 ) and odd ( @xmath107 ) parity resonances , respectively ; @xmath108 is the resonance mass and @xmath109 the coupling constant .
both parameters are fit parameters . _
@xmath110 lagrangian ( vector meson @xmath64 ) : _
+ @xmath111 where @xmath112 is the levi - civita tensor .
@xmath113 stands for the vector meson field ( @xmath114 ) .
the resulting @xmath115 vertex is multiplied by the form factor @xmath116 which describes the off - shell behavior of the intermediate vector meson with squared momentum transfer @xmath117 ( cf .
fourth diagram in fig .
[ fig : etaphotoprod ] ) . in general
, we use the dipole form @xmath118 ( see , however , fig . [
fig : e_mr2 ] below and its discussion in the text ) .
the cutoff @xmath119 , taken to be identical for both @xmath62 and @xmath24 , is a fit parameter .
the coupling constants @xmath120 in eq.([eq : gammaveta ] ) are taken from radiative decays @xcite ; their signs are inferred from @xmath32 symmetry considerations following ref .
@xcite in conjunction with the sign of the coupling constant @xmath121 determined from a study of pion photoproduction in the 1 gev energy region @xcite .
the following lagrangians describe the hadronic vertices . _
@xmath65 lagrangian : _
@xmath122\eta^\prime\right\ } n ~ , \label{nneta'}\ ] ] where @xmath123 .
_ @xmath124 lagrangian ( resonance @xmath104 ) : _ @xmath125\eta^\prime\right\ } n \nonumber\\ & \qquad + { \text{h.\,c.}}\label{nn^*eta'}\end{aligned}\ ] ] where the upper and lower signs on the left refer to the even ( @xmath106 ) and odd ( @xmath107 ) parity resonances , respectively . following refs .
@xcite , each of the @xmath126 vertices obtained from eqs .
( [ nneta ] ) and ( [ nn^*eta ] ) ( @xmath127 ) is multiplied by a phenomenological cutoff function @xmath128 which is normalized to unity , i.e. , @xmath129 .
the variable @xmath130 is the squared four - momentum of the intermediate off - shell baryon @xmath131 , whose mass @xmath132 is equal to either the nucleon mass @xmath100 or the mass of the resonance , @xmath108 .
the cutoff @xmath133mev is taken as the same for all baryons .
the parameters @xmath134 in eqs .
( [ nneta ] ) and ( [ nn^*eta ] ) describing the mixing of pseudoscalar and pseudovector contributions and the coupling constants @xmath135 are individual fit parameters for each of the three baryon states considered here .
( as the subsequent discussion shows , the fits prefer couplings that are almost entirely pseudovector for the nucleon , i.e. , @xmath136 , and almost entirely pseudoscalar for the resonances , i.e. , @xmath137 . )
_ @xmath138 lagrangian ( vector meson @xmath64 ) : _ @xmath139 where @xmath140 and @xmath141 stand for the @xmath62 and @xmath24 meson fields , respectively .
each of the resulting @xmath142 vertices is multiplied by a product of form factors , @xmath143 , where @xmath144 and @xmath145 ( cf .
last diagram in fig .
[ fig : ppetaprod ] ) .
the form factor @xmath146 associated with one of the intermediate off - shell vector mesons is the same as in eq .
( [ eq : dipoleff ] ) , with the same cutoff masses @xmath119 , except for the normalization point , consistent with the kinematics at which the coupling constants @xmath147 are extracted .
the @xmath147 are obtained from a systematic analysis of the radiative decay of vector and pseudoscalar mesons based on @xmath32 symmetry considerations in conjunction with vector - meson dominance arguments @xcite .
hence , there are no free _ independent _ parameters for this vertex . _
@xmath148 lagrangian ( resonance @xmath104 ) : _ @xmath149 where @xmath150 denotes the pion field .
again , each of the resulting @xmath148 vertices is multiplied by a product of form factors , @xmath151 ; @xmath152 is the pion s four - momentum and @xmath130 , as before , is the squared four - momentum of the intermediate @xmath52 state .
the form factor @xmath153 here is exactly the same as in eq .
( [ eq : ffhadron ] ) for @xmath154 , with @xmath155mev .
@xmath156 is the pion form factor parametrization from the bonn potential , with a cutoff - mass value of 900mev . for this vertex ,
therefore , the coupling constant @xmath157 is the only additional fit parameter . _
@xmath158 lagrangian ( vector meson @xmath64 ; resonance @xmath104 ) : _ @xmath159 where @xmath160 .
each resulting vertex is multiplied by a product of form factors , @xmath161 ; @xmath162 is the vector meson s four - momentum and @xmath130 has the same meaning as before .
the parameters of @xmath153 and @xmath163 are fixed already ; the coupling constant @xmath164 , therefore , is the only fit parameter here .
all of the remaining _ @xmath165 vertices _
( meson @xmath166 ) are parametrized as in the bonn potential @xcite .
the only exceptions are the values of the coupling constant @xmath167 , the pseudoscalar - pseudovector ( ps - pv ) mixing parameter @xmath168 , and the cutoff - mass value of 900mev at the @xmath169 vertex used in the resonance and meson exchange currents ( see discussion in refs .
@xcite ) . throughout this work ,
the widths of the nucleon resonances are fixed to be @xmath170mev ( @xmath171 ) .
we neglect their energy dependence in order to keep the analysis simple .
certainly , such a feature should be taken into account when aiming at a more quantitative extraction of the resonance parameters with data more accurate than what are available at present . in the present work we restrict ourselves to contributions from @xmath81 , @xmath62 , and @xmath24 meson exchanges in the resonance currents in describing the @xmath172 reaction . also , in contrast to ref .
@xcite , we omit the @xmath173-exchange current in the present work because it is much less under control than the dominant @xmath142-exchange contribution and its inclusion would introduce additional uncertainties in the model . employing form factors for the @xmath6- and @xmath7-channel contributions to the photoproduction amplitude containing an intermediate nucleon
( see the first two diagrams in fig .
[ fig : etaphotoprod ] ) and allowing for pseudovector couplings in the @xmath65 vertex in general destroys the gauge invariance of the production amplitude . within the present context of a model approach , to restore gauge invariance requires the introduction of phenomenological contact - type currents .
following here the prescription given in refs .
@xcite , there are two basic contributions necessary to ensure gauge invariance for the present application to @xmath1 .
the first contribution , @xmath174~ , \label{eq : jkr}\ ] ] corresponding to the kroll - ruderman current of pion photoproduction , cancels the gauge - invariance - violating terms arising from using pseudovector couplings .
the form factors @xmath175 here correspond to eq .
( [ eq : ffhadron ] ) . the second gauge - invariance preserving ( gip )
contribution , @xmath176 \nonumber\\ & \quad\mbox { } -e\,g_{nn\eta'}\,\gamma_5\,\frac{(2p'-k)^\mu)}{u - m_n^2 } \left [ g_n(u)-\hat{f } \right]~ , \label{eq : jgip}\end{aligned}\ ] ] is necessary because our model employs form factors at the vertices .
as far as gauge invariance is concerned , the function @xmath177 here is arbitrary .
analyticity , on the other hand , requires that this current be free of singularities , i.e. , it must be a _ contact _ current .
one of the simplest choices for @xmath177 in the present context then is in terms of the existing form factors of the problem at hand .
the most general ansatz then would be @xmath178^i [ g_u(u)]^j [ g_t(t)]^k \ , \ ] ] where @xmath179 , @xmath180 , and @xmath181 are the @xmath6- , @xmath7- , and @xmath5-channel form factors .
the simplest choice that is free of singularities is then given by restricting the sum to @xmath182 and putting @xmath183 .
equation ( [ eq : fhat ] ) follows with @xmath184 and @xmath185 . ] @xmath186\left [ g_n(u)-1 \right]~. \label{eq : fhat}\ ] ] this corresponds to the choice advocated in ref .
@xcite on the grounds of crossing symmetry .
the resulting gip current then is @xmath187 which evidently is free of any singularities . adding the sum of the kroll - ruderman term ( [ eq : jkr ] ) and the gip current ( [ eq : jgip1 ] ) restores gauge invariance for the present model ; in fig . [
fig : etaphotoprod ] , they correspond to the rightmost diagram in the top row of diagrams . [ cols="<,>,>,>,>",options="header " , ] [ tab3 ] in any case , we have also performed the calculation of the photoproduction reaction by replacing the @xmath5-channel vector meson exchanges by the regge trajectories following ref .
apart from the obvious differences in the details of the treatment of the resonance current , our calculation also differs in the sign of the @xmath188 coupling constant from that employed in ref . @xcite . in the present work the signs of the @xmath189 couplings
are inferred from a systematic analysis @xcite of the pseudoscalar and vector meson radiative decays based on an @xmath32 lagrangian in conjunction with the sign of the coupling constant @xmath121 determined from a study of pion photoproduction in the 1 gev energy region @xcite .
our results using the regge trajectories are shown in fig .
[ fig : regge ] ; the corresponding parameters are found in table [ tab3 ] .
we see that , overall , the results are basically the same as those of fig . [
fig : photodiff](c ) using the conventional @xmath62 and @xmath24 meson exchanges .
in particular , here also the interference between the mesonic and resonance currents is the underlying mechanism responsible for reproducing the observed angular distribution .
this corroborates the findings of ref .
@xcite . however , the resulting resonance masses of @xmath190mev and @xmath191mev differ considerably from those obtained in fig . [
fig : photodiff](c ) .
it is natural to ask whether this discrepancy is related to the uncertainties in the determination of the resonance mass using only the cross section data as illustrated in fig .
[ fig : resmass ] or whether it is due to different approaches used in the treatment of the @xmath5-channel exchange contribution . to address this question , in fig .
[ fig : regge_resmass ] we show the results of three different fits using the regge trajectories .
the solid curves are the same results shown in the left panel of fig .
[ fig : regge ] .
the dashed curves correspond to another fit resulting in the resonance masses of @xmath192mev .
one sees that the qualities of both fits are comparable to each other but the corresponding values of @xmath193 differ by more than @xmath194mev , revealing once more ( cf.fig .
[ fig : resmass ] ) that the cross section data alone are insufficient to constrain accurately the resonance masses .
however , we were unable to fit the data ( with a comparable quality ) with the mass of the @xmath3 resonance much smaller than @xmath190mev .
the dash - dotted curves in fig . [ fig : regge_resmass ] correspond to a fit with a fixed mass of @xmath195mev .
this yields a fitted mass of @xmath196mev for the @xmath4 resonance .
this set of the resonance masses @xmath197mev is more in line with the set @xmath198mev obtained using the conventional vector meson exchanges in the @xmath5-channel . here , however , the quality of the fit is inferior to that achieved in fig .
[ fig : photodiff](c ) , although the data are still reproduced within their uncertainties .
these considerations indicate that the determination of the resonance masses is also quite sensitive to different approaches used .
further studies of this issue are needed before a more unambiguous identification of the resonances can be made from the @xmath12 photoproduction process .
it should also be noted that our calculations based on regge trajectories differ in details from the results obtained in ref . @xcite . as mentioned above , these differences should ,
in part , be due to the different treatment ( in detail ) of the resonance current contributions .
( color online ) three fits based on regge trajectories resulting in different sets of the extracted resonance mass values .
all the fits include the @xmath3 and @xmath4 resonances as well as the mec .
the solid curves are the same ones shown in fig .
[ fig : regge ] with the mass values of @xmath199mev . for the dashed curves , the corresponding mass values are @xmath200mev .
the dash - dotted curves correspond to a fit with the mass values @xmath201mev . ]
( color online ) fig .
( a ) here corresponds to fig .
[ fig : photodiff](c ) , and ( b ) and ( c ) correspond to the two respective panels labeled ( c ) in fig .
[ fig : totang2 ] , the difference being that now the @xmath202 vertex functions of dipole form [ see eq .
( [ eq : dipoleff ] ) ] have been replaced by exponential form factors , @xmath203}$ ] , similar to ref .
@xcite ( however , with a different normalization ; see text ) .
, title="fig : " ] + ( color online ) fig .
( a ) here corresponds to fig .
[ fig : photodiff](c ) , and ( b ) and ( c ) correspond to the two respective panels labeled ( c ) in fig .
[ fig : totang2 ] , the difference being that now the @xmath202 vertex functions of dipole form [ see eq .
( [ eq : dipoleff ] ) ] have been replaced by exponential form factors , @xmath203}$ ] , similar to ref .
@xcite ( however , with a different normalization ; see text ) .
, title="fig : " ] + ( color online ) fig . ( a ) here corresponds to fig .
[ fig : photodiff](c ) , and ( b ) and ( c ) correspond to the two respective panels labeled ( c ) in fig .
[ fig : totang2 ] , the difference being that now the @xmath202 vertex functions of dipole form [ see eq .
( [ eq : dipoleff ] ) ] have been replaced by exponential form factors , @xmath203}$ ] , similar to ref .
@xcite ( however , with a different normalization ; see text ) .
, title="fig : " ] sibirtsev _ et al .
_ @xcite have also reported their study of the @xmath204 reaction quite recently .
in contrast to ref .
@xcite and the present work , they describe the forward rise of the angular distribution basically by the @xmath62 and @xmath24 meson exchanges in the @xmath5-channel .
they achieve this by including a ( @xmath5-dependent ) exponential form factor at the @xmath189 vertex .
moreover , the @xmath55 resonance was introduced in order to describe the steep rise and fall of the total cross section close to threshold .
we have also repeated our calculation employing an exponential form factor at the @xmath189 vertex instead of the dipole form factor ; the corresponding parameters are given in table [ tab3 ] .
the results are shown in fig .
[ fig : e_mr2](a ) which again exhibits the same features observed in the calculation using the dipole form factor [ see fig . [
fig : photodiff](c ) ] .
we were not able to reproduce their results using only the mesonic and @xmath55 resonance currents .
the chief difference between the results shown in fig .
[ fig : e_mr2 ] and those in ref .
@xcite is that we have used an exponential form factor normalized to unity at the on - mass - shell point @xmath205 , consistent with the kinematics at which the coupling constant @xmath206 is extracted . in ref .
@xcite , the form factor is normalized at @xmath207 instead .
figures [ fig : e_mr2](b ) and [ fig : e_mr2](c ) show the corresponding results for @xmath208 using the exponential form factor .
they are essentially the same as those using the dipole form factor .
the above considerations show that we arrive at the same conclusion , namely , that the interference between the meson exchange and resonance currents is the mechanism responsible for the angular distribution exhibited by the photoproduction data , irrespective of whether one uses the @xmath62 and @xmath24 meson exchanges in the @xmath5-channel ( with either dipole or exponential form factor ) or regge trajectories .
certainly , the problem of the regge trajectory versus form factor is an extremely important issue that needs to be addressed .
judging from our findings so far , it may well be that both simulate the same physics not accounted for explicitly in these calculations .
we have described consistently the @xmath209 and @xmath210 reactions within an approach based on a relativistic meson - exchange model of hadronic interactions .
the model includes the nucleonic and the mesonic , as well as the nucleon - resonance currents .
the photoproduction process is made gauge - invariant by adding a phenomenological contact current that parametrizes the effect of the final - state interactions .
the @xmath210 is described within the distorted - wave born approximation in which both the initial and final state @xmath79 interactions are taken into account explicitly . for @xmath12 photoproduction
, we have shown that the mesonic as well as the @xmath3 and @xmath4 resonance currents are important to describe the existing data .
our analysis , where the widths of the resonances were set to @xmath211mev , yields a position close to 1650 mev and 1870 mev for the @xmath3 and @xmath4 resonances , respectively .
this suggests that the former resonance may well be identified with the known @xmath8 resonance @xcite , whose quoted width is @xmath212mev .
the @xmath4 resonance , in contrast , does not correspond to any known resonance .
it is tempting to identify it with one of the missing resonances predicted at 1880 mev with the correspond width of @xmath213mev @xcite .
we emphasize , however , that one should be cautious with such an identification of the resonances .
as we have seen , the cross section data alone do not impose enough constraints for an unambiguous determination of the resonances .
to do so probably requires more exclusive data than just the cross sections .
moreover , the extracted values of the resonance masses are quite sensitive to the model used in the description of the photoproduction process .
in particular , the issue of regge trajectories versus conventional vector meson exchanges ( with form factors ) is of extreme importance .
these points require further investigation before a conclusive identification of the resonances can be made .
our study also shows that the nucleonic current should be relatively small .
indeed , the available photoproduction data prefer this current to be compatible with zero . in any case
, the @xmath9 coupling constant can not be much larger than @xmath10 .
the @xmath12 photoproduction reaction may impose a more stringent constraint on @xmath51 , provided one measures the cross sections at higher energies and backward angles . in this respect ,
as we have also shown , spin observables such as the photon asymmetry might be suited better than the cross sections .
it should be noted that the result pertaining here to the @xmath9 coupling constant is , of course , a model dependent one .
indeed , what is relevant in our calculations is the product of @xmath51 and the associated form factor .
we have also addressed the contradictory conclusions as to the underlying reaction mechanisms arrived in the recent work by two independent groups @xcite . in our consistent calculations , whether introducing a form factor at the electromagnetic vertex in the @xmath5-channel meson - exchange current or using the regge trajectories instead , one arrives at the same conclusion ; namely , the observed angular distribution is due to the interference between the @xmath5-channel and the nucleon resonance @xmath6- and @xmath7-channel contributions , irrespective of the particular approach one uses .
it is conceivable , therefore , that the phenomenological aspects of the various approaches ( including the present one ) may be simulating the same physics not taken into account explicitly . as for the @xmath11 reaction
, the present study yields the @xmath3 resonance as the dominant contribution to the production current .
the @xmath4 resonance , mesonic and nucleonic currents are much smaller than the @xmath3 resonance current . the combined analysis of this and the photoproduction reaction was crucial for these findings .
the details of the excitation mechanism of the @xmath3 resonance , however , are not constrained by the currently existing data . to learn more about the relevant excitation mechanism , observables other than the cross sections , such as the invariant mass distribution , are necessary @xcite , in addition to measuring the @xmath214 and/or @xmath215 process .
this process will help disentangle the isoscalar and isovector meson - exchange contributions .
the present model can not describe the flat angular distribution in @xmath216 measured by the disto collaboration @xcite at @xmath217mev , once the recently measured angular distribution by the cosy-11 collaboration @xcite at @xmath218mev is included in a global fitting for the relevant hadronic coupling parameters .
the calculated result exhibits a pronounced angular dependence . if we wish , however , a flatter angular distribution at @xmath217mev compatible with the disto data within the given experimental uncertainties can be achieved provided the @xmath218mev angular distribution data is excluded from fitting .
doing so , the predicted angular distribution at @xmath218mev comes out to be nearly completely isotropic .
although this seems still compatible with the cosy-11 data , the latter exhibits some angular dependence which is too disturbing to be ignored .
independent measurements of the angular distribution for excess energies @xmath219mev will help resolve this issue .
the authors would like to thank a. khoukaz and the cosy-11 collaboration for providing the new cosy-11 data prior to publication .
this work was supported by the cosy grant no.41445282(cosy-58 ) .
the work of h.h . was supported in part by grant no.de-fg02-95er-40907 of the u.s .
department of energy .
99 j. kogut and l. susskind , phys .
d * 10 * , 3468 ( 1974 ) ; phys . rev .
d * 11 * , 3594 ( 1975 ) ; e. witten , nucl .
* b149 * , 285 ( 1979 ) ; ann .
phys . * 128 * , 363 ( 1980 ) ; g. thooft , phys .
142 * , 357 ( 1986 ) .
t. feldmann , p. kroll , and b. stech , phys .
d * 58 * , 114006 ( 1998 ) ; phys .
* b449 * , 339 ( 1999 ) ; t. feldmann , int .
* a15 * , 159 ( 2000 ) ; t. feldmann and p. kroll , phys .
scripta * t99 * , 13 ( 2002 ) .
d. atwood and a. soni , phys .
b405 * , 150 ( 1997 ) ; phys .
lett . * 79 * , 5206 ( 1997 ) ; w .- s .
hou and b. tseng , _ ibid . _ * 80 * , 434 ( 1998 ) ; m.r .
ahmady , e. kou , and a. sugamoto , phys .
d * 58 * , 14015 ( 1998 ) ; e. kou , _ ibid .
_ * 63 * , 054027 ( 2001 ) .
et al . _ ,
* b438 * , 41 ( 1998 ) ; p. moskal _ et al .
_ , phys .
lett . * 80 * , 3202 ( 1998 ) ; p. moskal _
et al . _ ,
lett . * b474 * , 416 ( 2000 ) ; p. moskal _
_ , _ ibid .
_ * b482 * , 356 ( 2000 ) .
a. sibirtsev and w. cassing , eur .
j. * a2 * , 333 ( 1998 ) ; v. bernard , n. kaiser , and u. meiner , _ ibid . _ * a4 * , 259 ( 1999 ) ; e. gedalin , a. moalem , and l. razdolskaja , nucl .
* a650 * , 471 ( 1999 ) ; v. baru , j. haidenbauer , c. hanhart , a. kudryavtsev , and j. speth , eur . phys .
j. * a6 * , 445 ( 1999 ) . | the production of @xmath0 mesons in the reactions @xmath1 and @xmath2 is described consistently within a relativistic meson exchange model of hadronic interactions .
the photoproduction can be described quite well over the entire energy range of available data by considering an @xmath3 and a @xmath4 resonance , in addition to the @xmath5-channel mesonic current .
the observed angular distribution is due to the interference between the @xmath5-channel and the nucleon resonance @xmath6- and @xmath7-channel contributions .
our analysis yields positions close to 1650 mev and 1870 mev for the @xmath3 and @xmath4 resonances , respectively .
we argue that , at present , identifying these states with the known @xmath8 resonance and the missing @xmath4 resonance predicted at 1880 mev , respectively , would be premature .
it is found that the nucleonic current is relatively small and that the @xmath9 coupling constant can not be much larger than @xmath10 . as for the @xmath11 reaction ,
different current contributions are constrained by a combined analysis of this and the photoproduction reaction .
difficulties to simultaneously account for the 47-mev and 144-mev angular distributions measured by the cosy-11 and disto collaborations , respectively , are addressed . |
it has been reported that double - peaked broad balmer lines are seen in the optical spectra of some broad - line radio galaxies ( blrgs ) [ e.g. , osterbrock , koski , & phillips 1976 ; halpern & filippenko 1988 ; prez et al . 1988
; halpern 1990 ; veilleux & zheng 1991 ; colina , lpari , & macchetto 1991 ; eracleous & halpern 1994 ( eh94 ) ; halpern & eracleous 1994 ; sulentic et al . 1995
; halpern et al . 1996 ] and liners ( e.g. , storchi - bergmann , baldwin , & wilson 1993 ; bower et al . 1996 ; ho et al .
2000 ; shields et al .
since this remarkable feature may contain kinematical and geometrical information concerning the ionized gas in broad - line regions ( blrs ) , various models for the double - peaked emission line have been proposed up to now . among such models ,
the disk - blr model , in which the double - peaked profile is explained taking account of accretion disks as an emitter of broad emission lines , has been often investigated since prez et al .
( 1988 ) reported that the profile of the h@xmath1 emission of 3c 390.3 can be well fitted by a relativistic accretion disk model [ see also chen , halpern , & filippenko 1989 ( chf89 ) ; chen & halpern 1989 ( ch89 ) ] .
however , the gravitational energy released in a standard , geometrically - thin and optically - thick accretion disk is insufficient to account for the luminosity of the blr emission ( chf89 ; eh94 ) .
moreover , if the standard disk is assumed , the effective temperature in the region where the blr emission is expected to arise is calculated to be less than 5000 k , which is too low in blrs ( chf89 ; see also collin - souffrin 1987 ) .
therefore , an external heating source is required for an accretion disk to be a source of the blr emission .
when the mass accretion rate is small compared to the eddington value , ions in the inner disk become very hot because of lack of an efficient cooling process .
the resulting high pressure produces puffed - up structure with nearly spherical inflow , which is generally called advection - dominated accretion flow ( adaf ; e.g. , ichimaru 1977 ; rees et al . 1982 ; narayan & yi 1995 ) .
chf89 pointed out that this geometrically - thick adaf can illuminate an outer disk that remains geometrically thin ( see also ch89 ; shields et al .
this supplied energy enables the outer thin disk to radiate the blr emission , giving rise to the double - peaked broad balmer lines .
when the mass accretion rate is high , on the contrary , we see only single - peaked ( `` normal '' ) balmer lines which may arise in more distant regions from the nucleus compared to an accretion disk.these two situations are shown in figure 1 .
this scheme seems interesting because it explains why blrgs do not always show the double - peaked balmer lines .
since the adafs are optically thin , they do not exhibit a strong big blue bump ( bbb ) , i.e. , a thermal blue / uv component , in their spectra .
therefore , the above scheme ( hereafter `` illuminated - disk model '' ) predicts that an evident difference in the spectral energy distribution ( sed ) in the wavelength of uv to x - ray is expected between the blrgs with and without the double - peaked balmer lines ( e.g. , eh94 ) .
indeed , chf89 reported that there is no evidence for the strong bbb in arp 102b , which is a typical example of blrgs with double - peaked balmer lines ( see also edelson & malkan 1986 ) .
however , since most of the blrgs are rather faint , the difference in the sed is hard to be observed directly , except for a few exceptions .
eh94 mentioned that the difference in the sed between standard thin disks and adafs may affect the physical properties of ionized gas in narrow - line regions ( nlrs ) which are photoionized by the nuclear continuum radiation .
they found that the blrgs with double - peaked balmer lines exhibit stronger [ o i]@xmath06300 emission , which arises at nlrs , than those without double - peaked balmer lines .
this can be interpreted as the effect of the harder sed of adafs ( see also halpern & eracleous 1994 ; halpern 1999 ; ho et al .
however , there is no study to examine this interpretation quantitatively based on photoionization models .
therefore , we investigate the properties of the nlrs in the blrgs with / without the double - peaked balmer lines in this paper to examine whether or not the observed difference in the narrow emission - line flux ratios is owing to the difference in the sed .
this attempt seems crucially important toward the understanding of not only the origin of the double - peaked emission lines but also the configuration of blrs and accretion disks in active galactic nuclei ( agns ) .
eh94 classified 94 blrgs into five classes based on the profile of the h@xmath2 emission as follows ; ( 1 ) the blrgs with double - peaked h@xmath2 emission whose blue peak is stronger than the red peak and whose profile can be fitted by their relativistic disk models ( 12 blrgs ) , ( 2 ) the blrgs with double - peaked h@xmath2 emission whose blue peak is stronger than that of red peak but whose profile can not be fitted by their relativistic disk models ( 4 blrgs ) , ( 3 ) the blrgs with double - peaked h@xmath2 emission whose red peak is stronger than the blue peak ( 6 blrgs ) , ( 4 ) the blrgs without double - peaked h@xmath2 emission but their profile show a pronounced asymmetry or a single displaced peak ( 9 blrgs ) , and ( 5 ) the blrgs with normal single - peaked h@xmath2 emission ( 63 blrgs ) . here
it should be noted that eh94 adopted simple relativistic disk models described by chf89 and ch89 to fit the observed h@xmath2 profiles .
since these models can not explain the observed double - peaked profiles whose red peak is stronger than the blue peak , eh94 treated only blrgs categorized in the class of ( 1 ) as `` disk - like emitters '' .
eh94 compared the statistical properties of the 12 disk - like emitters with those of the other 82 blrgs in order to investigate the origin of the double - peaked balmer lines .
however , it is now recognized that some modified relativistic disk models can successfully reconstruct the profiles of double - peaked balmer lines which can not be explained by the models of chf89 and ch89 ; e.g. , disk plus hot patch models ( veilleux & zheng 1991 ; zheng , veilleux , & grandi 1991 ) , accretion disk with two - arm spiral waves ( chakrabarti & wiita 1994 ) , and elliptical disk models ( eracleous et al . 1995 ; see also syer & clarke 1992 ) . therefore
, we treat the first three classes of eh94 as `` double - peaked blrg '' regardless of the peak strength ratio .
these objects are expected to harbor a geometrically - thick adaf in their nuclei if the illuminated - disk model is the case .
then , we call the class ( 4 ) of eh94 as `` intermediate blrg '' .
it is not clear whether or not the objects in this class have the adafs and the illuminated thin disk in their nucleus . and
finally , we call the blrg with normal single - peaked h@xmath2 emission as `` single - peaked blrg '' . the objects in this class
are expected not to have an evolved geometrically - thick adaf .
note that a blrg , pictor a , is classified as a double - peak blrg in this paper though it was classified as ( 5 ) by eh94 , following some recent reports that pictor a also has double - peaked balmer emission ( halpern & eracleous 1994 ; sulentic et al . 1995 ) . in summary ,
the eh94 sample is divided into three groups : 23 double - peaked blrgs , 9 intermediate - type blrgs , and 62 single - peaked blrgs . in order to investigate the properties of ionized gas in nlrs for each group of blrgs
, we compiled data of forbidden emission - line flux ratios of the eh94 sample from the literature .
the compiled emission - line flux ratios are [ o i]@xmath06300/[o iii]@xmath05007 , [ o ii]@xmath03727/[o iii]@xmath05007 , [ s ii]@xmath36717,6731/[o iii]@xmath05007 , and [ ne iii]@xmath03869/[o ii]@xmath03727 , which are given in table 1 . for the case that a certain emission - line flux ratio of a certain object is given by more than one paper , the averaged value is given in this table . here
we do not use balmer lines , because the narrow balmer components of broad - line objects are often hard to measure correctly ( see , e.g. , nagao , murayama , & taniguchi 2001c ) .
for the same reason , estimating the amount of extinction by the balmer decrement method ( e.g. , osterbrock 1989 ) can not be adopted .
we can not use other methods ( e.g. , using the flux ratio of [ s ii]@xmath04071/[s ii]@xmath010320 ) because of small number of available emission - line data .
therefore , we do not make the reddening correction for the data .
note that the dust reddening of nlr emission of broad - line objects are small in general ( e.g. , de zotti & gaskell 1985 ; dahari & de robertis 1988 ) .
the data presented in table 1 are not corrected for galactic reddening .
the effect of dust extinction on our results is discussed in section 4.2 .
in figure 2 , we show the frequency distributions of the emission - line flux ratios of [ o i]@xmath06300/[o iii]@xmath05007 , [ o ii]@xmath03727/[o iii]@xmath05007 , [ s ii]@xmath36717,6731/[o iii]@xmath05007 , and [ ne iii]@xmath03869/[o ii]@xmath03727 , for the double - peaked blrgs , the intermediate blrgs , and the single - peaked blrgs , respectively .
the mean and median values of each flux ratio for each class of blrg are given in table 2 .
figure 2 and table 2 clearly show that the double - peaked blrgs exhibit larger ratios of [ o i]@xmath06300/[o iii]@xmath05007 than the single - peaked blrgs . in order to examine whether or not this difference is statistically real , we apply the kolmogorov - smirnov ( ks ) statistical test .
the ks test results in a probability that the observed frequency distributions of [ o i]@xmath06300/[o iii]@xmath05007 of the double - peak blrgs and the single - peak blrgs originate in the same underlying population of 1.21 @xmath4 10@xmath5 .
thus , we conclude that the flux ratio of [ o i]@xmath06300/[o iii]@xmath05007 of the double - peaked blrgs is statistically larger than that of the single - peaked blrgs .
note that this result is consistent with the earlier remark of eh94 .
although the frequency distribution of the flux ratio of [ o i]@xmath06300/[o iii]@xmath05007 for the intermediate blrgs appears to resemble that for the single - peaked blrgs and to be different from that for the double - peaked blrgs , it is not conclusive since the number of the sample is too small
. we can not conclude whether or not there is any statistical difference in the flux ratios of [ o ii]@xmath03727/[o iii]@xmath05007 , [ s ii]@xmath36717,6731/[o iii]@xmath05007 and [ ne iii]@xmath03869/[o ii]@xmath03727 among the classes of blrgs , because the number of the objects for which these flux ratios have been measured is too small .
we mention , however , that the flux ratio of [ o ii]@xmath03727/[o iii]@xmath05007 appears to show the same tendency as the flux ratio of [ o i]@xmath06300/[o iii]@xmath05007 , i.e. , the double - peaked blrgs exhibit larger flux ratios than the single - peaked blrgs .
note that we do not find significant difference in the flux ratio of [ s ii]@xmath36717,6731/[o iii]@xmath05007 between the double - peaked blrgs and the single - peaked blrgs .
this seems rather inconsistent with the result of eh94 that the disk - like emitters ( in the definition of eh94 ) present larger equivalent width of [ s ii]@xmath36717,6731 than the other blrgs though the equivalent width of [ o iii]@xmath05007 is similar between the disk - like emitters and the other blrgs .
eh94 mentioned that the higher [ o i]@xmath06300/[o iii]@xmath05007 ratio in double - peaked blrgs is attributed to the relatively harder ionizing sed arising from a nuclear adaf , which illuminates an outer thin disk and then causes the double - peaked profiles of balmer lines ( see also halpern 1999 ) . however , there are other possibilities that make the flux ratio of [ o i]@xmath06300/[o iii]@xmath05007 high . for instance , since the critical density of the [ o i]@xmath06300 transition ( 1.8 @xmath4 10@xmath6 @xmath7 ) is higher than that of the [ o iii]@xmath05007 transition ( 7.0 @xmath4 10@xmath8 @xmath7 ) , a high ratio of the [ o i]@xmath06300/[o iii]@xmath05007 is achieved when the typical gas density is high .
thus the difference in the flux ratio of [ o i]@xmath06300/[o iii]@xmath05007 between the double - peaked blrgs and the single - peaked blrgs may be attributed to the difference in the typical gas density in nlrs .
the lower ionization parameter ( i.e. , the ratio of the ionizing photon density to the hydrogen density ) in nlrs of the double - peaked blrgs may be also responsible for the observed difference in the narrow emission - line flux ratios .
these possibilities would not be discriminated without photoionization model calculations taking account of the sed difference predicted by the illuminated - disk model , which are presented in the following section .
as described in section 1 , the illuminated - disk model hypothesizes that the double - peaked blrgs harbor an adaf in their nucleus while the single - peaked blrgs do not . in order to investigate the effect of this difference on the gas in the nlrs and on the narrow emission - line flux ratios , we perform photoionization model calculations using the publicly available code @xmath9 version 94.00 ( ferland 1997 , 2000 ) . here
we assume uniform density gas clouds with a plane - parallel geometry .
the parameters for the calculations are ( i ) the hydrogen density of a cloud ( @xmath10 ) , ( ii ) the ionization parameter ( @xmath11 ) , ( iii ) the chemical composition of the gas , and ( iv ) the shape of the sed of the input continuum radiation .
we perform several model runs covering the following ranges of parameters : 10@xmath12 @xmath7 @xmath13 @xmath10 @xmath13 10@xmath14 @xmath7 and 10@xmath15 @xmath13 @xmath11 @xmath13 10@xmath16 .
we set the gas - phase elemental abundances to be the solar ones .
the adopted solar abundances relative to hydrogen are taken from grevesse & anders ( 1989 ) with extensions by grevesse & noels ( 1993 ) . for simplification , dust grains in nlrs
are not taken into account in our calculations .
we prepare the two sed templates for the input continuum radiation .
one is a typical sed for an optically - thick , geometrically - thin accretion disk .
this is characterized by a strong bbb in the wavelength range of uv to x - ray .
we adopt the empirically constructed sed ( see nagao , murayama , & taniguchi 2001a ) for this type of sed template ( hereafter `` sed with bbb '' ) .
this sed is described by the following function : @xmath17 ( see ferland 1997 ; nagao et al .
2001a ) . here
the following parameter set is adopted ( nagao et al .
2001a ) : ( i ) the infrared cutoff of the big blue bump component , @xmath18 = 0.01 ryd , ( ii ) the slope of the low - energy side of the big blue bump , @xmath19 = 0.5 , ( iii ) the uv to x - ray spectral index , @xmath20 = 1.35 , ( iv ) the slope of the x - ray power - law continuum , @xmath21 = 0.85 , and ( v ) the characteristic temperature of the big blue bump , @xmath22 = 490,000 k. note that the parameter @xmath23 in the equation ( 1 ) is determined from the adopted value of @xmath20 .
the last term in equation ( 1 ) is not extrapolated below 1.36 ev or above 100 kev .
below 1.36 ev the last term is simply set to zero . above 100 kev
the continuum is assumed to fall off as @xmath24 .
the other template is a typical sed for an optically - thin adaf .
the seds produced by the optically - thin adaf are expected to exhibit no strong bbb . recently
, many efforts have been made to estimate the sed generated by adafs theoretically ( e.g. , narayan & yi 1995 ; chen , abramowicz , & lasota 1997 ; narayan , kato , honma 1997 ; manmoto , mineshige , & kusunose 1997 ; kurpiewski & jaroszyski 1999 ; quataert & narayan 1999 ; manmoto 2000 ; kino , kaburaki , & yamazaki 2000 ) .
there are , accordingly , various shapes of the calculated seds . among them , we follow the sed presented by kurpiewski & jaroszyski ( 1999 ) for the template ( hereafter `` sed without bbb '' ) .
this sed exhibits no bbb component and is roughly described by single power - law continua in the range of 10@xmath25 hz to 10@xmath26 hz .
we choose the photon index of @xmath2 = 0.89 ( @xmath27 ) , which is predicted for the case that a non - rotating black hole is assumed ( see kurpiewski & jaroszyski 1999 ) .
this sed is expressed by the simple power - law spectra with the exponential cutoffs at 10@xmath15 ryd and 10@xmath28 ryd . in this way , we have prepared the two sed templates ( see figure 3 ) . note that these sed templates may be too simple to predict accurate emission - line spectra radiated from nlrs in blrgs .
however , these rough seds can be used to investigate the effect of the presence or absence of the bbb component , which is the main difference in the sed between the accretion disk with / without an adaf .
the calculations are stopped when the temperature falls to 3000 k , below which the gas does not contribute significantly to the observed optical emission - line spectra .
we show the results of the model calculations and compare them with the observations in figure 4 , which is a diagram of [ o i]@xmath06300/[o iii]@xmath05007 versus [ o ii]@xmath03727/[o iii]@xmath05007 .
this diagnostic diagram has been frequently used to discuss physical properties of gas in ionized regions ( e.g. , baldwin , phillips , & terlevich 1981 ) . as mentioned at section 2.2 ,
the plotted data are not corrected for dust extinction .
the data points will move on figure 4 as shown by the thick arrow if the extinction correction of @xmath29 = 1.0 mag is applied .
note that the direction of the extinction correction is perpendicular to the direction of the difference between the data of the double - peak and single - peak blrgs .
this suggests that the difference in the narrow emission - line flux ratios between the double - peaked and single - peaked blrgs is not caused mainly by the difference in the degree of the dust extinction .
although the data points are clearly separated between the double - peaked and single - peaked blrgs in figure 4 , there is an exceptional object , 3c 287.1 ; this object is classified as a single - peaked blrg while its locus on figure 4 is far from the other single - peaked blrgs .
this object , on the contrary , appears to belong to the class of the double - peaked blrg in terms of the narrow emission - line flux ratios .
we speculate that this is due to the misclassification of 3c 287.1 , because the red displaced peak of h@xmath2 is clearly seen between [ n ii]@xmath06583 and [ s ii]@xmath06717 in the optical spectra of this object , presented by eh94 . further observation for 3c 287.1 is necessary to examine this speculation .
as shown in figure 4 , the models adopting the sed without bbb predict a higher [ o i]@xmath06300/[o iii]@xmath05007 ratio by an order of magnitude than the models adopting the sed with bbb .
this tendency is consistent with the observations ; i.e. , the observed data of the single - peaked blrgs are explained by the models adopting the sed with bbb in the ranges of @xmath11 @xmath30 @xmath31 and 10@xmath28 @xmath7 @xmath32 @xmath10 @xmath32 10@xmath14 @xmath7 while the observed data of the double - peak blrgs appear to be explained introducing the models adopting the sed without bbb
. the larger flux ratios of [ o i]@xmath06300/[o iii]@xmath05007 in the models adopting the sed without bbb are attributed to the fact that the sed without bbb has harder spectra , which create larger partially - ionized regions in nlrs . to see the effect of the sed shape on the ionization structure of gas in nlrs , we show the ionization fractions of hydrogen and oxygen as functions of depth into the nebula for the cases of @xmath10 = 10@xmath12 @xmath7 and 10@xmath14 @xmath7 in figure 5 . here
@xmath11 = @xmath31 is adopted . as shown in the top panels of figure 5 , partially ionized regions become large when the sed without bbb is adopted ,
compared to the case that the sed with bbb is adopted .
accordingly , the volume emissivities of low - ionization emission lines such as [ o i]@xmath06300 and [ o ii]@xmath03727 are higher at larger radii from the ionization source in the case that the sed without bbb is adopted , as shown in the bottom panels of figure 5 . note that the ionization structure in the nebula is almost independent of the gas density as shown in figure 5 , though the volume emissivities of the forbidden emission lines depend on the density .
we also compare the observations with the model predictions in other diagnostic diagrams , which are presented in figure 6 .
they are diagrams of [ o i]@xmath06300/[o iii]@xmath05007 versus [ s ii]@xmath36717,6731/[o iii]@xmath05007 and [ o i]@xmath06300/[o iii]@xmath05007 versus [ ne iii]@xmath03869/[o ii]@xmath03727 .
similarly to the trend in figure 4 , the data of the double - peaked blrgs and the single - peaked blrgs are also well separated in these diagnostic diagrams .
this difference in the location of the data on the two diagrams is well explained by the idea that the gas in nlrs of the single - peaked blrgs is photoionized by the sed with bbb while that of the double - peaked blrgs is photoionized by the sed without bbb , which is expected in the framework of the illuminated - disk model .
it is noted that 3c 287.1 is again located far from the other single - peaked blrgs but beside the double - peaked blrgs in the two diagrams shown in figure 6 .
as presented in section 4.2 , the observed difference in the narrow emission - line flux ratios can be explained consistently by the photoionization models if we take account of the sed difference between the double - peaked and the single - peaked blrgs , which is predicted by the illuminated - disk model .
however , the sed without bbb may not be a unique way to explain the difference in the observed emission - line flux ratios between the double - peaked and the single - peaked blrgs .
for instance , systematic differences in the physical properties of gas in nlrs could be the origin of the difference in the narrow emission - line flux ratios .
as mentioned in section 3 , the difference in the flux ratio of [ o i]@xmath06300/[o iii]@xmath05007 can be also explained if the gas density in nlrs is systematically different between the two populations of blrgs .
this possibility is , however , rejected as we compare the observations with the model predictions in the diagnostic diagrams shown in figures 4 and 6 .
these diagnostic diagrams suggest that there is no difference in the gas density between the double - peaked and single - peaked blrgs if we adopted only the models with the sed with bbb ( 10@xmath28 @xmath7 @xmath32 @xmath10 @xmath32 10@xmath14 @xmath7 ) .
this means that the observed difference in the flux ratio of [ o i]@xmath06300/[o iii]@xmath05007 is not caused only by a difference in the gas density in nlrs .
there is another possibility which can explain the observed difference in the narrow emission - line flux ratios without introducing the sed difference . since the effect of the sed difference and the varying @xmath11 sequence
are degenerated as displayed in figures 4 and 6 , a systematic difference in the ionization parameter of gas in nlrs between the double - peaked and the single - peaked blrgs can be also responsible for the difference in the emission - line flux ratios .
note that pictor a and arp 102b , which are double - peaked blrgs , exhibit too strong low - ionization emission lines to be explained only by the effect of the sed difference . here
it should be noted that adafs are expected to radiate a small number of ionizing photons compared to the standard @xmath2 disks , which leads to a small ionization parameter of the gas in nlrs if the gas density and the distance from the nucleus of the nlrs are independent of the type of the accretion flow in their nuclei .
eh94 reported that the double - peaked blrgs exhibit weaker non - stellar continuum radiation in optical spectra than the single - peaked blrgs .
therefore , the difference in the narrow emission - line flux ratios between the double - peaked and the single - peaked blrgs may be attributed to the effects of both the sed shape and the ionization parameter .
unfortunately , the current information about the narrow emission - line spectra of blrgs is insufficient to examine the contribution of these two effects . in order to investigate this issue ,
future observations of optical spectra with sufficient quality ( i.e. , measurable rather faint emission lines ) for a large sample of blrgs is crucially necessary . for the sake of demonstration
, we show two diagnostic diagrams which are possibly useful to solve the degeneracy of the effects of the sed shape and of the ionization parameter in figure 7 , in which a diagram of [ ar iii]@xmath07136/[o iii]@xmath05007 versus [ o ii]@xmath03727/[o iii]@xmath05007 and a diagram of [ ar iii]@xmath07136/[o iii]@xmath05007 versus [ ne iii]@xmath03869/[o ii]@xmath03727 are displayed .
since the flux ratio of [ ar iii]@xmath07136/[o iii]@xmath05007 is almost independent of the shape of the sed especially in the range of @xmath33 , these diagnostic diagrams can be used to determine the ionization parameter .
then we will be able to discuss the effect of the sed shape solely , taking account of other information of the emission - line spectra .
the biconical radial outflow of gas in blrs has been often discussed as the possible origin of the double - peaked balmer lines as well as the illuminated - disk model ( e.g. , zheng , binette , & sulentic 1990 ; zheng et al . 1991 ) . here
we discuss whether or not the properties of nlr emission presented in this paper constrain this radial outflow model .
since this model does not require any special accretion mechanism to generate the double - peaked balmer lines , there is no reason to introduce the difference in the nuclear ionizing radiation between the double - peaked and the single - peaked blrgs .
however , narrow emission - line flux ratios would be different between the two groups of the blrg if the outflow in blrs of double - peaked blrgs affects the outer regions and then causes a shock wave in nlrs . since shocks generate a large partially - ionized region in the gas , strong low - ionization emission lines such as [ o i]@xmath06300 arise from such shock - heated gas ( mouri , kawara , & taniguchi 2000 and reference therein ) , which seems consistent with the fact that double - peaked blrgs exhibit stronger low - ionization emission lines than single - peaked blrgs .
note that stauffer , schild , & keel ( 1983 ) mentioned that the strong low - ionization emission lines exhibited in the spectrum of arp 102b , a prototype of double - peaked blrg , may be attributed to shock excitation . in order to examine whether or not the shock heating by the outflowing material of the double - peaked blrgs is responsible for the difference in the narrow emission - line flux ratios between the double - peaked and the single - peaked blrgs , we plot the predictions of the shock models presented by dopita & sutherland ( 1995 ) on the diagnostic diagrams of [ o i]@xmath06300/[o iii]@xmath05007 versus [ o ii]@xmath03727/[o iii]@xmath05007 , [ s ii]@xmath36717,6731/[o iii]@xmath05007 , and [ ne iii]@xmath03869/[o ii]@xmath03727 ( figure 8) .
the shock models appear to predict too large flux ratios of [ o ii]@xmath03727/[o iii]@xmath05007 and too small flux ratios of [ ne iii]@xmath03869/[o ii]@xmath03727 compared to the data of the double - peaked blrgs .
these results suggest that the difference in the narrow emission - line flux ratios is not caused by the shock - heated gas in nlrs .
therefore , the biconical radial outflow model seems less reasonable for the double - peaked balmer lines .
note , however , that this result is not conclusive because we can not exclude the possibility that the nlr emission of the double - peaked blrgs is contributed by the shock - heated gas .
for instance , the flux ratio of [ ne iii]@xmath03869/[o ii]@xmath03727 may be possibly lower in the double - peaked blrgs than in the single - peaked blrgs ( see table 2 ) , although it can not be examined statistically due to the small number of objects . then
, which emission - line flux ratio can be used to discriminate between shocks and photoionization ?
the flux ratio of [ o iii]@xmath04363/[o iii]@xmath05007 , which is one of temperature indicators , has been often regarded as a powerful tool to discriminate between shocks and photoionization ( e.g. , koski & osterbrock 1976 ; heckman 1980 ; ferland & netzer 1983 ; rose & cecil 1983 ; keel & miller 1983 ; rose & tripicco 1984 ) .
however , nagao , murayama , & taniguchi ( 2001b ) pointed out that high ratios of [ o iii]@xmath04363/[o iii]@xmath05007 do not always mean the shock excitation .
they showed that the photoionization models in which high - density gas clouds are taken into account can explain the observed high ratios of [ o iii]@xmath04363/[o iii]@xmath05007 ( see also nagao et al .
therefore , we focus on the emission - line flux ratio of [ s iii]@xmath09532/[s ii]@xmath36717,6731 .
this flux ratio is expected to be low when the gas is excited by shocks , because s@xmath34 ions cool predominantly by emission of uv lines due to the higher temperature and thus the [ s iii]@xmath09532 line becomes weak when the shock contributes to the excitation significantly ( e.g. , dopita 1977 ; daz , pagel , & wilson 1985 ; daz , terlevich , & pagel 1985 ; bonatto , bica , & alloin 1989 ; kirhakos & phillips 1989 ; simpson et al .
note that this tool can work even in the case that the gas density is rather high , being different from the flux ratio of [ o iii]@xmath04363/[o iii]@xmath05007 .
we compare the photoionization models adopting the sed with / without bbb with the shock models on the diagrams of [ s iii]@xmath09532/[s ii]@xmath36717,6731 versus [ o i]@xmath06300/[o iii]@xmath05007 , [ o ii]@xmath03727/[o iii]@xmath05007 , and [ s ii]@xmath36717,6731/[o iii]@xmath05007 ( figure 9 ) .
it is clearly shown that these diagnostic diagrams can be used to discriminate between shock and photoionization .
further observations of the [ s iii]@xmath09532 emission will make it clear whether or not the observed differences in the narrow emission - line flux ratios between the double - peaked and single - peaked blrgs are attributed to the shock excitation .
now we mention the intermediate blrgs .
on the diagnostic diagrams presented in figures 4 and 6 , the data of the intermediate blrgs appear to concentrate near the data of the single - peaked blrgs and to be separated from the data of the double - peaked blrgs .
moreover , it seems hard to explain the observed emission - line flux ratios of the intermediate blrgs by the photoionization models adopting the sed without bbb , since unreasonably large ionization parameters ( @xmath35 ) are required ( see figures 4 and 6 ) .
this suggests that the gas clouds in nlrs of the intermediate blrgs are photoionized by the ionizing continuum with bbb , which is expected in the case that the accretion disk in their nuclei is characterized by a standard optically - thick ( and geometrically - thin ) disk .
we thus speculate that an adaf occupies only a very small part of the inner accretion disk in the intermediate blrgs .
if this is the case , the thin disk can emit a substantial bbb emission ( see gammie , blandford , & narayan 1999 ) .
however , since the number of the intermediate blrgs whose nlr emission is investigated is too small , further observations are necessary for the discussion of this issue in detail .
recently , it has been reported that some of liners also show the double - peaked balmer lines ( e.g. , storchi - bergmann et al .
1993 ; bower et al . 1996
; ho et al . 2000 ; shields et al . 2001 ; see also barth et al .
2001 ; eracleous & halpern 2001 ) . since some lines of evidence that the liners harbor an adaf in their nucleus , the reason why the liners show lower ionization parameters than other agns such as seyfert nuclei may be the harder sed of the ionizing continuum radiation which is caused by the adafs . to investigate this issue , detailed photoionization model calculations for
the nlr emission of liners will be necessary .
we will investigate this issue in a subsequent paper .
based on the data compilation of narrow emission - line flux ratios from the literature , we find that the data of the double - peaked blrgs and those of the single - peaked blrgs are clearly separated on some diagnostic diagrams which consist of the narrow emission - line flux ratios of [ o i]@xmath06300/[o iii]@xmath05007 , [ o ii]@xmath03727/[o iii]@xmath05007 , [ s i]@xmath36717,6731/[o iii]@xmath05007 , and [ ne iii]@xmath03869/[o ii]@xmath03727 .
our photoionization model calculations show that the data of the single - peaked blrgs are well described by the models adopting the sed with bbb in the ranges of the parameters of @xmath11 @xmath30 @xmath31 and 10@xmath28 @xmath7 @xmath32 @xmath10 @xmath32 10@xmath14 @xmath7 while the data of the double - peaked blrgs are consistent with the models adopting the sed without bbb .
this is consistent with the illuminated - disk model that predicts the existence of an adaf , which illuminates the outer thin disk and causes the double - peaked emission lines , in a nucleus of a blrg with the double - peaked balmer lines .
we would like to thank gary ferland for providing his code @xmath9 to the public .
charles r. lawrence kindly provided us some useful information about spectroscopic properties of 3c 390.3 .
we also thank naohisa anabuki for fruitful discussion .
the anonymous referee and motoki kino gave us useful comments .
we also acknowledge yuko kakazu for her assistance .
ys is supported by a research fellowship from the japan society for the promotion of science for young scientists .
this work was financially supported in part by grant - in - aids for the scientific research ( nos .
10044052 , 10304013 , and 13740122 ) of the japanese ministry of education , culture , sports , science , and technology .
lcccccc /[o iii ] & 0.216 & 0.132 & 0.048 & 0.052 & 0.053 & 0.044 /[o iii ] & 0.395 & 0.288 & 0.075 & 0.074 & 0.218 & 0.114 /[o iii ] & 0.348 & 0.172 & 0.103 & 0.093 & 0.505 & 0.143 /[o ii ] & 0.762 & 0.515 & 1.581 & 1.565 & 1.095 & 0.846 lccccl 3c 17 & & 0.603 & & & 1 , 2 3c 59 & 0.070 & 0.139 & 0.169 & 0.478 & 3 iras 0236.6 - 3101 & 0.108 & & & & 4 pks 0340 - 37 & 0.110 & & & & 4 3c 93 & 0.077 & & & & 4 1e 0450.3 - 1817 & 0.731 & & & & 5 pictor a & 0.571 & 0.539 & 0.269 & & 1 , 6 , 7 , 8b2 0742 + 31 & 0.030 & & & & 4 pks 0857 - 19 & 0.186 & & & & 4 pks 1151 - 34 & 0.140 & 0.235 & & & 1 , 4 , 9 mrk 668 & 0.037 & & & & 4 3c 303 & 0.251 & & & & 4 3c 332 & 0.144 & 0.156 & 0.172 & 0.372 & 3 3c 382 & 0.074 & 0.288 & 0.164 & 0.552 & 2 , 10 3c 390.3 & 0.113 & 0.092 & 0.120 & 1.767 & 2 , 10 , 11 , 12arp 102b & 0.803 & 1.197 & 1.248 & 0.152 & 13 pks 1739 + 18 & 0.154 & & & & 4 pks 1914 - 45 & 0.123 & & & & 4 pks 2300 - 18 & 0.171 & 0.308 & 0.294 & 1.250 & 8 , 14 pks 0202 - 76 & 0.055 & & & & 4 4c 5.38 & 0.049 & & & & 4 pks 1232 - 24 & 0.017 & & & & 4 3c 227 & 0.071 & 0.088 & 0.093 & 1.268 & 1 , 10 , 15 4c 73.18 & 0.055 & 0.064 & 0.142 & 1.911 & 16 3c 445 & 0.042 & 0.074 & 0.073 & 1.565 & 1 , 8 , 10 , 17 b2 0110 + 29 & 0.019 & & & & 4 3c 48 & & 0.895 & 3.739 & 0.714 & 18 , 19 phl 1093 & @xmath360.306 & & & & 20 3c 61.1 & 0.047 & 0.174 & 0.209 & 0.457 & 16 pks 0214 + 10 & 0.042 & & & & 4 3c 109 & & 0.093 & & & 2 3c 120 & 0.035 & 0.097 & 0.060 & 2.376 & 1 , 21 , 22 pks 0736 + 01 & & @xmath360.080 & & & 1 3c 206 & 0.070 & 0.093 & & 2.222 & 4 , 20 4c 9.35 & 0.036 & & & & 4 3c 234 & 0.014 & 0.104 & 0.045 & 1.032 & 2
. 3 pks 1011 - 282 & & 0.093 & & & 23
3c 249.1 & & 0.222 & 0.138 & 0.787 & 18 , 24 , 25pks 1101 - 32 & 0.046 & 0.058 & & 1.578 & 4 , 26 pks 1217 + 02 & @xmath360.119 & 0.093 & & 1.915 & 20 b2 1223 + 25 & 0.012 & 0.117 & & 1.539 & 4 , 20 , 23 3c 273 & & @xmath360.161 & 0.148 & & 1 , 18 3c 277.1 & 0.040 & 0.324 & & 0.396 & 4 , 27 3c 287.1 & 0.241 & 0.720 & 0.348 & 0.204 & 3 pks 1355 - 41 & & 0.116 & & & 1 pks 1417 - 19 & 0.038 & 0.107 & 0.123 & 1.120 & 3 , 14 , 28 pks 1421 - 38 & 0.078 & & & & 4 4c 37.43 & 0.029 & 0.218 & 0.073 & 0.406 & 4 , 18 4c 35.37 & & 0.356 & & & 3 3c 323.1 & 0.030 & 0.112 & & 0.846 & 20 , 27 b2 1719 + 35 & 0.060 & & & & 4 mc 1745 + 16 & 0.046 & & & & 4 pks 2135 - 14 & @xmath360.052 & 0.091 & & 0.839 & 1 , 20 pks 2139 - 04 & 0.055 & & & & 4 pks 2208 - 13 & 0.073 & & & & 4 pks 2227 - 399 & 0.058 & & & & 4 pks 2302 - 71 & 0.047 & & & & 4 pks 2349 - 01 & 0.039 & 0.271 & 0.168 & & 3 | in order to examine the prediction that the broad - line radio galaxies ( blrgs ) with double - peaked balmer lines harbor an accretion disk characterized by an advection - dominated accretion flow ( adaf ) in their nuclei , we investigate narrow emission - line flux ratios of the narrow - line regions which are photoionized by the nuclear continuum radiation .
we compile data from the literature and confirm the pioneering work of eracleous & halpern that the blrgs with the double - peaked balmer emission exhibit larger flux ratios of both [ o i]@xmath06300/[o iii]@xmath05007 and [ o ii]@xmath03727/[o iii]@xmath05007 than the blrgs without the double - peaked balmer emission . to examine whether or not these properties are attributed to the difference in the spectral energy distribution ( sed ) of the ionizing radiation between the blrgs with and without the double - peaked balmer emission , we perform photoionization model calculations using two types of input continuum radiation ; one has the strong big blue bump which is expected for standard optically - thick accretion disks and another does not exhibit a strong big blue bump as expected for optically - thin adafs .
we find that the data of the blrgs with the double - peaked balmer lines are consistent with the models adopting the sed without a strong big blue bump while the data of the blrgs without the double - peaked emission lines are well described by the models adopting the sed with a strong big blue bump . on the other hand , the observed differences in the nlr emission
is hard to be explain by the difference in the contribution of shocks .
these results support the idea that the double - peaked balmer lines arise at an outer region of an accretion disk which is illuminated by an inner , geometrically - thick adaf . |
although the fractional quantum hall effect is originally a two - dimensional phenomenon @xcite , multi - layer systems have been studied as natural development @xcite .
multi - layer systems show some unusual effects because of new additional degrees of freedom , namely the inter - layer coulomb interaction and the inter - layer electron tunneling .
for example , in the case of the simplest double - layer system , the `` even denominator '' fractional quantum hall effect with the landau level filling @xmath2 has been observed @xcite , and is interpreted in terms of halperin s @xmath3 state @xcite . one may then naturally ask what happens in infinite multi - layer systems .
experimentally , the integer quantum hall effect is observed in a 30-layer gaas / algaas superlattice system @xcite and also in organic conductors @xmath4 @xcite and bedt - ttf salts @xcite .
motivated by these works we study the ground state of the infinite - layer system under a strong magnetic field and discuss quantum phase transitions between incompressible and compressible states .
we can have the following naive expectation in advance of quantitative analysis : in the limit of infinite inter - layer distance @xmath5 , layers are completely independent of each other and the laughlin state is realized for appropriate fillings , e.g. @xmath6 where @xmath7 is an odd integer .
however for @xmath8 where @xmath9 is the magnetic length , other fractional quantum hall states can be realized because of the inter - layer coulomb interaction .
furthermore three - dimensional states are also realizable with increasing electron tunnelings between layers .
the fractional quantum hall states are incompressible , but the last three - dimensional state is compressible because the bloch states are formed perpendicular to layers and the fermi surface should appear . therefore a transition between incompressible and compressible states is expected at a certain value of @xmath5 , provided that @xmath5 is a variable parameter . based on the above expectation we introduce three typical trial states for the ground state with the lowest landau level filling @xmath10 and compute energies of them in order to find which one is the ground state . under a strong magnetic field
we restrict the single - particle states within the lowest landau level , and consider only the coulomb energy and tunneling energy .
we calculate the coulomb energy by the variational monte carlo ( vmc ) method for the two - dimensional trial states @xcite , while the hartree - fock approximation is used for the three - dimensional trial state .
we consider a system with the area @xmath11 of each layer and the height @xmath12 as shown in fig .
[ config ] .
the total number @xmath13 of layers is given by @xmath14 .
then , with the total number @xmath15 of electrons in the system , we define the electron number per layer @xmath16 by @xmath17 . taking into account the layer structure of the system
, we assume the positive charge distribution @xmath18 this positive charge density corresponds to the uniform distribution in each layer and the discrete distribution along the layer . under a magnetic field
@xmath19 the landau level degeneracy @xmath20 is given by @xmath21 where @xmath22 is the magnetic length .
the average filling factor of each layer @xmath23 is defined as @xmath24 . under the strong magnetic field with @xmath25
we may deal with only the lowest landau level ( lll ) explicitly and neglect higher levels .
we will discuss possible ground states in the subspace of the lll in the following .
now we write down the hamiltonian of the system with @xmath26 as @xmath27 here the one - body part is given by @xmath28 where @xmath29 and @xmath30 are the effective masses in the layer and along the @xmath31-direction respectively . the squared effective charge @xmath32 is defined by use of the dielectric constant @xmath33 as @xmath34 .
@xmath35 is the binding potential of electrons to layers ( like a multi square well potential ) .
we choose two types of configuration as follows : one is the circular shape and the other is the rectangular shape .
these shapes are chosen for convenience of calculations .
let us consider a pile of infinitely large layers ( @xmath36 ) with the circular symmetry and take the symmetric gauge .
for the ( @xmath37)-space this geometry is the same as the one that laughlin chose @xcite .
we adopt for this geometry the periodic boundary condition in the @xmath31-direction .
this geometry is convenient for description of two - dimensional states and also convenient for the vmc calculation @xcite .
if we assume that electrons do not transfer between layers , then the one - particle state is written as @xmath38 \phi_{0}\left(z - ld_{s}\right),\ ] ] where @xmath39 is the complex coordinate in the ( @xmath37)-space and @xmath40 defined by ( [ phizero ] ) is the normalized wave function localized at the layer @xmath41 . since the uniform positive charge is inconvenient for finite particles which we actually study numerically , we take an alternative positive charge density @xmath42 which has the same distribution as that of electrons .
namely with @xmath43 this positive charge distribution tends to the uniform one in the infinite - size limit .
we now discuss the coulomb potential with periodic boundary condition along the @xmath31-axis .
taking account of mirror images we obtain the potential as @xmath44^{1/2}}.\ ] ] the summation over @xmath45 is divergent logarithmically .
this divergence should be canceled by the attractive part in the final result . in this paper
we replace it by the following one : @xmath46^{1/2}}.\ ] ] this potential has the right periodicity along the @xmath31-axis and becomes equivalent to the coulomb potential in the limit of @xmath47 . the hamiltonian with finite @xmath15
is represented as follows : @xmath48 the expectation value is given by @xmath49 where @xmath50 is a constant representing the single - particle energy .
it is given by @xmath51 where we have assumed that the tunneling along the @xmath31-direction is suppressed .
next we consider a pile of rectangular ( @xmath52 ) layers .
we impose the periodic boundary condition in each ( @xmath37 and @xmath31 ) direction and choose the landau gauge .
we obtain @xmath53 . in the case of three - dimensional states ,
this geometry is convenient . when there are tunnelings , the one electron state is written as @xmath54 where @xmath55 ( @xmath56 ) indicates the degree of freedom in the lll , and @xmath57 ( @xmath58 ) is the wave number along the @xmath31-direction .
the @xmath59-dependent part @xmath60 is the landau wave function @xcite : @xmath61 , \label{landauwf}\ ] ] where @xmath62 and @xmath63 . for the @xmath31-dependent part @xmath64 we make the bloch function by use of @xmath65 as @xmath66 where we take the following localized function : @xmath67 with @xmath68 being positive infinitesimal . under the tight - binding approximation
@xmath69 becomes the eigenstate of the one - particle hamiltonian @xmath70 .
we introduce the creation and annihilation operators , @xmath71 and @xmath72 , corresponding to @xmath69 .
then the total hamiltonian of this system is given in second quantized form by @xmath73 \right ] .
\label{hamiltonian2}\end{aligned}\ ] ] here @xmath74 is the eigenvalue of the one - electron state with the inter - layer transfer @xmath75 , @xmath76 is the riemann s zeta function , @xmath77 is the fourier coefficient of the coulomb potential , and @xmath78 is the density operator defined by @xmath79 where @xmath80 is the electron field operator projected onto the lll .
we introduce three trial states for the ground states with @xmath10 of the hamiltonians ( [ hamiltonian1 ] ) and ( [ hamiltonian2 ] ) .
as mentioned in the previous section , we take ( [ hamiltonian1 ] ) for states without the tunneling ( two - dimensional states ) and ( [ hamiltonian2 ] ) for states with the tunneling ( three - dimensional states ) . here
we remark on the index of position coordinates .
the position coordinates in the hamiltonian ( [ hamiltonian1 ] ) have the index @xmath81 ( @xmath82 ) of electrons .
however when we refer to two - dimensional states , we make the position coordinates have the layer index additionally because electrons belonging to different layers are distinguishable .
hence the position coordinates have the two indices @xmath41 and @xmath83 where @xmath41 ( @xmath84 ) is the layer index and @xmath83 ( @xmath85 ) is the particle index in a layer .
indices @xmath86 and @xmath87 are related to each other by @xmath88 .
the first trial state is a two - dimensional one without the tunneling represented by @xmath89 \phi\left(\left\{\left\{z_{l p}\right\}\right\}\right ) , \label{mltlaughlin}\ ] ] where @xmath90 and @xmath91 describe the motion in the ( @xmath37)-space and along the @xmath31-direction respectively .
they are given by @xmath92,\ ] ] @xmath93 here @xmath94 denotes @xmath95 and @xmath96 does @xmath97 .
we call the state ( [ mltlaughlin ] ) the `` multi - laughlin state '' .
it is the direct product of the laughlin state with @xmath10 @xcite in each layer and is expected to be stable for a large inter - layer distance : @xmath98 .
the second trial state is also a two - dimensional one without the tunneling .
it is represented by @xmath99 where @xmath100.\ ] ] we call this state the `` ( 111 ) state '' @xcite .
this state is expected to be stable for @xmath101 where the inter - layer coulomb interaction becomes comparable to the intra - layer one .
the numbers ( 111 ) indicate the exponents of the jastrow factor which controls the strength of the electron correlation .
the first `` 1 '' is the exponent of the jastrow factor between a given layer and the nearest lower layer , the second `` 1 '' is that within the layer , and the last `` 1 '' is that between the layer and the nearest upper layer
. generally we can define ( @xmath102 ) state for our system . for example
( 131 ) state is possible for @xmath103 , and ( 212 ) state is also possible for the same @xmath23 .
physically ( 212 ) state should be unstable because the inter - layer correlation becomes larger than the intra - layer one in this state . in the case of @xmath10 only the ( 111 )
state is allowed within the subspace of halperin - type states .
the third state is expected to be stable at @xmath104 when electrons transfer between layers .
we call such a state the `` itinerant state '' and define it as follows . with the tunneling , bloch waves are formed along the @xmath31-direction .
then the simplest candidate for the ground state is the one with the smallest kinetic ( tunneling ) energy .
namely the cosine band in the @xmath31-direction is filled up to the fermi level @xcite , with the filling of the landau level for each wave number being unity below the fermi level and zero above it .
we can write it by use of electron creation operators as @xmath105 where @xmath106 is the fermi wave number given by @xmath107 .
although we assume @xmath10 in this paper , the itinerant state can be defined for all @xmath108 $ ] .
we use the vmc method to calculate the expectation value of the coulomb energy numerically for the multi - laughlin state and the ( 111 ) state .
technical details are explained in appendix . in the case of the multi - laughlin state ,
@xmath109 are reduced to the sum of the two - dimensional result simply because all layers are independent of each other , and the inter - layer coulomb energy is exactly canceled .
we obtain @xmath110 on the other hand such cancellations do not occur for the ( 111 ) state , and the calculation becomes more complicated because of the inter - layer correlation .
the energy is given by @xmath111 figure [ graph1 ] shows the size dependence of the coulomb energy of the multi - laughlin state .
the extrapolated value is in good agreement with that in the literature @xcite .
the dependence of the coulomb energy for the ( 111 ) state on the inter - layer distance is shown in fig.[graph2 ] .
the calculation was performed for three cases of finite size systems : @xmath112 , @xmath113 and @xmath114 with @xmath115 .
we find that the energy decreases with decreasing inter - layer distance .
the dependence on @xmath16 is small in the region of small inter - layer distance .
we calculate the energy of the itinerant state for the hamiltonian ( [ hamiltonian2 ] ) with use of the hartree - fock approximation .
since we take the homogeneous trial state , the mean field is given by @xmath116 the mean field for the interaction part of the hamiltonian ( [ hamiltonian2 ] ) consists of the direct and the exchange interactions .
of these the energy due to the direct term is written as @xmath117 where we used the relation @xmath118 valid in the macroscopic limit .
here we have @xmath119 and @xmath120 with @xmath121 integers .
the direct term cancels exactly with the coulomb energy of the positive charge because of homogeneity of the itinerant state within each layer .
therefore the hartree - fock hamiltonian is given by @xmath122 where the matrix element for the exchange mean - field is given by @xmath123.\end{aligned}\ ] ] the energy of the itinerant state is obtained as @xmath124.\end{aligned}\ ] ] the region of summation over @xmath125 in the above formula is shown in fig . [ baaiwake ] . in order to derive the sum @xmath126
we analyze the cases as follows : ( 1 ) @xmath127 for @xmath128 , ( 2 ) @xmath129 for @xmath130 , ( 3 ) @xmath131 for @xmath132 , and ( 4 ) @xmath127 for @xmath133 . as a result
we summarize the four cases into a single expression given by @xmath134 therefore the energy per particle which contains the kinetic term is given in the thermodynamic limit by @xmath135 \\ & = & \alpha + \varepsilon_{kin } + \varepsilon_{int}^{hf}. \nonumber\end{aligned}\ ] ] at this stage we let @xmath136 .
it can be shown that the contribution of large @xmath137 ( @xmath138 ) to the integration is negligible . to see this we observe that @xmath139 becomes @xmath140 and summation over @xmath141 converges in the limit of @xmath136 .
therefore , @xmath142 becomes @xmath143 with use of the following formula : @xmath144 we can accomplish the summation over @xmath141 as @xmath145 consequently we obtain the following result : @xmath146 where @xmath147 denotes the inter - layer distance in units of the magnetic length .
figure [ graph3 ] shows @xmath142 as a function of the inter - layer distance .
we remark that @xmath142 decreases with decreasing @xmath5 .
this result originates from the fact that the inter - layer coulomb interaction becomes more effective as layers get closer .
we first compare the coulomb energy per particle of the three states .
figure [ graph4 ] shows the results .
the energy of the ( 111 ) state is obtained by the vmc calculation for 5 layers with 42 particles in each layer . for the multi - laughlin and
the itinerant states the result is shown for the infinite - size system .
we remark the energy of the multi - laughlin state does not depend on the inter - layer distance since all layers are independent of each other and electrons feel other layers neutral . in the region of @xmath148
the coulomb energy of the multi - laughlin state is the lowest .
this is evident because the system becomes a pile of almost independent two - dimensional layers in such a region . in the region of @xmath149 the ( 111 )
state has the lowest coulomb energy .
this result originates from the fact that the ( 111 ) state gains the inter - layer coulomb energy at the cost of the intra - layer one in the region where intra- and inter - layer interactions are comparable . with respect to the itinerant state , the coulomb energy decreases in the region of @xmath149 .
there is however no region where it is the lowest .
so far we have considered only the coulomb energy .
we must take into account the tunneling energy of the itinerant state in order to compare the total energies of the three states .
let us assume that the transfer energy @xmath150 depends on the inter - layer distance as @xmath151 where @xmath152 and @xmath153 are constants .
then as shown in fig .
[ graph5 ] we find a region of @xmath5 where the itinerant state has the lowest total energy . here
we put appropriate values into @xmath152 and @xmath153 , namely @xmath154 for a case shown as @xmath155 , and @xmath156 for another case shown as @xmath157 . the parameters for @xmath157 with @xmath158
are chosen to reproduce the band width of @xmath159 at @xmath160 which seems appropriate for a gaas / algaas superlattice @xcite .
on the basis of the comparison we propose a phase diagram as shown in fig .
[ graph6 ] for the ground state of the present system in the plane of @xmath150 and @xmath5 .
if we assume the relation ( [ tofd ] ) between @xmath150 and @xmath5 in the diagram , possible states follow a curve in fig .
[ graph6 ] as @xmath5 is varied . depending on the parameters @xmath152 and @xmath153
, there are either single transition ( @xmath155 ) or double transitions ( @xmath157 ) . consequently the phase diagram fig .
[ graph6 ] is interpreted as follows . 1 . in the region @xmath161 ,
the system is a pile of independent two - dimensional layers and the incompressible multi - laughlin state is stable .
2 . in the region of @xmath149 and
the tunneling energy is small ( @xmath162 ) , the ( 111 ) state which is also incompressible is stabilized instead of the multi - laughlin state by the inter - layer correlation .
when the tunneling energy is large , the compressible itinerant state is stabilized by competition of the tunneling energy and the coulomb energy .
we compared the result of the ( 111 ) state for a finite size system and that of the multi - laughlin and the itinerant states for the infinite size system .
the finite size ( @xmath16 and @xmath13 ) correction should be taken into account in the ( 111 ) state for more quantitative discussion . while we have no information with respect to the dependence on @xmath13 , the following remark is given from fig .
[ graph2 ] about the dependence on @xmath16 .
when one increases @xmath16 the coulomb energy much increases in the region of @xmath161 . while for @xmath8 the coulomb energy looks almost independent of @xmath16 within the error bars . since the boundaries of the ( 111 ) state occurs with @xmath163 in fig .
[ graph6 ] , the finite - size correction should not change the results qualitatively .
it should be possible to observe transitions in actual systems by adding a pressure to the system or tilting the magnetic field . since adding a pressure corresponds to control of @xmath5 , a change of parameters along a curve like @xmath155 or @xmath157 in the diagram fig .
[ graph6 ] is expected . on the other hand tilting
the magnetic field corresponds to control of @xmath150 independent of @xmath5 because the horizontal component of the magnetic field works on electrons to localize in the vertical direction and reduces @xmath150 .
then the parameter change along a vertical straight line at a certain @xmath5 in fig .
[ graph6 ] is expected .
the transition between incompressible and compressible states is analogous to the mott transition in the hubbard model .
namely the incompressible multi - laughlin and ( 111 ) states correspond to an insulator , while the compressible itinerant state corresponds to a metal .
the transition between incompressible and compressible states occurs due to competition of the transfer ( tunneling ) and the on site ( intra - layer ) coulomb energy . in our model
we estimate that the critical transfer energy is almost @xmath164 .
although we discussed only three trial states for @xmath0 in the present study , we can not exclude the possibility of other states . in terms of two - dimensional incompressible states for @xmath0 ,
it is difficult to construct states more stable than the multi - laughlin and the ( 111 ) states within the simple family of jastrow - type functions .
one may however ask whether other states with three - body correlation or even more correlations are more favorable energetically . for three - dimensional compressible states ,
the itinerant state is taken to be the simplest homogeneous state .
then inhomogeneous states like cdw @xcite should also be discussed for more refined argument .
furthermore we have neglected the spin degree of freedom which may be relevant to some systems .
for example the sdw state , which accounts for the qhe of @xmath4 @xcite , can be a relevant state depending on parameters in the system .
one of the authors ( s. s ) would like to thank h. yokoyama and s. tokizaki for instruction of the vmc method and useful discussion , and m. arikawa for pertinent suggestions .
we describe the method of numerical calculation for the ( 111 ) state in this appendix .
the expectation value of the coulomb energy for the ( 111 ) state is represented by @xmath165 we can obtain the numerical value of @xmath166 for finite size ( finite particles ) systems by the vmc method . in evaluating the second term with minimum statistical errors , we separate the variables in it and reduce the expectation value of the quantity between two bodies to that of one body . for this purpose
we introduce the fourier transform of the coulomb potential @xmath167 , one body distribution function @xmath168 , and its fourier transform @xmath169 as follows : @xmath170 @xmath171,\ ] ] @xmath172 @xmath173 where we wrote @xmath174 as @xmath175
. then the second term @xmath176 in ( [ 11hcoulomb ] ) is arranged into @xmath177 we use this formula to obtain the numerical value of @xmath176 .
figure [ gq13 ] shows an example of numerical results .
the integral with respect to @xmath178 is performed by the trapezoidal rule because only discrete values of @xmath179 are obtained by the vmc calculation .
we cut off the upper region in the integral at a finite value @xmath180 because the integrand converges to zero near @xmath181 and its contribution to the integral is negligible in the region @xmath182 .
the e - mail address is [email protected] . , edited by r.e .
prange and s.m .
girvin ( springer , new york , 1990 ) 2nd ed ; _ the quantum hall effects _
, t. chakraborky and p. pietilinen ( springer , new york , 1988 ) .
laughlin , phys .
* 50 * , 1395 ( 1983 ) .
tsui , h.l .
strmer , and a.c .
gossard , phys .
* 48 * , 1559 ( 1982 ) .
, edited by s. das sarma and a. pinczuk ( wiley , new york , 1997 ) .
d. yoshioka , a.h .
macdonald , and s.m .
girvin , phys .
* b39 * , 1932 ( 1989 ) .
x. qiu , r. joynt and a.h .
macdonald , phys .
, 11943 ( 1989 ) .
a.h . macdonald , surf .
sci . * 229 * , 1 ( 1990 ) .
song he , s. das sarma , and x.c .
xie , phys .
rev . * b47 * , 4394 ( 1993 ) .
eisenstein , g.s .
boebinger , l.n .
pfeiffer , k.w .
west , and song he , phys .
, 1383 ( 1992 ) .
strmer , j.p .
eisenstein , a.c .
gossard , w. wiegmann , and k. baldwin , phys .
lett . * 56 * , 85 ( 1986 ) .
chamberlin , m.j .
naughton , x. yan , l.y .
chiang , s. -y .
hsu , and p.m. chaikin , phys .
lett . * 60 * , 1189 ( 1988 ) .
cooper , w. kang , p. auban , g. montambaux , and d. jrome , phys .
lett . * 63 * , 1984 ( 1989 ) .
hannahs , j.s .
brooks , w. kang , l.y .
chiang , and p.m. chaikin , phys .
lett . * 63 * , 1988 ( 1989 ) .
w. kang , s.t .
hannahs , and p.m. chaikin , phys .
lett . * 70 * , 3091 ( 1993 ) .
n. harrison , a. house , m.v .
kartsovnik , a.v .
polisski , j. singleton , f. herlach , w. hayes , and n.d .
kushch , phys .
77 * , 1576 ( 1996 ) .
macmillan , phys .
rev . * 138 * a442 ( 1965 ) ; d. ceperley , g.v .
chester , and m.h .
kalos , phys .
* b16 * , 3081 ( 1977 ) .
r. morf and b.i .
halperin , phys .
rev . , 2221 ( 1986 ) .
d. yoshioka , b.i .
halperin , and p.a .
lee.phys . rev .
lett . * 50 * , 1219 ( 1983 ) .
the structure of the band may be modified in the presence of the magnetic field as discussed by m. abolfath et al .
b55*,10643(1997 ) , for the single wide quantum well .
this effect enters through modification of the transfer integral @xmath150 in our formulation .
we do not consider the change explicitly since the effect is shown to be small if the magnetic field is perpendicular to the layers .
y. kuramoto , j. phys .
. jpn . * 44 * , 1572 ( 1978 ) .
b. i. halperin , jpn .
* 26 - 3 * 1913 ( 1987 ) , and references therein . | possible phase transitions between incompressible quantum hall states and compressible three - dimensional states are discussed for infinite - layer electron systems in strong magnetic field . by variational monte carlo calculation , relative stability of some trial states is studied .
if the inter - layer distance is large enough , the laughlin state is stabilized for the landau - level filling @xmath0 of each layer .
when the inter - layer distance is comparable to the magnetic length , the laughlin state becomes unstable against a different fractional quantum hall state with inter - layer correlation , and/or against a three - dimensional compressible state .
it is discussed how the quantum phase transition between them can be controlled in actual systems .
# 1@xmath1#1 @align#1#2.6ex |
let @xmath0 be a smooth complex projective variety which admits a morphism with connected fibers @xmath5 onto a normal variety @xmath6 such that the anticanonical bundle @xmath7 is @xmath8-ample , @xmath9 and @xmath10 ( _ i.e. _ an elementary extremal contraction of fiber type ) .
+ it is well known , by fundamental results of mori theory , that through every point of @xmath0 there is a rational curve contracted by @xmath8 .
the numerical classes of these curves lie in an extremal ray of the cone @xmath4 . by taking a covering family of such curves
one obtains a _ quasi - unsplit _ family of rational curves , _ i.e. _ a family such that the irreducible components of all the degenerations of curves in the family are numerically proportional to a curve in the family .
it is very natural to ask if the converse is also true : given a covering quasi - unsplit family @xmath11 of rational curves , is there an extremal elementary contraction which contracts all curves in the family or , in other words , does the numerical class of a curve in the family span an extremal ray of @xmath4 ? as proved in @xcite ( see also @xcite and @xcite ) there is always a rational fibration , defined on an open set of @xmath0 , whose general fibers are proper , which contracts a general curve in @xmath11 .
more precisely , a general fiber is an equivalence class with respect to the relation induced by the closure @xmath12 of the family @xmath11 in the chow scheme of @xmath0 in the following way : two points @xmath13 and @xmath14 are equivalent if there exists a connected chain of cycles in @xmath12 which joins @xmath13 and @xmath14 . by a careful study of this fibration and of its indeterminacy locus ,
a partial answer to this question has been given in ( * ? ? ?
* theorem 2 ) ; namely , if the dimension of a general equivalence class is greater than or equal to the dimension of the variety minus three then the numerical class of a general curve in the family spans an extremal ray of @xmath4 . before the results in @xcite a special but very natural situation
in which the question arises has been studied in @xcite . in that paper manifolds covered by rational curves of degree
one with respect to an ample line bundle @xmath2 were considered , and it was proved that a covering family of such curves ( we will call them lines , by abuse ) of anticanonical degree greater than or equal to @xmath15 spans an extremal ray ( see also ( * ? ? ? * theorem 2.4 ) ) .
recently , in ( * ? ? ?
* theorem 7.3 ) , the extremality of a covering family @xmath11 of lines was proved under the weaker assumption that the anticanonical degree of such curves , denoted by abuse of notation by @xmath16 , is greater than or equal to @xmath17 .
the goal of the present paper is to prove the following let @xmath18 be a polarized manifold with a dominating family of rational curves @xmath11 such that @xmath19 . if @xmath20 , then @xmath21 $ ] spans an extremal ray of @xmath4 .
the main idea is , as in @xcite , to combine the ideas and tecniques of @xcite , especially taking into consideration a suitable adjoint divisor @xmath22 and studying its nefness , with those of @xcite , in particular regarding the existence of special curves in the indeterminacy locus of the rational fibration associated to @xmath11 .
let @xmath0 be a smooth projective variety defined over the field of complex numbers .
a _ contraction _
@xmath5 is a proper surjective map with connected fibers onto a normal variety @xmath6 .
+ if the canonical bundle @xmath23 is not nef , then the negative part of the cone @xmath4 of effective 1-cycles is locally polyhedral , by the cone theorem . by the contraction theorem , to every face in this part of the cone is associated a contraction . + unless otherwise stated , we will reserve the name _ extremal face _ for a face contained in @xmath24 , and we will call _ extremal contraction _ the contraction of such a face . +
an extremal contraction associated to an extremal face of dimension one , _
i.e. _ to an extremal ray , is called an _ elementary contraction _ ; an extremal ray @xmath25 is called _ numerically effective _ , and the associated contraction is said to be of _ fiber type _ , if @xmath26 ; otherwise the ray is called _ non nef _ and the contraction is _ birational_. + if the codimension of the exceptional locus of an elementary birational contraction is equal to one , the ray and the contraction are called _ divisorial _ , otherwise they are called _ small_. + a cartier divisor which is the pull - back of an ample divisor @xmath27 on @xmath6 is called a _ supporting divisor _ of the contraction @xmath8 . + if the anticanonical bundle of @xmath0 is ample , @xmath0 is called a fano manifold . for a fano manifold
, the _ index _ , denoted by @xmath28 , is defined as the largest natural number @xmath29 such that @xmath30 for some ( ample ) divisor @xmath2 on @xmath0 .
+ throughout the paper , unless otherwise stated , we will use the word _
curve _ to denote an irreducible curve .
[ rf ] a _ family of rational curves _ is an irreducible component @xmath31 ( see ( * ? ? ?
* definition 2.11 ) ) . given a rational curve
we will call a _ family of deformations _ of that curve any irreducible component of @xmath32 containing the point parametrizing that curve .
we will say that @xmath11 is _ unsplit _ if it is proper .
we define @xmath33 to be the set of points of @xmath0 through which there is a curve among those parametrized by @xmath11 ; we say that @xmath11 is a _ covering family _ if @xmath34 and that @xmath11 is a _ dominating family _ if @xmath35 .
+ we denote by @xmath36 the subscheme of @xmath11 parametrizing rational curves passing through @xmath37 and by @xmath38 the set of points of @xmath0 through which there is a curve among those parametrized by @xmath36 .
+ by abuse of notation , given a line bundle @xmath39 , we will denote by @xmath40 the intersection number @xmath41 , with @xmath42 any curve among those parametrized by @xmath11 .
( ( * ? ? ?
* iv.2.6))[iowifam ] let @xmath11 be an unsplit family of rational curves on @xmath0
. then * @xmath43 ; * every irreducible component of @xmath38 has dimension @xmath44 .
this last proposition , in case @xmath11 is the unsplit family of deformations of a rational curve of minimal anticanonical degree in an extremal face of @xmath4 , gives the _ fiber locus inequality _ : ( ( * ? ? ?
* theorem 0.4 ) , ( * ? ? ?
* theorem 1.1))[fiberlocus ] let @xmath8 be a fano
mori contraction of @xmath0 .
denote by @xmath45 the exceptional locus of @xmath8 and by @xmath46 an irreducible component of a non - trivial fiber of @xmath8 .
then @xmath47 where @xmath48 .
if @xmath8 is the contraction of an extremal ray @xmath25 , then @xmath49 is called the _ length of the ray_. [ cf ] we define a _ chow family of rational curves _ @xmath50 to be an irreducible component of @xmath51 parametrizing rational and connected 1-cycles .
+ we define @xmath52 to be the set of points of @xmath0 through which there is a cycle among those parametrized by @xmath50 ; notice that @xmath52 is a closed subset of @xmath0 ( ( * ? ? ?
* ii.2.3 ) ) .
we say that @xmath50 is a _ covering family _ if @xmath53 .
if @xmath11 is a family of rational curves , the closure of the image of @xmath11 in @xmath51 , denoted by @xmath12 , is called the _ chow family associated to _ @xmath11 . if @xmath11 is proper , _
i.e. _ if the family is unsplit , then @xmath11 corresponds to the normalization of the associated chow family @xmath12 .
let @xmath12 be the chow family associated to a family of rational curves @xmath11 .
we say that @xmath11 ( and also @xmath12 ) is _ quasi - unsplit _ if every component of any reducible cycle in @xmath12 is numerically proportional to @xmath11 .
let @xmath50 be a chow family of rational curves on @xmath0 and @xmath54 .
we define @xmath55 to be the set of points @xmath56 such that there exists a cycle @xmath57 among those parametrized by @xmath50 with @xmath58 and @xmath59 .
+ we define @xmath60 to be the set of points @xmath56 such that there exists a chain of cycles among those parametrized by @xmath50 connecting @xmath13 and @xmath6 .
notice that , a priori @xmath60 is a countable union of closed subsets of @xmath0 .
* notation * : if @xmath61 we will denote by @xmath62 the vector subspace generated by numerical classes of curves in @xmath63 ; we will denote by @xmath64 the subcone generated by numerical classes of curves in @xmath63 .
+ the notation @xmath65 will denote a linear subspace , while the notation @xmath66 will denote a subcone .
[ numeq ] ( ( * ? ? ?
* proposition iv.3.13.3 ) , ( * ? ? ?
* lemma 4.1 ) ) let @xmath61 be a closed subset and let @xmath50 be a chow family of rational curves . then every curve contained in @xmath67 is numerically equivalent to a linear combination with rational coefficients of a curve contained in @xmath63 and irreducible components of cycles among those parametrized by @xmath50 which intersect @xmath63 .
* proof of lemma 1.4.5 ) , ( * ? ? ?
* lemma 1 ) ) [ numequns ] let @xmath61 be a closed subset and let @xmath11 be a quasi - unsplit family of rational curves .
then every curve contained in @xmath68 is numerically equivalent to a linear combination with rational coefficients @xmath69 where @xmath70 is a curve in @xmath63 , @xmath42 is a curve among those parametrized by @xmath11 and @xmath71 .
* corollary 2.2 and remark 2.4))[locvf ] let @xmath72 be an extremal face of @xmath4 and denote by @xmath46 a fiber of the contraction associated to @xmath72 .
let @xmath11 be a quasi - unsplit family numerically independent from curves whose numerical class is in @xmath72
. then @xmath73 \rangle_c,\ ] ] _ i.e. _ the numerical class in @xmath0 of a curve in @xmath74 is in the subcone of @xmath4 generated by @xmath72 and @xmath21 $ ] .
[ efnef ] let @xmath75 be an effective divisor on @xmath0 and @xmath76 a nef divisor .
if @xmath77 is nef then @xmath78 is nef .
assume that @xmath79 is an effective curve on @xmath0 such that @xmath80 . by the nefness of @xmath76 we have @xmath81 , hence @xmath82 .
but @xmath78 is nef on @xmath75 , a contradiction .
let @xmath0 be a smooth complex projective variety and let @xmath50 be a covering chow family of rational curves .
the family @xmath50 defines a relation of rational connectedness with respect to @xmath50 , which we shall call _ rc@xmath83-relation _ for short , in the following way : @xmath13 and @xmath14 are in rc@xmath83-relation if there exists a chain of cycles among those parametrized by @xmath50 which joins @xmath13 and @xmath14 . to the rc@xmath83-relation
we can associate a fibration , at least on an open subset ( @xcite , ( * ? ? ? *
iv.4.16 ) ) ; we will call it
_ rc@xmath83-fibration_. in the notation of @xcite , by ( * ? ?
* theorem 5.9 ) there exists a closed irreducible subset of @xmath84 such that , denoting by @xmath85 its normalization and by @xmath86 the restriction of the universal family , we have a commutative diagram @xmath87^e \ar[d]_p & x \ar@{-->}[ld]^q \\ y & } \ ] ] where @xmath88 is the projection onto the first factor and @xmath89 is a birational morphism whose exceptional locus @xmath45 does not dominate @xmath85 .
moreover , a general fiber of @xmath90 is irreducible and is a rc@xmath83-equivalence class .
+ let @xmath91 be the image of @xmath45 in @xmath0 ; note that @xmath92 , as @xmath0 is smooth . if we consider a ( covering ) chow family @xmath12 , associated to a quasi - unsplit dominating family @xmath11 , then by ( * ? ? ?
* proposition 1 , ( ii ) ) @xmath91 is the union of all rc@xmath93-equivalence classes of dimension greater than @xmath94 .
moreover we have the following [ inb ] let @xmath11 be a quasi - unsplit dominating family of rational curves on a smooth complex projective variety @xmath0 .
let @xmath91 be the indeterminacy locus of the rc@xmath93-fibration @xmath95 & } { y}}$ ] , let @xmath75 be a very ample divisor on @xmath96 and let @xmath97
. then 1 .
@xmath98 ; 2 .
if @xmath99 is a curve not numerically proportional to @xmath21 $ ] , then @xmath100 ; 3 . if @xmath101 for every curve @xmath102 not numerically proportional to @xmath21 $ ] , then @xmath21 $ ] spans an extremal ray of @xmath4
see ( * ? ? ?
* proof of proposition 1 ) .
* proposition 3 ) .
[ incompreso ] let @xmath11 be a quasi - unsplit dominating family of rational curves on a smooth complex projective variety @xmath0 ; denote by @xmath91 the indeterminacy locus of the rc@xmath93-fibration and by @xmath103 the dimension of the general rc@xmath93-equivalence class .
+ if @xmath21 $ ] does not span an extremal ray of @xmath4 , then @xmath91 is not empty .
in particular there exist rc@xmath93-equivalence classes of dimension @xmath104 .
we now give a lower bound on the dimension of @xmath105 , depending on the position of the subvariety @xmath106 with respect to the indeterminacy locus of the rc@xmath93-fibration .
[ primaopoi ]
let @xmath11 be a quasi - unsplit dominating family of rational curves on a smooth complex projective variety @xmath0 ; denote by @xmath91 the indeterminacy locus of the rc@xmath93-fibration and by @xmath103 the dimension of the general rc@xmath93-equivalence class .
+ let @xmath107 be an irreducible subvariety such that @xmath21 \not \in { \operatorname{ne}\,(s , x)}$ ] .
then there exists an irreducible @xmath108 contained in @xmath105 such that 1 . if @xmath109 , then @xmath110 ; 2 . if @xmath111 , then @xmath112 .
moreover , @xmath108 is not rc@xmath93-connected .
we refer to diagram ( [ diagram ] ) . given any @xmath113 we will set @xmath114 .
let @xmath115 be an irreducible component of @xmath116 which dominates @xmath106 via @xmath89 .
+ by our assumptions on @xmath117 we have that @xmath118 meets any fiber of @xmath119 in points so , up to replace @xmath120 with @xmath121 , we can assume that @xmath118 is a section of @xmath119 .
+ let @xmath122 be an irreducible component of @xmath120 which contains @xmath118 .
we have @xmath123 moreover , notice that @xmath124 .
assume that @xmath125 .
then @xmath126 , hence the map @xmath127 is generically finite .
therefore , in view of ( [ zetaesse ] ) , @xmath128 ; moreover , since @xmath129 we have that @xmath130 is not rc@xmath93-connected .
assume now that @xmath111 .
assertion ( 2 ) will follow once we prove that the general fiber @xmath131 of @xmath132 has dimension strictly smaller than the general fiber of @xmath133 for at least one irreducible component @xmath134 of @xmath120 which dominates @xmath135 .
in fact , recalling also ( [ zetaesse ] ) , in this case we will have @xmath136 * claim .
* let @xmath131 be an irreducible component of a fiber of @xmath137 , let @xmath138 be any point and let @xmath139 be the intersection of the fiber of @xmath88 containing @xmath140 with @xmath118 ; then there exists an irreducible component @xmath46 of the fiber @xmath141 of @xmath133 containing @xmath142 such that @xmath143 .
to prove the claim , recall that , since @xmath144 , the image via @xmath89 of any curve in @xmath145 which is irreducible , being a section over @xmath146 must be a point , otherwise it would be a curve contained in @xmath147 , which is a contradiction , since curves in @xmath106 are numerically independent from @xmath21 $ ] .
+ therefore @xmath145 is contained in a fiber @xmath141 of @xmath133 .
to prove the claim we take as @xmath46 the irreducible component of @xmath141 containing @xmath145 .
let @xmath148 be the proper closed subset on which @xmath133 is not equidimensional and let @xmath149 be the proper closed subset of points in which the fiber of @xmath133 is not locally irreducible .
recalling that @xmath150 is a finite map we see that @xmath151 is a proper closed subset of @xmath135 .
+ let @xmath152 be a general point ; in particular there is only one irreducible component @xmath46 of the fiber @xmath141 of @xmath133 passing through @xmath153 and @xmath154 .
+ notice that @xmath155 , otherwise a one parameter family of fibers of @xmath88 meeting @xmath46 would have the same image in @xmath0 ( cf .
* end of proof of proposition 1 ) , where @xmath156 ) . +
this implies that , for an irreducible component @xmath157 of @xmath158 we have @xmath159 . taking as @xmath134 an irreducible component of @xmath120 containing @xmath160 we have that , for every point @xmath161 and any irreducible component @xmath131 of the fiber of @xmath132 passing through @xmath140 we have @xmath143 , hence @xmath162 ;
the same inequality then holds for the general fiber by semicontinuity of the local dimension .
+ noticing that @xmath106 is contained in @xmath163 the last assertion follows . both the bounds in lemma ( [ primaopoi ] ) are sharp .
an example for the second one is given by ( * ? ? ?
* example 2 ) : in that example @xmath164 ; taking as @xmath106 a fiber of the projection onto @xmath165 we have equality in @xmath166 .
in this section we consider the following situation , which will show up in the proof of theorem ( [ main ] ) : [ blow ] let @xmath18 be a polarized manifold with a dominating family of rational curves @xmath11 such that @xmath19
. denote by @xmath103 the dimension of the general rc@xmath93-equivalence class and assume that there exists an extremal face @xmath72 in @xmath4 whose associate contraction @xmath167 is a smooth blow - up along a disjoint union of subvarieties @xmath168 of dimension @xmath169 such that @xmath170 for every exceptional divisor @xmath171 and @xmath172 if @xmath173 is a line in a fiber of @xmath174 .
finally denote by @xmath175 a family of deformation of @xmath176 , with @xmath177 a general curve parametrized by @xmath11
. then 1 .
@xmath178 ; 2 .
there exists an ample line bundle @xmath179 on @xmath180 such that @xmath181 ; 3 . if @xmath182 is a curve parametrized by @xmath175 such that @xmath183 , then @xmath184 ; 4 .
@xmath185 ; 5 .
if @xmath186 $ ] spans an extremal ray of @xmath187 , then @xmath21 $ ] spans an extremal ray of @xmath4 .
it is enough to prove the statement in case @xmath188 , _
i.e. _ @xmath189 is the blow - up of @xmath180 along a smooth subvariety @xmath63 associated to the extremal ray @xmath72 .
in fact , if @xmath190 , the contraction of @xmath72 factors through elementary contractions , each one satisfying the assumptions in the statement .
denote by @xmath45 the exceptional locus of @xmath174 .
since @xmath191 the first assertion in the statement follows from the canonical bundle formula for blow - ups . moreover , the fact that @xmath191 also implies that any rc@xmath93-equivalence class meeting @xmath45 is actually contained in @xmath45 . therefore , if @xmath46 is a non - trivial fiber of @xmath174 , then @xmath192 . by lemma ( [ primaopoi ] )
@xmath193 hence @xmath194 and @xmath195 ; in particular , applying corollary ( [ locvf ] ) we get that @xmath196 , \sigma \rangle_c$ ] .
the line bundle @xmath197 is nef and it is trivial only on @xmath72 , since @xmath198 and @xmath199 .
then @xmath200 is nef by lemma ( [ efnef ] ) .
+ notice also that @xmath200 is trivial only on @xmath72 .
indeed , let @xmath79 be an effective curve on @xmath0 such that @xmath201 .
due to the ampleness of @xmath2 we have @xmath202 , hence @xmath203 .
this implies that @xmath204 \in \sigma$ ] .
therefore @xmath205 , with @xmath179 an ample line bundle on @xmath180 . by the projection formula @xmath181 ,
hence part ( 2 ) in the statement is proved .
now , let @xmath182 be a curve parametrized by @xmath175 meeting @xmath63 and assume by contradiction that @xmath182 is not contained in @xmath63 ; denote by @xmath206 its strict transform .
then @xmath207 which is a contradiction .
it follows that every curve parametrized by @xmath175 which meets @xmath63 is contained in it ; so we get part ( 3 ) in the statement
. as to part ( 4 ) , assume by contradiction that @xmath208 .
this implies that @xmath180 is rc@xmath209-connected , but this is impossible as , in view of part ( 3 ) , we can not join points of @xmath63 and points outside of @xmath63 with curves parametrized by @xmath175 . finally , to prove part ( 5 ) assume that @xmath186 $ ] spans an extremal ray of @xmath180 and let @xmath91 be the indeterminacy locus of the rc@xmath93-fibration .
we claim that @xmath210 .
assume by contradiction that this is not the case ; then @xmath45 meets ( and hence contains ) an rc@xmath93-equivalence class @xmath131 of dimension @xmath211 .
take a fiber @xmath46 of @xmath174 meeting @xmath131 . then @xmath212 . on the other hand , @xmath213 as @xmath21 \not \in \sigma$ ] .
so we get a contradiction .
let @xmath27 be a supporting divisor of the contraction associated to @xmath186 $ ] .
the pull - back @xmath214 defines a two - dimensional face @xmath215 of @xmath216 containing @xmath72 and @xmath21 $ ] .
let @xmath217 be as in lemma ( [ inb ] ) ; by the same lemma @xmath218 and @xmath219 .
+ assume that @xmath215 is not spanned by @xmath72 and @xmath21 $ ] ; in this case there exists a class @xmath220 belonging to @xmath215 such that @xmath221 and @xmath222 .
+ let @xmath223 be a sequence of effective one cycles such that the limit of @xmath224 $ ] is @xmath225 ; by continuity , for some @xmath226 we have @xmath227 and @xmath228 for @xmath229 , hence @xmath230 , and @xmath231 for @xmath229 , contradicting @xmath210 .
first of all we consider polarized manifolds @xmath18 with a quasi - unsplit dominating family of rational curves @xmath11 proving that if , for @xmath232 large enough , the adjoint divisor @xmath233 defines an extremal face containing @xmath21 $ ] , then @xmath21 $ ] spans an extremal ray of @xmath0 .
[ ivan ] let @xmath18 be a polarized manifold which admits a quasi - unsplit dominating family of rational curves @xmath11 ; denote by @xmath103 the dimension of a general rc@xmath93-equivalence class .
+ if , for some integer @xmath232 such that @xmath234 , the divisor @xmath22 is nef and it is trivial on @xmath21 $ ] , then @xmath21 $ ] spans an extremal ray of @xmath4 .
assume by contradiction that @xmath21 $ ] does not span an extremal ray in @xmath4 .
+ this implies that @xmath233 defines an extremal face @xmath72 of dimension at least two , containing @xmath21 $ ] . by (
* lemma 7.2 ) there exists an extremal ray @xmath235 whose exceptional locus is contained in the indeterminacy locus @xmath91 of the rc@xmath93-fibration . since @xmath236 ,
the length @xmath237 is greater than or equal to @xmath232 .
+ let @xmath46 be a non - trivial fiber of the contraction associated to @xmath238 ; since this contraction is small , being @xmath92 , then @xmath239 by proposition ( [ fiberlocus ] )
. + by part @xmath166 of lemma ( [ primaopoi ] ) , the dimension of @xmath240 is @xmath241 as the rc@xmath93-equivalence classes are either contained in @xmath91 or have empty intersection with it , @xmath242
. therefore we get @xmath243 which is a contradiction . as the last preparatory step , we consider the following special case .
[ premain ] let @xmath11 be a quasi - unsplit dominating family of rational curves on a smooth complex projective variety @xmath0 .
denote by @xmath103 the dimension of a general rc@xmath93-equivalence class .
assume that there exists an extremal ray @xmath238 , independent from @xmath21 $ ] , whose associated contraction has a fiber @xmath46 such that @xmath244
. then @xmath245 and @xmath246 , { \vartheta}\rangle_c$ ] .
in particular @xmath247 .
by part ( 1 ) of lemma ( [ primaopoi ] ) we have @xmath248 hence @xmath245 and @xmath249 ; so the assertion follows by corollary ( [ locvf ] ) .
[ main ] let @xmath18 be a polarized manifold with a dominating family of rational curves @xmath11 such that @xmath19 .
if @xmath20 , then @xmath21 $ ] spans an extremal ray of @xmath4 .
let @xmath91 be the indeterminacy locus of the rc@xmath93-fibration @xmath95 & } { y}}$ ] , let @xmath75 be a very ample divisor on @xmath96 and let @xmath97 .
denote by @xmath232 the anticanonical degree of @xmath11 and by @xmath103 the dimension of a general rc@xmath93-equivalence class .
notice that , since @xmath11 is a dominating family , we have @xmath250 .
+ by proposition ( [ iowifam ] ) @xmath251 ; since a general fiber of the rc@xmath93-fibration contains @xmath38 for every point @xmath13 in it , we have @xmath252 .
if @xmath22 is nef , then the assertion follows by proposition ( [ ivan ] ) ; therefore we can assume that @xmath233 is not nef .
+ let @xmath238 be an extremal ray such that @xmath253 and let @xmath254 be the associated contraction .
notice that @xmath238 has length @xmath255 , hence every non - trivial fiber of @xmath254 has dimension @xmath256 by proposition ( [ fiberlocus ] ) . on the other hand , by lemma ( [ premain ] ) we can confine to assume that all fibers of @xmath254 have dimension @xmath257 . + in particular
this implies that , denoted by @xmath258 a minimal degree curve whose numerical class belongs to @xmath238 , we have @xmath259 . indeed ,
if this were not the case , we would have @xmath260 , hence every non - trivial fiber of @xmath254 would have dimension @xmath261 , by proposition ( [ fiberlocus ] ) and the fact that @xmath262 .
if the picard number of @xmath0 is one the theorem is clearly true , so we can assume that @xmath263 .
now we split up the proof in two cases , according to the value of @xmath264 : first we consider the case @xmath247 and then the general one .
@xmath247 . the proof is based on different arguments , depending on the dimension of the fibers of the contraction associated to the extremal ray @xmath238 .
the contraction @xmath254 admits an @xmath265-dimensional fiber @xmath46 .
consider @xmath266 .
we have , by corollary ( [ locvf ] ) , that @xmath267,{\vartheta}\rangle_c$ ] and , by lemma ( [ primaopoi ] ) , that @xmath268 if @xmath269 , then the statement is proved .
so we can assume that an irreducible component @xmath270 of @xmath271 is a divisor and thus that @xmath272 .
notice that @xmath273 , otherwise we would have @xmath269 .
+ consider now the intersection number of @xmath274 with curves whose numerical class belongs to @xmath238 ; since @xmath247 and @xmath275 we can not have also @xmath276 .
let us show that we can not have @xmath277 , too . +
assume by contradiction that this is the case .
then @xmath278 , so @xmath254 is divisorial by proposition ( [ fiberlocus ] ) . by the same proposition ,
recalling that we are assuming that all the fibers of @xmath254 have dimension @xmath257 , every non - trivial fiber has dimension @xmath279 . +
then @xmath254 is the blow - up of a smooth variety @xmath180 along a smooth center @xmath63 by ( * ? ? ?
* theorem 4.1 ( iii ) ) .
the dimension of the center is @xmath280 we can thus apply part ( 4 ) of lemma ( [ blow ] ) and we get @xmath281 , reaching a contradiction
. therefore @xmath282 , hence @xmath283 is nef and thus , by lemma ( [ efnef ] ) , @xmath270 is nef . as @xmath273 and @xmath247 , @xmath270 is the supporting divisor of an elementary contraction of @xmath0 whose associated extremal ray is spanned by @xmath21 $ ] .
the contraction @xmath254 is equidimensional with @xmath232-dimensional fibers . by proposition ( [ fiberlocus ] ) ,
@xmath254 is of fiber type and @xmath284 .
hence , by ( * ? ? ?
* lemma 2.12 ) , @xmath0 is a projective bundle over a smooth variety @xmath85 , _
i.e. _ @xmath285 , where @xmath286 .
+ notice that @xmath85 has picard number one and is covered by rational curves the images of the curves parametrized by @xmath11
therefore @xmath85 is a fano manifold .
+ by the canonical bundle formula for projective bundles we have @xmath287 in particular , if @xmath42 is a curve among those parametrized by @xmath11 , by the projection formula we can compute @xmath288 it follows that @xmath289 and that @xmath290 is the ample generator of @xmath291 .
the ampleness of @xmath292 implies that @xmath293 ; therefore @xmath294 , hence the index @xmath295 of @xmath85 is greater than or equal to @xmath232 .
if @xmath296 , denoted by @xmath297 a rational curve of minimal degree in @xmath85 , then @xmath298 ; moreover , the splitting type of @xmath292 , which is ample and of rank @xmath279 , on rational curves of minimal degree is uniform of type @xmath299 .
+ we can thus apply ( * ? ? ?
* proposition 1.2 ) , so we obtain that @xmath300 .
it follows that the curves of @xmath11 are contained in the fibers of the first projection and that @xmath21 $ ] spans an extremal ray .
therefore we are left with @xmath301 . recalling that @xmath302 , by the kobayashi ochiai theorem ( @xcite )
we get that @xmath85 is a projective space or a hyperquadric .
assume by contradiction that @xmath21 $ ] does not span an extremal ray of @xmath0 .
+ by part ( 3 ) of lemma ( [ inb ] ) there exists a curve @xmath102 , whose numerical class is not proportional to @xmath21 $ ] , such that @xmath303 .
actually , since @xmath247 and @xmath304 , we have @xmath305 .
+ by part ( 2 ) of lemma ( [ primaopoi ] ) , there exists @xmath306 which is not rc@xmath93-connected such that @xmath307 .
+ by lemma ( [ numequns ] ) @xmath308 has non positive intersection number with every curve in @xmath309 and it is trivial only on curves which are numerically proportional to @xmath21 $ ] . + since @xmath310 , we have that @xmath254 does not contract curves in @xmath309 , hence @xmath311 and so @xmath312 . +
since @xmath309 is not rc@xmath93-connected , for every point @xmath313 of @xmath309 , the intersection @xmath314 of the rc@xmath93-equivalence class containing @xmath313 with @xmath309 has dimension @xmath315 .
in particular @xmath309 is the union of a one parameter family of rc@xmath93-connected subvarieties @xmath314 .
we claim that there exists a line @xmath297 in @xmath85 which is not contained in @xmath316 for any @xmath317 .
notice that , since @xmath254 does not contract curves in @xmath309 , through a general point @xmath14 in @xmath85 there is a finite number of such subvarieties .
+ if @xmath318 , a line joining @xmath14 with a point outside the union of these subvarieties has the required property .
+ assume now that @xmath319 ; for any @xmath320 the locus of the lines through @xmath14 is a quadric cone @xmath321 with vertex @xmath14 . therefore ,
if every line through @xmath14 is contained in @xmath316 for some @xmath317 , then @xmath321 is an irreducible component of @xmath316 ; since @xmath314 moves in a one - dimensional family , for the general point @xmath322 , the general line through @xmath14 has the required property .
the splitting type of @xmath292 on this line is one of the following : @xmath323 if @xmath324 and either @xmath325 or @xmath326 if @xmath318 .
recalling that @xmath250 we have that , among the summands of @xmath327 there is at least one @xmath328 .
+ consider @xmath329 ; its cone of curves is generated by the class of a line in a fiber of the projection onto @xmath297 and the class of a minimal section @xmath330 . by the discussion above we have that @xmath331 .
moreover , @xmath332 , hence @xmath333=[v]$ ] ; in particular @xmath308 is nef on @xmath329 .
+ consider an irreducible curve in @xmath334 ; by our choice of @xmath297 , this curve is not contained in a rc@xmath93-equivalence class contained in @xmath309 , so it is negative with respect to @xmath308 , a contradiction .
the case @xmath247 is thus completed .
@xmath335 .
notice that , in view of corollary ( [ incompreso ] ) , we can confine to assume that @xmath336 ; moreover , by part ( 3 ) of lemma ( [ inb ] ) , we can also assume the existence of a curve @xmath102 such that @xmath337 $ ] is not proportional to @xmath21 $ ] and @xmath303 .
we claim that @xmath338 is nef .
+ assume by contradiction that @xmath338 is not nef .
let @xmath25 be a ray such that @xmath339 , denote by @xmath340 a rational curve of minimal anticanonical degree in @xmath25 and by @xmath341 the contraction associated to @xmath25 .
+ notice that @xmath25 has length @xmath342 , hence every non - trivial fiber of @xmath341 has dimension @xmath343 by proposition ( [ fiberlocus ] ) .
+ on the other hand @xmath341 can not have fibers of dimension @xmath344 , otherwise , by lemma ( [ premain ] ) , we would have @xmath247
. therefore every non - trivial fiber of @xmath341 has dimension @xmath279 .
+ in view of proposition ( [ fiberlocus ] ) , we thus get that @xmath341 is of fiber type and that the length of @xmath25 is @xmath345 ; this last fact gives @xmath346 .
let us consider @xmath347 to be a minimal degree covering family of curves whose numerical class belongs to @xmath25 .
+ since @xmath91 is not empty , there are rc@xmath93-equivalence classes of dimension @xmath348 ; let @xmath131 be one of these classes .
notice that since @xmath341 is equidimensional with @xmath265-dimensional fibers , we have @xmath349 .
by part ( 1 ) of lemma ( [ primaopoi ] ) we have @xmath350 so by lemma ( [ numeq ] ) we deduce @xmath247 , a contradiction which proves the nefness of @xmath338 .
recall now that the extremal ray @xmath351 which we fixed at the beginning of the proof has length @xmath352 and is generated by a curve @xmath353 such that @xmath354 , therefore @xmath355 and @xmath338 is not ample .
+ let @xmath72 be the extremal face contracted by @xmath338 .
we now consider separately two cases , depending on the existence in @xmath72 of a fiber type extremal ray .
there exists a fiber type extremal ray @xmath356 in @xmath72 .
let @xmath357 be the contraction associated with @xmath356 and denote by @xmath358 a minimal degree covering family of curves whose numerical class belongs to @xmath356 .
+ by part ( 2 ) of lemma ( [ primaopoi ] ) , there exists an irreducible @xmath306 such that @xmath359 .
+ according to lemma ( [ numequns ] ) , every curve in @xmath309 can be written as @xmath360 + \beta [ v]$ ] with @xmath361 ; in particular , since @xmath219 by lemma ( [ inb ] ) , it follows that @xmath217 is not positive on any curve contained in @xmath309 . by the same lemma @xmath362 , hence @xmath363 \not \in { \operatorname{ne}\,(x_c , x)}$ ] . therefore part ( 1 ) of lemma ( [ primaopoi ] ) gives @xmath364 where @xmath365 is the dimension of the general rc@xmath366-equivalence class .
+ therefore , by applying twice lemma ( [ numequns ] ) , we get that the class of every curve in @xmath0 can be written as @xmath367 + \beta [ v ] ) + \mu[w_\varrho]\ ] ] with @xmath368 and @xmath369 + \beta [ v ] \in { \operatorname{ne}\,(x_c , x)}$ ] . +
this has some very important consequences : first of all , since we are assuming @xmath370 , this implies that @xmath371 ; in particular @xmath337 $ ] is not contained in the plane @xmath215 in @xmath372 spanned by @xmath363 $ ] and @xmath21 $ ] .
moreover the intersection of @xmath215 with @xmath4 is a face of @xmath4 .
+ we have to prove that @xmath373 , [ w_\varrho ] \rangle_c$ ] .
if this is not the case , then there exists a class @xmath374 such that @xmath375 \rangle_c$ ] and @xmath376 .
+ denote by @xmath377 a class , not proportional to @xmath21 $ ] , lying in the intersection of @xmath378 with the plane @xmath379 and by @xmath380 the plane spanned by @xmath363 $ ] and @xmath381 .
+ formula ( [ cone ] ) , traslated in geometric terms , says that @xmath382 is contained in the intersection of half - spaces determined by @xmath215 and by @xmath380 as in the figure below , which shows a cross - section of @xmath216 .
let @xmath223 be a sequence of effective one cycles such that the limit of @xmath224 $ ] is @xmath383 ; by continuity , for some @xmath226 we have @xmath228 for @xmath229 , hence @xmath230 for @xmath229 , and all the above arguments apply to @xmath384 , for @xmath229 .
in particular , defining @xmath385 and @xmath386 as above , we get that , for @xmath229 , @xmath382 is contained in the intersection of half - spaces determined by @xmath215 and by @xmath387 .
since @xmath388 as @xmath224 \rightarrow { \mathbb r}_+a$ ] and @xmath371 we get a contradiction .
every ray in @xmath72 is birational .
let @xmath389 be any ray in @xmath72 . by proposition ( [ fiberlocus ] ) ,
for every non - trivial fiber of its associated contraction @xmath390 we have @xmath391 . recalling that , by lemma ( [ premain ] )
, we can assume @xmath392 , we have @xmath393 .
this also implies that , if @xmath394 is a minimal degree curve whose numerical class is contained in @xmath389 we have @xmath395 .
+ by proposition ( [ fiberlocus ] ) , @xmath390 is a divisorial contraction , hence , by ( * ? ? ?
* theorem 4.1 ( iii ) ) , is the blow - up of a smooth variety along a smooth center @xmath63 of dimension @xmath396 .
let @xmath45 be the exceptional divisor of @xmath390 .
by part ( 2 ) of lemma ( [ primaopoi ] ) , there exists an irreducible @xmath306 with @xmath359 .
+ by lemma ( [ numequns ] ) @xmath308 has non positive intersection number with every curve in @xmath309 . + if @xmath397 , then there is a fiber @xmath46 of @xmath390 meeting @xmath309 .
counting dimensions , we find that @xmath398 , which is a contradiction as @xmath399 .
so @xmath400 , whence @xmath191 .
+ therefore @xmath45 contains rc@xmath93-equivalence classes and @xmath401 , since @xmath390 is finite - to - one on rc@xmath93-equivalence classes .
recalling that @xmath252 we derive @xmath402 .
assume that @xmath403 and let @xmath404 be the exceptional loci of two different extremal rays @xmath405 in @xmath72 ; since the fibers of the contractions @xmath406 and @xmath407 have dimension @xmath279 and @xmath408 we have that @xmath409 . +
therefore the contraction @xmath167 of the face @xmath72 verifies the assumptions of lemma ( [ blow ] ) , hence there exists an ample line bundle @xmath179 on @xmath180 and an unsplit dominating family @xmath175 on @xmath180 such that @xmath410 and @xmath411 .
+ denote by @xmath412 the dimension of the general rc@xmath209-equivalence class .
since a general fiber of the rc@xmath209-fibration contains @xmath413 , we have @xmath414 .
+ consider the adjoint divisor @xmath415 ; if it is nef , or an extremal ray @xmath416 such that @xmath417 has a fiber of dimension @xmath418 , then @xmath186 $ ] spans an extremal ray by proposition ( [ ivan ] ) or by lemma ( [ premain ] ) , so @xmath21 $ ] spans an extremal ray by lemma ( [ blow ] ) .
let us show that the remaining case does not happen .
assume that there is an extremal ray @xmath416 such that @xmath417 and every fiber of the associated contraction has dimension @xmath257 .
in particular we have @xmath419 , otherwise we would have @xmath420 , hence every non - trivial fiber of the associated contraction would have dimension @xmath261 by proposition ( [ fiberlocus ] ) .
moreover , we have @xmath421 , since @xmath422 .
+ on the other hand , recalling that @xmath423 and that @xmath424 , we have @xmath425 so , by the projection formula , @xmath426 is ample on @xmath180 , a contradiction .
let @xmath18 be a polarized manifold of dimension at most five , with a dominating family of rational curves @xmath11 such that @xmath19 .
then @xmath21 $ ] spans an extremal ray of @xmath4 .
in the paper @xcite , an application of the results about extremality of families of lines was a relative version of a theorem proved in @xcite , which was the first step towards a conjecture of mukai for fano manifolds .
+ this conjecture states that , for a fano manifold @xmath0 , denoted by @xmath264 its picard number and by @xmath28 its index , we have @xmath427 more precisely , in ( * ? ? ?
* theorem b ) it was proved that , if @xmath428 , then @xmath429 unless @xmath430 , while in ( * ? ? ?
* theorem 3.1.1 ) it was proved that a fiber type contraction @xmath431 supported by @xmath432 with @xmath433 is elementary , unless @xmath430 . in the last few years some progress has been made towards mukai conjecture ; in particular it was recently proved in ( * ? ? ?
* theorem 3 ) that it holds for a fano manifold with ( pseudo)index greater than or equal to @xmath434 . + it is therefore natural to ask if the corresponding relative statement is true , namely , given a fiber type contraction @xmath431 , corresponding to an extremal face @xmath72 , supported by @xmath432 with @xmath435 is it possible to find a bound on the dimension of @xmath72 ?
the answer to this question is negative , as we will show with an example in which @xmath436 ; it follows that ( * ? ? ?
* theorem 3.1.1 ) can not be improved .
let @xmath6 be a smooth variety of dimension @xmath437 , denote by @xmath85 the product @xmath438 and by @xmath439 the projections onto the factors .
let @xmath440 be points of @xmath6 and denote by @xmath441 the fibers of @xmath442 over @xmath443 .
+ let @xmath444 be the blow - up of @xmath85 along the union of @xmath441 . the canonical bundle of @xmath0 is @xmath445 denoting by @xmath446 and by @xmath447 , we can rewrite formula @xmath448 as @xmath449 it is easy to check that @xmath450 is @xmath451-ample .
let @xmath452 be an ample line bundle such that @xmath453 is ample ; then @xmath454 is an ample line bundle on @xmath0 ; moreover @xmath455 for a line @xmath297 in the strict transform of a fiber @xmath46 of @xmath442 not contained in the center of @xmath174 .
+ the contraction @xmath456 is supported by @xmath457 and contracts a face of dimension @xmath458 .
the difference between the relative and the absolute case is given by the existence of minimal horizontal dominating families of rational curves for proper morphisms defined on a open subset of a fano manifold ( for the definition and the references see ( * ? ? ?
* remark 6.4 ) ) .
such families do not exist in general in the relative case .
we learned of the results about extremality of families of lines in @xcite from an interesting series of lectures given by paltin ionescu .
we thank the referee for many useful suggestions and remarks , which helped to fix some issues in the proofs .
mauro c. beltrametti , andrew j. sommese , and jarosaw a. winiewski .
results on varieties with many lines and their applications to adjunction theory . in
_ complex algebraic varieties ( bayreuth , 1990 ) _ , volume 1507 of _ lecture notes in math .
_ , pages 1638 .
springer , berlin , 1992 .
takao fujita . on polarized manifolds
whose adjoint bundles are not semipositive . in _ algebraic geometry , sendai , 1985 _ , volume 10 of _ adv .
pure math .
_ , pages 167178 .
north - holland , amsterdam , 1987 . | let @xmath0 be a smooth complex projective variety and let @xmath1 be an ample line bundle .
assume that @xmath0 is covered by rational curves with degree one with respect to @xmath2 and with anticanonical degree greater than or equal to @xmath3 .
we prove that there is a covering family of such curves whose numerical class spans an extremal ray in the cone of curves @xmath4 .
[ section ] [ theorem]lemma [ theorem]proposition [ theorem]corollary [ theorem]example [ subsection ] [ teo]lemma [ teo]proposition [ teo]corollary [ theorem]definition [ theorem ] [ theorem]remark [ teo]definition [ teo]example [ teo]remark |
soft hadronic scattering , characterized by long distance phenomena , is one of the great challenges in high - energy physics .
the difficulty arises from the fact that perturbative qcd can not be applied to these processes and , presently , we do not know even how to calculate elastic hadron - hadron scattering amplitudes from a pure non - perturbative qcd formalism .
however , progresses have been achieved through the approach introduced by landshoff and nachtmann @xcite , developed by nachtmann @xcite and connected with the stochastic vacuum model ( svm ) @xcite . in that formalism
the low frequencies contributions in the functional integral of qcd are described in terms of a stochastic process , by means of a cluster expansion .
the most general form of the lowest cluster is the gauge invariant two - point field strength correlator @xcite @xmath1(1-{\kappa})d_{1 } ( { z^2/a}^{2})\ } , \end{aligned}\ ] ] where @xmath2 is the two - point distance , @xmath3 is a characteristic correlation length , @xmath4 a constant , @xmath5 the gluon condensate and @xmath6 the number of colours ( @xmath7 ) . the two scalar functions @xmath8 and @xmath9 describe the correlations and they play a central role in the application of the svm to high energy scattering @xcite .
once one has information about @xmath8 and @xmath9 , the svm leads to the determination of the elementary quark - quark scattering amplitude , which constitutes important input for models aimed to construct hadronic amplitudes .
numerical determinations of the above correlation functions , in limited interval of physical distances , exist from lattice qcd in both quenched approximation ( absence of fermions ) @xcite and full qcd ( dynamical fermions included ) @xcite . in a previous paper
, we determined the elementary amplitudes from lattice qcd in the quenched approximation @xcite , using as framework the svm . in this communication , we apply the same procedure , now taking into account the lattice results in full qcd .
our main goal is to investigate the differences in the elementary amplitudes associated with quenched theory and full qcd and also the effect of different bare quark masses .
in addition , we obtain analytical parametrizations for the amplitudes that are suitable for phenomenological uses , and discuss in some detail the region of validity of all the results .
the manuscript is organized as follows . in sec .
ii we recall the main formulas related with the elementary amplitudes in the context of the svm and in sec .
iii we review the parametrizations for the correlators from numerical simulations on the lattice . in sec .
iv we present the results for the elementary amplitudes from full qcd with different quark masses and discuss the similarities and differences with our previous result in the quenched approximation . the conclusions and some final remarks are the contents of sec . v.
in this section we briefly review the main steps of the calculation scheme that allows the determination of the elementary amplitudes through the svm @xcite .
we refer the reader to @xcite for more details concerning specific calculation .
the elementary amplitude @xmath10 in the momentum transfer space may be expressed in terms of the elementary profile @xmath11 in the impact parameter space , through the symmetrical two - dimensional fourier transform @xmath12 where @xmath13 is the momentum transfer , @xmath14 the impact parameter and @xmath15 is a bessel function . in the nachtmann approach @xcite , the study of the elementary scattering
is based on the amplitude of quarks moving on lightlike paths in an external field , picking up an eikonal phase in traveling through the nonperturbative qcd vacuum . in order to have gauge invariant dirac s wave function solutions a wilson loop
is proposed to represent each quark . in this context
the no - colour exchange parton - parton ( loop - loop ) amplitude can be written as @xcite @xmath16tr[{\cal{p}}e^{-ig{\int}_{loop2}d { \sigma}_{\rho\sigma}f_{\rho \sigma}(y;w)}-1]{\rangle } , \nonumber\end{aligned}\ ] ] where @xmath17 means the functional integration over the gluon fields ( the integrations are over the respective loop areas ) , and @xmath18 is a common reference point from which the integrations are performed .
this expression is simplified in the krmer and dosch description by taking the wilson loops on the light - cone . in the svm
the _ leading order _ contribution to the amplitude is given by @xcite @xmath19 where @xmath20 is a constant depending on normalizations and @xmath21 here @xmath22 is the minkowski version of the gluon correlator .
after a two - dimensional integration , @xmath23 may be expressed in terms of the correlation functions in ( 1 ) by @xcite @xmath24 where @xmath25(b'),\end{aligned}\ ] ] @xmath26(b ) , \label{epsilonii}\end{aligned}\ ] ] for @xmath27 or @xmath28 , @xmath29 $ ] , where @xmath30 denotes a n - dimensional fourier transform . with the above formalism , once one has inputs for the correlation functions @xmath31 and @xmath32 , the elementary amplitude in the momentum transfer space , eq .
( 2 ) , may , in principle , be evaluated through eqs .
( 3 - 6 ) .
it is important to stress that , as constructed , this approach is intended for the high energy limit and small momentum transfer region , namely , @xmath33 , where @xmath34 is the c.m .
energy and @xmath35 gev@xmath0 .
the determination of the correlation functions through numerical simulation on a lattice is made by means of the cooling technique , a procedure that removes the effects of short - range fluctuations on large distance correlators .
the numerical results with the associated error are usually parametrized with the functions @xcite @xmath36 @xmath37 where @xmath38 in the non - perturbative exponential terms is the correlation length of the gluon field strengths .
discussions on these choices , including the perturbative - like divergence at short distances , may be found in refs . @xcite and @xcite .
these correlation functions were first determined in the quenched su(3 ) theory and in the interval of physical distances ( euclidean space ) between 0.1 and 1 fm @xcite .
after that , the effects of dynamical fermions have also been included ( full qcd ) , for bare quark masses @xmath39 and @xmath40 ( @xmath3 is the lattice spacing ) and physical distances between 0.3 and 0.9 fm @xcite .
the above parametrizations are the same in all these cases and the numerical values of the parameters , obtained from references @xcite and @xcite , are displayed in table 1 . [ cols="^,^,^,^ " , ] and quenched approximation , in the region of large ( a ) and small ( b ) momentum transfer.,title="fig:",width=321,height=226 ] and quenched approximation , in the region of large ( a ) and small ( b ) momentum transfer.,title="fig:",width=321,height=226 ] .,title="fig:",width=321,height=226 ] .,title="fig:",width=321,height=226 ] an immediate conclusion from these results is the presence of an additional component in the case of full qcd .
this is a central point that we shall discuss in certain detail in what follows .
let us start with the components with the highest slopes , which appear in both cases in the small momentum transfer region , namely , below @xmath41 gev@xmath0 ( figs . 3 and 4 ) .
as mentioned before , the real " lattice results correspond to sets of discrete theoretical points with errors , in a finite interval of physical distances , roughly 0.1 - 1.0 fm .
therefore , the parametrizations ( 7 ) and ( 8) extrapolate this interval down and above .
the highest physical distance reached in the simulations was 0.85 fm , which corresponds to @xmath42 gev@xmath0 .
therefore , we conclude that the components with the highest slopes in figs 3 and 4 ( @xmath43 gev@xmath0 ) are associated with the extrapolations above the physical region with real " lattice results . in the same manner ,
the components with the smallest slopes are connected with extrapolations down the real " lattice points in physical distances and they are the responsible for the amplitudes in high momentum region ( @xmath44 gev@xmath0 in figs . 3 and 4 ) ; therefore , they are outsite the region of validity of the svm , namely , @xmath35 gev@xmath0 .
we conclude that in this context only the intermediate components , predominant in the interval , let us say , @xmath45 gev@xmath0 , can have physical meaning in the sense of being in agreement with the svm and the real " lattice results in both full qcd and quenched approximation . from figs . 3 and 4 this secure " region
is characterized by only one component in the case of the quenched approximation and two components in full qcd .
the transition between these two components occurs at @xmath46 gev@xmath0 ( fig .
4 - left ) , a limit region for which the svm is intended for .
in this work we have obtained analytical parametrizations for the quark - quark scattering amplitudes in a nonperturbative qcd framework ( svm ) and using as inputs the correlation functions , determined from numerical simulation on a lattice , in both quenched approximation and full qcd .
the formalism is intended for small momentum transfer ( @xmath35 gev@xmath0 ) , asymptotic energies @xmath33 and physical distances between 0.1 and @xmath47 0.9 fm . as discussed in sec .
iv , these conditions put some restrictions in the physical interpretations and , therefore , in the practical phenomenological uses of these amplitudes .
however , even under the above strictly conditions we can surely extract some novel results : ( 1 ) the amplitudes decrease smoothly as the momentum transfer increases and they do not present zeros ; ( 2 ) the decreasing is faster when going from quenched approximation to full qcd ( with decreasing quark masses ) , and this effect is associated with the increase of the correlation lengths ( @xmath38 and @xmath48 ) ; ( 3 ) the dynamical fermions generate two contributions in the region of small momentum transfer , which are of the same order at @xmath49 1 gev@xmath0 ( only one contribution is present in the case of quenched approximation ) .
we understand that result ( 3 ) may suggest some kind of change in the dynamics at the elementary level , near @xmath49 1 gev@xmath0 and at asymptotic energies .
if that is true , some signal could be expected at the hadronic level .
one possibility is that this effect can be associated with the position of the dip ( or beginning of the shoulder " ) in the hadronic ( elastic ) differential cross section data .
the asymptotic condition embodied in our result indicates that @xmath49 1 gev@xmath0 seems in agreement with limit of the shrinkage of the diffraction peak , empirically verified when the energy increases in the region 23 gev @xmath50 1.8 tev .
at last it should be noted that if there is no new effect above the physical distances presently investigated on the lattice ( @xmath51 0.9 fm ) , the extrapolations can be considered as a good representation of the lattice results . in this case
our analytical parametrizations may be usefull inputs for phenomenological uses in the region of small momentum transfer and asymptotic energies .
we are thankful to professor a. di giacomo for valuable discussions and suggestions and to dr .
j. montanha neto for reading the manuscript .
this work was supported by fapesp ( contract n. 01/08376 - 2 and 00/04422 - 7 ) . | making use of the stochastic vacuum model and the gluon gauge - invariant two - point correlation function , determined by numerical simulation on the lattice in both quenched approximation and full qcd , we calculate the elementary ( quark - quark ) scattering amplitudes in the momentum transfer space and at asymptotic energies . our main conclusions are the following : ( 1 ) the amplitudes decrease monotonically as the momentum transfer increases ; ( 2 ) the decreasing is faster when going from quenched approximation to full qcd ( with decreasing quark masses ) and this effect is associated with the increase of the correlation lengths ; ( 3 ) dynamical fermions generate two components in the amplitude at small momentum transfer and the transition between them occurs at momentum transfer near 1 gev@xmath0 .
we also obtain analytical parametrizations for the elementary amplitudes , that are suitable for phenomenological uses , and discuss the effects of extrapolations from the physical regions investigated in the lattice .
_ to be published - physics letters b _ |
a reduced model for slow variables of multiscale dynamics is a lower - dimensional dynamical system , which `` resolves '' ( that is , qualitatively approximates in some appropriate sense ) major large scale slow variables of the underlying higher - dimensional multiscale dynamics while at the same time being relatively simple and computationally inexpensive to work with .
this is important in real - world applications of contemporary science , such as geophysical science and climate change prediction @xcite , where the actual underlying physical process is impossible to model directly , and its reduced approximation has to be designed for such a purpose .
reduced dynamics were used to model global circulation patterns @xcite , and large - scale features of tropical convection @xcite . typically , reduced models of multiscale dynamics consist of simplified lower - dimensional dynamics of the original multiscale dynamics for the resolved variables , with additional terms and parameters which serve as replacements to the missing coupling terms with the unresolved variables of the underlying physical process . these extra parameters in the reduced model are usually computed to match a particular dynamical regime of the underlying multiscale dynamics @xcite .
in particular , if the underlying multiscale process changes its dynamical regime ( for example , in response to changes in its own forcing parameters ) , then the parameters of the corresponding reduced model have to be appropriately readjusted to match its dynamical regime to the new regime of the multiscale dynamics . in some real - world applications , such as the climate change prediction ,
the reduced models of complex multiscale climate dynamics are used to predict the response of the actual multiscale climate to changes in various global atmospheric and oceanic parameters . however , while a reduced model may be manually adjusted to match a particular dynamical regime of a multiscale process , it is unclear whether it should respond to identical external perturbations _ a priori _ in the same way as the multiscale process , without any extra readjustments . how do reduced models of multiscale dynamics , adjusted to a particular dynamical regime , respond to external perturbations which force them out of this regime ?
is their response similar to the response of the underlying multiscale dynamics to the same external perturbation ?
it is quite clear that the reduced dynamics evolve on a set with lower dimension than that of the full multiscale dynamics .
how do the properties of this limiting set respond to changing external forcing parameter , in comparison to the full multiscale attractor ? here
we develop a set of criteria for similarity of the response to small external perturbations between slow variables of multiscale dynamics and those of a reduced model for slow variables only , determined through statistical properties of the unperturbed dynamics .
we also carry out a computational study of the difference in responses of the full multiscale and deterministic reduced dynamics of the linearly coupled rescaled lorenz 96 model from @xcite to identical external perturbations .
we compare and contrast both the actual ( `` ideal '' ) responses of the multiscale and reduced models directly to finitely small perturbations of external forcing , and the linear response predictions of the reduced models via the fluctuation - dissipation theorem @xcite .
two different types of forcing perturbations are used : the time - independent heaviside forcing , and the simple time - dependent ramp forcing .
the manuscript is structured as follows . in section [ sec : averaged ] we formulate the standard averaging formalism to obtain the averaged slow dynamics from a general two - scale dynamical system .
section [ sec : response ] describes statistically tractable criteria to ensure similarity of responses between a two - scale system and its averaged slow dynamics . in section [ sec : implement ] we describe the first - order reduced model approximation to a two - scale dynamics with linear coupling between the slow and fast variables , previously developed in @xcite . in section [ sec : lorenz96 ] we introduce the two - scale lorenz 96 toy model which will be our testbed for this method . in section [ sec : results ]
we present comparisons of the large features of the multiscale and reduced systems , including statistical comparisons as well as the ability of the reduced model to capture perturbation response of the multiscale system .
section [ sec : summary ] summarizes the results and suggests future work .
a general two - scale dynamical system with slow variables @xmath0 and fast variables @xmath1 is usually represented as @xmath2 where , @xmath3 are the slow variables of the system , @xmath4 are the fast variables , and @xmath5 and @xmath6 are nonlinear differentiable functions .
the integer parameters @xmath7 are the dimensions of the slow and fast variable subspaces , respectively .
usually , a time - scale separation parameter is used to denote the difference in time scales between the slow and fast variables , however , here we omit it , as the framework for reduced models from @xcite , which we use here , does not require such a parameter to be explicitly present . under the assumption of `` infinitely fast '' @xmath1-variables , one can use the averaging formalism @xcite to write the averaged system for slow variables alone : @xmath8 where @xmath9 is the invariant distribution measure of the fast limiting system @xmath10 with @xmath0 above in being a constant parameter .
we express the slow solutions of the two - scale system in and the averaged system in in terms of differentiable flows : @xmath11 @xmath12 it can be shown ( see @xcite and references therein ) that if the time scale separation between @xmath0 and @xmath1 is large enough , then , for the identical initial conditions @xmath13 and generic choice of @xmath14 , the solution @xmath15 of the averaged system in remains near the solution @xmath16 of the original two - scale system in for finitely long time .
let @xmath17 and @xmath18 denote the invariant distribution measures for the two - scale system in and the averaged system in , respectively . also , let @xmath19 be a differentiable test function .
then , the statistically average values of @xmath20 for both two - scale and averaged systems are given via [ eq : h ] @xmath21 @xmath22 now , consider the two - scale system in , and the averaged system in , perturbed at the slow variables by a small time - dependent forcing @xmath23 : [ perturbed - eqn ] @xmath24 @xmath25 then , for small enough @xmath23 , the average responses @xmath26 and @xmath27 for the two - scale system in and the averaged system in , respectively , can be approximated by the following linear response relations : @xmath28 @xmath29 for details , see @xcite . above , it is clear that any differences between @xmath26 and @xmath30 are due to differences between @xmath31 and @xmath32 , since @xmath33 is identical in both cases .
the differences between @xmath31 and @xmath32 are , in turn , caused by the differences between the flows @xmath34 and @xmath35 , and the differences between the invariant distribution measures @xmath17 and @xmath18 , which are difficult to quantify in practice . in what follows
we express the differences between @xmath31 and @xmath32 via statistically tractable quantities .
first , we assume that the invariant measures @xmath17 and @xmath18 are absolutely continuous with respect to the lebesgue measure , with distribution densities @xmath36 and @xmath37 , respectively : @xmath38 @xmath39 while it is known that purely deterministic processes may not have lebesgue - continuity of their invariant measures @xcite , however , even small amounts of random noise , which is always present in real - world complex geophysical dynamics , usually ensure the existence of the distribution density .
integration by parts yields [ eq : r ] @xmath40 @xmath41 at this point , let us express @xmath36 as the product of its marginal distribution @xmath42 , defined as @xmath43 and conditional distribution @xmath44 , given by @xmath45 it is easy to check that the conditional distribution @xmath44 satisfies the identity @xmath46 now , the formula for the linear response operator @xmath31 above can be written as @xmath47 we now denote @xmath48 where @xmath49 is small compared to either @xmath50 or @xmath51 for relevant values of @xmath52 , @xmath0 and @xmath1 .
then , for the second integral in the right - hand side of we write @xmath53 where the first integral in the right - hand side is zero due to the condition in . neglecting the @xmath54 terms in , we write @xmath55 @xmath41 at this point
, we express @xmath42 and @xmath37 as exponentials @xmath56 where @xmath57 and @xmath58 are smooth functions , growing to infinity as @xmath0 becomes infinite .
the latter yields @xmath59 @xmath60 replacing invariant measure averages with long - term time averages yields the following time correlation functions : [ eq : r2 ] @xmath61 @xmath62 taking into account the arbitrariness of @xmath20 , we conclude that , in order for @xmath32 to approximate @xmath31 despite the fact that , for long times @xmath63 , @xmath64 diverges from @xmath65 even for identical initial conditions , we generally need three conditions to be approximately satisfied : 1 . for identical initial conditions
, @xmath15 should approximate @xmath16 ( that is , @xmath49 in should indeed be small ) on the finite time scale of decay of the correlation functions in ; 2 .
@xmath58 should approximate @xmath57 , which means that the invariant distribution @xmath37 of the averaged system in should be similar to the @xmath0-marginal @xmath42 of the invariant distribution of the two - scale dynamical system in ; 3 . the time autocorrelation functions of the averaged system in should be similar to the time autocorrelation functions of the slow variables of the two - scale system in . as a side note , observe that the nature of dependence of the conditional distribution @xmath44 on @xmath0 does not play any role in the criteria for the similarity of responses .
in particular , the exact factorization of @xmath36 into its @xmath0- and @xmath1-marginals ( which means that @xmath44 is independent of @xmath0 ) is not required , unlike what was suggested in @xcite for the gaussian invariant states .
as formulated above in sections [ sec : averaged ] and [ sec : response ] , the criteria of the response similarity are applicable for a broad range of dynamical systems with general forms of coupling and their averaged slow dynamics . however , the practical computation of the reduced model approximation to averaged slow dynamics depends on the form of coupling in the two - scale system @xcite . in this work , we consider the linear coupling between the slow and fast variables in the two - scale system .
the linear coupling is the most basic form of coupling in physical processes , however , because of that it is also probably the most common form of coupling . for the linear coupling
, the reduced model is constructed according to the method developed previously in @xcite , which we briefly sketch below .
we consider the special setting of with linear coupling between @xmath0 and @xmath1 : @xmath66 where @xmath67 and @xmath68 are nonlinear differentiable functions , and @xmath69 and @xmath70 are constant matrices of appropriate sizes . the corresponding averaged dynamics for slow variables from simplifies to @xmath71 where @xmath72 is the statistical mean state of the fast limiting system @xmath73 with @xmath0 treated as constant parameter . in general , the exact dependence of @xmath72 on @xmath0 is unknown , except for a few special cases like the ornstein - uhlenbeck process @xcite . here , like in @xcite , we approximate @xmath72 via the linear expansion @xmath74 where @xmath75 is the statistical average state of the full multiscale system in , and @xmath76 . the constant matrix @xmath77 is computed as the time integral of the correlation function @xmath78 where @xmath79 is the solution of for @xmath80 .
the above formula constitutes the quasi - gaussian approximation to the linear response of @xmath81 to small constant forcing perturbations in , and is a good approximation when the dynamics in are strongly chaotic and rapidly mixing @xcite . with
, the reduced system in becomes the explicitly defined deterministic reduced model for slow variables alone : @xmath82 where @xmath83 . in what follows , the `` zero - order '' model refers to with the last term set to zero ( such that the coupling is parameterized only by the constant term @xmath84 ) . for details , see @xcite and references therein .
in the current work , we test the response of the reduced model for slow variables on the rescaled lorenz 96 system with linear coupling @xcite , which is obtained from the original two - scale lorenz 96 system @xcite by appropriately rescaling dynamical variables to approximately set their mean states and variances to zero and one , respectively .
below we present a brief exposition of how the rescaled lorenz 96 model is derived . the original two - scale lorenz 96 system @xcite is given by @xmath85 , \end{aligned}\right.\ ] ] where @xmath86 and periodic boundary conditions @xmath87 , @xmath88 and @xmath89 here @xmath90 and @xmath91 are constant forcing terms , @xmath92 and @xmath93 constant coupling parameters , and @xmath94 is the time scale separation parameter . throughout this paper
we will consider systems with twenty slow variables @xmath95 and eighty fast variables @xmath96 . in lorenz s original formulation
@xcite studying predictability in atmospheric - type systems , he begins with the uncoupled system @xmath97 with periodic boundary conditions .
this system has generic features of geophysical flows , namely a nonlinear advection - like term , linearly unstable waves , damping , forcing , mixing , and chaos @xcite .
the simple formulation , with invariance under index translation and a uniform forcing term @xmath98 , allows for straightforward analysis - in particular the long - time statistics of each variable should be identical and will only depend on @xmath98 .
additionally , the chaos and mixing of the system are simply regulated by the forcing , with decaying solutions for @xmath98 near zero , periodic solutions for @xmath98 slightly larger , weakly chaotic quasi - periodic solutions around @xmath99 , and chaotic and strongly mixing systems around @xmath100 and higher .
lorenz s two - time coupled system was introduced to study predictability and lyapunov exponents of systems with subgrid phenomena on faster timescales , and one of the authors of the current work has recent results showing that coupling two chaotic systems can suppress chaos in the slower system @xcite . to simplify the analysis of coupling trends for the two - time system
, we will scale out the dependence of the mean state and mean energy on the forcing term @xmath98 .
due to the translational invariance , the long - term mean @xmath101 and standard deviation @xmath102 for the uncoupled system are the same for all @xmath103 .
so we rescale @xmath0 and @xmath52 as @xmath104 where the new variables @xmath105 have zero mean and unit standard deviation , while their time autocorrelation functions have normalized scaling across different dynamical regimes ( that is , different forcings @xmath98 ) for short correlation times .
this rescaling was previously used in @xcite . in the rescaled variables ,
the uncoupled lorenz model becomes @xmath106 where @xmath101 and @xmath102 are functions of @xmath98 .
we similarly rescale the coupled two - scale lorenz 96 model : @xmath107 where @xmath108 and @xmath109 are the long term means and standard deviations of the uncoupled systems with @xmath90 or @xmath91 as constant forcing , respectively .
it is this rescaled coupled lorenz 96 system that we focus on for the closure approximation . before any numerical tests
, one can already anticipate that the zero - order reduced system will be inadequate for this model even with such simple coupling .
once the reference state @xmath75 is determined and @xmath110 computed , the zero - order reduced system is given , according to , by @xmath111 this is equivalent to perturbing @xmath90 by @xmath112 , which we expect to be small since @xmath105 and @xmath113 have zero mean in the uncoupled setting . in particular , we expect this perturbation to have only a small effect on the dynamics .
however , in the multiscale dynamics it has been shown that a chaotic regime in the fast system can suppress chaos when coupled to the slow system @xcite , and this phenomenon is completely lost in the zero - order model .
here we compare numerical results of the rescaled two - scale lorenz 96 system with its corresponding reduced system .
in particular we look at the ability of the reduced system captures some statistical quantities and how well it captures mean response to perturbations in the slow variables . in all parameter regimes
considered , we have a slow system consisting of twenty variables @xmath95 coupled with a fast system of eighty variables @xmath114 .
we use a fourth order runge - kutta method with timestep @xmath115 in the multiscale system and @xmath116 in the reduced system . to compute the mean response , an ensemble of @xmath117 points
is sampled from a single trajectory which has been allowed to settle onto the attractor .
using the translational symmetry of the lorenz 96 system , we rotate the indices to generate an ensemble twenty times larger . on a modern laptop , the initial calculation to generate the reduced system for the lorenz 96 system takes only a few minutes ; once computed , numerical simulation of the reduced system is faster than the multiscale system by a factor on the order of @xmath118 .
computing the mean response for a single forcing for 5 time units with a sufficiently large ensemble size ( @xmath119 trajectories ) takes over an hour in the multiscale system with @xmath120 but less than three minutes for the corresponding reduced system . in section [ sec : response ] we outlined the main requirements for correctly capturing the response of the two - scale system by its reduced model .
those were the approximation of joint distribution density functions ( ddf ) for slow variables , and the time autocorrelation functions of the time series .
it is , of course , not computationally feasible to directly compare the 20-dimensional ddfs and time autocorrelations for all possible test functions .
however , it is possible to compare the one - dimensional marginal ddfs and simple time autocorrelations for individual slow variables , to have a rough estimate on how the statistical properties of the multiscale dynamics are reproduced by the reduced model . in figure [ ddfs ]
we compare the distribution density functions and autocorrelation functions of the slow variables .
the ddfs are computed using bin - counting , and the autocorrelation function @xmath121 , averaged over @xmath52 , is normalized by the variance @xmath122 .
results from three parameter regimes are presented , and in all three regimes the fast system is chaotic and weakly mixing @xmath123 and the coupling strength is chosen to be large enough @xmath124 so that the multiscale dynamics are challenging to approximate .
of particular interest are timescale separations of @xmath125 and @xmath120 .
first we consider a chaotic and strongly mixing slow regime @xmath126 .
figures are presented for the timescale separation @xmath125 only , because in this regime the picture is very similar for @xmath120 .
we also consider a weakly chaotic and quasi - periodic slow regime @xmath127 . in this regime , the coupled dynamics are more dependent on the timescale separation so we present results for both @xmath125 and @xmath120
. statistical quantities of other regimes , including regimes with more periodic behavior , have been presented in @xcite .
@xmath128 + [ cols="^,^ " , ] we observe that the first - order reduced system ideal response is a much closer approximation to the multiscale ideal response than the corresponding zero - order ideal response . in these regimes
the relative error of the first - order response is limited to about @xmath129 for the heaviside forcing and less for the ramp forcing , while in the zero - order system the error is around @xmath130 for the step forcing and @xmath131 for ramp forcing at time @xmath132 .
remark that in the third plot for ramp forcing response in figure [ fdt_ideal_ramp_error ] there is a small bump in the relative error shortly after the onset of forcing .
this plot corresponds to a weakly chaotic regime ( @xmath133 ) with a large timescale separation ( @xmath120 ) in the multiscale system . in this regime
the small nonlinear fluctuations of the multiscale system are relatively large compared to the ramp forcing for @xmath52 near zero , so the relative error of the reduced system responses is large . above in section [ meanresponse_subsection ]
we discussed the actual responses of the statistical mean states of both the two - scale and reduced models to small heaviside and ramp forcings .
for completeness of the study , we also attempt to predict the response of the mean state of the two - scale system via the quasi - gaussian linear response approximation @xcite of the reduced system . in the quasi - gaussian response approximation ,
the terms @xmath57 and @xmath58 in are replaced with the gaussian approximations with same mean state and covariance matrices as in the actual dynamics .
this , and the fact that @xmath134 in yields the following formula for the linear response approximation of the mean state response : [ fdt - resp ] @xmath135 @xmath136 where @xmath137 and @xmath138 are the mean state and covariance matrix of the corresponding unperturbed systems ( two - scale and reduced ) , computed as @xmath139 @xmath140 for large multiscale problems the mean response may be difficult to compute directly , since the large ensemble size needed for an accurate average is compounded by an already large number of variables and small timestep discretization . in the case where the mean response of the slow variables is desired , one might prefer to compute the fdt response ( [ fdt - resp ] ) using a time series from the reduced system for a `` quick and dirty '' approximation to the ideal response operator for the multiscale slow system .
we show the accuracy of this fdt response approximation for the lorenz 96 system , using a quasi - gaussian approximation with time series data from the zero- and first - order reduced systems .
the quasi - gaussian response snapshots for the response times @xmath141 and @xmath142 are shown in figures [ fdt_ideal ] and [ fdt_ideal_ramp ] for the heaviside and ramp forcing , respectively .
qualitatively , the quasi - gaussian response does capture the large features of the actual response , although most noticeable in these snapshots is the large exaggeration of the quasi - gaussian response calculated from the first - order reduced system , which predicts a much larger off - diagonal response than what is observed .
the possible reason for that is that the distribution densities of the first - order reduced model are more strongly non - gaussian than those of the zero - order reduced model , while the time autocorrelation functions are more weakly decaying ( see figure [ ddfs ] ) .
it was observed previously in @xcite that in these conditions the quasi - gaussian linear response approximation tends to overshoot the off - diagonal response by a large margin . in other words ,
the better precision of the quasi - gaussian linear response of the zero - order model is the result of mutual cancellation of the two errors : first one is the error in the distribution density of the zero - order reduced model ( significantly more gaussian than in in the multiscale dynamics ) , while the second one is the error in the quasi - gaussian linear response due to non - gaussianity of the statistical state ( less in the case of the zero - order model ) .
the relative error and cosine similarity are measured against the multiscale ideal response and can be seen for heaviside step forcing in figure [ fdt_ideal_error ] and for ramp forcing in figure [ fdt_ideal_ramp_error ] .
the ideal response of the first - order reduced system is clearly the best of the four responses at capturing response in the slow variables .
it is interesting , but perhaps not too surprising , that the least accurate estimate is given by the fdt response of the first - order reduced system .
this should be expected since the quasi - gaussian approximation is only valid for well - mixing systems whose distribution densities are close to gaussian , which in particular is the case for the uncoupled lorenz 96 systems in a chaotic regime @xmath143 .
however , such a system exhibits suppressed chaos when coupled to another chaotic systems , and the resulting distribution density will be far from gaussian @xcite .
since the first - order reduced system matches more closely the multiscale system , and the zero - order system will behave as an uncoupled lorenz 96 system , the first - order system will be less chaotic and will be a poor candidate for the quasi - gaussian fdt response .
in fact , for non - chaotic regimes , as in the case of @xmath144 where spatially periodic solutions emerge in the two - scale and first - order reduced systems , the long - time covariance matrix @xmath138 will be singular , so the quasi - gaussian response as presented will not be applicable
. for further reading on blended fdt responses which might be more effective in these cases , see @xcite .
in this work we studied the response to small external perturbations of multiscale dynamics and their reduced models for slow variables only . we elucidated a set of criteria for statistical properties of the multiscale and reduced systems which facilitated similarity of responses of both systems to small external perturbations .
it was shown that the similarity of marginal distribution densities of slow variables and their time autocorrelation functions controlled the similarity of responses to small external perturbations of both systems . like in @xcite , here we demonstrated that including a first - order correction term to a standard closure approximation for a nonlinear chaotic two - time system offered distinct improvements over the zero - order closure in capturing large - scale features of the slow dynamics .
in particular , this reduced system was able to accurately capture the distribution density of solutions as well as the mean state response of the system to simple forcing perturbations .
this correction term was relatively easy to generate , requiring only simple statistical calculations of the uncoupled fast system for an appropriate set of fixed parameters , and the resulting reduced system required much less computational resources than the corresponding multiscale system .
focusing on the mean state linear response of the slow variables , we showed that forcing perturbations in the reduced systems have similar responses as in the two - time system .
furthermore , we showed that using the unperturbed dynamics of the reduced systems for linear response prediction is also possible .
however , in the parameter regimes we present here the first - order reduced systems are not rapidly mixing and do not follow a gaussian distribution , but the zero - order reduced systems do have these properties , so this fluctuation - dissipation response is effective only using the zero - order system . a linear response method which takes into account the non - gaussianity of the invariant statistical state ( such as the blended response algorithm @xcite , based on the tangent map linear response approximation )
is apparently needed to capture the response for strongly non - gaussian dynamical regimes in reduced models . here
the linear response closure derivation and numerical results have been presented only for the special case of linear coupling between slow and fast systems , but this derivation has been extended to systems with nonlinear and multiplicative coupling @xcite . in future work
we hope to extend similar results to these more general systems and to test the robustness of this method for application to a large variety of problems
. * acknowledgment .
* rafail abramov was supported by the national science foundation career grant dms-0845760 , and the office of naval research grants n00014 - 09 - 0083 and 25 - 74200-f6607 .
marc kjerland was supported as a research assistant through the national science foundation career grant dms-0845760 .
g. papanicolaou .
introduction to the asymptotic analysis of stochastic equations . in r.
diprima , editor , _ modern modeling of continuum phenomena _ , volume 16 of _ lectures in applied mathematics_. american mathematical society , 1977 . | in real - world geophysical applications ( such as predicting the climate change ) , the reduced models of real - world complex multiscale dynamics are used to predict the response of the actual multiscale climate to changes in various global atmospheric and oceanic parameters . however , while a reduced model may be adjusted to match a particular dynamical regime of a multiscale process , it is unclear why it should respond to external perturbations in the same way as the underlying multiscale process itself . in the current work ,
the authors study the statistical behavior of a reduced model of the linearly coupled multiscale lorenz 96 system in the vicinity of a chosen dynamical regime by perturbing the reduced model via a set of forcing parameters and observing the response of the reduced model to these external perturbations .
comparisons are made to the response of the underlying multiscale dynamics to the same set of perturbations .
additionally , practical viability of linear response approximation via the fluctuation - dissipation theorem is studied for the reduced model .
multiscale dynamics ; reduced models ; response to external forcing * ams subject classifications . * 37 m , 37n |
the bulk of the observed x - ray emission from galaxy clusters is due to thermal bremsstrahlung in a hot gas ( @xmath10 ) enriched in heavy elements . in the past two decades
a great deal of attention has been devoted to the study of the abundances of heavy elements ( mostly iron ) in the intracluster medium ( hereafter icm ) . from the hot x - ray emitting intergalactic gas in clusters of galaxies a universal abundance of iron of roughly 1/3
solar has been derived ( renzini 1997 and references therein ; fukazawa et al .
attempts to explain this icm iron abundance can be traced back to the early 1970s , when for the first time the iron - emission line in the x - ray spectra of clusters of galaxies was discovered ( mitchell et al .
1976 ; serlemitsos et al .
some interpreted the presence of iron in the icm to be due to gas ejected from galaxies , either by means of galactic winds or ram pressure stripping ( gunn and gott 1972 ; larson and dinerstein 1975 ; vigroux 1977 ; de young 1978 ; sarazin 1979 ; himmes and biermann 1980 ; matteucci and vettolani 1988 ; renzini et al .
1993 ; matteucci and gibson 1995 ; gibson and matteucci 1997 ) , while others suggested pregalactic objects such as population iii stars as the origin ( white and rees 1978 ) .
recently , thanks to the results obtained with the asca satellite , more detailed observations of iron and @xmath0-element abundances in local galaxy clusters are becoming available . before the launch of asca ,
spectroscopic measures of elements such as si and s were known for the perseus cluster ( mushotzky et al .
1981 ) and a576 ( rothenflug et al . 1984 ) , indicating that the abundances of these elements as well as that of iron are roughly solar , although their level of accuracy did not allow one to make strong statements about possible differences between the @xmath0-elements and iron . concerning oxygen
there were only a few of measures of the o viii line in virgo and perseus ( canizares et al .
1988 ) indicating a higher than solar [ o / fe ] .
more recent measurements from asca by mushotzky et al .
( 1996 ) implied @xmath4 \approx$ ] + 0.2-+0.3 dex , thus indicating a general overabundance of @xmath0-elements relative to iron in the icm .
we derived from the mushotzky data an average @xmath11 ( we estimated a formal statistical uncertainty and no systematic effect ) .
this value corresponds ( by adopting the photospheric abundances of anders and grevesse ( 1989 ) ) to [ o / fe]@xmath12 dex .
this means , by considering the errors , a marginal ( if any ) overabundance of oxygen relative to iron .
the average si / fe ratio estimated from mushotzky data is @xmath13 corresponding to [ si / fe]= @xmath14 dex , which agrees with the same ratio obtained by fukazawa et al .
( 1998 ) in their analysis of about 40 nearby poor and rich clusters .
in fact , from this latter compilation of data we derived @xmath15 for clusters with @xmath16 kev ( we consider these clusters as rich according to our definition in tab .
2 ) and @xmath17 for clusters with @xmath18 kev ( poor clusters ) . ishimaru and arimoto ( 1997a , b ) and arimoto et al . ( 1997 ) claimed that the present uncertainties both on the assumed solar abundances ( used to derive the quantity [ o / fe ] ) and on the derived x - ray abundances are consistent with almost solar values of the @xmath4 $ ] ratios in the icm .
in particular , the use of the meteoritic abundances for the sun ( anders and grevesse 1989 ) instead of the photospheric ones used in the previous studies , would reduce the @xmath4 $ ] values quoted above by @xmath19 dex ( ishimaru & arimoto 1997a , b ; renzini 1977 ) . therefore , we consider all of these observational results still preliminary and subject to variation . concerning the evolution of the abundances and abundance ratios as a function of redshift
very little is known at the moment . however , preliminary results by mushotzky and lowenstein ( 1997 ) seem to indicate no evolution in the fe abundance for @xmath20 .
in this paper we study the chemical evolution of the icm , in particular we predict the evolution of fe and @xmath4 $ ] ratios as functions of redshift .
this is because the value of the @xmath4 $ ] ratio as well as the fe abundance in the icm impose strong constraints on the evolution of the galaxies in clusters as well as on the roles of supernovae of different types in the enrichment of the icm , as discussed by renzini et al .
( 1993 ) and matteucci and gibson ( 1995 ) .
since the heavy elements in the icm are likely to come from ellipticals , as shown by arnaud ( 1994 ) , it is important to explore different kinds of models for the evolution of ellipticals . in particular , we will use either a one - zone model ( matteucci and gibson 1995 , hereafter referred to as `` burst model '' ) and a multi - zone model ( martinelli et al .
1998 , hereafter `` continuous model '' ) for elliptical galaxies . the continuous model , which predicts abundance gradients in ellipticals in very good agreement with the observed ones , assumes that the outermost regions of these galaxies experience early galactic winds whereas the innnermost ones keep an active , although low , star formation rate until late times .
the physical reason for this resides in the fact that the binding energy of the gas is lower in the outermost galactic regions and this is supported by the observational finding of a correlation between metallicity and escape velocity in ellipticals ( carollo and danziger , 1994 ) .
the late occurrence of galactic winds in the most internal regions of ellipticals , if true , will have an effect on the chemical evolution of the icm , certainly at variance with the predictions by models with only early winds ( matteucci and gibson 1995 ; gibson and matteucci , 1997 ) .
the plan of this paper is the following .
in section 2 we present the models and the computational method . in section 3
the results are discussed and some conclusions are drawn .
we use throughout the paper an adimensional hubble parameter @xmath21 .
the mass in the form of specific chemical elements as well as the total gas masses ejected by the elliptical galaxies both at early and late galactic lifetimes into the icm , can be computed in detail once a chemical evolution model is assumed .
two different models for the evolution of elliptical galaxies have been taken into account : 1 . _ burst models _ , namely those with only one major episode of galactic wind which occurs simultaneously all over the galaxy ; 2 . _
continuous models _ , where the galactic wind starts at an early epoch in the most external regions of the galaxies and continues until the present time in the most internal ones .
the burst model is the same as in matteucci and gibson ( 1995 ) , where we direct the reader for details , in the case of the arimoto & yoshii ( 1987 ) imf .
this is their best model and we refer to it as cwm . the continuous wind model is described in martinelli et al .
( 1998 ) , where we direct the reader for details .
the basic difference of this approach , relative to the model of matteucci and gibson ( 1995 ) , is that the elliptical galaxy is divided in several shells not interacting . for each shell
the potential energy of the gas is calculated and compared with the thermal energy of the gas due to supernova explosions .
a galactic wind develops first in the outermost shells and then progressively in the most inner ones . an improved expression for the cooling time of sn ( of both types ) remnants as well as adetailed description of the potential energy of the gas
are used in this model .
the sn remnant cooling time strongly influences the occurrence of a galactic wind as well as the amount of energy that sne can transfer into the interstellar medium ( ism ) . in the cwm model ,
a cooling time not dependent on time was used , whereas in the continuous models the cooling time depends on the gas density ( @xmath22 ) which is a function of time . in both models , after the occurrence of a galactic wind it is assumed that star formation is no longer taking place , and the galaxy or the galactic region evolves passively after that . in the case of the continuous model , the innermost regions ( inside 30 pc )
never develop a galactic wind and the star formation continues until late times although it is so low that it would be hardly detectable .
the chemical evolution of the icm is caused by two different processes : 1 .
in the burst model , galaxies begin to eject the gas only after the time of occurrence of a galactic wind ( @xmath23 ) , and this process lasts only for few @xmath24 years .
the time @xmath23 is a function of the initial galactic mass which influences the potential well .
therefore , when the time of the first occurrence of the galactic wind increases , namely when more and more massive galaxies achieve the conditions to develop a wind , the number of galaxies contributing to the enrichment of the icm also increases , as shown in figure 1 .
this type of evolution stops at the time @xmath25 , corresponding to the onset of the galactic wind for the galaxy with the highest mass considered .
2 . in the continuous models ,
the icm evolution is caused by the gas ejected continuously by each galaxy after the time @xmath23 .
this continuous ejection process , not present in the burst models , has the characteristic of acting also at low redshift and , in particular , now .
the assumed imf is the same as in the burst model .
we now elucidate , in some detail , how we have calculated the chemical abundances in the icm for several values of redshift , according to the two elliptical galaxy models described above .
we computed the ejected masses of each element ( fe , si , o , and total gas ) as functions of several initial galactic masses ( the procedure we describe here can be also found in matteucci and vettolani , 1988 ) .
we found that the relation between the generic chemical element @xmath26 and the initial galactic mass can be approximated by a power law : @xmath27 where @xmath28 represents the mass ejected in the form of the chemical species @xmath26 by a galaxy of initial mass @xmath29 , and @xmath30 and @xmath31 are two constants ( that can be fixed by a least squares fit ) .
we have normalized the relations ( 1 ) to a galaxy at the _ break _ of the luminosity function of a cluster ( schechter 1976 ) having a blue luminosity @xmath32 , absolute blue magnitude @xmath33 and total mass @xmath34 . in particular : @xmath35 where @xmath36 represents the mass ejected in the form of an element @xmath26 by the galaxy with mass @xmath34 .
.age of the elliptical galaxies for several values of the redshift @xmath6 in the two cosmological models examined [ cols="<,<,<",options="header " , ] the evolution of the elliptical galaxies is considered in the framework of different cosmological models : a flat , scale invariant cdm cosmology with @xmath37 , @xmath38 and a vacuum dominated cdm model with @xmath39 , @xmath40 and @xmath41 . in all the models
considered we set the galaxy formation epoch , @xmath42 , and their age @xmath43 gyr ( we consider the same age for all the galaxies in our code ) .
the models are defined in table 3 .
in this paper we have computed the evolution of the abundances of fe and @xmath0-elements in the icm of poor and rich clusters as a function of redshift under different cosmologies . to do that we have used two different models for the chemical evolution of elliptical galaxies , which are the main contributors to the icm enrichment . in particular , we used a model where the galactic wind develops at the same time all over the galaxy ( the burst model ) and after that the galaxy evolves only passively not contributing any longer to the icm enrichment , and a model where galactic winds develop first at large galactocentric distances and then progressively later in the innermost zones ( continuous model ) .
our results are shown in figures 2 - 7 , where the behaviour of the @xmath44 $ ] , @xmath45 $ ] and @xmath46 $ ] as functions of the redshift @xmath6 in the cases of burst models ( fig . 2 , 3 , 6 ) and continuous models ( fig .
4 , 5 , 7 ) is indicated .
it should be noted that the [ fe / h ] and the [ @xmath0/h ] abundances refer to the abundances expected in the total gas ejected by ellipticals and not mixed with the pristine gas in the cluster .
previous calculations ( matteucci and vettolani , 1988 ; matteucci and gibson , 1995 ; gibson and matteucci 1997 ) have shown that while the predicted total iron mass ejected by ellipticals agrees with the observed iron mass in clusters the total mass of gas ejected from ellipticals is much lower than the observed total gas mass in clusters .
this is interpreted as due to the fact that most of the gas in clusters should be pristine gas .
the models presented here also predict a too small amount of ejected total gas as compared to the observed one . a larger production of total gas from galaxies
can be obtained by adopting a steep faint end slope of the luminosity function .
however , gibson and matteucci ( 1997 ) showed that in this case the contribution of the dwarf ellipticals is at most @xmath47 of the total icm gas , so that the conclusion about most of the icm gas being pristine gas remains unaffected . in order to obtain the abundances in the icm
we should rescale the abundances shown in figures 2 and 4 to the total observed mass of gas in a typical poor and a rich cluster by assuming that most of it is pristine gas . in order to compute the evolution of the abundances as a function of redshift we also assumed that the amount of pristine gas in the clusters did not change during the cluster lifetime , in agreement with the results of white et al .
( 1993 ) which found that the baryon fraction in clusters should have been constant . after doing that by taking @xmath48 as representative of the total gas of a rich cluster ( coma ) and @xmath49 as representative of a poor cluster ( virgo ) we found that , in order to represent the @xmath50_{icm}$ ] , the curves of figure 2 should be lowered by -1.48 and -1.35 dex for a rich and a poor cluster , respectively , whereas the curves of figure 4 should be lowered by -1.37 and -1.30 for a rich and a poor cluster , respectively .
the present time iron abundance , @xmath51 , obtained in this way from the continuous models is in very good agreement with the observational estimates of the iron abundance in local clusters .
in fact , the continuous models predict @xmath52 both for poor and rich clusters , whereas the burst model predicts a lower abundance ( @xmath53 . on the other hand ,
the abundance ratios [ @xmath0/fe ] are completely unaffected by the amount of gas in the clusters and therefore the abundance ratios in the gas ejected from ellipticals represent also the abundance ratios in the icm .
for this reason we want to stress the importance of considering abundance ratios rather than absolute abundances .
* we observe that in the burst models the chemical evolution of the icm stops entirely at @xmath54 when the cosmological parameters are @xmath39 , @xmath40 , @xmath41 and at @xmath55 when @xmath37 , @xmath56 and @xmath38 , respectively .
* in the case of continuous models there is a continuous evolution up to @xmath57 and negligible evolution from @xmath58 up to @xmath3 . *
the predicted [ fe / h ] at the present time in the icm is in very good agreement with the observations ( 1/3 solar ) for the continuous models whereas the burst models predict a lower abundance .
* for the @xmath4_{icm}$ ] ratios we predict @xmath4_{icm}>0 $ ] at @xmath3 in the burst models and @xmath4_{icm } \le 0 $ ] at @xmath3 for the continuous models , the exact values of the @xmath4_{icm}$ ] ratio depending on the assumed cosmology . in particular , at @xmath59 we predict @xmath8_{icm } \simeq + 0.27 \div + 0.37 $ ] dex and @xmath9_{icm } \simeq + 0.14 \div + 0.15 $ ] dex for the bursts models and @xmath8_{icm } \simeq -0.12 \div -0.05 $ ] dex and @xmath9_{icm } \simeq + 0.11 \div + 0.13 $ ] dex for the continuous models .
these values were derived by adopting the meteoritic solar abundances ( anders and grevesse , 1989 ) .
+ in both cases ( different normalization to the solar abundances ) the predicted @xmath4_{icm}$ ] ratios at the present time are higher in the burst than in the continuous model .
the main reason for this behaviour of the @xmath4_{icm}$ ] ratio in the two different cases is the fact that in the burst model the galactic wind develops at early times in all galaxies and therefore it is mainly enriched by the products of sne ii ( high @xmath4 $ ] ratio ) . on the other hand , in the continuous models
the galactic wind continues to develop until the present time and therefore it is enriched also by the products of sne ia ( low @xmath4 $ ] ratio ) .
this result is quite important since it depends uniquely on the evolution of the ellipticals and not on the particular history of the cluster evolution ( infall , etc .. ) as it is for absolute abundances .
we note that the predicted present time @xmath9_{icm}$ ] ratios are larger than the present time @xmath8_{icm}$ ] for the continuous models .
this is due to the different nucleosynthesis of si and o : o , in fact , is entirely produced by sne ii whereas si is mostly produced by sne ii but a fraction , not entirely negligible , arises from sne ia and this contributes to avoid that the si / fe ratio decreases too much , since the bulk of fe coming from sne ia is partly compensated by the si which comes also from sne ia .
for this reason si is not the best @xmath0-element to be used to check the @xmath4 $ ] ratio , whereas mg and o should be preferred .
+ finally , it is worth noting that these @xmath4_{icm}$ ] ratios have been obtined by adopting a flat imf in the galactic models ( @xmath60 ) .
these values are therefore upper values especially in the case of the continuous model , since the adoption of the salpeter imf ( @xmath61 ) would lead to lower present time values of these ratios , due to the larger contribution from type ia sne relative to type ii sne . * for both galactic models these results do not differ for rich or poor clusters . *
the effect of a change in the cosmological parameters is a simple translation of the abundance ratios ( namely the @xmath4 $ ] ratios are lower by at most @xmath62 dex when @xmath39 than when @xmath37 ) .
a change in the cosmology produces also a delay in the chemical evolution , in the sense that when @xmath37 the chemical evolution of the icm stops a little later than when @xmath39 .
this is a consequence of the different correspondence between the redshift @xmath6 and the galactic age @xmath63 as shown in table 1 . from the previous results we can conclude that if we want to see an evolution in the icm we have to be able to perform observations of abundances in very high redshift clusters ( at least @xmath64 ) . preliminary results from asca ( mushotsky and lowenstein , 1997 ) seem to indicate no evolution in the abundance of fe for @xmath65 in agreement with our results related to all models here considered
therefore , on this basis alone we can not distinguish among the two models . concerning our predictions on the @xmath4_{icm}$ ] ratios , recent studies ( ishimaru and arimoto , 1997a , b ) adopting the meteoritic solar abundances of anders and grevesse ( 1989 ) as in this work , seem to favor a low value of this ratio in very good agreement with our continuous wind model predictions .
anders , e. , grevesse , n. , 1989 geochim.cosmochim .
acta 53 , 197 arimoto n. , yoshi y. , 1987 , a&a , 173 , 23 arimoto n. , matsushita , k. , ishimaru , y. , ohashi , t. , renzini , a. , 1997 , ap.j . , 477 , 128 arnaud , m. , 1994 , in `` clusters of galaxies '' , eds . f. durret et al . ,
editions frontieres , gif - sur - yvette , p.211 canizares , c.r . , markert , t.h . , donahue , m.e . , 1988 , _
cooling folws in clusters and galaxies _ , ed .
fabian , kluwer : dordrecht p.63 carollo , m. , danziger , i.j .
, 1994 , mnras , 270 , 523 de young , d.s . , 1978 , apj , 223 , 47 fukazawa , y . , makishima , k . ,
tamura , t. et al .
, 1998 , pasj , 50 , 187 gibson b.k . , matteucci , f. , 1997 , apj , 457 , 47 gunn , j.e . and
gott , j.r .
, 1972 , ap.j .
, 176 , 1 himmes , a. and biermann , p. , 1980
, a&a , 86 , 11 ishimaru , y. and arimoto , n. , 1997a , pasj , 49 , 11 ishimaru , y. and arimoto , n. , 1997b , iaus 187 , _ cosmic chemical evolution _ , kyoto , japan larson r. b. and dinerstein , h.l . , 1975 ,
pasp , 87 , 911 martinelli a. , matteucci f. and colafrancesco s.,1998 mnras 298 , 42 .
matteucci , f. and vettolani , g. , 1988 , a&a , 202 , 21 matteucci f. and gibson b. , 1995 , a&a , 304 , 11 mitchell , r.j . ,
culhane , j.l . , davison , p.j.n . , ives , j.c . , 1976 , mnras , 176 , 29 mushotzky r.f . ,
holt , s.s . ,
smith , b.w .
, boldt , e.a . , 1981 , apj , 225 , 21 mushotzky r.f . , loewenstein , m. , arnaud , m. , tamura , t. fukazawa , y. et al .
1996 , ap.j . , 466 , 686 mushotzky r.f . , lowenstein m. , 1997 , apj , 481 , l63 renzini , a. , ciotti , l. , dercole , a. , pellegrini , s. , 1993 , ap.j . , 419 , 52 renzini , a. , 1997 , ap.j . , 488 , 35 rothenflug , r. , vigroux , l. , mushotsky r.f . ,
holt , s.s . , 1984 , ap.j . , 279 , 53 schechter , p , 1976 , ap.j . , 203 , 297 selermitsos , p.j . ,
smith , b.w . ,
boldt , e.a . ,
holt , s.s . ,
swank , j.h .
, 1977 , ap.j . , 211 , l63 sarazin , c.l . , 1979 ,
lett . , 20 , 93 vigroux , l. , 1977 , a&a , 56 , 473 white , s.d.m . and rees , m.j . , 1978 , mnras , 183 , 341 white , s.d.m . ,
navarro , j.f . ,
evrard , a.e . , frenk , c.s .
1993 , nature , 366 , 429 | we predict the behaviour of the abundances of @xmath0-elements and iron in the intracluster medium as a function of redshift in poor and rich clusters . in order to do
that we calculate the detailed chemical evolution of elliptical galaxies by means of one - zone and multi - zone models and then we integrate the contributions to the total gas and single elements by ellipticals over the cluster mass function at any given cosmic time which is then transformed into redshift according to the considered cosmological model . in the case of the multi - zone model for ellipticals
the more external regions evolve much faster than the internal ones which maintain a very low level of star formation almost until the present time . in other words ,
the outermost regions develop a galactic wind , after which the region evolves passively , at much earlier times than the innermost regions as opposed to the classic one - zone model where the galactic wind develops at early times over the whole galaxy .
we refer to the one - zone model as to _ burst model _ and to the multi - zone model as to _
continuous model_. we find that in the case of the burst model the icm abundances should be quite constant starting from high redshifts ( @xmath1 ) up to now , whereas in the continuous model the icm abundances should increase up to @xmath2 and are almost constant from @xmath2 up to @xmath3 .
particular attention is devoted to the predictions of the @xmath4_{icm}$ ] ratio in the icm : for the burst model we predict @xmath4_{icm}>0 $ ] over the whole range of redshifts and in particular at @xmath5 , whereas in the case of the continuous model we predict a decreasing @xmath4_{icm}$ ] ratio with decreasing @xmath6 and @xmath4_{icm } \le 0 $ ] at @xmath7 .
in particular , we predict @xmath8_{icm}(z=0 ) \le + 0.35 $ ] dex and @xmath9_{icm}(z=0 ) \le + 0.15 $ ] dex for the bursts models and @xmath8_{icm}(z=0 ) \le -0.05 $ ] dex and @xmath9_{icm}(z=0 ) \le + 0.13 $ ] dex for the continuous models , the precise values depending on the assumed cosmology . finally , we discuss the influence of different cosmologies on the results .
psfig.tex # 1to 0pt#1 |
the origin of young stars near sgra * has been a major puzzle of the galactic center ( gc ) for over a decade ( phinney 1989 , sanders 1992 , morris 1993 ) .
one of the few promising explanations assumes that the stars were originally bound to an intermediate - mass black hole ( imbh ) which has inspiraled from large distances ( few parsecs ) to the immediate vicinity ( @xmath1pc ) of sgra * ( hansen and milosavljevic , 2003 ) . in this scenario
the imbh itself would be a compact remnant of a superstar of several thousand solar masses .
this superstar could be a product of stellar mergers at the center of dense young massive cluster ( portegies zwart et .
2004 , gurkan et .
al . 2004 ) or could be produced in a fragmenting massive accretion disc ( levin and beloborodov 2003 , levin 2003 , goodman and tan 2004 ) .
regardless of its origin , the inpiraling imbh will eject a number of stars from the gc and send them flying through ( and out of ) the galaxy at high velocities ( yu and tremaine 2003 , see also gualandris et .
brown et .
2005 have recently identified a high - velocity star in sdss data and have argued that it may have been ejected from the gc .
the european satellite gaia , scheduled to fly in 2012 , will detect and measure proper motions of stars to magnitude 20 ; the sun has magnitude of 19.6 at @xmath2kpc from earth ( perryman et .
al . 2001 ) .
the threshold for radial velocity measurement by gaia is higher magnitude 17 ; however the few high - proper - motion stars detected by gaia could be observed by the ground - based telescopes and their radial velocities could be determined .
thus , many high proper - motion - velocity stars could potentially be detected , their spatial position in the galaxy could be measured and their radial velocities could be determined from follow - up observations . with their 6-d phase - space information determined , their trajectories could be traced back to the gc and their times of ejection could be approximately determined ; this point was recently and independently emphasized by gnedin et .
( 2005 ) . in this paper
we derive the phase - space distribution of the ejected stars , and explain qualitatively how 6-d information could be used to constrain the imbh orbital evolution . in the next paper
we will carry out the rigorous modeling of the signal expected by gaia from an inspiraling imbh .
the typical mass ratio of an imbh and sgra * ranges from @xmath3 to @xmath4 .
therefore the gravitational interaction of a star with the imbh which results in the star s ejection occurs very close to the imbh and can be considered instantaneous , in a sense that the imbh velocity is not affected by the central black hole ( sgra * ) during the interaction .
moreover , the imbh has an immediate effect on the distribution function of stars in the cusp only in it s immediate vicinity ; therefore the stars interacting with the imbh have initial velocities drawn from the distribution function of the cusp at the instantaneous location of the imbh .
the differential cross - section of instantaneous encounters and the associated rates of stellar ejections were derived by michel henon in a series of seminal papers ( 1960a , b , 1969 , see also lin and tremaine 1980 ) .
our calculations are different from henon s , since we care about the distribution of the velocity vectors of ejected stars , and not just their energies . also , we consider a single ejector ( the imbh ) and do not average over the population of ejectors .
the derivations below are of geometric nature ; they are informed by but do not follow henon s work .
consider an imbh flying through the stellar cusp with velocity @xmath5 at radius @xmath6 from sgra*. it is reasonable to assume that the cusp is composed of stars bound to the central black hole .
therefore , in velocity space , all the stars interacting with the imbh are inside the sphere bounded by the escape velocity at @xmath6 , @xmath7 see fig . 1 . in the equation above we have assumed that the potential is dominated by central black hole of mass @xmath8 , which is an excellent approximation at the location of the puzzling young stars .
first , let us calculate the rate of ejection of stars with velocity @xmath9 , where @xmath10 and @xmath11 . here and further @xmath12 , @xmath13 , etc .
stand for the magnitudes of the corresponding velocity vectors .
figure 1 illustrates the velocity - space of the problem at hand .
let @xmath14 and @xmath15 be the velocity vector of a star before and after scattering by the imbh , respectively .
it is convenient to consider the problem in the frame of reference of the imbh , and define @xmath16 the dashed region of the velocity space marks the stars which can be scattered to velocities @xmath17 ; analytically it is defined by two conditions : @xmath18 the vector of the outgoing relative velocity @xmath19 must satisfy @xmath20 which can be translated into a constraint on the angle @xmath21 between @xmath5 and @xmath19 : @xmath22 the rate of ejection with velocities greater than @xmath23 is given by @xmath24 where the integral is performed over the dashed volume in fig . 1 , @xmath25 is the velocity distribution function , and @xmath26 is the cross - section for the star with initial relative velocity @xmath27 to be scattered into the velocity cone described by eq .
( [ phiout ] ) .
this cross - section is evaluated in the appendix : @xmath28 where @xmath29 is the mass of the imbh and @xmath30 is the angle between @xmath5 and @xmath27 .
now , let us find the differential ejection rate @xmath31 , where @xmath32 is the incremental volume of the velocity space centered on @xmath15 and @xmath33 is the rate of ejections of stars into this volume .
it is convenient to express @xmath32 as @xmath34 where @xmath35 and @xmath36 is the increment of the solid angle around the vector @xmath19 .
the differential ejection rate is then given by @xmath37 here @xmath38 is the differential crossection for gravitational scattering from @xmath27 to @xmath19 , and @xmath39 is the solid angle increment of the integrated vector variable @xmath27 .
the differential ejection rate is particularly easy to evaluate if the distribution function @xmath40 is axially symmetric about the vector @xmath5 .
the ejection rate is then also axially symmetric . by averaging eq .
( [ diffrate ] ) over velocity vectors @xmath15 which are connected by rotation about @xmath5 one can show that @xmath41 where @xmath42 here @xmath43 , and @xmath44 as given by eq .
( [ a ] ) . for a general distribution function @xmath25 , the integrals in eqs .
( [ rate1 ] ) and ( [ diffrate1 ] ) must be integrated numerically . in the next section
, we consider an important particular case where these integrals can be evaluated analytically .
a commonly considered cusp distribution function which is convenient for analytical work is the peebles - young distribution function @xmath46 ; this corresponds to an isotropic cusp with the density profile @xmath45 ( young 1980 ) .
this profile is also very close to that of the observed cusp in the galactic center ( genzel et .
al . 2003
; see below ) , and hence in this section we will consider in some detail the distribution of the stars ejected by the imbh inspiraling through the peebles - young cusp .
the ejection rate is obtained by evaluating the integral in eq .
( [ rate1 ] ) .
we have found @xmath47 and @xmath30 to be convenient integration variables ; it is easiest to integrate first over @xmath30 and then over @xmath47 .
one then finds the following for the ejection rate : @xmath48 where @xmath49 , @xmath50 , @xmath51 is the local number density of stars and @xmath52\\ & + & \left[{1\over q - p}-{1\over 1+p}\right]{q^2-p^2\over 2p}\left[1-{q^2-p^2\over q^2 - 1}\right]\nonumber\\ & -&{(1+p)^3-(q - p)^3\over
6p(q^2 - 1)}-\ln \left({1+p\over q - p}\right).\nonumber\label{lambda}\end{aligned}\ ] ] the differential ejection rate is then obtained by evaluating the integral in eq .
( [ diffrate1 ] ) , and is given by @xmath53\right\},\label{diffrate2}\end{aligned}\ ] ] where @xmath54 and @xmath55 is the angle between @xmath15 and @xmath5 .
let us put make a numerical estimate for eq .
( [ rate2 ] ) with the parameters relevant for the galactic center .
we are interested in stars with luminosity no less than solar , since we want the star to be observable by gaia at the distance of @xmath56 kpc . at the moment the observational constraint on the number of such stars in the central cusp is poor , as the faintest stars observable by genzel s and ghez groups with the vlt and keck have masses @xmath57 ( r. genzel and a. ghez 2005 , private communications ) . among the stars which are observable , the mass function does not follow salpeter s law , and is clearly top - heavy . the imf favoring high masses
is consistent both with stars born in a dense accretion disc ( artymowicz et .
al . 1993 , levin 2003 , levin and beloborodov 2003 ) or brought in as a part of a core of a dense young cluster ( gerhard 2001 ) .
to make further progress we need to parametrize the distribution of stars with masses @xmath58 in the central cusp : @xmath59 the slope @xmath60 was measured by genzel et .
al . ( 2003 ) and shodel et .
( 2005 ) to be @xmath61 and @xmath62 , respectively , with both of the analyses assuming constant mass - to - light ratio .
thus the cusp slope is close to the peebles - young profile , and in the fiducial example below we will assume @xmath63 . as explained above , the normalization factor @xmath64 is not well known at present ; @xmath65 corresponds to the case when the mass density at @xmath66pc quoted in eq .
( 2 ) of genzel et .
al . ( 2003 ) is dominated by solar - mass stars . with this parametrization
the ejection rate becomes @xmath67 consider now an imbh of mass @xmath68 which is inspiraling into sgra * on a circular orbit . for a circular imbh orbit , @xmath69 and the function @xmath70 is plotted as a function of @xmath71 in fig . 2 .
for the peebles - young profile the inspiral proceeds at constant rate , and the radius and the escape velocity of the orbit are given by @xmath72 where @xmath73 is the dynamical friction timescale , with @xmath74 being the coulomb logarithm [ cf .
( 13 ) and ( 14 ) of levin , wu , and thommes 2005 ; our @xmath74 is their @xmath75 .
let us now calculate , as a function of time , the number of stars which are ejected during the inspiral and which will have velocities greater than @xmath76km / sec once they climb through the bulge potential .
we assume that the escape velocity from the bottom of the bulge is @xmath77km / sec ( see yu and tremaine 2003 for discussion ) . therefore , after collision with the imbh the star must have velocity greater than @xmath78 to evaluate the number of ejected solar - type stars as a function of time passed since the beginning of the inspiral , we integrate eqs .
( [ rate2.5 ] ) , ( [ rinsp ] ) , and ( [ vej4 ] ) with respect to the time variable .
the result is shown in fig .
3 , for a particular case of @xmath65 and @xmath79 .
the rate of ejections as a function of time is shown in fig .
the solid part of the curve , i.e. the part with @xmath80myr , marks the first stage of the inspiral when the imbh mass is smaller than the mass of the stellar cusp out to the radius of the imbh orbit .
therefore , the black - hole inspiral does not significantly affect the stellar distribution function of the cusp , and our analysis is reliable . for @xmath81myr , which corresponds to orbital radius @xmath82pc
, the imbh dominates the mass budget inside its orbit and depletes its neighborhood of stars on the timescale of @xmath83 .
therefore , within our analytical formalism we can not reliably predict the rate of ejections since the original distribution function of the cusp is strongly altered by the imbh ) mass ratios considered here . ] .
we model very roughly this imbh - dominated regime by positing that the overall density of stars and the rate of inspiral are declining exponentially on the timescale of @xmath83 , while assuming that the overall form of the cusp distribution function remains the same .
the dashed part of the curve shows the number of ejections calculated within this rough model . to get a better intuition for the parameters of ejected stars , we have generated mock data based on the ejection rates computed above .
we present this data as a scatter plot in fig . 5 :
on the @xmath84-axis we show the time of ejection and on the @xmath85-axis we show the velocity of the ejected star after it escapes from the bulge . with gaia
one will be able to clock the high - velocity stars back to their ejections and thus measure both @xmath86 and 3-d velocity for many of the high - velocity stars within 10 kpc from the earth whose proper - motion vector is pointing to the galactic center . building a scatter plot like the one in fig . 5 for the ejected high - velocity stars will provide a clear diagnostic for the imbh inspiral as the origin of the ejected stars . on the theoretical side
, we have a problem - more than half of the ejected stars are produced during the part of the inspiral which can not be modeled analytically .
therefore , we have conceived , but not yet carried out , a series of numerical experiments aimed at addressing the part of the inspiral where imbh mass dominates that of the stellar cusp . for
now , we limit ourselves to two qualitative remarks:(a ) the number of the stars ejected during a single inspiral is dependent on the poorly - known distribution of stars in the cusp but may be counted in hundreds .
these stars are potential targets for gaia.(b ) the stars are ejected in a burst the most intense part of which lasts 23 dynamical friction timescales . if no more than one imbh per 10 myr inspirals into sgra * , than the high - velocity stars seen by gaia should be all generated by a single inspiral event .
how anisotropic are the ejected stars ?
first we must note that the velocity vector @xmath87 of a star after it escapes the potential of the central black hole is different from @xmath15 , which is the stellar velocity right after gravitational collision with the imbh .
generally the velocity of the escaped star @xmath88 can be found by solving for the keplerian hyperbolic trajectory of the star ; here @xmath89 is the position of the imbh during the gravitational collision .
the mathematical expressions for keplerian hyperbolae can be found in a standard textbook on celestial mechanics , e.g. danby ( 1988 ) .
a useful way to quantify the anisotropy of ejections is to compute the dispersion of angles @xmath90 between the velocity vectors of the ejected stars and the inspiral plane .
this dispersion @xmath91 can be calculated by evaluating the following integral over a single period of the imbh orbit : @xmath92 , \label{sigma}\ ] ] where @xmath93 is the time , @xmath94 is the period of the orbit , @xmath95 is given by eq .
( [ diffrate ] ) [ and in appropriate cases by eq .
( [ diffrate1 ] ) or ( [ diffrate2 ] ) ] , and @xmath96 is the number of stars ejected during an orbital period : @xmath97 in fig . 6 we plot @xmath91 vs time for the circular imbh inspiral , with the thick horizontal line corresponding to isotropic distribution of the velocity vectors . at early times the stars ejected with velocities @xmath98km
/ sec will be clustering towards the inspiral plane , while at later times the velocity vectors of the ejected stars have no strong anisotropy .
unfortunately , at early times the ejection rate is low , and thus in the example we are considering only the first 10 or so ejections will show strong anisotropy .
gaia will see at most about 1/4 of the ejected stars ; we think therefore that detection of spatial anisotropy for circular inspiral is unlikely . [ [ eccentric - inspiral . ] ] eccentric inspiral .
+ + + + + + + + + + + + + + + + + + + there is a stronger chance to detect spatial anisotropy of the ejected stars if the imbh inspirals on a strongly eccentric orbit .
most high - velocity stars will be ejected when imbh passes the pericenter .
these stars will then form a broad jet which will have the general direction of the imbh velocity vector at the pericenter , i.e. perpendicular to the runge - lenz vector of the imbh orbit .
the elliptical orbit of the imbh will precess in the potential of the cusp , and the jet will rotate with the orbital ellipse . therefore , the rate of stars ejected into the region of the galaxy observable by gaia ( the gaia sphere ) will be modulated with the period of the imbh precession . _
it is this periodic modulation that is potentially observable , and that would be a tell - tale of the eccentric imbh inspiral_. for the peebles - young cusp the period of precession depends only on the imbh eccentricity and does not depend on its semimajor axis .
thus during the first stage of the inspiral , when the cusp is not affected by the imbh , the period of modulation will be constant and equal to @xmath99 see munyaneza et .
al . ( 1999 ) .
here we have taken the imbh eccentricity @xmath100 .
the galactic center cusp has spatial distribution similar to the peebles - young profile , and therefore the period of precession will be nearly constant during the first stage of the inspiral . during the second stage ,
however , when the imbh depletes the stellar density of the cusp , the precession period becomes longer .
detection of this period change would clearly provide a strong probe of the cusp structure . in fig .
7 we show the rate of stars injected into the gaia visibility sphere as a function of the injection time ( all computed for the fiducial cusp parameters described above ) . the star is defined to be in the gaia visibility sphere if at the current time , taken to be @xmath2myr , the star is within @xmath2kpc from the earth .
the orbit of the imbh is taken to lie in the galactic plane .
we clearly see in fig .
7 both the periodicity in the injection rate and the change in this periodicity when the imbh begins to empty its local region of the cusp .
in this paper we have derived the analytic expression for the time - dependent phase - space distribution of stars ejected from the galactic center as a result of the imbh inspiral .
we have discussed how gaia , together with the follow - up ground - based radial velocity measurements , may be used to track the history of ejections from the galactic center ; this point was recently and independently made by gnedin et .
al . 2005 .
we find that a single imbh inspiral produces a burst in the rate of ejection which has a duration a few dynamical friction timescales . in the very beginning of the burst ,
the velocities of the ejected stars cluster around the inspiral plane , however this anisotropy will be hard to detect and most of the stars will be ejected isotropically .
if the imbh orbit is eccentric , then the ejected stars form a broad `` jet '' roughly aligned with imbh velocity in the pericenter .
because of the cusp potential , the imbh orbit will precess with the period of @xmath0 years , and the rate of stellar ejections into the gaia visibility zone will be strongly periodically modulated . when the imbh begins to modify stellar distribution in its neighborhood ( i.e. , when its mass becomes comparable to the mass of stars inside its orbit ) ,
the period of precession will increase , as shown in fig .
therefore , the burst in the ejection rate and potential modulation of this rate during the burst , if detected by gaia , would be a clear signature that an imbh has inspiraled into the galactic center in the past 10 - 20 myr .
we thank tim de zeeuw for carefully reading the manuscript and making a number of important suggestions , and scott tremaine for encouraging discussions .
this research was supported by nserc .
in this appendix we evaluate the crossection for a star with the initial velocity @xmath27 relative to the imbh to scatter into the cone given by eq .
( [ phiout ] ) @xmath102 here @xmath21 is the angle between @xmath19 and the velocity vector of the imbh .
let @xmath103 be the plane which is perpendicular to @xmath27 and which contains the imbh .
let @xmath104 be the impact parameter vector , i.e. a vector directed from the imbh to the projection of the star onto @xmath103 before the star is scattered by the imbh .
after the scattering is complete , the stellar velocity relative to the imbh is given by @xmath105 where @xmath55 is the angle of deflection given by @xmath106 from eq .
( [ wf ] ) we see that equations ( [ cosphiout ] ) and ( [ theta2 ] ) together with eq .
( [ phiout1 ] ) define a finite domain in the @xmath104-plane .
it is straightforward but somewhat tedious to find the boundary of this domain ; this calculation is not shown here .
the area of the domain is the crossection we seek and is given by @xmath109 which is the equation ( [ a ] ) in the text .
artymowicz , p. , lin , d. n. c. , & wampler , e. j. 1993 , apj , 409 , 592 brown , w. b. , et .
2005 , apj , 622 , l33 danby , j. m. a. 1988 , fundamentals of celestial mechanics , richmond , va .
, usa , willmann - bell , 2nd ed .
gerhard , o. 2001 , apj , 546 , l39 gnedin , o. y. , et .
2005 , astro - ph/0506739 goodman , j. , & tan , j. c. 2004 , apj , 608 , 108 gualandris , a. , portegies zwart , s. , & sipior , m. 2005 , accepted to mnras , astro - ph/0507365 gurkan , m. a. , freitag , m. , & rasio , f. a. 2004 , apj , 604 , 632 hansen , b. m. s. , & milosavljevic , m. 2003 , apj , 593 , l77 henon , m. 1960a , anap , 23 , 467 henon , m. 1960b , anap , 23 , 668 henon , m. 1969 , a&a , 2 , 151 levin , y. 2003 , astro - ph/0307084 levin , y. , & beloborodov , a. m. 2003 , apj , 590 , l33 levin , y. , wu , a. s. p. , thommes , e. w. 2005 , submitted to apj , astro - ph/0502143 lin , d. n. c. , & tremaine , s. 1980 , apj , 264 , 364 morris , m. 1993 , apj , 408 , 496 munyaneza , f. , tsiklauri , d. , viollier , r. d. 1999 , apj , 526 , 744 phinney , e. s. 1989 , in iau symp .
136 , the center of the galaxy , ed .
m. morris ( dordrecht : kluwer ) , 543 perryman , m. a. c. , et .
2001 , a&a , 369 , 339 portegies zwart , s. f. , et .
2004 , nature , 428 , 724 sanders , r. h. 1992 , nature , 359 , 131 shodel , r. , et .
2005 , in preparation young , p. 1980
, apj , 242 , 1232 yu , q. , & tremaine , s. 2003 , apj , 599 , 1129 | the presence of young stars in the immediate vicinity and strong tidal field of sgra * remains unexplained .
one currently popular idea for their origin posits that the stars were bused in by an intermediate - mass black hole ( imbh ) which has inspiraled into the galactic center a few million years ago .
yu and tremaine ( 2003 ) have argued that in this case some of the old stars in the sgra * cusp would be ejected by hard gravitational collisions with the imbh .
here we derive a general expression for the phase - space distribution of the ejected high - velocity stars , given the distribution function of the stars in the cusp .
we compute it explicitly for the peebles - young distribution function of the cusp , and make a detailed model for the time - dependent ejection of stars during the imbh inspiral .
we find that ( 1 ) the stars are ejected in a burst lasting a few dynamical friction timescales ; if the ejected stars are detected by gaia they are likely to be produced by a single inspiral event , ( 2 ) if the inspiral is circular than in the beginning of the burst the velocity vectors of the ejected stars cluster around the inspiral plane , but rapidly isotropise as the burst proceeds , ( 3 ) if the inspiral is eccentric , then the stars are ejected in a broad jet roughly perpendicular to the runge - lenz vector of the imbh orbit . in a typical cusp
the orbit will precess with a period of @xmath0 years , and the rate of ejection into our part of the galaxy ( as defined by e.g. the gaia visibility domain ) will be modulated periodically .
gaia , together with the ground - based follow - up observations , will be able to clock many high - velocity stars back to their ejection from the galactic center , thus measuring some of the above phenomena .
this would provide a clear signature of the imbh inspiral in the past 1020 myr . |
it took more than four decades since the publication of the maxwell equations to establish their conformal invariance .
thereafter the conformal symmetry became a subject of continuous interest in theoretical physics and the realm of its applications vast.@xcite the conformal algebra in two - dimensional spacetime has the special property of having an infinite number of generators .
this makes the symmetry especially powerful and makes conformal field theories in two - dimensional spacetime appealing .
much work on them has been done in recent years ( see e.g. , and references therein ) .
the purpose of this report is to discuss a novel classical conformal field theory , on an arbitrary two - dimensional riemannian spacetime background , in which the dynamical object is a space - filling string .
if one considers an arbitrary but fixed system of coordinates @xmath0 on the background , a generic ( parametrized ) space - filling string is described by a function @xmath1 , and an infinitesimal deformation is the change , @xmath2 in two spacetime dimensions , and only then , that deformation may be described in a conformally invariant manner as , @xmath3 provided the normal @xmath4 is chosen to be weyl invariant , instead of being the customary unit normal .
then the deformation parameters @xmath5 , @xmath6 describe a conformal transformation , because the general integrability conditions of surface deformations@xcite imply that the corresponding generators @xmath7 , @xmath8 , obey the algebra of the conformal group,@xcite @xmath9 & = \left(s_{1}\left(x\right)+s_{1}\left(x'\right)\right)\delta'\left(x , x'\right),\label{eq : hphp-1 - 1}\\ \left[s_{1}\left(x\right),s_{\perp}\left(x'\right)\right ] & = \left(s_{\perp}\left(x\right)+s_{\perp}\left(x'\right)\right)\delta'\left(x , x'\right)-\zeta\delta'''\left(x , x'\right),\label{eq : hphi-1 - 1}\\ \left[s_{1}\left(x\right),s_{1}\left(x'\right)\right ] & = \left(s_{1}\left(x\right)+s_{1}\left(x'\right)\right)\delta'\left(x , x'\right),\label{eq : hihi-1 - 1}\end{aligned}\ ] ] where the central charge @xmath10 is a constant with the units of an action.)([eq : hihi-1 - 1 ] ) can be decomposed as two copies of the virasoro algebra , @xmath11=\left(l\left(x\right)+l\left(x'\right)\right)\delta'\left(x , x'\right)-\frac{\zeta}{2}\delta'''\left(x , x'\right)$ ] , whose generators are given by @xmath12 .
with the standard convention the virasoro central charge @xmath13 is given by @xmath14 . ]
therefore if one fixes a reference cut @xmath15 there is a one - to - one correspondence between the functions @xmath1 and the conformal transformation whose effect is to deform @xmath15 into @xmath1 .
, then ( [ eq : yl ] ) is equivalent to a change of the coordinates @xmath16 . therefore , a quantity , for example an action integral , will be invariant under changes in the spacetime coordinates if and only if it is conformally invariant . for this reason we will use both terminologies interchangeably in the text , depending on the desired emphasis . ]
we will discuss herein a realization of the algebra ( [ eq : hphp-1 - 1])([eq : hihi-1 - 1 ] ) on a two - dimensional spacetime background , endowed with an arbitrary but given metric @xmath17 , in which _ the conformal symmetry generators are built out of the @xmath1 themselves and their canonical conjugates_. in virtue of the fact that the deformation is described in a conformally invariant manner , one may say that the configuration space of the theory is the group space of the conformal group .
the action and symmetry generators of the theory in question were found in ref .
, as those of a particular and especial``g - brane . '' the purpose of the present discussion is to bring out in a self - contained manner the ( 1 + 1 ) g - brane as a classical conformal field theory , focusing on aspects that were not covered or emphasized in ref . .
the plan of the paper is the following : sec .
[ sec : surface - deformation - algebra ] recalls the results obtained in ref . for the conformal generators of the ( 1 + 1 ) g - brane , while sec .
[ sec : action - principle ] discusses the action principle , the ( lack of ) conformal and weyl invariance of the action , and the equations of motion .
[ sec : stress - tensor .- coupling ] is devoted to analyzing the energy - momentum tensor , the coupling to liouville matter , and the cancellation of anomalies .
finally sec .
[ sec : concluding - remarks ] contains concluding remarks .
to relate the infinitesimal deformation ( [ eq : yl ] ) to the conformal algebra , it is essential to bring in the metric @xmath18 in order to split @xmath19 into its normal and tangential components , @xmath5 and @xmath6 according to eq .
( [ eq : deltay ] ) .
the normal @xmath4 appearing in that equation is defined by , @xmath20 where , @xmath21 is the metric on the cut @xmath1 . on account of ( [ eq : normal ] ) and ( [ eq : indmetric ] ) the normal @xmath4 is invariant under weyl transformations , @xmath22 the fields @xmath0 , @xmath23 , and @xmath24 are also weyl invariant because the background metric @xmath18 does not enter into their definition .
we now recall the expressions for the generators @xmath7 and @xmath8 obtained in ref . .
if the canonical momenta conjugate to @xmath0 are denoted by @xmath25 , so that in terms of poisson brackets , @xmath26=\delta_{\rho}^{\lambda}\delta\left(x , x'\right),\ ] ] the canonical generators that span the algebra ( [ eq : hphp-1 - 1])([eq : hihi-1 - 1 ] ) are realized in terms of the fields @xmath27 as , @xmath28 here , @xmath29 and , @xmath30 the fields @xmath31 and @xmath32 are the local relative scale and extrinsic curvature of the space - filling string respectively .
they are expressed in terms of the @xmath0as , @xmath33 @xmath34 the derivative @xmath35 is the covariant derivative in the external space projected onto the string .
if in ( [ eq : l ] ) the field @xmath31 is regarded as a fundamental variable , and @xmath32 as its `` invariant velocity '' ; as opposed to being expressed in terms of the @xmath0 through ( [ eq : phi ] ) and ( [ eq : k ] ) , then @xmath36 appearing in ( [ eq : l ] ) becomes the lagrangian for an analog in two - dimensional spacetime of einstein s theory of gravitation , , , which is related to the liouville theory .
if the replacements ( [ eq : phi ] ) and ( [ eq : k ] ) are implemented the `` gravitational ancestor '' ( [ eq : l ] ) gives rise to its descendant , the space - filling string .
the action integral is given by , @xmath37 the weyl invariant lapse and shift functions @xmath38 are considered as external fields which are not varied in the action principle .
they parametrize the most general conformal transformation whose unfolding in time is the dynamics of the system , generated by the hamiltonian , @xmath39 a particular case is , @xmath40 , @xmath41 , which might be termed a `` rigid time translation '' , the corresponding hamiltonian is the integral of @xmath7 .
the hamiltonian is linear in the momenta because it is a linear combination of the generators @xmath42 @xmath8 , which have that property .
this is a distinctive feature of the ( 1 + 1 ) g - brane .
geometrically , it captures the fact that the motion may be considered as a displacement on the manifold of the conformal group ( see eq .
( [ eq : eom1 ] ) below ) .
dynamically it implies that , independently of the sign of the constant @xmath10 , the `` energy density '' @xmath7 can be arbitrarily large and positive or negative . as a consequence of the linearity in the momenta they can not be expressed in terms of the velocities to obtain an action principle in terms of a lagrangian which would depend on @xmath43 and @xmath44 .
this situation changes however when one couples the ( 1 + 1 ) g - brane to matter as discussed in section [ lag ]
. it will be useful here to employ a standard compact notation to account for both continuous and discrete indices , so that , for example summation over @xmath45 includes an integration over an accompanying continuous index @xmath46 .
the conformal algebra ( [ eq : hphp-1 - 1])([eq : hihi-1 - 1 ] ) then reads , @xmath47=\kappa_{ab'}^{\;\;\ ; c''}s_{c''}+\zeta_{ab'},\ ] ] where the structure constants @xmath48 are products of @xmath49-functions and the central term @xmath10 is proportional to @xmath50 .
the conformal transformation of any functional @xmath51 of the dynamical fields @xmath0 and @xmath25 is given by its poisson bracket with the corresponding generator : @xmath52,\label{eq : deltaf}\ ] ] whereas the change of the `` conformal lapse '' and shift @xmath53 is given by , @xmath54 if one employs equations ( [ eq : deltaf ] ) and ( [ eq : deltaeta ] ) to evaluate the change in the action ( [ eq : action ] ) for a region @xmath55 with boundary @xmath56 , one finds @xmath57 therefore , due to the central term , even when the boundary is mapped onto itself ( @xmath58 over @xmath56 ) , the action is not conformally invariant .
this anomaly does not affect the equations of motion because the central term does not include the dynamical fields . under an infinitesimal weyl transformations of the background metric ,
@xmath59 the action ( [ eq : action ] ) transform as : @xmath60 the change ( [ eq : deltaweyl ] ) may be obtained by direct calculation , or , better , by realizing that it must be the same as the one of its gravitational ancestor mentioned at the end of the previous section . in that context eq .
( [ eq : deltaweyl ] ) is at the heart of the gravitation theory in two spacetime dimensions : demanding that it should vanish gives then the equation of motion , @xmath61 of the ancestor theory.@xcite for general @xmath62 and @xmath63 the equations of motion are given by , @xmath64 these equations are conformally invariant because , as it was stated at the end of subsection [ sub : change - of - the ] the change in the action under a conformal transformation does not depend on the dynamical fields .
comparison of equation ( [ eq : eom1 ] ) with equation ( [ eq : deltay ] ) shows that the motion may be considered as a displacement on the manifold of the conformal group . to obtain insight
we now consider the case in which the background is minkowski space and the external coordinate system is cartesian , that is @xmath65 furthermore we take at time @xmath66 @xmath67 while leaving the @xmath68 arbitrary , and further specialize to a `` rigid time translation '' , i.e. , @xmath69 for all times .
then the solution of the equations of motion is : @xmath70 it is important to emphasize that the above solution _ is not _ the most general solution of the equations of motion when @xmath71 because the initial conditions ( [ eq : initial ] ) are very particular corresponding to a simple cut through the spacetime .
the most general solution is obtained by acting on ( [ eq : solution ] ) with the conformal group through an iteration of transformations generated by @xmath7 and @xmath8 .
the result of this iteration is to go from the cut ( [ eq : initial ] ) to a generic spacelike cut @xmath72 , but the corresponding result for @xmath25 can not written in a closed form .
it is important to keep in mind that different cuts @xmath72 do not correspond to different gauge choices , because @xmath7 and @xmath8 do not generate a gauge symmetry since they are not constrained to vanish .
note that for the solution ( [ eq : solution ] ) the conformal energy density @xmath7 does not vanish when @xmath73 , but is given by the background energy , @xmath74
one may define an energy - momentum tensor for the ( 1 + 1 ) g - brane through the standard formula , @xmath75 as a consequence of the lack of invariance of the action under changes of the spacetime coordinates expressed by ( [ eq : lack ] ) , @xmath76 defined by ( [ eq : t ] ) not only is not conserved when the equations of motion ( [ eq : eom1])([eq : eom3 ] ) hold but its components do not even transform as those of a tensor under a change of the spacetime coordinates .
this conformal anomaly is brought in by the presence of the central charge ( `` schwinger term '' ) in the commutation rule for the energy and momentum densities @xmath7 , @xmath8 , which are related to @xmath76 by , @xmath77 the remaining component , @xmath78 where , @xmath79 may be obtained from ( [ eq : deltaweyl ] ) to be , @xmath80 and it is a world scalar although @xmath76 is not a tensor .
one may couple the ( 1 + 1 ) g - brane by adding to ( [ eq : action ] ) the action of a `` matter field '' also described by a conformal field theory . that matter action will have a form similar to ( [ eq : action ] ) : @xmath81 the generators @xmath82 will obey the conformal algebra ( [ eq : hphp-1 - 1])([eq : hihi-1 - 1 ] ) . since they will be built out of @xmath83 and @xmath84 they will commute with those of the ( 1 + 1 ) g - brane . therefore the generators of the complete theory , @xmath85 will obey the conformal algebra with a total central charge which is the sum of the central charges of the ( 1 + 1 ) g - brane and that of the matter theory .
there is one theory for a matter field which has a central charge at the classical level , that is the liouville theory for which , @xmath86 obey the algebra ( [ eq : hphp-1 - 1])([eq : hihi-1 - 1 ] ) with , @xmath87 the total central charge is then @xmath88 an interesting possibility now appears , namely that of adjusting the ( 1 + 1 ) g - brane central charge @xmath10 for a given @xmath89 so that the sum ( [ eq : sum ] ) vanishes , @xmath90 for physical reasons , such as the positivity of the energy of the liouville field , it is desirable to have @xmath91 , however no similar restriction appears for @xmath10 because , as it was pointed out above , for any sign of @xmath10 the ( 1 + 1 ) g - brane energy is unbounded from above and from below . when ( [ eq : zetazero ] ) holds , and only then , one has the option of constraining the total generators to vanish ,
@xmath92 if the total central charge were not to vanish equations ( [ eq : sptot ] ) and ( [ eq : s1tot ] ) would not be preserved under the action of @xmath93 and @xmath94 . before the constraints are imposed
the complete theory is invariant under two independent conformal symmetries , one acting on the ( 1 + 1 ) g - brane , the other on the liouville field .
there are independent parameters @xmath95 for each , which are not varied in the action principle .
the two theories are decoupled and there are three degrees of freedom per space point . if the constraints are imposed the only symmetry is a simultaneous conformal transformation of the same magnitude on both the brane and the liouville field .
the common parameter @xmath95 is now varied in the action principle in order to yield the constraint .
the number of degrees of freedom per point of the combined theory then decreases from three to one , and the brane and the liouville field become coupled through the constraint , although their equations of motion remain uncoupled . when the total central charge vanishes , according to eq .
( [ eq : zetazero ] ) the complete action is invariant under spacetime reparametrizations because the anomalous term in ( [ eq : lack ] ) is proportional to @xmath96 . on the other hand
the change of the complete action under a weyl transformation is still given by ( [ eq : deltaweyl ] ) because the liouville action is weyl invariant if one assumes @xmath84 to have that property .
for the coupled action one may pass from the hamiltonian formulation to a lagrangian formulation by eliminating the constraints and the momenta from their equations of motion .
the resulting lagrangian density which is a function of @xmath0 , @xmath97 , @xmath84 , @xmath98 and @xmath99 has the form , @xmath100 where , @xmath101 with @xmath36 given by ( [ eq : l ] ) , and @xmath102^{2}-\eta^{\perp}\psi'^{2}+4\eta^{\perp}\psi''+2\eta^{\perp}\lambda e^{\psi}\right\}.\ ] ] ( see ref . ) . here
@xmath62 and @xmath63 stand as abbreviations for,@xmath103 and
the constant @xmath89 is related to the ( 1 + 1 ) g - brane central charge @xmath10 through ( [ eq : sum ] ) and ( [ eq : zetazero ] ) . since the conformal anomalies of the ( 1 + 1 ) g - brane and the liouville field mutually cancel the action,@xmath104 is invariant under spacetime reparametrizations , therefore the total energy - momentum tensor , @xmath105 is indeed a tensor and it is conserved .
the energy and momentum densities ( of weight two ) are given by @xmath106 whereas the trace is still given by ( [ tracestress ] ) .
one sees that the energy of the system acquires a `` background contribution '' @xmath107 in addition to the liouville energy density .
furthemore it is only the sum of the two which has covariant meaning .
if one changes the spatial coordinate on the surface @xmath108 , the two pieces do not transform separately as a density because of the conformal anomaly , but the sum does . in the above discussion the liouville field @xmath84 is unrelated to the field @xmath31 appearing in the ancestor gravitation theory referred to after eq .
( [ eq : k ] ) . if it where identified with it , so as to couple the descendant to its ancestor , the variation of the combined action with respect to @xmath84 would be , @xmath109 so that the action principle would either be contradictory , if @xmath110 , or empty , if @xmath111 .
incest is either forbidden or it does not happen .
we have presented a conformal field theory on an arbitrary two - dimensional spacetime background .
the dynamical object is a space - filling string , and the evolution may be thought as occurring on the group space of the conformal group .
the theory is a `` descendant '' in the sense of ref .
, of an analog of the theory of gravitation proposed in refs .
, for two dimensional spacetime .
the discussion has remained classical throughout , including the treatment of the central charge which already appears at that level
. one would expect many delicate issues to appear in the passage to the quantum theory , not least in the discussion of anomalies .
the centro de estudios cientficos ( cecs ) is funded by the chilean government through the centers of excellence base financing program of conicyt .
c.b . wishes to thank the alexander von humboldt foundation for a humboldt research award .
the work of a.p .
is partially funded by the fondecyt grant n 11130262 .
e. s. fradkin and m. y. palchik , `` new developments in d - dimensional conformal quantum field theory , '' _ phys .
_ * 300 * , 1 ( 1998 ) .
doi:10.1016/s0370 - 1573(97)00085 - 9 p. di francesco , p. mathieu and d. senechal , `` conformal field theory , '' doi:10.1007/978 - 1 - 4612 - 2256 - 9 c. teitelboim , `` the hamiltonian structure of two - dimensional space - time and its relation with the conformal anomaly , '' in * christensen , s.m .
( ed . ) : quantum theory of gravity * , adam hilger ltd , bristol , 327 - 344 . c. teitelboim ,
`` gravitation and hamiltonian structure in two space - time dimensions , '' _ phys .
lett . _ * b126 * , 41 ( 1983 ) .
doi:10.1016/0370 - 2693(83)90012 - 6 c. bunster and a. perez , `` space - filling branes of gravitational ancestry , '' to be published in _ phys .
, arxiv:1508.03340 [ hep - th ] .
r. jackiw , `` liouville field theory : a two - dimensional model for gravity ?
, '' in * christensen , s.m .
( ed . ) : quantum theory of gravity * , adam hilger ltd , bristol , 403 - 420 . | we present a classical conformal field theory on an arbitrary two - dimensional spacetime background .
the dynamical object is a space - filling string , and the evolution may be thought as occurring on the manifold of the conformal group .
the theory is a `` descendant '' of the theory of gravitation in two - dimensional spacetime .
the discussion is based on the relation of the deformations of the space - filling string with conformal transformations .
the realization of the conformal algebra in terms of surface deformations possesses a classical central charge .
the action principle , the conformal and weyl invariances of the action , and the equations of motion are studied .
the energy - momentum tensor , the coupling to liouville matter , and the cancellation of anomalies are analyzed .
the quantum theory is not discussed . |
we consider a multiple testing scenario encountered in many current applications of statistics .
given a large index set @xmath0 and a family @xmath1 of null hypotheses about the distribution of a high - dimensional random vector @xmath2 , we wish to design a procedure , basically a family of test statistics and thresholds , to estimate the subset @xmath3 over which the null hypotheses are false .
we shall refer to @xmath4 as the `` active set '' and write @xmath5 for our estimator of @xmath4 based on a random sample @xmath6 of size @xmath7 from @xmath8 .
the hypotheses in @xmath9 ( namely the ones for which the null is rejected ) are referred to as `` detections '' or `` discoveries . ''
naturally , the goal is to maximize the number @xmath10 of detected true positives while simultaneously controlling the number @xmath11 of false discoveries . there are two widely used criteria for controlling false positives : + * fwer : * assume that @xmath12 is defined on the probability space @xmath13 . the family - wise error rate ( fwer ) is @xmath14 which is the probability of making at least one false discovery .
this is usually controlled using bonferroni bounds and their refinements @xcite , or using resampling methods or random permutation . + * fdr : * the false discovery rate ( fdr ) is the expected ratio between the number of false alarms @xmath11 and the number of discoveries @xmath15 @xcite .
+ in many cases , including the settings in computational biology which directly motivate this work , we find @xmath16 , @xmath17 as well as small `` effect sizes . ''
this is the case , for example , in genome - wide association studies ( gwas ) where @xmath18 and the dependence of the `` phenotype '' @xmath19 on the `` genotype '' @xmath20 is often assumed to be linear ; the active set @xmath4 are those @xmath21 with non - zero coefficients and effect size refers to the fraction of the total variance of @xmath19 explained by a particular @xmath22 . under these challenging circumstances , the fwer criterion is usually very conservative and power is limited ; that is , number of true positive detections is often very small ( if not null ) compared to @xmath23 ( the `` missing heritability '' ) .
this is why the less conservative fdr criterion is sometimes preferred : it allows for a higher number of true detections , but of course at the expense of false positives . however , there are situations , such as gwas , in which this tradeoff is unacceptable ; for example , collecting more data and doing follow - up experiments may be too labor intensive or expensive , and therefore having even one false discovery may be deemed undesirable . to set the stage for our proposal , suppose we are given a family @xmath24 of test statistics and can assume that deviations from the null are captured by small values of @xmath25 ( e.g. , p - values ) .
we make the usual assumption , easily achieved in practice , that the distribution of @xmath26 does not depend on @xmath21 when @xmath27 , and individual rejection regions are of the form @xmath28 for a constant @xmath29 independent of @xmath21 . defining @xmath30 , the bonferroni upper - bound is @xmath31 to ensure that @xmath32 , @xmath33 is selected such that @xmath34 whenever @xmath35 .
the bonferroni bound can only be marginally improved ( see , in particular estimator @xcite , which will be referred to as bonferroni - holm in the rest of the paper ) in the general case . while alternative procedures ( including permutation tests ) can be designed to take advantage of correlations among tests ,
the bound is sharp when @xmath36 and tests are independent .
* coarse - to - fine testing : * clearly some additional assumptions or domain - specific knowledge is necessary to ameliorate the reduction in power resulting from controlling the fwer . motivated by applications in genomics , we suppose the set @xmath0 has a natural hierarchical structure . in principle , it should then be possible to gain power if the active hypotheses are not randomly distributed throughout @xmath0 but rather have a tendency to cluster within cells of the hierarchy .
in fact , we shall consider the simplest example consisting of only two levels corresponding to individual hypotheses indexed by @xmath37 and a partition of @xmath0 into non - overlapping subsets @xmath38 , which we call `` cells . ''
we will propose a particular multiple testing strategy which is coarse - to - fine with respect to this structure , controls the fwer , and whose power will exceed that of the standard bonferroni - holm approach for typical models and realistic parameters when a minimal degree of clustering is present .
it is important to note that clustering property is not a condition for a correct control of the fwer at a given level using our coarse - to - fine procedure , but only for its increased efficiency in discovering active hypotheses .
our estimate of @xmath4 is now based on two families of test statistics : @xmath39 , as above , and @xmath40 .
the cell - level test @xmath41 is designed to assume small values only when @xmath42 is `` active , '' meaning that @xmath43 .
our estimator of @xmath4 is now @xmath44 one theoretical challenge of this method is to derive a tractable method for controlling the fwer at a given level @xmath45 .
evidently , this method can only out - perform bonferroni if @xmath46 ; otherwise , the coarse - to - fine active set is a subset of the bonferroni discoveries .
a key parameter is @xmath47 , an upper bound on the number of active cells , and in the next section we will derive an fwer bound @xmath48 under an appropriate compound null hypothesis . the main results of the paper are in the ensuing analysis for different models for @xmath8 . in each case , the first objective is to compute @xmath49 for a given @xmath50 and @xmath51 and the second objective is to maximize the power over all pairs @xmath52 which satisfy @xmath53 .
the smaller our upper bound on @xmath47 , the stronger is the clustering of active hypotheses in cells and the greater is the gain in power compared with the bonferroni bound . in particular , as soon as @xmath54 , the coarse - to - fine strategy will lead to a considerably less conservative score threshold for individual hypotheses relative to the bonferroni estimate and the coarse - to - fine procedure will yield an increase in power for a given fwer .
again , our assumptions about clustering are only expressed through an upper bound on @xmath47 ; no other assumptions about the distribution of @xmath4 are made and the fwer is controlled in all cases .
the main technical difficulty arises from the correlation between the corresponding test statistics .
this must be taken into account since it increases the likelihood of an individual index @xmath21 being falsely declared active when the cell @xmath55 that contains it is falsely discovered ( survivorship bias ) .
more specifically , we require sharp estimates of quadrant probabilities under the _ joint distribution _ of @xmath56 and @xmath26 when @xmath55 , the cell containing @xmath21 , is inactive .
all these issues will be analyzed in two cases .
first , we will consider the standard linear model with gaussian data . in this case
@xmath49 is expressed in terms of centered chi - square distributions and the power is expressed in terms of non - centered chi - square distributions .
the efficiency of the coarse - to - fine method in detecting active hypotheses will depend on effect sizes , both at the level of cells and individual @xmath21 , among other factors .
a non - parametric procedure will then be developed in section [ sec:5 ] based on generalized permutation testing and invariance assumptions .
finally , we shall derive a high - confidence upper bound on @xmath47 based on a martingale argument .
extensive simulations comparing the power of the coarse - to - fine and bonferroni - holm appear throughout .
* applications and related work : * as indicated above , our work ( and some of our notation ) is inspired by statistical issues arising in gwas @xcite and related areas in computational genomics . in the most common version of gwas , the `` genotype '' of an individual
is represented by the genetic states @xmath22 at a very large family of genomic locations @xmath37 ; these variations are called single nucleotide polymorphisms or snps . in any given study
the objective is to find those snps @xmath57 `` associated '' with a given `` phenotype '' , for example a measurable trait @xmath19 such as height or blood pressure .
the null hypothesis for snp @xmath21 is that @xmath19 and @xmath22 are independent r.v.s , and whereas @xmath58 may run into the millions , the set @xmath4 of active variants is expected to be fewer than one hundred .
( ideally , one seeks the `` causal '' variants , an even smaller set , but separating correlation and causality is notoriously difficult . )
control of the fwer is the gold standard and the linear model is common .
if the considered variants are confined to coding regions , then the set of genes provides a natural partition of @xmath0 ( and the fact that genes are organized into pathways provides a natural three - level hierarchy ) @xcite another application of large - scale multiple testing is variable filtering in high - dimensional prediction : the objective is to predict a categorical or continuous variable @xmath19 based on a family of potentially discriminating features @xmath59 . learning a predictor @xmath60 from i.i.d .
samples of @xmath61 is often facilitated by limiting _ a priori _ the set of features utilized in training @xmath60 to a subset @xmath57 determined by testing the features one - by - one for dependence on @xmath19 and setting a signficance threshold . in most applications of machine learning to artificial perception , no premium is placed on pruning @xmath4 to a highly distinguished subset ; indeed , the particular set of selected features is rarely examined or considered of significance .
in contrast , the identities of the particular features selected and appearing in decision rules are often of keen interest in computational genomics , e.g. , discovering cancer biomarkers , where the variables @xmath22 represent `` omics '' data ( e.g. , gene expression ) , and @xmath19 codes for two possible cellular or disease phenotypes .
obtaining a `` signature '' @xmath62 devoid of false positives can be beneficial in understanding the underlying biology and interpreting the decision rules . in this case the gene ontology ( go ) @xcite provides a very rich hierarchical structure , but one example being the organization of genes in pathways .
indeed , building predictors to separate `` driver mutations '' from `` passenger mutations '' in cancer would appear to be a promising candidate for coarse - to - fine testing due to the fact that drivers are known to cluster in pathways .
there is a literature on coarse - to - fine pattern recognition ( see , e.g. , @xcite and the references therein ) , but the emphasis has traditionally been on computational efficiency rather than error control .
computation is not considered here . moreover , in most of this work , especially applications to vision and speech , the emphasis is on detecting true positives ( e.g. , patterns of interest such as faces ) at the expense of false positives .
simply `` reversing '' the role of true positives and negatives is not feasible due to the loss of reasonable invariance assumptions ; in effect , every pattern of interest is unique .
finally , in @xcite , a hierarchical testing approach is used in the context of the fwer .
however , the intention is to improve the power of detection relative to the bonferroni - holm methods only at level of clusters of hypotheses ; in contrast to our method , the two approaches have comparable power at the level of individual hypotheses .
* organization of the paper : * the paper is structured as follows : in section [ sec:2 ] we present a bonferroni - based inequality that will be central for controlling the fwer using the coarse - to - fine method in different models . in section [ sec:3 ] will consider a parametric model that will illustrate precisely the way we control the fwer at a fixed level and permit a power comparison between coarse - to - fine and bonferroni - holm .
we then propose a non - parametric procedure in section [ sec:5 ] under general invariance assumptions .
a method for estimating an upper bound on the number of active cells and incorporating it into the testing procedure without violating the fwer constraint is derived in section [ sec:6 ] .
finally , some concluding remarks are made in the discussion .
the finite family of null hypotheses will be denoted by @xmath63 , where @xmath64 is either true or false .
we are interested in the active set of indices , @xmath65 and will write @xmath66 for the set of inactive indices
. suppose our data @xmath6 takes values in @xmath67 .
the set @xmath9 is commonly designed based on individual rejection regions @xmath68 , with @xmath69 as indicated in the previous section , in the conservative bonferroni approach , the @xmath70 is controlled at level @xmath45 by assuming @xmath71 .
if the rejection regions are designed so that this probability is independent of @xmath21 whenever @xmath72 , then the condition boils down to @xmath73 for @xmath74 .
generally , @xmath75 for a constant @xmath76 for some family of test statistics @xmath77 .
while there is not much to do in the general case to improve on the bonferroni method , it is possible to improve power if @xmath0 is structured and one has prior knowledge about way the active hypotheses are organized relative to this structure . in this paper
, we consider a coarse - to - fine framework in which @xmath0 is provided with a partition @xmath78 , so that @xmath79 , where the subsets @xmath80 ( which we will call cells ) are non - overlapping . for @xmath81 , we let @xmath55 denote the unique cell @xmath42 that contains it .
the `` coarse '' step selects cells likely to contain active indices , followed by a `` fine '' step in which a bonferroni or equivalent procedure is applied only to hypotheses included in the selected cells .
more explicitly , we will associate a rejection region @xmath82 to each @xmath83 and consider the discovery set @xmath84 we will say that a cell @xmath42 is active if and only if @xmath85 , which we shall also express as @xmath86 , implicitly defining @xmath87 as the logical and " of all @xmath88 . we will also consider the double null hypothesis @xmath89 of @xmath21 belonging in an inactive cell ( which obviously implies that @xmath21 is inactive too ) , and we will let @xmath90 be the set of such @xmath21 s .
let @xmath91 denote the size of the largest cell in @xmath78 and @xmath47 be the number of active cells .
we will develop our procedure under the assumption that @xmath47 is known , or , at least bounded from above . while this can actually be a plausible assumption in practice
, we will relax it in section [ sec:5 ] in which we will design a procedure to estimate a bound on @xmath47
. then under these assumptions we have the following result : [ prop : first.bound ] with @xmath92 defined by : @xmath93 this is just the bonferroni bound applied to the decomposition @xmath94 so that @xmath95 and the proposition results from @xmath96 and @xmath97 .
the sets @xmath82 and @xmath98 will be designed using statistics @xmath99 and @xmath26 setting @xmath100 $ ] and @xmath101 $ ] for some constants @xmath50 and @xmath51 , and assuming that the distribution of @xmath102 ( resp .
@xmath26 ) is independent of @xmath21 for @xmath103 ( resp .
@xmath104 ) . letting @xmath105 for @xmath106 and @xmath107 for @xmath108 , the previous upper bound becomes @xmath109 in the following sections our goal will be to design @xmath50 and @xmath51 such that this upper bound is smaller than a predetermined level @xmath45 .
controlling the second term will lead to less conservative choices of the constant @xmath51 ( compared to the bonferroni estimate ) , as soon as @xmath110 ( or @xmath54 if all cells have comparable sizes ) .
depending on the degree of clustering , the probability @xmath111 of false detection in the two - step procedure can be made much smaller than @xmath112 without harming the true detection rate and the coarse - to - fine procedure will yield an increase in power for a given fwer .
we require tight estimates of @xmath111 and taking into account the correlation between @xmath56 and @xmath26 is necessary to deal with `` survivorship bias . ''
in this section , the observation is a realization of an i.i.d .
family of random variables @xmath113 where the @xmath114 s are real - valued and the variables @xmath115 is a high - dimensional family of variables indexed by the set @xmath0 .
we assume that the distribution of @xmath116 , are independent and centered gaussian , with variance @xmath117 , and that @xmath118 where @xmath119 are i.i.d .
gaussian with variance @xmath120 and @xmath121 , are unknown real coefficients
. we will denote by @xmath122 the vector @xmath123 and by @xmath124 where @xmath125 is the vector composed by ones repeated @xmath7 times .
we also let @xmath126 and @xmath127 , so that @xmath128 finally , we will denote by @xmath129 the common variance of @xmath130 and assume that it is known ( or estimated from the observed data ) . for @xmath81 ,
we denote by @xmath131 the orthogonal projection on the subspace @xmath132 spanned by the two vectors @xmath133 and @xmath125 .
we will also denote by @xmath134 ( @xmath83 ) the orthogonal projection on the subspace @xmath135 spanned by the vectors @xmath133 , @xmath136 , and @xmath125 .
the scores at the @xmath42 level and @xmath21 level will be respectively : @xmath137 and @xmath138 ( the projections are simply obtained by least - square regression of @xmath122 on @xmath139 , for @xmath134 and on @xmath133 for @xmath131 . ) we now provide estimates of @xmath140 for @xmath106 and @xmath141 and @xmath142 for @xmath104 .
note that , because we consider residual sums of squares , we here use large values of the scores in the rejection regions ( instead of small values in the introduction and other parts of the paper ) , hopefully without risk of confusion .
[ prop:1 ] for all @xmath50 and @xmath51 : @xmath143 where @xmath144 is the cdf of a @xmath145 distribution evaluated at @xmath146 and : @xmath147 moreover @xmath148 where @xmath149 is the c.d.f . of a chi - squared distribution with @xmath150 degrees of freedom . for @xmath106 and @xmath151 , we can write @xmath152 because @xmath153 and @xmath154 .
consider the conditional probability : @xmath155 the conditional distribution of @xmath122 given @xmath156 is gaussian @xmath157 ( where @xmath158 is the @xmath7-dimensional identity matrix ) .
denote by @xmath159 the projection on the orthogonal complement of @xmath160 in @xmath132 and by @xmath161 the projection on the orthogonal complement of @xmath162 in @xmath135 , so that @xmath163 and @xmath164 this implies that : @xmath165 at this stage , applying cochran s theorem to @xmath166 and @xmath167 , which are conditionally independent given @xmath168 , reduces the problem to finding an upper bound for : @xmath169 where @xmath170 is @xmath171 and @xmath172 is @xmath173 , and the two variables are independent .
let us write this probability as @xmath174 which is less than : @xmath175 ( here , @xmath176 refers to the expectation with respect to @xmath177 . ) consider the first term in the sum : @xmath178
. this term can be re - written as : @xmath179 at this stage , we will use the following tail inequality for @xmath180 random variables : @xmath181 for any @xmath182 .
we apply this result to @xmath183 and @xmath184 to get the upper bound : @xmath185 since the density of a @xmath173 is proportional to @xmath186 , the term in @xmath187 will cancel in the last integral ( expectation ) . using a simple change of variables in the remaining integral , we have as a final upper bound : @xmath188 where @xmath144 is the cdf of a beta(a , b ) evaluated at @xmath146 .
+ the second upper - bound , for @xmath189 , is easily obtained , the proof being left to the reader .
this leads us immediately to the following corollary : [ cor:2 ] with the thresholds @xmath50 and @xmath51 , an upper bound of the fwer is : @xmath190 figure [ fig : level.curves ] provides an illustration of the level curves associated to the above fwer upper bound .
more precisely , it illustrates the tradeoff between the conservativeness at the cell level and the individual index level . in the next section , the optimization for power will be made along these level lines .
figure 1 also provides the value of the bonferroni - holm threshold . for the coarse - to - fine procedure to be less conservative than the bonferroni - holm approach , we need the index - level threshold to be smaller , i.e. , the optimal point on the level line to be chosen below the corresponding dashed line .
level curves of the upper bound of the fwer for the levels 0.2 ( blue ) , 0.1 ( green ) and 0.05 ( red ) .
the horizontal dashed lines represent the thresholds at the individual level for a bonferroni - holm test , with corresponding colors . ] the derivation of is based on the assumption that we have a fixed cell size ( across all the cells ) , which is not needed . in the case where the size of the cell is varying , it is easy to generalize the previous upper bound . letting @xmath191 it suffices to replace @xmath192 in with @xmath193 where @xmath50 does not depend on the cell @xmath42 .
equation provides a constraint on the pair @xmath52 to control the fwer at a given level @xmath194 .
we now show how to obtain `` optimal '' thresholds @xmath195 that maximize discovery subject to this constraint .
the discussion will also help understanding how active indices clustering in cells improve the power of the coarse - to - fine procedure .
the conditional distribution of @xmath122 given @xmath196 is @xmath197 with @xmath153 .
it follows from this that , conditionally to these variables , @xmath198 follows a non - central chi - square distribution @xmath199 , with @xmath200 where @xmath201 .
using the fact that @xmath202 converges to @xmath203 we will work with the approximation @xmath204 with a similar analysis , and letting for @xmath205 , @xmath206 , we will assume that @xmath207 with @xmath208 we now have the simple lemma [ prop:3 ] if @xmath209 and @xmath210 , then for all @xmath211 such that @xmath212 , i=1,2 @xmath213 this is based on the inequality @xcite , valid for @xmath214 : @xmath215 which implies @xmath216 as soon as @xmath217 , and on the simple lower - bound @xmath218 this proposition can be applied , in our case , to @xmath219 and @xmath220 .
more concretely , we fix a target effect size @xmath170 ( the ratio of the effect of @xmath133 compared to the total variance of @xmath122 ) , and a target cluster size , @xmath150 , that represents the number of active loci that we expect to find in an active cell , and we take @xmath221 and @xmath222 to optimize the upper - bound in subject to the fwer constraint and @xmath223 and @xmath224 to find optimal constants @xmath52 for this target case .
this is illustrated with numerical simulations in the next section .
figures [ fig : digraph.1 ] compares the powers of the coarse - to - fine procedure and of the bonferroni - holm procedure under the parametric model described in the section .
comparison of the power of different methods .
we compare the detection rate of an individual active index for different number of active indices in the cell containing that index .
the coarse - to - fine method is more powerful when the number of active indices is two or greater .
this confirms the intuition that the more the clustering assumption is true , the more powerful is the coarse - to - fine method compared to bonferoni - holm approcach . ]
the parameters chosen in our simulations were taken with our motivating application ( to gwas ) in mind .
thinking of @xmath19 as a phenotype , and @xmath0 as a set of snp s , we assimilate cells @xmath83 to genes .
we used @xmath225 and @xmath226 .
the true number of active variables is 50 with a corresponding coefficient @xmath227 for each of them , and we generate the data according to the linear model described in this section with a variance noise that is equal to 10 .
we assumed that we knew an upper bound @xmath47 for the number of active sets ( this assumption is relaxed in section [ sec:6 ] ) . to compute the optimal thresholds , some values for @xmath228 and @xmath229 have to be chosen ( this should not be based on observed data , since this would invalidate our fwer and power estimates ) . in our experiments , we optimize the upper bound on the probability for an active variable to be detected in an active cell by choosing @xmath230 and @xmath231 .
this corresponds to an `` almost non noisy case '' where the effect size of the `` gene '' is two times the effect size of the `` snp '' .
recall that @xmath12 denotes the random variable representing all the data , taking values in @xmath67 .
we will build our procedure from user - defined _ scores _ , denoted @xmath229 ( at the locus level ) and @xmath228 ( at the cell level ) , both defined on @xmath67 , i.e. , functions of the observed data .
moreover , we assume that there exists a group action of some group @xmath232 on @xmath67 , which will be denoted @xmath233 for example , if , like in the previous paragraph , one takes @xmath234 and @xmath235 , we will take @xmath232 to be the permutation group of @xmath236 with @xmath237 to simplify the discussion , we will assume that @xmath238 is finite and denote by @xmath239 the uniform probability measure on @xmath238 , so that @xmath240 we note , however , that our discussion remains true if @xmath238 is a compact group , @xmath239 the right - invariant haar probability measure on @xmath238 and @xmath241 are continuous in @xmath242 .
our running assumption will be that , 1 . for any @xmath106 , the joint distribution of @xmath243 is independent of @xmath244 .
2 . for any @xmath245 , the joint distribution of @xmath246 is independent of @xmath244 .
we will also use the following well - known result . [
lem : basic ] let @xmath247 be a random variable and let @xmath248 denote the left limit of its cumulative distribution function , i.e. , @xmath249 .
then , for @xmath250 $ ] , one has @xmath251 ( with equality if @xmath252 is continuous ) .
we define the asymptotic scores at the cell and variable level by @xmath253 and @xmath254 @xmath255 and @xmath256 are the typical statistics used in randomized tests , estimating the proportion of scores that are higher than the observed one after randomization . for the coarse - to - fine procedure , we will need one more `` conditional '' statistic . for a given constant @xmath50 , we define @xmath257 we then let @xmath258 we call our scores asymptotic in this section because exact expectations over @xmath239 can not be computed in general , and can only be obtained as limits of monte - carlo samples .
the practical finite - sample case will be handled in the next section .
we this notation , we let @xmath259 which depends on the choice of three constants , @xmath260 and @xmath261 .
we then have : [ th : ideal ] for all @xmath104 : @xmath262 and for all @xmath106 , @xmath263 this result tells us how to control the fwer for a two - level permutation test based on any scores in the ( generally intractable ) case in which we can exactly compute the test statistics , when we declare an index @xmath21 active if and only if @xmath264 and @xmath265 and @xmath266 . for
, we use a standard argument justifying randomization tests , that we provide here for completeness .
if @xmath104 , we have @xmath267 + from the invariance assumption , we have @xmath268 it now remains to remark that @xmath269 where @xmath172 is the random variable on @xmath238 defined by @xmath270 , so that , by lemma [ lem : basic ] , @xmath271 which proves .
+ let us now prove , assuming @xmath106 and letting @xmath141 .
we write @xmath272 and find an upper bound for the right - hand side of the inequality . using the invariance assumption , we have , for all @xmath273 , @xmath274 notice that , since @xmath239 is right - invariant , we have @xmath275 and @xmath276 let @xmath277 denote the probability @xmath239 conditional to the event @xmath278 ( @xmath279 being fixed )
. then @xmath280 where @xmath281 hence , lemma [ lem : basic ] implies that , for each @xmath279 : @xmath282 hence , @xmath283 applying lemma [ lem : basic ] to the random variable @xmath284 for the probability distribution @xmath239 , we immediately get @xmath285 so that @xmath286 as an immediate corollary , we have : [ cor : fwer.ideal ] @xmath287 [ rem:1 ] even though @xmath238 is finite , it is a huge set in typical applications , and while lemma only provides inequalities for discrete distributions , we can safely ignore the discontinuity in practice and work as if the distributions to which we applied this lemma were continuous .
doing so , it is easy to convince oneself , by inspecting the previous proof that our estimates become ( for @xmath106 , @xmath141 ) @xmath288 this implies that @xmath289 because we have also : @xmath290 this tells us that , conditional to @xmath291 , @xmath292 is uniform distributed on @xmath293 $ ] .
as mentioned above , this result does not have practical interest since it requires applying all possible permutations to the data . in practice ,
a random subset of permutations is picked instead , and we will develop the related theory in the next section ( using these inequalities as intermediary results in our proofs ) .
we now replace @xmath41 , @xmath294 and @xmath295 with monte - carlo estimates and describe how the upper bounds in theorem [ th : ideal ] need to be modified . for a positive integer @xmath296 , we let @xmath297 denote the @xmath296-fold product measure on @xmath238 , whose realizations are @xmath296 independent group elements @xmath298
. we will use the notation @xmath299 and @xmath300 to denote probability or expectation for the joint distribution of @xmath8 and @xmath242 ( i.e. , @xmath301 ) .
we will also denote by @xmath302 the empirical measure @xmath303 with this notation , we let @xmath304 @xmath305 and @xmath306 where @xmath307 we can now define @xmath308 and state : [ th : sample ] making the continuous approximation described in remark [ rem:1 ] , the following holds . for @xmath104 , @xmath309 and , for @xmath106 and @xmath141 , @xmath310 [ cor : sample ] the fwer for the randomized test is controlled by @xmath311 we start with which is simpler and standard .
let @xmath104 .
conditionally to @xmath312 , @xmath313 follows a binomial distribution @xmath314 ( with @xmath295 defined by equation ) , so that @xmath315 } } \binom{k}{j } \mathbb e\left(t_v({\mathbf u})^j(1-t_v({\mathbf u}))^{k - j}\right)\ ] ] where @xmath316}$ ] denotes the integral part .
the continuous approximation implies that @xmath256 is uniformly distributed , so that @xmath317 yielding @xmath318}}{k } \leq { \theta_v}'.\ ] ] + we now consider and take @xmath106 , @xmath141 .
we need to prove that : @xmath319 given @xmath312 , @xmath320 is the empirical mean of @xmath296 i.i.d bernoulli random variables ( @xmath321 ) with success probability @xmath255 .
hoeffding s inequality @xcite implies that @xmath322 so : @xmath323 we now fix @xmath324 in @xmath325 and consider @xmath326 since this probability does not depend on which @xmath324 is chosen , we will estimate it , without loss of generality , for @xmath327 , which will simplify the notation . for this , notice that @xmath328 and also that @xmath329 now we use hoeffding s inequality to obtain @xmath330 the same upper - bound applies to @xmath331 by taking the expectation with respect to @xmath332 , and one deduces from this that @xmath333 introduce @xmath334 on the event @xmath335
one has @xmath336 moreover , we have @xmath337 , by applying lemma [ lem : basic ] to the random variable @xmath338 under the distribution @xmath339 .
this implies that , on @xmath340 , we have @xmath341 which implies @xmath342 we now provide an estimate for the last term in .
we have @xmath343 now write @xmath344 given @xmath312 , the probability ( for @xmath239 ) of the event @xmath345 is @xmath346 and @xmath347 follows a binomial distribution @xmath348 .
we make the continuous approximation ( remark [ rem:1 ] ) to write @xmath349 , and to use the fact that @xmath350 is uniformly distributed on @xmath293 $ ] conditionally to @xmath351 , so that @xmath352 is uniformly distributed on @xmath353 $ ] given @xmath354 .
this implies @xmath355}}\binom{k } { i } p^{i}{(1-p)}^{k - i}\right)dp\\ & \leq \frac{1}{(k+1)({\theta_g}+{\varepsilon_g})}\sum_{i=0}^{{[k({\theta_g}+{\varepsilon_g}){\theta_v}]}}\left(\int_{0}^{1}\binom{k+1 } { i } p^{i}{(1-p)}^{k - i}dp\right)\\ & = \frac{{[k({\theta_g}+{\varepsilon_g}){\theta_v}]}+1}{(k+1)({\theta_g}+{\varepsilon_g})}\end{aligned}\ ] ] since each of the integrals is equal to 1 ( they are densities of beta distributions ) . from
, we find @xmath356 which , combined with equations and completes the proof of the theorem . for the simulations
, we generated data according to the parametric model described in section [ sec:3 ] , and compared the detection rate using the non - parametric approach to the one obtained with thresholds optimized as described in section [ sec:3 ] .
we fixed the ratio between the non - parametric thresholds @xmath50 and @xmath51 to coincide with the ratio between the type i errors at the cell and index levels in the parametric case , i.e. , @xmath357 for @xmath106 , where @xmath358 are the scores and thresholds designed for the parametric case .
fixing @xmath359 and the fwer uniquely determines the thresholds .
figure [ fig : digraph.0 ] provides a comparison between the coarse - to - fine approaches ( parametric and non - parametric ) and bonferroni - holm . in figures [ fig : digraph.4 ] and [ fig :
digraph.5 ] , the non - parametric coarse - to - fine method is compared with bonferroni - holm for two different index - level effect sizes . using the notation of section [ sec:3 ] ,
the cell - level effect size is defined by @xmath360 while @xmath361 at the index level .
the range of index - level effect sizes we consider in these simulation is similar to the one observed in typical genome - wide association studies . improves the performance of the bonferroni - holm method .
the coarse - to - fine method is performs better at levels that correspond to two or more active indices per cell . ]
we now focus on the issue of estimating the number of active cells , @xmath47 , from observed data , since this number intervenes in our fwer estimates .
we use a method inspired from @xcite for the estimation of false discovery rates , adapted to our context . our estimation will be made based on cell statistics @xmath362 under the following setting . * if @xmath363 ( @xmath42 is inactive ) , then @xmath41 is uniformly distributed . * if @xmath364 are inactive , then @xmath365 , @xmath366 are independent .
these assumptions will be justified in section [ sec : justify ] .
we will also assume that @xmath41 takes large values when @xmath42 is active , so that , for a suitable non conservative threshold @xmath367 , we have @xmath368 . to simplify the argument
, we will actually make the approximation that : * there exists @xmath369 such that @xmath370 if @xmath371 . for @xmath250 $ ] , we define @xmath372 .
let @xmath373 be the set of active cells , and @xmath374 .
note that @xmath375 .
then , for @xmath376 , @xmath377 we therefore have , for @xmath376 , @xmath378 where @xmath379 is a centered random variable . the following proposition states that the process @xmath380 for @xmath376 has the covariance structure of a brownian bridge .
+ [ prob : cov ] under assumptions a1 to a3 , we have , for @xmath381 , @xmath382 since @xmath383 , one has @xmath384 if @xmath385 , then @xmath386 , and for @xmath387 : @xmath388 finally , from , we get @xmath389 which concludes the proof .
we now make a gaussian approximation for large @xmath78 of the vector @xmath390 with @xmath391 , with @xmath392 .
[ prop : clt ] using the previous notation , @xmath393 when @xmath394 diverges to infinity , where @xmath395 is the covariance matrix with entries : @xmath396 our assumptions ensure that @xmath172 satisfies a central limit theorem conditionally to @xmath19 , with a limit , @xmath397 that is independent of the value of @xmath8 .
this implies that the limit is also unconditional .
we are now able to present our principal result which provides a high - probability upper bound for @xmath47 .
[ th : j ] let @xmath398 be a randomization variable , independent of @xmath362 . for @xmath399 and @xmath400 ,
define @xmath401 then @xmath402 where @xmath403 . as a consequence ,
given @xmath404 , let @xmath405 . then @xmath406 is such that @xmath407 ( we use a randomly sampled value of @xmath408 ) .
let us now prove this result .
we know from proposition [ prop : clt ] that the vector @xmath409 where @xmath410 then , since min is a continuous function on @xmath411 , we deduce that : @xmath412 but : @xmath413 the process @xmath414 , @xmath415 is a martingale and @xmath416 is a submartingale for all @xmath417 s . applying doob s inequality ,
then optimizing over @xmath418 s finally gives : @xmath419 it remains to prove that @xmath420 but : @xmath421 then : @xmath422 solving the quadratic inequality for @xmath423 , one finds that @xmath424 is equivalent to @xmath425 , which completes the proof .
the previous section provided us with an estimator @xmath426 in such that @xmath427 with probability larger than @xmath428 , which implies that @xmath429 with probability @xmath428 at least .
we previously chose constants @xmath50 and @xmath51 by optimizing the detection rate on a well - chosen alternative hypothesis subject to the upper - bound being less than a significance level @xmath45 .
this was done using a deterministic upper - bound of @xmath47 , but can not be directly applied with a data - based estimation of @xmath47 since this would yield data - dependent constant @xmath50 and @xmath51 , which can not be plugged into the definition of the set @xmath92 without invalidating our estimation of the fwer . in other terms , if , for a fixed number @xmath430 , one defines @xmath431 to be the discovery set obtained by optimizing @xmath50 and @xmath51 subject to @xmath432 , our previous results imply that @xmath433 for all @xmath434 , but not necessarily that @xmath435 .
a simple way to address this issue is to replace @xmath436 with @xmath437 because @xmath438 with probability at least @xmath428 , we have @xmath439 so that @xmath440 controls the fwer at level @xmath441 as intended .
we check that conditions a1 and a2 are satisfied for the two situations that we consider in this paper . in the example from section [ sec:3 ] , we can take ( using the same notation and introducing the c.d.f . of a chi - square distribution ) @xmath442 ( recall that @xmath134 is the orthogonal projection on the space generated by @xmath443 and @xmath444 .
we also let @xmath445 be the empirical variance of @xmath446 . )
note that the conditional distribution of @xmath41 given @xmath447 is always uniform over @xmath293 $ ] and therefore does not depend on @xmath448 , which proves that @xmath41 and @xmath446 are independent .
similarly , taking @xmath449 , @xmath365 and @xmath366 are conditionally independent given @xmath446 ( because @xmath450 and @xmath451 are independent ) .
but @xmath452 and @xmath453 being conditionally independent given @xmath446 and each of them independent of @xmath446 implies that the three variables are mutually independent . + the same argument can be applied to the non - parametric case , when ( now using notation from that section ) one assumes that scores are such that @xmath454 , and uses to simplify the discussion the statistic @xmath455 assuming , in addition , the following .
if we denote by @xmath456 the space where the random variable @xmath457 takes its values , there exists a group @xmath458 , a group isomorphism @xmath459 between @xmath238 and @xmath458 and a group action of @xmath458 on @xmath456 that we will denote by @xmath460 satisfying the two following conditions : * the distribution of @xmath457 is invariant under the action @xmath460 .
* @xmath461 for example , for permutation tests , the group @xmath458 is simply the group of permutations @xmath238 itself .
the isomorphism @xmath459 is the inverse map @xmath462 .
the group action @xmath460 is just the permutation of the observations .
finally , @xmath228 can be any score that is symmetric with respect to the observations . + assuming these conditions
, one can immediately apply lemma @xmath463 to conclude .
given a partition of the space of hypotheses , the basic assumption which allows the coarse - to - fine multiple testing algorithm to obtain greater power than the bonferroni - holm approach at the same fwer level is that the distribution of the numbers of active hypotheses across the cells of the partition is non - uniform .
the gap in performance is then roughly proportional to the degree of skewness .
the test derived for the parametric model can be seen as a generalization to coarse - to - fine testing of the f - test for determining whether a set of coefficients is zero in a regression model ; the testing procedure derived for the non - parametric case is a generalization of permutation tests to a multi - level multiple testing .
this scenario was motivated by the situation encountered in genome - wide association studies , where the hypotheses are associated with genetic variations ( e.g. , snps ) , each having a location along the genome , and the cells are associated with genes . in principle , our coarse - to - fine procedure will then detect more active variants to the extent that these variants cluster in genes .
of course this extent will depend in practice on many factors , including effect sizes , the representation of the genotype ( i.e. , the choice of variants to explore ) as well as the phenotype , and complex interactions within the genotype .
it may be very difficult and uncommon to know anything specific about the expected nature of the combinatorics between genes and variants . in some sense
, `` the proof is in the pudding , '' in that one can simply try both the standard and coarse - to - fine approaches and compare the sets of variants detected .
given tight control of the fwer , everything found is likely to be real .
indeed , the analytical bounds obtained here make this comparison possible , at least under linear model commonly used in gwas and in a general non - parametric model under invariance assumptions .
looking ahead , we have only analyzed the coarse - to - fine approach for the simplest case of two - levels and a true partition , i.e. , non - overlapping cells .
the methods for controlling the fwer for both the parametric and non - parametric cases generalize naturally to multiple levels assuming nested partitions .
the analytical challenge is to generalize the coarse - to - fine approach to overlapping cells , even for two levels : while our methods for controlling the fwer remain valid , they are likely to become overly conservative if cell overlap .
this case is of particular interest in applications , where genes are grouped into overlapping `` pathways . '' for example , in `` systems biology , '' cellular phenotypes , especially complex diseases such as cancer , are studied in the context of these pathways and mutated genes and other abnormalities are in fact known to cluster in pathways ; indeed , this is the justification for a pathway - based analysis .
hence the clustering properties may be stronger for variants or genes in pathways than for variants in genes .
michael ashburner , catherine a ball , judith a blake , david botstein , heather butler , j michael cherry , allan p davis , kara dolinski , selina s dwight , janan t eppig , et al .
gene ontology : tool for the unification of biology .
, 25(1):2529 , 2000 . | we analyze control of the familywise error rate ( fwer ) in a multiple testing scenario with a great many null hypotheses about the distribution of a high - dimensional random variable among which only a very small fraction are false , or `` active '' . in order to improve power relative to conservative bonferroni bounds ,
we explore a coarse - to - fine procedure adapted to a situation in which tests are partitioned into subsets , or `` cells '' , and active hypotheses tend to cluster within cells .
we develop procedures for a standard linear model with gaussian data and a non - parametric case based on generalized permutation testing , and demonstrate considerably higher power than bonferroni estimates at the same fwer when the active hypotheses do cluster .
the main technical difficulty arises from the correlation between the test statistics at the individual and cell levels , which increases the likelihood of a hypothesis being falsely discovered when the cell that contains it is falsely discovered ( survivorship bias ) .
this requires sharp estimates of certain quadrant probabilities when a cell is inactive . , |
nnser is a detached close binary system consisting of a dwarf m star in a very close orbit ( @xmath3hr ) around the hot white dwarf primary .
such close binaries are formed when the separation of the original main sequence stars are smaller than the extent of a red giant : when the envelope of the evolved primary expands past the orbit of the secondary , frictional forces transfer the orbital angular momentum and energy from the secondary to the envelope . during the resulting `` common - envelope '' ( ce ) phase
, the secondary plunges down into the final tight orbit and @xmath4 of the primary s mass is cast off in the form of a planetary nebula . with an unusually high temperature of about @xmath5k for the white dwarf in nnser
, this event must have happened only about a million years ago .
eventually , the loss of angular momentum due to stellar winds and gravitational radiation will bring the white dwarf and m star into contact , after which nnser will become a classic cataclysmic variable ( cv ) .
the light curves of eclipsing pre - cv s are particular simple , since the light curves are affected only by the varying visibilities of the highly irradiated dwarf and the dramatic eclipses of the white dwarf .
this fact , along with the very short orbital periods , makes it particularly easy to measure the ephemerides of such systems . for several of these systems
, long - term monitoring has resulted in the detection of highly significant orbital period variations of the order of tens of seconds relative to linear long - term mean ephemerides ( parsons et al .
there are various plausible origins for these variations , including applegate s mechanism ( applegate 1992 ) , apsidal motion or the presence of one or more bodies around the binary .
historically , there has been some hesitance about attributing the observed timing effects to one or more circumbinary objects : not only does one have to exclude other possible explanations , there is also a deep - set skepticism about the idea of the formation of planets around binary star systems .
thus , the search for circumbinary planets is interesting on several accounts : how many are out there , what binary systems can support them , and even when are such planets formed ?
post - ce systems like nnser provide a unique opportunity to answer some of these questions .
the timing data in the literature for nnser consisted of the early photoelectric measurements by haefner et al .
( 1989 ) on the eso danish 1.5 m , ccd photometry on a smaller telescope by pigulski & michalska ( 2002 ) , vlt observations by haefner et al .
( 2004 ) , wht+ultracam observations by brinkworth et al .
( 2006 ) , and finally ccd photometry on a @xmath6 m by qian et al .
the latter group looked at all of the very heterogeneous timing measurements and suggested the deviations might be due to a 3rd body a massive jupiter - like planet in a @xmath7year orbit but they ignored the fact that some of the measurements were very precise and did not fit their simple 3-rd body orbit at all ( @xmath8 deviations ) .
part of the problem was the fact that there were too few early measurements with accuracies comparable with the latter measurements , particularly an inexact measurement by haefner et al .
( 2004 ) using a trailed image and the vlt , for which haefner et al . estimated an accuracy of only @xmath9sec .
in the hopes of extracting a more accurate timing measurement from the 1999 vlt observations , we re - analyzed the trailed images of nnser taken by haefner et al .
only the image taken on 11 june 1999 was good enough for a detailed analysis ( fig.1 ) .
there are several problems associated with the time - calibration of trailed images , including the intrinsic accuracy of the system clock and potential absolute delays caused by camera operations as well as the relative timing accuracy posed by the characterization of the intrinsic length of the trail .
the absolute precision of the times were estimated by eso staff to be better than @xmath10ms .
the measurement of the length of the trail was performed using several methods . using star images in acquisition images taken immediately before and after the trailed image
, it was possible to estimate the intrinsic smearing of the trail along the trail axis , from which the starting and ending points could be determined ; together with the integration time , this determined the time - scaling .
alternatively , one can simply take the known plate scale and integration time and calculate the time scaling , checking only for a comparable seeing conditions at the start and end of the exposure . using the lucy deconvolution algorithm and the acquisition images , the trails of neighboring stars
were deconvolved to produce trails with very sharp boundaries , which then could be compared with the nnser trail .
finally , the length of the eclipse in the extracted pixel light curve can be calibrated to the known shape of the eclipse , producing a nearly independent time scaling .
all of these methods led to a very exact common time calibration and a final central eclipse time of bjd(tt ) @xmath11 , i.e. with an accuracy of about @xmath12sec .
in addition , we re - analyzed the ccd photometric data of pigulski & michalska ( 2002 ) , kindly provided in the form of a light curve by the authors .
although they had fit a formal eclipse model profile to the data , the finite integration times had not been taken into account .
since the eclipse is of known width and almost exactly representable by a trapezoid of known width and central depth , it was easy to correct for this effect , resulting in a significant improvement in this timing point as well ( from @xmath13 to @xmath14sec ) : bjd(tt ) @xmath15 .
we have obtained additional observations using several telescopes .
a small number of highly accurate timings was obtained with ultracam on the eso 3.5 m nnt telescope with a sloan g filter ( orbits @xmath16 & @xmath17 ) and with the argus camera on the mcdonald observatory @xmath18 m with a bg40 ( six eclipses between @xmath19 and @xmath20 ) .
these measurements were supplemented with white - light observations made with the remotely controlled the monet / north telescope at mcdonald ( seven eclipses between @xmath21 and @xmath22 ) .
the resulting table of revised and new timing observations can be found in beuermann et al .
( 2010 ) .
the timing points from the literature , the revised timing values , and our new observations are shown in fig.2 . with the revised timing points ,
the variations no longer look random but show a quasi- but not perfectly sinusoidal change in the effective orbital period on a timescale of two decades .
in the case of nnser , it is very easy to see that the timing variations can only be due to at least one additional body .
the observed timing variations are much too large to be due to the applegate mechanism ( chen 2009 ) . while apsidal motion is possible , the amplitude and form depend upon the eccentricity of the central binary orbit : since the observed eccentricity is essentially zero ( parsons et al .
2010a ) , the apsidal effect should be small and sinusoidal .
we first attempted a one - planet fit ( @xmath23 free parameters , including the binary ephemeris ) to the timing variations .
the resulting orbit has a large eccentricity ( @xmath24 ) and an orbital period of @xmath25years .
however , the reduced @xmath26 is @xmath27 , showing that the residuals on shorter timescales are highly significant . a full two - planet fit ( @xmath28 free parameters ) suffers from the bunching of the data points into @xmath29 loose groups , so a full mapping of parameter space was necessary in order to find a reasonable family of solutions .
since the eccentricity of the main planetary component is always small , we set it to zero and fit the remaining @xmath30 parameters .
two solutions remained , both having reduced @xmath26 of about @xmath31 : one with an orbital period ratio of @xmath32 and another with a ratio of @xmath33 , suggesting that the two planets are in a resonant configuration .
the planet with the largest effect has an orbital period of @xmath34 ( or @xmath35 ) years , a semi - major axis of @xmath36 ( @xmath37 ) a.u . , and a mass of @xmath38 ( @xmath39 ) @xmath40 ( assuming that the planets have the same inclination of @xmath41 as the binary ) .
the companion planet has a shorter period of @xmath42 ( @xmath43 ) years , a semi - major axis of @xmath44 ( @xmath45 ) a.u . , and a mass of @xmath46 ( @xmath47 ) @xmath40 .
the latter also has a significant eccentricity of @xmath48 ( @xmath49 ) , so that the second timing component can not be due to apsidal motion but must be due to a true additional body .
the detailed parameters and errors can be found in beuermann et al .
( 2010 ) .
nnser thus joins the very small group of binary star systems with circumbinary planets : dpleo ( one planet ; qian et al .
2010 ) , huaqr ( one planet ; schwarz et al .
2009 ) , qsvir ( probably one planet in addition to a brown dwarf ; parsons et al .
2010b ) , and hwvir ( 2 planets ; lee et al .
in order to characterize the planetary system in nnser , one needs to know how , where , and when the planets were formed .
unfortunately , it is not at all trivial to estimate what the binary system looked like before the close binary was created : the ce phase is very difficult to simulate .
the frictional cross - sections of the secondary star and planets are largely gravitational , for which the bondi - hoyle drag coefficients are only a rough approximation .
the secondary star plunging into the envelope of the red giant primary is not simply a question of a transfer of energy and angular momentum because the envelope as a quasi - hydrostatic body has it s own internal energy .
the ce itself assumes a strongly bipolar shape as it expands but the actual density and velocity of the envelope in the orbital plane needed to calculate the drag effects is difficult to estimate . in the late stages ,
the hot core of the red giant creates an enormous bipolar hii region and thus dramatically increases the inner sound speed relative to the cool expanding outer envelope .
zorotovic et al .
( 2010 ) have studied the simple parameterizations of ce and found that the ce efficiency parameter , @xmath50 , must be in the range @xmath51@xmath52 . adopting this parameterization and a value of @xmath53 , we estimate that the former binary star system consisted of a @xmath0 a star orbited by the present m4 dwarf secondary at a separation of about @xmath54a.u .
one possibility is that the planets in nnser existed before the ce phase and were created with the stars out of an original circumbinary proto - planetary disk
so - called `` 1st - generation '' planets .
stability requires that such planets have orbits with semi - major axes of at least @xmath55a.u . due to the loss of @xmath4 of the mass of the binary system during the ce
, one would naively expect any 1st - generation planets to either be lost or at least to have their orbits expanded quasi - adiabatically by a factor of @xmath56 .
thus , 1st - generation planets must have originally been at much larger orbital separations and have suffered the orbital drag during the ce phase which produced the close binary , even though the gravitational cross - section is much smaller and hence the intrinsic effect less strong .
if drag is more important than one would naively expect , does this mean that the original planets may have been formed in very distant orbits and how many close - in planets were lost ?
the alternative scenario is that the planets did not exist before the ce phase but were created afterwards .
this is not implausible , since the amount of matter cast off is substantial much more than is typically in a proto - planetary disk .
however , it is not at all clear how the planetary nebula caused by the ce evolves over long timescales : if the expansion of the nebula is sub - sonic , does this mean the matter `` returns '' , forms a circumbinary disk , and creates 2nd - generation planets ? does the enriched composition of the ejecta make it simpler to form planets ?
is it possible to turn 1st - generation small planets / asteroids into massive 2nd - generation jupiters ? given the maximum age of 1 million years , 2nd - generation planets in nnser would be the youngest known by far , making them planets particularly interesting for a wide variety of reasons .
the recognition that nnser has a complex circumbinary planetary system with an enigmatic history and that there are bound to be many more of these systems out there opens up a totally new perspective on the formation of planets .
before these systems appeared , we naively thought it was very unlikely that circumbinary planets might form .
now we know that they exist , but realize how little we actually know about their cosmogony .
j. h. applegate , _ apj _ * 385 * , 621 ( 1992 ) .
k. beuermann et al . , _ a&a _ , in press ( 2010 ) .
c. s. brinkworth , t. marsh , v. s. dhillon , c. knigge , _ mnras _ * 365 * , 287 ( 2006 ) .
r. haefner , _
a&a _ * 213 * , l15 ( 1989 ) .
r. haefner , a. fiedler , k. butler , h. barwig , _
a&a _ * 428 * , 181 ( 2004 ) .
k. z. kaminski et al . , _ aj _ * 134 * , 1206 ( 2007 )
. j. lee et al . , _ aj _ * 137 * , 3181 ( 2009 ) .
s. g. parsons et al . , _ mnras _ * 402 * , 2591 ( 2010a ) .
s. g. parsons et al . , _ mnras _ * 407 * , 2362 ( 2010b ) .
a. pigulski , g. michalska , _ ibvs _ * 5218 * , 1 ( 2002 ) .
r. schwarz et al . , _ a&a _ * 496 * , 805 ( 2009 ) .
qian et al . , _ apj _ * 706 * , l96 ( 2009 ) . s .-
qian et al .
, _ apj _ * 708 * , l66 ( 2010 ) .
m. zorotovic , m. r. schreiber , b. t. gnsicke , _ a&a _ * 520 * , 86 ( 2010 ) . | we have detected 2 circumbinary planets around the close binary system nn serpentis using the orbital time delay effect measured via the sharp eclipses of the white dwarf primary .
the present pre - cataclysmic binary was formed when the original @xmath0 primary expanded into a red giant , causing the secondary star to drop from its original orbit at a separation of about @xmath1a.u . down to its current separation at @xmath2a.u .
a quasi - adiabatic evolution of the circumbinary planets orbits during the common - envelope phase would have placed them in unstable configurations , suggesting that they may have suffered significant orbital drag effects and were originally in much larger orbits .
alternatively , they may have been created as 2nd - generation planets during the last million years from the substantial amount of material lost during the creation of the binary , making them the youngest planets known .
either solution shows how little we actually understand about planetary formation .
address = georg - august - universitt gttingen , institut fr astrophysik , friedrich - hund - platz 1 , 37077 gttingen , germany address = georg - august - universitt gttingen , institut fr astrophysik , friedrich - hund - platz 1 , 37077 gttingen , germany address = georg - august - universitt gttingen , institut fr astrophysik , friedrich - hund - platz 1 , 37077 gttingen , germany address = department of physics , university of warwick , coventry , cv47al , uk address = department of physics , university of warwick , coventry , cv47al , uk address = department of physics , university of warwick , coventry , cv47al , uk address = department of astronomy , university of texas at austin , rlm 16.236 , austin , tx 78712 , usa address = department of astronomy , university of texas at austin , rlm 16.236 , austin , tx 78712 , usa address = department of astronomy , university of texas at austin , rlm 16.236 , austin , tx 78712 , usa address = departamento de fsica y astronoma , facultad de ciencias , universidad de valparaso , av . gran bretaa 1111 , valparaso , chile address = institute fr astronomie und astrophysik , kepler center for astro and particle physics , eberhard - karls - universitt tbingen , morgenstelle 10 , 72076 tbingen , germany address = department of physics and astronomy , university of sheffield , sheffield s3 7rh , uk address = department of physics and astronomy , university of sheffield , sheffield s3 7rh , uk |
coherence is ultimately the most distinctive feature of quantum systems .
finding a proper measure of the coherence present at different times scales in a quantum dynamical system is the first essential step for assessing the role of quantum interference in natural and artificial processes .
this is of particular relevance for those quantum evolutions in which information or energy are transformed and transferred in order to achieve a given task with high efficiency . in this context
the relevant questions are : how much and what kind of coherence is created vs destroyed by the dynamical evolution ? how does coherence determine / enhance the performance of the given process ?
these are in general difficult questions and to be answered they require an appropriate and sufficiently comprehensive framework . a general and fundamental formalism to describe quantum interference is provided by the decoherent histories ( dh ) approach to quantum mechanics .
dh have mainly found applications to foundational issues of quantum mechanics such as the formulation of a consistent framework to describe closed quantum systems , the emergence of classical mechanics from a quantum substrate , the solution of quantum paradoxes , decoherence theory , quantum probabilities @xcite . however , dh can also be a systematic tool for quantifying interference in quantum processes , and discussing its relevance therein . indeed ,
dh provide a precise mathematical formalization of interference by means of the the so called _ decoherence matrix _ @xmath0 .
the latter is built on the elementary notion of _ histories _ and allows one to describe the quantum features vs the classical ones in tems of interference between _ histories _ , or _ pathways _ if one resorts to the mental picture of the double slit experiments .
it is however difficult to quantify in a compact and meaningful way the content of @xmath0 and its implications for the dynamics of specific systems .
our first main goal is therefore to define and test appropriate measures allowing for the investigation of how interference can determine the performance of a given quantum information processing task .
starting from @xmath0 and by its sub - blocks we define different functionals .
in particular , we introduce a global measure of coherence @xmath1 able to describe the coherence content of a general quantum evolution at its various time scales ; an average ( over different time - scales ) measure of coherence ; and average mesure of interference between histories leading to a specific output . while the tools we introduce are of general interest and application , in order to test them we apply them to a specific but relevant instance of quantum dynamics taking place in photosynthetic membranes of bacteria and plants : quantum energy transport . here
the basic common mechanism is the following : a quantum excitation is first captured by the system and then migrates through a network of sites ( chromophores ) towards a target site , e.g. , a reaction center , where the energy is transformed and used to trigger further chemical reactions .
there is now an emerging consensus that efficient transport in natural and biologically - inspired artificial light - harvesting systems builds on a finely tuned balance of quantum coherence and decoherence caused by environmental noise @xcite , a phenomenon known as environment - assisted quantum transport ( enaqt ) .
this paradigm has emerged with clarity in recent years , as modern spectroscopic techniques first suggested that exciton transport within photosynthetic complexes might be coherent over appreciable timescales @xcite .
indeed , a growing number of experiments has provided solid evidence that coherent dynamics occurs even at room temperature for unusually long timescales ( of the order of @xmath2 ) @xcite .
efforts to describe these systems have led to general models of enaqt @xcite , depicting the complex interplay of three key factors : coherent motion , i.e. , quantum delocalization of the excitation over different sites , environmental decoherence , and localization caused by a disordered energy landscape .
so far , the presence of coherence in light - harvesting systems has been qualitatively associated to the observation of distinctive ` quantum features ' .
originally , coherence was identified with ` quantum wavelike ' behavior as reflected by quantum beats in the dynamics of chromophore populations within a photosynthetic complex .
later works , employing quantum - information concepts and techniques , have switched attention towards quantum correlations between chromophores , in particular quantum entanglement@xcite .
besides being open to criticism ( see , e.g. , @xcite ) , these approaches do not provide direct quantitative measures of coherence in the presence of noise .
therefore , in what follows , we shall apply the novel tools based on dh to a simple yet fundamental model of quantum energy transfer .
we will focus on a relevant trimeric subunit of the fenna - matthews - olson ( fmo ) complex , the first pigment - protein complex to be structurally characterized @xcite .
the trimer is virtually the simplest paradigmatic model retaining the basic charcteristics of a disordered transfer network and it can also be conceived as an essential building block of larger networks . for simplicity , we will use the well - known haken - strobl model @xcite to describe the interplay between hamiltonian and dephasing dynamics .
while the model is an oversimplified description of the actual dynamics taking place in real systems , it allows to spot out the essential features that may determine the high efficiency of the transport .
we shall initially focus on a new coherence measure @xmath1 , based on the decoherence matrix , and characterize its behavior verifying that it can consistently identify the bases and timescales over which quantum coherent phenomena are present during the evolution of the system .
we shall then show how the average coherence exhibited on those time scales can be connected with the delocalization process .
a more detailed analysis will be aimed at distinguishing between constructive and destructive interference affecting the histories ending at the site where the excitation exits the photosynthetic structure . by using the decoherence functional
, we will show that the beneficial role of dephasing for the transport efficiency lies in a selective suppression of destructive interference , a fact that has been systematically suggested in the literature , but never expressed within a general and comprehensive framework that allows the quantitative evaluation of coherence and its effects .
the application of the introduced tools and methods based on dh to a simple yet paradigmatic system shows how one can properly quantify the coherence content of a complex quantum dynamics and elucidate the role of coherence in determining the overall efficiency of the process @xcite .
the paper is structured as follows . in section [ sec : decoherent - histories ] we review the basic decoherent histories formalism . in section
[ sec : the - coherence - measure ] we define the measure of coherence @xmath1 and describe its meaning and properties . in section [ sec : trimer ]
we first introduce the used model for describing the energy transport in the selected trimeric complex .
we then discuss the coherence properties of the excitonic transport : by means of the appropriate measures based on the decoherent histories formalism we identify the essential features that may determine the high efficiency of the transport . in section [ sec : fmo ] we briefly discuss how to extend our results to the whole fmo complex . in section [ sec : conclusions ] we summarize our results an draw our conclusions .
the formalism of decoherent ( or consistent ) histories was developed in slightly different flavors by griffiths @xcite , gell - mann @xcite , hartle@xcite and omns @xcite .
dh provide a consistent formulation of quantum mechanics where probabilities of measurement outcomes are replaced by probabilities of _ histories_. in this formulation , external measurement apparatuses are not needed , and then one does not need to postulate a classical domain of observers . as a consequence ,
quantum mechanics becomes a theory that allows the calculation of probabilities of sequences of events within any closed system , including the whole universe , without the necessity of invoking postulates about the role of measurement . in this framework ,
the classical domain can be seen to emerge as the description of the system becomes more and more coarse - grained .
the idea of ` histories ' stems from feynman s ` sum - over - histories ' formulation of quantum mechanics .
as is known , any amplitude @xmath3 between an initial and a final state can be expressed as a sum over paths , or histories : upon inserting the identity decomposition @xmath4 at differerent times @xmath5 we get @xmath6 where we use the heisenberg notation @xmath7 .
thus the total amplitude @xmath3 is decomposed as a sum of amplitudes , each one corresponding to a different _ history _ identified by a sequence of projectors @xmath8 .
the decoherent histories formalism assumes that histories are the fundamental objects of quantum theory and gives a prescription to attribute probabilities to ( sets of ) histories .
a history is defined as a sequence of projectors at times @xmath9 .
probabilites can be assigned within exhaustive sets of exclusive histories , i.e. , sets of histories @xmath10 where subscripts @xmath11 label different alternatives at times @xmath12 .
histories are exhaustive and exclusive in the sense that the projectors at each time satisfy relations of orthogonality @xmath13 , and completeness , @xmath14 . in other words ,
the projectors @xmath15 define a projective measurement . within a specified set
, any history can be identified with the sequence of alternatives @xmath16 realized at times @xmath12 .
different alternative histories can be grouped together with a procedure called _ coarse - graining_. starting from histories @xmath17 and @xmath18 we can define a new , _ coarse - grained _
history @xmath19 by summing projectors for all times @xmath20 such that @xmath21 and @xmath22 differ : @xmath23 for all @xmath24 . by iterating this procedure
, one can obtain more and more coarse - grained histories .
a special type of coarse - graining is the _ temporal coarse - graining _ : we group together histories @xmath25 such that such that at some time @xmath20 we have @xmath26 .
then the coarse - grained history @xmath27 contains only one - projector ( equal to the identity ) at time @xmath20 , that can be neglected and hence removed from the string of projectors defining the history . on the other hand _ temporal fine - graining _ can be implemented for example by allowing different alternatives at a times @xmath28 . in particular
, one can create new sets of histories @xmath29 from a given one @xmath30 by adding different alternatives at time @xmath31 ; the sets @xmath32 are fine grained versions of the sets @xmath30 .
+ once we specify the initial state @xmath33 and the ( unitary ) time evolution @xmath34 , we can assign any history @xmath17 a _ weight _ @xmath35,\quad\mbox{with}\quad c_{{\mathbf{j}}}=p_{j_{n}}(t_{n})\dots p_{j_{1}}(t_{1})\ ] ] where we use the heisenberg notation @xmath36 .
when the initial state is pure , @xmath37 and the final projectors are one dimensional , @xmath38 , this formula takes the simple form of a squared amplitude @xmath39 weights can not be interpreted as true probabilities , in general . indeed ,
due to quantum interference between histories , the @xmath40 do not behave as classical probabilities .
indeed , consider two exclusive histories @xmath41 and the relative coarse - grained history @xmath42 by : @xmath43 .
if the @xmath44 were real probabilities , we would expect @xmath45 .
instead , what we find is @xmath46).\ ] ] due to the non - classical term @xmath47)$ ] , representing quantum interference between the histories @xmath17 and @xmath18 , the classical probability - sum - rule is violated .
the matrix @xmath48={\mbox{tr}}[p_{j_{n}}u(t_{n}-t_{n-1})\dots p_{j_{1}}u(t_{1})\varrho u(t_{1})^{\dagger}p_{k_{1}}\dots u(t_{n}-t_{n-1})^{\dagger}p_{k_{n}}]\label{eq : decomatrix}\ ] ] is called _ decoherence functional _ or _
decoherence matrix_. the decoherence matrix can be thought of as a `` density matrix over histories '' : its diagonal elements are the weights of histories and its off - diagonal elements are interferences between pairs of histories .
the decoherence matrix has the following properties : i ) it is hermitian ii ) it is semipositive definite iii ) it is trace one iv ) it is block - diagonal in the last index , @xmath49
. weights of coarse - grained histories can be obtained by summing matrix entries in an @xmath50 block of the decoherence matrix corresponding to the original fine - grained histories . for istance ,
the weight of history @xmath42 is obtained by summing entries of a @xmath51 block of the decoherence matrix : @xmath52 a necessary and sufficient condition to guarantee that the probability sum rule @xmath53 apply within a set of histories is @xmath54=0,\forall{\mathbf{j}}\neq{\mathbf{k}}.\ ] ] this condition is termed as _
weak decoherence@xcite _ ; the necessary and sufficient condition that is typically satisfied _ @xcite _ and that we will adopt in the following is the stronger one termed as _ medium decoherence _
@xmath55 medium decoherence implies weak decoherence .
any exhaustive and set of exclusive histories satisfying medium decoherence is called a _
decoherent set_. the fundamental rule of dh approach is that probabilities can be assigned within a decoherent set , each history being assigned a probabability equal to its weight .
if medium decoherence holds , the diagonal elements of the decoherence matrix can be identified as real probabilities for histories and we can write @xmath56 . + + due to property iv ) , if we perform a temporal coarse - graining over all times except the last , we obtain ` histories ' with only one projection , @xmath57 at the final time @xmath58 .
these histories automatically satisfy medium decoherence : @xmath59 & \equiv\delta_{j_{n}k_{n}}p_{j_{n}}\end{aligned}\ ] ] where @xmath60 $ ] is the probability that the system is in @xmath61 at time @xmath58 .
due to interference , the probability of being in @xmath61 at time @xmath58 is _ not _ simply the sum of probabilities of all alternative paths leading to @xmath61 , i.e , of all alternative histories with final projection @xmath57 . in formulas , @xmath62 the probability and the global interference of histories @xmath63 ending in @xmath61 can be thus expressed as @xmath64 with @xmath65 .
destructive interference will happen when @xmath66 , constructive interference when @xmath67 .
the decoherent histories formalism is consistent with and encompasses the model of environmentally induced decoherence@xcite . given a factorization of the hilbert space into a subsystem of interest and the rest ( environment ) , @xmath68
, the events of a history take the form @xmath69 where @xmath70 and @xmath71 are projectors onto hilbert subspaces of @xmath72 and @xmath73 respectively .
histories for @xmath74 alone can be obtained upon considering appropriate coarse - grainings over the degrees of freedom of the environment , such that the events are @xmath75 where @xmath76 is the identity over @xmath73 . upon introducing the time - evolution propagator @xmath77 as @xmath78 $ ]
, we can rewrite the decoherence matrix as : @xmath79\dots]\tilde{p}_{k_{n-1}}]\tilde{p}_{k_{n } } ] .
\nonumber \end{aligned}\ ] ] if the initial state is factorized @xmath80 , then the reduced density matrix @xmath81 $ ] evolves according to @xmath82 where @xmath83 is the ( non - unitary ) reduced propagator defined by @xmath84=\tilde{\mathcal{k}}_{t\ t_{0}}[\tilde{\varrho}_s(t_{0 } ) ] .
\end{small}\ ] ] if the evolution of the system and environment is markovian , we can write @xmath85 .
as proved by zurek @xcite , under the assumption of markovianity we can rewrite the decoherence matrix in terms of reduced quantities alone , i.e. , quantities pertaining to the system only : @xmath86\dots]\tilde{p}_{k_{n-1}}]\tilde{p}_{k_{n } } ] .
\label{eq : reduceddec}\end{aligned}\ ] ] that is , the model of environmentally induced decoherence can be obtained by applying the decoherent histories formalism to system and environment together , and by coarse - graining over the degrees of freedom of the environment .
the dh approach provides the most fundamental framework in which the transition from the quantum to the classical realm can be expressed .
indeed , it is based on the most basic feature characterizing the quantum world : interference and the resulting coherence of the dynamical evolution . despite being a well developed field of study
, the dh history approach lacks for a proper global measure of the coherence produced by the dynamics at the different time scales .
we therefore introduce a measure that quantifies the global amount of coherence within a set of histories .
assume projectors for all times @xmath88 are taken in a fixed basis @xmath89 , @xmath90 .
assume further that histories are composed by taking equally spaced times between consecutive projections i.e. , @xmath91 .
( in other words , histories correspond to projections applied in the same basis and repeated at regular times ) .
for such a set of histories , consider the decoherence matrix @xmath92\ ] ] where @xmath93 .
take the von neumann entropy of the decoherence matrix , @xmath94.\end{aligned}\ ] ] due to coherence between histories , @xmath95 differs from the ` classical - like ' shannon entropy of history weights @xmath96 where @xmath97 $ ] are the diagonal elements of @xmath98 , i.e. , the weights .
the difference between the two quantities is wider if off - diagonal elements of the decoherence matrix are bigger , i.e. , if the set of histories is more coherent . let us define : @xmath99 we argue that @xmath100 is suitable to be used as a general measure of coherence within the set of histories defined by @xmath101 .
indeed , we can readily prove the following properties : + i ) @xmath102 .
@xmath103 is obvious . to prove @xmath104 ,
let us define a matrix @xmath105 where off - diagonal entries are set to zero .
since @xmath106 & = \sum_{{\mathbf{j}}}d_{{\mathbf{j}},{\mathbf{j}}}^{(n , p,\delta t)}\log d_{{\mathbf{j}},{\mathbf{j}}}^{(n , p,\delta t)}= \\ & = { \mbox{tr}}[\tilde{\mathcal{d}}^{(n , p,\delta t)}\log\tilde{\mathcal{d}}^{(n , p,\delta t)}]\end{aligned}\ ] ] we obtain that the numerator of ( [ eq : cfunction ] ) can be expressed as a quantum relative entropy : @xmath107+{\mbox{tr}}[{\mathcal{d}}^{(n , p,\delta t)}\log{\mathcal{d}}^{(n , p,\delta t ) } ] \\ & = { \mbox{tr}}[\tilde{\mathcal{d}}^{(n , p,\delta t)}(\log\mathcal{d}^{(n , p,\delta t)}-\log\tilde{\mathcal{d}}^{(n , p,\delta t)})]\\ & = h(\mathcal{d}^{(n , p,\delta t)}||\tilde{\mathcal{d}}^{(n , p,\delta t)})\geq0\end{aligned}\ ] ] where @xmath108 is the relative entropy between @xmath109 and @xmath110 .
+ ii ) @xmath111 iff @xmath112 , i.e. , @xmath100 vanishes if medium decoherence holds for the set of histories , since the two quantities @xmath113 and @xmath114 coincide in this case .
thus @xmath100 is in essence a ( statistical ) distance between the decoherence matrix @xmath0 and the corresponding diagonal matrix @xmath115 , renormalized so that its value lies between @xmath116 and @xmath117 .
the greater are the off - diagonal elements of @xmath118 , the greater the distance .
the meaning of @xmath100 can be easily understood if we use the linear entropy , a lower bound to the logarithmic version : @xmath119,\\ \ & 1-h_{l}^{(c)}(p , n,\delta t)={\mbox{tr}}[(\tilde{{\mathcal{d}}}^{(n , p,\delta t)})^{2}].\end{aligned}\ ] ] in this case , we obtain a ` linear entropy ' proxy of @xmath120 as : @xmath121 which is a simplified version that , by avoiding the diagonalization of @xmath98 , helps containing the numerical complexity .
the measure introduced is well grounded on physical considerations . in the following
we will apply it to a simple system in order to check its consistency , and later use it to characterize the coherence properties of the evolutions induced by various regimes of interaction with the environment .
first , one has to check whether the measure properly takes into account the action of the bath .
in particular , if the bath is characterized by a decoherence time @xmath122 , it is known ( @xcite ) that on time scales @xmath123 the decoherence matrix becomes diagonal : the probability of a history at time @xmath124 can be fully determined by its probability at time @xmath58 , since no interference can occur between different histories . indeed , the action of the bath is to create a _
decoherent set of histories _ that are defined by a proper projection basis : the pointer basis ( @xcite ) .
therefore , the fine - graining procedure obtained by constructing a set of histories @xmath32 via the addition of a new complete set of projections in the same basis at time @xmath125 to the set @xmath30 , should leave the coherence functional @xmath87 invariant , i.e. , @xmath126 .
if instead @xmath127 the same fine - graining procedure should lead to @xmath128 . before passing to analyze a specific system , we want to focus on the complexity of the evaluation of @xmath0 and @xmath129 .
the dimension of the decoherent matrix grows with the dimension @xmath130 of the basis @xmath131 and the number @xmath132 of time instants that define each history as @xmath133 . this exponential growth in principle limits the application of the dh approach to small systems .
however , as for the system considered in this paper the computational effort is contained due to the small number of subsystems ( chromophores ) and the small dimension of the hilbert space which is limited to the single - exciton manifold . as we shall see , by limiting the choice of @xmath132 to a reasonable number , the analysis can be fruitfully carried even on a laptop .
we now start to analyze decoherent histories in simple models of energy transfer comprising a small number @xmath130 of chromophores ( sites ) . neglecting higher excitations
, each site @xmath134 can be in its ground @xmath135 or excited @xmath136 state .
we work in the single - excitation manifold , and define the site basis as @xmath137 i.e. , state @xmath138 represents the exciton localized at site @xmath134 . on - site energies and couplings
are represented by a hamiltonian @xmath139 that is responsible for the unitary part of the dynamics .
interaction with the environment is implemented by the haken - strobl model , that has been extensively used in models of enaqt @xcite .
the effect of the environment is represented by a markovian dephasing in the site basis , expressed by lindblad terms @xmath140 in the evolution , as follows : @xmath141+\sum_{i}\gamma_{i}[2l_{i}\varrho
l_{i}^{\dag}-l_{i}^{\dag}l_{i}\varrho-\varrho l_{i}^{\dag}l_{i}]\label{eq : hakenstrobl}\ ] ] where @xmath142 are projectors onto the site basis , and @xmath143 are the ( local ) dephasing rates .
furthermore , site @xmath130 can be incoherently coupled to an exciton sink , represented by a linblad term @xmath144\ ] ] where @xmath145 and @xmath146 is the trapping rate .
contrary to other works , we neglect exciton recombination , as it acts on much longer timescales @xmath147 than dephasing and trapping .
+ the global evolution is markovian and can be represented by means of the liouville equation @xmath148 that can be simply solved by exponentiation , @xmath149 in the notation above , the propagator has the form @xmath150 .
the efficiency of the transport can be evaluated as the leak of the population @xmath151 of the exit site @xmath152 towards the sink : @xmath153 the overall efficiency of the process is obtained by letting @xmath154 while its markovianity limits the faithful description of decoherence processes actually taking place in real photosynthetic systems , _ _ the model retains the basic and commonly accepted aspects of decoherence , that acts in the site basis : albeit in a complex non - markovian way , the protein enviroment measures the system locally ( i.e. , on each site ) , thus _ _ destroying the coherence in the site basis and creating it in the exciton basis . note that the formalism can also be applied to a ` dressed ' or polaronic basis where we include strong interactions between chromophores and vibrational modes .
that is , to apply the dh method , one only needs a model in which an exciton hops between sites , dressed or undressed .
the model is therefore suitable to readily implement the decoherent histories paradigm and to spot the main basic features we are interested in and that are at the basis of the success of enaqt .
the fmo unit has @xmath155 chromophores and a complex energy and coupling landscape with no symmetries .
energies and couplings ( i.e. , the hamiltonian @xmath139 ) can be obtained by different techniques : they can be extracted by means of 2d spectroscopy as in @xcite or computed through ab initio calculations as in @xcite , with similar but not exactly equal results .
this very complex struxture makes fmo far from ideal as a first example to study .
we thus prefer to start by working with a much simpler , yet fully relevant subsystem : the trimeric unit composed by the sites @xmath156 and @xmath157 of the fmo complex in the notation of @xcite ) .
the first chromophore is the site in which the energy transfer begins , while the third chromophore is the site from which the excitation leaves the complex .
the hamiltonian of the trimeric subunit is @xcite @xmath158 the eigenenergies of the system ar given by * @xmath159@xmath160 * @xmath161 which yields the eigenperiods @xmath162 @xmath163 . due its structure ,
the trimer is a chain composed by a pair of chromophores ( @xmath156 ) , degenerate in energy and forming a strongly coupled dimer , and a third chromophore moderately coupled with the second one only . since in the following
we suppose that the exciton starts from site @xmath117 , we expect a prominent role of the dimer in the dynamics , at least in the first tens of femtoseconds . + in order to show how the dh analysis can be implemented , in the following we are going to consider histories in the site and the energy bases , with @xmath132 projections at times @xmath164 .
we first use the coherence function @xmath100 introduced above ( [ eq : cfunction ] ) to evaluate the global coherence of the exciton transport process . in order to test the behavior of @xmath87 for different values of dephasing , in fig [ fig : trimer renger c function ] we first plot @xmath87 as a function of the time interval @xmath165 between projections for two values of the dephasing rate : @xmath166 , corresponding to the full quantum regime ( fig.[fig : trimer renger c function ] ) for the site basis ( a ) ; @xmath167 corresponding to an intermediate value of dephasing ( fig . [ fig : trimer renger c function ] ) for the site basis ( b ) and the energy basis ( c ) . before entering the discussion of the various regimes , we note that as a function of the number of projections @xmath132 all curves display the expected behavior : the increase ( decrease ) of the number of projections corresponds to a temporal fine - graining ( coarse - graining ) of the evolution ; therefore , an increase ( decrease ) of @xmath132 should imply an increase ( decrease ) of the amount of coherence between histories . as shown in fig .
( [ fig : trimer renger c function ] ) the function @xmath87 correctly reproduces the fine ( coarse ) graining feature : the qualitative behavior of @xmath87 as a function of @xmath165 is not affected by the choice of @xmath132 , while an increase of @xmath132 corresponds , at fixed @xmath165 , to an increase of @xmath87 .
we will therefore use in the following the value @xmath168 that allows for a neat description of the phenomena and for a reasonable computational time . as for the behavior at fixed @xmath132
, we have that in the full quantum regime ( @xmath169 ) , the system obviously displays coherence in the site basis only since @xmath170={\langle e_{i}|}\rho{|e_{i}\rangle}\delta_{i , j}\ ] ] @xmath171and the decoherence matrix @xmath0 in the energy basis is diagonal and independent on @xmath165 and @xmath132 .
this simply means that in the full quantum regime histories in the exciton basis are fully decohered , * * since the system is not able to create coherence among excitons . still in the full quantum regime , in the site basis , the coherence oscillates as the exciton , starting at site @xmath117 , goes back an forth along the trimer , and the evolution builds up coherence in this basis , see fig.[fig : trimer renger c function ] ( a ) . in this regime ,
the trimer can be approximately seen as a dimer composed by the first two chromophores , and the exciton performs rabi oscillations with a period given by @xmath172 ; @xmath87 oscillates with half the period : for @xmath173 the exciton is migrated mostly on site @xmath174 and @xmath87 has a minumum which is different from zero since the exciton is partly delocalized on site @xmath157 , and the system therefore exhibits a non vanishing coherence . for intermediate values of @xmath175 , fig.[fig : trimer renger c function ] ( b ) , the coherence in site basis as measured by @xmath87 correctly drops down at @xmath123@xcite .
the dephasing has a strong and obvious effect on the coherence between pathways : coherence in this basis is a monotonically decreasing function of @xmath176 .
this is well highlighted by the global coherence function @xmath87 , whose maximal values are reduced by a factor of @xmath177 with respect to those corresponding to full quantum regime . after a time @xmath178 the histories are fully decohered . indeed , due to the specific model of decoherence ( [ eq : hakenstrobl ] ) , which amounts to projective measurements on @xmath179 at each site with a rate @xmath176 , the system kills the coherence in the site basis , which in turn corresponds to the stable pointer basis for this model @xcite , i.e. , the basis in which the density matrix is forced to be diagonal by the specific decoherence model . on the other hand , and for the same reason , the dynamics starts to build up coherence in the exciton basis @xmath180 , see fig .
[ fig : trimer renger c function](c ) .
however , this coherence is later destroyed - on a time scale of approximately @xmath181 - since the stationary state of the model is the identity .
this effect is even more evident if one compares the behavior of @xmath87 in the exciton basis for different values of @xmath176 , as shown in fig .
[ fig : trimer renger c function](d ) : @xmath87 grows with @xmath176 and it lasts over longer time scales .
this feature is coherent with the expectations : the equilibrium state for high @xmath176 is the identity . due to the projections implemented by the environment in the site basis ,
the system is forced to create coherence in the exciton basis .
when @xmath176 is very high a quantum zeno effect takes in , the dynamics is blocked , and the time required to reach the equilibrium , and to destroy coherences in all bases , consequently grows .
this first analysis therefore shows that @xmath87 is indeed a good candidate for assessing the global coherence properties of quantum evolutions . for a fixed number of projections @xmath132 ,
@xmath182 can be interpreted as a _ measure of the global coherence exhibited by the dynamics over the time scale _ @xmath165 .
we now analyze in detail the specific features of quantum transport for the trimer .
the dynamics starts at site @xmath117 and evolves by delocalizing the exciton on the other chromophores . in order to study this process
, we first use a measure of delocalization introduced in@xcite for the study of lhcii complex dynamics : @xmath183 that is simply the shannon entropy of @xmath184 , the populations of the three chromophores .
this measure allows one to follow how much the exciton gets delocalized over the trimer with time and in different dephasing situations : @xmath185 , i.e. , @xmath186 is zero when the exciton is localized on a chromophore and it takes its maximal value @xmath187 when the population of the three sites are equal . in fig .
: pop site 3 and delocalization ] we plot both @xmath188 and the population @xmath189 of site @xmath157 for different values of @xmath176 . due to the presence of interference , in the mainly quantum regime ( @xmath190 ) , the exiton first delocalizes mainly over the dimer and partly on the third site : the first maximum corresponds to @xmath191 when the system builds up a ( close to uniform ) coherent superposition between sites @xmath117 and @xmath174 , while a non negligible part of the exciton is found in site @xmath157 ; indeed @xmath192 , the last value corresponding to @xmath193 i.e. , to a uniform superposition over the sites @xmath117 and @xmath174 only .
as the dynamics of the sytems extends to later times we see that @xmath188 and @xmath189 have an oscillatory behavior , whose main period is @xmath194 , and which approximately corresponds to rabi oscillations between site @xmath117 and @xmath174 , although the initial state fully localized in site @xmath117 can not be rebuilt due to the presence of site @xmath157 . as for the transport , we see that in this regime the system can not take advantage of the initial fast and high delocalization : the exciton bounces back and forth over the trimer . in the intermediate regime @xmath195 , due ,
as we will later see , to the selective suppression of interference processes , the initial speed up in delocalization is sustained by the dynamical evolution , and the transfer rate to site @xmath157 is correspondingly increased . for very high values of decoherence ( @xmath196 )
the role of initial interference is suppressed and the initial speed - up disappears : the environment measures the system in site basis at high rates and the delocalization process is highly reduced .
the optimal delocalization occurs in correspondence of @xmath197 and it can be interpolated with a double exponential function @xmath198 with @xmath199 .
the first time scale @xmath200 describes the initial fast quantum delocalization process described above , while the second time scale @xmath201 the slower subsequent delocalization and the reaching of the equilibrium situation , @xmath202 .
we now pass to systematically analyze the behavior of the coherence of the evolution with respect to the strength of the interaction with the environment and its relevance for the energy transport process . as a first step we plot both @xmath203 and @xmath204 for different values of @xmath176 , fig .
( [ fig . :
h and c various gamma ] ) .
the plots show that the coherence function exhibits the required behavior : for small @xmath205 , @xmath204 oscillates with period @xmath194 , following the rabi oscillations of the dimer .
the minima occur at @xmath206 , showing that the exciton is `` partially '' localized on site @xmath117 or @xmath174 , and partially delocalized on site @xmath157 .
as @xmath176 grows , the system becomes unable to create coherence on large time scales ; the decay of @xmath204 is mirrored by a the reduction of the amplitude in the oscillations of @xmath203 . and coherence function @xmath204 for @xmath207 @xmath208,scaledwidth=40.0% ]
we now focus on the relevant time scales @xmath209 for the initial fast delocalization process highlighted by our previous analysis , which are of the order of tens to hundreds of femtoseconds .
we therefore introduce the following _ average measure of global coherence of the evolution _
@xmath211 is the average of the coherence exhibited by the dynamics of the system at the time scales @xmath212 . in fig .
average c vs gamma ] we show @xmath211 for the trimer ( [ eq : hamtrimer renger ] ) in the site basis for different values of @xmath209 .
* * we first focus on the behavior of @xmath211 for values of dephasing in the range @xmath213 @xmath208 . in this range , for small time
scales @xmath214 to @xmath215 the average global coherence @xmath211 is approximately constant and equals the value attained in the full quantum regime i.e. , @xmath216 . for larger time scales ( @xmath217 ) @xmath211 rapidly decreases with @xmath176 .
this analysis shows that the behavior of @xmath211 matches the expectations : the higher @xmath176 the smaller the time scales over which decoherence takes place , the lower the global coherence of the dynamics .
along with @xmath182 the functional @xmath211 is therefore in general a good candidate for the evaluation of the global coherence of open quantum systems evolution .
as for the transport dynamics , we focus on the timescale identified with the analysis of @xmath188 for optimal dephasing ; for @xmath218 and @xmath219 we see that the system indeed retains most of the average coherence of the purely quantum regime up to the optimal values of decoherence ( @xmath195 in the figure ) , losing it afterwards ; this is a clear indication that this phenomenon is at the basis of the the fast initial delocalization process . over longer time scales ,
the relevance of coherence is highly suppressed**. * * for the trimer ( [ eq : hamtrimer renger ] ) for different values of @xmath209,scaledwidth=40.0% ] we now deepen our analysis about the relevance of the coherence of the evolution for the energy tranfer efficiency . to this aim
we focus on the basic feature that distingushes the classical and the quantum regime : interference .
in particular we focus on the sub - block @xmath220 of the decoherence matrix @xmath0 pertaining to the third chromophore , which describes the set of histories in site basis ending at site @xmath157 . due to interference the probability of occupation of the site @xmath157 at time @xmath221 can be written in terms of the the histories ending at site @xmath157 @xmath222 , see ( [ eq : interference and probability at ending site j ] ) . in fig .
: interference with gamma trimer ] we show @xmath223 for different values of dephasing .
one has different regimes : for @xmath224 , the set of histories in site basis is fully decohered ; @xmath225 , the histories do not interfere with each other and @xmath226 i.e , the probability is simply the sum of diagonal elements of @xmath220 . in the mainly quantum regime @xmath227 ,
@xmath228 : after the initial positive peak the histories interfere with each other , globally the interference is mostly negative and therefore @xmath229 . for intermediate values of decoherence @xmath230
the interference has a positive peak and then reduces to zero . while the first initial fingersnap of positive interference that takes place in the first @xmath231 is common for all curves corresponding to small and intermediate values of @xmath176 , the _
main effect of the bath is displayed after this initial period of time : the decoherence gradually suppresses interference , both the positive and the negative one _ ; _ however , for intermediate values of @xmath176 the effect is stronger as for the negative part of the interference patterns_. the environment thus implements what can be called a _ quantum recoil avoiding effect : _ it prevents the part of the exciton that thanks to constructive interference
has delocalized on site @xmath157 to flow back to the the other sites . ) * ( a ) * average positive ( green ) , negative ( blue ) , total ( red ) interference @xmath232 of histories ending in site @xmath157 as a function of @xmath176 ; * ( b ) * total average interference @xmath233 for site @xmath234 ( red ) , @xmath235 ( blue ) , @xmath236 ( green ) @xmath176.,title="fig:",scaledwidth=40.0% ] ) * ( a ) * average positive ( green ) , negative ( blue ) , total ( red ) interference @xmath232 of histories ending in site @xmath157 as a function of @xmath176 ; * ( b ) * total average interference @xmath233 for site @xmath234 ( red ) , @xmath235 ( blue ) , @xmath236 ( green ) @xmath176.,title="fig:",scaledwidth=40.0% ] in order to evaluate a possible advantage provided by the initial speed - up in the delocalization process and by the interference phenomena showed above one has to take into account another relevant time scale of the transport process : the trapping time . indeed ,
if the system is to take advantage of the fast delocalization due to the coherent behavior , the exit of the exciton should take place on time scales of the order of the delocalization process .
the theoretical and experimental evidences show that this is the case : the trapping time @xmath237 for the fmo complex is estimated in the literature to be of the order of @xmath181 i.e. , the exit of the exiton starts soon after the fast delocalization due to quantum coherence has taken place .
the role of the interference between paths , in particular those leading to site @xmath157 , can therefore be appreciated by numerically evaluating @xmath238 i.e. , the average over the trapping time scale of @xmath239 of the total ( @xmath240 ) , negative ( @xmath241 ) and positive ( @xmath242)@xmath171 average interference between the histories ending in site @xmath134 , with @xmath243 . in particular , in fig .
[ fig . : average interference site 3](a ) the different kinds of interference are plotted for histories terminating at site @xmath157 : on average , the negative interference highly reduces the total interference for small values of decoherence strength ; when @xmath244 , @xmath245 vanishes , the average total interference equals the positive one @xmath246 , and it is maximal for values of @xmath176 comparable to those that maximize @xmath188 ( @xmath247 ) . in fig .
[ fig . : average interference site 3](b ) , we compare the behavior of @xmath248 for all sites .
the results again suggest that decoherence acts on the interference provided by the quantum engine in order to favor the flow of the exciton towards the exit chromophore : the average positive interference between histories ending at sites @xmath174 and @xmath157 grows in modulus with @xmath176 and attains a maximum for intermediate values of decoherence ; while the average negative interference between histories ending at site 1 decreases and attains a minimum for intermediate values of @xmath176 .
the combined effect of decoherence and interference thus helps depopulating site 1 and populating site 2 and 3 .
we can now tackle one of the most relevant aspects of our discussion : the net effect of the above described phenomena on the overall efficiency of the transport .
the latter can be fully appreciated by evaluating the efficiency of the process ( [ eq : efficiency ] ) and by recognizing that , in the decoherent histories language , it can be expressed as : @xmath249 where @xmath221 and @xmath250 .
this split allows one to appriciate the role of interference for the efficiency . in fig .
( [ fig . : efficiency vs interference trimer ] ) @xmath251 is plotted for different values of dephasing . in agreement with what discussed above
, we have three regimes : for very small values of @xmath176 the overall efficiency is poor ; this is due to the presence of high negative interference that in average prevents the exciton to migrate to the exit site . for large values of @xmath176
the interference processes are completely washed out and the system can not take advantage of the fast quantum delocalization . for intermediate ( optimal ) values of @xmath176
only the negative interference has been washed out : @xmath252 is positive , it acts on short time scales , and it provides on average an enhancement of the global efficiency . ):
transport efficiency @xmath253 , integrated weight @xmath254 and integrated interference @xmath255 for pathways ending at site @xmath157 for different values of @xmath256 @xmath208 and @xmath257,scaledwidth=40.0% ] these results , within the limits of the simple model of decoherence taken into account , undoubtedly show for the first time that the so called enaqt phenomenon can well and properly be understood both qualitatively and quantitatively within the decoherence histories approach , i.e. , in terms of very the basic concepts of coherence and interference between histories .
the often recalled `` convergence '' of time scales or `` goldilocks '' effect ( @xcite ) in biological quantum transport systems seems therefore to be well rooted in the processes discussed above : if decoherence is too small the system shows both positive and negative interference ( see fig .
: pop site 3 and delocalization ] ) , the delocalization has an ocillatory behavior , and the exciton bounces back and forth along the network thus preventing its efficient extraction .
if instead decoherence is very high one has that the complete washing out of intereference and coherence implies the delocalization process to be very slow , no matter how fast the trapping mechanism try to suck the exciton out of the system . in order to take advantage of the effects of quantum coherent dynamics : @xmath258 the bath must act on the typical time scales of quantum evolution in order to implement the quantum recoil avoiding process ; @xmath259 the extraction of the exciton from the complex , characterized by @xmath146 , must then start soon after the initial fast delocalization has taken place .
should the extraction take place on longer time scales , the benefits of the fast initial delocalization would be spoiled : waiting long enough , the system would eventually reach together with equilibrium a decent delocalization even for moderately high values of @xmath176 , but in this case the transfer would be obviously much slower .
the above arguments can be easily applied to the whole fmo complex . fig . [
fig : fmo c various gammas and c vs h ] and [ fig : fmo average i on 200 fs 3 site and efficiency ] show the application of the decoherent histories method to excitonic transport in fmo .
the main features of the behavior of @xmath260 and @xmath251 are maintained although obvious differences can be found since the dynamics in now determined by the interplay of different eigenperiods and interference paths are more complex .
in particular , fig.[fig : fmo c various gammas and c vs h](b ) , one can observe a revival of positive interference @xmath261 for small values of @xmath262 , that does enhance the efficiency for @xmath263 , fig .
[ fig : fmo average i on 200 fs 3 site and efficiency](a ) ; but this is not sufficient to compensate the initial and subsequent negative interference , thus impeding the reach of optimal values of @xmath251 . in general , compared to the trimer and as suggested by fig .
[ fig : fmo average i on 200 fs 3 site and efficiency](a ) the maximum average positive coherence on short time scales is attained for smaller values of @xmath176 .
the overall picture is not significantly affected if one decides to start the dynamics from site @xmath264 instead of site @xmath117 , as it often is reported in the literature .
the decoherent histories approach provides a general theory to study the distinctive feature exhibited by quantum systems : coherence . however , despite its generality and foundational character , in order to measure the effects of coherence and decoherence the dh approach needs to be complemented with a quantitative way to condense the information contained in the basic object of the theory , i.e. , the decoherence matrix @xmath265 . in this paper
we introduce a set of tools that allow one to assess the ( global ) coherence properties of quantum ( markovian ) evolution and that can be used to relate the coherence content of a general quantum dynamical process to the relevant figure of merits of the given problem .
we first define the _ coherence functional _
@xmath100 , that can be interpreted as a measure of the global coherence exhibited by the dynamics in the basis @xmath131 over the time scale _ @xmath165 .
_ while this measure is completely general , one can further introduce other relevant tools tailored to the specific system and type of system - environment interaction at hand .
we thus focus on a simple yet paradigmatic model of environmentally assisted energy transfer where coherence effects have been shown to play a significant role in determining the efficiency of the process : a trimeric subunit of the fenna - matthews - olson photosynthetic complex .
based on @xmath265 and @xmath100 we define : @xmath266 a measure @xmath267 able to characterize the average coherence exhibited by the dynamics of the system over the time scales @xmath268 for a fixed value of the dephasing @xmath176 ; @xmath269 a measure of the average interference @xmath270 occurring between the histories ending at a given `` site '' @xmath134 .
+ within the specific model , we first thoroughly assess the consistency of the behavior of @xmath100 in the various regimes .
we then show how the introduced tools allow to study the intricate connections between the efficiency of the transport process and the coherence properties of the dynamics .
in particular we show that the delocalization of the exciton over the chromophoric subunit is strongly affected by the amount of ( average ) coherence allowed by the interaction with the bath in the first tens to hundreds of femtoseconds .
if the system - bath interaction is too strong , coherence is suppressed alongside the interference between different histories , in particular those ending at the site where the excitation leaves the complex .
if the interaction is too weak the system exhibits high values of coherence even on long time scales , but it also exhibits negative interference between pathways ending at the exit site , a manifestation of the fact that the exciton bounces back and forth over the network thus preventing its efficient extraction . in the intermediate regime
i.e. , when the different time scales of the system ( quantum oscillations , decoherence and trapping rate ) converge , the system shows high values of coherence on those time scales .
the action of the bath has a _ quantum recoil avoiding effect _ on the dynamics of the excitation : the _ benefits of the fast initial quantum delocalization of the exciton over the network are preserved and sustained in time by the dynamics _ ; in terms of pathways leading to the exit site , the action is to s__electively kill the negative interference between pathways , while retaining the initial positive one . _
_ these effects can be explicitly connected to the overall efficiency of the environment - assisted quantum transport : the gain in efficiency for intermediate ( optimal ) values of decoherence can thus be traced back to the basic concepts of coherence and interference between pathways as expressed in the decoherent histories language .
+ while the specific decoherence model used ( haken - strobl ) is an oversimplified description of the actual dynamics taking place in real systems , we believe that our analysis allows to spot out the essential features that may determine the high efficiency of the transport even in more complex system - environment scenarios .
+ in conclusion , the tools introduced in this paper allow to thoroughly assess the coherence properties of quantum evolutions and can be applied to a large variety of quantum systems , the only limits being the restriction to markovian dynamics and the computational efforts required for high dimensional systems .
however , the extension to non - markovian realms is indeed possible @xcite , and the use of parallel computing may allow the treatement of reasonably large systems .
giorda and m. allegra would like to thank dr .
giorgio `` giorgione '' villosio for his friendship , his support , his always reinvigorating optimism , and his warm hospitality at the institute for women and religion - turin , where this paper was completed ( _ cogitato , mus pusillus quam sit sapiens bestia , aetatem qui non cubili uni umquam committit suam , quin , si unum obsideatur , aliud iam perfugium elegerit _ ) .
+ + p.
giorda would like to thank prof .
a. montorsi , prof .
paris and prof .
m. genovese for their kind help .
+ + s. lloyd would like to thank m. gell - mann for helpful discussions .
m. gell - mann and j. b. hartle , in _ complexity , entropy , and the physics of information _
( addison - wesley , reading , massachusetts , 1990 ) ; in proc . of the 25th international conference on high energy physics , singapore ( world scientific , singapore , 1990 ) .
h. park , n. heldman , p. rebentrost , l. abbondanza , a. iagatti , a. alessi , b. patrizi , m. salvalaggio , l. bussotti , m. mohseni , f. caruso , h. c. johnsen , r. fusco , p. foggi , p. f. scudo , s. lloyd & a. m. belcher , nature materials ( 2015 ) .
h. haken , g. strobl , in _ the triplet state _ , proceedings or the international symposium , am .
beirut , lebanon ( 1967 ) , a.b .
zahlan , ed . , cambridge university press , cambridge ( 1967 ) ; h. haken , p. reineker , z. phys . * 249 * , 253 ( 1972 ) . | assessing the role of interference in natural and artificial quantum dyanamical processes is a crucial task in quantum information theory . to this aim ,
an appopriate formalism is provided by the decoherent histories framework .
while this approach has been deeply explored from different theoretical perspectives , it still lacks of a comprehensive set of tools able to concisely quantify the amount of coherence developed by a given dynamics . in this paper
we introduce and test different measures of the ( average ) coherence present in dissipative ( markovian ) quantum evolutions , at various time scales and for different levels of environmentally induced decoherence . in order to show the effectiveness of the introduced tools ,
we apply them to a paradigmatic quantum process where the role of coherence is being hotly debated : exciton transport in photosynthetic complexes .
to spot out the essential features that may determine the performance of the transport we focus on a relevant trimeric subunit of the fmo complex and we use a simplified ( haken - strobl ) model for the system - bath interaction .
our analysis illustrates how the high efficiency of environmentally assisted transport can be traced back to a _ quantum recoil avoiding effect _ on the exciton dynamics , that preserves and sustains the benefits of the initial fast quantum delocalization of the exciton over the network .
indeed , for intermediate levels of decoherence , the bath is seen to selectively kill the negative interference between different exciton pathways , while retaining the initial positive one .
the concepts and tools here developed show how the decoherent histories approach can be used to quantify the relation between coherence and efficiency in quantum dynamical processes . |
we prove that for random walks of non zero drift on the euclidian lattice , killed when leaving a convex two dimensional cone with vertex in @xmath0 , there are uncountably many positive harmonic functions .
the main assumption is finiteness of the jump generating function of the step of the random walk .
the proof is constructive and an adaptation of the similar proof in @xcite , which considers the special case of the positive quadrant .
we also make a conjecture about the martin boundary of such random walks and comment on the difficulties in translating the @xcite proof to the more general setting we are considering .
to begin , take @xmath1 to be a set of two points in @xmath2 , @xmath3 , so ordered that the angle @xmath4 between them is in @xmath5 .
the rays from ( 0,0 ) to infinity going through the @xmath6-sector between the two vectors in @xmath1 enclose a convex cone .
we will call the interior of such a cone @xmath7 .
it depends on the vectors we chose , i.e. @xmath8 .
we also note the unit vectors @xmath9 and @xmath10 , respectively perpendicular to @xmath11 and @xmath12 , pointing inwards .
see figure [ fig : f1 ] for a typical example .
we consider a random walk on the euclidean two dimensional lattice @xmath13 with step distribution @xmath14 which satisfies the following assumptions : a1 : : the homogeneous random walk @xmath15 is irreducible and has @xmath16 a2 : : the random walk killed when leaving @xmath7 is irreducible in @xmath7 .
a3 : : the jump generating function @xmath17 is finite everywhere in @xmath18 .
a4 : : @xmath19 is an aperiodic random walk on its respective lattice @xmath20
. .,title="fig:",scaledwidth=80.0% ] [ fig : f1 ] we note here that assumption * a3 * is indispensable for studying nontrivial cases of the martin boundary of random walks , killed when leaving cones in euclidean spaces . indeed , as @xcite proves : for a one dimensional random walk on the integers , with negative drift and such that * a3 * is not fulfilled , which is killed when leaving the set of nonnegative real numbers , there does nt exist any nonnegative nontrivial harmonic function .
we will denote @xmath21 for the measure describing the distribution of random walks started at @xmath22 , i.e. with @xmath23 . under these assumptions
it is well - known ( see @xcite and references therein ) , that @xmath24 .,title="fig:",scaledwidth=80.0% ] [ fig : f2 ] is a strictly convex and compact set , the gradient @xmath25 exists everywhere and does not vanish on @xmath26 , and the mapping @xmath27 is a homeomorphism between @xmath28 and @xmath29 . the inverse mapping is denoted by @xmath30 and we extend this map to nonzero @xmath31 by setting @xmath32 .
this definition implies that @xmath33 is the only point in @xmath28 where @xmath34 is normal to @xmath35 .
see figure [ fig : f2 ] for a typical picture of @xmath35 .
fixing a cone @xmath7 of the type described at the beginning and defining @xmath36 as well as the stopping time @xmath37 we want to prove the following .
[ prop : harm ] for every @xmath38 and @xmath39 @xmath40 , & \mbox{if } q(a ) = c_i,\text { } i = 1,2\\ \exp(a\cdot z ) - \mathbb{e}_z[\exp(a\cdot s(\tau ) ) , \tau < \infty ] , & \mbox{if } q(a ) \in int(\sigma ) \end{array}\right.\ ] ] are strictly positive and harmonic for the random walk , killed when leaving the cone .
these harmonic functions are just a generalization of the functions found in @xcite .
intuitively , a look at figure [ fig : f1 ] and at their paper suggests , that these functions must be the harmonic functions for the general case .
finally , one can see how the result in @xcite immediately follows from this proposition by taking @xmath41 and @xmath42 .
[ prop : harmq ] for every @xmath43 and @xmath44 @xmath45 , & \mbox{if } q(a ) = ( 0,1),\\ x_2 \exp(a\cdot z ) - \mathbb{e}_z [ s_2(\tau)\exp(a\cdot s(\tau ) ) , \tau < \infty ] , & \mbox{if } q(a ) = ( 1,0)\\\exp(a\cdot z ) - \mathbb{e}_z[\exp(a\cdot s(\tau ) ) , \tau < \infty ] , & \mbox{otherwise } \end{array}\right.\ ] ] are strictly positive and harmonic for the random walk , killed when leaving the positive quadrant .
the rest of the paper is organized as follows .
the next section states the natural conjecture about the martin boundary of random walk , killed when leaving a two - dimensional convex cone .
we also underline where the proof in @xcite , which considers only the positive quadrant , breaks down for the general case .
in the last section proposition [ prop : harm ] is proven by adapting the proof of proposition [ prop : harmq ] , contained in @xcite , to the general setting we are considering .
the first significant work on the martin boundary of random walks in euclidean lattices is @xcite , where the authors show that every positive harmonic function @xmath46 for the random walk can be expressed as @xmath47 here @xmath14 is a positive borel measure on some suitable set @xmath48 .
these types of functions and the types considered in remark [ remk : rmk1 ] of the next section are not harmonic for killed random walk on the quadrant .
to `` make '' them harmonic , one has to consider the correction term .
therefore the form of the functions in proposition [ prop : harmq ] .
the main contribution of @xcite is to show that these functions are the whole martin boundary for the case of the positive quadrant ( see theorem 1 there).judging from the analogy between proposition [ prop : harm ] and [ prop : harmq ] , one can conjecture the following ( stated analoguously to theorem 1 in @xcite ) .
[ [ conjecture ] ] * conjecture * + + + + + + + + + + + + for the cone encoded by @xmath11 and @xmath12 as in section [ sec : intro ] and under the assumptions * a1 - a4 * made there , we have that : 1 : : a sequence of points @xmath49 in @xmath7 with @xmath50 converge to a point of the martin boundary for the killed random walk when leaving the cone , if and only if @xmath51 for some @xmath52 .
2 : : the full martin compactification of @xmath53 is homeomorphic to the closure of the set @xmath54 in @xmath18 . in short , proposition
[ prop : harm ] characterizes fully the martin boundary of random walks on the two dimensional euclidean lattice , killed when leaving convex cones with vertex in zero .
if one tries to carry over the methods of @xcite to this general case , one sees that the _ communication condition _ contained there and the _ large deviations result _ can be modified to work for the more general setting as well .
we will not give details how this is done , but we mention shortly that both can be proven if one replaces assumption * a2 * by the following .
[ [ strong - local - irreducibility ] ] * `` strong local '' irreducibility : * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + there exists some uniform @xmath55 such that for every @xmath56 such that @xmath57 we have : there exists a path of measure non zero within @xmath58 from @xmath22 to @xmath59 .
this assumption seems to be neccessary , if one wants to work with the communication condition and is fulfilled in the positive quadrant setting due to irreducibility .
the actual obstacle for generalizing the proof in the case of the positive quadrant is the lack of markov - additivity for local processes for the general case .
we recall that a markov chain @xmath60 on a countable space @xmath61 is called _ markov - additive _ if for its transition matrix @xmath62 it holds : @xmath63 @xcite make extensive use of this property when showing the above conjecture for the case of the positive quadrant .
one idea for the general case of convex cones would be to look at local processes `` deep '' inside the cone , where the random walk is markov - additive in two directions . but
approaching the boundary of the cone , this property disappears in general in both directions . for the positive quadrant
this happens only for one direction and this is crucial for the proof in @xcite . without markov - additivity
it seems impossible to come to a usable ratio limit theorem as was done for the positive quadrant in @xcite . on the other hand ,
the proof of proposition [ prop : harm ] does not use markov - additivity .
this suggests that there should be more general methods than those of @xcite for proving the conjecture made in this section .
before starting with a series of lemmas , which will lead to the proof of proposition [ prop : harm ] we define for @xmath64 @xmath65 and @xmath66 then of course @xmath67 since @xmath68 .
finally , we introduce the family of twisted random walks @xmath69 with ( substochastic ) transition matrix @xmath70 and @xmath71 the respective exit time from @xmath7 .
we start the proof of proposition [ prop : harm ] by proving the following . from previous lemma ,
@xmath81 $ ] is finite in @xmath35 .
we also have @xmath82 on @xmath83 .
now fix an @xmath84 .
we have the existence ( recalling figure [ fig : f2 ] and the definition of the function @xmath30 ) of an @xmath85 small enough , such that for every @xmath86 there exist @xmath87 with @xmath88 .
then we have of course that @xmath89 on the event @xmath90 and therefore @xmath91\leq \mathbb{e}_z[e^{\tilde{a_1}\cdot s(\tau_1 ) } , \tau = \tau_1 < \infty ] + \mathbb{e}_z[e^{\tilde{a_2}\cdot s(\tau_2 ) } , \tau = \tau_2 < \infty]\\ & \leq \mathbb{e}_z[e^{\tilde{a_1}\cdot s(\tau ) } , \tau < \infty ] + \mathbb{e}_z[e^{\tilde{a_2}\cdot s(\tau ) } , \tau < \infty ] < \infty \end{split}\ ] ] due to previous lemma . [ remk : rmk1 ] for every @xmath92 and @xmath93 perpendicular to @xmath34 we have that @xmath94 is harmonic for the original random walk @xmath95 . indeed @xmath96 = \mathbb{e}_z[\tilde{q}\cdot
s(1 ) e^{a(q)\cdot s(1 ) } ] \\&= \mathbb{e}_z[\tilde{q}\cdot ( s(1)-z)e^{a(q)\cdot ( s(1)-z)+a(q)\cdot z } + \tilde{q}\cdot z e^{a(q)\cdot ( s(1)-z ) + a(q)\cdot z}]\\ & = e^{a(q)\cdot z}\tilde{q } \cdot ( \nabla \varphi ( a)|_{a = a(q ) } + z ) = f_q(z ) , \end{split}\ ] ] since @xmath97 for @xmath98 .
[ remk : rmk2 ] for @xmath80 and @xmath72 @xmath99 \\&= \mathbb{e}_z[e^{a\cdot ( s(\tau)-z))},\tau=\tau_2<\tau_1 ] e^{a\cdot z } \leq e^{a\cdot z } , \end{split}\ ] ] since the expectation in the second line is just @xmath100 with the same reasoning as in lemma [ lemma : l1 ] .
take w.l.o.g . @xmath104 and @xmath105 . then in the event that @xmath106 we have @xmath107 and @xmath108 .
also ( look again at figure [ fig : f2 ] ) for small enough @xmath109 there always exists some suitable @xmath110 so that @xmath111 lies in @xmath35 .
this yields @xmath112 \leq \frac{1}{\delta } \mathbb{e}_z [ e^{a(c_1)+\delta f_1\cdot s(\tau)},\tau=\tau_2<\tau_1]\\ & \leq \frac{1}{\delta}\mathbb{e}_z[e^{a(c_1 ) + ( \delta f_1 -\epsilon
f_2 ) s(\tau ) } , \tau = \tau_2 < \tau_1]\\ \end{split}\ ] ] since @xmath113 .
now the result follows from remark [ remk : rmk2 ] .
[ lemma : l4 ] for a random walk with jump @xmath114 of mean zero , @xmath115>0 $ ] and @xmath116<\infty$ ] and @xmath117 we have @xmath118
< \infty$ ] for @xmath119 .
of algebraic dimension at most 2 and this would suffice for our purposes , but the statement here is more general .
i thank dr .
wachtel for suggesting me this proof . ]
we define @xmath120 as the negative ladder heights of the random walk @xmath121 and look at @xmath122 where @xmath123 and @xmath124 we have @xmath125 < 0 $ ] and @xmath126 = \mathbb{e}[w_{\tau_0}]$ ] . using @xcite , more exactly theorem 2.1 .
there , we get @xmath127 = \int_{0}^{\infty } \mathbb{p}(\chi ( x ) > t ) dt \\ & \leq c \int_{0}^{\infty } \left(f[t , + \infty ) + \int_{t}^{t+x } f[u,+\infty)du\right)dt \\ & = c(1+x ) \int_{0}^{+\infty } f[u,+\infty)du , \end{split}\ ] ] where we have used tonelli in the second equality and @xmath128 and @xmath129 is the distribution function of @xmath130 .
now we are done if @xmath131<\infty$ ] . but
this is clear from the assumption @xmath116<\infty$ ] and results in @xcite .
take @xmath104 w.l.o.g . then @xmath134 = \mathbb{e}_z[|f_1 \cdot s(\tau)|e^{a(c_1)\cdot s(\tau)},\tau = \tau_2 <
\tau_1 ] \\ & + \mathbb{e}_z[|f_1 \cdot s(\tau)|e^{a(c_1)\cdot s(\tau)},\tau = \tau_1 < \infty ] \end{split}\ ] ] note that the first term in the sum above is finite due to corollary [ cor : cor1 ] .
the second one is smaller than @xmath135 = -\mathbb{e}_z[f_1\cdot s(\tau_1 ) e^{a(c_1)\cdot s(\tau_1)},\tau_1 < \infty]\ ] ] now we have that @xmath136 = f_1 \cdot \mathbb{e}_0[s(1)e^{a(c_1 ) \cdot
s(1)}]\\ & = f_1\cdot \nabla\varphi ( a)|_{a = a(c_1 ) } = f_1 \cdot c_1 = 0 \end{split}\ ] ] which means in short @xmath137 = 0\ ] ] now the random walk @xmath138 takes its values in the abelian subgroup @xmath139 of @xmath140 and due to our assumptions on the original random walk , it surely holds that @xmath141 < \infty\ ] ] with this and @xmath142 = \mathbb{e}_{f_1\cdot z}[|f_1\cdot s_a(\tau_{a1})|]\ ] ] we can use lemma [ lemma : l4 ] and finish the proof . for @xmath145 fixed and @xmath146
we have due to lemma [ lemma : l1 ] @xmath147 = \mathbb{p}_z(\tau_a <
\infty ) = 1\ ] ] since also @xmath148 = 0 $ ] i.e. the respective one dimensional random walk is recurrent with the same calculation as before .
furthermore for @xmath84 @xmath149 this means that @xmath150 , since not collinear to any of the two @xmath151-s .
the strong law of large numbers implies @xmath152 regardless of the starting point , so that there exists some @xmath153 and @xmath154 so that @xmath155 is contained in @xmath7 .
together with [ eq : slln ] this implies the existence of some @xmath156 such that for @xmath157 we have @xmath158 .
therefore @xmath159 is almost surely finite if @xmath160 .
for some fixed @xmath161 ( recall , then we have @xmath162 for @xmath132 ) we get with help of lemma [ lemma : l1 ] @xmath163 = \mathbb{p}_{\hat{z}}(\tau_a = \infty ) \\ & = \mathbb{p}_0 ( \min_{n \in \mathbb{n } } s_a ( n ) \cdot f_i > - \hat{z } \cdot f_i , i=1,2 ) > 0 \end{split}\ ] ] now we use * a2 * to get through the markov property for general @xmath39 @xmath164 = \mathbb{p}_z(\tau_a = \infty ) \\ & \geq \mathbb{p}_z(s_a(t ) = \hat{z } , \tau_a > t)\mathbb{p}_{\hat{z}}(\tau_a = \infty ) > 0 \end{split}\ ] ] if @xmath165 is chosen such that the first probability is not zero .
take w.l.o.g .
due to remark [ remk : rmk1 ] we have that @xmath167 is a martingale and by the stopping time theorem for every @xmath80 we have @xmath168 \\ & = \mathbb{e}_z[f_1 \cdot s(\tau\wedge n ) e^{a(c_1)\cdot s(\tau\wedge n ) } ] -\mathbb{e}_z[f_1 \cdot s(n ) e^{a(c_1)\cdot s(\tau ) } , \tau >
n ] \\ & = f_{c_1}(z ) - \mathbb{e}_z[f_1 \cdot s(n ) e^{a(c_1)\cdot s(\tau ) } , \tau > n ] \leq f_{c_1}(z ) \end{split}\ ] ] with the notation of remark [ remk : rmk1 ] .
now lemma [ lemma : l5 ] justifies dominated convergence and the result follows .
take first @xmath169 . by lemma
[ lemma : l6 ] @xmath170 is strictly positive in @xmath7 .
set @xmath171\ ] ] for @xmath172 one has @xmath173 which implies @xmath174 and with it @xmath175 = 0 $ ] .
for @xmath80 we have @xmath176 = \mathbb{e}_z\left[\mathbb{e}_{s(1)}[e^{a\cdot s(\tau)},\tau < \infty]\right ] \\ & = \mathbb{e}_z[e^{a\cdot s(1)},\tau = 1 ] + \mathbb{e}_z\left[\mathbb{e}_z [ e^{a\cdot s(\tau)},\tau<\infty|\mathcal{f}_1],\tau > 1 \right ] \\ & = f(z ) \end{split}\ ] ] as one can easily see .
but this means @xmath177 = f(z)$ ] , which also means that for @xmath178 , the equality @xmath179 = \mathbb{e}_z[h_a(s(1)),\tau>1 ] = h_a(z)$ ] holds .
here we have implicitly used that @xmath180=e^{a\cdot z } $ ] since @xmath181 . with this
, the case @xmath169 is solved .
take now w.l.o.g .
@xmath182 . we know from lemma [ lemma : l7 ] that @xmath170 is well - defined and nonnegative in @xmath7 .
take first @xmath183 .
then , it is clear that @xmath174 as is @xmath175 $ ] .
take now @xmath80 .
we have first @xmath184 = 0 $ ] and therefore @xmath185 = \mathbb{e}_z[h_a(s(1 ) ) ] \\&= f_1 \cdot e^{a \cdot z } -\mathbb{e}_z\left[\mathbb{e}_{s(1)}[f_1 \cdot s(\tau ) e^{f_1 \cdot s(\tau)},\tau < \infty]\right ] = h_a(z ) \end{split}\ ] ] since the second term in the sum after the second equality is equal to @xmath186 $ ] by the similar reasoning as in [ eq : ka ] . with this , harmonicity of @xmath170
is proved and it remains to show that @xmath170 is strictly positive in @xmath7 .
we have @xmath187\\ & - \mathbb{e}_z[f_1 \cdot s(\tau ) e^{a\cdot ( s(\tau)-z)},\tau = \tau_2 < \tau_1 < \infty ] = f_1\cdot z -a - b \end{split}\ ] ] where of course @xmath188 since @xmath80 . for @xmath189 and @xmath109
we have @xmath190 \\ & \leq \frac{1}{\delta}\mathbb{e}_z[e^{a\cdot ( s(\tau)-z)+ \delta f_1\cdot
s(\tau ) -\epsilon f_2\cdot s(\tau_2)},\tau = \tau_2<\tau_1 ] \end{split}\ ] ] where @xmath191 are chosen such that @xmath192 ( note that this is possible , see figure [ fig : f2 ] to get a grasp of this ) and therefore due to lemma [ lemma : l1 ] @xmath193e^{(\epsilon f_2 -\delta f_1)\cdot z } \leq \frac{1}{\delta } e^{(\epsilon f_2 -\delta f_1)\cdot z}\ ] ] note now that there exists some @xmath80 such that @xmath194 . fix such a @xmath22 and set @xmath195 and the respective @xmath189 and @xmath196 evaluated at @xmath49 with @xmath197 and @xmath198 .
there certainly exists @xmath199 such that @xmath200 .
now use * a2 * and harnack s classical inequality for non - negative harmonic functions , here @xmath201 to get the positivity result for all @xmath80 . | we prove the existence of uncountably many positive harmonic functions for random walks on the euclidean lattice with non - zero drift , killed when leaving two dimensional convex cones with vertex in @xmath0 .
our proof is an adaption of the proof for the positive quadrant from @xcite .
we also make the natural conjecture about the martin boundary for general convex cones in two dimensions .
this is still an open problem and here we only indicate where the proof technique for the positive quadrant breaks down . |
the idea of dependence logic , introduced in @xcite on the basis of hodges @xcite , is the following : if truth in the traditional sense , i.e. as tarski defined it , is defined with respect to a _ set _ of assignments , rather than just _ one _ assignment , it becomes possible to talk meaningfully about variables being _ dependent _ or _ independent _ from each other .
the set of assignments can be thought of as a _ data set _ , giving evidence much as in statistics about mutual dependencies between the variables . if the set of assignments the data is thought of as a _ database _ we arrive at the concept of _ functional dependency _ , an important concept in database theory since codd s pioneering paper @xcite .
we can also consider the set of assignments as expressing _ uncertainty _
about one true " assignment , as in _ inquisitive logic _ ( @xcite ) , or as indicating _ belief _ about an unknown assignment , as in _
doxastic logic _
( @xcite ) . finally , considering truth as given by sets of assignments leads naturally to the concept of the _ probability _ that a randomly chosen assignment ( from the given set ) satisfies a given propositional formula .
this idea is developed in @xcite to analyse so - called bell s inequalities of quantum physics .
following @xcite , we call sets of assignments _
teams_. teams have been previously used to study dependence concepts in predicate logic @xcite and modal logic @xcite .
we now focus this study to propositional logic .
the fundamental concept of dependence logic is the concept @xmath0 of a variable @xmath1 depending on a sequence @xmath2 of other variables , which is taken as a new atomic formula .
the meaning of such atomic dependence formulas is given via _
teams_. studying the logics of dependence concepts in propositional logic resembles the case of predicate logic in that we use the method of teams .
a _ team _ in this case is defined to be _ a set of valuations _ of propositional variables .
there are , however , also significant differences between the predicate logic and the propositional cases .
notably , propositional logics of dependence are _
decidable_. this is because for any given formula of the logics with @xmath3 propositional variables , there are in total @xmath4 valuations and @xmath5 teams .
the method of truth tables has its analogue in these logics , but the size of such tables grows exponentially faster than in the case of classical propositional logic , rendering it virtually inapplicable .
this emphasizes the role of the axioms and the completeness theorem in providing a manageable alternative for establishing logical consequence .
classical propositional logic is based on propositions of the form [ cols= " < " , ] if the @xmath6th symbol of a formula @xmath7 starts a string @xmath8 which is a subformula of @xmath7 , we denote the subformula by @xmath9_\phi$ ] , or simply @xmath9 $ ]
. we will sometimes refer to an occurrence of a formula @xmath10 inside a subformula @xmath8 of @xmath7 . in this case
, we decide to be sloppy about the notations and will use the same counting for the subformula @xmath8 .
we write @xmath11)$ ] for the formula obtained from @xmath7 by replacing the occurrence of the subformula @xmath12 $ ] by @xmath13 .
for example , for the formula @xmath14 , we denote the second occurrence of the dependence atom @xmath15 by @xmath16 $ ] , and the same notation also designates the occurrence of @xmath15 inside the formula @xmath17 . the notation @xmath18 $ ] designates the formula @xmath19 .
let us now present our natural deduction system of @xmath20 . as with the other logics considered in this paper ,
the consequence relation @xmath21 of @xmath20given by the system is _ not _ structural . in the following definition , only the symbols @xmath7 and @xmath22 stand for arbitrary formulas .
the propositional variables @xmath23 in the axioms or rules can not be substituted by arbitrary formulas .
[ pd_rules ] atomic excluded middle + + the rules @xmath24 , @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31and @xmath32from + dependence atom introduction + + dependence atom elimination + + + + strong elimination + + all rules that do not involve the intuitionistic disjunction @xmath33 in the natural deduction system of @xmath34 ( ) are included in the above system of @xmath20 . thus all clauses in that do not involve @xmath33 are derivable in the above system .
we have discussed in that the behavior of dependence atoms can be characterized completely using intuitionistic connectives : @xmath35 in view of the deduction theorem , these equivalences are equivalent to @xmath36 @xmath37 @xmath38 it is not hard to see that the induction and elimination rules for dependence atoms characterize the above clauses .
the strong elimination rule @xmath39characterizes the interaction between constancy dependence atoms and other formulas , and the rule @xmath40is a special case of the rule @xmath39 . for any formula @xmath7
, we have @xmath41)$ ] .
but as @xmath33 distributes over all connectives , this is reduced to @xmath42)\vee\phi(\neg p_i/[{\ensuremath{\mathop{=\!}}\xspace}(p_i),m])$ ] ( see the proof of the soundness theorem for details ) .
if the intuitionistic disjunction @xmath33 was a connective in our logic @xmath20 , then we could derive in its sound and complete deduction system using the usual introduction and elimination rules for @xmath33 that @xmath43)\vee\phi(\neg p_i/[{\ensuremath{\mathop{=\!}}\xspace}(p_i),m])$ ] .
then , we could derive @xmath22 from @xmath7 as follows : evidently every such derivation can be simulated in our deduction system using the rule @xmath39 . before we prove the soundness theorem , as a demonstration of the system ,
let us derive armstrong s axioms @xcite in the example below . to simplify notations
, we use @xmath44 as metalanguage variables that stand for arbitrary propositional variables .
[ armstrong_ax_pd ] the following clauses , known as armstrong s axioms , are derivable in the natural deduction system of @xmath20 .
( i ) : : @xmath45 ( ii ) : : @xmath46 ( iii ) : : @xmath47 ( iv ) : : @xmath48 ( v ) : : @xmath49 the derivations are as follows . for ( i ) : note that in the above derivation the formula @xmath50 is both the assumption and the conclusion of a sub - derivation . for ( ii ) : for ( iii ) : for ( iv ) : for ( v ) : next , we prove the soundness theorem for the deduction system of @xmath20 .
[ pdbor_soundness ] for any set @xmath51 of formulas in the language of @xmath20 , we have @xmath52 we show that @xmath53 holds for each derivation @xmath54 with conclusion @xmath7 and hypotheses from @xmath55 .
we only verify the case when the strong elimination rule @xmath39is applied .
the other cases are left to the reader . put @xmath56)$ ] and @xmath57)$ ] .
assume that @xmath58 and @xmath59 are derivations for @xmath60 , @xmath61 and @xmath62 , respectively .
we show that @xmath63 follows from the induction hypothesis @xmath64 , @xmath65 and @xmath66 .
this is reduced to showing that @xmath67 .
we prove by induction on subformulas @xmath8 of @xmath7 that @xmath68 holds .
the case when @xmath8 is an occurrence of an atom different from @xmath69 $ ] trivially holds . since @xmath70 , the case
@xmath71 $ ] also holds . if @xmath72 and without loss of generality we assume that the occurrence of the dependence atom @xmath69 $ ] is in the subformula @xmath22 , then by the induction hypothesis we have @xmath73 .
thus @xmath74 .
the case @xmath75 is proved similarly .
the remainder of this section is devoted to the proof of the completeness theorem for our system .
our argument makes heavy use of the notion of the _ realization _ of a formula
. we will give the definition of realizations in two steps .
firstly we define realizations for dependence atoms .
let @xmath76 be a dependence atom and @xmath77 .
we call a function @xmath78 a _ realizing function _ for @xmath79 , where we stipulate that @xmath80 . a _ realization _
@xmath81 of the dependence atom @xmath79 over @xmath82 is defined as @xmath83 by , @xmath79 is semantically equivalent to the intuitionistic disjunction of all its realizations , i.e. , @xmath84 .
the function @xmath82 in a realization @xmath85 characterizes one type of functional dependency between @xmath86 and @xmath87 , and thus the formula @xmath81 `` realizes '' such dependency . here the formula @xmath88 is not in the language of @xmath20 , as the intuitionistic disjunction @xmath33 is not an eligible connective in the logic @xmath20 .
nevertheless , we will show that in the deduction system of @xmath20one derives _ essentially _ , in a sense that becomes apparent below , the equivalence between @xmath79 and @xmath88 .
we first prove the easy direction that @xmath79 essentially follows from @xmath88 .
this , in our deduction system of @xmath20 , is simulated by the derivation that @xmath79 follows from each intuitionistic disjunct @xmath85 .
[ depezk_derive ] if @xmath82 is a realizing function for a dependence atom @xmath79 , then @xmath89 .
suppose that @xmath82 is a realizing function for @xmath76 .
we will show @xmath90 from which @xmath89 will follow by applying the rule @xmath91 . for all @xmath92
, we have @xmath93 . if @xmath94 and @xmath95 for some @xmath96 , then @xmath97 by the rule @xmath31and @xmath98 .
now , assume @xmath99 for all @xmath96 with @xmath94 .
the derivation is as follows : @xmath100 next , we prove ( [ depezk_derive_eq1 ] ) . by the claim , for all @xmath101 we have @xmath102
then , by the rule @xmath40 , we derive that for all @xmath101 , @xmath103 repeating this argument @xmath104 times , we obtain ( [ depezk_derive_eq1 ] ) .
this completes the proof .
the converse direction that @xmath79 essentially implies @xmath88 will be a special case of a general result ( ) in what follows .
now , we are ready to define realizations for arbitrary formulas @xmath7 .
let @xmath105,\,\dots,\,[\alpha_c , m_c]\rangle\ ] ] be a sequence of some occurrences of dependence atoms in @xmath7 .
a _ realizing sequence of @xmath7 over @xmath106 _ is a sequence @xmath107 such that each @xmath108 is a realizing function for @xmath109 .
we call the formula @xmath110 defined as follows a _ realization of @xmath7 over @xmath106 _ :
@xmath111,\dots,(\alpha_c)_{f_c}^\ast/[\alpha_c , m_c]).\ ] ] let @xmath112 be the sequence of all of the occurrences of dependence atoms in @xmath7 . a realizing sequence of @xmath7 over @xmath112 is called a _ maximal realizing sequence _ , and a realization @xmath113 over @xmath112
is called a _ complete realization _ of @xmath7 .
note that a complete realization @xmath113 of a formula @xmath7 in the language of @xmath20is a classical formula .
for example , consider the formula @xmath14 that we discussed earlier .
let @xmath114 be two realizing functions for @xmath15 . over @xmath115\rangle$ ]
the sequence @xmath116 is a realizing sequence of @xmath7 and the formula @xmath117 is a realization of @xmath7 .
both @xmath118 and @xmath119 are complete realizations of @xmath7 . note that if a formula does not contain dependence atoms , then its maximal realizing sequence is the ( unique ) empty sequence @xmath120 .
having defined the crucial notion for the proof of the completeness theorem , we now check in the next lemma that every formula is semantically equivalent to the intuitionistic disjunction of its realizations over some @xmath106 . in the proof of our main theorem ( )
, we are only interested in complete realizations , but the lemmas leading to this theorem will be stated in their most general forms . [ wnf_pd_sound ] let @xmath7 be a formula in the language of @xmath20and @xmath121 the set of all its realizing sequences over a sequence @xmath106 .
then @xmath122 we prove the lemma by induction on @xmath7 . the case that @xmath7 is a dependence atom follows from .
the case that @xmath7 is some other atom is trivial .
the induction case @xmath123 follows from the induction hypothesis and the fact that @xmath124 \longrightarrow a{\ensuremath{\otimes}\xspace}b\models a'{\ensuremath{\otimes}\xspace}b'$ ] and that @xmath125 .
similarly for the case @xmath126 .
a crucial fact to be used in the proof of the completeness theorem is that a formula @xmath7 is essentially provably equivalent to @xmath127 .
we first prove the easy direction that @xmath7 follows essentially from @xmath128 .
as with realizations for dependence atoms , we will show , in the absence of the intuitionistic disjunction , that each realization of @xmath7 implies @xmath7 .
[ si_derive ] if @xmath129 is a realizing sequence of a formula @xmath7 over a sequence of some occurrences of dependence atoms in @xmath7 , then @xmath130 .
we prove the lemma by induction on @xmath7 .
the case that @xmath7 is a dependence atom follows from .
the case that @xmath7 is some other atom is trivial .
the induction case @xmath123 is proved by applying the rule @xmath28and the case @xmath126 is proved by applying the rule @xmath131 .
we now turn to the converse direction that @xmath7 essentially implies @xmath127 .
let us ponder on what we should settle here in the logic @xmath20 in the absence of the intuitionistic disjunction .
assume that the intuitionistic disjunction was a connective in our logic and we have proved that @xmath132 .
if we have a derivation of @xmath7 , then to derive a formula @xmath22 from the same assumptions , it is sufficient to prove that @xmath22 follows from each realization @xmath113 of @xmath7 .
our idea is to discard the intermediate step that involves the intuitionistic disjunction and conclude @xmath22 directly from @xmath7 as soon as the fact that @xmath22 follows from each @xmath113 is settled . to be precise , we will prove the lemma below .
[ sek_derive]let @xmath121 be the set of all realizing sequences of @xmath7 over a sequence @xmath106 of some occurrences of dependence atoms in a formula @xmath7 .
let @xmath133 be a collection of sets of formulas .
for any formula @xmath22 , if @xmath134 then @xmath135 .
we will now work towards a proof of the above lemma .
clearly , if we consider only a single occurrence of a constancy dependence atom in a formula , i.e. , if @xmath136\rangle$ ] , then the implication in the lemma follows immediately by applying the rule @xmath39 .
the idea is then to reduce the general case to this special case in a certain way .
first we show that dependence atoms with multiple arguments can be reduced to formulas with constancy dependence atoms only . if @xmath76 , we write @xmath137 where @xmath77 . by ,
@xmath79 and @xmath138 are semantically equivalent .
let @xmath139,\,\dots,\,[\alpha_c , m_c]\rangle$ ] be a sequence of some occurrences of dependence atoms in a formula @xmath7 and let @xmath140,\dots,\alpha_c^\ast/[\alpha_c , m_c]).\ ] ] [ phi_derive_phiast ] let @xmath7 and @xmath106 be as above
. then @xmath141 .
we prove the lemma by induction on @xmath7 .
if @xmath142 and @xmath77 , then we first derive @xmath143 as follows : @xmath144 next , for each @xmath145 , by the rules @xmath146and @xmath147 , we have @xmath148 hence @xmath149 follows from ( [ pd_nf_derive_eq3 ] ) and @xmath28 . if @xmath150 , then the statement is trivial .
the induction steps can be derived using the rules @xmath24and @xmath28 . in the next lemma
we derive the special case of in which the sequence @xmath106 consists of occurrences of constancy dependence atoms only . for a constancy dependence atom @xmath151
there are only two realizing functions , namely the constant functions @xmath152 and @xmath153 defined as @xmath154 and @xmath155 . and @xmath156 and @xmath157 .
a realizing sequence @xmath158 over a sequence @xmath106 of occurrences of constancy dependence atoms is a sequence of constant functions @xmath159 and @xmath160 , and can , therefore , be characterized by a function @xmath161 defined as @xmath162 .
we identify the sequence @xmath129 with the function @xmath163 and sometimes write @xmath164 instead of @xmath165 .
[ sezk_derive ] for each function @xmath166 , let @xmath167 be a set of formulas .
for any formulas @xmath7 and @xmath22 , if @xmath168 holds for all @xmath166 , i.e. , @xmath169,\dots , p_{j_k}^{g(k)}/[{\ensuremath{\mathop{=\!}}\xspace}(p_{j_k}),m_k])\vdash\theta,\ ] ] then @xmath170 . by the assumption , for all
@xmath171 we have @xmath172,\dots , p_{j_k}^{h_1(k)}/[{\ensuremath{\mathop{=\!}}\xspace}(p_{j_k}),m_k])\vdash\theta\ ] ] and @xmath173,\dots , p_{j_k}^{h_1(k)}/[{\ensuremath{\mathop{=\!}}\xspace}(p_{j_k}),m_k])\vdash\theta,\ ] ] where @xmath174 is the function @xmath175 and @xmath176 is the function @xmath177 . then by the rule @xmath39we derive for all @xmath171 , @xmath178,\dots , p_{j_k}^{h_1(k)}/[{\ensuremath{\mathop{=\!}}\xspace}(p_{j_k}),m_k])\vdash\theta.\ ] ] repeating this argument @xmath104 times , we obtain @xmath179 as required .
having analyzed the simple realizations that involve constancy dependence atoms only , let us now give the proof of .
our main argument is to apply to transform dependence atoms with multiple arguments to formulas with constancy dependence atoms only so as to apply to finish the proof .
let @xmath180,\dots,[\alpha_c , m_c ] \rangle$ ] , @xmath181 and @xmath182 for any @xmath183 . in view of , to prove the lemma it suffices to show that the assumption ( [ sek_derive_eq1 ] ) implies @xmath184 . for each @xmath185 ,
let @xmath186 \rangle_{s\in 2^{k_i}}$ ] be the sequence of all occurrences of the ( constancy ) dependence atoms in the subformula @xmath187 of @xmath188 .
each function @xmath189 induces a realizing sequence of @xmath188 over the sequence @xmath190 of occurrences of constancy dependence atoms . for each @xmath191 , define a function @xmath192 as @xmath193 observe that @xmath194 is a realizing function for @xmath109 and @xmath195 is a realizing sequence of @xmath7 over @xmath106 .
@xmath196 . for each @xmath183
, we have @xmath197 induces a realizing sequence of @xmath198 over @xmath199 and @xmath200 now , by ( [ sek_derive_eq1 ] ) we have @xmath201,\dots,(\alpha_c)_{f_{g , c}}^\ast/[\alpha_c , m_c])\vdash\theta,\ ] ] which is @xmath202,\dots,(\alpha_c^\ast)_{g}^\ast/[\alpha_c , m_c])\vdash\theta,\text { i.e. , } \pi_{\omega_g},(\phi^\ast_o)^\ast_{g}\vdash\theta.\ ] ] put @xmath203 . by the claim and , we obtain that @xmath204 .
it is not hard to see that there is a one - to - one correspondence between functions @xmath205 and realizing sequences @xmath129 of @xmath7 over @xmath106 .
this means that @xmath206 , which completes the proof .
now , we are in a position to prove the ( strong ) completeness theorem for the deduction system of @xmath20 .
[ pd_completeness ] for any set @xmath51 of formulas in the language of @xmath20 , we have @xmath207 by the compactness theorem ( ) we may assume that @xmath55 is a finite set and @xmath208 .
suppose @xmath209 . by lemma [ wnf_pd_sound ]
we have @xmath210 where @xmath121 and @xmath211 are the ( nonempty ) sets of all maximal realizing sequences of @xmath8 and @xmath7 , respectively . for each @xmath212
, we have @xmath213 .
since @xmath129 and @xmath214 are maximal realizing sequences , @xmath215 and @xmath216 are classical formulas . by the standard argument
these two classical formulas can be transformed into formulas in the classical disjunctive normal form in the deduction system of @xmath217and therefore by also in the deduction system of @xmath20 , i.e. , @xmath218 where @xmath219 , @xmath220 are some teams on @xmath221 , and the formula @xmath222 is defined as in lemma [ x_thetax ] . by the soundness theorem we obtain that @xmath223 .
now by the same argument as that in the proof of one derives that @xmath224 for some @xmath225 , which implies that @xmath226 by ( [ pd_cmpl_eq1 ] ) .
hence we obtain by that @xmath227 for all @xmath212 , which by yields @xmath228 as desired .
we have shown that there is a robust expressively complete propositional dependence logic for reasoning about dependencies on the propositional level .
this logic can be fairly called _ robust _ because it has five different expressively equivalent formulations : @xmath20 , @xmath34 , @xmath229 , @xmath230and @xmath231 , and each of these different formulations has its complete axiomatization .
this logic adopts the team semantics , which extends the single - valuation semantics of classical propositional logic @xmath217 in a conservative manner : the team semantics reduces to the single - valuation semantics for formulas in the language of @xmath217 , and over singleton teams all connectives of the logic behave classically . yet the move to _ teams _ allows us to meaningfully characterize the dependencies between propositions , which can not be expressed in @xmath217 .
also , two non - classical connectives ( the intuitionistic disjunction and implication ) are identified in the team semantics setting .
exactly with these different connectives ( and atoms ) we formed our different formulations of the logic .
these five formulations of the same logic are well - behaved also in the sense that all of them have disjunctive or conjunctive normal forms that resemble the same normal forms in @xmath217 . making essential use of the disjunctive normal form of the logics , @xcite proved the craig s interpolation theorem for all of the five logics , and @xcite showed that the logics are _ structurally complete _ with respect to flat substitutions ( i.e. , the admissible rules of these logics with respect to flat substitutions are derivable ) .
the study of dependence in logic originates from the first - order context .
when quantifiers are added to our logics , many nice properties obtained in this paper fail immediately . as already noted
, first - order dependence logic is not axiomatizable , it is much weaker in expressive power than first - order intuitionistic dependence logic , and the equivalence given in does not apply to first - order dependence atoms . the results in this paper show that the dependence notion on the propositional level is simple "
this is justified by and the simplicity of the sound and complete deduction systems we defined in .
we finish by mentioning some relevant related areas .
one area is the study of the computational complexity of propositional logics of dependence . a downward closed formula @xmath7 being satisfiable by some team is equivalent to @xmath7 being satisfiable by some singleton team .
since over singleton teams dependence atoms are always true and the logics have the same semantics as classical propositional logic @xmath217 , the satisfiability problem ( sat ) for @xmath231and its fragments has the same complexity as sat for @xmath217 , which is well - known to be np - complete . as for the model checking problem ( mc ) , @xcite shows that mc for @xmath20and @xmath34are np - complete , and it follows from @xcite that mc for @xmath229and @xmath230are conp - complete .
@xmath20is closely related to _ dependency quantified boolean formulas _ , whose validity problem is shown in @xcite to be nexptime - complete .
on the basis of this , @xcite showed that the computational complexity of the validity problem of @xmath20is nexptime - complete . adding modalities to propositional dependence logic , one obtains _
modal dependence logic _ , which was introduced in @xcite .
this field has been very active in recent years .
another recent development is the rise of _ independence logic _ ( introduced in @xcite ) , which is , _ a priori _ , stronger than dependence logic .
the authors have a paper on _ propositional independence logic _ under preparation .
the authors would like to thank ivano ciardelli , pietro galliani , lauri hella , taneli huuskonen , rosalie iemhoff , dick de jongh , juha kontinen , tadeusz litak , floris roelofsen and dag westersthl for inspiring conversations related to this paper .
we also thank the anonymous referee for reading the paper carefully and making many useful suggestions , which led to a great improvement in the presentation of our results .
10 dependency as question entailment . in _
dependence logic : theory and application _ ,
h. v. s. abramsky , j. kontinen and j. vnnen , eds . , progress in computer science and applied logic .
birkhauser , to appear .
complexity of model checking for modal dependence logic . in _ sofsem _ , m. bielikov , g. friedrich , g. gottlob , s. katzenbeisser , and g. turn , eds .
7147 of _ lecture notes in computer science_. springer , 2012 , pp . 226237 .
model checking for modal intuitionistic dependence logic . in _ post - proceedings of the ninth tbilisi symposium on language , logic and computation _ ( 2013 ) , vol .
lecture notes in computer science _ , pp . 231256 .
the doxastic interpretation of team semantics . in _ logic without borders : essays on set theory , model theory , philosophical logic and philosophy of mathematics _ ,
a. villaveces , r. kossak , j. kontinen , and .
hirvonen , eds .
de gruyter , 2015 , pp .
167192 .
inquisitive semantics and pragmatics . in _ meaning , content , and
argument : proceedings of the ilcli international workshop on semantics , pragmatics , and rhetoric _
, j. m. larrazabal and l. zubeldia , eds .
university of the basque country publication service , may 2009 , pp .
4172 . the expressive power of modal dependence logic . in
_ advances in modal logic 10 , invited and contributed papers from the tenth conference on advances in modal logic _ ( 2014 ) , college publications , london , pp .
. some strange quantifiers . in _ structures in logic and computer science : a selection of essays in honor of a. ehrenfeucht _ , j. mycielski , g. rozenberg , and a. salomaa , eds .
lecture notes in computer science_. london : springer , 1997 , pp . 5165 . axiomatizing propositional dependence logics . in
24th eacsl annual conference on computer science logic _ ( 2015 ) , vol .
41 of _ leibniz international proceedings in informatics _ , schloss dagstuhl - leibniz - zentrum fuer informatik , pp . 292307 .
complexity of validity for propositional dependence logics . in _
proceedings fifth international symposium on games , automata , logics and formal verification , gandalf 2014 _ ( 2014 ) , no . 161 in eptcs , pp . | in this paper , we study logics of dependence on the propositional level . we prove that several interesting propositional logics of dependence , including _ propositional dependence logic _ , _ propositional intuitionistic dependence logic _ as well as _
propositional inquisitive logic _ , are expressively complete and have disjunctive or conjunctive normal forms .
we provide deduction systems and prove the completeness theorems for these logics . propositional dependence logic , inquisitive logic , team semantics , non - classical logic 03b60 |
the population of known extrasolar planets which transit the face of their parent stars has been growing steadily in recent years .
it currently includes @xcite , ogle - tr-56b @xcite , ogle - tr-113b @xcite , ogle - tr-132b @xcite , tres-1 @xcite , ogle - tr-111b @xcite and ogle - tr10b @xcite . for , the transit lightcurve was observed with the _ hubble space telescope ( hst ) _ to be symmetric around its centroid to an exquisite photometric precision of @xmath7 magnitude per data point for a few hundred data points tracing both ingress and egress @xcite . in this _
letter _ we show that transit lightcurves must have a time - reversal asymmetry even if the star and the planet are perfectly symmetric and the orbit is circular .
the asymmetry originates from the persistent acceleration of the planet towards the star during the transit . for a circular orbit
, the planet moves towards the observer at the beginning of the transit ( ingress ) and away from the observer at its end ( egress ) .
this net acceleration introduces a change in the relative rate by which ingress and egress advance in the observer s frame of reference .
the effect simply follows from the unsteady change in the propagation delay of photons which distorts the transformation of time between the planet and the observer .
the variation rate of the delay is changing most rapidly at the middle of the transit when the full newtonian acceleration vector points straight along the line - of - sight . in the following sections , we derive the magnitude of the resulting lightcurve asymmetry ( 2 ) and discuss its implications ( 3 ) .
the observed arrival time of photons @xmath8 is given by the time they left the planet @xmath9 plus the propagation delay over the distance of the planet @xmath10 , @xmath11 the @xmath9-derivative of equation ( [ eq : doppler ] ) gives @xmath12 where @xmath13 is the velocity of the planet along the line - of - sight .
since @xmath14 changes from negative to positive during the transit of a planet in a circular orbit , it is now obvious that ingress would advance at a faster rate ( i.e. a shorter @xmath15 per @xmath16 interval ) than egress .
figure 1 illustrates schematically this generic behaviour .
note that equation ( [ eq : deriv ] ) is accurate to all orders in @xmath17 but we will keep only first order terms in subsequent derivations .
even though the planet is not emitting the observed photons , the timing of its occultation is dictated by the instant at which stellar photons graze its outer surface ; this timing is distorted by the same propagation delay effect that would exist if the grazing photons were emitted by the planet itself rather than the star since the propagation history of the photons before they graze the planet s boundary is irrelevant .
the transverse spatial coordinates of the planet orbit , @xmath18 , are the same in the observer and planet reference frames .
the apparent transverse velocity of the planet is therefore different from its actual transverse velocity , @xmath19 @xcite , @xmath20 where the transverse position vector @xmath18 corresponds to angular coordinates on the sky times the distance to the planetary system . throughout our discussion ,
the terms parallel ( @xmath21 ) or transverse ( @xmath22 ) are relative to the line - of - sight axis that starts at the observer and goes through the center of the stellar image .
we focus on the time when the occulting planet crosses this spatial point of symmetry . taking the @xmath8derivative of both sides of equation ( [ eq : velocity ] ) and keeping terms to leading - order in @xmath23 , we get two new contributions to the difference between the observed and keplerian values of the transverse acceleration of the planet across the sky @xcite , @xmath24 where @xmath25 is the observed acceleration and @xmath26 is the actual keplerian acceleration of the planet .
the last term on the right - hand - side of equation ( [ eq : acceleration ] ) implies that the apparent _ transverse _ acceleration of the planet gets a contribution from its keplerian acceleration _ along _ the line - of - sight , @xmath27 .
this term is the source of our effect . for simplicity ,
we assume that the orbital plane is viewed edge - on by the observer ; this geometry is exact for transits that cross the center of the star ( `` central transits '' ) and is a very good approximation more generally as long as the orbital radius is much larger than the radius of the star . for a circular planetary orbit ,
newtonian dynamics implies no transverse acceleration of the planet at the spatial center of symmetry of the transit where @xmath28 here @xmath2 is the stellar mass and @xmath3 is the planet s orbital radius .
however , equation ( [ eq : acceleration ] ) implies that at the same time there would be an apparent transverse deceleration of the planet in the observer s frame of reference @xmath29 which is directed opposite to @xmath30 .
this persistent apparent deceleration implies that the planet would cross ingress and egress at different apparent rates .
thus , even if the stellar image is perfectly symmetric , the orbit is circular and the planet is perfectly spherical , there is an inherent asymmetry in the transit lightcurve due to the nearly steady value of @xmath31 during the transit .
newtonian dynamics alone predicts a non - vanishing @xmath32 as soon as the planet moves away from the transit center , but for a circular orbit this deviation would maintain the time - reversal symmetry between ingress and egress .
the fractional asymmetry in the transit lightcurve is of order the fractional change in @xmath33 over the transit duration , since the rates by which ingress and egress proceed are proportional to @xmath33 . for
a total transit duration @xmath34 and an orbital period @xmath35 , the slope of the initial drop and final rise in the transit lightcurve will differ by a fractional amplitude @xmath36 for a `` hot jupiter '' like , the ingress phase ( @xmath37 ) would proceed at a rate that is higher by fractional amplitude of @xmath7 than the egress phase ( @xmath38 ) .
planets that are closer in by a factor of a few could produce an asymmetry of up to @xmath39 ; ogle - tr-56b for which @xmath40 provides @xmath41 [ see @xcite for a compilation of all known transit systems and the prospects for detecting others ] .
the observed time @xmath8 is a slightly distorted version of the time axis @xmath9 along which the lightcurve is symmetric .
we may write @xmath42 , where @xmath43 and we shifted @xmath44 to be at the transit centroid .
the observed photon flux , @xmath45 , corresponds to the flux at an undistorted time @xmath46 $ ] , while @xmath47 corresponds to @xmath48=-t_{\rm obs}[1+\epsilon(t)]$ ] . for small deviations
, we may expand the photon flux as a function of time to leading order , @xmath49 . since @xmath50 , @xmath51 , and @xmath52 , the time - dependent photometric asymmetry of the lightcurve is given by @xmath53 where we have set @xmath54 at the transit centroid ( for which @xmath55 ) and kept terms to leading order in @xmath56 .
given a preliminary transit lightcurve , @xmath45 , and an approximate knowledge of the system parameters , it is possible to predict the lightcurve asymmetry under the assumption of a circular orbit .
figure 2 shows the expected @xmath57 for based on the lightcurve data in @xcite .
for a transit depth of @xmath58 magnitudes produced by a close - in planet , the net photometric asymmetry would typically amount to a few percent of @xmath59 or equivalently to a peak value of @xmath60 in the range of ( 0.11)@xmath61 .
the photometric sensitivity of @xmath7 magnitudes per data point achieved with _ hst _ for @xcite provides a net sensitivity to a time - reversal asymmetry of @xmath62 , where @xmath63 is the total number of independent data points during ingress and egress .
our effect appears to be on the borderline of being detectable with existing techniques .
a statistically significant ( @xmath64 ) detection of the asymmetry in equation ( [ eq : min ] ) for would require an additional order of magnitude improvement in sensitivity , which is achievable with future space missions such as _ kepler _ or _ corot _ and only marginally with the existing instrument _
most_. a simple way to reach the required sensitivity would be to maintain the existing photometric precision per data point and increase the number of data points from @xmath65 up to @xmath66 ( e.g. by increasing the number of transits from the 4 observed by _ hst _ to @xmath67 ) over a period of several years .
the orbital period scales as @xmath5 while the fractional asymmetry in equation ( [ eq : min ] ) scales as @xmath68 .
thus , detection of the lightcurve asymmetry would allow to determine @xmath2 and @xmath3 separately .
additional data on the reflex motion of the star would then provide the planet mass , @xmath69 , based on dynamical information alone .
currently , the stellar mass , @xmath2 , is inferred indirectly based on theoretical fitting of the stellar spectrum which is subject to modeling uncertainties at the level of @xmath70 [ see fig . 3 in @xcite ] . if the planetary orbit is nearly circular but has a finite eccentricity , @xmath71 , then it is easy to show that at the transit center @xmath72 to leading order in @xmath73 , where @xmath74 is the angle between the shortest radius vector of the orbital ellipse ( along the apsidal line ) and the line to the observer .
as long as @xmath75 , the eccentricity effect is smaller than the propagation delay effect in equation ( [ eq : dopa ] ) . in the absence of perturbers ,
any initial orbital eccentricity is expected to decay exponentially through tidal dissipation on an e folding timescale @xcite @xmath76 where the quality parameter @xmath77 ( @xmath78 for jupiter ) is inversely proportional to the dissipation rate in the planet s interior @xcite , and @xmath79 is the planet s radius .
aside from the unknown @xmath77 , the parameter values in equations ( [ eq : min ] ) and ( [ eq : dis ] ) were chosen to match .
since the orbital circularization timescale is shorter by up to two orders of magnitude than the age of planetary systems such as hd 209458 [ @xmath80 gyr , see @xcite and @xcite ] , the eccentricity - driven asymmetry is expected to diminish for a close - in planet unless a third body pumps its orbital eccentricity .
incidentally , such a perturber was hypothesized as the driver of the inflated radius of @xcite but its existence has not been demonstrated yet . for close - in planets such as
, our effect is larger than a different source of lightcurve asymmetry that was already discussed in the literature , namely the planet s obliquity @xcite .
this is because close - in planets are expected to rotate slowly and possess a small projected obliquity as a result of strong tidal locking with their orbital revolution .
their rotation axis is expected to be normal to their orbital plane which is viewed nearly edge - on [ in the case of , the orbital inclination angle is @xmath81 @xcite ] .
another source of asymmetry is rotation of the star .
typical projected rotation speeds of @xmath82 @xcite would produce a doppler offset @xcite between the flux emitted by the approaching and receding sides of the star of order @xmath83 [ the photometric analog of the _ rossiter - mclaughlin effect _ ; see @xcite ] , or an occultation contrast of @xmath84 magnitudes which is much smaller than the effect considered here .
if the stellar rotation axis is significantly misaligned with the normal of the orbital plane and the transit is not central , then the rotation - induced oblateness of the stellar image could generate a fractional ( @xmath59equivalent ) asymmetry of less than @xmath85 , which is again well below our effect for where the misalignment angle must be small @xcite .
the propagation delay relation between @xmath8 and @xmath9 in equation ( [ eq : doppler ] ) can be easily incorporated into a computer program that searches for the best - fit keplerian orbit under the constraints of a given data set .
the _ delay
corrected _ keplerian fit would provide new dynamical constraints on the planetary system .
such a fit would involve the same number of free parameters as the standard keplerian fit .
we have shown that the propagation delay of light introduces an _ apparent _ transverse deceleration of a planet on the sky during its transit across the face of its parent star ( eq .
[ eq : dopa ] ) .
this persistent deceleration breaks the time - reversal symmetry of the transit lightcurve for a spherical planet in a circular orbit around a spherical star . throughout the transit
, the planet velocity along the line - of - sight , @xmath14 , is changing at a nearly steady rate , @xmath86 .
this produces a steady change in the transformation of time to the observer s frame ( see eq .
[ eq : deriv ] ) .
the net change in @xmath14 between ingress and egress , @xmath87 , introduces a fractional difference of magnitude @xmath88 in the slopes of their lightcurves .
it is possible to search for this difference in slopes by folding the lightcurve over its centroid . for close - in planets
, ingress should typically proceed at a rate that is faster by a fractional amplitude @xmath89@xmath1 than egress ( see eq .
[ eq : min ] ) .
this level of asymmetry will be detectable with forthcoming space telescopes such as _ corot _ or _
@xcite , which are scheduled for launch within 2 - 3 years .
because this asymmetry has a unique scaling with stellar mass and orbital radius ( @xmath90 ) , its detection together with the reflex motion of the star will allow to determine the star and planet masses as well as the orbital radius using purely dynamical data .
this method evades the theoretical uncertainties inherent in the existing approach for determining stellar masses based on modeling of spectroscopic data [ e.g. @xcite ] .
orbital eccentricity could induce a stronger asymmetry in the lightcurve but is expected to decay exponentially to negligible levels through tidal dissipation for close - in planets like , unless it is being pumped by the gravitational perturbation of another planet .
similarly to other transit timing residuals @xcite , our effect would be contaminated by noise from inhomogeneities on the face of the star and would compete against other small effects involving asymmetries from the oblateness of the planet @xcite or the rotation of the star .
additional special - relativistic or general - relativistic effects are of the order of @xmath91 or @xmath92 , or smaller , where @xmath93 is the gravitational potential produced by the star ( @xmath94 ) ; these corrections are orders of magnitude smaller than the propagation delay effect discussed here .
finally , we note that a change of the opposite sign in @xmath14 occurs when the planet goes behind the star . in this case
, ingress would be slower than egress .
there is no net change in the orbital period over a full closed orbit .
however when the planet enters its own ( secondary ) eclipse by the star , the photons it emits will be delayed relative to primary eclipse ( the transit ) by the difference in emission times plus the light travel time across the orbital diameter , as implied by equation ( [ eq : doppler ] ) . for a circular orbit , the time interval between the centroids of the primary and secondary
eclipses would be longer by @xmath95 than half of the full orbital period .
future detection of a planet s infrared emission would then allow to determine the orbital radius from the light propagation delay of @xmath96 seconds .
the full orbital period @xmath97 would then yield the stellar mass through @xmath98 $ ] .
an eccentricity could change the lightcurve history of the illuminated planet and in particular make the time from primary eclipse to secondary eclipse different from the time back to primary eclipse .
i thank david charbonneau , scott gaudi , matt holman , dimitar sasselov , george rybicki , and josh winn for helpful discussions .
this work was supported in part by nasa grant nag 5 - 13292 and nsf grants ast-0071019 and ast-0204514 . | as a planet transits the face of a star , it accelerates along the line - of - sight . the changing delay in the propagation of photons produces an apparent deceleration of the planet across the sky throughout the transit .
this persistent transverse deceleration breaks the time - reversal symmetry in the transit lightcurve of a spherical planet in a circular orbit around a perfectly symmetric star .
for `` hot jupiter '' systems , ingress advances at a higher rate than egress by a fraction @xmath0@xmath1 .
forthcoming space telescopes such as _ kepler _ or _ corot _ will reach the sensitivity required to detect this asymmetry . the scaling of the fractional asymmetry with stellar mass @xmath2 and planetary orbital radius @xmath3 as @xmath4 is different from that of the orbital period as @xmath5 .
therefore , this effect constitutes a new method for a purely dynamical determination of the mass of the star , which is currently inferred indirectly with theoretical uncertainties based on spectral modeling .
radial velocity data for the reflex motion of the star can then be used to determine the planet s mass .
although orbital eccentricity could introduce a larger asymmetry than the light propagation delay , the eccentricity is expected to decay by tidal dissipation to negligible values for a close - in planet with no perturbing third body .
future detection of the eclipse of a planet s emission by its star could be used to measure the light propagation delay across the orbital diameter , @xmath6 seconds , and also determine the stellar mass from the orbital period . |
in this note we address some features of the limiting behavior of the minimizers of a class of second - order singular perturbation energies . the model we analyze
was introduced in @xmath15 by coleman and mizel in the context of the theory of second - order materials and was then studied in @xcite in collaboration with marcus .
coleman and mizel proposed a model for nonlinear materials in which the free energy depends on both first and second order spatial derivatives of the mass density . in this way they expected to prove the occurrence of layered structures of the ground states ( as observed in concentrated soap solutions and metallic alloys ) without appealing to non - local energies ( such as , for example , the otha - kawasaki functional @xcite ) .
specifically , they introduced the free - energy functional @xmath16 $ ] given by @xmath17 where @xmath18 ( the mass density ) is the order parameter of the system , @xmath19 and @xmath4 is a double - well potential with two potential wells of level zero at @xmath5 .
as @xmath0 goes to zero , the functional accounts for the energy stored by a one - dimensional physical system occupying the bounded open interval @xmath20 .
this model can be viewed as a scaled second - order landau expansion of the classical cahn - hillard model for sharp phase transition ; _ i.e. , _
@xmath21 for the cahn - hillard model the lack of uniqueness is usually solved in the context of first - order gradient theory of phase transitions considering the simplest diffuse phase transition model ; _ i.e. , _ the van der waals model .
the latter is obtained by adding a singular gradient perturbation to the previous functional . after scaling , the new functional @xmath22
$ ] reads as @xmath23 if @xmath8 grows at least linearly at infinity , modica and mortola @xcite proved that sequences @xmath24 with equi - bounded energy ( _ i.e. _ such that @xmath25 ) can not oscillate as , up to subsequences , they converge in @xmath26 to a function @xmath27 .
moreover , the @xmath9-limit of @xmath28 is given by @xmath29 for a suitable constant @xmath30 depending on the double - well potential @xmath8 .
the above phenomenon characterizes first order phase transitions of every material having positive surface energy .
on the other hand , in nature there are materials that relieve energy whenever the measure of their surface is increased .
these materials have a so - called negative surface energy . to give a mathematical description of this kind of materials within the framework of the gradient theory of phase transitions , coleman and mizel introduced the energy @xmath14 .
the requirement for an energy of the form of @xmath14 to be bounded from below forces the coefficient in front of the highest gradient squared to be nonnegative . on the other hand
different phenomena can occur depending on the coefficient @xmath12 in front of @xmath31 . specifically , for negative constants @xmath12 , different authors showed that this model leads to the same asymptotic behavior of the first order perturbation , avoiding oscillations and converging to a sharp interface functional .
the case @xmath32 was settled by hilhorst , peletier and schtzle in @xcite , where the authors proved that the functionals @xmath1 @xmath33-converge to a limit functional of type .
the case @xmath34 was instead considered by fonseca and mantegazza . in @xcite
the authors established the same limit behavior thanks to a compactness result for sequences with equi - bounded energy obtained exploiting some _ a priori _ bounds given by the growth assumption on the double - well potential @xmath8 and by a gagliardo - nirenberg interpolation inequality . in this paper
we investigate the case @xmath3 .
the presence of a negative contribution due to the term involving the first order derivative makes the problem quite unusual in the context of higher - order models of phase transitions .
in particular , since the three different terms in the energy are of the same order , their competition s outcomes strongly depend on the value of @xmath12 .
heuristically , large values of @xmath12 make the phases highly unstable favoring oscillations between them and correspond to negative surface tensions .
this was rigorously proved by mizel , peletier and troy in @xcite .
the authors considered the classical potential @xmath35 and showed that , for @xmath36 , @xmath37 and that there exists a class of minimizers of @xmath1 which are non - constant periodic functions oscillating between the two potential wells .
finer properties of these minimizers have been studied also in @xcite .
it is worth mentioning here that analogous results have been obtained for the non - local perturbations of the van der waals model in one - dimensional space , as the already mentioned otha
kawasaki model ( see , for example , the forerunner study of mller @xcite in the context of coherent solid - solid phase transitions ) .
these energies , when viewed as functionals of a suitable primitive of the order parameter of the system , become second - order functionals with a potential constraint on the first derivative , and lead to similar results .
what was left open by the analysis carried out in @xcite is the case of `` small '' , positive constants @xmath12 .
we prove here that , under the assumptions that the potential @xmath38 is quadratic in a neighborhood of the wells and grows at least as @xmath39 at infinity ( both hypothesis being necessary as discussed in section [ s : cpt ] ) , small values of @xmath12 make the phases stable and correspond to positive surface tensions ; _ i.e. , _ the asymptotic behavior of @xmath14 is again described by a sharp interface limit as in .
the main difficulty in the achievement of the above result lies in the proof of a compactness theorem analog to the one obtained in the case of the modica
mortola functional . indeed , the negative term in the energy @xmath1 when @xmath3 gives no _ a priori _ bounds on minimizing sequences .
here we solve this problem showing the existence of constants @xmath40 such that a new nonlinear interpolation inequality holds ( see lemma [ nonli ] ) : @xmath41 for every @xmath42 and @xmath43 .
this inequality enables us to estimate from below our functionals with @xmath44 ( the one corresponding to @xmath34 ) for which the compactness result has been proved in @xcite .
therefore , for @xmath7 every sequence of functions with equi - bounded energy @xmath1 converges in @xmath26 ( up to subsequences ) to a function @xmath45 ( see proposition [ comp2 ] ) . on account of this result
, we then complete the @xmath46-convergence analysis of the family of functionals @xmath14 by proving in theorem [ glimit ] that , for every @xmath7 , the functionals @xmath1 @xmath9-converge to @xmath47 where @xmath48 is given by the following optimal profile problem ( whose solution s existence is part of the result ) , @xmath49 in the last part of the paper we address the problem of estimating the constants @xmath12 for which there are no oscillations in the asymptotic minimizers . in order to compare our results with those in @xcite ,
we focus here on the explicit potential @xmath13 they considered . since the estimate we could derive for @xmath6 are very rough , in section [ sec : vs ] we investigate the following different problem , @xmath50 where for every interval @xmath51 and every @xmath52 , @xmath53 is the rayleigh quotient defined as @xmath54 problem is clearly related to the computation of the optimal constant in the nonlinear interpolation inequality and seems to be a challenging open problem . clearly , for @xmath55 ,
the minimizers exhibit an oscillating behavior .
but , more importantly , we are able to show that , for @xmath56 , there are no oscillations , because of a @xmath57 compactness result in @xmath58 for sequences of functions equi - bounded in energy and having at least one zero of the first derivative ( see proposition [ comp - ottimale ] and notice that this condition is fulfilled by any sequence of functions which is supposed to oscillate ) . this compactness result , although analogous to the previous one provided for @xmath7 , is actually more difficult , because in this last case the energy is not everywhere positive , but can be in principle negative in a boundary layer ( see lemma [ interp - bordo ] ) .
the benefit of this more refined compactness is that we can provide an upper bound and a lower bound on @xmath59 which have the same order of magnitude .
the lower bound we obtain for @xmath59 follows by carefully tracing the constants in the linear interpolation inequality and amounts to @xmath60 ; the upper bound @xmath61 follows , instead , from a test with quadratic polynomials and gives an improved estimate with respect the one given in @xcite ( @xmath62 ) .
what the present analysis does not still settle is a better understanding of the interpolation constants @xmath6 and @xmath59 .
we conjecture , indeed , that for every @xmath63 the functionals @xmath14 do not develop microstructures and @xmath46-converge to a sharp interface functional of type up to an additive constant depending on the presence of possible boundary layers energies .
similarly , we plan to address the analogous analysis in any space dimension in a future work ( see remark [ + d ] for the proof of the compactness in any dimension as a consequence of the one - dimensional result ) .
in this section we set our notation and we recall some preliminary results we employ in the sequel . with @xmath64 we always denote an open bounded interval and with @xmath65 two positive constants .
moreover , we fix a class of double - well potentials with the following properties : @xmath66 is continuous , @xmath67 ( the location of the wells clearly can be fixed arbitrarily ) , and satisfies * there exists @xmath68 such that @xmath69 for @xmath70 . note that in particular the standard double - well potential @xmath13 belongs to this class .
we consider the functionals @xmath14 defined in and , whenever the domain of integration is clear from the context , we simply write @xmath71 in place of @xmath72 . we denote by @xmath73 the functional introduced in @xcite ; that is @xmath74 as we heavily use it in the sequel , we recall here the statement of one of the main results of @xcite ( see ( * ? ? ? * proposition 2.7 ) ) .
[ p : cptfm ] let @xmath75 satisfy @xmath76 . then
, there exist a subsequence ( not relabeled ) and a function @xmath77 such that @xmath78 in @xmath79 .
we also recall two classical interpolation inequalities ( see ( * ? ? ?
* theorem 1.2 and ( 1.22 ) pag .
10 ) and @xcite ) .
[ standard - interp ] for every @xmath5 , @xmath80 , and every function @xmath81 , the following inequalities hold : \(i ) ( optimal constant ) @xmath82 for every @xmath68 , with @xmath83 ; \(ii ) there exists a constant @xmath68 such that @xmath84 finally , we prove the following interpolation inequality with boundary terms .
[ bdry - interp ] for every @xmath5 with @xmath80 , @xmath85 and @xmath68 , it holds @xmath86 we have the following identity @xmath87 then , integrating both sides of we find @xmath88 hence , dividing by @xmath68 we get the thesis .
in this section we prove one of the main result of this paper , namely the existence of a constant @xmath89 such that , for every @xmath7 , the functional @xmath90 satisfy the same compactness property of proposition [ p : cptfm ] . as an easy consequence ,
we then obtain the existence of the solution to the optimal profile problem for @xmath14 . in this subsection
we prove a nonlinear version of the standard @xmath91-interpolation inequality of type ( i ) proposition [ standard - interp ] .
[ nonli ] there exists a constant @xmath92 such that @xmath93 for every @xmath85 and for every @xmath94 with @xmath95 . up to translations and rescalings ,
it is enough to prove for @xmath96 . to this end , we set @xmath97 and , by the symmetry of , up to exchanging @xmath18 with @xmath98 , we assume that @xmath99 . from the fundamental theorem of calculus , it follows that , @xmath100 therefore , to prove it is enough to show the existence of a constant @xmath68 such that@xmath101 if @xmath102 , then clearly follows .
if this is not the case , applying jensen s inequality in , we get @xmath103 this implies that @xmath18 is strictly increasing in @xmath104 and , therefore , @xmath18 does not vanish in at least one of the two intervals @xmath105 and @xmath106 . without loss of generality , we may assume that @xmath107 in @xmath105 .
hence , by and hypothesis ( w ) , we have @xmath108 since implies @xmath109 , from we get and thus the thesis
. [ interp - on - r ] dividing @xmath110 into disjoint intervals of length @xmath111 and applying we may deduce @xmath112 for every @xmath113 with @xmath92 as in lemma [ nonli ] .
now we prove that lemma [ nonli ] together with a simple decomposition argument yield a lower bound for @xmath14 in terms of the functional @xmath114 .
[ p : stima ] for every interval @xmath20 and @xmath115 , there exists @xmath116 such that , for every @xmath3 , @xmath43 , and @xmath117 , @xmath118 a change of variable gives @xmath119 where @xmath120 . set @xmath121 $ ]
; we divide the interval @xmath122 into @xmath123 pairwise disjoint open intervals @xmath124 , @xmath125 , of length @xmath126 .
then , by applying on each interval @xmath124 we get @xmath127 since @xmath128 we get the thesis just by unscaling .
the following compactness result is now an immediate consequence of proposition [ p : stima ] and proposition [ p : cptfm ] .
[ comp2 ] let @xmath7 and let @xmath129 be a sequence satisfying @xmath130 .
then there exist a subsequence ( not relabeled ) and a function @xmath45 such that @xmath131 in @xmath26 .
[ + d ] proposition [ p : stima ] can be easily generalized to any space dimension @xmath132 .
namely , an immediate consequence of it is the existence of a constant @xmath133 such that , for every smooth bounded domain @xmath134 , @xmath135 , @xmath136 , and @xmath0 small @xmath137 indeed , by a standard covering argument it is enough to discuss the case of a rectangle @xmath138 and then the conclusion follows by an easy application of fubini s theorem .
let @xmath139 and @xmath140 , we have @xmath141 here we briefly comment on the assumption ( w ) on the double - well potential @xmath8 .
we show with two explicit examples that the two following conditions * @xmath142 , * @xmath143 , ( which together are equivalent to ( w ) ) are necessary to establish .
indeed , let @xmath144 be two parameters to be fixed later and such that @xmath145 .
consider the two families of periodic functions , of period @xmath146 , @xmath24 and @xmath147 defined in @xmath104 in the following way . on a half period ,
both @xmath148 and @xmath149 are defined by a line of slope @xmath150 , an arc of parabola , and another line of slope @xmath151 , as in the figure [ f : ex ] ; moreover , @xmath152 and @xmath153 . for the sake of simplicity , to shorten the present computation ,
assume that @xmath8 is monotone on the intervals @xmath154 , @xmath155 , @xmath104 , @xmath156 ( note that this hypothesis is not necessary ) .
it is readily verified that : * @xmath157 always and @xmath158 on a set of measure @xmath159 , * @xmath160 , so that , for @xmath161 large enough , we have @xmath162 , * @xmath163 , so that @xmath164 . now ,
if holds true , from estimates ( a ) , ( b ) , ( c ) it follows that @xmath165 then , if ( i ) does not hold true , taking the limit as @xmath166 goes to @xmath167 and then as @xmath168 goes to @xmath167 gives contradiction . similarly , if ( ii ) is not satisfied , taking the limit as @xmath166 goes to @xmath169 and then @xmath168 tends to @xmath167 yields a contradiction as well . as a consequence of lemma [ nonli ] ,
we prove here the existence of a solution to the optimal profile problem for @xmath14 , with @xmath7 .
specifically , we consider the following set of functions @xmath170 and we define @xmath171 we have the following result .
[ equiv - min ] let @xmath6 be as in lemma [ nonli ]
. for every @xmath172 the constant @xmath11 is positive and @xmath173 before proving proposition [ equiv - min ] , we introduce the functions @xmath174 given by @xmath175),\\ g(0)=w,\ , g(1)=1,\ , g'(0)=z,\ , g'(1)=0 \bigg\}\end{gathered}\ ] ] and @xmath176),\\ h(0)=-1,\ , h(1)=w,\ , h'(0)=0,\ , h'(1)=z \bigg\}.\end{gathered}\ ] ] if @xmath177 and @xmath178 are the corresponding functions for @xmath34 it is easy to check ( see also ( * ? ? ?
* section 2 ) ) that @xmath179 then , by the positivity of @xmath12 and by virtue of we have @xmath180 which lead immediately to @xmath181 by virtue of the nonlinear interpolation inequality of lemma [ nonli ] , the proof of this proposition is an easy modification of that of ( * ? ?
* lemma 2.5 ) .
the positivity of @xmath11 follows from remark [ interp - on - r ] and ( * ? ? ?
* lemma 2.5 ) , since @xmath182 now we prove that @xmath183 , where @xmath184 clearly , @xmath185 .
for the converse inequality , fix @xmath186 and let @xmath187 be an admissible function for @xmath188 such that @xmath189 we show that it is possible to find two sequences @xmath190 and @xmath191 converging to @xmath167 and @xmath192 respectively , and such that @xmath193 as @xmath194 . indeed , in view of remark [ interp - on - r ] we have @xmath195 thus , since @xmath7 we deduce that @xmath196 and so there exist two sequences of points @xmath197 and @xmath198 such that @xmath199 let @xmath200 and @xmath201 be two admissible functions for @xmath202 and @xmath203 , respectively , such that @xmath204 @xmath205 and set @xmath206 we define @xmath207 clearly , @xmath208 is a test function for @xmath11 and for @xmath172 we have @xmath209 hence we conclude letting @xmath194 and appealing to . finally , it remains to prove that @xmath188 admits a minimizer . to this end , let @xmath210 be a sequence which realizes @xmath188 .
then , by remark [ interp - on - r ] we have @xmath211 hence , again by interpolation and appealing to the sobolev embedding theorem , we deduce that ( up to subsequence ) the sequence of @xmath212 functions @xmath213 converges in @xmath214 to a @xmath212 function @xmath187 with @xmath215 by , it follows that @xmath216 for every @xmath217 , by the @xmath214-convergence of @xmath213 , fatou lemma and the lower semicontinuity of the @xmath91-norm of the second derivative , we have @xmath218 where the last inequality in is a consequence of written for the two half lines @xmath219 and @xmath220 .
then , taking into account and passing to the sup on @xmath217 we get @xmath221 thus , it remains only to show that the limit function @xmath187 is admissible . since this is a direct consequence of the third step of the proof of ( * ? ? ?
* lemma 2.5 ) , we leave some minor details to the reader and conclude the proof .
on account of the compactness result proposition [ comp2 ] , in this section we compute the @xmath46-limit of the functionals @xmath14 when @xmath7 . [
t : glimit ] for every @xmath7 , the functionals @xmath1 @xmath9-converge to @xmath223 where @xmath224 is the number of jumps of @xmath18 in @xmath20 and @xmath11 is as in .
we divide the proof into two parts , proving the @xmath46-liminf and the @xmath46-limsup inequality , respectively . _
part i : @xmath46-liminf .
_ let @xmath225 and @xmath226 such that @xmath131 .
we want to show that @xmath227 clearly , it is enough to consider the case @xmath228 . from
we immediately deduce @xmath229 , so that gives @xmath230 let @xmath231 , @xmath232 with @xmath233 , and set @xmath234 .
fix @xmath235 .
then ( up to subsequences ) @xmath236 for every @xmath237 . hence if we let @xmath238 , for every @xmath237 we may find two points @xmath239 such that , for sufficiently small @xmath240 , @xmath241 and set @xmath242 with @xmath243 and @xmath244 admissible for @xmath245 and @xmath246 , respectively , and satisfying @xmath247 and @xmath248 now we suitably modify the sequence @xmath24 `` far '' from each jump point @xmath249 . to this end , for every @xmath237 we define on @xmath110 the functions @xmath250 as @xmath251 since each @xmath250 is a test function for @xmath11 , we have @xmath252 then , as the intervals @xmath253 are pairwise disjoint for @xmath237 , in view of the non - negative character of @xmath1 for @xmath7 , we get @xmath254 finally , letting @xmath255 , we conclude by . _
part ii : @xmath46-limsup .
_ let @xmath45 with @xmath256 as in part @xmath20 , and set @xmath257 , @xmath258 . for @xmath237 define @xmath259 $ ] and @xmath260 .
fix @xmath261 and @xmath262 such that @xmath263 if @xmath264 , @xmath265 if @xmath266 , for some @xmath217 , and @xmath267 starting from this @xmath187 we construct a recovery sequence @xmath24 for our @xmath46-limit .
there exists @xmath116 such that for every @xmath268 we have @xmath269 . for @xmath270 , we define @xmath271(s_i)>0 , \cr\cr { \displaystyle}f\big(-\frac{x - s_i}{\e}\big ) & \text{if}\ ; x\in i_i\ ; \text{and}\ ; [ u](s_i)<0,\cr\cr u(x ) & \text{otherwise } , \end{cases}\ ] ] where @xmath272(s_i):=u(s_i)-u(s_{i-1})$ ] , for @xmath273 .
it can be easily shown that @xmath129 and @xmath274 in @xmath26 .
moreover , @xmath275(s_i)>0 } \int_{i_i/\e}(w(f(x))-k(f'(x))^2+(f''(x))^2){\,dx}\\ & + \sum_{i=1,\ldots , n \colon [ u](s_i)<0 } \int_{i_i/\e}(w(f(-x))-k(f'(-x))^2+(f''(-x))^2){\,dx}\big\}\\ \leq { } & { \bf m}_k n+\d = { \bf m}_k \#s(u)+\d , \end{aligned}\ ] ] hence we conclude by the arbitrariness of @xmath115 .
throughout the last two sections we fix @xmath13 . in the spirit of mizel , peletier and troy @xcite , in this section we provide a compactness result , alternative to that of proposition [ comp2 ] , which asserts the existence of a range of values of @xmath12 such that sequences with equi - bounded energy @xmath1 , whose derivative vanishes at least in one point of @xmath20 , do not develop oscillations .
the reason for this new compactness result , as explained in the introduction , is to give reasonable bounds on these values of @xmath12 .
the key parameter for our analysis is the following @xmath276 where for every interval @xmath277 and every @xmath85 , @xmath278 is the rayleigh quotient defined as @xmath279 it can be easily proved that @xmath59 can be equivalently rewritten as @xmath280 clearly , for @xmath55 , there are functions in the class defining @xmath59 for which the functionals @xmath14 are non - positive
. therefore , there are minimizers of @xmath14 developing an oscillating structure , finer and finer as @xmath0 approaches @xmath169 .
a thorough study of oscillating minimizers has been carried out in @xcite .
on the other hand , when @xmath63 and the function @xmath42 is such that it is possible to divide its domain @xmath20 into subintervals @xmath281 in which @xmath18 has constant sign , is strictly monotone , and @xmath282 , the definition of @xmath59 directly implies @xmath283 hence , this case falls in the analysis performed in the previous sections and the development of oscillations for minimizers is ruled out , as implied by theorem [ t : glimit ] .
this conclusion applies , for instance , when we prescribe homogeneous neumann boundary conditions or periodic boundary conditions on @xmath18 . on the contrary , if we do not impose any boundary conditions on @xmath18 , the estimate of the energy of @xmath18 in a neighborhood of the extrema of the interval @xmath20 require a further investigation .
such investigation is the main issue of this section .
the main result asserts that for @xmath56 , even without prescribing boundary conditions , minimizers of @xmath1 can not develop an oscillatory structure .
[ comp - ottimale ] let @xmath284 ; let @xmath129 be a sequence such that @xmath285 at least in one point of @xmath20 and satisfying @xmath286 , then there exist a subsequence ( not relabeled ) and a function @xmath45 such that @xmath131 in @xmath287 .
the proof of proposition [ comp - ottimale ] is a straightforward consequence of proposition [ stima - ottimale ] below and of proposition [ p : cptfm ] .
unfortunately , at this stage it is still unclear if in proposition [ comp - ottimale ] taking the minimum between @xmath59 and @xmath288 is really necessary or it is a technical hypothesis .
[ stima - ottimale ] for every @xmath284 there exist two constants @xmath289 such that @xmath290 for every @xmath240 and for every @xmath42 such that @xmath291 vanishes at least in one point of @xmath20 .
the following lemma is the main ingredient in the proof of proposition [ stima - ottimale ] .
[ interp - bordo ] let @xmath42 and let @xmath292 , @xmath80 be the smallest points in @xmath20 such that @xmath293 .
then , there exists @xmath294 such that , for every @xmath295 we have @xmath296 let @xmath297 , with @xmath298 as in the statement , thus @xmath18 is monotone in @xmath299 and has constant sign in @xmath277 . by the symmetry property of the problem , without loss of generality ,
we assume @xmath18 strictly increasing in @xmath299 and @xmath107 in @xmath277 . therefore , we are in the following hypotheses : @xmath300 ( see also figure [ f : bordo ] ) . in a neighborhood of @xmath301 . ]
the interpolation inequality on the interval @xmath302 , with @xmath303 , gives @xmath304 while the same computation on @xmath277 yields @xmath305 from we deduce @xmath306 since for every @xmath307 we have @xmath308 , we may write @xmath309 then , gathering and , we find @xmath310 by estimating in the quantity @xmath311 with , we get @xmath312 thus finally @xmath313 then , to get the thesis for @xmath314 for a suitable constant @xmath68 , it is suffice to show that @xmath315 for every @xmath307
. we prove .
consider @xmath316 to be fixed later .
if @xmath317 , by exploiting @xmath318 and the fundamental theorem of calculus , we have @xmath319 from which follows with @xmath320 .
so now suppose @xmath321 .
again we distinguish two cases .
if @xmath322 then , arguing as above we find @xmath323 thus directly yields again with @xmath320 .
if does not hold true , then , from @xmath324 , to get it is enough to estimate @xmath325 . by the young inequality @xmath326 we have @xmath327 on the other hand , using the jensen inequality we find @xmath328 therefore @xmath329 where in the last inequality we also used the fact that @xmath330 , and @xmath331 in @xmath277 .
eventually we have @xmath332 taking @xmath333 for a suitable constant @xmath68 , follows and thus the thesis .
the proof is straightforward combining lemma [ interp - bordo ] and .
indeed , let @xmath334 : then , applies for two suitable neighboring intervals of @xmath301 and @xmath335 , while applies on internal intervals @xmath336 such that @xmath337 have constant sign in @xmath336 and @xmath282 .
in order to compare our results with those by mizel , peletier and troy @xcite , in this section we provide two estimates , one from below and one from above , on the interpolation constant @xmath59 , the one from above improving their bound . to establish an estimate from below on @xmath59 ,
the idea is to use the interpolation inequality ( i ) of proposition [ standard - interp ] , which gives a good bound on @xmath59 on `` large '' intervals , and to combine it with an inequality of jensen type which is good on `` small '' intervals .
then , recalling the definition of @xmath344 , by we get the first bound @xmath345 on the other hand , proposition [ standard - interp ] ( i ) gives @xmath346 hence gathering and yields the second bound @xmath347 finally , combining and , recalling the definition of @xmath59 and choosing @xmath348 , an explicit calculation yields @xmath349 concerning the estimate from above , we test the value of the rayleigh quotient @xmath350 on the quadratic polynomials @xmath351 , with @xmath352 .
then , a straightforward computation gives @xmath353 then , minimization of @xmath354 on @xmath355 and @xmath356 yields the bound @xmath357 .
a slightly better upper bound on @xmath59 can be obtained testing @xmath350 on functions as in figure [ f : ex ] .
nevertheless , since the value we find in this way ( @xmath358 ) is again larger than @xmath288 , this does not substantially improve the statement of proposition [ stima - ottimale ] , hence we omit this further computation .
: some results and conjectures in the gradient theory of phase transitions , in `` metastability and incompletely posed problems , proc .
workshop , minneapolis / minn .
1984/85 '' , _
i m a vol .
appl . _ * 3 * ( 1987 ) , 135 - 146 . | we consider the problem of the asymptotic description , as @xmath0 tends to zero , of the functionals @xmath1 introduced by coleman and mizel in the theory of nonlinear second - order materials ; _ i.e. , _
@xmath2 where @xmath3 and @xmath4 is a double - well potential with two potential wells of level zero at @xmath5 . by proving a new nonlinear interpolation inequality ,
we show that there exists a positive constant @xmath6 such that , for @xmath7 and for a class of potentials @xmath8 , @xmath1 @xmath9-converges to @xmath10 where @xmath11 is a constant depending on @xmath8 and @xmath12 .
moreover , in the special case of the classical potential @xmath13 , we provide an upper bound on the values of @xmath12 such that the minimizers of @xmath14 can not develop oscillations on some fine scale , thus improving previous estimates by mizel , peletier and troy . |
in the absence of exact solutions non - linear differential equations of field theory manifest features that they can be simulated by some known potentials .
linear equations with complicated potentials also can be treated similarly . in this
regard we recall the @xmath1atom endowed with non - linear electromagnetic field due to born and infeld @xcite . in this approach basic idea was to eliminate the divergences due to point charges .
the resulting regular electromagnetic potential expected to contribute to the solution of schrdinger equation and the ground state energy level was obtained remarkably as a simulation @xcite of the morse potential morse .
being motivated by that study we wish to employ the same method to another well - known differential equation representing an analog of global monopole in flat spacetime .
global monopoles are topological structures that emerge from spontaneous symmetry breaking and are believed to exist from the big bang .
such a monopole was considered first by barriola and vilenkin bv as source of gravity which may go to the extend of forming black holes . in @xmath2dimensions
the symmetry that has been broken is @xmath3 which ultimately reduces to @xmath4 . in this approach
@xcite the scalar field triplet is given by @xmath5 where @xmath6 are the gauge parameters , @xmath7 is the monopole charge and @xmath8 is the function to be determined .
unfortunately the differential equation satisfied by @xmath9 does nt admit an exact analytical solution .
the asymptotic behaviors satisfied by @xmath10 are such that @xmath11 for @xmath12 and @xmath13 with @xmath14 for @xmath15 .
numerical integration of the resulting differential equation is at our disposal and our aim is to simulate such a global monopole equation by using a morse type potential . before extending our formalism to higher dimensions we consider the @xmath16dimensional analogue of the @xmath2dimensional global monopole . in @xmath16dimensions
the gauge group to be broken becomes now @xmath17 so that the scalar doublet is represented by @xmath5 where @xmath18 the resulting differential equation for @xmath9 in @xmath16dimensions is as difficult as in the @xmath2dimensional case . as a matter of fact asymptotically the resulting equation is reminiscent of a minor variation of the third painlev transcendent @xcite encountered in the study of cylindrical gravitational waves @xcite .
this oscillatory solution , however , does nt represent the monopole character but yet arises as a distinct solution to the same differential equation satisfied by the monopole .
in other words both types of solutions arise from the same differential equation with different initial conditions .
morse potential fitting to the global monopole equation in @xmath16dimensions is observed to work perfectly .
going to the lower dimension @xmath19 , the differential equation satisfied by @xmath9 is solved exactly as @xmath20 this is , however , a singlet field and it ca nt be interpreted as a global monopole since there is no symmetry breaking in the usual sense .
reflection symmetry is the only available symmetry to be in @xmath21dimensional space .
next , we interpolate our analysis to higher dimensions @xmath22 and repeat our method of morse simulation .
we observe that as @xmath23 gets higher the fitting of our morse type potential becomes less reliable in comparison with the dimensions @xmath24 and @xmath25 .
let s start with the lagrangian of a global - monopole in @xmath16dimensional flat , spherically symmetric spacetime which is given by@xmath26here @xmath27 , @xmath28 is a coupling constant , @xmath7 is the monopole parameter and@xmath29for @xmath30 and @xmath31 to find the field equation for @xmath9 we express the field lagrangian in terms of @xmath9 only , i.e. , @xmath32now , variation of the action with respect to @xmath33 yields@xmath34or in more convenient form@xmath35 in which a prime stands for the derivative with respect to @xmath36 and @xmath37 is the size of the global monopole in flat spacetime .
we note that a scaling of the form @xmath38 absorbs the parameter , hence without loss of generality one can set @xmath39 note also that the term @xmath40 is as a result of symmetry breaking .
let us note that eq .
( 5 ) admits an asymptotic solution in analogy with the cylindrical gravitational waves expressed exactly in terms of painlev type iii @xcite .
our case turns out to be@xmath41(for @xmath42constant ) which is comparable with the asymptotic solution given in @xcite . in fig .
1 we plot the solution of latter equation for @xmath43 and @xmath44 the solution is oscillatory for @xmath45 , asymptotes to one for @xmath46 and diverges for @xmath47 interestingly the decisive factor between the two types of solutions is the initial slope of the curve to start with . below certain value
we have oscillations while above that we have the frozen global monopole solution . with respect to @xmath36 for @xmath48 in flat spacetime .
the initial conditions are set as @xmath49 and @xmath50 @xmath51 and @xmath52 from the top / black - solid to the bottom / red - long dash .
note that the shaded region corresponds to the inner region of global - monopole .
we add also that depending on the initial slope eq .
( 5 ) admits a wave like solution which is apart from the monopole . for completeness
we have shown the latter as well.,width=264 ] with respect to @xmath36 for @xmath48 in flat spacetime for the initial conditions are set as @xmath49 and @xmath53 and the morse function ( red - dash).,width=264 ] in fig .
2 we plot the numerical solution of eq . ( 5 ) which represents a global monopole with @xmath43 and the morse type function@xmath54 in which @xmath55 .
obviously @xmath9 satisfies the conditions at @xmath56 and @xmath57
in higher dimensional spherical symmetric flat spacetime the action of a global monopole field is given by ( 1 ) and ( 2 ) , with @xmath58 and @xmath59where @xmath60 $ ] and @xmath61 $ ] with @xmath62 the arbitrary dimensional action can be expressed in the form@xmath63 in which @xmath64 with @xmath23 the dimension of the spacetime .
the field equation following the above lagrangian is found to be@xmath65with @xmath66 the size of the higher dimensional global monopole . in fig .
3 we plot the global monopole solution of the latter equation for @xmath67 ( top / black - long dash ) , @xmath68 and @xmath69 ( bottom / purple - dot ) .
the case when the dimension is two i.e. , @xmath67 is analytically solvable .
the field equation for @xmath67 and @xmath43 can be written as@xmath70with a first integral@xmath71 in which @xmath72 is an integration constant . from the top to the bottom .
, width=264 ] and the numeric solution for @xmath73 the two solutions agree highly which can not be distinguished from each other.,width=264 ] as we are looking for a global monopole type solution , we impose the constraint that when @xmath15 , @xmath74 and @xmath75 this in turn implies that @xmath76 upon considering @xmath72 in ( 12 ) we obtain@xmath77furthermore the global monopole field admits a positive derivative and is less than one which imply@xmath78this final equation admits an exact solution given by@xmath79 in which @xmath80 is an integration constant . imposing @xmath81 one finds @xmath82 which makes the final solution simply as@xmath83 in fig .
4 we plot the exact solution ( 16 ) and the numerical solution found from the field equation .
we see that they agree to high precision .
this accuracy guides our numerical method in @xmath84 also to the higher dimensions .
similar to @xmath85 for @xmath22 we can simulate the solution with a morse type function given by@xmath86 in which @xmath87 is the slope of the solution at @xmath56 and @xmath88 is an adjusting parameter . in table .
1 we give @xmath87 for dimensions @xmath22 and in fig .
5 we plot the numerical solution with the simulated function for @xmath89 .@xmath87 for different dimensions . [ cols="^,^ " , ] and the numeric solution ( black - solid ) for @xmath89 in the morse function we set @xmath90 the two solutions agree in both small and large @xmath36 to high accuracy.,width=264 ]
having the solution of eq . ( 10 ) simulated by the morse type function is not a coincidence .
the initial form of the morse potential can be written as @xcite@xmath91 in which @xmath92 , @xmath93 and @xmath94 are free positive parameters of the potential .
this function has a minimum at @xmath95 and for large @xmath36 it approaches to @xmath96 setting @xmath97 and @xmath98 we find a function whose asymptotic behaviours are as we are expecting for @xmath9 in eq .
( 10 ) i.e. , @xmath99 and @xmath100 furthermore , for large @xmath36 i.e. , @xmath101 eq . ( 10 ) becomes independent of dimensions and therefore the solution must be asymptotically the same as the exact solution of ( 10 ) in @xmath67 given by ( 16 ) .
hence we get@xmath102where @xmath103 is a positive constant which could be set according to the dimensions .
the asymptotic behavior of the right hand side in ( 19 ) surprisingly is of the morse type function i.e. , for large @xmath36@xmath104finally , for small @xmath36 i.e. , @xmath105 the field equation ( 10 ) becomes@xmath106with a solution given by@xmath107 in which @xmath108 and @xmath109 are constants .
a regular solution at @xmath56 requires @xmath110 and the asymptotic solution for small @xmath36 becomes @xmath111 from the other hand the asymptotic behaviour of the morse potential ( 18 ) with @xmath97 and @xmath98 near @xmath56 becomes@xmath112which is same as @xmath113 in conclusion the morse type function in both limits i.e. , @xmath114 and @xmath15 matches very well with the solution of the field equation ( 10 ) . as we expect from the fig .
5 , the maximum agreement occurs in these limits .
irrespective of dimensions ( @xmath115 ) the geometry of a global monopole emerges complicated enough to be tackled with , even in a flat spacetime .
the reason is the transcendental differential equation satisfied by the global monopole function @xmath113 asymptotically the equation satisfied by @xmath9 has similarity with the painlev type iii , whose behavior can be studied numerically .
we have shown that the numerical plot of the function can be simulated by a morse type function to great accuracy .
it is shown that the accuracy of our simulation works better in lower dimensions such as @xmath116 .
@xmath67 is the only available case that admits exact integration , however , its significance as a monopole remains questionable .
recalling that a global monopole constitutes a particular case of @xmath117 theories similar simulations may find application also in the latter theories .
further , question can naturally be raised about the importance of such a simulation advocated herein : the physics of particles in the field of a monopole becomes tractable through the morse function .
our project next is to extend the idea to global monopole in a curved spacetime . | we show that a morse type potential simulates an analytic solution for the highly non - linear global monopole field equation in three and higher dimensional flat spacetimes . owing to the fact that in the flat space limit the similar equation remains intact we wish to borrow the curved space terminology of global monopole also in flat spacetime .
this may provide a compelling example that can be used effectively in different non - linear theories such as flat space @xmath0 as well as in curved spacetimes . |
distributed bragg reflectors ( dbrs ) are typically applied to enhance@xcite the reflectance ( @xmath5 ) , but the opposite result a very low @xmath5 can be also designed and obtained with the use of dbrs@xcite .
such antireflection ( ar ) stacks are particularly useful in resonant optical spectroscopy , where the reflected laser light dominates the investigated signal substantially .
in particular , using four - wave mixing ( fwm ) for probing the coherence of single quantum dots ( qds ) , necessitates efficient injection of the optical excitation into the hetero - structure , so as to reach the field amplitude at the qd location sufficient to induce its fwm .
this condition has been achieved by employing ar planar samples@xcite and , more recently , by suitably designed nanophotonic devices@xcite . here , we realize half - cavity ar planar structures and employ them to investigate fwm of individual cdte qds . as the same structure contains qds with incorporated single mn@xmath1 ions , the appealing prospect of our work is to reveal the exciton s coherence coupled with a magnetic dopant via @xmath2,@xmath3-@xmath4 exchange interaction .
there are two basic approaches to ar structures on the top of high refractive index ( @xmath6 ) semiconductors .
firstly , one can deposit a single layer with a lower refractive index equal to @xmath7 and thickness of @xmath8 , where @xmath9 is the targeted wavelength for the lowest reflectance .
while this method offers a broad ar spectral range , its implementation is conditioned by the availability of materials with a required value of @xmath6 .
secondly , one can use symmetric microcavities composed of two dbrs with the nominally the same @xmath5 .
the cavity mode exhibits a pronounced dip in the @xmath5 spectrum , which should be accurately fitted to the wavelength of interest . in this work
, we implement the ar strategy based on a cavity system . in order to maximize the spectral width of the cavity resonance , we have designed and grown a low q - factor structure ( @xmath10 ) : a half - cavity is formed by four bottom bragg pairs and the spacer containing the qd layer , also acting as a semiconductor / air interface at the top .
the sample was grown using molecular beam epitaxy ( mbe ) model svt associates .
our molecular sources were standard knudsen cells loaded with commercially available materials : zn , cd , and te with 7 n purity ; mg and mn with 6 n purity .
chamber pressure was at level of about @xmath11 torr .
the growth was assisted by _ in - situ _ optical reflectance ( model filmetrics f-30 ) with a white light incident beam perpendicular to the sample surface .
the system is equipped with a heated optical viewport , so that it was possible to perform spectroscopic measurement during a long growth sequence , while avoiding material deposition on the window . for the sample described in this work
, we monitored @xmath5 versus time at @xmath12 nm , as shown in fig.[fig : structure](a ) . in our method , this choice defined the resonant wavelength of the cavity for a given spot of the sample ( the substrate was not rotated during the growth ) .
however , other zones of the sample exhibit either thinner layers with cavity resonance at shorter wavelength or thicker layers with resonance at longer wavelength . the growth was performed at constant temperature t = 345@xmath13 c , on a 2-inch gaas : si substrate oriented ( 100 ) with 2@xmath14 off .
we started by depositing a 3@xmath15 m thick cdte buffer .
since cdte is absorbing at @xmath16 nm , the corresponding oscillations of @xmath5 were rapidly suppressed and a characteristic @xmath17 was measured before time @xmath18s . as depicted in fig.[fig : structure](b ) , we next grew a cd@xmath19zn@xmath20mg@xmath21te buffer lattice matched to subsequent dbr layers . due to the presence of mg and zn , such buffer exhibits a slightly lower refractive index than cdte and @xmath22mev larger energy gap than cdte . therefore , cd@xmath19zn@xmath20mg@xmath21te buffer is transparent for @xmath23 nm , generating interferences in the @xmath24 trace ( time between 10400s and 13100s in fig.[fig : structure]a ) during the growth of this material .
observation of @xmath24 interferences allowed us to determine growth rate v = 1.1 @xmath25m / h .
next , when @xmath24 reached a local maximum , we started the growth of lattice matched dbrs composed of cd@xmath26zn@xmath27mg@xmath28te ( low @xmath6 layer ) and cd@xmath19zn@xmath20mg@xmath21te ( high @xmath6 layer)@xcite .
adjacent layers with various composition of mg were obtained using a pair of mg sources , such that either one or two sources were opened at a time . an exact time - point for growing dbrs
was not predefined in the programmed sequence . instead , it was adjusted manually by closing one of the mg shutters , as soon as the minimum of @xmath5 was reached , while keeping opened both mg shutters when the maximum of @xmath5 was attained ( see fig.[fig : structure](a ) ) . as expected , during the initial stage of a dbr growth , after each pair , the contrast between the maximum and minimum value of @xmath5 was increasing .
interestingly , @xmath5 reached the minimum value below 1@xmath29 already after depositing the fourth dbr pair .
since this is what required for an efficient ar structure and further dbr growth would only increase reflectance , we completed the growth of dbrs already after four pairs and we completed @xmath30 cavity composed of the high @xmath6 material cd@xmath19zn@xmath20mg@xmath21te , enclosing cdte : mn qds in the middle .
as shown in fig.[fig : structure](a ) , fabricating thickness equal to @xmath31 or @xmath32 of high index material does not affect @xmath5 at @xmath23 nm .
our half - cavity structure is formed and terminated by the semiconductor / air interface on the top .
qds are formed out of about 1 nm ( 3 monolayers ) thick cdte : mn layer grown at the same conditions as cd@xmath19zn@xmath20mg@xmath21te and the remaining top part of the sample .
basing on giant zeeman effect@xcite measured for thicker cdte : mn layers grown in similar conditions , expected concentration of mn in our cdte : mn layer is about @xmath33 . as a matter of fact
, this layer forms a thin quantum well exhibiting interface fluctuations , leading to formation of qds .
the growth of similar qds , but in cd@xmath34mg@xmath35te barriers , was previously reported@xcite . here
, we used even lower barriers and consequently we shifted the qd emission energy to spectral region where the absorption of ( cd , zn , mg)te is low enough to realize microcavities .
microscopy image of the most important parts of studied structure is shown in fig .
[ fig : tem ] .
dbr layers , seen as alternating bright and dark layers , are well resolved .
also , the qd layer is resolved in the middle of the cavity , it is seen as a thin trace in the magnified part of the cavity .
[ [ photoluminescence - and - reflectance . ] ] photoluminescence and reflectance .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + fig.[fig : pl - ref ] shows low temperature photoluminescence ( pl ) and reflectance spectra of studied sample , both measured using experimental setup equipped in a microscope objective with spatial resolution of about 1@xmath25m@xcite .
pl was excited nonresonantly at 532 nm using cw power of about 1@xmath25w , well below the qd saturation .
the reflectance was measured using a white ( halogen ) lamp .
k. pl shows a multitude of sharp lines corresponding to excitonic transitions in individual qds .
qds are in resonance with cavity mode observed in reflectance .
reflectance spectrum displays a broad ( about 40 nm wide ) and deep ( below 3% of reflectance ) cavity mode , showing that the dbr acts as an ar photonic structure for spectral region of qds emission . ]
zn@xmath20mg@xmath21te qd with a single mn ion , measured for higher excitation power ( upper spectrum ) and lower excitation power ( lower spectrum ) .
as expected , biexciton ( xb@xmath36 ) transitions exhibit higher sensitivity on excitation power than the exciton ( gx@xmath37 ) ones .
both clusters are sixfold split due to @xmath2,@xmath3-@xmath4 exchange interaction between individual mn@xmath1 ion and carriers confined in a qd . ]
the pl reveals a signal from a thin cdte layer at 665 nm and sharp lines of individual cdte qds in range 670 - 700 nm . in our sample ,
cdte qds are placed in cd@xmath19zn@xmath20mg@xmath21te barrier , which is also visible as a weak peak close to 650 nm in the pl spectrum and as a spectral wiggle in @xmath38 between 635 and 650 nm .
we therefore estimate the confinement energy of around 50mev for the excitons localized in qd - like states .
reflectance spectrum at the investigated area of the wafer shows also a broad ( about 40 nm wide ) and deep ( below 3@xmath29 of reflectance ) cavity mode centered at 670 nm , in resonance with the emission of qds . the obtained broad and deep cavity mode , mimics an ar coating , but also provides a local excitation intensity enhancement @xmath39 for the spectral and spatial location of qds . during the growth @xmath5
was minimized for 730 nm ( fig.[fig : structure](a ) ) for the center of the sample , but as shown in fig [ fig : pl - ref](b ) , owing to the thickness gradients , other parts of the wafer display a spectrally shifted cavity mode .
this permits to adjust low @xmath5 for the desired spectral range by varying the position on the sample .
the sample offers a possibility to investigate and control an exciton coherence coupled , via @xmath2,@xmath3-@xmath4 exchange interaction , with an individual spin of a magnetic dopant@xcite .
this is possible thanks to mn @xmath40-doping introduced within the cdte qd layer . via micro - photoluminescence spectroscopy
we can recognize qds containing mn ions , and in particular distinguish those incorporating exactly one such dopant . to be more specific ,
qds without magnetic ions exhibit sharp lines ( typical full - width at half - maximum ( fwhm ) below 0.1 mev ) , originating from the exciton recombination .
qds with more than two magnetic ions exhibit spectrally broader lines ( typically fwhm above 1 mev ) , which significantly shift in a magnetic field.@xcite for qds with exactly two mn ions , each excitonic transition gets split into a manifold of 36 lines , owing to 6 * 6 possible spin states of two mn ions.@xcite finally , a spectral fingerprint of a qd with a single mn ion is very characteristic : exciton transitions display 6-fold splitting , due to @xmath2,@xmath3-@xmath4 exchange interaction with the mn ion .
the latter exhibits 6 possible spin states , corresponding to spin projection @xmath41 , @xmath42 , and @xmath43 on quantization axis of excitons ( growth axis).@xcite in our sample , majority of qds do not contain mn ions and thus show single lines .
we have not observed qds with two or more mn ions .
however , there is a small fraction ( well below 1% ) of qds , which exhibit the pl pattern characteristic for qds containing exactly one individual mn ion .
a typical example of pl measured at such magnetic qd is shown in fig.[fig : mn ] : the excitonic spectral clusters corresponding to recombination of exciton - to - ground ( gx@xmath37 ) and biexciton - to - exciton states xb@xmath37 do display the characteristic 6-fold splitting @xcite .
spectra are measured as a function of the excitation power . as expected , the biexciton transitions exhibits a much stronger intensity dependence , than the exciton ones ( see fig .
[ fig : mn ] ) .
observation of a pronounced and spectrally sharp biexciton complex , with a binding energy of around 7mev , indicates that qds forming from fluctuations of thin cdte layer in cd@xmath19zn@xmath20mg@xmath21te barrier are in fact well defined zero dimensional objects , which efficiently localize even four carriers ( two holes and two electrons ) .
other excitonic complexes , corresponding to a negatively and/or positively charged qds , doped with single mn@xmath1 ions have also been identified ( not shown ) . [ [ four - wave - mixing . ] ] four - wave mixing . + + + + + + + + + + + + + + + + + to retrieve the coherent nonlinear response of individual qds , we employed the heterodyne spectral interferometry @xcite .
technical details regarding the current implementation of the setup are provided in ref .
[ ] . to resonantly excite the qds at 680 - 690 nm with 0.3ps pulses , we use an optical parametric oscillator ( opo model inspire provided by radiantis ) , pumped with a ti : sapphire laser ( model tsunami femto provided by spectra - physics ) . the excitation spectrum generated by the opo in shown in fig.[fig : fwm]b ( green ) .
we form a pair of pulse trains @xmath44 and @xmath45 with a variable delay @xmath46 controlled with a mechanical stage providing @xmath46 up to 1ns .
@xmath44 and @xmath45 are modulated using acousto - optic deflectors , optimized for a vis range and driven at @xmath47mhz and @xmath48mhz , respectively .
both beams are recombined into the same spatial mode and are focussed to the diffraction limit at the sample surface using a microscope objective ( na=0.65 ) installed on a xyz piezo - stage .
we interfere the reflectance with the frequency shifted reference beam @xmath49 . in particular , by selecting the heterodyne mixing at @xmath50mhz , we select the beating frequency corresponding to the fwm field , proportional to @xmath51 in the lowest order . the fwm-@xmath49 interference is spectrally resolved and detected by a ccd camera .
zn@xmath20mg@xmath21te barrier .
( a ) pl from a single qd , showing a high emission flux of 40kcounts / s , indicating an increased extraction efficiency .
( b ) fwm spectral interference ( blue ) of the same transition as in ( a ) driven with the @xmath44 and @xmath45 intensities @xmath52w , @xmath53ps .
the @xmath49 spectrum is given with a green trace .
( c ) time - integrated fwm amplitude as a function of @xmath44 amplitude @xmath54 .
( d ) coherence dynamics of the transition shown in ( b ) measured with fwm at @xmath55k : the applied pulse sequence is depicted in the inset .
the @xmath46-dependence of time - integrated fwm amplitude is shown in blue , the noise level is given with open circles .
we estimate the dephasing time @xmath56ps [ fig : fwm ] . ] as candidates for fwm measurements , we pick up the transitions as the one shown in fig.[fig : fwm]a , which feature an enhanced flux of the pl , thus indicating an increased extraction efficiency @xmath57 with respect to other qds . by monitoring the reflectance
, we assure that the targeted transition lies at the minimum of the cavity mode . as a result
, we expect the weak fwm field to be enhanced by a factor @xmath58 with respect to bare qds , and thus capable to be induced with a respectively lowered resonant intensity of @xmath59and @xmath60 .
a representative example of the fwm interference at @xmath61ps is presented in fig.[fig : fwm ] , with a signal - to - noise of 10 after 300s integration .
this is comparable with previous fwm measurements on single qds employing ar structures @xcite . in fig.[fig : fwm]c we present fwm amplitude , as a function of @xmath44 amplitude , proportional to its area @xmath62 . with increasing the latter ,
the fwm is expected to yield rabi oscillations @xcite , with the first maximum ( minimum ) at @xmath63 @xmath64 . in our case
however , the accessible range of intensities is not sufficient to attain significant @xmath62 , so as to approach @xmath65 area .
the flopping is not observed and we remain in the @xmath66 driving regime of the fwm . in the inset of fig.[fig : fwm]d we illustrate the two - pulse sequence performed to infer the coherence dynamics of the same transition ( @xmath49 is set fixed , 1ps prior to @xmath45 ) : @xmath44 induces the coherence , which is converted into the density grating by @xmath45 .
the latter is then self - diffracted into the fwm polarization .
by measuring fwm as a function of @xmath46 one can access the homogeneous broadening @xmath67 ( where @xmath68 denotes dephasing time ) within the inhomogeneously broadened distribution via inferring photon echoes , which in case of a single qd are created by a random spectral wandering @xcite .
the data are fit by the model including both homogeneous and inhomogeneous contributions to the spectral shape @xcite .
the result , shown with a black trace , yields the dephasing time of @xmath69ps and inhomogeneous broadening of @xmath70ev ( fwhm ) .
a large uncertainty of @xmath68 stems from a limited range of the accessible delays , restricted by the time - resolution of the spectrometer - the issue which could be overcome by introducing the delay between @xmath49 and @xmath45.@xcite note the decrease of the fwm during initial several ps , which is due to dephasing with acoustic phonons.@xcite we observe no fwm signal for @xmath71 , which could rise from a two - photon coherence@xcite between the ground state and a two - exciton ( biexciton ) state in a qd .
the missing signal for @xmath72 , indicates that the qd is charged , allowing only for the trion transition @xcite .
finally , it is worth to point out an order of magnitude longer @xmath68 , with respect to the previous , and until now sole report available for the cdte - platform @xcite .
our structure represents the third successful approach to the growth of photonic structures with qds containing single mn ions .
the first report has been presented in ref . [ ] , where a three - step process , involving two remote mbe facilities , was established to grow znte / mgte / mgse based microcavities with cdte / znte qds containing single mn ions . in that case , growth interruptions were due to missing molecular sources in growth chambers : the first growth chamber was used to realize dbrs , the second one was employed for the growth of mn - doped qds .
recently , the same material system was used for the growth of mn - doped qds on the top of 10 dbr pairs .
the process was carried out using a single mbe machine , but growth interruptions were anyway required to clean the chamber from the residual se atoms , which incorporated into qds and impaired their optical properties @xcite . comparing to previous works , here the procedure is significantly simplified , since se is not involved in the growth of our microcavity based on ( cd , zn , mg)te . therefore , the whole process is realized in a single growth chamber , without growth interruptions
. it can be reproduced in any mbe growth chamber devoted to te compounds .
one more important advantage of the present system is that cdte /cd@xmath19zn@xmath20mg@xmath21te
qds with single mn ion presented here could be brought toward the spectral range of al@xmath73o@xmath74:ti tunable femtosecond laser , i.e. above 700 nm .
this could be particularly handy as it eliminates the requirement of using opo sources to perform resonant , ultrafast spectroscopy of this material system .
finally , the design of antireflective photonic structure presented in this work will be necessary for achieving a satisfactory retrieval sensitivity of coherent nonlinear responses of magnetic qds , where the exciton oscillator strength is distributed over a set of spin - assisted transitions .
this work was partially supported by the ministry of science and higher education under grants iuventus plus ip2014 040473 and ip2014 034573 , and by the polish national science centre under decisions dec-2015/18/e
/ st3/00559 , dec-2011/02/a / st3/00131 , dec-2014/13/n / st3/03763 , dec-2013/09/b / st3/02603 , dec-2015/16/t / st3/00506 , umo-2012/05/b / st7/02155 .
the research was carried out with the use of cept , cezamat and nltk infrastructures financed by the european union - the european regional development fund within the operational programme `` innovative economy '' .
vd , tj and jk acknowledge the financial support by the european research council ( erc ) starting grant picsen ( grant no .
306387 ) .
wang , s. _ ieee journal of quantum electronics _ * 1974 * , _ 10 _ , 413427 sadowski , j. ; mariette , h. ; wasiela , a. ; andr , r. ; merle daubign , y. ; dietl , t. _ phys .
b _ * 1997 * , _ 56 _ , r1664r1667 benisty , h. ; de neve , h. ; weisbuch , c. _ ieee j. quant . electron . _ * 1998 * , _ 34 _ , 16121631 ng , h. m. ; doppalapudi , d. ; iliopoulos , e. ; moustakas , t. d. _ appl . phys .
lett . _ * 1999 * , _ 74 _ , 10361038 tawara , t. ; yoshida , h. ; yogo , t. ; tanaka , s. ; suemune , i. _ j. cryst . growth _ * 2000 * , _ 221 _ , 699 703 guo , s. p. ; maksimov , o. ; tamargo , m. c. ; peiris , f. c. ; furdyna , j. k. _ appl .
_ * 2000 * , _ 77 _ , 41074109 pacuski , w. ; kruse , c. ; figge , s. ; hommel , d. _ appl .
lett . _ * 2009 * , _ 94 _ , 191108 gust , a. ; kruse , c. ; otte , k. ; kalden , j. ; meeser , t. ; sebald , k. ; gutowski , j. ; hommel , d. _ nanotechnology _ * 2009 * , _ 20 _ , 015401 rousset , j .- g . ; kobak , j. ; slupinski , t. ; jakubczyk , t. ; stawicki , p. ; janik , e. ; tokarczyk , m. ; kowalski , g. ; nawrocki , m. ; pacuski , w. _ j. cryst . growth _ * 2013 * , _ 378 _ , 266 269 koba , m. ; suffczyski , j. _ j. electromagnet . wave _ * 2013 * , _ 27 _ , 700706 koba , m. ; suffczyski , j. _ europhys .
lett . _ * 2014 * , _ 108 _ , 27004 jakubczyk , t. ; franke , h. ; smoleski , t. ; ciesiek , m. ; pacuski , w. ; golnik , a. ; schmidt - grund , r. ; grundmann , m. ; kruse , c. ; hommel , d. ; kossacki , p. _ acs nano _ * 2014 * , _ 8 _ , 99709978 schubert , m. f. ; mont , f. w. ; chhajed , s. ; poxson , d. j. ; kim , j. k. ; schubert , e. f. _ opt .
express _ * 2008 * , _ 16 _ , 52905298 chhajed , s. ; schubert , m. f. ; kim , j. k. ; schubert , e. f. _ appl .
lett . _ * 2008 * , _ 93 _ , 251108 langbein , w. ; patton , b. _ opt . lett .
_ * 2006 * , _ 31 _ , 11511153 patton , b. ; langbein , w. ; woggon , u. ; maingault , l. ; mariette , h. _ phys . rev
* 2006 * , _ 73 _ , 235354 kasprzak , j. ; portolan , s. ; rastelli , a. ; wang , l. ; plumhof , j. d. ; schmidt , o. g. ; langbein , w. _ new j. phys . _
* 2013 * , _ 15 _ , 055006 kasprzak , j. ; reitzenstein , s. ; muljarov , e. a. ; kistner , c. ; schneider , c. ; strauss , m. ; hfling , s. ; forchel , a. ; langbein , w. _ nat .
mater . _ * 2010 * , _ 9 _ , 304308 fras , f. ; mermillod , q. ; nogues , g. ; hoarau , c. ; schneider , c. ; kamp , m. ; hfling , s. ; langbein , w. ; kasprzak , j. _ nat .
photon . _ * 2016 * , _ 10 _ , 155158 mermillod , q. ; jakubczyk , t. ; delmonte , v. ; delga , a. ; peinke , e. ; grard , j .- m . ; claudon , j. ; kasprzak , j. _ phys .
* 2016 * , _ 116 _ , 163903 rousset , j .- g . ;
pitka , b. ; krl , m. ; mirek , r. ; lekenta , k. ; szczytko , j. ; borysiuk , j. ; suffczyski , j. ; kazimierczuk , t. ; goryca , m. ; smoleski , t. ; kossacki , p. ; nawrocki , m. ; pacuski , w. _ appl .
_ * 2015 * , _ 107 _ , 201109 mirek , r. ; krl , m. ; lekenta , k. ; rousset , j .- g . ;
nawrocki , m. ; kulczykowski , m. ; matuszewski , m. ; szczytko , j. ; pacuski , w. ; pitka , b. _ phys .
b _ * 2017 * , _ 95 _ , 085429 gaj , j. a. ; kossut , j. _ introduction to the physics of diluted magnetic semiconductors _ ; springer science & business media , 2011 ; vol . 144 besombes , l. ; boukari , h. _ phys . rev .
b _ * 2014 * , _ 89 _ , 085315 sawicki , k. ; malinowski , f. k. ; gakowski , k. ; jakubczyk , t. ; kossacki , p. ; pacuski , w. ; suffczyski , j. _ appl .
lett . _ * 2015 * , _ 106 _ , 012101 besombes , l. ; lger , y. ; maingault , l. ; ferrand , d. ; mariette , h. ; cibert , j. _ phys .
lett . _ * 2004 * , _ 93 _ , 207403 kudelski , a. ; lematre , a. ; miard , a. ; voisin , p. ; graham , t. c. m. ; warburton , r. j. ; krebs , o. _ phys .
lett . _ * 2007 * , _ 99 _ , 247209 kobak , j. ; smoleski , t. ; goryca , m. ; papaj , m. ; gietka , k. ; bogucki , a. ; koperski , m. ; rousset , j .- g . ; suffczyski , j. ; janik , e. ; nawrocki , m. ; golnik , a. ; kossacki , p. ; pacuski , w. _ nat . commun . _
* 2014 * , _ 5 _ , 3191 smoleski , t. ; kazimierczuk , t. ; kobak , j. ; goryca , m. ; golnik , a. ; kossacki , p. ; pacuski , w. _ nat .
commun . _ * 2016 * , _ 7 _ , 10484 lafuente - sampietro , a. ; utsumi , h. ; boukari , h. ; kuroda , s. ; besombes , l. _ phys .
b _ * 2016 * , _ 93 _ , 161301 fainblat , r. ; barrows , c. j. ; hopmann , e. ; siebeneicher , s. ; vlaskin , v. a. ; gamelin , d. r. ; bacher , g. _ nano lett . _ * 2016 * , _ 16 _ , 63716377 wojnar , p. ; suffczy ski , j. ; kowalik , k. ; golnik , a. ; karczewski , g. ; kossut , j. _ phys . rev .
b _ * 2007 * , _ 75 _ , 155301 besombes , l. ; cao , c. l. ; jamet , s. ; boukari , h. ; fernndez - rossier , j. _ phys . rev .
b _ * 2012 * , _ 86 _ , 165306 jakubczyk , t. ; delmonte , v. ; fischbach , s. ; wigger , d. ; reiter , d. e. ; mermillod , q. ; schnauber , p. ; kaganskiy , a. ; schulze , j .- h . ;
strittmatter , a. ; rodt , s. ; langbein , w. ; kuhn , t. ; reitzenstein , s. ; kasprzak , j. _ acs photonics _ * 2016 * , _ 3 _ , 24612466 pacuski , w. ; jakubczyk , t. ; kruse , c. ; kobak , j. ; kazimierczuk , t. ; goryca , m. ; golnik , a. ; kossacki , p. ; wiater , m. ; wojnar , p. ; karczewski , g. ; wojtowicz , t. ; hommel , d. _ cryst .
growth des . _ * 2014 * , _ 14 _ , 988992 rousset , j .- g . ;
kobak , j. ; janik , e. ; parlinska - wojtan , m. ; slupinski , t. ; golnik , a. ; kossacki , p. ; nawrocki , m. ; pacuski , w. _ j. appl . phys . _ * 2016 * , _ 119 _ , 183105 | this work presents epitaxial growth and optical spectroscopy of cdte quantum dots ( qds ) in ( cd , zn , mg)te barriers placed on the top of ( cd , zn , mg)te distributed bragg reflector .
the formed photonic mode in our half - cavity structure permits to enhance the local excitation intensity and extraction efficiency of the qd photoluminescence , while suppressing the reflectance within the spectral range covering the qd transitions .
this allows to perform coherent , nonlinear , resonant spectroscopy of _ individual _ qds .
the coherence dynamics of a charged exciton is measured via four - wave mixing , with the estimated dephasing time @xmath0ps . the same structure contains qds doped with single mn@xmath1 ions , as detected in photoluminescence spectra .
our work therefore paves the way toward investigating and controlling an exciton coherence coupled , via @xmath2,@xmath3-@xmath4 exchange interaction , with an individual spin of a magnetic dopant . |
the present impact crater size frequency distribution , n is the result , on one hand , of a rate of crater formation , @xmath0 , and , on the other hand , the elimination of craters , as time goes by , due to effects like erosion and obliteration .
therefore if we want to understand the crater formation history we will need to know how these forming and erasing factors combine to create @xmath1 .
thus , in this work the above problem is analyzed , and in section 2 we find that n can be expressed in terms of @xmath0 and the fractional reduction of craters per unit of time , c. then , a simple model is discussed that describe the crater size distribution in mars data , collected by barlow @xcite , where it is assumed that @xmath0 is independent of time .
the above model is realistic , since according to several investigations @xmath0 has remained nearly constant for the last 3 to 3.5 billion years @xcite .
the simplest interpretation of this model implies that @xmath0 and @xmath2 are given as the following inverse power of the diameter , @xmath3 , of the crater : @xmath4 , @xmath5 . in section 3 the model
is applied to craters data on earths , and it is concluded that also in our planet @xmath5 .
this result is interpreted to mean that on mars and earth we have @xmath6 , or equivalently the crater mean life @xmath7 , with @xmath8 .
investigations of geometric properties of martian impact craters reflect values of the average height @xmath9 consistent with the above conclusion .
in what follows we will present theoretical and analytical curves which will reproduce the essential features of the martian crater - size frequency distribution empirical curves ( figure 1 ) , based on barlow s ( 1987 ) @xmath10 of about 42,000 impact craters .
the models will be derived using reasonable simple assumptions , that will allow us to relate the present crater population with the crater population at each particular epoch . to this end , let @xmath11 represents the number of craters of diameter @xmath12 formed during the epoch @xmath13 , where we are assuming that @xmath14 and @xmath15 are sufficiently large that is justified treating @xmath16 as a statistical continuous function , but , on the other hand , they should be sufficiently small ( @xmath17,@xmath18 ) to be able to treat them as differentials in the following discussion . this initial population will change as time goes on due to climatic and geological erosion , and the obliteration of old craters by the formation of new ones .
then , we expect that the change in @xmath19 during a time interval @xmath20 will be proportional to itself and @xmath20 : @xmath21 where c is the factor that takes into account the depletion of the craters , and should be a function of the diameter , since the smaller a crater is the most likely it will disappear . furthermore , c could also depend on time however , we will ignore such changes here , which we believe is a good starting approximation to the general problem .
it is easy to integrate equation ( 1 ) in time to obtain : @xmath22,\ ] ] @xmath23 equation ( 2 ) gives the number of craters , as a function of @xmath3 , observed at time t , that were produced at the time interval @xmath24 .
therefore the total contribution to the present ( t=0 ) population due to all the epochs @xmath25 is : @xmath26,\ ] ] or in the continuous limit @xmath27 , @xmath28 \,d{\tau},\ ] ] where @xmath29 @xmath30 is the rate of crater formation of diameter @xmath3 at the epoch @xmath31 , and @xmath32 is the total time of crater formation . in the next section
we will determine the function @xmath33 and @xmath0 for a model where we assumed that the rate of crater formation , @xmath0 , is independent of @xmath31 .
investigations of the time dependance of cratering rate of meteorites have concluded @xcite that the impact rate went through a heavy bombardment era that decayed exponentially until about 3 to 3.5 gy , and since then has remained nearly constant until the present .
therefore , for surfaces that are younger than 3 to 3.5 gy we can reasonably assume that @xmath0 is independent of @xmath31 , and hence from equation ( 5 ) immediately obtain @xmath34.\ ] ] we then find that the simplest model that essentially reproduces the data in figure 1 , for @xmath35 km , is given by equations ( 8)and ( 9 ) : @xmath36 @xmath37 we see that the theoretical curve ( 7 ) , shown in figure(2 ) , differs significantly from the observed curves for @xmath3 less than about 6 km .
however , according to barlow @xcite the empirical data is undercounting the actual crater population for @xmath3 less than @xmath38 km , and therefore no meaningful comparison is then possible between models and data for this region of small craters . equation ( 2 ) implies that the fraction of craters of diameter @xmath3 formed at each epoch @xmath31 that still survive at the present time @xmath39 is given by : @xmath40\,\approx\,{\rm exp}\left[{-\left(\frac{57}{d}\right)}^{2.5}\,\frac{\tau}{\tau_f } \right]\ ] ] and thus we have that the mean life for craters of diameter @xmath3 , @xmath41 , is @xmath42 hence , craters with @xmath43 57 km have @xmath44 , while @xmath45 the region @xmath46 km is approximately described by the limit of equation ( 7 ) when @xmath47 : @xmath48 which corresponds to a straight line of slope -4.3 in a log n vs log d plot , and that would be the form of equation ( 7 ) in the absence of erosion and obliterations ( @xmath49 ) .
hence , we have that the bending of the empirical curve ( figure 1 ) for @xmath50 is explained in this model as the result of the elimination of smaller craters as they get older .
we also see from equations(13 ) that when the effect of @xmath2 can be ignored we have @xmath51 , and therefore the actual crater density @xmath1 is proportional to the age of the underlying surface @xmath32 . on the other hand , when for smaller craters @xmath52\,<<\,1 $ ] we will have from ( 7 ) that @xmath53 and in this limit the crater density @xmath54 is proportional to the survival mean life , @xmath41 , of the craters of size @xmath3 .
thus , when saturation occurs and hence n is independent of @xmath32 , we have , instead , that @xmath1 is proportional to @xmath41 .
this feature is called by hartmann @xcite``crater retention age '' , and in mars this effect shows , according to this model , in craters smaller than about 57 km .
the model given by equations ( 7 ) , ( 8) , and ( 9 ) assumed a simple polynomial form for @xmath0 and @xmath2 , however , alternative models can be also considered .
for instance , by assuming that @xmath55,\ ] ] we will reproduce the mars crater data , exactly as in model given by equations ( 7),(8),(9 ) but now with @xmath56 , and the change in slope in figure ( 1 ) around @xmath57 km will now be interpreted as intrinsic behavior of @xmath58 rather than due to the erosion and obliteration of smaller craters .
how can we then discriminate between these two alternative views ? .
we see that in the model given by equation ( 7 ) the fraction of craters of a given diameter , @xmath3 , produced at a time @xmath31 , decreases with time according to equation ( 10 ) as @xmath59,\ ] ] while in the model of equation ( 15 ) this fraction is independent of time . therefore we can put to test the validity of equation ( 16 ) by studying crater size frequency distributions as a function of time .
this is possible to do in our planet , and in this section we will investigate the consistency of the hypothesis ( 16 ) with the earth craters data . thus consider the average diameter of craters observed today that were formed during a given time @xmath60 , which is given , according to equation ( 16 ) , by @xmath61 assuming that @xmath2 and @xmath0 behave in the form @xmath62 @xmath63 we can rewrite equation ( 17 ) in the form ( appendix ) @xmath64 where @xmath65 and @xmath66 is the gamma function
. equation ( 20 ) can be rewritten as @xmath67 which represents a linear relation between @xmath68 and @xmath69 with slope @xmath70 . in figure ( 3 ) we plot @xmath71 vs @xmath69 from data of crater size vs @xmath31 on earth , and the straight line best fitting gives @xmath72 , which is the value determined for model ( 7 ) for mars .
this result is interpreted as follows .
if we assume that , as expected , @xmath73 is a function of the volume of the crater , @xmath74 , that decreases with decreasing @xmath74 , then it is reasonable to expand it in terms of powers of @xmath75 , and thus we will have @xmath76 furthermore , for sufficiently small volumes we would have , as a good approximation to @xmath2 , that @xmath77 where we are writing @xmath78 with @xmath79 as the average height of the crater of size @xmath3 .
the comparison of equation ( 25 ) with equations ( 19 ) , with @xmath80 , imply that @xmath81 which is a prediction that can be investigated , and we have found that indeed equation ( 27 ) is consistent with results from studies of impact crater geometric properties on the surface of mars , by j.b .
garrin @xcite .
therefore it appears that the age distribution of craters on earth favor the simple model considered for mars , where there is an erosion and obliteration factor @xmath2 with the approximate form @xmath82 it is also suggested here that the above behavior for @xmath2 follows from a relation of the form @xmath83 with @xmath84 further investigations and observations of the crater data on the terrestrial planets , the moon and the asteroids are necessary for additional tests of the validity of the model ( 7 ) and its interpretation .
lets define barlow , n.g .
1988 , , 75 , 285 hartmann w.k .
1966b , , 5 , 406 neukum g. et al .
2001 , chronology and evolution of mars 55 bern : international space science intitute neukum g. 1983 , ludwig - maximillians - university of munich 186 .
ryder g. 1990 , eos , 71 , 313 hartmann w.k .
2002 , lunar and planetary science xxxiii 1876 garrin j.b .
2002 , lunar and planetary science xxxiii 1255 | we present a theoretical and analytical curve with reproduce essential features of the frequency distributions vs. diameter , of the 42,000 crater contained in the barlow mars catalog .
the model is derived using reasonable simple assumptions that allow us to relate the present craters population with the craters population at each particular epoch .
the model takes into consideration the reduction of the number of craters as a function of time caused by their erosion and obliteration , and this provides a simple and natural explanation for the presence of different slopes in the empirical log - log plot of number of craters ( n ) vs. diameter ( d ) . |
this paper grew out of a larger project , devoted to the investigation of so - called _ geometric curvature energies _ which include various types of geometric integrals , measuring the degree of smoothness and bending for objects that do not , at least a priori , have to be smooth . here
, we study the energy functional @xmath9 defined for a class @xmath10 of admissible , @xmath1-dimensional sets in @xmath11 . the precise definition of @xmath10 is given in section [ sec:2 ] ; we just mention now that for each @xmath12 a weak counterpart of the classic tangent plane is defined almost everywhere with respect to the @xmath1-dimensional hausdorff measure @xmath13 on @xmath2 . in other words , for @xmath13-a.e .
@xmath14 there is an @xmath1-plane @xmath15 such that the portion of @xmath2 near the point @xmath16 is close to the affine plane @xmath17 . the quantity @xmath18 in the integrand is referred to as the _ tangent - point radius _ and denotes the radius of the smallest sphere tangent to the affine plane @xmath19 and passing through @xmath20 .
( if @xmath20 happens to be contained in @xmath19 , then we set @xmath21 . )
thus , @xmath22 is defined a.e . on @xmath23 with respect to the product measure @xmath24
. notice that for any compact embedded manifold of class @xmath25 this repulsive potential @xmath4 is finite . for two - dimensional surfaces in @xmath26 , i.e. @xmath27 , @xmath28 , banavar et al .
@xcite suggested , in fact , the use of such tangent - point functions to construct self - interaction energies with non - singular integrands that do not require any sort of ad hoc regularization , in contrast to standard repulsive potentials .
the latter would penalize any two surface points that are close in euclidean distance , no matter whether these points are adjacent on the surface ( leading to singularities ) or belong to different sheets of the same surface .
our aim here is to show that for the infinite range of exponents @xmath5 finiteness of @xmath0 has three sorts of consequences for any admissible set @xmath29 : measure - theoretic , topological , and analytical . to see them in a proper perspective ,
let us give a plain description of the surfaces we work with .
our class @xmath10 consists of @xmath1-dimensional sets @xmath30 with finite measure @xmath31 on which we impose ( 1 ) a certain degree of flatness in the neighbourhood of many ( but a priori not all ! ) points of @xmath2 , and ( 2 ) some degree of connectivity . a priori , we allow for various self intersections of @xmath2 , and for singularities along low dimensional subsets . for the purposes of this introduction ,
however , it is enough to keep in mind the following examples of admissible surfaces ( more general examples are presented in section [ sec:2.3 ] ) : 1 .
if @xmath32 , where @xmath33 is arbitrary and all @xmath34 are compact , closed , embedded @xmath1-dimensional submanifolds of class @xmath6 such that @xmath35 whenever @xmath36 , then @xmath37 is admissible ; 2 .
if @xmath37 is as above , then @xmath38 is admissible whenever @xmath39 is a bilipschitz homeomorphism of @xmath11 .
the dimension @xmath1 and the codimension @xmath40 of @xmath2 in @xmath11 are fixed throughout the paper but otherwise arbitrary .
the reader may adopt for now the temporary definition @xmath41 it is easy to see that @xmath42 is a critical exponent here : for @xmath43 the energy @xmath0 is scale invariant , and for each @xmath44 a surface @xmath2 with a conical singularity at one point must have @xmath45 .
we prove in this paper that for @xmath46 all kinds of singularities are excluded .
in fact , upper bounds for @xmath0 lead to three kinds of effects .
firstly , measure - theoretic effects : the measure of @xmath2 contained in a ball or radius @xmath47 is comparable to @xmath48 on small scales that depend solely on the energy .
secondly topological effects : an admissible surface @xmath2 with finite @xmath4-energy has no self - intersections , it must be an embedded manifold , and finally , far - reaching analytical consequences : we have precise @xmath8 bounds for the charts in an atlas of @xmath2 .
-energy for all @xmath49 ) . ]
let us first state the results precisely and then comment on the proofs and discuss the relations of this paper to existing research .
keep in mind , though , that all results stated in the introduction will be proved for a more general class @xmath50 of admissible sets much larger than the preliminary class @xmath10 defined above ; see section [ sec:2.3 ] .
to get a first impression of other admissible sets have a look at figure [ fig:1 ] .
[ thm : uar ] assume that @xmath29 is an admissible @xmath1-dimensional surface in @xmath11 with @xmath51 , @xmath5 .
there exists a constant @xmath52 , depending _ only _ on @xmath53 and @xmath1 , such that @xmath54 for all @xmath55 and all radii @xmath56 [ conj1 ] ( here , and throughout the paper , @xmath57 denotes the closed ball of radius @xmath58 centered at @xmath16 . ) in other words : if @xmath12 has finite energy for some @xmath5 , then up to the length scale given by @xmath59 which depends _ only _ on the energy bound @xmath60 and the parameters @xmath61 , but _ not _ on @xmath2 itself isolated _ thin fingers _ , narrow tubes , and the likecannot form on @xmath2 .
the measure of the portion of @xmath2 inside the ball @xmath62 is at least as large as half of the measure of the @xmath1-dimensional equatorial cross - section of @xmath62 .
a similar lower estimate on the ahlfors regularity was proven by l. simon for smooth two - dimensional surfaces with finite willmore energy ( * ? ? ?
* corollary 1.3 ) ; see also the work of p. topping @xcite which even contains _ sharp _ lower bounds for the sum of local @xmath63-norm of the classic mean curvature and the area of the surface in a small ball .
mean curvature at a particular point @xmath16 on a smooth surface in @xmath26 may be viewed as the arithmetic mean of minimal and maximal normal curvature at @xmath16 .
@xmath64 , on the other hand , is a two - point function taking non - local interactions into account as well , but if one looks at the coalescent limits @xmath65 one obtains absolute values of intermediate normal curvatures at @xmath16 depending on the direction of approach as @xmath20 tends to @xmath16 ( cf .
* section 3.2 ) ) .
so , the local portion of our energy @xmath4 near @xmath16 may be regarded as another kind of averaging normal curvatures at @xmath16 , leading to density estimates as does the willmore functional .
the next result gives a quantitative description of flatness of @xmath2 , in terms of the so - called @xmath66-numbers introduced by p. jones .
[ thm : betas ] let @xmath29 be an admissible @xmath1-dimensional surface in @xmath11 with @xmath67 for some @xmath5 .
there exist two constants @xmath68 and @xmath69<\infty$ ] , both depending _ only _ on @xmath70 and @xmath71 , such that whenever the radius @xmath72 and the bound @xmath73 satisfy the balance condition @xmath74 then we have @xmath75 where @xmath76 denotes the grassmannian of all @xmath1-dimensional subspaces of @xmath11 .
thus , for small @xmath58 we have @xmath77 it is known that this condition alone does not suffice to conclude that @xmath2 is a topological manifold .
d. preiss , x. tolsa and t. toro @xcite , extending an earlier work of g. david , c. kenig and t. toro @xcite , study reifenberg flat sets @xmath2 whose @xmath66-numbers satisfy such estimates , see e.g. ( * ? ? ?
2.4 ) where it is proved that a decay bound for @xmath66 s combined with reifenberg flatness and an @xmath1-plane @xmath78 in @xmath57 is uniformly controlled , and small . ]
implies that @xmath2 must be a submanifold of class @xmath79 .
since @xmath12 might , at least a priori , have transversal self intersections , we do not have reifenberg flatness here , and a quick direct use of the results of @xcite is impossible . however , we are able to use the energy estimates and the information given by theorem [ thm : betas ] iteratively . extending the ideas from our earlier work ( * ? ? ?
* section 5 ) devoted to surfaces in @xmath26 , we prove here that at every point @xmath14 there exists the classic tangent plane @xmath80 , and that the oscillation of tangent planes along @xmath2 satisfies uniform hlder estimates .
this implies that each @xmath29 with @xmath81 must be an embedded @xmath1-dimensional manifold of class @xmath79 .
later on , working with graph patches of @xmath2 , we use slicing techniques and a bootstrap reasoning to improve and sharpen this information .
the following theorem is the main result of this paper .
[ thm : bootstrap ] let @xmath29 and @xmath82 for some @xmath5 .
then @xmath2 is an embedded submanifold of class @xmath8 , where @xmath83 .
in fact , there exist constants @xmath84 , depending only on @xmath85 and @xmath71 , with the following property : for each @xmath55 and each @xmath86 there exists an @xmath1-plane @xmath87 and a function @xmath88 of class @xmath8 such that @xmath89 where @xmath90 denotes the graph of @xmath91 , and @xmath92 where @xmath93 [ conj:2.3 ] we believe that the exponent @xmath94 , strictly larger than @xmath95 , is optimal here .
it is clear that finiteness of @xmath4 does not lead to @xmath96 regularity : consider a rotational cylinder closed with two hemispherical caps as an admissible surface @xmath2 of class @xmath25 but not in @xmath96 . for this particular surface inequality
is qualitatively optimal and , due to the factor @xmath97 and boundedness of @xmath98 , yields in fact _ lipschitz _ estimates for the gradient of local graph representations of @xmath2 .
please note two more things .
first , the exponent @xmath99 is computed according to the recipe used in the classic sobolev
morrey imbedding theorem in the supercritical case .
here , the dimension of the domain of integration , i.e. of @xmath100 , equals @xmath101 .
we have @xmath102 as @xmath103 ; for two - dimensional surfaces , the limiting case @xmath104 has been treated earlier in our papers @xcite and @xcite .
second , what we have learnt about @xmath2 is not limited to embeddedness and purely qualitative @xmath8 estimates .
it is clear that the bounds given by theorem [ thm : bootstrap ] are uniform in any class of surfaces with uniformly bounded energy @xmath4 . in other words ,
if @xmath105 satisfies @xmath106 then we can find two constants @xmath107 , depending only on @xmath108 and @xmath71 , such that each @xmath109 , where @xmath110 and @xmath111 , is obtained by a rigid motion of @xmath11 from a graph of a function @xmath112 which satisfies the uniform estimate @xmath113 , no matter how @xmath110 and @xmath111 have been chosen .
thus , a uniform upper bound on @xmath4 allows us to fix a uniform size of charts for all @xmath114 , and forces the equicontinuity of gradients of local graph representations of the surfaces @xmath115 . in a forthcoming paper @xcite
we show how to use this idea to obtain finiteness theorems for classes of @xmath116 embedded manifolds @xmath115 in @xmath11 satisfying a volume constraint and a uniform energy bound .
we do not know what happens in the critical case @xmath117 .
let us mention here one plausible conjecture that we can not prove at this stage .
every immersed @xmath1-dimensional @xmath6-manifold in @xmath118 with finite @xmath119-energy is embedded .
another , probably more difficult , question that we can not handle at present is the following : how regular are the minimizers of @xmath4 ( say , with upper bounds for the total measure , to prevent the decrease of energy caused by rescaling ) in isotopy classes of @xmath6 embedded manifolds ? are they @xmath25 ( this is optimal for _ ideal links _
@xcite corresponding to the case @xmath104 in dimensions @xmath27 and @xmath120 where contact phenomena are present ) ? or maybe @xmath121 , as minimizers of a mbius invariant knot energy in @xcite , @xcite ; see also @xcite , @xcite ?
in addition , s. blatt @xcite characterized all curves with finite mbius energy as embeddings in certain sobolev - slobodecki classes ; such a _ characterization _ of finite energy submanifolds for the tangent - point energy @xmath4 is presently not known .
our interest in this topic has been triggered by several factors .
they include manifold applications of menger curvature in harmonic analysis and geometric measure theory ( see e.g. the survey articles of p. mattila @xcite , @xcite , g. david @xcite and x. tolsa @xcite , and the literature cited therein , including j.c .
lger @xcite and the relation between @xmath122-rectifiability and @xmath63-integrability of menger curvature ) .
there are also works of different origin , investigating another geometric concept , the so - called global curvature introduced by gonzalez and maddocks @xcite .
the second author of the present paper took part in laying out the strict mathematical foundations for global curvature of rectifiable loops and its variational applications to elastic curves and rods with positive thickness ; see @xcite , @xcite , @xcite , @xcite , @xcite , @xcite .
part of this work , in turn , has been a starting point for our subsequent joint research devoted to various energies that , roughly speaking , interpolate between global curvature and menger curvature .
finiteness of these energies ; see e.g. @xcite , @xcite , @xcite , analogously to the case that we consider here , leads to an increase of regularity , to compactness effects , and yields a tool to control the amount of bending of non - smooth objects in purely geometric terms .
of the present paper , along with a few knot - theoretic applications of @xmath4 for curves . ] the novelty in the present paper is that we work in full generality , overcoming the difficulty that both the dimension and the codimension may be arbitrary . in an ongoing research
@xcite s. kolasiski obtains analogues of our results for basically the same admissible class of surfaces that we consider here , but for a different integral energy , defined as an @xmath123-fold integral ( with respect to @xmath13 ) over the set of all simplices with vertices on @xmath2 , directly extending our results in @xcite to surfaces of arbitrary dimension and codimension .
closely related research includes also g. lerman and j.t .
whitehouse @xcite , @xcite , who investigate a number of ingenious high - dimensional curvatures of menger type and obtain rectifiability criteria for @xmath58-dimensional subsets of hilbert spaces .
last but not least , the deep and classic paper of w. allard @xcite sets forth a regularity theory for @xmath124-dimensional varifolds whose first variation ( roughly : the distributional counterpart of mean curvature ) is in @xmath125 for some @xmath126 .
our regularity results bear some resemblance to his theorem 8.1 .
there are many differences , though , that remain to be fully understood .
it is clear that without some extra topological assumptions on @xmath2 finiteness of can not lead to the conclusion that @xmath2 is locally ( on a scale depending only on the energy ! ) homeomorphic to a disc ; one could punch an arbitrary number of ` holes ' in a smooth surface and this would just decrease the energy we work with . in allard
s case , once we fix a ball where appropriate density estimates hold and the weight @xmath127 of the varifold @xmath128 is close to the hausdorff measure of a disk , then the ` lack of holes ' is built into his assumption on the first variation @xmath129 of @xmath128 .
on the other hand , @xmath4 as a non - local energy in contrast to the locally defined distributional mean curvature averages over all global tangent - point interactions , which leads to self - avoidance and control over topology of the given surface .
admissible sets with finite @xmath4-energy are differentiable manifolds , which allard s result can not guarantee for varifolds with distributional mean curvature in @xmath125 , @xmath130 : there is a remaining ( small ) singular set , such that there is no control on the topology of the support of the varifold measure .
to possibly bridge the apparent gap between allard s work and our results we should note that versions of @xmath4 can be defined for general @xmath1-dimensional varifolds @xmath128 , via double integrals : the integrand @xmath98 can be treated as a function on _ points and planes_. it is an intriguing question whether finiteness of such integrals for some @xmath71 s lead to rectifiability criteria or to an improved regularity in the case of varifolds .
let us now informally sketch the main thread of our reasoning , and describe the organization of the paper in more detail .
we want to exclude self intersections and to have a quantitative description of flatness ; for this , theorem [ thm : betas ] would be a good starting point . the main idea behind its proof is pretty straightforward : if the @xmath66-numbers were too large , i.e. if @xmath55 but @xmath131 were not confined to a narrow tube @xmath132 around some affine @xmath1-plane @xmath133 , then , a simple argument shows that we would have two much smaller balls @xmath134 , say with @xmath135 such that for all @xmath136 and a nonzero proportion of @xmath137 the distance @xmath138 would be comparable to @xmath139 .
this yields @xmath140 , and a lower bound for the energy follows easily , leading to a contradiction , _ if _ the @xmath141 and the bound for the @xmath66 s are chosen in a suitable way which happens to be precisely the balance condition .
there is only one serious catch here : in order to make the resulting estimate uniform , and to be able to iterate it later on , we must guarantee that @xmath142 with some _ absolute _ constant @xmath143 .
and we want _ both _ @xmath144 and @xmath143 independent of a particular @xmath2 . for this
, we need theorem [ thm : uar ] which serves as the backbone for all the later constructions and estimates of the paper .
the overall idea here is somewhat similar to an analogous result in our work @xcite on menger curvature for surfaces in @xmath26 .
the main difference , however , leading to crucial difficulties , is that the _ codimension _ of @xmath2 may be arbitrary .
the proof of theorem [ thm : uar ] has two stages .
first , for a fixed generic point @xmath55 and all radii @xmath47 below a stopping distance @xmath145 , we control the size of projections of @xmath146 onto some @xmath1-plane @xmath147 ( which may vary as @xmath47 varies ) . here , topology comes into play . to grasp the essence of our idea , it is convenient to think of @xmath148 as in example ( i ) at the beginning of the introduction . for @xmath149 and for infinitesimally small radii @xmath47
we start with the tangent planes @xmath150 , and note that small @xmath151-spheres that are perpendicular to @xmath152 are nontrivially linked with @xmath153 .
then , for a sequence of growing radii @xmath154 , we rotate @xmath78 if necessary by a controlled angle to a new position @xmath155 in order to keep the projections large . at the same time , we construct a growing connected _ excluded region _
@xmath156 which does not contain any point of @xmath2 in its interior .
the size of the projections is controlled via a topological argument , involving the homotopy invariance of the linking number mod 2 of submanifolds .
the construction stops at some stopping distance @xmath157 , and yields another point @xmath158 with @xmath159 and two smaller balls @xmath160 , @xmath161 , where @xmath162 is an explicit absolute constant , such that @xmath163 for all @xmath164 and a significant proportion of @xmath165 . in the second stage we use the energy bounds to show that @xmath166 is positive and satisfies @xmath167 , where @xmath59 is the uniform constant given in theorem [ thm : uar ] .
the details of that part are given in section [ sec:4 ] ; see lemma [ 1/r - est ] , lemma [ mainlemma ] and their corollaries .
sections [ sec:2 ] and [ sec:3 ] contain all the necessary prerequisites and are included for the sake of completeness . in section [ sec:2 ]
we gather elementary estimates of angles between planes spanned by nearby almost orthogonal bases , and introduce the class of admissible sets whose definition is designed so that the above sketchy idea can be made precise . in section [ sec:3 ] we explain how the linking number mod 2 can be used for elements of @xmath10 , providing specific statements ( and short proofs ) for sake of further reference .
once theorem [ thm : uar ] is proved , we use the hausdorff convergence of excluded regions defined for generic points @xmath55 to obtain a corollary which , roughly speaking , ascertains that for _ every _ @xmath14 and @xmath168 there is some plane @xmath169 such that @xmath146 has large projection onto @xmath170 and is contained either in @xmath171 ( where , a priori , at this stage of the reasoning , @xmath2 might behave in a pretty wild way ) or in a narrow tubular region @xmath172 , for some specific constant @xmath173 .
a use of energy bounds yields now theorem [ thm : betas ] , and an iterative argument implies that in fact @xmath2 must locally be a @xmath79 graph .
all this is done in section [ sec:5 ] .
embeddedness of @xmath2 is established here , too .
finally , in section [ sec:6 ] , we prove theorem [ thm : bootstrap ] and sharpen the hlder bounds . to this end , we show that if @xmath174 is a graph of @xmath175 , then @xmath176 satisfies an improved estimate , @xmath177 where @xmath178 stands for the supremum of oscillations of @xmath176 over all possible balls of radius @xmath179 , and @xmath60 is the portion of energy coming from some ball containing @xmath180 .
the point is that holds for some @xmath181 , so that for @xmath175 the first term of the right hand side can be viewed as an unimportant , small scale perturbation .
the main idea behind is that when the integral average of @xmath182 is bounded by @xmath183 , then there are numerous points @xmath184 in small balls around the @xmath185 , @xmath186 , where @xmath187 .
a geometric argument implies that for such points @xmath188 can be controlled by the second term in the right hand side of , and a routine iterative reasoning , with a certain morrey
campanato flavour , allows us to get rid of the @xmath189 and finish the whole proof
. * acknowledgement . *
the authors would like to thank the deutsche forschungsgemeinschaft , polish ministry of science and higher education , and the alexander von humboldt foundation , for generously supporting this research .
substantial parts of this work have been written while the first author has been staying at the rwth aachen university in the fall of 2009 ; he is very grateful to his german colleagues for their hospitality .
we write @xmath62 to denote the _ closed _ ball in @xmath190 , with center @xmath16 and radius @xmath191 .
the volume of the unit ball in @xmath192 is denoted by @xmath193 . for a closed set @xmath39 in @xmath190 we set @xmath194
@xmath76 denotes the grassmannian of all @xmath1-dimensional linear subspaces of @xmath190 . if @xmath87 , then @xmath195 denotes the orthogonal projection of @xmath190 onto @xmath78 , and @xmath196 is the orthogonal projection onto @xmath197 . for two planes
@xmath198 we define their distance ( or angle ) @xmath199 where the right hand side is the usual norm of the linear map @xmath200 .
the grassmannian @xmath76 equipped with this metric is compact . finally , we use the following variant of p. jones
beta - numbers ( see david s and semmes monograph ( * ? ? ?
* chapter 1 , sec .
1.3 ) for a discussion ) : @xmath201 [ sec:2.2 ] throughout most of the paper , we shall work with estimates of various geometric quantities related to two planes in @xmath76 that form a small angle . for sake of further reference
, we gather here several such estimates .
we also fix specific constants ( which in all cases are far from being optimal ) that are needed later , in more involved computations in sections [ ahlfors][slicing ] .
all proofs are elementary , but we provide them to make the exposition complete . [ sequences ] assume that @xmath202 and a sequence of nonnegative numbers @xmath203 satisfies @xmath204 , @xmath205 then for each @xmath206 we have @xmath207 , @xmath208 . * proof .
* one proceeds by induction . clearly , for @xmath209 we just need @xmath210 . for each @xmath211 the recursive condition for @xmath212 yields , under the inductive hypothesis , @xmath213 now , @xmath214 guarantees that @xmath215 and@xmath216 .
thus , yields @xmath217 .
@xmath218 [ bases - angle ] if @xmath219 have orthonormal bases @xmath220 and @xmath221 such that @xmath222 for each @xmath223 , then @xmath224 . *
proof . * take an arbitrary unit vector @xmath225 and estimate @xmath226 , expressing both projections in orthonormal bases @xmath227 and @xmath228 .
@xmath218 [ gramschmidt ] assume that @xmath229 . if @xmath230 is an orthonormal basis of a subspace @xmath231 and @xmath232 satisfy @xmath233 , then @xmath234 are linearly independent
. moreover , the gram - schmidt orthogonalization process @xmath235 yields vectors @xmath236 @xmath237 that satisfy @xmath238 where @xmath239 . if @xmath240 , then @xmath241 with @xmath242 . [ bases ] * proof . * as @xmath243 for all @xmath244 , we have @xmath245 . therefore , @xmath246 for @xmath247 and @xmath248 . using this observation ,
one proves by induction ; assuming for @xmath249 and all @xmath250 , we obtain @xmath251 where the last inequality follows from elementary computations ( the estimate is not sharp ) .
this yields the first part of for @xmath252 ; the second one follows from the triangle inequality .
in particular , we also have @xmath253 , and therefore @xmath254 are linearly independent .
setting @xmath255 , we easily conclude the proof of the whole lemma .
( to check inequality , apply lemma [ bases - angle ] and note that @xmath256 . )
@xmath218 [ bases2 ] let @xmath257 be the constant defined in lemma [ bases ] above .
assume that 1 .
there exist orthonormal @xmath258 such that @xmath259 for @xmath260 , and @xmath261 ; 2 .
@xmath262 for all @xmath260 , and @xmath263 .
then the subspaces @xmath264 and @xmath265 belong to @xmath76 , and we have @xmath266 with @xmath267 .
* proof . *
it follows from lemma [ bases ] that @xmath268 .
we use again the gram - schmidt algorithm and set @xmath269 , @xmath270 , @xmath271 then , @xmath272 and @xmath184 form orthogonal bases of @xmath170 and @xmath273 , respectively .
inequality yields @xmath274 satisfies the assumptions of lemma [ sequences ] with @xmath275 and @xmath276 .
for @xmath209 we have @xmath277 .
let @xmath278 . for all @xmath279 in the annulus @xmath280
we have @xmath281 and , for @xmath282 , @xmath283 thus , since @xmath284 , we obtain @xmath285 for each @xmath286 , in particular for @xmath287
. therefore certainly @xmath288 , @xmath289 , by lemma [ sequences ] . keeping in mind that @xmath290 , we obtain @xmath291 the inequality @xmath292 follows now from lemma [ bases - angle ] . @xmath218 the next two lemmata are concerned with the set @xmath293 where @xmath294 form a small angle so that @xmath295 restricted to @xmath296 is bijective . since @xmath297 is convex , closed and centrally symmetric in @xmath11 . ] for each @xmath186 , we immediately obtain the following : [ cccs ] @xmath298 is a convex , closed and centrally symmetric set in @xmath190 ; @xmath299 is a convex , closed and centrally symmetric set in @xmath300 .
the next lemma and its corollary provide a key tool for bootstrap estimates in section [ slicing ] .
[ proj - slab ] let @xmath301 and @xmath302 denote the constants defined in lemma [ bases ] .
if @xmath303 satisfy @xmath304 , then there exists an @xmath305-dimensional subspace @xmath306 such that @xmath307 * proof .
* let @xmath308 ; we have @xmath309
. for @xmath186 set @xmath310 .
then , @xmath311 is the orthogonal sum of @xmath170 and @xmath312 .
let @xmath313 ; by construction , @xmath314 .
finally , let @xmath315 be the orthogonal complement of @xmath316 in @xmath11 , so that @xmath11 is equal to @xmath317 , and the spaces @xmath318 are pairwise orthogonal .
it is now easy to see , directly by definition , that @xmath319 . * step 1 .
* we shall first show that there exists a vector @xmath320 such that @xmath321 fix an orthonormal basis @xmath322 of @xmath323 . since @xmath324
, we have @xmath325 . applying lemma [ bases ] with @xmath326 to @xmath323 and @xmath327
, we check that the @xmath328 , @xmath329 form a basis of @xmath327 , and @xmath330 for at least one @xmath331 .
assume w.l.o.g . that this is the case for @xmath332 .
thus , there are no points of @xmath327 in the interior of @xmath333 , and therefore there are no points of @xmath327 in the interior of the cone @xmath334 set @xmath335 .
then @xmath336 , so that the closed ball @xmath337 . hence , @xmath338 let @xmath339 denote the segment @xmath340 ; we claim that @xmath341 . to check this
, we argue by contradiction . if @xmath342 and @xmath343 , then , decomposing @xmath344 where @xmath345 , @xmath346 , and @xmath347 , we have @xmath348 for some @xmath179 , @xmath349 . as @xmath350 and @xmath351 , this yields @xmath352 and @xmath353 for some @xmath354
now , shows that if @xmath355 , then we must have @xmath356 .
this , however , yields @xmath357 , which contradicts the assumption @xmath342 .
thus , @xmath358 satisfies .
now , in order to prove the existence of the desired subspace @xmath306 , consider the function @xmath359 defined on the unit sphere in @xmath360 .
if @xmath361 , then @xmath362 .
since @xmath363 , we have @xmath364 everywhere .
note that if @xmath365 then @xmath366 and @xmath367 for every @xmath368 .
( thus , @xmath369 is the ` exit time ' that we need to leave @xmath299 , travelling with unit speed from @xmath370 in the direction given by @xmath371 . )
* we shall first show that there exists a vector @xmath372 , @xmath373 , such that @xmath374 since , by step 1 , we have @xmath375 , it is of course enough to show that @xmath376 is achieved on the unit sphere of @xmath360 . take a sequence of unit vectors @xmath377 such that @xmath378 ; passing to a subsequence , we can assume @xmath379 as @xmath380 .
suppose now that @xmath381 .
then , for some fixed @xmath73 we have @xmath382 for all @xmath383
. consider the points @xmath384 and @xmath385 in @xmath299 .
then , @xmath386 as @xmath380 . by definition of @xmath47 and convexity
, @xmath387 thus , for all @xmath388 such that @xmath389 and @xmath390 the point @xmath391 is _ in the interior of _ @xmath392 .
then , however , by definition of @xmath393 we obtain @xmath394 , a contradiction which shows that @xmath395 .
* step 3 . *
@xmath396 satisfies the desired condition .
note that @xmath273 is chosen so that the set @xmath397 is the ` narrowest strip in @xmath360 ' containing @xmath398 .
indeed , if there was a point @xmath399 , then , taking the straight line through @xmath20 and @xmath400 ( with @xmath401 ) , one could easily reach a contradiction : take a unit vector @xmath402 in @xmath403 , @xmath404 , and use convexity of @xmath398 to show that then @xmath405 ( for otherwise the straight segment connecting the points @xmath20 and @xmath406 contained in @xmath398 would intersect the ray @xmath407 contradicting the definition of @xmath408 ) .
this completes the proof of lemma [ proj - slab ] . @xmath218
the next lemma is practically obvious .
[ strip - ball ] suppose that @xmath409 and a set @xmath410 is contained in @xmath411 for some @xmath412 , where @xmath273 is an @xmath305-dimensional subspace of @xmath170 .
then @xmath413 for each @xmath414 and each @xmath415 .
* decomposing each @xmath416 as @xmath417 , one sees that @xmath418 is contained in a rectangular box with @xmath305 sides parallel to @xmath273 and of length @xmath419 and the remaining side perpendicular to @xmath273 and of length @xmath420.@xmath218 [ proj - meas ] if two planes @xmath303 satisfy @xmath421 , then @xmath422 for every @xmath13-measurable set @xmath423 .
* proof . *
it is enough to prove the inequality when @xmath424 is the @xmath1-dimensional unit cube in @xmath296 , and @xmath425 .
fix an orthonormal basis @xmath426 of @xmath296 and let @xmath427 for @xmath428 . then , by hadamard s inequality , @xmath429 let us now give a precise definition of the class of admissible surfaces
. intuitively speaking , the energy functional @xmath4 can be defined for all @xmath430 compact , @xmath31 , which are a union of continuous images of @xmath1-dimensional closed manifolds of class @xmath6 , satisfying two additional conditions .
one of them ensures that @xmath2 is pretty flat near @xmath13almost all its points @xmath16 , so that we have , in a sense , a ` mock ' tangent plane @xmath15 to @xmath2 at @xmath16 .
a priori , @xmath15 does not even have to coincide with the classic tangent plane .
the second condition guarantees , as we shall see later , that small @xmath151-dimensional spheres centered at @xmath16 and parallel to @xmath431 are nontrivially linked with the surface @xmath2 .
as we have already said in the introduction , it might be convenient to think of the following example .
assume that @xmath432 are embedded , compact , closed @xmath1-dimensional @xmath6-submanifolds of @xmath11 .
they might intersect each other but only along sets of @xmath1-dimensional measure zero , so that @xmath433 whenever @xmath36 . then , for any bilipschitz homeomorphism @xmath434 , @xmath435 is an admissible surface .
the definition of admissible surfaces involves the notion of degree modulo 2 ; here are its relevant properties .
deg2_2 [ degmod2 ] let @xmath436 be compact manifolds of class @xmath6 without boundary and of the same dimension @xmath249 .
assume that @xmath437 is connected .
there exists a unique map @xmath438 such that : 1 . if @xmath439 , then @xmath440 is surjective ; 2 .
if @xmath441\to n$ ] is continuous , @xmath442 and @xmath443 , then @xmath444 ; 3 .
if @xmath445 is of class @xmath6 and @xmath446 is an arbitrary regular value of @xmath91 , then @xmath447 for a proof , see e.g. the monograph of m.w .
hirsch ( * ? ? ?
* chapter 5 ) , theorem 1.6 and the surrounding comments .
blatt gives a detailed presentation of degree modulo 2 ( even for noncompact manifolds ) in his thesis @xcite .
now , let @xmath448 and let @xmath339 be a finite or countable set of indices .
[ admissible ] we say that a compact set @xmath449 is an admissible ( @xmath1-dimensional ) surface of class @xmath50 if the following conditions are satisfied .
( h1 ) ahlfors regularity .
: : @xmath31 and there exists a constant @xmath450 such that @xmath451 ( h2 ) structure .
: : there exist compact , closed @xmath1-dimensional manifolds @xmath153 of class @xmath6 and continuous maps @xmath452 , @xmath110 , where @xmath339 is at most countable , such that @xmath453 , where @xmath454 .
( h3 ) mock tangent planes and @xmath455-flatness .
: : there exists a dense subset @xmath456 with the following property : @xmath457 and for each @xmath458 there is an @xmath1-dimensional plane @xmath459 and a radius @xmath460 such that @xmath461 ( h4 ) linking . : : if @xmath458 and @xmath462 is given by ( h3 ) above , then there exists an @xmath463 such that the map is well defined , as @xmath464 , and @xmath465 by virtue of ( h3 ) . ]
@xmath466 satisfies the condition @xmath467.(here we use the notation @xmath468 for @xmath469 , @xmath470 and @xmath471 . )
if @xmath2 is a compact , connected manifold of class @xmath6 without boundary , embedded in @xmath11 , then @xmath472 for every @xmath473 .
we can take @xmath474 , @xmath475 , @xmath476 , and @xmath477 ; ( h1 ) and ( h2 ) follow .
it is clear that condition ( h3 ) is satisfied if we choose @xmath478 for @xmath55 .
condition ( h4 ) is then satisfied , too . in this simple model case ( h4 )
ascertains that small @xmath151-dimensional spheres centered at the points of an embedded manifold @xmath2 and contained in planes that are normal to @xmath2 are linked with that manifold ; see e.g. @xcite for the definition of linking coefficient .
( we do not assume orientability of @xmath153 ; this is why degree modulo 2 is used . ) note that if @xmath479 is fixed , then we are not forced to set @xmath480 ; conditions ( h3 ) and ( h4 ) in this example would be satisfied also if @xmath15 were sufficiently close to @xmath80 .
thus , for given @xmath2 satisfying ( h1 ) and ( h2 ) the choice of @xmath15 _ does not have to be unique .
_ the next two examples show that we can allow @xmath2 to have several @xmath6-pieces that intersect along sets of @xmath1-dimensional measure zero , and are embedded away from those sets .
[ c1union ] if @xmath2 is connected , @xmath481 , where @xmath115 are compact , connected manifolds of class @xmath6 without boundary , embedded in @xmath11 , and moreover @xmath482 then @xmath483 for every @xmath473 .
the set @xmath339 is now equal to @xmath484 and we set @xmath485 for each @xmath458 there is a unique @xmath388 such that @xmath486 and we take @xmath487 .
conditions ( h1 ) and ( h2 ) are clearly satisfied with @xmath474 and @xmath488 for @xmath489 , and the verification of ( h3 ) and ( h4 ) is similar to the previous example ; one just has to ensure that for @xmath490 the radius @xmath491 is chosen so that @xmath492 .
let the @xmath153 , @xmath493 , be compact , connected @xmath1-dimensional @xmath6-manifolds without boundary .
let @xmath494 be @xmath6-immersions , and let @xmath495 for @xmath489 .
if @xmath496 is connected , @xmath497 and @xmath498 then @xmath483 for every @xmath473 .
we leave the verification to the reader .
it is also clear that the condition that all maps @xmath499 in the previous example be of class @xmath6 is too strong .
we can allow @xmath500 to have large intersections with other @xmath501 as long as the flatness condition in ( h3 ) is satisfied , and we need @xmath15 only for a.e .
thus , it is relatively easy to give more examples of admissible surfaces .
if @xmath502 is a bilipschitz homeomorphism , and we take @xmath2 as in example [ c1union ] , then @xmath503 indeed , we then set @xmath504 .
let @xmath505 be given by .
to define @xmath506 , we use compactness and smoothness of the @xmath115 to fix a radius @xmath191 with the following property : for each @xmath507 and each point @xmath508 there is a function @xmath509 , @xmath510 , such that @xmath511 and @xmath512 we also let @xmath513 for @xmath514 .
now , a point @xmath515 is in @xmath506 if @xmath516 ( we exclude the intersections ) , and moreover there is @xmath517 and an @xmath508 such that @xmath518 for some @xmath519 which is a point where @xmath520 is differentiable .
it follows from rademacher s theorem that @xmath521 has full measure and is dense in @xmath522 .
condition ( h1 ) is also satisfied , since bilipschitz maps distort the measure @xmath13 at most by a constant factor . to check ( h3 ) ,
one notes that as @xmath523 is bilipschitz , its differential @xmath524 must have maximal rank @xmath1 at all points where it exists ; it is then a simple exercise to check that for @xmath525 the plane @xmath526 satisfies all requirements of condition ( h3 ) for all @xmath527 to check ( h4 ) , one can use the homotopy invariance of the degree ; we leave the details to the reader .
we do not have a simple characterization of the class of admissible surfaces .
however , it contains weird countably rectifiable sets , too .
\(a ) for @xmath528 let @xmath529 , @xmath530 , @xmath531 , @xmath532 ( so that the spheres @xmath115 and @xmath533 touch each other at @xmath534 ) , and let @xmath535 .
set @xmath536 . then , @xmath2 is an admissible surface , belonging to @xmath50 for each @xmath537 .
all points of @xmath2 except @xmath370 and the @xmath391 for @xmath538 belong to @xmath539 . for @xmath458 , one verifies ( h3 ) and ( h4 ) just as as in example [ c1union ] .
moreover , ( h1 ) is also valid . to see this , fix @xmath540 . if @xmath111 and @xmath541 , then @xmath542 by the standard formula for the area of a spherical cap .
if @xmath543 but @xmath544 , then it is possible to check that the largest of all @xmath501 completely contained in @xmath545 has @xmath546 $ ] .
estimating @xmath547 from below by @xmath548 , we obtain ( h1 ) for @xmath549 ; a similar argument works for @xmath550 .
\(b ) a modification of the above example yields the following ( see figure [ fig:1 ] ) : set @xmath551 where @xmath552 is the surface of a cube of side length @xmath553 , and @xmath554 is a segment of length @xmath122 .
to be more specific , @xmath555 ^ 3\bigr)\subset { { { { \mathbb r}}}}^3 $ ] , and we let @xmath552 be a translated copy of @xmath556 ^ 3 $ ] , @xmath557 ^ 3\bigr ) + ( 1 - 2^{-i})(e_1 + 2e_3 ) + ( k-1 ) 2^{-i}e_2,\ ] ] so that , for fixed @xmath388 , the @xmath552 with @xmath558 form a layer of touching cubes stacked on top of the union of all the previous @xmath559 , @xmath560 and @xmath561 . finally , set @xmath562 $ ] ; we add this segment to the union of all @xmath552 to make @xmath2 closed .
it is possible to check that if @xmath539 is equal to the union of the interiors of all the faces of the cubes ( which is a dense set of full surface measure in @xmath2 ) , then ( h3 ) and ( h4 ) are satisfied .
ahlfors regularity of @xmath2 can be checked as in ( a ) above .
the mock tangent planes @xmath15 are not unique in the definition of the class @xmath50 but _ a posteriori _ it follows from theorem [ thm : bootstrap ] that if @xmath563 , then for any @xmath564 there is at most _ one _ choice of the @xmath15 ( up to a set of zero measure ) if one wants @xmath0 to be finite .
thus , finiteness of the energy is a very strong assumption : it forces us to abandon the apparent freedom of choice of the @xmath15 , and forces @xmath2 to be a single embedded manifold , with a controlled amount of bending at a given length scale , depending only on the energy .
to guarantee the existence of big projections later on , we shall need a topological invariant , which is a version of the linking number modulo 2 .
assume that @xmath2 is an admissible surface of class @xmath50 and @xmath565 is a compact , closed @xmath151-dimensional manifold of class @xmath6 , embedded in @xmath11 and such that @xmath566 . for each @xmath110 and for the manifolds
@xmath153 which satisfy ( h2 ) and ( h4 ) of definition [ admissible ] , let @xmath567 we set @xmath568 0 & \quad\mbox{if $ \ { { \rm deg}\,}2 g_i=0 $ for all $ i\in i$. } \end{array}\right.\ ] ] we shall use this definition mostly in the case where @xmath565 is a round sphere ( or an ellipsoid with ratio of axes very close to @xmath122 ) contained in some @xmath569-affine plane in @xmath11 .
we need the following four properties of this invariant .
[ h - inv - link ] let @xmath563 and let @xmath437 be a compact , closed @xmath151-dimensional manifold of class @xmath6 , and let @xmath570 for @xmath571 , where @xmath572 is a @xmath6 embedding of @xmath437 into @xmath11 such that @xmath573 . if there is a homotopy @xmath574\to { { { { \mathbb r}}}}^n\setminus \sigma\ ] ] such that @xmath575 and @xmath576 , then @xmath577 * proof . * note that the mappings @xmath578 are such that @xmath579 is homotopic to @xmath580 for each @xmath581 .
thus , the lemma follows directly from theorem [ degmod2 ] ( ii ) .
@xmath218 [ lk=1 ] assume that @xmath582 and @xmath458 .
then for all @xmath583 and for @xmath584 we have @xmath585 where @xmath462 is the constant in condition ( h3 ) of definition [ admissible ] .
* proof . * due to condition ( h3 ) , each sphere @xmath586 with @xmath587 can be deformed homotopically to @xmath588 ; we simply adjust the radius , changing it linearly . since the image of that homotopy is disjoint from @xmath2 , the lemma follows from ( h4 ) and lemma [ h - inv - link ] .
@xmath218 [ lk=0 ] if @xmath563 , @xmath589 and @xmath590 , then @xmath591 for each plane @xmath592 . * proof .
* fix an arbitrary @xmath110 .
set @xmath593 and let @xmath594 be defined by .
we shall prove that @xmath595 . to this end , consider the homotopy @xmath596 \to { { \mathbb s}}^{n-1}\ ] ] given by @xmath597\ , .
\label{hwzt}\ ] ] it is easy to see that @xmath170 is well defined and continuous ; we have @xmath598 .
thus , by theorem [ degmod2 ] ( ii ) , @xmath599 for each @xmath600 $ ] .
if @xmath601 , then the image of @xmath602 in @xmath603 is the same as image of @xmath115 under the map @xmath604 which is lipschitz in a neighbourhood of @xmath115 . since @xmath605 , @xmath602 can not be surjective , since the @xmath606-measure of its image is zero .
thus , we obtain @xmath607 .
if @xmath608 , we first approximate @xmath499 by a smooth map @xmath609 , so that @xmath610 . then , @xmath611 satisfies to @xmath612 along a segment , which avoids @xmath565 , as @xmath610 and @xmath613 . ] @xmath614 .
next , we define @xmath615 by with @xmath499 s replaced by @xmath616 s .
if @xmath617 has no regular points (= points where the differential has rank equal to @xmath618 ) , then its jacobian is zero , and @xmath617 is not surjective . if @xmath617 has at least one regular point , then since @xmath565 consists of two distinct points @xmath619 and @xmath620 , we see each regular value of @xmath621 has an even number of preimages in @xmath622 . hence , in either case @xmath623 . [ link->proj ] if @xmath563 and for some @xmath624 , @xmath191 and @xmath592 we have @xmath625 then the disk @xmath626 contains at least one point of @xmath2 .
* suppose this were not the case .
then @xmath627 for some @xmath73 .
we deform continuously the sphere @xmath628 to @xmath629 , staying all the time in @xmath630 , at the distance at least @xmath631 to @xmath2 .
this yields @xmath632 by lemma [ h - inv - link ] and lemma [ lk=0 ] , a contradiction . @xmath218
[ sec:4 ] we introduce here the notion of a _
good couple_. it expresses in a quantitative way the following rough idea : if there are two points @xmath633 such that the distance from @xmath20 to a substantial portion of the affine planes @xmath634 ( where @xmath635 is very close to @xmath16 ) is comparable to @xmath636 , then a certain portion of energy comes _ only from the neighbourhood of points forming such a configuration_. quantifying this , and iterating the resulting information in the next section , we eventually are able to pinpoint some of the local and global properties of the surface .
recall that @xmath637 stands for the orthogonal projection onto @xmath638 .
[ couples ] we say that @xmath639 is a @xmath640good couple if and only if the following two conditions are satisfied : 1 .
@xmath641 ; 2 . the set @xmath642 we shall be using this definition for fixed @xmath643 depending only on @xmath644 and @xmath1 . intuitively , good couples force the energy to be large .
once we have a @xmath640good couple , then @xmath98 must be @xmath645 on a set in @xmath23 of @xmath24-measure roughly @xmath646 .
thus , for @xmath5 , one can not have @xmath58 small and @xmath0 small simultaneously .
we quantify that in lemma [ low - de - bounds ] .
[ 1/r - est ] if @xmath647 is a @xmath640good couple with @xmath648 and an arbitrary @xmath649 $ ] , then @xmath650 for all @xmath651 and @xmath652 . *
* for @xmath653 as above we have @xmath654 moreover , @xmath655 .
thus , by , @xmath656 to prove uniform ahlfors regularity , we shall demonstrate that each @xmath2 with finite energy can not penetrate certain conical regions of @xmath11 .
the construction of those regions will guarantee that in a neighbourhood of each point @xmath657 the projections of @xmath2 onto suitably chosen @xmath1-planes passing through @xmath16 are large , and a bound on the energy will allow us to prove that such neighbourhoods have to be uniformly large , independent of the particular point @xmath458 we have chosen . for a plane @xmath409 and @xmath448 we set @xmath658 ( these are closed ` double cones ' with ` axis ' equal to @xmath659 .
note that if @xmath660 , then the interior of @xmath661 and of @xmath662 is connected . )
we shall also use the intersections of cones with annuli , @xmath663 [ mainlemma]@xmath664 there exist constants @xmath665 which depend only on @xmath70 , and have the following property . for every @xmath563 and every @xmath458 there exist @xmath666 and @xmath158 such that 1 .
@xmath667 is a @xmath668good couple ; 2 .
for each @xmath669 $ ] there exists a plane @xmath670 such that @xmath671 and therefore @xmath672 for all @xmath673 ; 3 .
the plane @xmath674 is such that @xmath675 .
each disk @xmath676 , and radius @xmath47 such that @xmath677 contains at least one point of @xmath2 .
the number @xmath145 is referred to as the _ stopping distance_. it can be checked that the condition for the radii of disks containing points of @xmath2 is equivalent to @xmath678 [ low - de - bounds ] let @xmath679 be the constant of lemma [ mainlemma ] . if @xmath563 for some @xmath680 $ ] and @xmath681 for some @xmath682 , then the numbers @xmath145 satisfy @xmath683 moreover , we have @xmath684 where @xmath685 for @xmath686 and @xmath687 as in lemma [ mainlemma ] .
the rest of this section is organized as follows .
we prove lemma [ mainlemma ] in the next subsection .
then , in subsection [ sec:4.4 ] , we derive lemma [ low - de - bounds ] from lemma [ mainlemma ] , and prove theorem [ thm : uar ] .
the proof of lemma [ mainlemma ] is similar to the proof of theorem 3.3 in our paper @xcite .
it has algorithmic nature .
proceeding iteratively , we construct an increasingly complicated set @xmath505 which is centrally symmetric with respect to @xmath16 and its intersection with each sphere @xmath688 is equal to the union of two or four spherical caps .
the size of these caps is proportional to @xmath47 but their position may change as @xmath47 grows from @xmath370 to the desired stopping distance @xmath145 .
the interior of @xmath505 contains no points of @xmath2 but it contains numerous @xmath151-dimensional spheres which are nontrivially linked with @xmath2 .
eventually , this ensures parts ( ii)(iv ) of the lemma . to find a good couple @xmath667 , we construct @xmath505 so that @xmath689 is nonempty , and one of the points in this intersection , or one of nearby points of @xmath2 will be good enough for our purposes . the rest of this subsection is organized as follows .
we first list the conditions that have to be satisfied by @xmath690 , @xmath455 , and @xmath691 .
then , we set up the plan of the whole inductive construction and describe the first step in detail .
next , we give the stopping criteria . analyzing them ,
we demonstrate that when the iteration stops , then ( i)(iv ) of the lemma are satisfied .
if the stopping criteria do not hold , then we perform the iterative step .
due to the nature of stopping criteria the total number of steps in the iteration must be finite , since @xmath2 is compact .
we fix a sufficiently small @xmath479 ( to be specified soon ) and assume that @xmath2 belongs to the class @xmath50 of all admissible surfaces defined in section [ sec:2 ] . for the sake of simplicity ,
we assume throughout the whole proof that @xmath692 . * the constants .
* we fix the three constants @xmath693 in @xmath694 so that several conditions are satisfied .
we first pick @xmath690 and @xmath455 so small that @xmath695 where @xmath257 , @xmath696 and @xmath697 denote the constants ( depending only on @xmath1 ) introduced in lemma [ bases ] and lemma [ bases2 ] in section [ sec : linear ] . , is needed later , in applications in section [ sec:5 ] . here , in this proof , just the condition @xmath698 would be sufficient . ] without loss of generality we can also assume that @xmath699 and @xmath700 next , we let @xmath701 be the minimal number such that there exist @xmath701 balls @xmath702 that form a covering of the whole grassmannian @xmath76 .
since @xmath690 depends only on @xmath1 , this number @xmath701 depends in fact only on @xmath70 . finally ,
once @xmath701 is fixed , we let @xmath703 * the construction . * proceeding iteratively , we shall construct three finite sequences : * of compact , connected , centrally symmetric sets @xmath704 , * of @xmath1-planes @xmath705 and @xmath706 such that the angle @xmath707 for each @xmath708 , where @xmath257 is the small constant of lemma [ bases ] , * and of radii @xmath709 , where @xmath710 , so @xmath711 will provide the desired stopping distance for @xmath16 as claimed in the statement of lemma [ mainlemma ] . everywhere below in this subsection
, we write @xmath712 and @xmath713 .
these sequences will be shown to satisfy the following properties : * ( diameter of @xmath714 grows geometrically ) .
* we have @xmath715 and @xmath716 for @xmath717 .
moreover @xmath718 * * ( large ` conical caps ' in @xmath714 and @xmath719).**@xmath720 and @xmath721 * * ( @xmath2 does not enter the interior of @xmath714 or @xmath722).**@xmath723 moreover , we have @xmath724 * * ( points of @xmath725 on @xmath726).**the intersection @xmath727 is nonempty for each @xmath489 . * ( linking ) . *
if @xmath728 satisfies @xmath729 and the radius @xmath191 is chosen such that the @xmath151-dimensional sphere @xmath730 is contained in the interior of @xmath731 , then @xmath732 for @xmath733 . * * ( big projections of @xmath734 onto @xmath311).**for @xmath735 $ ] , @xmath489 , we have @xmath736 * start of the iteration . *
we set @xmath737 , @xmath738 , @xmath739 and @xmath740 , where @xmath15 stands for the _ mock tangent _ plane at @xmath741 , satisfying ( h3 ) of definition [ admissible ] .
moreover , we use the convention that our closed balls are defined as @xmath742 so that the closed ball @xmath743 of radius zero is the empty set .
notice that for a complete iteration start we need to define @xmath744 and @xmath745 in order to check conditions in ( a ) , in ( b ) , for @xmath746 , and constituting conditions ( e ) and ( f ) .
all the other conditions within the whole list are immediate for @xmath747 we set @xmath748 with growing radii @xmath749 the sets @xmath750 describe larger and larger double cones with ` axis ' perpendicular to @xmath360 and fixed opening angle which is very close to @xmath751 when @xmath455 is small .
now we define @xmath752 and notice that since @xmath539 satisfies of condition ( h3 ) by definition , one has @xmath753 .
this yields in ( a ) for @xmath746 . set @xmath754 ; in other words we have @xmath755 with @xmath756 , so that all properties mentioned in ( a ) are satisfied for @xmath746 .
moreover , since we have adopted the convention that @xmath743 is an empty set and @xmath757 , condition in ( b ) does hold for @xmath746 .
the definition of @xmath744 guarantees that there are no points of @xmath2 in @xmath758 , implying in ( c ) for @xmath746 . condition ( d ) for @xmath746 follows from the definition of @xmath744 , as @xmath2 is a closed subset of @xmath11 .
let us now take care of ( e ) and ( f ) for @xmath746 . to check ( e )
, note that by lemma [ lk=1 ] we have @xmath759 for every @xmath760 , @xmath761 .
any sphere @xmath762 with @xmath635 and @xmath47 specified in ( e ) for @xmath746 which is contained in @xmath758 can be homotopically deformed to , say , @xmath763 with @xmath764 ; to this end , we just first move the base point @xmath635 to @xmath370 along the segment @xmath765\}$ ] in @xmath360 , and then adjust the radius .
notice that all @xmath151-spheres used to define such a homotopy are contained in @xmath758 and therefore stay away from @xmath2 by in ( c ) for @xmath746 .
thus , by lemma [ h - inv - link ] , ( e ) follows for @xmath746 .
( note that in this first step we have even proved more .
in fact , every sphere @xmath762 with @xmath766 , @xmath767 and radius @xmath47 such that @xmath768 is nontrivially linked with @xmath2 ; for @xmath746 we do not have to restrict ourselves to spheres in @xmath758 intersected with the annulus .
this restriction , however , will be necessary at later steps . ) invoking lemma [ link->proj ] , we conclude that each @xmath569-dimensional disk @xmath769 , with @xmath635 and @xmath47 as in ( e ) for @xmath746 , must contain at least one point of @xmath2 .
therefore , @xmath770 since all disks @xmath769 are contained in @xmath771 we conclude @xmath772 this is the big projection property ( f ) for @xmath746 . to summarize this first step ,
we have defined the sets @xmath773 , and the planes @xmath774 , @xmath775 and @xmath360 which , up to now , are all identical , so that the desired estimate for the angle @xmath776 holds trivially for @xmath777 .
we also have defined @xmath778 , postponing the decision whether @xmath779 or @xmath780 .
note that we have not defined @xmath781 yet .
however , ( e)(f ) do hold for @xmath746 , and all those items in the list ( a)(d ) for @xmath746 which do not involve statements about @xmath782 or @xmath781 also do hold .
we shall now discuss the stopping criteria and show how to pass to the next step of the iteration when it is necessary .
* stopping criteria and the iteration step .
* for the decision whether to stop the iteration or to continue it with step number @xmath783 for @xmath784 , we may now assume that the sets @xmath785 and the @xmath1-planes @xmath786 , @xmath787 with @xmath788 for @xmath789 , have already been defined .
we also have at this point a sequence of radii @xmath790 satisfying the growth condition for @xmath791 . the first two conditions in ( a ) may be assumed to hold for @xmath792 . in ( b ) we may suppose for @xmath793 , in contrast to which holds only for @xmath794 similarly , we may now work with in ( c ) and in (
d ) for all @xmath795 whereas in ( d ) can be assumed for @xmath796 .
the statements in ( e ) and ( f ) can be used for @xmath793 .
we are going to study the geometric situations that allow us to stop the iteration right away ; if this is the case , then we set @xmath797 and @xmath798 .
basically , there are two cases when we can stop the construction because then there is a point @xmath799 such that @xmath667 form a @xmath800good couple . in the third case
it turns out that @xmath801 is contained in a thin tubular neighbourhood of some plane @xmath802 , which is close to @xmath803 and _ very _ close to many of the mock tangent planes @xmath804 for points
@xmath635 in @xmath805 a priori , possibly even to all of these tangent planes .
when this happens , then we set @xmath806 , define a new radius @xmath807 , new sets @xmath808 containing @xmath809 , and finally check all the properties listed in ( a)(f ) .
the different geometric situations depend on the position of the point where the surface hits the current centrally symmetric set @xmath809 .
( first hit immediately gives a good couple . ) : : this occurs if there exists at least one point @xmath810 such that the set @xmath811 , cf .
definition [ couples ] ( ii ) , satisfies @xmath812 if case 1 holds , then , directly by definition , @xmath667 is a @xmath800good couple .
we then set @xmath797 , @xmath813 , and stop the construction .
it is easy to see that all conditions of lemma [ mainlemma ] are satisfied .
if case 1 fails , then we define the new plane @xmath802 which , roughly speaking , gives a very good approximation of a significant portion of the mock tangent planes @xmath804 for @xmath635 close to @xmath16 , and examine the portion of @xmath2 contained in the closed set @xmath814 to distinguish two more cases .
in one of them the iteration can be stopped in a similar way . in the second one
, the whole intersection @xmath815 might be very close to all mock tangent planes @xmath804 so that there is no chance of finding a good couple ; we have to continue the iteration then .
we begin with the definition of @xmath802 .
the choice of @xmath691 in comes into play here . in one of the two remaining cases @xmath802 will become the new @xmath816 . in the other case
we can stop the iteration , setting @xmath817 .
fix @xmath818 .
cover the grassmannian @xmath76 by finitely many balls @xmath819 let @xmath820 since we already can use the big projection property of condition ( f ) for all @xmath821 , it follows that @xmath822 thus , we can estimate @xmath823 now , let @xmath824 since the @xmath825 cover @xmath826 , there exists at least one @xmath827 such that @xmath828 we set @xmath829 , and distinguish two more cases .
( some points of @xmath815 are far from @xmath802 . ) : : by this we mean that there exists a point @xmath830 such that @xmath831 if holds , then , as in case 1 , we set @xmath797 , @xmath813 , and stop the iteration .
it remains to check that @xmath667 is a @xmath800good couple .
condition ( i ) of definition [ couples ] is clearly satisfied . to check ( ii ) of that definition we estimate for each @xmath832 , using the triangle inequality , @xmath833 ( for the last inequality we just use @xmath834 and @xmath835 . )
therefore , @xmath836 .
moreover , guarantees that @xmath837 is large enough .
it follows that @xmath667 is a @xmath800good couple . as before in case 1 , it is easy to see now that all conditions of lemma [ mainlemma ] are satisfied with @xmath838 for all @xmath839 $ ] .
if neither case 1 nor case 2 occurs , then we have to deal with case 3 .
( flat position ; the whole @xmath840 is very close to @xmath802 . )
: : this happens if and only if for each point @xmath841 we have @xmath842 intuitively , case 3 corresponds to the following situation : most points of @xmath843 are close to some fixed @xmath1-plane which is a very good approximation of @xmath804 for many ( possibly all ! ) points @xmath844 close to @xmath16 .
we then set @xmath806 and have to continue the iteration .
* flat position and the passage to the next step .
* we shall first check that if case 3 has occurred , then @xmath845 in order to prove that this is indeed the case , we shall check that @xmath846 indeed , assume were false .
fix a unit vector @xmath847 such that @xmath848 is empty .
let @xmath849 for @xmath850 pick @xmath851 is contained in the interior of the intersection of @xmath852 and the annulus @xmath853 .
thus , we may use condition ( e ) , for @xmath854 , and lemma [ link->proj ] to conclude that the disk @xmath855 contains at least one point @xmath856 .
we also have @xmath857 ; this follows from the choice of @xmath635 and @xmath47 . invoking and above , we have @xmath858 since @xmath859 and @xmath849 , by scaling we have also @xmath860 so that the triangle inequality gives , by , @xmath861 this
, however , is a contradiction to condition which holds in case 3 .
hence , holds too , and for every orthonormal basis @xmath862 the vectors @xmath863 form a basis of @xmath816 which satisfies @xmath864 .
lemma [ bases ] implies that @xmath865 which is . as the angle @xmath866 is small , the cones @xmath867 and @xmath868 have a large intersection .
indeed , for any unit vector @xmath225 with @xmath869 we have @xmath870 by definition of @xmath871 .
thus , @xmath872 whenever @xmath873 . in particular
, every unit vector @xmath874 belongs to the interior of @xmath852 .
we now define @xmath875 according to , this immediately gives the missing conditions in ( b ) and in ( c ) for @xmath854 . to check in ( c ) for @xmath854 ,
note that in case 3 we have @xmath876 for each point of @xmath2 in the annulus @xmath853 .
however , when @xmath877 for some @xmath878 , then @xmath879 so that @xmath20 can not be a point of @xmath2 .
this gives for @xmath854 .
now the crucial thing is to define the next radius @xmath807 and take care of the linking condition for @xmath880 . * the next radius and homotopies from large spheres to smaller tilted ones . *
set @xmath881 and define @xmath882 notice that condition guarantees that @xmath883 .
this verifies in condition ( a ) for @xmath884 now we define @xmath885 and check that conditions ( a)(f ) are satisfied
. indeed , @xmath886 by condition ( a ) for @xmath887 which implies that ( a ) holds for @xmath880 as well .
next , @xmath888 since @xmath889 by condition ( a ) for @xmath854 . hence holds for @xmath880 .
the inclusion @xmath890 and other conditions involving @xmath891 have already been checked ; they follow directly from the definition of @xmath891 ; see . using for @xmath854 and the definition of @xmath892 in we infer that holds for @xmath880 , and for @xmath854 .
we also have @xmath893 also condition ( d ) follows directly from the definition of @xmath807 .
now we turn to the proof of the linking condition , , for @xmath884 the definition of @xmath807 implies that each sphere @xmath894 where @xmath895 , @xmath896 and the radius @xmath144 is such that @xmath897 can be homotopically deformed to @xmath898 without meeting any points of @xmath2 , so that the linking invariant used in is preserved .
one of the possible homotopies is to move the base point @xmath635 to 0 along the segment @xmath899 , @xmath900 $ ] , at the same time increasing the radius from @xmath901 to @xmath902 so that @xmath903 remains constant for all @xmath900 $ ] ; in this way , we simply slide the @xmath151-dimensional spheres along the surface of a fixed @xmath904-sphere , staying all the time in the interior of @xmath905 intersected with the annulus @xmath906 . by lemma [ h - inv - link ] we have @xmath907 next
, we may homotopically deform the sphere @xmath908 to another sphere of radius @xmath909 , @xmath910 we just shrink the radius linearly , staying all the time in the @xmath569-dimensional subspace @xmath911 .
it is clear that all the flat spheres realizing this homotopy @xmath912 stay in the interior of @xmath913 ( by for @xmath880 and the definition of @xmath891 in ) and do not contain any points of @xmath2 , so that , again by lemma [ h - inv - link ] , @xmath914 but @xmath915 can be homotopied still in the interior of @xmath913 to another sphere , @xmath916 which has the same radius @xmath909 but is slightly tilted ; therefore , @xmath917 to check this , we perform two steps .
first we move each point @xmath20 of @xmath918 along the segment that joins @xmath20 to its projection @xmath919 .
this gives an ellipsoid which is nearly spherical and has all axes at least @xmath920 because of the condition for the angle between @xmath921 and @xmath911 .
next , we continuously blow up this ellipsoid , moving each of its points along the rays that emanate from @xmath370 to points of @xmath922 . because of the smallness condition for the constants that we use , each segment @xmath923 with one endpoint at @xmath924 , @xmath925 , and the other at @xmath919 is certainly contained in the interior of @xmath809 ( i.e. far away from @xmath2 ) , as @xmath926 so that @xmath927 . thus , invoking lemma [ h - inv - link ] one more time , and applying the inductive assumption , i.e. the linking condition for @xmath854 , we finally obtain @xmath928 this gives of ( e ) for @xmath880 .
it is now easy to establish the big projection property of ( f ) for @xmath880 .
we do this as in the first step of the proof : invoking lemma [ link->proj ] , we conclude that each flat @xmath569-dimensional disk @xmath929 , with @xmath635 and @xmath47 as in ( e ) for @xmath880 , must contain at least one point of @xmath2 .
therefore , @xmath930 all disks @xmath929 are contained in @xmath931 so that @xmath932 this gives in ( f ) for @xmath880 , and finishes the proof of all conditions in the list ( a)(f ) in the iteration step .
since we have established condition ( e ) in the iteration step and holds , too , we can deduce that case 3 can happen only finitely many times , depending on the position @xmath16 on @xmath2 and on the shape and size of @xmath2 : @xmath933 whence the maximal number of iteration steps is bounded by @xmath934 this concludes the consideration of case 3 , and the whole proof of lemma [ mainlemma ] .@xmath218
we shall now derive lemma [ low - de - bounds ] and theorem [ thm : uar ] from lemma [ mainlemma ] .
this is a relatively easy task at this stage .
we shall just relay on estimates for the @xmath4-energy in the neighbourhooud of a good couple @xmath935 .
* proof of lemma [ low - de - bounds ] . * since @xmath936 for @xmath937 , we assume from now on that @xmath938 is the constant of lemma [ mainlemma ] .
fix @xmath73 small ( to be specified later on ) .
assume that @xmath939 and select a point @xmath940 such that @xmath941 .
use lemma [ mainlemma ] to select a @xmath668good couple @xmath935 .
let @xmath942 be as in definition [ couples ] ( ii ) , and let @xmath943 .
applying lemma [ 1/r - est ] we estimate @xmath944 which implies @xmath945 a contradiction for @xmath946 this proves the first part of the lemma .
now , for an arbitrarily small @xmath947 pick @xmath948 such that @xmath949 .
select @xmath950 so that @xmath951 is a @xmath952good couple .
we have @xmath953 , so that by lemma [ mainlemma ] ( ii ) @xmath954 certainly holds for @xmath955 since @xmath956 by . estimating the energy one more time , as before , we obtain @xmath957 thus , @xmath958 where @xmath959 , as in . letting @xmath960
, we obtain and conclude the whole proof.@xmath218 * proof of theorem [ thm : uar ] .
* by the lower bound for stopping distances , the inequality @xmath961 holds for each @xmath458 and each @xmath962 . by density of @xmath539 in @xmath2
, we obtain @xmath963 for _ all _ @xmath55 and @xmath964 .
this implies theorem [ thm : uar ] with @xmath965 where @xmath686 and @xmath687 are the constants introduced in lemma [ mainlemma ] . as we have already mentioned in the introduction , the proof above yields a result which is stronger than the formal statement of theorem [ thm : uar ] .
in fact , the result holds also for all @xmath966 with @xmath967 , where @xmath968 is the positive constant of lemma [ mainlemma ] , and this is a wider class of sets than the one we used in the introduction .
in this section we prove that for each point @xmath55 there exists a plane @xmath969 such that @xmath970 for @xmath971 , @xmath972 .
moreover , the mapping @xmath973 is of class @xmath974 , @xmath975 .
a posteriori it turns out that if @xmath537 is small enough and @xmath563 is an admissible surface with @xmath976 for some @xmath5 , then the mock tangent planes @xmath15 defined a.e . on @xmath2 must coincide with the classically understood @xmath80 .
the idea is to combine the results of the previous section with energy bounds and show that the p. jones @xmath66-numbers of @xmath2 satisfy a decay estimate of the form @xmath977 .
this alone would not be enough , but we already know that @xmath2 has big projections . adding this ingredient , we are able to prove that @xmath2 is in fact a @xmath79-manifold . moreover , in each ball of radius @xmath978 centered at @xmath55 the surface @xmath2 is a graph of a @xmath79 function @xmath979 over @xmath980 .
the core of this section is formed by an iterative construction , presented in section [ iteration ] , which yields the existence of tangent planes and estimates for their oscillation . at each iteration step , we need to check that the @xmath66-numbers decrease sufficiently fast as the length scale shrinks to zero . at the same time
, we have to guarantee that the linking conditions which imply the existence of big projections are also satisfied .
to make the presentation of that proof easier to digest , we introduce an ad - hoc notion of _ trapping boxes _
( section [ boxes ] ) and prove an auxiliary lemma which is then used in the iteration .
[ boxes ] everywhere in this section @xmath59 denotes the radius specified in theorem [ thm : uar ] ascertaining the uniform ahlfors regularity of surfaces with bounded energy . for the rest of the whole section , we fix @xmath981 small
so that is satisfied and all claims of lemma [ mainlemma ] are fulfilled .
[ tb ] assume that @xmath563 , @xmath55 , @xmath982 , @xmath983 $ ] and @xmath409 .
we say that a closed set @xmath984 is _ a @xmath985-trapping box for @xmath2 in @xmath986 _ if and only if the following conditions are satisfied : 1 .
@xmath987 ; 2 .
@xmath988 ; 3 .
if @xmath989 satisfies @xmath990 , then there exists a @xmath991 such that @xmath992 , the sphere @xmath993 is contained in the interior of @xmath994 and @xmath995 thus , informally , a trapping box is a subset of @xmath996 which is at least as large as a cylindrical neighbourhood of @xmath997 in @xmath998 ( of size specified by the parameter @xmath999 ) , and gives us some control of the location of @xmath1000 and of its projections onto @xmath170 .
if @xmath563 , @xmath458 and @xmath1001 is given by condition ( h3 ) of definition [ admissible ] , then for radii @xmath1002 a simple example of a trapping box is provided by the cylinder @xmath1003 it satisfies all conditions of definition [ tb ] for @xmath1004 ; in particular , lemma [ lk=1 ] guarantees condition ( iii ) .
another example is given by the following .
[ startbox ] let @xmath679 be the small constant of lemma [ mainlemma ] .
assume that @xmath563 , @xmath680 $ ] , @xmath51 , and @xmath59 denotes the radius specified in theorem [ thm : uar ] .
then , for each @xmath55 and each @xmath1005 there exists a plane @xmath409 such that @xmath1006 is a @xmath1007-trapping box for @xmath2 in @xmath986 .
* one can check that conditions ( a)(f ) stated at the beginning of the proof of lemma [ mainlemma ] combined with the lower bound for stopping distances obtained in lemma [ low - de - bounds ] imply the statement of proposition [ startbox ] for all points @xmath657 .
( to see this , look at condition of case 3 , which is the only case when the iterative construction is continued .
it has been designed in such a way that the union of @xmath1008 and @xmath1009 be a trapping box for @xmath2 in @xmath1010 ; condition ( e ) , cf .
, implies the existence of _ many _ spheres linked with @xmath2 so that ( iii ) of definition [ tb ] is also satisfied . since @xmath1011 by
, the claim of the proposition holds for all @xmath1012 $ ] with @xmath1013 , and we can certainly increase @xmath47 up to the infimum @xmath1014 of all stopping distances , which satisfies @xmath1015 by lemma [ low - de - bounds ] . ) assume now that @xmath1016 .
fix @xmath1005 and select a sequence @xmath1017 , @xmath1018 as @xmath1019 . for each @xmath1020 ,
let @xmath1021 whose existence is given by the statement of the proposition at points @xmath1017 .
passing to a further subsequence , we may assume that @xmath1022 as @xmath1019 .
the trapping boxes @xmath1023 corresponding to @xmath1024 and @xmath1021 via converge then in hausdorff distance to a closed set @xmath39 given by for @xmath16 and @xmath170 . since @xmath2 is closed , @xmath1025 must be contained in @xmath39 .
condition ( ii ) of definition [ tb ] is trivially satisfied , and condition ( iii ) is easily verified by using homotopical invariance of the linking number as we already did before ( one has to slightly tilt the spheres in @xmath994 to obtain spheres in @xmath1026 ) . @xmath218
the main idea of this section is to show that once we have a trapping box of the form , possibly with @xmath455 replaced by some smaller number @xmath1027 , then , under a certain balance condition for @xmath1028 , @xmath47 and the energy of @xmath2 , we can perturb the plane @xmath170 slightly to a new position @xmath360 and find a smaller , cylindrical @xmath1029-trapping box .
we make this precise in the next subsection .
we introduce two new constants @xmath1030 recall from lemma [ bases2 ] that the constant @xmath1031 depends on @xmath1 only .
[ shrink - box ] assume that @xmath409 , @xmath1032 , @xmath1033 , @xmath1034 , @xmath5 .
let @xmath1035 , @xmath680 $ ] be an admissible surface with @xmath1036 .
suppose that @xmath1037 is a @xmath985-trapping box for @xmath2 in @xmath1038 .
if @xmath1039 satisfies the _ balance condition _
@xmath1040 then there exists a plane @xmath1041 such that 1 .
@xmath1042 ; 2 . the cylinder @xmath1043 is a @xmath1044-trapping box for @xmath2 in @xmath1045 .
the main point is that once we fix a finite energy level @xmath60 , and @xmath47 sufficiently small , then the condition @xmath5 guarantees that there are numbers @xmath1046 which satisfy the balance condition and are such that @xmath1047 is ( much ) smaller than @xmath999 . since the angle @xmath1048 is controlled due to ( i ) , the lemma can be applied iteratively .
this will be done in the next subsection .
[ 2pt - plane ] if we fix an arbitrary point @xmath1049 such that @xmath1050 , then the plane @xmath360 in lemma [ shrink - box ] can be chosen so that @xmath1051 , as can be seen from the first step of the following proof .
* proof of lemma [ shrink - box ] .
* fix an arbitrary orthonormal basis @xmath1052 of @xmath170 and let @xmath1053 since @xmath1054 , we have @xmath1055 by .
set @xmath1056 , @xmath1057 . * step 1 .
choice of @xmath360 .
* using condition ( iii ) of definition [ tb ] , lemma [ h - inv - link ] and lemma [ link->proj ] , we conclude that each disk @xmath1058 contains in all disks with slightly larger radii , and @xmath1059 is closed . ]
a point @xmath1060 . set @xmath1061 . letting @xmath1062 , we use @xmath1054 and to estimate @xmath1063 and invoke lemma [ bases ] to obtain @xmath1064 .
( this initial step of the proof shows why remark [ 2pt - plane ] is satisfied .
we can work with an orthonormal basis @xmath1065 such that @xmath1066 . ) now , set @xmath1067 . * step 2 . for @xmath635
near 0 , most of the @xmath804 are close to @xmath360 .
* we shall establish the following : for each @xmath428 , the couple of points @xmath550 and @xmath1068 is not a @xmath1069good couple .
assume that the opposite were true and for some @xmath428 we had a @xmath1069good couple @xmath1070 .
then , using the two estimates @xmath1071 where comes from theorem [ thm : uar ] , and the inequality of lemma [ 1/r - est ] to estimate @xmath98 , we would obtain a lower bound for the energy , @xmath1072 by and the balance condition ; this contradiction proves that the claim of step 2 does hold .
in particular , since the condition @xmath1073 is satisfied for each @xmath388 , we have @xmath1074 * step 3 .
the new box contains @xmath1075 .
* we shall show that the cylinder @xmath39 defined by contains @xmath1076 .
again , we argue by contradiction .
suppose that there exists @xmath1077 such that @xmath1078 .
set @xmath1079 by theorem [ thm : uar ] and , we have @xmath1080 and due to the definition of @xmath1081 we know that @xmath1082 fix @xmath1083 .
yields @xmath1084 given by @xmath1085 satisfies @xmath1086 for each @xmath388 . letting @xmath1087
, we check that @xmath1088 as @xmath1089 . invoking lemma [ bases2 ] for @xmath1090 and @xmath1091
, we conclude that @xmath1092 now , since @xmath1093 , we have @xmath1094 , and @xmath1095 thus , for @xmath1096 and @xmath1097 @xmath1098 since @xmath1099 satisfies and @xmath1100 . on the other hand , we certainly have @xmath1101 for every point @xmath1096 .
this yields @xmath1102 we may now estimate the energy analogously to and obtain would obtain a lower bound for the energy , @xmath1103 this is again a contradiction , proving that @xmath1104 .
* step 4 . the linking condition . *
since we have established @xmath1105 in the first step of the proof , the sphere @xmath1106 is contained in the interior of @xmath1107 and we have @xmath1108 , since all points of @xmath1076 are in the cylinder @xmath39 defined in , and @xmath1109 .
thus , we may deform @xmath908 homotopically to @xmath1110 so that the whole family of spheres realizing the homotopy stays in @xmath1111 , i.e. far away from @xmath2 .
( this can be done precisely as in the verification of at the end of the proof of lemma [ mainlemma ] : we move the points of @xmath908 to their projections onto @xmath659 , and then deform the resulting ellipsoid to obtain the round sphere @xmath1112 . )
thus , @xmath1113 by lemma [ h - inv - link ] .
now , every other sphere @xmath1114 , with @xmath766 , @xmath1115 and @xmath1116 , i.e. every @xmath151-sphere parallel to @xmath1117 and contained in the interior of @xmath1118 , can obviously be deformed homotopically to @xmath908 without hitting points of @xmath2 , since @xmath1119 .
thus , again by lemma [ h - inv - link ] , we conclude that condition ( iii ) of definition [ tb ] is satisfied for @xmath39 in @xmath1120 .
this completes the whole proof of the lemma.@xmath218 [ iteration ] in this subsection , we apply lemma [ shrink - box ] iteratively and prove the following .
[ tx - short ] let @xmath679 be the constant of lemma [ mainlemma ] .
assume that @xmath563 for some @xmath680 $ ] , @xmath1121 , @xmath1122 .
then @xmath2 is an embedded @xmath1-dimensional submanifold of class @xmath79 , @xmath95 .
in fact , theorem [ tx - short ] will be just a corollary of another result , which gives a lot of more precise , quantitative information .
let @xmath679 be the constant of lemma [ mainlemma ] .
assume that @xmath563 for some @xmath680 $ ] , @xmath1121 , @xmath5 .
then for each @xmath14 there exists a unique plane @xmath969 ( which we refer to as _ tangent plane of @xmath2 at @xmath16 _ ) such that @xmath1123 moreover , there exists a constant @xmath1124 with the following property .
whenever @xmath633 are such that @xmath1125 then @xmath1126 depending only on @xmath70 and @xmath71 .
moreover , @xmath1127 is an open @xmath1-dimensional topological disk , the orthogonal projection @xmath1128 onto @xmath80 restricted to @xmath1129 is injective , and each cylinder @xmath1130 with @xmath1131 is a @xmath1132-trapping box for @xmath2 in @xmath1133 .
[ thm:5.6 ] * proof .
* a rough plan of the proof is the following .
we shall first show , using lemma [ shrink - box ] iteratively , that for each @xmath55 there exists a plane @xmath1134 such that for @xmath279 sufficiently close we have @xmath1135 . as a byproduct
, we shall obtain a sequence of trapping boxes around each @xmath1136 , allowing us to show that @xmath1137 is in fact unique . finally , we set @xmath1138 and verify the statements concerning @xmath1128 . * step 1 . * fix @xmath633 and assume that does hold for a sufficiently small positive constant @xmath1139 that shall be specified later on .
fix @xmath1140 such that latexmath:[\[\label{r1 } \frac 23 r_1
< \frac{9}{10 } ( 1-\delta^2)^{1/2 } r_1 \le @xmath16 and @xmath1142 , we obtain a plane @xmath409 such that @xmath1143 is a @xmath1007-trapping box for @xmath2 in @xmath1144 .
now , for @xmath1145 we set @xmath1146 we have @xmath1147 as @xmath1148 ; the constant @xmath1139 will be chosen later , in below , so small that @xmath455 and @xmath1149 shall satisfy the assumptions of lemma [ shrink - box ] .
the choice of @xmath1149 guarantees that @xmath1150 apply lemma [ shrink - box ] and remark [ 2pt - plane ] with @xmath1151 , @xmath1142 and @xmath1152 to choose @xmath1041 such that @xmath1153 in @xmath360 and the cylinder @xmath1154 is a @xmath1155-trapping box for @xmath2 in @xmath1156 .
( the plane @xmath360 will serve , roughly speaking , as a sort of average position for all tangent planes to @xmath2 in @xmath1144 . ) * step 2 .
the choice of @xmath1157 .
* since @xmath1158 , we have @xmath1159 , and the intersection @xmath1160 provides a @xmath1161-trapping box for @xmath2 in @xmath1162 . invoking lemma [ shrink - box ] again , we find a plane @xmath1163 such that @xmath1164 and the cylinder @xmath1165 is a @xmath1166-trapping box for @xmath2 in @xmath1167 .
proceeding inductively , we find a sequence of planes @xmath1168 such that for each @xmath1145 the cylinder @xmath1169 is a @xmath1170-trapping box for @xmath2 in @xmath1171 and we have the estimate @xmath1172 since @xmath1173 , the planes @xmath1174 converge to some plane @xmath1175 such that @xmath1176 with @xmath1177 for the last inequality above , we have used an elementary estimate @xmath1178 which holds for each @xmath1179 for all @xmath1179 and @xmath1180 , as @xmath1181 and @xmath1182 on @xmath1183 . ] . now , note that since @xmath1051 the initial cylinder @xmath1184 is such that @xmath1185 provides a @xmath1161-trapping box for @xmath2 in @xmath1186 .
thus , replacing the roles of @xmath16 and @xmath20 from the second step on , we may run a similar iteration and obtain a plane @xmath1187 such that @xmath1188 together with a sequence of planes @xmath1189 ( with @xmath1190 ) and appropriate trapping boxes determined by those planes . by the triangle inequality , for @xmath780 and yield @xmath1191
once the uniqueness of @xmath1157 is established , we identify @xmath1157 with @xmath80 .
the estimate combined with will yield the desired ( note that @xmath1192 up to a constant factor which is less than 2 ) . * step 3 . trapping boxes around @xmath1157 . *
it is now easy to check that tilting the cylinders @xmath1193 and enlarging them slightly , we can obtain new trapping boxes @xmath1194 for @xmath2 in @xmath1171 .
fix @xmath1195 . for sake of brevity ,
let @xmath1196 and @xmath1197 denote the orthogonal projections of @xmath11 onto @xmath1198 and @xmath1199 .
we have @xmath1200 as @xmath1201 .
hence , by , @xmath1202 therefore , if @xmath1203 is defined by with @xmath1204 then we have @xmath1205 for each @xmath1195 .
thus the cylinder @xmath1206 contains @xmath1207 .
it follows that @xmath1208 , as @xmath1193 was a trapping box for @xmath2 in a larger ball @xmath1171 .
it is easy to see that the linking condition of definition [ tb ] is also satisfied ( we just take a smaller set of spheres that are slightly tilted ) so that @xmath1194 indeed is a @xmath1209-trapping box for @xmath2 in @xmath1133 .
let us now specify @xmath1139 .
we choose this constant so that @xmath1210 where @xmath1211 is the constant of theorem [ thm : uar ]
. then , by , @xmath1212 for these choices of @xmath1213 and @xmath1139 all applications of lemma [ shrink - box ] were justified .
now , returning to , we obtain @xmath1214 in particular , as @xmath1215 , we also have @xmath1216 to finish the whole proof , it remains to demonstrate that @xmath1157 is indeed unique and that @xmath1217 is an open @xmath1-dimensional disk such that the projection @xmath1218 is injective . * step 4 .
uniqueness of @xmath1157 . * since formally lemma [ shrink - box ] alone does not guarantee that the choice of each new plane @xmath1174 is unique , we must now show that @xmath1219 _ is _ unique .
suppose that this were not the case , and that choosing @xmath1220 in some steps of the iteration we could obtain a different limiting plane @xmath315 , with @xmath1221 .
select @xmath1222 with @xmath1223 such that @xmath1224 .
set @xmath1225 and without loss of generality suppose that @xmath550 .
the spheres @xmath1226 are contained in @xmath1227 , away from @xmath2 since @xmath1228 , and by lemma [ h - inv - link ] , are nontrivially linked with @xmath2 since @xmath1194 is a @xmath1209-trapping box for @xmath2 in @xmath1229 . since @xmath315 has been obtained by an analogous iteration process , the cylinders @xmath1230 should also provide @xmath1231-trapping boxes for @xmath2 in @xmath1229 .
however , taking @xmath437 so large that @xmath1232 , we obtain @xmath1233 .
thus , the sphere @xmath1234 is contained in the interior of @xmath1235 and satisfies the assumptions of lemma [ lk=0 ] with @xmath1236 and therefore _ is not _ linked with @xmath2 , a contradiction which proves that @xmath1137 has to be unique .
moreover , since @xmath1194 is a @xmath1209-trapping box for @xmath2 in @xmath1133 and @xmath1237 one easily concludes that for @xmath158 we have @xmath1238 which justifies the definition @xmath1239 . * step 5 .
injectivity of the projection . *
again , we argue by contradiction .
suppose that there exist @xmath1240 such that @xmath1241 . without loss of generality suppose that @xmath1242 and let @xmath1243 be defined by .
set @xmath1244 and let @xmath1245 , @xmath1246 denote the projections onto @xmath1247 , @xmath1248 , respectively . as @xmath1249 , we have @xmath1250 and @xmath1251 by . however , fixing @xmath437 so that @xmath1252 , i.e. the cylinders @xmath1253 which contain @xmath1254 , to estimate by virtue of @xmath1255 a contradiction . since for each @xmath1256
the cylinder @xmath1257 is a @xmath1258-trapping box for @xmath2 in @xmath1259 , and @xmath1260 , we know by now as @xmath1261 can be taken very close to @xmath1262 that the image of @xmath1128 restricted to , say , @xmath1263 , certainly contains the disk with center at @xmath1264 and radius @xmath1262 .
it follows that @xmath1265 is a topological disk , since @xmath1128 was also shown to be injective.@xmath218 * proof of theorem [ thm : betas ] . * as @xmath1266 , it can be checked that theorem [ thm : betas ] stated in the introduction follows from and the definition of a trapping box .
one can use the plane @xmath1267 to estimate the infimum in the definition of @xmath66-numbers .
@xmath218 we shall now use theorem [ thm:5.6 ] to construct the graph representations of an admissible surface @xmath2 with @xmath1268 for some @xmath5 . in section
[ slicing ] , this will be used to show that @xmath2 is in fact a manifold of class @xmath8 for @xmath1269 .
[ graphpatch ] suppose that @xmath563 for some @xmath680 $ ] , @xmath1270 , @xmath5 .
let @xmath1271 and @xmath1272 denote the constants introduced in theorem [ thm:5.6 ] .
set @xmath1273 .
then , for each @xmath55 , the following is true .
there exists a function @xmath1274 of class @xmath79 , @xmath1275 , such that @xmath1276 and @xmath1277 , and @xmath1278 where @xmath1279 denotes the graph of @xmath91 , and @xmath1280 for some constant @xmath1281 depending only on @xmath1282 .
* proof . * without loss of generality suppose that @xmath1283 and @xmath1284 , where @xmath1285 , @xmath1286 form the standard orthonormal basis of @xmath11 . by theorem [ thm:5.6 ] we know that @xmath1287 is invertible . by and for @xmath780 , the image of this map contains an @xmath1-dimensional disk of radius @xmath1288 .
* we now let @xmath1289 so that @xmath1290 is a natural parametrization of @xmath2 .
note that @xmath1291 contains @xmath1292 and that @xmath1276 .
both @xmath91 and @xmath39 are continuous .
* step 2 . * to prove that @xmath1293 exists and equals @xmath370 , use now the definition of @xmath91 to see that , and of theorem [ thm:5.6 ] yield @xmath1294 whenever @xmath1295 .
( recall that @xmath1296 ; we are free to use any @xmath1256 here . )
set @xmath1297 ; by , @xmath1298 .
thus , we also have @xmath1299 whenever @xmath1300 . as @xmath1301 for @xmath1302 ,
this gives @xmath1303 near @xmath370 and consequently @xmath1277 .
we shall now show that @xmath39 ( and hence @xmath91 ) is differentiable at each @xmath1304 .
fix @xmath1304 and @xmath1305 with @xmath1306 small .
set @xmath1307 we have @xmath1308 where the error @xmath1309 satisfies , by definition of @xmath315 and @xmath39 , @xmath1310 .
thus , @xmath1311 , so that @xmath1312 absorbing the first term and using now theorem [ thm:5.6 ] at @xmath1313 , we obtain @xmath1314 therefore , @xmath1315 . using this and , we now write @xmath1316 now , for all @xmath1306 sufficiently small we have @xmath1317 , as @xmath39 is continuous at @xmath635 . thus , the second term can be absorbed , yielding @xmath1318 as @xmath1319 .
plugging this into the right hand side of , we obtain the desired error estimate @xmath1320 as @xmath1319 . therefore , @xmath39 is differentiable at @xmath635 with @xmath1321 .
the uniform hlder bound for @xmath176 results now from one more application of the oscillation estimate for tangent planes : * step 3 . * with @xmath1322- \big[\begin{array}{c } e_i\\ \partial_if(z ) \end{array}\big]\right|= \left|\pi_{t_{f(w)}\sigma}\left(\big[\begin{array}{c } e_i\\ \partial_if(w ) \end{array}\big]\right)- \pi_{t_{f(z)}\sigma}\left(\big[\begin{array}{c } e_i\\ \partial_if(z ) \end{array}\big]\right)\right|\ ] ] we can estimate @xmath1323\right)- \pi_{t_{f(z)}\sigma}\left(\big[\begin{array}{c } e_i\\
\partial_if(w ) \end{array}\big]\right)\right|\\ & & \qquad + \left|\pi_{t_{f(z)}\sigma}\left(\big[\begin{array}{c } e_i\\ \partial_if(w ) \end{array}\big]\right)- \pi_{t_{f(z)}\sigma}\left(\big[\begin{array}{c } e_i\\ \partial_if(z ) \end{array}\big]\right)\right|\\ & & \hspace{-2 cm } \le { \mathop{\mbox{$<\!\!\!)$}}\nolimits}(t_{f(w)}\sigma , t_{f(z)}\sigma)(1+|\nabla f(w)|^2)^{1/2}+ \left|(\pi_{t_{f(z)}\sigma}-\pi_{t_0\sigma})\left(\big[\begin{array}{c } 0\\ \partial_if(w)-\partial_if(z ) \end{array}\big]\right)\right|\\ & & \hspace{-2.2cm}\overset{\eqref{osc - tan } } { < } c_6e^{1/(q+4m)}|w - z|^\kappa + { \mathop{\mbox{$<\!\!\!)$}}\nolimits}(t_{f(z)}\sigma , t_0\sigma)|\partial_if(w)-\partial_if(z)|.\end{aligned}\ ] ] since @xmath1324 by and our choice of constants , we can absorb the right term on the left - hand side to conclude .
now , using a standard cutoff technique , we leave @xmath91 unchanged on @xmath1325 , and extend it to the whole plane @xmath78 , so that the extension vanishes off @xmath1326 .
the corollary follows.@xmath218
[ slicing ] in this section we assume that @xmath2 is a flat @xmath1-dimensional graph of class @xmath79 having finite tangent - point energy @xmath0 .
the goal is to show how to bootstrap the hlder exponent @xmath1327 to @xmath1328 .
relying on corollary [ graphpatch ] , without loss of generality we can assume that @xmath1329 for a fixed number @xmath1330 , where @xmath1331 is of class @xmath79 and satisfies @xmath1277 , @xmath1276 , @xmath1332 to achieve , we use of corollary [ graphpatch ] and shrink @xmath1333 by a constant factor if necessary . the number @xmath1334 is chosen so that @xmath1335 for the constants @xmath257 and @xmath696 used in lemma [ bases ] and other auxiliary estimates in section [ sec : linear ] .
we let @xmath1336 be the natural parametrization of @xmath1337 , given by @xmath1338 for @xmath1339 ; outside @xmath1340 the image of @xmath39 does not have to coincide with @xmath2 .
the choice of @xmath1334 guarantees that , due to lemma [ bases ] ( ii ) , @xmath1341 whenever @xmath1342 .
thus , @xmath1343 as in our paper ( * ? ? ?
* section 6 ) , we introduce the maximal functions controlling the oscillation of @xmath176 at various places and scales , @xmath1344 where the supremum is taken over all possible closed @xmath1-dimensional balls @xmath1345 of radius @xmath1346 that are contained in a subset @xmath1347 , with @xmath1348 .
since @xmath1349 , we have a priori @xmath1350 for some constant @xmath1351 which does not depend on @xmath1352 . to show that @xmath1353 for @xmath99
, we check that locally , on each scale , the oscillation of @xmath176 is controlled by a main term which involves the local energy and resembles the right hand side of , up to a small error , which itself is controlled by the oscillation of @xmath176 on a much smaller scale .
[ key ] let @xmath91 , @xmath39 , @xmath2 , @xmath1330 and @xmath78 be as above . if @xmath1354 with @xmath1355 , then for any @xmath1356 we have @xmath1357 where @xmath1358 is an @xmath1-dimensional disc in @xmath78 , @xmath1359 , and @xmath1360 is the local energy of @xmath2 over @xmath998 .
* once this lemma is proved , one can fix an @xmath1-dimensional disk @xmath1361 and use to obtain for @xmath1362 @xmath1363 where @xmath1364 fixing @xmath1356 such that @xmath1365 we obtain @xmath1366 as @xmath1367 . using this
, one can iterate and show that @xmath1368 combining this estimate with corollary [ graphpatch ] , we obtain theorem [ thm : bootstrap ] stated in the introduction .
note that in fact the result holds for all surfaces @xmath563 for @xmath680 $ ] , where @xmath679 is the constant of lemma [ mainlemma ] .
* proof of lemma [ key ] .
* fix @xmath619 and the disk @xmath998 as in the statement of the lemma ; we have @xmath1369 .
pick @xmath1356 and let @xmath1370 be the local energy of @xmath2 over @xmath998 , defined by .
assume that @xmath1371 on @xmath998 , for otherwise there is nothing to prove .
we now estimate the local energy to obtain a bound for @xmath1374 , shrinking the domain of integration , as follows : @xmath1375 the last equality follows from .
thus , we obtain @xmath1376 and since the radius of @xmath998 equals @xmath749 , we obtain @xmath1377 now , select two points @xmath1378 ( @xmath186 ) . by the triangle inequality , @xmath1379 thus , it remains to show that the last term , @xmath1380 , does not exceed a constant multiple of @xmath1381 . to achieve this goal
, we assume that @xmath1382 and work with the portion of the surface parametrized by the points in @xmath1383 by , @xmath1384 satisfies @xmath1385 to conclude the whole proof , we shall derive an upper estimate for the measure of @xmath1384 , @xmath1386 where @xmath1387 and @xmath1388 denotes the tangent plane to @xmath2 at @xmath1389 for @xmath1390 combining and , we will then obtain @xmath1391 ( by a reasoning analogous to the proof of corollary [ graphpatch ] , this also yields an estimate for the oscillation of @xmath176 . ) * step 2 .
* by , we have @xmath1392 . by lemma [ proj - meas ]
applied to @xmath1393 , we obtain @xmath1394 so that would follow from @xmath1395 now , for @xmath1396 and @xmath186 we have by @xmath1397 let @xmath1398 be the affine tangent plane to @xmath2 at @xmath1399 . since @xmath39 is lipschitz with constant @xmath1400 and @xmath1401 , @xmath1402 for @xmath1403 , @xmath186 .
select the points @xmath1404 , @xmath186 , so that @xmath1405 .
the vector @xmath1406 is then orthogonal to @xmath360 and to @xmath296 , and since @xmath1384 is nonempty by , we have @xmath1407 by .
set @xmath1408 , pick a parameter @xmath1403 and consider @xmath1409 .
we have @xmath1410 so that @xmath1411 , and @xmath1412 therefore , since @xmath1413 and by , @xmath1414 . thus , @xmath1415 where @xmath1416 is the intersection of two slabs considered in section [ sec : linear ] . applying lemma [ proj - slab ] which is possible due to the estimate for @xmath1417
, we conclude that there exists an @xmath305-dimensional subspace @xmath306 such that @xmath1418 on the other hand , since @xmath39 is lipschitz , we certainly have @xmath1419 and therefore @xmath1420 combining and , we invoke lemma [ strip - ball ] to @xmath1421 , @xmath1422 , and @xmath1423 , to obtain @xmath1424 which is , implying and thus completing the proof .
banavar , j.r . ; gonzalez , o. ; maddocks , j.h . ; maritan , a. self - interactions of strands and sheets .
j. statist .
* 110 * ( 2003 ) , 3550 .
blatt , s. boundedness and regularizing effects of ohara s knot energies .
preprint ( 2010 ) .
pontryagin , l. s. selected works .
vol . 3 , _ algebraic and differential topology .
_ translated from the russian by p. s. v. naidu
. classics of soviet mathematics .
gordon & breach science publishers , new york , 1986 .
xxii+252 pp .
schuricht , f. ; von der mosel , h. euler - lagrange equations for nonlinearly elastic rods with self - contact .
arch . rat .
168 * ( 2003 ) , 3582 .
schuricht , f. ; von der mosel , h. characterization of ideal knots .
calc . var .
partial differential equations * 19 * ( 2004 ) , 281305 .
strzelecki , p. ; szumaska , m. , von der mosel , h. : regularizing and self - avoidance effects of integral menger curvature .
preprint nr .
29 , institut f. mathematik , rwth aachen ( 2008 ) . to appear in ann
. scuola norm .
di pisa .
strzelecki , p. ; von der mosel , h. on a mathematical model for thick surfaces . in : calvo ,
millett , rawdon , stasiak ( eds . ) _ physical and numerical models in knot theory _ , pp .
547564 . ser . on knots and
everything 36 , world scientific , singapore 2005 . | we consider repulsive potential energies @xmath0 , whose integrand measures tangent - point interactions , on a large class of non - smooth @xmath1-dimensional sets @xmath2 in @xmath3 finiteness of the energy @xmath0 has three sorts of effects for the set @xmath2 : topological effects excluding all kinds of ( a priori admissible ) self - intersections , geometric and measure - theoretic effects , providing large projections of @xmath2 onto suitable @xmath1-planes and therefore large @xmath1-dimensional hausdorff measure of @xmath2 within small balls up to a uniformly controlled scale , and finally , regularizing effects culminating in a geometric variant of the morrey - sobolev embedding theorem : any admissible set @xmath2 with finite @xmath4-energy , for any exponent @xmath5 , is , in fact , a @xmath6-manifold whose tangent planes vary in a hlder continuous manner with the optimal hlder exponent @xmath7 .
moreover , the patch size of the local @xmath8-graph representations is uniformly controlled from below only in terms of the energy value @xmath0 . |
consider @xmath0-steps _ self avoiding walk _ ( saw ) on the seven regular infinite planar triangulation .
the goal of this note it to prove that the saw is ballistic in expectation , i.e. , that the expected distance from the starting point grows linearly with @xmath0 .
this was already established for sufficiently large planar lattices in @xcite . in @xcite
it is proved that euclidean saw is sub - ballistic and see @xcite for ballisticity of saw on cayley graphs with more than one end .
recall that the @xmath0-steps saw is the uniform measure on the set of all length @xmath0 self avoiding paths starting at a fixed root . for background on saw
see @xcite , for background on gromov hyperbolic spaces see e.g. @xcite .
let @xmath1 be the seven regular infinite planar triangulation .
then the expected displacement of @xmath0 steps saw on @xmath1 is bigger than @xmath2 , where @xmath3 is a constant determined by @xmath1 .
question : improve the theorem to an asymptotically almost sure ballisticity .
no bounds on the number of self avoiding walks are used , unlike the argument in @xcite , and therefore no bounds on the spectral gap are needed .
the proof here works as is , for every @xmath4-skeleton of any hyperbolic lattice in @xmath5 , which is generate by reflections of the fundamental domain .
see @xcite for a survey on such lattices , which exist also in dimensions bigger than @xmath6 , yet there are no such of dimension bigger than @xmath7 . we leave it as an open _ question _ to adapt the proof to any finite volume lattice .
another model in which the proof works as is , is that of saw defined directly on @xmath5 as follows : consider continuous @xmath0 steps random walk on @xmath5 , where the next step is chosen uniformly and independently on the unit sphere around the current location .
look at the restriction of this uniform product measure to sequences in which the distance between any pair of vertices is bigger than @xmath3 , for some @xmath8 , fixed .
twenty years ago oded schramm briefly commented over lunch , that abundance of reflections will imply that hyperbolic saw is ballistic .
conceivably the argument here follows what oded had in mind .
[ fig:7hyp ] here is an outline of the proof . given a self avoiding walk @xmath9 in a space of constant negative curvature , a constant portion of the vertices are at the boundary of the convex hull of the walk . after reflection the part of @xmath9 from @xmath10 to @xmath11 , at a tangent of the convex hull at @xmath10 , the inverse triangle inequality is satisfied at @xmath10 , that is @xmath12 where @xmath13 depends only on the lattice , and @xmath14 denotes the reflected path .
we will see that the counting measure on the space of saw of length @xmath0 is up to multiplicative constant , invariant with respect to reflections at a tangent of the convex hull , as these are order @xmath0 abundant .
therefore a constant fraction of the vertices on a constant fraction of the self avoiding walks , satisfy the inverse triangle inequality .
now a path in bounded degree gromov hyperbolic graph , with inverse triangle inequality holding for portion of the vertices must be ballistic .
question : show that saw on @xmath15 is ballistic ? where @xmath16 is the three regular tree .
saw should be ballistic on any cayley graph in which simple random walk is ballistic .
in particular all non amenable cayley graphs . we expect saw to be ballistic on any infinite planar triangulation , in which all degrees are at least @xmath17 ( non positive curvature ) and there is @xmath18 , so that every @xmath19-ball in the triangulation contains a vertex of degree at least @xmath20
. such triangulations are non amenable ( see @xcite ) and for any finite set , a constant fraction of the vertices are at the boundary of the convex hull .
consider the uniform measure on anchored connected subsets of size @xmath0 in a hyperbolic lattice , or non amenable hyperbolic transitive graphs .
conjecture : the scaling limit of uniform connected subsets is brownian tree .
both when taking the ambient metric on the connected set , as an embedded subset , or the internal metric .
see @xcite for a related result and definition of the scaling limit .
similar conjecture should hold for the scaling limit of _ self avoiding loop _ in the ambient metric .
in @xcite it was shown that the probability euclidean saw makes a loop decays polynomially to @xmath21 .
show exponential decay for the analogous event on the seven regular triangulations .
this will give almost sure ballsticity of the saw .
simple counting gives this for @xmath22-regular triangulation with large @xmath22 .
recall ( see e.g. @xcite ) that a graph is gromov hyperbolic if for some @xmath23 , all geodesic triangles are @xmath24-thin . where a geodesic triangle is called @xmath24-thin if the distance from any point on one of the geodesics to the union of the two other geodesics is bounded by @xmath24 .
from now on @xmath1 denotes the seven regular infinite planar triangulation . the following two lemmas are essentially from @xcite .
[ convex2 ] let @xmath25 a finite set of vertices in a hyperbolic lattice .
then there are @xmath28 depending only on the lattice , so that at least @xmath27 vertices of @xmath25 are at distance at most @xmath29 from the boundary of the convex hull of @xmath25 .
proof : let @xmath30 be the minimal distance between distinct lattice points . for any finite subset @xmath25 of the lattice , @xmath31 , where @xmath32 . according to @xcite ,
the volume of the convex hull @xmath33 of @xmath34 differs from @xmath35 by at most a constant factor .
pick a random point @xmath36 . since the lattice is non amenable , with constant probability
, it lies in @xmath34 , and its distance from @xmath37 is at most a constant .
therefore a constant fraction of @xmath34 is of constant distance from @xmath37 , which implies the same for @xmath25 .
@xmath38 [ two convex sets ] assume @xmath39 , where @xmath40 is @xmath24-hyperbolic graph are disjoint convex sets with @xmath41 , then the following holds : for any @xmath42 , the geodesic between @xmath43 and @xmath44 intersects the ball around @xmath45 of radius @xmath46 $ ] .
proof of lemma : in any gromov hyperbolic space where all triangles are @xmath24-thin , if the distance between a pair of two sided infinite geodesics , is bigger than @xmath24 , then the distance between the pair of geodesics outside a ball of radius @xmath19 , grows at least exponentially in @xmath19 .
separate the two convex sets by two such geodesic to finish .
( in higher dimension separate the two convex sets by geodesic hyperplanes , to get a similar result ) .
@xmath38 @xmath1 is obtained by reflecting an heptagon ( the dual graph of the triangulation ) in @xmath53 , which is a fundamental domain for the seven regular lattice .
pick the @xmath20 reflections along geodesics which extends the edges of the fundamental domain . for each vertex in the graph fix @xmath20 such reflections , and ordered them arbitrarily .
define @xmath55 as follows .
pick the first reflection , in the order on reflections at @xmath10 , which reflect the part of @xmath9 from @xmath10 to @xmath11 , along a tangent to the convex hull of the self avoiding walk at @xmath10 , and apply it on @xmath9 .
if there is no such reflection at @xmath10 , then @xmath56 does nt change the walk .
* remark : * ( other lattices of hyperbolic reflection groups ) note that for @xmath4-skeletons of other lattices arising from reflections , when we only have [ convex2 ] and [ two convex sets ] with @xmath58 and @xmath23 , a local modification of the reflection at @xmath59 , @xmath56 is needed : reflect the self voiding walk along a geodesic tangent to a vertex @xmath60 of distance @xmath61 away from @xmath10 and the convex hull , and then locally modify the reflected self avoiding walk by connecting @xmath10 to @xmath60 , by a @xmath61 long path , and remove @xmath61 long segment from the end of the walk . by lemma [ convex ] uniformly over all walks in @xmath54 , a constant portion of the vertices are on the the boundary of the convex hull , so it is possible to apply a reflection at @xmath10 at a portion of the indexes , and for any walk the total number of reflections is order @xmath0 .
proof of claim : fix @xmath9 and pick @xmath59 uniformly between @xmath21 and @xmath0 . with probability bigger than @xmath3 , @xmath10 is at the boundary of the convex hull . observe that for any @xmath9 the number of reflections out side of the convex hull is proportional to the number of all the reflections , which is order @xmath0 .
the probability of the event @xmath63 , is up to a constant factor invariant with respect to @xmath56 and when @xmath56 reflects at @xmath59 , the event above holds by lemma [ two convex sets ] .
therefore , @xmath69
@xmath38 since geodesic triangles in gromov hyperbolic graphs are @xmath24-thin , if @xmath62-inverse triangle inequality holds at @xmath59 , by gromov hyperbolicity , it has to be in a bounded neighborhood of the geodesic , depending only on the @xmath24-hyperbolicity and @xmath62 .
the number of vertices in @xmath1 in a bounded neighborhood of the geodesic interval between @xmath71 and @xmath11 is bounded from above by @xmath72 . by ( 2.2 )
this neighborhood contains @xmath2 vertices of @xmath9 in expectation .
r. bauerschmidt , h. duminil - copin , j. goodman , and g. slade , lectures on self - avoiding walks .
( english summary ) probability and statistical physics in two and more dimensions , 395-467 , clay math .
proc . , 15 , amer . math
providence , ri , 2012 . | it is shown that self avoiding walk on the seven regular infinite planar triangulation has linear expected displacement . |
the bulk of the baryons in galaxy clusters are in their hot gas component , the intracluster medium ( icm ) , which typically contains 5 - 10 times the mass found in cluster galaxies .
the metal content of the icm is sensitive to the star formation histories of the cluster galaxies and to the way in which metals are ejected from galaxies in supernova explosions and winds . in particular ,
the abundances of different species can be used to constrain the relative numbers of type ia ( sne ia ) and type ii ( sne ii ) supernovae that occurred in the cluster galaxies .
this is because the abundances of the @xmath0 elements ( e.g. o , mg ) are driven primarily by the rate of sne ii , whereas the production of iron is dominated by sne ia .
the abundance of @xmath0 elements in the icm is found , from x - ray measurements , to be comparable to abundances in the solar neighbourhood , whereas the abundance of iron is only about 30% of that found locally ( e.g. * ? ? ?
this discrepancy has led several authors to propose that the initial mass function ( imf ) of stars forming in cluster galaxies could be substantially different from that inferred for star formation in less dense environments ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the chemical evolution of galaxies has been modelled extensively in the monolithic collapse scenario ( e.g. , * ? ? ?
simple infall models with delayed metal enrichment due to sne ia have been used to demonstrate that the metallicity distribution of stars in the solar neighbourhood can be explained by bottom - heavy imfs as proposed by @xcite @xcite .
the metal enrichment of the icm has also been studied in the framework of monolithic collapse , with the conclusion reached that salpeter - like imfs produce too few heavy elements to match the observed metal abundances in the icm @xcite . to reconcile the predictions with the observed icm abundances
, these studies concluded that a flatter imf than salpeter is required , such as the arimoto - yoshii imf , with @xmath3 @xcite , where the number of stars per logarithmic interval in mass is given as @xmath4 .
( for comparison , the salpeter imf has @xmath5 . )
this conclusion has been confirmed by numerical simulations @xcite . there is now overwhelming evidence supporting the gravitational instability paradigm for the formation of structure in the universe ( e.g. * ? ?
* ; * ? ? ?
* ) . any realistic model of galaxy formation
should therefore be placed in the proper context of a model that follows the hierarchical formation of cosmic structures , such as the successful @xmath6-dominated cold dark matter ( @xmath6cdm ) model @xcite .
semi - analytic models of galaxy formation provide a framework within which the physics of galaxy formation can be followed at the same time as dark matter halos are built up through mergers and the accretion of smaller objects ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these models are able to reproduce a wide range of properties of the local galaxy population and to make predictions for the high redshift universe ( e.g. * ? ? ?
to date , the majority of semi - analytic models have used the instantaneous recycling approximation for chemical enrichment , and only considered metal production by sne ii .
such models have been used to investigate the stellar and gas metallicities of disk and elliptical galaxies and of damped ly@xmath0 systems , with broad agreement obtained between model predictions and observations @xcite .
semi - analytic models based on instantaneous recycling have also been applied to enrichment of the icm by kauffmann & charlot ( 1998 ) and @xcite , as well as by nagashima & gouda ( 2001 ) .
the former two papers find that the observed icm metallicities are only reproduced in models in which the chemical yield is twice the standard value for a solar neighbourhood imf , in agreement with the conclusions from monolithic collapse models .
the first attempt to take into account metal enrichment due to sne ia in hierarchical galaxy formation models was that of @xcite and @xcite . however , their calculations used only the star formation histories for spiral and elliptical galaxies extracted from the munich group s semi - analytical model , and assumed closed - box chemical evolution .
they did not include the effects of gas inflows and outflows on the chemical enrichment histories of galaxies .
the metal enrichment due to sne ia was consistently included in a semi - analytical model for the first time by @xcite , who took into account the recycling of metals between stars , the cold gas in the disk and the hot halo gas . while the lifetime of sne ia s was simply assumed to be a constant 1.5gyr , their model provides good agreement with the observed distributions of [ fe / h ] and of [ o / fe ] vs. [ fe / h ] for both solar neighbourhood and bulge stars , based on a salpeter - like imf .
recently , @xcite showed that the observed number counts of faint sub - mm sources can be matched in hierarchical models if starbursts triggered by galaxy mergers are assumed to take place with a top - heavy imf ( with slope @xmath7 ) . since varying the slope of the imf affects the abundance ratios of @xmath0 elements to iron , an important test of this model
is to examine its predictions for the metallicity of the icm . in this paper , using essentially the same model as used by @xcite , we investigate the metal abundances of the icm .
we adopt a standard , `` bottom - heavy '' imf proposed by @xcite , with @xmath8 for @xmath9 , for quiescent star formation in disks , and a top - heavy imf with @xmath7 in starbursts . while the imf for quiescent star formation is reasonably well determined from observations in the solar neighbourhood ( e.g. * ? ? ?
* ) , the form of the imf appropriate to starbursts is still unclear
. there is , however , some tentative observational evidence supporting a top - heavy imf @xcite . in this paper , we incorporate metal enrichment due to sne ia as well as sne ii into the galform semi - analytic galaxy formation model @xcite , to see whether the model with a top - heavy imf is able to reproduce the metal abundances of the icm .
the model includes both inflows of gas and metals into galaxies due to gas cooling in halos , and ejection of gas and metals by galactic winds .
we calculate the production of different elements by sne ia and sne ii based on theoretical yields as a function of stellar mass , integrated over the model imf .
the lifetimes of sne ia are consistently calculated for the adopted stellar imfs , while instantaneous recycling is assumed for sne ii .
we compute icm abundances for o , mg , si and fe for clusters of different masses , and compare these with observational data . in making these comparisons , we correct the observed abundances for radial metallicity gradients ; these corrections are important for si and fe .
the paper is set out as follows . in section 2 ,
we briefly describe the galform model . in section 3
, we provide a detailed explanation of how we model metal enrichment . in section 4
, we show what our model predicts for the present - day luminosity function of galaxies , for different assumptions about feedback and thermal conduction . in section 5
, we present our predictions for the abundances of different elements in the icm , and for the ratios of icm mass to stellar luminosity and to total mass of dark halos , as functions of cluster x - ray temperature , and compare with observational data on present - day clusters .
section 6 compares the model with observational data on the evolution of icm abundances with redshift .
section 7 examines the predicted metal enrichment histories of clusters in more detail .
finally , section 8 presents our conclusions .
the appendix describes how we correct the observational data for radial abundance gradients , to estimate global icm metallicities . throughout this paper , the cosmological parameters are fixed to be @xmath10 km s@xmath11mpc@xmath11 = 0.7 and @xmath12 .
the solar abundances used are as given by @xcite .
some observational data based on the old solar abundances given by @xcite have been corrected to be compatible with the new more recently determined solar values .
we calculate galaxy formation in the framework of the cdm model using the semi - analytical model galform described in ( * ? ? ?
* hereafter clbf ) , with some modifications as described in ( * ? ? ?
* hereafter b04 ) .
the semi - analytical model calculates various processes as follows : ( a ) dark matter halos form by hierarchical merging ; ( b ) gas falls into these halos and is shock - heated ; ( c ) gas which cools in halos collapses to form rotationally supported galactic disks ; ( d ) stars form _ quiescently _ in disks ; ( e ) supernova explosions inject energy into the gas , ejecting some of it from the galaxy ; ( f ) dynamical friction brings galaxies together within common dark halos , leading to galaxy mergers ; ( g ) galaxy mergers can cause the transformation of stellar disks into spheroids , and can also trigger _ bursts _ of star formation .
the model also computes the luminosities and colours of galaxies from their star formation histories and metallicities , using a stellar population model and including a treatment of dust extinction .
the clbf model assumed a cosmic baryon fraction @xmath13 , that all stars formed with a kennicutt ( 1983 ) imf , and that gas ejected from galaxies by supernova feedback was retained in the host dark halo .
@xcite showed that if the more recent estimate @xmath14 was used , the original model predicted too many very luminous galaxies .
they proposed two possible solutions to this problem : either ( i ) _ thermal conduction _ suppresses the cooling of gas in high mass halos , or ( ii ) gas is ejected from the halos of galaxies by _
superwinds_. b04 adopted the _ superwind _ solution , but showed that in order to explain the numbers of lyman - break and sub - mm galaxies observed at high redshift , further changes are required in the model : they proposed changes in the star formation timescale in galaxy disks , in the triggering of starbursts by galaxy mergers , and in the imf of stars formed in bursts . in this paper , we adopt the b04 version of our galaxy formation model as the standard one , except that we retain both the _ thermal conduction _ and _ superwind _ variants .
our implementation of these two processes is similar to , but somewhat simpler than that of @xcite . for comparison purposes
, we will also present some results for the original unmodified clbf model . in the remainder of this section , we give some more details about some of the ingredients of our galaxy formation model .
our modelling of chemical enrichment is described in section [ metals ] . _
halo mergers and gas cooling : _ the merger histories of dark halos are generated for _ parent _ halos at @xmath15 , for a given power spectrum of density fluctuations , using a monte carlo scheme based on the extended press - schechter ( ps ) formalism ( press & schechter 1974 ; clbf ) . the lowest mass progenitor dark halos are assigned baryons in the form of diffuse gas only , according to the universal baryon fraction @xmath16 .
the diffuse gas is raised by shock heating to the virial temperature corresponding to the depth of the halo potential well .
a fraction of the diffuse hot gas cools and falls into the centres of the host dark halos , where it accretes onto the disk of the central galaxy .
the rate at which hot gas cools within a dark halo is computed by assuming a spherically symmetric gas distribution and using metallicity - dependent cooling functions .
the predictions of this simple cooling model compare favourably with the results of direct gas - dynamics simulations @xcite .
_ star formation in disks : _ the rate at which stars form from the cold gas in a galactic disk is given by @xmath17 , where @xmath18 is the mass of cold gas .
clbf used a star formation timescale that depended on the dynamical time of the galaxy . here , we adopt a prescription without any scaling with dynamical time , following b04 : @xmath19 where @xmath20 and @xmath21 are free parameters independent of redshift , whose values are set by requiring that the model reproduce the observed gas mass to luminosity ratios of present - day disk galaxies .
we employ the supernova feedback model described by clbf , in which cold gas is reheated and ejected into the halo by supernovae explosions according to the prescription : @xmath22 where @xmath23 and @xmath24 and @xmath25 are free parameters .
the latter is fixed to @xmath26 and @xmath24 is tabulated in table [ tab : param ] for the models considered in this paper .
as described above , for the standard baryon fraction employed in the model , @xmath27 , our simple cooling and feedback models result in the formation of too many giant luminous galaxies compared with recent determinations of the galaxy luminosity function ( e.g. * ? ? ?
we have investigated two different approaches to solving this problem , which we now describe : ( i ) the suppression of gas cooling by thermal conduction and ( ii ) the suppression of star formation by galactic superwinds . _
thermal conduction : _ in the first approach , thermal conduction of energy from the outer to the inner parts of the gas halo acts to balance the energy removed from the gas by radiative cooling .
this scenario is suggested by the lack of emission lines from low temperature gas in the x - ray spectra of clusters ( e.g. * ? ? ?
* ) . to model this phenomenon in a simple way
, we completely suppress cooling in halos whose circular velocity exceeds @xmath28 .
an estimate of the value of @xmath28 and its dependence on redshift can be derived by balancing the cooling rate with the heating rate due to thermal conduction : @xmath29 where @xmath30 and @xmath31 are the number densities of electrons and ions respectively , estimated here as their mean values within the virial radius , @xmath32 is the cooling function , @xmath33 is the thermal conductivity , @xmath34 is the temperature of hot gas , and @xmath35 is the radius of the halo ( e.g. , * ? ? ?
assuming a mean halo density given by the spherical collapse model , the spitzer form of conductivity without saturation , @xmath36 , and a power law cooling rate for thermal bremsstrahlung , @xmath37 , we obtain @xmath38 here @xmath39 is given by @xmath40 where @xmath41 is the ratio of hot gas to total baryon mass in a halo , @xmath42 gives the ratio of the conductivity to the spitzer value , and @xmath43 is the mean overdensity of a virialized halo relative to the average density of the universe .
( we have assumed @xmath44 and @xmath45 . ) for a flat universe with present day density parameter @xmath46 , @xmath47 @xcite .
assuming @xmath48 ( a realistic value in our models ) then gives @xmath49 for conduction at the spitzer rate ( @xmath50 ) . to match the bright end of the observed luminosity function , we show in section 4 that we require @xmath51 ; this implies @xmath52 if @xmath48 .
while the cut - off circular velocity required increases to @xmath53 km s@xmath11 if the redshift dependence of @xmath28 is neglected , the implied efficiency of conduction still seems unrealistically high . if conduction at this level is not viable , another mechanism needs to be invoked in order to prevent too much gas from cooling in massive halos .
one plausible , but as yet relatively unexplored alternative mechanism is the heating of the hot gaseous halo by emission from an active galactic nucleus ( e.g. * ? ? ?
. this may be particularly relevant in the more modest mass halos in which the merger rate peaks @xcite .
here we do not enter into the the details of the growth and fuelling of supermassive black holes in galaxy mergers .
however , we retain the freedom to choose apparently implausibly low values of @xmath28 on the grounds that other phenomena in addition to conduction may be operating within the halo , with the same consequences , namely the suppression of gas cooling in halos above a certain mass .
_ superwinds : _ in the second approach we have taken to prevent the formation of too many bright galaxies , we assume that gas is ejected out of galaxy halos by superwinds .
the superwind is driven by star formation , and ejects cold gas out of the galaxy and out of the halo at a rate given by @xmath54 where @xmath55 , \label{eq : sw}\ ] ] @xmath56 is a parameter , and @xmath57 is the circular velocity of the galactic disk for quiescent star formation or the circular velocity of the bulge in a starburst . in practise , we adopt the same values for the strength of the superwind , @xmath56 , in the quiescent and burst modes of star formation .
this is a simplified version of the scheme proposed by @xcite .
the form adopted for @xmath56 means that the superwind becomes less effective at ejecting cold gas from galaxies with @xmath58 .
the gas ( and associated metals ) expelled from the halo are assumed to be recaptured once the halo is subsumed into a more massive object with a halo circular velocity in excess of @xmath59 .
the free parameters in the superwind feedback are thus @xmath56 and @xmath60 .
_ galaxy mergers and starbursts : _ following a merger of dark matter halos , the most massive progenitor galaxy becomes the central galaxy in the new halo , and the remaining galaxies become its satellites .
the orbits of the satellite galaxies decay due to dynamical friction .
if the dynamical friction timescale of a satellite is shorter than the lifetime of the new halo ( defined as the time taken for its mass to double ) , the satellite will merge with the central galaxy .
the consequences of the galaxy merger are determined by the mass ratio of the satellite to the primary , @xmath61 . if @xmath62 , the merger is classified as a _ major merger _ ; the merger remnant is a spheroid and any cold gas is consumed in a starburst .
the duration of the starburst is proportional to the dynamical timescale of the bulge . for mergers with @xmath63 ,
the merger is classified as a _
minor merger_. in this case , the stars of the satellite are added to the bulge of the primary , but the stellar disk of the primary remains intact . if @xmath64 and the ratio of cold gas mass to total stellar and gas mass exceeds @xmath65 , the minor merger triggers a burst of star formation in which all the gas from the primary disk participates .
we adopt the same values for the parameters governing the physics of galaxies mergers as used by b04 : @xmath66 and @xmath67 .
we refer the reader to that paper for further discussion of the merger model .
the standard model presented in this paper in both its _ thermal conduction _ and _ superwind _ variants is able to successfully reproduce many observed properties of present - day galaxies , such as luminosity functions , gas fractions , and sizes of galaxy disks , as will be shown in @xcite .
the comparison of the _ superwind _ variant of the model with observations of lyman - break and sub - mm galaxies at high redshift has already been published in @xcite .
the predictions for stellar metallicities of galaxies will be presented in a subsequent paper @xcite .
our galaxy formation model thus has baryons in four different components : stars in galaxies , cold gas in galaxies , hot gas in halos , and hot gas outside halos .
there are exchanges of material between stars and cold gas , cold gas and hot gas inside halos , cold gas and hot gas outside halos , and hot gas outside and inside halos , as described in the previous section .
in addition , mergers of halos mix their reservoirs of hot gas , and mergers of galaxies mix the stellar and cold gas components of the two galaxies .
the exchanges of mass and metals between the different baryonic components are shown schematically in fig .
[ fig : exchanges ] .
our model keeps track of the effects of all of these mass exchanges and merging events on the metallicities of the different components .
finally , we must calculate the production of new heavy elements inside stars and their ejection back into the gas phase .
we describe now how we model this last process .
stars lose mass through stellar winds as they evolve .
some stars ultimately undergo supernova explosions which pollute the interstellar medium with newly produced metals .
the rates of both of these phenomena depend upon the mass of the star .
the total amount of gas recycled and metals produced per unit stellar mass depends , therefore , on the adopted imf . in this section
we explain how we obtain the gas restitution rate , the rates of sne ii and sne ia and the chemical yield for a single stellar population .
this information is then incorporated directly into the galform model , which predicts the star formation histories of individual galaxies , in order to obtain these quantities for the model galaxies .
the discussion below is based upon that given by @xcite with slight differences .
stellar data are extracted from @xcite for massive stars and @xcite for intermediate and low mass stars .
the chemical yields of sne ia are taken from @xcite .
= we define the imf as the number of stars per logarithmic interval of stellar mass per unit total mass of stars formed : @xmath68 normalised to unity when integrated over the whole range of mass @xmath69 so that @xmath70 has units of @xmath71 . in this paper
, we use the imf proposed by @xcite for quiescent star formation in disks , and a flat imf in starbursts triggered by galaxy mergers .
the @xcite imf has @xmath72 for @xmath73 and @xmath8 for @xmath74 . in starbursts , @xmath7 . in both modes of star formation , the lower and
upper stellar mass limits of the imf are @xmath75 .
[ fig : imf ] shows the quiescent ( solid line ) and starburst ( dashed line ) imfs .
in addition to the visible stars with @xmath76 , we allow for brown dwarfs with @xmath77 . the mass fraction of brown dwarfs is quantified by @xmath78 , which is the ratio of the total mass in stars to the mass in visible stars for a newly formed population of stars .
thus , the mass fraction in visible stars is @xmath79 .
clbf adopted the kennicutt imf in both the quiescent and burst modes of star formation , and required @xmath80 to match the predictions of their fiducial model to the observed present - day galaxy luminosity function around @xmath1 .
the other models we consider in this paper have @xmath81 .
the recycled mass fraction from an evolving population of stars is a function of time . for a population of stars formed with an imf @xmath70 at time @xmath82 , the total mass fraction released by time @xmath83 is @xmath84 \
, \phi(m)\frac{dm}{m } , \label{eq : recycle}\ ] ] where @xmath85 is the mass of a star reaching the end of its life at time @xmath83 , and @xmath86 is the mass of the remnant left by a star with initial mass @xmath87 . in our model , since the timestep we use is roughly comparable to the lifetime @xmath88yr of an 8@xmath89 star , the lowest mass star that results in a sne ii , we can justifiably use the instantaneous recycling approximation for sne ii .
we therefore instantaneously release the mass recycled by stars with @xmath90 . in fig .
[ fig : e ] , we show @xmath91 for both imfs considered .
the dotted lines show how these predictions change if the assumption of instantaneous recycling is relaxed .
= the cumulative number of sne ii explosions up to time @xmath83 ( per @xmath89 of stars formed ) , @xmath92 , and the mass fraction of metals of the @xmath93-th element , @xmath94 , released up to time @xmath83 are given respectively by @xmath95 where @xmath96 is the mass of the @xmath93-th element released from a star with mass @xmath87 . because @xmath94 includes metals that already existed when the stars formed
, these pre - existing metals must be subtracted in order to estimate the chemical yields , @xmath97 .
if the metallicity of stars with respect to the @xmath93-th element at their birth is @xmath98 , @xmath99 can be divided into two terms , @xmath100 where @xmath101 and @xmath102 indicates the mass of newly produced @xmath93-th element in a star with mass @xmath87 .
portinari et al . (
1998 ) show that the dependence of @xmath103 and @xmath104 on initial stellar metallicity is small , so for simplicity we neglect this dependence and use the values for solar initial metallicity throughout . in figs .
[ fig : r2 ] and [ fig : p ] , we show @xmath92 and @xmath105 , respectively . in both cases ,
the solid and dashed lines denote the kennicutt and top - heavy imfs respectively .
the initial time for which results are plotted corresponds to @xmath106 and the final time to @xmath107 , where @xmath108 is the lifetime of a star of mass @xmath87 . in fig .
[ fig : p ] , the lines indicate the yields of o , si , mg , fe from top to bottom at the right hand side of the panel . in this paper , because instantaneous recycling is assumed for sne ii , we only use the values of chemical yields at @xmath107 .
note that we rescale the mg yields from sne ii , @xmath109 , upwards by a factor of 4 , so that the mg abundances predicted by the model are more consistent with observed abundances in solar neighbourhood stars .
this correction is standard practice in chemical evolution modelling , and is justified because the theoretical calculations of the yield of mg are known to have a large ( factor @xmath110 ) uncertainty @xcite . in reality ,
any correction to the yields @xmath111 is likely to depend on stellar mass , so the correction factor for the integrated yield @xmath109 could be different for the two imfs we use . in our model , the enrichment of galactic disks is dominated by stars formed with the kennicutt imf , while the @xmath0-element enrichment of the icm is predicted to be dominated by stars formed with the top - heavy imf , and in fact , comparison with solar neighbourhood abundances favours a lower correction factor ( @xmath1122 - 3 ) for the mg yield than do the icm abundances .
however , for simplicity we have assumed the same correction factor for both imfs .
= = the standard scenario @xcite for sne ia is that they occur in binary systems , in which a c - o white dwarf produced by evolution of the primary star accretes gas from the secondary star when the secondary overflows its roche lobe at the end of its own evolution .
this gas accretion drives the mass of the primary white dwarf over the chandrasekhar limit , causing it to explode .
we model this using the scheme of @xcite , with parameters updated according to @xcite .
the progenitors of sne ia are assumed to be binary systems with initial masses in the range @xmath113 , where @xmath114 , and @xmath115 and @xmath116 are the initial masses of the primary and secondary ( @xmath117 ) respectively .
we assume @xmath118 , where @xmath119 is the largest single - star mass for which the endpoint is a c - o white dwarf .
the binary star systems which are the progenitors of sne ia are assumed to have an initial mass function @xmath120 , where @xmath70 is the same as for the single - star imf .
the distribution of mass fractions for the secondary , @xmath121 , is assumed to have the form ( normalised over the range @xmath122 ) @xmath123 for a single generation of stars formed at @xmath82 , the cumulative number of sne ia explosions up to time @xmath83 is then given by @xmath124 \frac{dm_{b}}{m_{b } } , \label{eq : ratesni}\ ] ] where the lower limit @xmath125.\ ] ] is set by the conditions that the secondary has evolved off the main sequence and that @xmath126 .
following portinari et al .
( 1998 ) , we adopt @xmath127 , @xmath128 and @xmath129 .
the factor @xmath130 is usually chosen ( for given @xmath131 and @xmath132 ) to reproduce the observed ratio of rates of sne ia to sne ii in spiral galaxies .
following portinari et al . , we adopt @xmath133 .
this observational normalisation is derived for an imf similar to kennicutt , but for simplicity we assume that the same normalisation applies to the top - heavy imf also . in fig .
[ fig : r1 ] , we show the cumulative number of sne ia as a function of time , @xmath134 , for the kennicutt imf ( solid line ) and the top - heavy imf ( dashed line ) . the number of intermediate mass stars responsible for sne ia is larger in the case of the kennicutt imf than for the top - heavy imf ( as shown by fig .
[ fig : imf ] ) .
this is in contrast to the situation for sne ii .
the yield of metals from sne ia is computed as @xmath135 where @xmath136 is taken from the w7 model of @xcite and takes the values of 0.143 , 8.5@xmath137 , 0.154 and 0.626 @xmath89 for o , mg , si and fe respectively . = note that if the imf includes a fraction of brown dwarfs ( with @xmath138 ) , the values in equations ( [ eq : recycle ] ) , ( [ eq : ratesnii ] ) , ( [ eq : yieldsnii ] ) , ( [ eq : ratesni ] ) and ( [ eq : yieldsni ] ) must be multiplied by @xmath79 , because @xmath138 corresponds to having a fraction @xmath79 of visible stars with @xmath139 .
we treat the metal enrichment due to sne ii in the same way as clbf , except that the instantaneously returned mass fraction and yields ( denoted @xmath35 and @xmath140 in clbf ) are replaced here by @xmath141)$ ] and @xmath142)$ ] .
the mass released from evolved stars after @xmath107 and the metals produced by sne ia are computed at every timestep .
the gas and metals recycled from disk stars are put into the cold gas in the galaxy , while those from bulge stars are put into the hot gas in the halo .
this picture is consistent with many evolutionary models of disk galaxies , in which metals from sne ia are put into the gaseous disk ; observations suggest that elliptical galaxies do not keep the metals released from sne ia , because the metal abundances in the interstellar medium of ellipticals are either smaller than or at best similar to solar abundances @xcite .
the mass released from stars during the time interval @xmath143 is computed from @xmath144\psi(t')dt',\ ] ] where @xmath145 is the star formation rate at time @xmath83 for the galaxy under consideration . in order to solve this equation ,
the star formation histories in both modes of star formation , quiescent and starburst , must be recorded separately for individual galaxies . due to the memory overhead involved
, the star formation histories are rebinned into 30 linear steps in time .
we have confirmed that our results are insensitive to the precise number of steps used . as well as the returned mass ,
the metals produced by evolving stars , including sne ia , are computed in a similar way .
the mass of the @xmath146 element released from evolved stars and sne ia over the time interval [ @xmath147 is @xmath148\psi(t')dt'\nonumber\\ + \hspace{-1mm}\int_{0}^{t_{i+1}}\hspace{-1mm}z(t ' ) [ e(\leq t_{i+1}-t')-e(\leq t_{i}-t')]\psi(t')dt',\end{aligned}\ ] ] where @xmath149 is the metallicity of stars at their birth at time , @xmath83 ; this quantity is stored in the same fashion as the star formation histories .
the metals are also assumed to be put into cold and hot phases along with the recycled gas .
in this paper we consider three principal models : the fiducial model of clbf ( which assumed @xmath150 ) and two new models which have @xmath151 .
the first of these new models is _ superwind _ model , in which gas ejection from the halo is invoked to suppress the formation of bright galaxies . in the second new model , which we refer to as the _ conduction _ model , cooling in massive halos
is suppressed by thermal conduction . in the clbf model ,
all stars form with a kennicutt imf . in the _ superwind _ and _ conduction _ models ,
the kennicutt imf is adopted for quiescent star formation in disks and a flat imf is used in starbursts .
the superwind model is similar to the model used by baugh et al .
( 2004b ) .
we also consider a variant of the superwind model ( which we denote _ superwind / kennicutt _ ) in which all star formation produces stars with a kennicutt imf .
we list the parameters for the different models in table [ tab : param ] . [
cols="^,^,^,^,^,^,^,^,^,^,^,^ " , ] = an important constraint on the models is that they should be consistent with the observed present - day galaxy luminosity function . in fig .
[ fig : lf ] we compare the model predictions for the present - day @xmath152-band luminosity function with observational estimates .
the predictions of the superwind model with @xmath153 km s@xmath11 are shown by the solid line , those of the conduction model with @xmath154 km s@xmath11 by the short dashed line , and the luminosity function of the clbf model is plotted using a long dashed line .
the symbols show observational data .
although it is challenging for models with the currently favoured value of @xmath155 to match the observed luminosity function as well as in the clbf model , as shown by @xcite , the superwind and conduction models are in broad agreement with the observations . in the conduction model
, the value of the parameter @xmath39 which is required to produce a good match with the observed luminosity function seems unphysically small .
the value of @xmath156 requires a thermal conductivity that is many times larger than the spitzer value . as explained in section 2 , processes other than thermal conduction could also contribute to the suppression of cooling , so we retain the freedom to adopt such a low value for this parameter . in fig . [
fig : lfvrecap ] we show how the luminosity function in the superwind model depends on the parameter @xmath59 , the threshold circular velocity above which halos recapture gas previously ejected in a superwind .
clearly if @xmath59 is too low , then too many very bright galaxies are produced .
on the other hand , the galaxy luminosity function becomes insensitive to increases of this parameter above @xmath157 km s@xmath11 .
however , as we will show later , the value of @xmath59 has a strong impact on metal abundances in the icm .
the predicted icm metal abundances are too low unless the metals ejected in the superwind can be recaptured by cluster - sized halos .
thus , we require a value @xmath158 km s@xmath11 in order to reproduce both the observed field galaxy luminosity function and the metal abundance of the icm .
a basic test for any model of the chemical enrichment of the icm is that it should reproduce the observed ratios of the icm mass to total stellar luminosity and to total mass in clusters .
this is because the metals are produced in stars , so a model which produces the observed icm metallicity is invalid unless it predicts the correct number of stars at the same time . in fig .
[ fig : ml ] , we plot the ratio @xmath159 of the mass of the icm , @xmath160 , to the total @xmath161-band stellar luminosity of the galaxies , @xmath162 , as a function of the icm temperature @xmath34 . we show predictions for the superwind , conduction and clbf models . in the models ,
the icm temperature is assumed to equal the halo virial temperature , @xmath163 while for the observations we use the measured x - ray temperature .
the relationship between cluster mass and virial temperature in the @xmath6cdm model at @xmath15 is then @xmath164 .
we compare the models with observational data from @xcite and @xcite .
we have combined the data from individual clusters in bins in x - ray temperature , and computed the mean and dispersion in each bin , weighted according to the errors on the individual datapoints .
the error bars on the plotted datapoints are calculated from the dispersion in each bin , and do not include any other measurement errors .
the conduction and clbf models predict a ratio @xmath165 which increases only slightly with cluster temperature .
the superwind model predicts a somewhat larger increase , because of the expulsion of gas from lower - mass halos , but again is almost flat for halos with @xmath166 , corresponding to @xmath167kev which is indicated by the vertical dashed line in the top panel .
the models are in approximate agreement with the observational data for @xmath168kev , but for lower temperatures , the observations suggest icm fractions significantly smaller than what the models predict . however , there are still important uncertainties in the observational data ( not included in the error bars plotted ) resulting from the need to extrapolate x - ray measurements out to the cluster virial radius , so the icm fractions in cooler clusters are not yet securely determined as shown below .
we note that the clbf model predicts very similar @xmath169 values to the superwind and conduction models , even though it has a global baryon fraction only half that of the other models .
this agreement comes about because the clbf model predicts a luminosity density which is a factor @xmath170 lower in all environments compared to the other models . in fig .
[ fig : mgasmhalo ] we show the ratio of the icm mass to the total cluster mass including dark matter as a function of cluster temperature .
the dots and crosses indicate the hot gas fractions for the superwind and the conduction models , respectively .
the dashed horizontal line corresponds to the cosmic baryon - to - total mass ratio ; the model points would all lie on this line if all baryons were retained in the cluster in the hot gas component .
we also show observational data from @xcite , binned in temperature as in fig .
[ fig : ml ] .
the superwind model reproduces the qualitative trend seen in the observations that the gas fraction drops in low temperature clusters .
however , we caution that the observational estimates rely on extrapolating both the icm and dynamical masses from the outermost measured x - ray datapoint out to the halo virial radius .
future more accurate x - ray observations will allow more precise tests of models for gas ejection from halos .
= = we now compare the icm metallicities predicted by our models with observational data .
our model predicts global metallicities , averaged over the entire icm within the cluster virial radius .
in contrast , the icm metallicities measured from x - ray data are often only representative of the central regions of the cluster , because these dominate the x - ray emission . some clusters , known as _
cooling flow _ clusters , show metallicity gradients for some elements , in the sense that metallicities increase towards the cluster centre .
it is therefore essential to correct the observational data to estimate mean metallicities averaged over the entire cluster , before comparing with the models .
we compare our model predictions with the following observational data on icm abundances : ( * ? ? ?
* _ bepposax _ ) for fe [ see also @xcite and @xcite ] ; ( * ? ? ?
* _ xmm - newton _ ) for o , fe , mg and si [ see also @xcite ] ; ( * ? ? ?
* _ asca _ ) for fe and si ; and ( * ? ? ?
* _ asca _ ) for si .
we explain in the appendix why we have omitted some observational data from our comparison .
we use the results of the systematic study of iron gradients by @xcite to convert the measured central metallicities of iron and silicon to global average values , assuming that silicon has the same gradient as iron , as is suggested by observational data .
we present the details of how we make these corrections in the appendix .
the size of these estimated corrections is up to a factor of two .
oxygen and magnesium do not show metallicity gradients @xcite , so we assume that the global abundances of these elements are equal to the measured central values .
we note that previous studies ( e.g. de lucia et al . 2004 ) have not included any correction for metallicity gradients when comparing models with observed icm abundances .
= 0.7 we first show how the predicted metal abundances in the icm depend on the imf adopted for starbursts .
[ fig : mh ] shows the o , fe , mg and si abundances ( relative to solar values ) as functions of cluster temperature for the superwind model .
the dots show the predicted values for the standard superwind model with a top - heavy imf ( @xmath7 ) in bursts and a kennicutt imf in disks , and the crosses show a variant superwind model in which disks and bursts both have a kennicutt imf . for clusters with @xmath171kev , the abundances of the @xmath0-elements o , mg and si are 34 times higher in the case of the top - heavy imf in bursts , while the abundance of fe is only about 60% higher .
this difference reflects that fact that the @xmath0-elements are produced mostly by sne ii , and the number of sne ii is larger for a top - heavy imf , while fe is produced by both sne ia and sne ii , and the number of sne ia is actually reduced for the top - heavy imf . in fact , the models predict that most fe is produced by sne ia for a kennicutt imf , but by sne ii for the top - heavy imf . when in fig .
[ fig : mh ] we examine the dependence of icm abundances on cluster temperature ( or mass ) in the superwind model , we see that the abundances of the different elements all show a break at @xmath172 kev .
this temperature corresponds exactly to the halo circular velocity @xmath153 km s@xmath11 above which gas and metals ejected by superwinds are recaptured .
halos with @xmath173 effectively trap all of the metals ever produced by the galaxies they contain . for clusters hotter than the break , the abundances
are predicted to be almost constant .
the jump in abundances across the break is about a factor 2 for o , mg and si , but is much smaller for fe , especially if a kennicutt imf is used in bursts .
this again reflects the fractions of these different elements produced in either sne ii or sne ia .
since sne ii also drive the superwinds , while sne ia are typically delayed until after the strong superwind phase has subsided , superwinds are much more effective in expelling from halos the heavy elements produced in sne ii than those produced in sne ia . if @xmath59 is varied , then the metal abundances in clusters with @xmath166 are insensitive to the value of @xmath60 for @xmath174 km s@xmath11 , but increase slightly if @xmath175 km s@xmath11 , because more stars are produced in that case ( c.f .
fig.[fig : lfvrecap ] ) .
[ fig : mh ] also shows the observed icm abundances of o , mg , si and fe , corrected to global average values as we have described .
we see that the superwind model with a top - heavy imf in bursts agrees well with the observed abundances of o , mg and si , while the superwind model with a kennicutt imf predicts abundances of these elements which are too low by factors 23 .
for fe , the model with the top - heavy imf again agrees better with the observations , although the abundance is still slightly low .
we note that the current abundance measurements for o , mg and si are all for clusters hotter than the break predicted in the models .
however , the value of @xmath59 can not be significantly increased above our standard value of 600 km s@xmath11 without moving the abundance break into the temperature range where the observations indicate a plateau in abundances , thus conflicting with the data .
the measured fe abundances , which extend down to somewhat lower cluster temperatures , do in fact suggest a break close to the temperature predicted by the superwind model , but more measurements for cool ( @xmath176kev ) clusters are needed to confirm this .
it is clear that more abundance measurements in cool clusters , especially of o , mg and si , would put stringent constraints on our model of ejection of gas and metals from halos by superwinds .
the models also predict a very small scatter in global icm abundances at a given cluster temperature , especially above the break .
the observed abundances show a larger scatter , but most of this results from measurement errors . in fig .
[ fig : mh2 ] we show the icm metal abundances predicted by our other two models : the conduction model with a top - heavy imf in bursts , and the clbf model with a kennicutt imf for all modes of star formation .
the models are compared with the same observational data as in fig .
[ fig : mh ] .
for both of these models , halos retain all of their metals , and consequently the predicted icm abundances are almost constant with cluster temperature .
the clbf model predicts abundances too low compared to observed values , by factors 23 for o , mg and si , and by around 40% for fe , and is thus ruled out .
the conduction model predicts abundances in much better agreement with observations , although slightly too high for o and si .
the abundances in the conduction model are slightly higher overall than those in the superwind model .
= 0.7 it should be remembered that we have applied a correction to the yield of mg from sne ii predicted by stellar models , in order to better match observational constraints .
the model yields have been multiplied by a factor of four throughout , a value which is a compromise between that required to reproduce mg abundances in galactic disks ( specifically , in the solar neighbourhood ) and in the icm ; a separate correction factor could be applied to each imf because of the mass - dependent uncertainties in estimating theoretical yields .
no correction factors have been applied to the yields of the other elements : o , fe and si .
examining the abundance ratios between different elements in the icm provides an alternative method to constrain the form of the imf , in which the effects of uncertainties in the icm to stellar mass ratio cancel out . in fig .
[ fig : mfe ] we show the abundance ratios of different @xmath0 elements to iron , [ o / fe ] , [ mg / fe ] and [ si / fe ] , for the standard superwind model , with a top - heavy imf in bursts , and the variant superwind model , with a kennicutt imf in bursts .
we also show observational data from @xcite , which in the case of si and fe have been corrected for abundance gradients in the way already described .
we see that the predicted o / fe and mg / fe ratios agree better with observations for the model with the top - heavy imf , while on the other hand the si / fe ratio agrees better with the model having a kennicutt imf in bursts . the overall level of agreement of the abundance ratios with observational data in the top - heavy imf model could be improved if the fe abundances in the model were slightly larger ( this can also be seen in fig . [
fig : mh ] ) , or if the observed si abundances were slightly larger .
the latter would occur if the abundance gradients in si were weaker than we have assumed , since we would then infer larger global si abundances from the same measured central values .
however , recent data from @xcite , in which metal abundances are measured out to @xmath177 kpc from cluster centres , seem to support our original assumption that the radial gradients in si and fe are the same . = fig .
[ fig : mfe2 ] is the equivalent of fig [ fig : mfe ] for the conduction and clbf models .
since the conduction model predicts slightly higher fe abundances than the superwind model , the agreement with the observed abundance ratios improves .
the clbf model fails to reproduce the observed abundance ratios .
the models predict very little variation in the abundance ratios with cluster temperature , apart from the break in the superwind model where recapture sets in .
they also predict very small scatter for the abundance ratios at a given cluster temperature .
= in fig .
[ fig : metalcomp ] we compare the predicted mass - weighted mean iron metallicities for the different baryonic components in clusters : stars in galaxies , cold gas in galaxies , and hot intracluster gas .
results are shown for the standard superwind and conduction models as functions of cluster temperature . in both models ,
the mean stellar and cold gas metallicities are close to solar , while the hot gas metallicity is 23 times lower .
the mean total metallicity is only slightly larger than the hot gas metallicity , reflecting the fact that most of the baryons are in the hot gas component . for the superwind model
, the onset of recapture of gas by halos produces the expected jump in the hot gas and total metallicities , but does not produce any corresponding feature in the metallicities of the stars or cold gas .
this is because the enrichment of these components has almost finished by the formation epoch of the cluster and the epoch of recapture of expelled metals ( see section 7 ) .
= = fig .
[ fig : metalfrac ] shows the fractions of the total mass of iron which are in stars , cold gas , hot gas and ` lost ' gas which is expelled from halos by superwinds , for the same models as in fig .
[ fig : metalcomp ] . in all cases ,
more of the iron is in the hot gas than in galaxies . in the superwind model , superwinds put about 40% of the iron outside halos for low mass clusters below the threshold for recapture .
thus , according to this model , a significant mass of metals should be observed in the surroundings of low temperature clusters .
thus , in summary , from the icm results at @xmath15 , we conclude that a top - heavy imf with @xmath178 in starbursts is required for the model predictions to match the observed icm metallicities , as long as we adopt the kennicutt or a similar imf for disk star formation .
the use of a kennicutt or a similar imf for disk star formation is supported by many observations and models of solar neighbourhood stars , including a semi - analytic model by @xcite in which a salpeter - like imf provides good agreement with the observed metallicities of solar - neighbourhood stars .
we have also confirmed that while using the arimoto - yoshii imf @xcite with @xmath179 for both star formation modes substantially increases metallicities compared with the kennicutt imf , a combination of the kennicutt imf for quiescent star formation and the arimoto - yoshii imf for starbursts does not reach the observed metallicities .
with the advent of the new generation of @xmath180-ray satellites , measurements of icm metallicities are now becoming possible in high - redshift clusters , thus allowing direct tests of icm evolution . in fig .
[ fig : fez ] , we show what our models predict for the evolution of the iron abundance in the icm with redshift in the range @xmath181 . at each redshift , we calculate the mean fe abundance for clusters in two different ranges of temperature , @xmath182kev and @xmath183kev .
we show results for the standard conduction and superwind models . in the same figure
we also show observational data from @xcite , based on measurements of the iron abundances in 18 distant clusters with @xmath184 using _ chandra _ and _ xmm - newton_. we have binned this data in redshift and in the same ranges of temperature as used for the models .
note that we have no information about metallicity gradients for these distant clusters , so we have not applied any corrections to the measured abundances to get global values .
we also show observational data for @xmath15 clusters binned in the same ranges of temperature .
the @xmath15 data are corrected for metallicity gradients as previously described .
our models predict a weak decline in the fe abundance with increasing redshift , by about 30% over the range @xmath181 , and furthermore this decline is predicted to be the same for both low and high temperature clusters .
@xcite concluded from their own data that the evolution in fe abundance with redshift is negligible , but we see that the data for the hotter ( @xmath183kev ) clusters are also quite consistent with the weak decline predicted by our models .
the data for the cooler ( @xmath182kev ) clusters suggest higher fe abundances and a weak increase with redshift relative to the hotter clusters , both of which would conflict with our model , but the uncertainties are very large .
= the abundances of elements other than fe have not yet been measured in high - redshift clusters , so we simply present predictions for the evolution of o abundances from our model . for clusters selected in the same temperature range at different redshifts , the mean o abundance is predicted to decline with redshift even more weakly than for fe , by only @xmath185 in over the range @xmath181 .
the difference in behaviour is caused by the difference in the timescales for enrichment by sne ia and sne ii .
the o / fe ratio in the icm is therefore predicted to increase with redshift .
this is shown in fig .
[ fig : ofez ] . for the superwind model
, there is a significant difference in o / fe between low ( 0.7 - 1 kev ) and high ( 3 - 8.5 kev ) temperature clusters , caused by recapture of superwind ejecta in clusters with @xmath186kev .
in this section , we investigate what our model predicts for the enrichment histories of individual clusters . for this purpose ,
we divide clusters into three populations based on their _ present - day _ icm temperatures : low ( 0.7 - 1 kev ) , intermediate ( 1.5 - 4 kev ) and high ( 4 - 8 kev ) , corresponding to cluster masses @xmath187 , @xmath188 and @xmath189 respectively .
these ranges are chosen to illustrate different behaviours regarding recapture of gas by halos in the superwind model .
for the low temperature clusters , a significant quantity of metals and gas have been expelled from the progenitors of these halos without recapture .
the intermediate temperature clusters experience recapture of gas and metals at low redshifts , so we can see how the recapture process affects the evolution of the icm metallicity .
the high temperature clusters have already recaptured gas and metals at high redshifts . in fig .
[ fig : hist1 ] we show the evolution of the halo mass and icm metallicity for the main progenitors of present - day clusters , for a randomly chosen set of clusters in each range of present - day temperature .
we define the main progenitor of a present - day cluster at any earlier epoch as follows @xcite : starting from the present - day halo , we follow the halo merger history back in time , and at each merger event , we follow the branch in the merger tree corresponding to the more massive progenitor involved in that merger event .
we define the formation redshift @xmath190 of a cluster as the earliest redshift at which the main progenitor has at least half of the present cluster halo mass @xcite . for @xmath191 ,
main _ progenitor is also the _ most massive _ progenitor , but at earlier times this may not be the case . the top row of panels in fig .
[ fig : hist1 ] shows the evolution of the halo mass of the main progenitor relative to the present - day halo mass , illustrating the spread in cluster assembly histories .
we plot each cluster trajectory in fig .
[ fig : hist1 ] as a solid line for @xmath192 and as a dotted line for @xmath193 . in the @xmath6cdm model
, clusters typically form around @xmath194 , but with a wide dispersion , and with more massive clusters typically forming more recently .
the medians and the 10% and 90% percentiles of the distribution of @xmath190 values are ( 0.64,0.38,0.95 ) for low temperature , ( 0.57,0.27,0.79 ) for medium temperature , and ( 0.46,0.21,0.75 ) for high temperature clusters respectively .
the lower panels of fig .
[ fig : hist1 ] show the evolution of the icm abundances in the main progenitor of each cluster , for the same set of cluster halo assembly histories as in the top row .
the second and third rows show the fe abundance in the superwind and conduction models , while the fourth row shows the o abundance in the superwind model only .
the fe abundances typically evolve only weakly with redshift over the range @xmath195 , which is similar to the behaviour seen in fig .
[ fig : fez ] . the dispersion in the final metallicity and in the metallicity evolution is larger in the lower temperature clusters . for the intermediate temperature clusters ,
the dispersion in the metallicity evolution is larger in the superwind model than in the conduction model , due to the ejection and then recapture of gas in the former case .
these recapture events can be seen as abrupt upward jumps in metallicity in panel ( e ) .
the evolution of the o abundance in the superwind model shows a somewhat larger dispersion than the evolution in fe abundance .
the jumps in o metallicity due to recapture of superwind material are larger than for fe ( compare panels ( k ) and ( e ) ) , because the superwinds are preferentially enriched in products of sne ii rather than sne ia .
some of the evolution tracks for the o abundance show a _ decline _ in mean icm abundance with _ increasing _ time .
this behaviour can be understood as follows : the dispersion in o abundance between halos of the same mass increases with decreasing halo mass , due to variations in the number of @xmath196 galaxies within a halo resulting from the stochastic nature of the halo assembly and galaxy formation processes .
as the halo mass builds up , the metallicity tends to converge towards the average value for clusters with that present - day temperature : progenitor halos with above - average icm metallicity tend to merge with other halos having lower metallicity , and the resulting mixing of the hot gas components produces the declining tracks seen for instance in panel ( j ) .
= in fig .
[ fig : hist2 ] we show the build - up of the total mass of metals found in the present - day icm for the same set of clusters as in fig .
[ fig : hist1 ] .
we calculate these evolutionary tracks by summing over the icm metal masses of all of the progenitors of a present - day cluster .
we include in this sum metals which have been ejected by superwinds , if these metals have been recaptured by the present day .
these tracks thus allow us to determine the redshifts at which metals in the icm of the present - day cluster were actually produced ( in the sense of being ejected from stars in supernova explosions ) .
the top two rows of panels show the evolution of the iron mass in the superwind and conduction models , for different ranges of present - day cluster temperature .
the bottom row shows the evolution of the oxygen mass in the superwind model .
most of the metals are typically produced well before the clusters are assembled , with the oxygen being produced even earlier than the iron .
for example , in the superwind model , half of the o is produced by redshifts @xmath197 for the low , medium and high temperature clusters respectively , compared to @xmath198 for producing half of the fe .
the corresponding redshifts in the conduction model are not much different : @xmath199 and @xmath200 .
thus , the icm enrichment occured mostly in many smaller progenitors of a present - day cluster .
we also see that metal production typically occured slightly earlier in high mass than low mass clusters , even though high mass clusters were assembled ( in the sense of reaching half of their present mass ) slightly later .
we can compare our results for the redshift of metal production in clusters with the results from the semi - analytical model of @xcite .
de lucia et al .
use the instantaneous recycling approximation throughout , so their results can be considered to apply to oxygen , but are not valid for iron , for which the delayed production by sne ia is important .
furthermore , they only compute the icm evolution for the single example of a @xmath201 cluster . for three different feedback / ejection schemes
, they calculate that 60 - 80% , 35 - 60% , and 20 - 45% of the metals are incorporated into the icm at redshifts larger than 1 , 2 , and 3 , respectively .
note that in their definition , metals are `` incorporated '' into the icm either when they are ejected from galaxies directly into the icm , or when metals ejected from the cluster are recaptured . comparing with the predictions for high temperature clusters in our own models , we find that the fraction of oxygen incorporated into the icm at redshifts larger than 1 , 2 and 3 is 47% , 23% , and 10% respectively for the superwind model , and 79% , 51% , and 24% for the conduction model .
thus in our superwind model , o is incorporated into the icm later than in the models of de lucia et al . , while in our conduction model , the incorporation happens at similar redshifts .
( note that the `` metal incorporation '' redshifts given here are different from the `` metal production '' redshifts derived from fig .
[ fig : hist2 ] , since the latter include ejected metals among the metals produced by a given redshift , provided these ejected metals have been recaptured by the present day . ) since we assume the same @xmath6cdm cosmology as de lucia et al .
, the differences must arise from using somewhat different assumptions about gas cooling , star formation , feedback and gas ejection .
mergers of halos are expected to lead to extensive mixing of the icm , which will tend to erase any metallicity gradients .
the prediction of the models that typically most of the iron in the icm is produced well before half of the cluster mass has been assembled then seems to make it harder to understand the iron abundance gradients seen in `` cooling flow '' clusters .
however , since the production of iron by sne ia is long - delayed , it may still be possible to build up iron abundance gradients in the central regions of clusters by ejection of iron from the central galaxy after most of the cluster mass has been assembled .
this problem needs to be investigated using gas - dynamical simulations of cluster formation .
we have also calculated the build - up of metals within cluster galaxies , counting metals in both stars and in the cold gas .
when we compute trajectories for the evolution of the total mass in metals in galaxies summed over all cluster progenitors , relative to the present metal mass , we find that there is very little dispersion between different clusters of the same mass ( much less than for the icm metal mass shown in fig .
[ fig : hist2 ] ) .
the trajectories are also very similar for different ranges of cluster temperature , and for the superwind and conduction models . for clusters in the temperature range @xmath202kev , half of the o mass in cluster galaxies is produced by redshifts @xmath203 for the superwind model , and by @xmath204 in the conduction model .
for the fe mass in cluster galaxies , the corresponding redshifts are @xmath205 in the superwind model , and @xmath206 in the conduction model .
the metals in cluster galaxies are thus produced significantly earlier than the metals in the icm .
= finally , we examine which galaxies contribute the most to cluster metal enrichment . fig .
[ fig : metaldist ] shows the total amount of iron which galaxies have produced ( either retained within the galaxy , injected into the icm , or ejected from the cluster ) as a function of their @xmath152-band luminosities at @xmath15 .
the contribution of each galaxy is calculated by summing the iron production over its formation history .
the distributions over luminosity @xmath207 are then normalized to unit area .
results are shown for both the superwind and conduction models , and also split according to present - day cluster temperature .
the solid line in each panel shows the iron production averaged over halos of all masses ( including halos much smaller than clusters ) .
these distributions are quite similar to the distributions of total stellar mass , which is expected since the mass of iron produced is closely related to the mass of stars formed .
we also find that the normalized distributions for the mass of iron injected into the icm are very similar to the distributions in fig .
[ fig : metaldist ] for the total mass of iron produced .
the distributions of iron production in fig .
[ fig : metaldist ] for the clusters show pronounced peaks at the high luminosity end .
as the cluster mass or temperature increases , the position of the peak moves up in luminosity , but its amplitude decreases .
this peak is produced by the central galaxies in the cluster halos , and reflects the distribution of total stellar mass @xcite .
the peak is most prominent for the lowest temperature clusters , because mergers are most effective at concentrating the stellar mass into a single central galaxy in such systems .
since the position of the peak due to central galaxies depends on halo mass , it is smeared out when we average over all halos , producing a smoother distribution ( see the solid lines in fig .
[ fig : metaldist ] ) .
we find that in the superwind model , half of the iron is produced by galaxies with luminosities @xmath208 brighter than @xmath209 , @xmath210 and @xmath211 for low , medium and high temperature clusters respectively , corresponding to galaxy stellar masses @xmath212 , @xmath213 and @xmath214 respectively . the median galaxy masses and luminosities for oxygen production are almost identical to those for iron . in the conduction model ,
the median galaxy masses and luminosities for metal production are somewhat higher : for iron production , the median k - band luminosities are @xmath215 , @xmath216 and @xmath217 , and the median stellar masses are @xmath218 , @xmath219 and @xmath220 respectively , with very similar values for oxygen production .
when we look at metal production averaged over halos of all masses , we find median luminosities @xmath208 for producing fe or o of @xmath210 and @xmath221 respectively for the superwind and conduction models , and median stellar masses of @xmath222 and @xmath223 respectively .
we thus see that the median k - band luminosity ( or stellar mass ) for metal production is similar ( within a given model ) for high mass clusters and field galaxies , but significantly larger in low mass clusters . for comparison ,
the characteristic luminosity of field galaxies in the k - band is @xmath224 @xcite .
the metal production in clusters and in the field is thus predicted to be dominated by galaxies with @xmath225 . in our models ,
the median galaxy luminosities and masses for metal injection into the icm are almost identical to those for metal production .
our median stellar masses for metal production are 23 times larger in the case of the most massive clusters than the median galaxy mass of @xmath226 for metal ( effectively oxygen ) injection into the icm calculated by @xcite for a @xmath201 cluster .
however , we are in agreement de lucia et al . on the general conclusion that the icm metal enrichment is dominated by fairly massive galaxies , in contrast to suggestions ( e.g. * ? ? ?
* ) that metal ejection by dwarf galaxies has dominated the chemical enrichment of the icm .
we have investigated the metal enrichment of the icm for clusters of different masses using a semi - analytical model of galaxy formation and evolution , galform ( cole et al .
we have implemented in this model a detailed treatment of metal production by both sne ia and sne ii , based on the predictions of stellar evolution models for both stellar lifetimes and yields of different heavy elements .
this enables us to follow the evolution of different elements ( o , mg , si and fe ) , including the long time delays for sne ia . our model consistently calculates the cycling of gas and metals between stars , cold gas and hot gas .
in order to reproduce the observed present - day galaxy luminosity function using a realistic baryon fraction in the models , we have considered two alternative mechanisms to prevent the overproduction of bright galaxies : ( a ) galactic superwinds which expel gas and metals from dark halos , with subsequent recapture in very massive halos , and ( b ) thermal conduction which transfers heat from the outer to the central regions of hot gas halos , thereby supressing gas cooling .
we considered two different imfs for bursts of star formation : the kennicutt imf that we also adopt for quiescent disk star formation , and a top - heavy , flat imf with @xmath7 .
we have found that the form of the imf is strongly constrained by the metal abundances of o , mg , si and fe in the icm .
our main conclusions can be summarized as follows : 1 .
a top - heavy imf in starbursts is required in order to obtain icm metal abundances as high as those observed , when the imf for quiescent star formation in disks is fixed to be a standard , bottom - heavy imf , such as the kennicutt imf .
a burst imf with @xmath7 reproduces the observed metal abundances in the icm .
we have also tried the arimoto - yoshii imf with @xmath179 in bursts , and confirmed that it fails to produce sufficient metals .
considering that the imf for quiescent star formation has been well constrained by many studies , including a semi - analytical model @xcite , this is strong evidence in support of a top - heavy imf with @xmath178 in starbursts .
2 . both the superwind and conduction models reproduce the observed icm metal abundances and abundance ratios .
the superwind model predicts lower metal abundances in low temperature ( @xmath227 kev ) clusters .
therefore if the drop in [ fe / h ] for clusters with @xmath227 kev found using _ asca _ is real
, this could be evidence for the action of superwinds .
the current data are still limited , however , so future observations , particularly of low temperature clusters , would strongly constrain the superwind model .
our model is in broad agreement with the observed ratio of icm gas mass to the total @xmath161-band stellar luminosity in clusters , @xmath159 . in particular ,
the superwind model predicts a decrease of @xmath228 in low temperature clusters , as is observed , because the superwind expels gas from these clusters .
the iron metallicity , [ fe / h ] , is predicted to decrease only slowly with increasing redshift .
the difference in [ fe / h ] between @xmath15 and 1 is at most 0.2 dex .
this mild evolution is broadly consistent with the recent observations by @xcite .
in contrast with [ fe / h ] , [ o / h ] is almost independent of redshift over this range , so [ o / fe ] increases with redshift .
the metal production histories of clusters having the same present - day mass are very similar .
in contrast , the cluster mass assembly histories show a wide variety .
this suggests that the variety of metallicity gradients observed in clusters may be caused not only by observational errors but also by the variety of assembly histories .
presumably this is the origin of the diversity in the observed metallicities of present - day clusters , because most observations measure metallicities only at cluster centres .
6 . in clusters and in the field ,
metal production is dominated by galaxies with k - band luminosities similar to or brighter than @xmath1 . in low - mass clusters ,
the contribution of the central galaxy to metal production is very important , but this effect is weaker in high - mass clusters . throughout this paper , we have neglected any contribution of intracluster stars to the metal enrichment of the icm , since in our model all stars are in galaxies .
based on recent measurements of the intracluster light , it has been suggested that intracluster stars might have produced a large fraction of the iron observed in the icm @xcite .
the intracluster stars would presumably originate from material tidally stripped from galaxies or ejected during violent mergers , so the luminosity function of galaxies in clusters would also be affected . in our model ,
tidal stripping of stars from galaxies would not affect the total mass of iron produced ( if all other parameters remain the same ) , but it might affect the fraction of the iron which is ejected into the icm .
this possibility clearly deserves further observational and theoretical study .
it is also evident that , in order to better constrain our model , we need more accurate icm metallicity measurements for larger samples of clusters .
these measurements need to cover the outer parts of clusters as well as the central regions , in order derive accurate global metallicities .
the observational situation will be substantially improved by the _ astro - e2 _ satellite scheduled for launch in 2005 , as well as by further data from _ xmm - newton _ and _ chandra_. further constraints on the model can be obtained by examining the abundances of different heavy elements in galaxies , and we will investigate this latter topic in future papers .
we thank takashi okamoto , richard bower , stefano ettori , simon d. m. white , yasushi fukazawa and takeshi go tsuru for useful discussions . we also acknowledge support from the pparc rolling grant for extragalactic astronomy and cosmology at durham .
mn is supported by the japan society for the promotion of science for young scientists ( no.207 ) .
cmb is supported by a royal society university research fellowship .
anders e. , grevesse n. , 1989 , geochim .
acta , 53 , 197 arimoto n. , yoshii y. , 1987 , a&a , 173 , 23 baugh c.m . , et al .
( the 2dfgrs team ) , 2004a , mnras , 351 , l44 baugh c.m . ,
lacey c.g .
, frenk c.s .
granato g.l .
silva l. , bressan a. , benson a.j .
, cole s. , 2004b , mnras , in press ( astro - ph/0406069 ) ( b04 ) baumgartner w.h . , loewenstein m. , horner d.j . ,
mushotzky r.f . , 2003 ,
apj , submitted ( astro - ph/0309166 ) bell e.f . ,
mcintosh d.h . , katz n. , weinberg m.d . , 2003 , apj , 149 , 289 benson a. j. , bower r.g .
, frenk c. s. , lacey c. g. , baugh c. m. , cole s. , 2003a , apj , 599 , 38 benson a.j . ,
frenk c.s .
, baugh c.m . , cole s. , lacey c.g .
, 2003b , mnras 343 , 679 cole s. , 1991 , apj , 367 , 45 cole s. , aragon - salamanca a. , frenk c. s. , navarro j. f. , zepf s. e. , 1994 , mnras , 271 , 781 cole s. , lacey c. g. , baugh c. m. , frenk c. s. , 2000 , mnras , 319 , 168 ( clbf ) cole s. et al .
2001 , mnras , 2001 , 326 , 225 david l.p . , forman w. , jones c. , 1991 , apj , 380 , 39 de grandi s. , molendi s. , 2002 , apj , 567 , 163 de grandi s. , ettori s. , longhetti m. , molendi s. , 2004 , a&a , 419 , 7 de lucia , g. , kauffmann , g. & white , s.d.m . , 2004 , mnras , 349 , 1101 eke v.r . , cole s. , frenk c.s . , 1996 , mnras , 282 , 263 enoki m. , nagashima m. , gouda n. , 2003 , pasj , 55 , 133 ettori s. , de grandi s. , molendi s. , 2002 , a&a , 391 , 841 fukazawa y. , makishima k. , tamura t. , ezawa h. , xu h. , ikebe y. , kikuchi k. , ohashi t. , 1998 , pasj , 50 , 187 garnett , d.r . , 2002 ,
apj , 581 , 1019 gibson b.k . ,
matteucci f. , 1997a , mnras , 291 , l8 gibson b.k .
, matteucci f. , 1997b , apj , 475 , 47 granato g.l . , de zotti g. , silva l. , bressan a. , danese l. , 2004 , apj , 600 , 580 .
greggio l. , renzini a. , 1983 , a&a , 118 , 217 grevesse n. , sauval a.j . , 1998 , space sci .
, 85 , 161 hatton , s. , devriendt , j.e.g . ,
ninin , s. , bouchet , f.r . ,
guiderdoni , b. , vibert , d. , 2003 , mnras , 343 , 75 helly j.c . , cole s. , frenk c.s . ,
baugh c.m . ,
benson a.j . ,
lacey c.g . , 2003 , mnras , 338 , 903 huang j .- s . , glazebrook k. , cowie l.l . , tinney c. , 2003 , apj , 584 , 203 irwin j.a .
, bregman j.n . , 2001 , apj , 546 , 150 jenkins a. , frenk c. s. , white s. d. m. , colberg j. m. , cole s. , evrard a. e. , couchman h. m. p. , yoshida n. 2001 , mnras , 321 , 372 kauffmann g. , 1996 , mnras , 281 , 475 kauffmann g. , charlot s. , 1998 , mnras , 294 , 705 kauffmann g. , haehnelt m. , 2000 , mnras , 311 , 576 kauffmann g. , white s. d. m. , guiderdoni b. , 1993 , mnras , 264 , 201 kennicutt r.c . , 1983 , apj , 272 , 54 kroupa p. , 2002
, 295 , 82 lacey c.g .
, cole s. , 1993 , mnras 262 , 627 lacey c.g . ,
2005 , in preparation larson r.b . , 1969 ,
mnras , 145 , 297 lia c. , portinari l. , carraro g. , 2002 , mnras , 330 , 821 loewenstein m. , mushotzky r.f . , 1996 , apj , 466 , 695 maoz d. , gal - yam a. , 2004 , mnras , 347 , 951 marigo p. , 2001
, a&a , 370 , 194 matsushita k. , ohashi t. , makishima k. , 2000 , pasj , 52 , 685 matsushita k. , finoguenov a. , bhringer h. , 2003 , a&a , 401 , 443 matteucci f. , gibson b.k . , 1995 , a&a , 304 , 11 matteucci f. , greggio l. , 1986 , mnras , 154 , 279 matteucci f. , franois p. , 1989 ,
mnras , 239 , 885 menci n. , cavaliere a. , fontana a. , giallongo e. , poli f. , 2002 , apj , 575 , 18 moretti a. , portinari l. , chiosi c. , 2003 , a&a , 408 , 431 mushotzky r. , loewenstein m. , arnaud k.a . , tamura t. , fukazawa y. , matsushita k. , kikuchi k. , hatsukade i. , 1996 , apj , 466 , 686 nagashima m. , gouda n. , 2001 , mnras , 325 , l13 nagashima m. , okamoto t. , 2004 , submitted ( astro - ph/0404486 ) nagashima m. , totani t. , gouda n. , yoshii y. , 2001 , apj , 557 , 505 nagashima m. , yoshii y. , totani t. , gouda n. , 2002 , apj , 578 , 675 nagashima m. , yoshii y. , 2004 , apj , 610 , 23 nagashima m. , et al .
2005 , in preparation navarro j.f . , frenk c.s . , white s.d.m .
, 1997 , apj , 490 , 493 nomoto k. , iwamoto k. , nakasato n. , thielemann f .- k . , brachwitz f. , tsujimoto t. , kubo y. , kishimoto n. , 1997 , nuclear physics , a621 , 467c norberg p. , et al .
( the 2dfgrs team ) , 2002 , mnras , 336 , 907 okoshi k. , nagashima m. , gouda n. , yoshioka s. , 2004 , apj , 603 , 12 pagel b.e.j . ,
tautvaiien g. , 1995 , mnras , 276 , 505 peacock j.a . , et al . (
the 2dfgrs team ) , 2001 , nature , 410 , 169 .
percival w.j . , et al .
( the 2dfgrs team ) , 2002 , mnras , 337 , 1068 peterson j.r . , et al .
, 2001 , a&a , 365 , l104 peterson j.r . ,
kahn s.m . ,
paerels f.b.s . ,
kaastra j.s .
, tamura t. , bleeker j.a.m . , ferrigno c. , jernigan j.g . , 2003 , apj , 590 , 207 portinari l. , chiosi c. , bressan a. , 1998 , a&a , 334 , 505 press w. , schechter p. , 1974
, apj , 187 , 425 renzini a. , ciotti l. , dercole a. , pellegrini s. , 1993 , apj , 419 , 52 .
salpeter e.e . , 1955 ,
apj , 121 , 161 sanderson a.j.r .
, ponman t.j . ,
finoguenov a. , lloyd - davies e.j . ,
markevitch m. , 2003 , mnras , 340 , 989 sanderson a.j.r . , ponman t.j .
, 2003 , mnras , 345 , 1241 smith l.j .
, gallagher iii j.s . , 2001 , mnras , 326 , 1027 somerville r.s .
, primack j. r. , 1999 , mnras , 310 , 1087 somerville r.s .
, primack j. r. , faber s.m .
, 2001 , mnras , 320 , 504 spergel d.n . , et al .
( the wmap team ) , 2003 , apjs , 148 , 175 takahara m. , takahara f. , 1981 , ptp , 65 , 369 tamura t. , bleeker j.a.m . , kaastra j.s . , ferrigno c. , molendi s. , 2001 , a&a , 379 , 107 tamura t. , kaastra j.s . , den herder j.w.a .
, bleeker j.a.m .
, peterson j.r . , 2004 , a&a , 420 , 135 thomas d. , 1999 , mnras , 306 , 655 thomas d. , greggio l. , bender r. , 1998 , mnras , 296 , 119 thomas d. , kauffmann g. , 1999 , in spectrophotometric dating of stars and galaxies , ed
. i. hubeny , s. heap & r. cornett , vol .
192 ( asp conf . ser . ) , 261 timmes f.x . ,
woosley s.e . ,
weaver t.a .
, 1995 , apjs , 98 , 617 tornatore l. , borgani s. , matteucci f. , recchi s. , tozzi p. , 2004 ,
mnras , 349 , 19 tozzi p. , rosati p. , ettori s. , borgani s. , mainieri v. , norman c. , 2003 , apj , 593 , 705 tsujimoto t. , nomoto k. , yoshii y. , hashimoto m. , yanagida y. , thielemann f .- k .
, 1995 , mnras , 277 , 945 tsuru t.g . , awaki h. , koyama k. , ptak a. , 1997 , pasj , 49 , 619 valvarnini r. , 2003 , mnras , 339 , 1117 whelan , j. , iben , i. , 1973 , apj , 186 , 1007 white , s.d.m .
& frenk , c.s .
, 1991 , apj , 379 , 52 yoshida n. , stoehr f. , springel v. , white s.d.m . , 2002 , mnras , 335 , 762 yoshii y. , tsujimoto t. , nomoto k. , 1996 , mnras , 462
, 266 zaritsky d. , gonzalez a.h . ,
zabludoff a.i . , 2004 ,
apj , in press ( astro - ph/0406291 ) zepf s.e . , silk j. , 1996 , apj , 466 , 114
some clusters are known to have substantial metallicity gradients .
the so - called ` cooling flow ' clusters show enhancements in the metallicities of some elements toward the cluster centre ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
@xcite have reported that in cooling flow clusters o and mg among the @xmath0 elements do not show the metallicity gradients , whereas fe and si tend to show gradients , at least from the cluster centre to around half of the virial radius , despite the fact that si is an @xmath0 element .
non - cooling flow clusters generally do not show metallicity gradients for any elements .
our model predicts the global metallicity of the icm , averaged over the whole cluster .
therefore we need to correct the observed metallicities to obtain values averaged over the whole cluster , because most observations give only central values .
@xcite compiled fe metallicity gradients in 12 cooling flow clusters . based on their results , we assume that all cooling flow clusters have the same shape of metallicity gradients in both fe and si , @xmath229^{-\alpha } , \label{eqn : metalgrad}\ ] ] where @xmath230 , @xmath231 is the radius within which the mean overdensity is 200 times the critical density , @xmath232 and @xmath233 .
we define the average metallicity within radius @xmath234 as , @xmath235 where @xmath236 is the gas density .
we assume that the observed metallicities correspond to average values within some radius @xmath234 , @xmath237 , and then use the assumed form of the metallicity profile ( equation ( eqn : metalgrad))to correct these to mean values @xmath238 within the virial radius @xmath239 .
the central metallicity @xmath240 is thus treated as a parameter chosen so as to match an observed metallicity @xmath237 .
@xcite give the iron masses of individual clusters as a function of @xmath241 , which is a spherical overdensity , @xmath242 where @xmath243 is the total mass within a radius @xmath234 and @xmath244 is the critical density of the universe at redshift @xmath245 .
using the data on the distributions of total and gas mass given by @xcite , we obtain an averaged metallicity for each cluster within @xmath246 , @xmath247 . the gas density distribution , @xmath236 is assumed to follow an isothermal profile with a core , @xmath248^{-1}.\ ] ] to obtain @xmath246 , we use the structural parameters given by @xcite assuming an nfw profile @xcite for the total density distribution .
we have found that the use of structural parameters assuming a king profile hardly affects averaged metallicities .
@xmath249 is assumed to be the same as the scale radius of the nfw profile , but the results are not sensitive to the exact value chosen .
the metallicity at the cluster centre , @xmath240 in eq.([eqn : metalgrad ] ) , depends on the individual cluster . using the above equations and the values at @xmath250 , at which the observed data are given in the above papers , we obtain the average metallicity @xmath251 within the virial radius ( @xmath252 , corresponding to @xmath253 in a @xmath6cdm universe ) , @xmath254 the average [ fe / h ]
is easily estimated from @xmath251 .
peterson et al . ( 2003 ) observed only the central parts of clusters within a @xmath255 angular radius .
their observed metallicities of fe and si are assumed to correspond to @xmath256 , where @xmath257 is the radius at which the angular size is equal to 1 arcmin .
the average metallicity within the virial radius is then obtained from @xmath258 the parameters regarding redshift and structure required to obtain averaged metallicities are tabulated in their paper .
recent observations suggest that the metallicity gradients of si are very similar to those of fe @xcite .
therefore we assume that si has the same shape of metallicity profile as fe .
as already mentioned , because o and mg seem not to have any metallicity gradients , we regard the observed values of central metallicities as equivalent to the metallicities averaged over the whole cluster . the measurements of fe and si abundances by fukazawa et al .
( 1998 ) include data from the outer parts of the clusters .
we therefore assume that their abundance measurements are close to global averages , and do not apply any corrections .
@xcite obtained average abundances of several elements , including fe and si , in bins of cluster temperature by performing a stacking analysis of observations of 273 clusters from the _ asca _ archives .
although their cluster sample is large , metallicity gradients combined with the large point spread function for _ asca _ cause problems for interpreting their results , especially for fe .
the iron abundance is usually estimated from the fe - k or fe - l lines . for high - temperature clusters ,
the photon flux in the fe - k line is larger than that in the fe - l line , and the fe - k line yields accurate measurements of the iron abundance . on the other hand , for low and intermediate temperature clusters , the fe - k line becomes very weak , and the photon flux is larger in the fe - l line , so the fe - l line dominates in abundance estimates .
lower temperature gas emits more photons in the fe - l line than in the fe - k line .
this has the consequence that for clusters with cooling cores , the observationally - estimated iron abundance is weighted towards the value in the low temperature component , that is , the central region , where the iron abundance is typically higher than the average over the whole cluster .
thus , as @xcite have done , the data for the central region should be removed to get a fairer estimate of the metallicity averaged over the whole cluster .
@xcite have not done this , and this may be the reason why their data show an enhancement of the iron metallicity at intermediate cluster temperatures .
for this reason , we do not include their data on fe abundances in our comparison with the models .
their data on si abundances seem to be less affected by this problem , so we do include those data in our comparison . | we investigate the metal enrichment of the intracluster medium ( icm ) in the framework of hierarchical models of galaxy formation .
we calculate the formation and evolution of galaxies and clusters using a semi - analytical model which includes the effects of flows of gas and metals both into and out of galaxies . for the first time in a semi - analytical model
, we calculate the production of both @xmath0 and iron - peak elements based on theoretical models for the lifetimes and ejecta of type ia and type ii supernovae ( sne ia and sne ii ) .
it is essential to include the long lifetimes of the snia progenitors in order to correctly model the evolution of the iron - peak elements .
we find that if all stars form with an imf similar to that found in the solar neighbourhood , then the metallicities of o , mg , si and fe in the icm are predicted to be 2 - 3 times lower than observed values . in contrast ,
a model ( also favoured on other grounds ) in which stars formed in bursts triggered by galaxy mergers have a top - heavy imf reproduces the observed icm abundances of o , mg , si and fe . the same model predicts ratios of icm mass to total stellar luminosity in clusters which agree well with observations . according to our model , the bulk of the metals in clusters are produced by @xmath1 and brighter galaxies .
we predict only mild evolution of [ fe / h ] in the icm with redshift out to @xmath2 , consistent with the sparse data available on high - z clusters .
by contrast , the [ o / fe ] ratio is predicted to gradually decrease with time because of the delayed production of iron compared with oxygen .
we find that , at a given redshift , the scatter in global metallicity for clusters of a given mass is quite small , even though the formation histories of individual clusters show wide variations .
the observed diversity in icm metallicities may thus result from the range in metallicity gradients induced by the scatter in the assembly histories of clusters of galaxies .
stars : luminosity function , mass function galaxies : clusters : general galaxies : formation large - scale structure of the universe |
having crossed the boundaries between theory and experiment , i attest that the experimenter feels like a hapless ant crawling on a desk , as far as searching for new physics ( np ) @xmath0 violation ( cpv ) is concerned .
take , for example , the belle _
_ paper @xcite on the difference in direct cpv between @xmath7 and @xmath8 ( @xmath9 ) . in reporting a large deviation ,
belle cited the baryon asymmetry of the universe ( bau ) as the reason to pursue cpv studies , but immediately admitted that all data support the unique kobayashi
maskawa ( km , @xcite ) source of cpv in the standard model ( sm ) , which is known to be too small " ( by @xmath1 @xcite at least ) for bau .
because the gap ( which the general experimenter knows only vaguely ) between sm and the heavenly bau is so large , it appears insurmountable , no matter what is found in the laboratory .
it is truly remarkable that the sm has @xcite all the necessary ingredients for baryogenesis , i.e. the sakharov conditions of baryon number violation , cpv , and deviation from equilibrium ( in the very hot early universe ) .
the agony is the insufficiency in the latter two : cpv is way to small , while the electroweak phase transition ( ewpht ) seems only a crossover .
we see no antibaryons in our universe , i.e. @xmath10 , while @xmath11 ( wmap ) ; bau is 100% . but
the folklore is that sm falls short by @xmath1 .
the source of this is the jarlskog invariant @xcite , @xmath12 which incorporates all requirements for cpv to be nonvanishing , where @xmath13 is twice the area of any unitarity triangle .
note that @xmath14 has dimensions @xmath15 . to compare with @xmath16 ,
one typically normalizes by the ewpht temperature @xmath17 gev ( or roughly the v.e.v .
putting in quark masses , and our knowledge that @xmath18 , one immediately finds @xmath19 , which falls short by @xmath1 .
the main source of suppression is the smallness of light quark masses .
the situation is in general much worse , since there are coupling constant factors as well .
_ we observe that , by extending from 3 to 4 quark generations , one can enhance eq . [ eq : j3 ] by over @xmath3 _
the thread that lead to this observation appeared concurrent with the 2004 observation of direct cpv in @xmath20 mode , i.e. the first hint of @xmath21 .
written into the belle paper @xcite at that time , it was noted that if the electroweak penguin @xmath22 ( @xmath23 penguin really ) is the source of this apparent difference , then np cpv is implied .
however , as is well known , the so - called color - suppressed tree diagram @xmath24 could also generate @xmath25 .
although peskin @xcite stressed the equal possibility of @xmath24 vs @xmath26 origins in his companion _ nature _ paper , privately he is very skeptical that the new belle result is new physics " .
so , with the gap of @xmath1 in mind , the hapless ant crawls on .
noticing that the @xmath22 ( or the @xmath23 penguin ) , where the @xmath23 is radiated off a virtual top or @xmath27 in a @xmath28 loop and turns into a @xmath29 ( but not a @xmath30 ) , i recalled my first b paper on @xmath31 @xcite .
naive counting would lead one to conclude that the photonic penguin diagram , at @xmath32 , would dominate over the @xmath23 penguin , at @xmath33 . even if one notes that the two differ by @xmath34 in dimensions , one would still have @xmath35 .
but it turns out , by direct computation , or by argument of conserved vector current vs spontaneous electroweak symmetry breaking ( ewsb ) , that the @xmath23 penguin behaves as @xmath36 and actually dominates .
this is called _ nondecoupling _ of heavy chiral quark masses in sm .
i therefore embarked on crawling up the little hill of adding a 4th generation .
but this usually appears as running against a wall in a quixotic way ; the 4th generation has long been viewed by many as ruled out already , by neutrino counting and electroweak precision tests ( ewprt ) .
however , we now know that neutrinos have mass , which calls for new physics , while kribs _
et al . _
@xcite recently pointed out that the 4th generation is not in such great conflict with ewprt . in any case
, we demonstrated , both at lo @xcite and nlo @xcite in pqcd factorization approach ( the only one that predicted the size and sign of @xmath37 ) , that the 4th generation can generate the observed @xmath38 . in ref .
@xcite we showed that @xmath39 and @xmath40 moved downwards by @xmath41 , which is the right direction and consistent with current data .
both the sign and strength are nontrivial .
@xmath42 is rather large for a np effect .
given that the @xmath43 box diagram has similar @xmath44 dependence as in the @xmath28 @xmath23 penguin , with the cdf measurement of @xmath45 mixing , a very sizable , and _ negative _ , mixing - dependent cpvis predicted @xcite for @xmath46 , i.e. @xmath47 compared with @xmath48 .
the range of @xmath49 to @xmath50 was already predicted in ref .
@xcite and reported @xcite at ichep 2006 in moscow .
the improvement of eq .
[ eq : sin2phibs ] came with the more precisely determined @xmath45 mixing , while the sign is determined by the sign of @xmath9 .
so at ichep 2006 , i already asked can large cpv in @xmath45 mixing be measured @ tevatron ? " , and pronounced that the case is good for the tevatron ( vs lhcb , which comes on later ) .
interestingly , by end of 2007 , cdf reported @xcite indications for @xmath51 that is consistent with eq .
[ eq : sin2phibs ] , but less consistent with the sm expectation . by this conference
, the d@xmath52 measurement @xcite and a cdf update @xcite both confirm this trend , and the combined deviation from sm is now @xcite more than 2@xmath53 , with central value of @xmath54 !
this incredible development makes 20092010 very interesting , whether lhcb arrives on the scene or not .
heavy sm chiral quark effects are nondecoupled in the box and @xmath23 penguin diagrams .
the source is both because of the subtleties of spontaneous ewsb , and that heavy quark masses are due to large yukawa couplings to the v.e.v .
this i knew since twenty some years .
stimulated by large @xmath9 , in the past 4 years i could not stop from pushing the work on the 4th generation , utilizing large @xmath55 yukawa couplings , and cpv phase in @xmath56 .
i can not remember when and how , but one day the
yureka(wa ) " moment came : large yukawa couplings can modify eq .
[ eq : j3 ] !
if one shifts by one generation with 4th generation sm ( sm4 ) , then eq .
[ eq : j3 ] becomes @xcite @xmath57 the notation will be clarified soon , but it is clear that the difference of light quark mass pairs , @xmath58 now all drop out , and one gains in the mass factors ( assuming @xmath59 gev ) by @xmath3 ! for the change in cpv area " @xmath13 , if the hints from @xmath9 and @xmath6 hold up , one could gain a further factor of 30 . to illustrate this last point , we show in figure [ bdbstri ] the @xmath28 quadrangle corresponding to the sm4 unitarity relation @xmath60 , together with the sm3 @xmath61 triangle @xmath62 .
the latter is from the current 3 generation fit to all data , the success of which lead to km receiving the 2008 nobel prize . for the former ,
it comes from the program @xcite that started with @xmath9 ( fixes @xmath56 for given @xmath63 ) , but incorporating the @xmath64 and rare kaon constraints on @xmath65 and @xmath66 , using unitarity of @xmath67 ckm matrix . -1.8 cm
-1.5 cm we note that , if one draws the line linking @xmath68 and @xmath69 in figure [ bdbstri ] , the rather squashed and elongated triangle corresponds to @xmath70 for sm3 , the 3 generation sm .
this triangle has the same area @xmath71 as the @xmath72 triangle .
it is the very tiny phase angle of the @xmath28 triangle in sm3 at the vertex @xmath68 that gives rise to the very small value of @xmath73 .
the sign , which is opposite to @xmath74 , is because the orientation " is opposite that of the sm3 @xmath72 triangle .
the large phase angle in sm4 at vertex @xmath68 leads to the large area of the quadrangle , or @xmath75 , hence our prediction of eq .
[ eq : sin2phibs ] . why do the @xmath72 processes give a triangle , rather than a quadrangle , if there are 4th generation effects lurking in @xmath76 transitions ?
this question was dealt with in ref .
@xcite : with large @xmath56 ( including cpv phase ) , after taking into account the @xmath64 and rare kaon constraints , the actual @xmath72 quadrangle mimics the sm3 triangle , or @xmath72 transitions are sm - like .
this is a nontrivial test , and indeed , another possible solution is rejected by this .
the original jarlskog invariant of eq .
[ eq : j3 ] was derived @xcite using @xmath77 \equiv { \rm im}\det\bigl[s,\;s'\bigr]$ ] .
jarlskog generalized to @xmath78 generations @xcite , and found the invariant cpv measure in terms of 3 cycles " , the trace of the cube of commutators of quark mass squares , or @xmath79 ^ 3 $ ] , which looks considerably more complicated . to cut a long story short ( and somehow never invoked by jarlskog in actual detail ) , note that we are close to the @xmath80-@xmath81 ( and @xmath82-@xmath83 as well ) degeneracy limit on the v.e.v . scale . in this degeneracy limit ,
the 4 generation world actually becomes the effectively 3 generation world of 2 - 3 - 4 generation quarks !
one sees now why eq .
[ eq : j234 ] would turn out to be by far the dominant , and why @xmath14 in eq .
[ eq : j3 ] , which could be written as @xmath84 , is so tiny ( the @xmath1 gap ! ) . out of the 3 independent phases in sm4 , one
is already measured in @xmath72 transitions , one could be emerging in a spectacular way in @xmath28 transitions .
a third subdominant phase can be glimpsed from figure [ bdbstri ] . since @xmath85 is small , the resulting triangle by shrinking @xmath86 is not much different from the quadrangle .
thus , we have been a little cavalier in the notation of @xmath87 , but again there is no doubt that @xmath88 of eq .
[ eq : j234 ] is the predominant cpv effect in sm4 , and the relevant one for bau .
judging from the combined @xmath89 enhancement from @xmath14 to @xmath88 , it seems sufficient to overcome the large gap of @xmath1 , even taking into account the gauge factors that we have alluded to .
what about ewpht ?
it is claimed that a first order transition is not possible for sm4 @xcite .
but perhaps strong yukawa couplings , beyond the unitarity limit of heavy @xmath55 and @xmath90 masses ( perturbativity is lost ) , opens a new possibility , as ewsb itself could be through the nambu jona - lasinio @xcite mechanism with large yukawa couplings .
given the developments at the tevatron on @xmath6 in the past year @xcite , 2009 appears extremely interesting , while lhcb may not deliver physics even by 2010 .
judging from the recent performance of the tevatron accelerator and experiments , if the current central value ( consistent with eq .
[ eq : sin2phibs ] ! ) stays , we would have evidence in 2009 , perhaps even observation in 2010 , _ if tevatron could continue running beyond 2009_. regardless , once lhcb has of order 0.5 fb@xmath91 data _ analyzed _ , whether one has np cpv enhancement or not , the whole situation would precipitate . but measurement of large @xmath6 , while exciting , does not constitute proof for a 4th generation . the real litmus test , as always , would be direct search .
current cdf limit gives @xmath92 gev at 90% c.l .
, using 2.8 fb@xmath91 data .
again , once lhc data becomes available , the full terrain can be covered in a straightforward way .
the gain of @xmath3 ( @xmath89 if @xmath93 gev is used ) in mass factors of eq .
[ eq : j234 ] with 4 generations , over eq .
[ eq : j3 ] with only 3 generations , seem to give enough cpv for generating bau .
_ maybe there is a 4th generation . _ in several years we should know whether the km mechanism with 4th generations could provide sufficient cpv for bau .
it would be amazing if what we find on earth really has something to do with ( baryo-)genesis ! 9 s.w .
lin , y. unno , w.s .
hou , p. chang _
et al . _ [ belle collaboration ] , _ nature _ * 452 * , 332 ( 2008 ) .
m. kobayashi and t. maskawa , _ prog .
_ * 49 * , 652 ( 1973 ) . m.e .
peskin , _ nature _ * 452 * , 293 ( 2008 ) ; and private communications .
kuzmin , v.a .
rubakov , and m.e .
shaposhnikov , _ phys .
_ b * 155 * , 36 ( 1985 ) . c. jarlskog , _ phys .
_ * 55 * , 1039 ( 1985 ) ; _ z. phys .
_ c * 29 * , 491 ( 1985 ) .
y. chao , p. chang _
et al . _ [ belle collaboration ] , _ phys .
_ * 93 * , 191802 ( 2004 )
hou , r.s .
willey and a. soni , _ phys .
lett . _ * 58 * , 1608 ( 1987 ) .
kribs , t. plehn , m. spannowsky , and t.m.p .
tait , _ phys .
_ d * 76 * , 075016 ( 2007 ) .
hou , m. nagashima , and a. soddu , _ phys .
_ * 95 * , 141601 ( 2005 ) .
hou , h.n .
li , s. mishima , and m. nagashima , _ phys .
lett . _ * 98 * , 131801 ( 2007 ) .
hou , m. nagashima , and a. soddu , _ phys .
_ d * 76 * , 016004 ( 2007 ) .
hou , arxiv : hep - ph/0611154 .
t. aaltonen , _ et al . _
[ cdf collaboration ] , _ phys . rev .
_ * 100 * , 161802 ( 2008 ) .
v.m . abazov _
[ d@xmath52 collaboration ] , arxiv:0802.2255 [ hep - ex ] .
see the parrelel session talk by d. tonelli , arxiv:0810.3229 [ hep - ex ] , with future projections .
see the plenary talk by m. paulini , this proceedings .
hou , arxiv:0803.1234 [ hep - ph ] .
hou , m. nagashima , and a. soddu , _ phys .
_ d * 72 * , 115007 ( 2005 ) .
c. jarlskog , _ phys .
_ d * 36 * , 2128 ( 1987 ) .
r. fok and g.d .
kribs , arxiv:0803.4207 [ hep - ph ] .
y. nambu and g. jona - lasinio , phys .
* 122 * ( 1961 ) 345 . | how to account for the matter predominance of our universe is a fundamental issue at the core of our existence .
one condition is @xmath0 violation , but the standard model falls short by more than @xmath1 . taking cue from a recent result from the @xmath2 factories , we find that a fourth quark generation can provide enhancement by a factor of @xmath3 or more
. this could be the source of @xmath0 violation for the baryon asymmetry of the universe .
the main source of enhancement is the large yukawa couplings of the heavy @xmath4 and @xmath5 quarks . with indications for a new
, large @xmath0 violating phase @xmath6 emerging at the tevatron , our suggestion can be verified or refuted at the lhc in the next few years . |
[ [ interlayer - phonon - modes - in - rese_2-and - res_2 ] ] * _ interlayer phonon modes in rese@xmath1 and res@xmath1 _ * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + figure [ fig1](a),(b ) shows optical images of @xmath0-layer rese@xmath1 and res@xmath1 , respectively .
the 1t structure of monolayer rex@xmath1 is shown in figure [ fig1](c),(d ) .
in contrast to the 1 t phase , the 1t phase displays covalent bonding between metal ( here re ) atoms . the covalently bound re atoms form diamond - like patterns that are connected to each other , leading to quasi one - dimensional re chains @xcite . in practice ,
rex@xmath1 crystals preferentially cleave along these chains ( see figure [ fig1](a),(b ) ) .
the 1t phase has considerably lower symmetry than the 1h phase @xcite .
monolayer 1t-rex@xmath1 belongs to the @xmath3 space group and is only endowed with one inversion center .
the monolayer unit cell of rex@xmath1 has 12 atoms giving rise to 36 zone - center phonon modes .
these modes have @xmath4 or @xmath5 symmetry and are infrared or raman active , respectively @xcite . in monolayer rex@xmath1 ,
one thus expects 18 raman - active modes .
figure [ fig1](e),(f ) shows the raman spectra of @xmath6 to @xmath2 layer and bulk rese@xmath1 and res@xmath1 , respectively , recorded at @xmath7 in the parallel ( xx ) configuration for randomly oriented samples . as previously reported @xcite , we can identify the complex manifold of intralayer modes in the range @xmath8 . in the following , we will only consider the fifth lowest frequency intralayer mode ( denoted mode v as in ref . ) at approximately @xmath9 ( respectively @xmath10 ) in rese@xmath1 ( respectively res@xmath1 ) .
this mode predominantly involves atomic displacement of the re - re bonds @xcite .
we now concentrate on the low - frequency raman response of @xmath0-layer rex@xmath1 , at raman shifts below 40 @xmath11 ( see left panels in fig .
[ fig1](e),(f ) and fig .
[ fig2 ] ) . for increasing @xmath12 ,
one observes well - defined stokes and anti - stokes peaks that we assign to the lsm and lbm branches .
as expected , no interlayer mode - features are observed for @xmath6 .
only two interlayer mode - features are resolved for @xmath13 and @xmath14 . as in 2hc - tmd
, two new features appear at @xmath15 ( see also figure [ fig2 ] ) and again at @xmath16 .
interestingly , the frequency of all the observed lsm and lbm seemingly decreases when @xmath0 increases , and ultimately merge into the rayleigh background for large @xmath0 .
this observation is in stark contrast with 2hc - tmd , where the frequencies of the lsm ( respectively , lbm ) have been shown to increase ( respectively , decrease ) with increasing @xmath0 @xcite and complicates the identification of the lsm and lbm .
remarkably , bulk rex@xmath1 crystals do not display any signature of finite frequency interlayer modes , as expected from group theory predictions for the 1t structure @xcite .
remarkably , bulk rex@xmath1 crystals do not display any signature of finite frequency interlayer modes .
this observation is consistent with a bulk 1t structure with one rex@xmath1 layer per bulk unit cell @xcite and contrasts with the case of bulk 2hc - tmd , which have two layers per bulk unit cell and display a raman active lsm @xcite . in the spectra shown in figure [ fig1](e),(f )
no clear splitting of the interlayer mode - features can be resolved .
figure [ fig2 ] shows higher resolution spectra recorded at a lower photon energy of @xmath17 . in both res@xmath1 and rese@xmath1 , peaks belonging to the lowest- and third - lowest frequency branches display small but measurable splittings of @xmath18 , while the features belonging to second - lowest frequency branch do not split . based on these observations ,
we assign the branches that do not exhibit any splitting to lbm and the branches that split to lsm , respectively .
we note that our assignment is consistent with previous measurements and calculations on layered compounds , which , for a given normal mode , systematically reported lower frequencies for lsm than for the lbm @xcite .
in addition , our direct measurement of the lsm splittings in rese@xmath1 are in - line with recent density functional theory ( dft ) calculations @xcite . [ [ fan - diagrams - and - interlayer - force - constants ] ] * _ fan diagrams and interlayer force constants _ * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + , @xmath19 , ) of ( a ) rese@xmath1 and ( b ) res@xmath1 and of the interlayer breathing modes ( @xmath20 , @xmath21 , ) of ( c ) rese@xmath1 and ( d ) res@xmath1 as a function of the number of layers @xmath0 ( solid symbols ) .
the measured frequencies are fit to eq .
[ eqlcm ] ( solid lines ) .
open symbols represent the infrared - active modes ( @xmath22 , @xmath23 , , and @xmath24 , @xmath25 , ) also predicted by eq .
[ eqlcm ] . ] based on these observations , we may now plot , in fig .
[ fig3 ] the fan diagrams associated with the two non - degenerate lsm and with the lbm in @xmath0-layer rex@xmath1 . as previously introduced for other layered materials @xcite , a linear chain model can be used to fit the frequencies of the observed modes and extract the interlayer force constants .
the lsm and lbm frequencies write @xmath26 with @xmath27 ( @xmath28 corresponds to the zero frequency acoustic mode at @xmath29 ) .
the observed raman - active modes correspond to branches with @xmath30 for the lsm and lbm .
the latter are denoted @xmath31 , @xmath19 , @xmath32, and @xmath33 , @xmath34 , @xmath35 ,
, respectively ( see fig .
[ fig3 ] ) . in eq .
[ eqlcm ] , @xmath36 denotes the non - degenerate lsm ( @xmath31 ) frequencies and the lbm ( @xmath37 ) frequency for bilayer rex@xmath1 , with @xmath38 the associated interlayer force constant and @xmath39 the mass per unit area in rex@xmath1 .
the resulting force constants are reported in table [ tabfc ] and compared to measurements on 2hc - tmd . and 2hc - mos@xmath1 in ( c ) and ( d ) , respectively ) .
the presence of an inversion center ( for any value of @xmath0 in the 1t phase and for even @xmath0 in the 2hc phase ) or of a mirror symmetry plane ( for odd @xmath0 in the 2hc phase ) is indicated .
the irreducible representation of each mode is indicated , with raman - active modes in red , infrared active modes and raman modes that are not observable in a backscattering geometry in black .
the oblique arrows symbolize the decrease ( increase ) of the frequency of one given mode as @xmath0 increases . ] [ ! htb ]
.force constants per unit area of the low - frequency ( lsm , lbm ) modes in 1t and 2hc transition metal dichalcogenides , and graphite , extracted from a fit to a finite linear chain model ( see eq . ) .
[ cols="^,^,^",options="header " , ] the interlayer force constants in 1t-rex@xmath1 appear to be significantly smaller , yet on the same order of magnitude as in 2hc - tmd .
one may even notice that @xmath40 is larger in rex@xmath1 than in graphite @xcite .
as a result , albeit the modest variations of the high - frequency intralayer raman response@xcite and of the pl spectra @xcite on @xmath0 may suggest marginal interlayer coupling in rex@xmath1 , our direct investigation of interlayer raman modes establishes that @xmath0-layer rex@xmath1 must be regarded as a van der waals coupled system , with obvious fingerprints of interlayer interactions .
we now propose a simple symmetry analysis to explain why the frequency of the lsm and lbm from a given branch both decrease as @xmath0 augments , in contrast to the case of 2hc - tmd @xcite .
first , in keeping with recent studies in twisted tmd bilayers and van der waals heterostructures made of 2hc - tmd , we would like to point out that the observation of prominent lsm features in fig .
[ fig1 ] and [ fig2 ] suggests that there is a well - defined stacking order in our @xmath0-layer rese@xmath1 and res@xmath1 samples @xcite .
let us then assume that @xmath0-layer rese@xmath1 and res@xmath1 both have a 1t structure , which has no mirror symmetry plane but preserves an inversion center for any value of @xmath0 .
in contrast , @xmath0-layer 2hc compounds exhibit a mirror symmetry plane ( but no inversion center ) for odd @xmath0 and an inversion center ( but no mirror symmetry plane ) for even @xmath0 @xcite .
then , as illustrated in figure [ fig4 ] , the ( symmetric ) @xmath5 raman - active lsm and lbm in rex@xmath1 correspond the lowest frequency mode , third - lowest frequency mode , , leading to branches that soften as @xmath0 increases from @xmath41 for @xmath13 down to zero frequency . in contrast , for 2hc compounds , the raman - active lsm correspond to the highest frequency mode , third - highest frequency mode , , leading to branches that stiffen with increasing @xmath0 , from @xmath41 for @xmath13 up to @xmath42 in the bulk limit , whereas the raman - active lbm correspond to the lowest , third lowest, frequency mode , leading to similar trends as for rex@xmath1 .
[ [ angular - dependence - of - the - raman - response ] ] * _ angular dependence of the raman response _ * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + next , we address the angular dependence of the raman response in rex@xmath1 .
figure [ fig5 ] displays the evolution of the lsm ( _ i.e. _ , the @xmath31 modes ) and lbm ( _ i.e. _ , the @xmath20 mode ) in bilayer rese@xmath1 , recorded as a function of the angle @xmath43 between the linearly polarized laser electric field and the re chains at @xmath17 in the xx configuration .
following the empirical method proposed by chenet _
et al . _ for res@xmath1
, the orientation of the re chains in rese@xmath1 has been preliminarily determined by inspecting the angular dependence of the integrated intensity of the mode v ( @xmath44 ) introduced above @xcite ( see fig .
[ fig1 ] ) , allowing us to define the reference angle @xmath45 for a laser polarization parallel to the re chains ( see solid grey lines in fig . [ fig5]- [ fig7 ] and supporting information ) . as @xmath43 varies , one clearly observes large , @xmath46-periodic variations of the lsm and lbm integrated intensities ( denoted @xmath47 and @xmath48 , respectively ) , which are direct fingerprints of in - plane anisotropy .
the frequencies and linewidth of the @xmath31 and @xmath37-mode features are found to be independent on @xmath43 .
remarkably , as shown in fig . [ fig5](c ) , we notice that at @xmath49 , @xmath50 reaches a maximum at @xmath51 and exhibits approximately a @xmath52 dependence , whereas @xmath53 displays a more complex angular dependence with a maximum at @xmath54 and secondary maxima near @xmath55 , close enough to the direction perpendicular to the re chains .
finally , the polar plot of @xmath48 also displays prominent maxima at @xmath56 , as well as secondary transverse maxima at @xmath57 that also match the maxima of @xmath53 .
interestingly , as it has also been reported in few - layer ( @xmath58-thick ) black phosphorus films @xcite , the observed angular patterns appear to be strongly dependent of the laser photon energy . indeed , as shown in fig .
[ fig6 ] and further outlined in the supporting information , the same measurements performed at a slightly higher photon energy of @xmath7 on the same rese@xmath1 bilayer reveal completely different angular dependencies for @xmath47 , @xmath48 , @xmath59 and , importantly , no raman intensity maximum at @xmath45 for any of these modes .
remarkably , we find that the total integrated intensity of the @xmath60 mode ( @xmath61 ) and @xmath59 display similar patterns at both photon energy . remarkably , the modes that exhibit _ two - lobe _ angular patterns at @xmath17 ( _ i.e. _ , the @xmath60 mode and mode v ) , exhibit _ four - lobe _ patterns at @xmath7 and _ vice versa _ ( cf . the @xmath20 mode ) .
furthermore , as shown in fig .
[ fig7 ] , we performed a similar study on a res@xmath1 bilayer with known crystal orientation and found angular patterns for the raman intensities that bear no particular similarity with their counterparts in rese@xmath1 .
the angular patterns of @xmath62 and @xmath63 depend on the incoming photon energy , whereas similar angular patterns are observed for @xmath59 at @xmath17 and @xmath7 , with a maxima in the vicinity of @xmath64 .
one may also notice that @xmath65 still exhibits a maximum near @xmath66 at @xmath17 , but not at @xmath7 , where a maximum is found near @xmath67 .
first , our observations provide useful empirical guidelines to determine the crystal orientation . in practice ,
the re chains are most currently found to coincide with one of the cleaved edges of the @xmath0-layer rex@xmath1 crystals , providing a first hint of the crystal orientation .
then , at @xmath68 , @xmath65 reaches a maximum near @xmath45 both for rese@xmath1 ad res@xmath1 .
nevertheless , we should stress that the present experimental results do not make it possible to unambiguously correlate the @xmath69 mode - features with the direction of the in - plane rigid layer displacements with respect to the re chains .
this observation invites further theoretical analysis of the normal displacements associated with the interlayer phonon modes .
second , the diverse angular patterns shown in fig .
[ fig5]-[fig7 ] and in the supporting information shed light on the extremely complex raman response of rex@xmath1 . in order to model the angular patterns described above
, one has to consider the raman tensor in rex@xmath1 ( see ref . and the supporting information ) . since these materials efficiently absorb visible light @xcite , the raman tensor elements are complex numbers @xcite , which are dependent on the incoming laser photon energy . using the general form for a symmetric , second rank tensor , we are able to fit the experimental angular dependence of the raman intensities ( see fig .
[ fig5]-[fig7 ] and supporting information ) . moreover , our observation of distinct angular patterns on rese@xmath1 and res@xmath1 bilayers at two close - lying photon energies points towards resonance effects .
we note that a recent theoretical study by zhong _
et al . _ predicts a complex manifold of strongly anisotropic excitons in monolayer rese@xmath1 and res@xmath1 , with several associated optical resonances in the spectral window ( 1.96 ev 2.33 ev ) investigated here @xcite .
a further theoretical analysis of anisotropic exciton - phonon coupling together with an experimental study of the anisotropic dielectric function and resonance raman profile in @xmath0-layer rex@xmath1 is mandatory to rationalize the anomalous photon energy dependence of the anisotropic raman response unveiled in fig .
[ fig6 ] and [ fig7 ] .
third , we would like to emphasize that due to the photon energy dependence of the anisotropic raman response in @xmath0-layer rex@xmath1 , great care must be taken while correlating the angular dependence of the raman response of a @xmath0-layer rex@xmath1 sample to its crystal orientation .
in particular , the behavior of @xmath44 shown in fig . [ fig6](c ) restricts the applicability of the empirical analysis of chenet _ et al . _
@xcite , which remains valid for res@xmath1 at @xmath17 and @xmath7 and for rese@xmath1 at @xmath17 but not at @xmath7 .
we have reported the observation of prominent , interlayer raman modes in distorted octahedral ( 1t ) rese@xmath1 and res@xmath1 .
the splitting of the interlayer shear modes and the strong angular dependence of the raman intensity of the interlayer shear and breathing modes are a direct fingerprint of the pronounced in - plane anisotropy in these materials .
noteworthy , the observation of similar intralayer raman spectra and fan diagrams of the interlayer modes in @xmath0-layer rese@xmath1 and res@xmath1 , suggests that the two materials studied here are ( quasi-)isostructural , bringing a key contribution to a longstanding debate summarized in ref . .
our findings will be useful to improve our understanding of other anisotropic layered materials such as black phosphorus @xcite .
finally , @xmath0-layer rese@xmath1 and res@xmath1 appear as a stimulating playground for further investigations of symmetry - dependent resonant exciton - phonon coupling @xcite , whereas from a more applied perspective the anisotropic optical response @xcite and electron transport properties @xcite of these systems hold promise for novel optoelectronic devices .
_ note : while finalizing this manuscript and during the review process , we became aware of related raman studies on res@xmath1 @xcite .
although no splitting of the interlayer shear mode is resolved in ref . , ref .
reports a shear mode splitting of @xmath70 in bilayer res@xmath1 .
the discrepancy between the @xmath71 splittings observed here and ref
. is likely due to different stacking orders with nearly equal stability , _ i.e. _ , polytypism @xcite . here
, the stacking order in our @xmath0-layer res@xmath1 crystals would correspond to the type 3 stacking studied in ref . .
noteworthy , this stacking displays an inversion center for any number of layers . _
@xmath0-layer rex@xmath1 crystals were prepared by mechanical exfoliation of commercially available bulk crystals ( 2d semiconductors ) onto si wafers covered with a 90-nm or 285-nm - thick sio@xmath1 epilayer .
our rex@xmath1 samples remained stable for several months in ambient conditions .
the number of layers was first estimated using optical contrast and atomic force microscopy , and unambiguously confirmed by the present raman study .
micro - raman scattering measurements were carried out in ambient conditions , in a backscattering geometry using a monochromator equipped with a 2400 grooves / mm holographic grating , coupled to a two - dimensional liquid nitrogen cooled charge - coupled device ( ccd ) array .
the samples were excited using a linearly polarized laser at two different photon energies ( @xmath7 and @xmath17 ) .
these energies are significantly above than the band - edge direct excitons that are near 1.3 ev and 1.5 ev in bulk rese@xmath1 and res@xmath1 , respectively @xcite .
a moderate laser intensity below @xmath72 was used for all measurements .
spectral resolutions of @xmath73 and @xmath74 were obtained at @xmath7 and @xmath17 , respectively . in order to attain the low - frequency range , a combination of one narrow bandpass filter and two narrow notch filters ( optigrate ) was used . after optimization , raman features at frequencies as low as 4.5 @xmath11 could be measured . in order to uncover the anisotropic raman response of rex@xmath1 layers , the samples were rotated with respect to the fixed linear polarization of the incoming laser field . in practice , each sample was mounted onto a rotating stage attached onto a xyz piezo - scanner and placed as close as possible to the rotational axis .
the region of interest was then parked under the fixed , diffraction limited laser spot using the piezo stage .
after each rotation step , the position of region of interest was re - adjusted under the laser spot ( with an accuracy better than 100 nm ) using the piezo - stage .
polarization - dependent raman studies were performed using an analyzer placed before the entrance of our spectrometer , followed by a polarization scrambler in order to avoid any artifacts due to polarization sensitive optics in our spectrometer .
thereafter , we will only consider parallel ( xx ) polarizations of the incoming and raman - scattered photons .
data recorded in the perpendicular ( xy ) configuration are shown in the supporting information . finally , the measured raman features are fit to voigt profiles , taking into account our spectral resolution .
we thank f. ben romdhane and f. banhart for recording transmission electron microscopy images of a bulk rese@xmath1 crystal , f. fernique for early contributions to this study , and the stnano clean room staff for experimental help .
we acknowledge financial support from the agence nationale de la recherche ( under grant quandogra 12 js10 - 001 - 01 and h2dh ) , from the cnrs and universit de strasbourg , and from the labex nie ( under grant who ) .
48ifxundefined [ 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 10.1080/00018736900101307 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.136805 [ * * , ( ) ] link:\doibase 10.1021/nl903868w [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.115438 [ * * , ( ) ] link:\doibase 10.1364/oe.21.004908 [ * * , ( ) ] http://pubs.acs.org/doi/abs/10.1021/nn305275h [ * * , ( ) ] link:\doibase 10.1021/nl502557 g [ * * , ( ) ] link:\doibase 10.1021/nl5045007 [ * * , ( ) ] link:\doibase 10.1038/nnano.2012.193 [ * * , ( ) ] link:\doibase 10.1038/nphys2942 [ * * , ( ) ] link:\doibase 10.1038/ncomms4252 [ * * , ( ) ] link:\doibase 10.1021/nn505354a [ * * , ( ) ] link:\doibase 10.1002/adfm.201500969 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.054110 [ * * , ( ) ] link:\doibase 10.1038/ncomms7991 [ * * , ( ) ] link:\doibase 10.1021/acsnano.5b04851 [ * * , ( ) ] link:\doibase 10.1039/c4nr01741b [ * * , ( ) ] link:\doibase 10.1021/nn5053926 [ * * , ( ) ] link:\doibase 10.1007/s12274 - 015 - 0865 - 0 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.365357 [ * * , ( ) ] link:\doibase 10.1103/physrevb.58.16130 [ * * , ( ) ] http://stacks.iop.org/0953-8984/16/i=32/a=026 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0038-1098(99)00240-9 [ * * , ( ) ] link:\doibase 10.1021/acs.nanolett.5b00910 [ * * , ( ) ] link:\doibase 10.1002/pssr.201510412 [ ( ) , 10.1002/pssr.201510412 ] link:\doibase 10.1063/1.4751266 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.241301 [ * * , ( ) ] link:\doibase 10.1021/nl304169w [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.115413 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.195316 [ * * , ( ) ] link:\doibase 10.1021/acs.nanolett.5b02683 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.094303 [ * * , ( ) ] http://www.nature.com/nmat/journal/v11/n4/abs/nmat3245.html [ * * , ( ) ] link:\doibase 10.1021/nl501678j [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.165403 [ * * , ( ) ] link:\doibase 10.1021/acsnano.5b07807 [ * * , ( ) ] , link:\doibase 10.1039/c5nr04349b [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.115438 [ * * , ( ) ] link:\doibase 10.1021/acsphotonics.5b00486 [ * * ( ) , 10.1021/acsphotonics.5b00486 ] link:\doibase 10.1021/acsnano.5b00698 [ * * , ( ) ] link:\doibase 10.1021/acs.nanolett.5b01117 [ * * , ( ) ] link:\doibase 10.1002/anie.201410108 [ * * , ( ) ] link:\doibase 10.1021/acs.nanolett.5b00775 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.136403 [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.235409 [ * * , ( ) ] link:\doibase 10.1039/c4nr05785f [ * * , ( ) ] link:\doibase 10.1021/acs.nanolett.5b04925 [ * * , ( ) ] , @noop * * ( ) * supplementary information * this document contains the following sections : * determination of the crystal orientation in rese@xmath1 * additional raman data on rese@xmath1 * additional raman data on res@xmath1 * angular dependence of the raman intensities
to verify whether the empirical approach introduced by chenet _ et al . _ on res@xmath1
@xcite is still holding for rese@xmath1 we recorded the angular dependence of the intensity of the mode v at @xmath75 in a few - layer rese@xmath1 crystal .
prior to the raman measurements , the rese@xmath1 crystal was transferred onto a transmission electron microscopy ( tem ) grid ( see fig .
[ figs1](a ) , with one freestanding region that could be imaged using by tem ( see fig .
[ figs1](b),(c ) ) , allowing a direct observation of the re chains .
raman measurements were then performed on the same region at @xmath76 . as shown in fig.[figs1](d),(e ) , the integrated intensity of the mode v is indeed maximal when the laser polarization is parallel to the re chains .
this empirical approach was then used , only at @xmath76 , to confirm the orientation of the re chains in the @xmath0-layer rese@xmath1 samples shown in the main manuscript . for res@xmath1 samples , we relied on the results by chenet _
et al . _ , obtained at @xmath77 and found that similar angular dependencies of the mode v integrated intensities were obtained at @xmath77 and at @xmath76 .
crystal placed onto a tem grid .
( c ) high - resolution tem image of the region highlighted with the red arrow in ( b ) , showing the re chains .
( d ) angular dependence of the raman mode v recorded as a function of the angle @xmath43 between the linearly polarized laser field and the re chains on the region highlighted in ( b ) in the parallel ( xx ) polarization configuration . the raw spectra ( spheres ) are fit to voigt profiles ( red lines ) and are vertically offset for clarity .
( e ) polar plot of the raman mode v integrated intensity .
the solid lines in ( e ) is a fit based on eq . .
@xmath66 corresponds to a laser polarization parallel to the re chains . ]
to @xmath2 layers ( a ) rese@xmath78 and ( b ) res@xmath1 , recorded at @xmath17 in the xx configuration , showing the lsm and the lbm features .
the grey bar in ( b ) masks a residual contribution from the laser beam .
the mode labels correspond to the notations in fig . 3 in the main manuscript .
] bilayer recorded as a function of the angle @xmath43 between the linearly polarized laser field and the re chains at @xmath17 in the perpendicular ( xy ) configuration .
the raw spectra ( spheres ) are fit to voigt profiles ( red lines ) and are vertically offset for clarity .
( b ) optical image of the rese@xmath1 bilayer showing the direction of the rhenium chains . at @xmath66 ,
the laser polarization is parallel to the re chains .
( c ) polar plot of the integrated raman intensity of the @xmath31 and @xmath20 modes .
the solid lines in ( c ) are fits to the experimental data based on eq . .
] bilayer recorded as a function of the angle @xmath43 between the linearly polarized laser field and the re chains at @xmath7 in the parallel ( xx ) configuration .
the raw spectra ( spheres ) are fit to voigt profiles ( red lines ) and are vertically offset for clarity .
( b ) optical image of the rese@xmath1 bilayer showing the direction of the rhenium chains .
the laser polarization for the reference at @xmath66 is parallel to the re chains .
( c ) polar plot of the integrated raman intensity of the @xmath31 and @xmath20 modes .
the solid lines in ( c ) are fits to the experimental data based on eq . .
] bilayer recorded as a function of the angle @xmath43 between the linearly polarized laser field and the re chains at @xmath7 in the parallel ( xx ) configuration . at @xmath66 , the laser polarization is parallel to the re chains .
the frequency of the lsm is minimal when the laser polarization is parallel or perpendicular to the re chains and reaches a maximum when the laser polarization makes an angle @xmath79 with the re chains .
this observation provides a useful criterion to identify the re chains . ]
bilayer recorded as a function of the angle @xmath43 between the linearly polarized laser field and the re chains at @xmath7 in the perpendicular ( xy ) configuration .
the raw spectra ( spheres ) are fit to voigt profiles ( red lines ) and are vertically offset for clarity .
( b ) optical image of the rese@xmath1 bilayer showing the direction of the rhenium chains . at @xmath66 ,
the laser polarization is parallel to the re chains .
( c ) polar plot of the integrated raman intensity of the @xmath31 and @xmath20 modes .
the solid lines in ( c ) are fits to the experimental data based on eq . .
] trilayer recorded as a function of the angle @xmath43 between the linearly polarized laser field and the re chains at @xmath17 in the parallel ( xx ) configuration .
the raw spectra ( spheres ) are fit to voigt profiles ( red lines ) and are vertically offset for clarity .
( b ) optical image of the rese@xmath1 trilayer showing the direction of the rhenium chains . at @xmath66 ,
the laser polarization is parallel to the re chains .
( c ) polar plot of the integrated raman intensity of the @xmath31 and @xmath20 modes .
the solid lines in ( c ) are fits to the experimental data based on eq . .
] bilayer recorded as a function of the angle @xmath43 between the linearly polarized laser field and the re chains at @xmath17 in the parallel ( xx ) configuration .
the raw spectra ( spheres ) are fit to voigt profiles ( red lines ) and are vertically offset for clarity .
( b ) optical image of the res@xmath1 bilayer showing the direction of the rhenium chains . at @xmath66 ,
the laser polarization is parallel to the re chains .
( c ) polar plot of the integrated raman intensity of the @xmath31 and @xmath20 modes .
the solid lines in ( c ) are fits to the experimental data based on eq . .
] bilayer recorded as a function of the angle @xmath43 between the linearly polarized laser field and the re chains at @xmath7 in the parallel ( xx ) configuration .
the raw spectra ( spheres ) are fit to voigt profiles ( red lines ) and are vertically offset for clarity .
( b ) optical image of the res@xmath1 bilayer showing the direction of the rhenium chains . at @xmath66 ,
the laser polarization is parallel to the re chains .
( c ) polar plot of the integrated raman intensity of the @xmath31 and @xmath20 modes .
the solid lines in ( c ) are fits to the experimental data based on eq . .
since the optical absorption spectrum of rex@xmath1 is dominated by strongly anisotropic excitons @xcite , the raman response of these systems is expected to exhibit a complex angular dependence @xcite . for an incoming and scattered polarizations ( in the xy plane ) denoted by @xmath80 and @xmath81
, the raman intensity writes with @xmath43 , the angle between @xmath80 and the rhenium chains , @xmath83 the angle between @xmath81 and @xmath80 and the raman tensor @xmath84 with @xmath85 ( @xmath86 ) , @xmath87 ( @xmath88 ) , @xmath89 ( @xmath90 ) the amplitude and phase of the raman tensor elements . as introduced in ref .
@xcite , the finite phases stem the anisotropic absorption of rex@xmath1 . in our experiments , we can not fit all the experimentally observed angular patters in fig . 5 - 7 using real tensor elements .
in practice and are used to fit the experimental angular dependences of the raman intensities .
the observation of a very distinct angular dependence of the raman intensities for two different laser photon energies is a direct proof that the raman tensor elements strongly depend on the laser photon energy . | we investigate the interlayer phonon modes in @xmath0-layer rhenium diselenide ( rese@xmath1 ) and rhenium disulfide ( res@xmath1 ) by means of ultralow - frequency micro - raman spectroscopy .
these transition metal dichalcogenides exhibit a stable distorted octahedral ( 1t ) phase with significant in - plane anisotropy , leading to sizable splitting of the ( in - plane ) layer shear modes .
the fan - diagrams associated with the measured frequencies of the interlayer shear modes and the ( out - of - plane ) interlayer breathing modes are perfectly described by a finite linear chain model and allow the determination of the interlayer force constants .
nearly identical values are found for rese@xmath1 and res@xmath1 .
the latter are appreciably smaller than but on the same order of magnitude as the interlayer force constants reported in graphite and in trigonal prismatic ( 2hc ) transition metal dichalcogenides ( such as mos@xmath1 , mose@xmath1 , mote@xmath1 , ws@xmath1 , wse@xmath1 ) , demonstrating the importance of van der waals interactions in @xmath0-layer rese@xmath1 and res@xmath1 .
in - plane anisotropy results in a complex angular dependence of the intensity of all raman modes , which can be empirically utilized to determine the crystal orientation .
however , we also demonstrate that the angular dependence of the raman response drastically depends on the incoming photon energy , shedding light on the importance of resonant exciton - phonon coupling in rese@xmath1 and res@xmath1 .
* keywords * : transition metal dichalcogenides , two - dimensional materials , rese@xmath1 , res@xmath1 , raman spectroscopy , interlayer interactions , anisotropy .
although well - documented in their bulk forms for decades , transition metal dichalcogenides ( tmd ) @xcite have recently attracted considerable attention since the observation of intense photoluminescence ( pl ) from monolayers of molybdenum disulfide ( mos@xmath1 ) @xcite .
so far , the majority of works on tmds has focused on trigonal prismatic compounds , with aba / bab stacking ( 2hc phase @xcite ) such as mos@xmath1 , mose@xmath1 , mote@xmath1 , ws@xmath1 , wse@xmath1 @xcite .
the latter exhibit an indirect - to - direct bandgap transition when their thickness is reduced down to the monolayer limit and hold promise for opto - electronic @xcite and valleytronic applications @xcite .
in contrast to these high - symmetry structures , other semiconducting tmd , such as rhenium disulfide ( res@xmath1 ) @xcite and rhenium diselenide ( rese@xmath1 ) @xcite ( thereafter denoted rex@xmath1 ) exhibit a distorted octahedral phase ( denoted 1t ) with much lower symmetry and significant in - plane anisotropy @xcite .
@xmath0-layer rex@xmath1 have been shown to remain indirect bandgap semiconductors , with faint pl irrespective of the number of layers @xmath0 @xcite . in spite of these poor emission properties , rex@xmath1 offer exciting perspectives for anisotropic electronic devices @xcite .
however , such developments require unambiguous fingerprints of in - plane anisotropy as well as fast and accurate methods to determine the crystal orientation , the strength of interlayer interactions , and the stacking order .
recently , raman scattering studies of rex@xmath1 have revealed that the raman response of these anisotropic systems depends subtly on the relative orientation of the crystal axis and of the exciting laser field and also on the angle between the polarizations of the incoming and raman scattered photons @xcite .
nevertheless , most studies on rex@xmath1 have focused on the complex manifold of intralayer raman modes , which exhibits minor variations as a function of the number of layers @xcite .
in contrast ,
the low - frequency interlayer shear modes ( lsm ) and breathing modes ( lbm ) in tmd make it possible to unambiguously determine the number of layers and evaluate the strength of van der waals interlayer coupling @xcite . in the case of @xmath0-layer rex@xmath1 , a splitting of the ( in - plane ) lsm
is expected @xcite and should provide an invaluable tool to investigate in - plane anisotropy .
however an experimental observation of this splitting is still lacking . in this article , we report a detailed raman study of @xmath0-layer rese@xmath1 and res@xmath1 .
we resolve a clear splitting of the ( in - plane ) lsm .
in contrast the ( out - of plane ) lbm do not exhibit any splitting , allowing an unambiguous distinction between lsm and lbm in the raman spectra .
we report the fan diagrams associated with all the raman active lsm and lbm up to @xmath2 .
the observed branches of lsm and lbm are very well described using a simple linear chain model .
noteworthy , in contrast to the case of 2hc - tmd , the frequency of both lbm _ and _ lsm seemingly decrease with increasing @xmath0 and we rationalize this observation using a symmetry argument .
finally , angle - dependent raman studies of the lsm and lbm in bilayer rex@xmath1 recorded at two different laser energies reveal complex patterns which , on the one hand , can empirically be used to identify the crystal orientation but , on the other hand , shed light on the importance of resonance raman effects in these strongly anisotropic layered materials . |
symplectic capacities are an important tool in the study of symplectic rigidity phenomena .
these were introduced and axiomatized by ekeland and hofer @xcite , though the first one was constructed previously by gromov @xcite .
many such capacities exist arguably , each phenomenon of symplectic rigidity gives rise to its own capacity .
a large number of examples are described in @xcite , and relationships between them , notably energy - capacity inequalities , lead to interesting connections between hamiltonian dynamics and symplectic topology .
very little is known about capacities defined relative to special submanifolds @xmath0 of a symplectic manifold , and , indeed , even in the lagrangian case there are many open questions .
barraud and cornea introduced the first relative capacities ( in the lagrangian case ) , the lagrangian gromov width and relative packing numbers @xcite .
upper bounds for the lagrangian gromov width of the clifford torus in @xmath1 were computed by biran and cornea@xcite .
buhovsky @xcite later showed lower bounds for the clifford torus .
schlenk s constructions @xcite also work in the relative case , and therefore compute the relative packing numbers for @xmath2 balls in @xmath3 . in @xcite ,
the second author defined a blow - up and blow - down procedure for lagrangian submanifolds , and used it to compute the relative gromov radius of a class of lagrangians that are fixed point sets of real , rank-@xmath4 symplectic manifolds . in @xcite and @xcite ,
zehmisch constructed a capacity of a manifold @xmath5 from embeddings of @xmath6-disk bundles over a lagrangian submanifold and subsequently related it to the geometry of the lagrangian .
finally , in @xcite , borman and mclean construct a spectral capacity for wrapped floer homology , and use it to study the relative gromov width of closed lagrangian submanifolds in liouville manifolds . in this paper
, we study a notion of a capacity relative to a coisotropic submanifold , which we call a coisotropic capacity . in a heuristic sense ,
if a symplectic capacity measures the width of a symplectic manifold , our coisotropic capacity similarly measures the symplectic size of a coisotropic embedding inside an ambient symplectic manifold . after discussing the general notions ,
we define a new capacity relative to a coistropic submaifold via hamiltonian dynamics on the ambient symplectic manifold in a manner analogous to the definition of the hofer - zehnder capacity @xcite .
our proof is obtained as a modification of their variational proof .
the main innovations in the proof are to consider a function space better suited to the relative problem , which also requires us to consider a different extension of the hamiltonian .
our extension introduces some new subtleties to the argument , which we emphasize in our discussion . as an application of the existence of our capacity ,
we show the following non - squeezing result for coisotropic balls and cylinders , which is the natural analogue of the gromov non - squeezing theorem @xcite in the coisotropic case .
we define the function @xmath7 to be the area of the region @xmath8 .
( our reasons for choosing @xmath9 instead of @xmath10 here will become clear in the theorem .
for now , note that @xmath11 \times \{0\ } = \mathcal{b } \cap ( { \mathbb{r}}\times \{0\})$ ] , and @xmath12 . )
we now have [ thm : relative non - squeezing ] consider @xmath13 , with @xmath14 the standard symplectic structure and let @xmath15 , @xmath16 , be the @xmath17 dimensional coisotropic subspace given by @xmath18 let @xmath19 denote the ball of radius @xmath4 centered at the origin in @xmath20 .
suppose @xmath21 satisfies @xmath22 , and let @xmath23 be the coisotropic ball @xmath24 where @xmath25 denotes the radius of the coisotropic ball .
let @xmath26 .
define @xmath27 and let @xmath28 . if there exists a symplectic embedding @xmath29 such that @xmath30 , then @xmath31 .
observe that @xmath32 has infinite gromov width , so this embedding is only obstructed by the coisotropic constraint . and of @xmath33 with the @xmath34
coordinate plane . ]
we construct a capacity for pairs @xmath35 consisting of a symplectic manifold @xmath36 and a properly embedded coisotropic submanifold @xmath37 , i.e. with @xmath38 ( or @xmath39 ) .
all of our symplectic manifolds will be assumed to be of the same dimension @xmath40 .
we begin with a few definitions .
let @xmath41 and @xmath42 be ( possibly open ) symplectic manifolds , and let @xmath43 and @xmath44 be properly embedded coisotropic submanifolds .
we call @xmath45 a _ relative symplectic embedding _ , or _ relative embedding _ , if @xmath46 is a symplectic embedding and @xmath47 .
[ defn : coisotropic objects ] recall that @xmath48 is an @xmath49-dimensional coisotropic linear subspace of @xmath20 .
we define @xmath50 as the @xmath40 dimensional symplectic ball of radius @xmath9 centered at @xmath51 and @xmath52 as the coisotropic ball of radius @xmath9 centered at @xmath53 : @xmath54 we likewise denote by @xmath55 the symplectic cylinder centered at @xmath56 , and by @xmath57 the coisotropic cylinder @xmath58 note that @xmath59 and @xmath57 are properly embedded in @xmath60 , @xmath61 , respectively , when @xmath62 , and @xmath63 .
a _ coisotropic capacity _ is a map @xmath64 which associates to every tuple @xmath65 consisting of a symplectic manifold @xmath5 and a properly embedded coisotropic submanifold @xmath37 a non - negative number that satisfies the following axioms : 1 .
[ axiom : monotonicity]monotonicity . if there exists a relative symplectic embedding @xmath66 for @xmath67 and @xmath68 of the same dimension , then @xmath69 2 .
[ axiom : conformality]conformality .
@xmath70 3 .
[ axiom : nontriviality]nontriviality .
denote the unit ball centered at the origin by @xmath71 , let @xmath72 denote the coisotropic ball @xmath73 , let @xmath74 be the symplectic cylinder , and let @xmath75 then @xmath76 .
when the symplectic form in @xmath65 is understood , we will abbreviate this to @xmath35 .
observe that the nontriviality axiom is not the strongest such axiom one might consider .
if one lets @xmath77 denote the hofer - zehnder capacity of the symplectic manifold @xmath78 , and @xmath79 is an arbitrary properly embedded coisotropic submanifold , then defining @xmath80 gives a coisotropic capacity by satisfying these axioms .
we therefore introduce a variant of the nontriviality axiom , where we say that @xmath81 is a _ strict _ coiostropic capacity if @xmath82 for @xmath83 .
we recall that @xmath84 is the region in @xmath20 given as the union of the disk with a half - infinite strip @xmath85 and @xmath86 .
observe that by considering embeddings of the form @xmath87 where @xmath88 is symplectic , it is easy to construct an embedding of @xmath89 to @xmath84 so that @xmath90 is mapped to @xmath91 . in particular then , a strict coisotropic capacity is a coisotropic capacity .
we will now construct a specific map @xmath92 , which we denote by @xmath93 , and will show that it satisfies the axioms for a strict coisotropic capacity .
our definition relies on the following sets of hamiltonians .
an autonomous hamiltonian @xmath94 is * simple * if 1 .
there exists a compact set @xmath95 ( depending on @xmath96 ) and a constant @xmath97 such that @xmath98 , @xmath99 , and @xmath100 2 .
there exists an open set @xmath101 ( depending on @xmath96 ) , with @xmath102 , and on which @xmath103 .
@xmath104 for all @xmath105 .
denote the set of simple hamiltonians by @xmath106 .
let @xmath5 be a symplectic manifold and let @xmath37 be a coisotropic submanifold .
let @xmath107 denote the characteristic foliation of @xmath79 , and let @xmath108 denote the hamiltonian vector field of the function @xmath109 .
suppose @xmath110 is a solution to @xmath111 , and let @xmath112 denote the leaf containing @xmath113 .
we call the value @xmath114 the _ return time _ of the orbit @xmath115 . if @xmath116 for all @xmath117 , then we say that @xmath118
. a function @xmath119 will be called _ admissible _ if all of the solutions of @xmath120 are either such that @xmath110 is constant for all @xmath121 , or the return time @xmath122 for the orbit @xmath115 .
we denote the collection of all admissible functions by @xmath123 .
we now define @xmath124 .
our main theorem in this section is the following .
the map @xmath93 is a coisotropic capacity .
[ thm : capacity ] we prove the monotonicity and conformality properties below , which proceed as in @xcite .
the proof of non - triviality will be completed in section [ sec : capacity proof ] .
the map @xmath93 satisfies the monotonicity axiom [ axiom : monotonicity ] .
[ lem : monotonicity ] let @xmath125 be a relative embedding .
we define the map @xmath126 by @xmath127 note that if @xmath128 is a compact set with @xmath129 and @xmath130 , then @xmath131 and @xmath132 .
@xmath133 by definition , and if @xmath134 is an open set on which @xmath135 , then , since @xmath136 is an embedding @xmath137 is an open set on which @xmath138 . also , by definition , @xmath139 , and therefore @xmath140 .
we now check that @xmath141 .
since @xmath136 is symplectic , @xmath142 and the flows are conjugate . in particular , since @xmath143 , if a flow @xmath144 for @xmath121 , then @xmath145 for @xmath146 as well , which completes the proof .
[ lem : conformality ] @xmath93 satisfies the conformality axiom define a map @xmath147 by @xmath148 and let @xmath149 denote @xmath150 .
note that @xmath136 is clearly injective , and @xmath151 , so the lemma follows if @xmath152 is a bijection .
we now remark that @xmath153 therefore , the hamiltonian flows for @xmath96 and @xmath149 have the same orbits with the same return times , which proves the lemma . in the next proposition
, we give a lower bound for @xmath154 @xmath155 .
[ prop : lower bound ] recall that @xmath156 , and suppose that @xmath63 . for @xmath157 , @xmath158 choose @xmath159 , and let @xmath160\to \mathbb{r}$ ] be a function with the following properties : @xmath161 , \\ & 0 \leq f'(t ) < \arccos \left(\frac{|a|}{\sqrt{t } } \right ) \quad & & \text { for } t >
|a|^2 + \epsilon,\\ & 0 <
f'(t ) \quad & & \text { for } t \in ( |a|^2 + 2\epsilon , 1 - \epsilon ) , \\ &
f'(t ) = 0 \quad & & \text { for } t \in [ 1-\epsilon , 1].\end{aligned}\ ] ] let @xmath162 be the hamiltonian defined by @xmath163 .
observe first of all that @xmath164 is simple by construction .
we will now show it is admissible . to this end , consider the associated hamiltonian ode given by @xmath165 where @xmath166 is the standard almost complex structure induced by the identification of @xmath20 with @xmath167 . since @xmath168
, we see that @xmath169 is constant along solutions of the equation .
thus , with @xmath170 we have for @xmath171 , @xmath172 thus , @xmath173 .
let @xmath174 be the return time to the coisotropic leaf of @xmath175 .
writing @xmath176 , the boundary condition for @xmath175 gives @xmath177 for @xmath178 to lie in the same coisotropic leaf as @xmath179 , we have : @xmath180 condition is satisfied if and only if either @xmath181 or @xmath182 .
condition is satisfied if and only if either @xmath183 or @xmath184 . finally , condition is easiest to verify in polar coordinates .
we eliminate the easy case of @xmath185 first : in this case , condition is satisfied if and only if @xmath186 or @xmath184 . now , if @xmath187 , we write @xmath188 .
let @xmath189 be the angle so that @xmath190 and @xmath191 .
in particular then , @xmath192 if @xmath193 and @xmath194 if @xmath195 .
condition now gives that @xmath196 , which holds if and only if @xmath197 or @xmath198 for some @xmath199 odd .
thus , if @xmath200 and for each @xmath201 , @xmath202 , and there is no @xmath203 $ ] where @xmath179 and @xmath178 are in the same leaf .
this is satisfied if @xmath204 observe now that by choosing @xmath159 sufficiently small , we may arrange for @xmath205 to be arbitrarily close to @xmath206 ( recall that @xmath207 . ) from this , we conclude @xmath208
the relative symplectic embedding of the ball @xmath210 combined with the monotonicity axiom ( lemma [ axiom : monotonicity ] ) and lemma [ prop : lower bound ] above , give a lower bound on the capacity @xmath211 . in this section
, we will show that this lower bound is also an upper bound .
recall that @xmath84 is the region in @xmath20 given as the union of the disk with a half - infinite strip @xmath85 and @xmath86 . in the following , we will write @xmath212 and @xmath213 . [
prop : upper bound ] the map @xmath93 verifies @xmath214 note that by the conformality and monotonicity properties , i.e. lemmas [ lem : monotonicity ] and [ lem : conformality ] above , both theorems [ thm : capacity ] and [ thm : relative non - squeezing ] are immediate corolaries
. the remainder of this section will prove proposition [ prop : upper bound ] .
let @xmath14 denote the standard symplectic form on @xmath20 , and recall that @xmath15 is the linear coisotropic subspace of @xmath215 consisting of the first @xmath17 coordinates , i.e. @xmath216 let @xmath217 and @xmath218 be the linear subspaces @xmath219 let @xmath220\right)$ ] denote the space of smooth maps @xmath221\to { \mathbb{r}}^{2n}$ ] such that @xmath222 for some isotropic leaf @xmath223 in the characteristic foliation @xmath224 of @xmath225 .
let @xmath226 be the standard inner product on @xmath20 , and define the functional @xmath227\right ) \to { \mathbb{r}}$ ] by @xmath228 note that any leaf @xmath223 in @xmath107 has the form @xmath229 , for @xmath230 . in order to study the critical points of @xmath231
, we will extend the definition of @xmath231 to the hilbert space of @xmath232 paths .
the hilbert space is constructed so the paths have boundary in @xmath15 , even though @xmath232 does not embed in @xmath233 , and thus a pointwise constraint can not be imposed .
the key observation we use is that @xmath15 is the fixed point locus of an involution on @xmath20 , which then induces an isometry on @xmath234 .
our path space is then an eigenspace of this isometry , though we also describe it explicitly .
we first show the following .
[ lem : fourier decomposition ] any element @xmath235\right)$ ] is given by @xmath236 where @xmath237 equivalently , @xmath238 with @xmath239 for odd @xmath199 and @xmath240 for even @xmath199 .
( i.e. @xmath241 with @xmath242 for all odd @xmath199 . )
we begin by identifying @xmath243 with @xmath244 , and we consider a smooth map @xmath245 \to { \mathbb{c}}$ ] such that @xmath246 .
we now extend this map to a piecewise smooth map @xmath247 by @xmath248 \\ \overline{\gamma(2 - 2 t ) } & t \in \left(\frac{1}{2},1\right ] , \end{cases}\ ] ] where the bar indicates complex conjugation .
note that @xmath249 is continuous by definition .
writing @xmath249 in terms of its fourier decomposition , we have @xmath250 however , since @xmath251 , and therefore @xmath252 which implies that @xmath253 , and therefore @xmath254 .
our original function @xmath110 is recovered by @xmath255 , where @xmath256 .
now consider a function @xmath245 \to { \mathbb{r}}^{2n}$ ] such that @xmath257 , where @xmath223 is a leaf of the characteristic foliation of @xmath15 .
write a point @xmath258 by @xmath259 , where @xmath260 , @xmath261 , for @xmath262 the standard complex structure on @xmath20 , and define @xmath263 by @xmath264 let @xmath15 denote the set of points @xmath265 in the special case of a lagrangian , i.e. for @xmath266 , we note that @xmath267 is a real structure for @xmath14 , i.e. @xmath268 . any leaf @xmath223 of @xmath107 is a set of the form @xmath269 for some fixed @xmath270
. we may write @xmath110 as a function @xmath271 , where each function @xmath272 \to { \mathbb{r}}^{2n}$ ] is a map @xmath273 for real functions @xmath274 \to{\mathbb{r}}$ ] . from the above , we see that if @xmath275 , then @xmath276 where @xmath277 for constants @xmath278 , @xmath279 a vector with @xmath4 in the @xmath280-th position and @xmath281s elsewhere . for @xmath282 , @xmath283 , and we have @xmath284 the conclusion of the lemma now follows immediately . note that if @xmath285,{\mathbb{r}}^{2n } ) \cap l^{1}$ ] and is of the form @xmath286 with @xmath287 as in [ eq : fourier decomposition ] above , then necessarily @xmath288 .
let @xmath289)$ ] be the hilbert space @xmath290,{\mathbb{r}}^{2n } ) \ , \bigg\vert \ , & \gamma = \sum_{k\in { \mathbb{z } } } a_k e^{k\pi j t } + \sum_{k\in 2{\mathbb{z } } } b_{k } e^{k\pi j t},\\ & \ , a_k\in v_0 , \,b_{k } \in v_1,\\ & \sum_{k\in { \mathbb{z } } } \vert a_k \vert^2 + \vert b_k \vert^2
< \infty \bigg \ } \end{split}\ ] ] with inner product @xmath291 define @xmath292\right)$ ] to be the space @xmath293 ) = \left\lbrace x\in l^2_{n , k } \left([0,1],{\mathbb{r}}^{2n}\right)\ : \bigg\vert \ ; \sum_{k\in { \mathbb{z } } } |k|^{2s}|z_k|^2 < \infty \right\rbrace\ ] ] where @xmath294 for odd @xmath199 and @xmath240 for even @xmath199 .
in the following lemmas , we collect several standard results from @xcite concerning the bilinear form @xmath10 and the spaces @xmath295 .
the proofs are identical to those in @xcite , replacing the spaces considered there with the corresponding spaces in our setting . for the convenience of the reader
, we have tried to keep our notation compatible with the notation of @xcite*sections 3.3 , 3.4 , though our @xmath296 is denoted by @xmath297 in their book .
that is , @xmath298 $ ] .
some of the more immediate results are stated without proof .
[ lem : hilbert space and compact embedding ] for each @xmath299 , @xmath300)$ ] is a hilbert space with the inner product @xmath301 , then the inclusion of @xmath300)$ ] into @xmath302)$ ] is compact .
recall that @xmath303 is a hilbert space .
the involution on @xmath20 given by @xmath304 induces an isometry on @xmath303 by acting on each fourier coefficient .
observe now that @xmath300)$ ] can be identified with the @xmath305 eigenspace of this operator , and thus identifies @xmath306 as a closed subspace of a hilbert space .
the compactness of the inclusion follows by considering the finite rank truncation operators @xmath307 let @xmath308 denote the inclusion @xmath309 .
then , in the operator norm for @xmath310 , @xmath311 , and thus the inclusion is the uniform limit of finite rank operators , and is thus compact .
let @xmath312 .
if @xmath313 ) \to h^{t}_{n , k}([0,1])$ ] is the inclusion operator , then the hilbert space adjoint @xmath314 ) \to h^{s}_{n , k}([0,1])$ ] is compact .
[ lem : smoothness ] if @xmath315)$ ] for @xmath316 , where @xmath9 is an integer , then @xmath317)$ ] .
@xmath318 , and @xmath319 .
[ def : projections ] the hilbert space @xmath320 admits a decomposition into negative , zero and positive fourier frequencies : @xmath321 let @xmath322 and @xmath323 denote the orthogonal projections onto each of these subspaces .
given a simple hamiltonian @xmath324 with @xmath325 , we will analyze an associated hamiltonian @xmath326 , and find a solution of @xmath327 which is also a non - trivial solution of @xmath328 . in the following ,
we construct the hamiltonian @xmath329 .
we consider @xmath330 fixed and the simple hamiltonian @xmath96 with @xmath325 fixed .
we now set some notation .
1 . @xmath331 , 2 .
3 . let @xmath333 be the projection orthogonal to the last coordinate , i.e. @xmath334 .
4 . let @xmath335 be the quadratic function @xmath336 and let @xmath337 where @xmath79 is chosen sufficiently large so that the support of @xmath338 @xmath339 + for notational convenience , let @xmath340 and @xmath341 so that @xmath342 .
now , given a small @xmath343 such that @xmath344 , we define @xmath345 to be a function such that @xmath346 we define the extended hamiltonian @xmath329 by @xmath347 in the next lemma , we give a criterion to determine which orbits of the hamiltonian @xmath329 are actually orbits of @xmath96 .
[ lem : real solution ] suppose @xmath348 $ ] is a solution of @xmath349 such that @xmath350 .
if @xmath351 , then @xmath352 is non - constant and @xmath352 is an orbit of @xmath96 . first , note that if @xmath353 is constant , then @xmath354 , since @xmath355 . to show the orbit of @xmath356 is an orbit of @xmath96 , we will show that @xmath357 at each point of the orbit .
denote by @xmath358 the strip @xmath359 $ ] . by definition of @xmath329 , on @xmath360 ,
the @xmath361 component of the hamiltonian vector field @xmath362 is zero , and the @xmath363 component of @xmath362 is constant in this region .
therefore , if @xmath364 for some @xmath365 , then the @xmath363 coordinate of @xmath352 , which we denote @xmath366 , will satisfy @xmath367 it follows that the only orbits @xmath352 with @xmath368 are those with @xmath369 for all @xmath370 $ ] .
note now that by construction , the region @xmath371 is flow invariant .
we consider then a leaf - wise chord @xmath352 with @xmath372 for some @xmath146 and thus for all @xmath146 .
for such a solution , @xmath353 satisfies @xmath373 and in particular , @xmath374 is an integral of motion .
it follows that @xmath375 for all @xmath370 $ ] , and therefore , we have @xmath376 which completes the proof .
[ def : aa ] for @xmath377)$ ] , we define @xmath378 we show the following simple lemma . [ lem : pseudo - orthogonality ] for any @xmath379 , @xmath380 , @xmath381 first , note that , since @xmath382 , @xmath383 and therefore @xmath384 [ lem : a ] for @xmath385)$ ] , @xmath386 where @xmath387 first , recall that , by lemma [ lem : fourier decomposition ] , that for @xmath388 , the fourier expansions @xmath389 and @xmath390 have that @xmath391 for odd @xmath199 and @xmath392 for even @xmath199 . substituting equations [ eq : phipsi ] into the expression for @xmath10 and using lemma [ lem : pseudo - orthogonality ] ,
we get @xmath393 [ def : a ] given @xmath394 , we define @xmath395 by equation [ eq : a ] , and @xmath396 .
lemma [ lem : a ] gives that definitions [ def : a ] and [ def : aa ] are consistent , i.e. they coincide for smooth paths , @xmath397 .
the function @xmath398 given by @xmath399 is therefore differentiable with derivative @xmath400 and
therefore the gradient @xmath401 is @xmath402 for @xmath403 , consider the expression @xmath404 since , by construction , @xmath405 for @xmath406 large , we have that @xmath407 may be extended to @xmath408 , and therefore also on @xmath409 .
the following results follow immediately from the proofs in @xcite .
[ lem : b ] the map @xmath410 is differentiable .
its gradient is continuous and maps bounded sets into relatively compact sets .
moreover , @xmath411 and @xmath412 for all @xmath413 .
we now see that the functional @xmath414 ) \to { \mathbb{r}}$ ] given by @xmath415 is well - defined . furthermore , since @xmath416,{\mathbb{r}}^{2n})$ ] and @xmath10 and @xmath407 are differentiable , @xmath417 is differentiable with gradient @xmath418 the results below summarize some of its properties that we will use in the following sections .
the proofs follow those given in @xcite .
[ lem : c1 ] assume @xmath419 is a critical point of @xmath417 , i.e. @xmath420 . then @xmath353 is in @xmath421)$ ] . if , in addition , @xmath422 for all @xmath423 , then @xmath424)$ ] .
the proof given in hofer and zehnder @xcite also applies in this case .
that is , we write @xmath353 and @xmath425 by their fourier series , we have @xmath426 since @xmath427 , this implies that @xmath428 substituting the fourier series of @xmath353 and @xmath429 into this expression , we obtain @xmath430 therefore @xmath431 and @xmath432 we conclude that @xmath433)$ ] , and therefore @xmath434)$ ] by lemma [ lem : smoothness ] .
it follows that @xmath435)$ ] , so @xmath436 however , it follows from the fourier expansions that @xmath437 , and therefore @xmath438)$ ] and solves @xmath439 if @xmath440 for all @xmath146 , then @xmath441)$ ] , so @xmath442)$ ] . repeating this ,
the second part of the lemma follows .
[ lem : palais - smale ] @xmath417 satisfies the palais - smale condition .
we recall that , for @xmath417 to satisfy the palais - smale condition , we must have that , for every sequence @xmath443 with @xmath444 , there exists a convergent subsequence .
if @xmath445 is bounded , then this follows from the compactness of @xmath446 and of @xmath447 .
we now assume that the sequence of norms @xmath445 is unbounded .
consider the rescaled paths @xmath448 , so that @xmath449 .
now , by assumption , @xmath450 now note that there exists an @xmath36 such that @xmath451 for all @xmath452 .
it follows that the sequence @xmath453 is bounded in @xmath408 .
since @xmath454 is compact , @xmath455 is relatively compact , and since @xmath456 is bounded in @xmath20 , the sequence @xmath457 is relatively compact in @xmath296 .
let @xmath458 be as in the definition of @xmath329 in equation [ def : h extension ] .
define @xmath459 after taking a subsequence we may assume that @xmath460 in @xmath296 and therefore @xmath460 in @xmath408 .
we now have @xmath461 since @xmath462 for all @xmath463 , and since @xmath464 defines a continuous operator on @xmath408 such that for @xmath465 , @xmath466 we may conclude that @xmath467 and therefore , @xmath468 it follows from this convergence that @xmath469 satisfies the following system of equations in @xmath296 @xmath470 one may verify as in lemma [ lem : c1 ] , that @xmath471,{\mathbb{r}}^{2n})$ ] and also satisfies the hamiltonian equation @xmath472 by construction of @xmath473 , however , there are no non - trivial solutions of .
this , however , contradicts the assumption that @xmath474 , and we conclude that the sequence @xmath475 must be bounded , proving the lemma .
[ lem : global flow ] the equation @xmath476 defines a unique global flow @xmath477 .
this follows immediately from the global lipschitz continuity of @xmath478 as a vector field on @xmath296 .
[ lem : flowrepresentation ] the flow of the ode @xmath479 has the following form @xmath480 where @xmath481 is continuous and maps bounded sets into precompact sets and @xmath482 , @xmath483 and @xmath484 .
the proof of this lemma follows exactly the proof in hofer and zehnder @xcite .
the key point is that if we explicitly define @xmath485 by the formula @xmath486 we may verify directly that this has the required properties .
we will now complete the proof of proposition [ prop : upper bound ] . to do this
, we will prove the following : [ thm : upper bound ] if @xmath96 is a simple hamiltonian on @xmath487 and @xmath325 , then there exists an orbit of the system @xmath488 with return time @xmath489 and @xmath490
. the remainder of this section will prove the theorem .
the proof follows closely the proofs in @xcite , though introduces some new subtleties .
first , we produce two `` half - infinite '' dimensional subsets of @xmath296 , @xmath491 and @xmath492 and show that the action @xmath493 , the action @xmath494 , and @xmath495 .
we then use the minimax lemma over the family of sets @xmath496 to establish the existence of a critical point .
as established in lemma [ lem : real solution ] , it suffices to show this for @xmath329 as the resulting orbit will be an orbit of @xmath96 .
we begin with the following lemma .
[ lem : h zero at origin ] let @xmath497 .
then there exists a compactly supported hamiltonian diffeomorphism @xmath498 with @xmath499 such that @xmath500 and @xmath501 vanishes in a neighborhood of @xmath281 .
observe that in order for a hamiltonian @xmath485 to have a hamiltonian vector field whose flow preserves @xmath502 , the following derivatives @xmath503 must vanish along @xmath504 . by hypothesis ,
@xmath96 is admissible , so there exists an interior point @xmath505 in whose neighborhood @xmath96 vanishes . let @xmath506 be a neighborhood of the ray @xmath507 \}$ ] that is invariant under the involution c_n , k ( x_1 , , x_n , y_1 , , y_k , y_k+1 , , y_n ) ( x_1 , , x_n , y_1 ,
, y_k , -y_k+1 , , -y_n ) .
let @xmath201 be a @xmath508-invariant cut - off function , identically equal to @xmath4 on the neighbourhood @xmath506 and whose support is compactly contained in the interior of @xmath509 .
now define a hamiltonian by @xmath510 by @xmath511 let @xmath512 be its associated hamiltonian vector field and @xmath513 its time @xmath4 map .
observe first that the hamiltonian vector field @xmath514 for any @xmath515 , so @xmath516 and thus @xmath517 vanishes in a neighbourhood of @xmath281 .
a computation of @xmath518 for @xmath519 shows that the vector field is tangent to @xmath520 ( using both that @xmath505 and that @xmath201 is @xmath93-invariant ) .
[ prop : existence of an orbit ] there exists @xmath521 satisfying @xmath522 and @xmath523 .
the proof of proposition [ prop : existence of an orbit ] follows from the following lemmas .
we start by recalling the minimax lemma ( see @xcite , page 79 for a proof ) , which will play a key role .
let @xmath524 be a differentiable function on a hilbert space @xmath296 , i.e. @xmath525 , and let @xmath107 be a family of subsets @xmath526 .
we call the value @xmath527 the * minimax * of @xmath528 on the family @xmath107 .
suppose @xmath529 , where @xmath296 is a hilbert space , and that @xmath528 satisfies the following conditions : 1 .
@xmath528 is palais - smale , 2 .
@xmath530 defines a global flow @xmath531 on @xmath296 , 3 .
the family @xmath107 is positively invariant under the flow , i.e. , @xmath532 for all @xmath533 and all @xmath534 , 4 .
@xmath535 , then the real number @xmath536 is a critical value of @xmath528 , that is , there exists an element @xmath537 with @xmath538 and @xmath539 .
we now set some notation for the discussion which follows .
1 . @xmath540 2 .
@xmath541 3 .
@xmath542 4 .
@xmath543 [ lem : integral inequality ] for @xmath544 and @xmath545 @xmath546 note first that @xmath547 for any @xmath548,{\mathbb{r}}^{2n}_+)$ ] . also recall that @xmath342 where @xmath340 and @xmath341 .
if @xmath549 and @xmath550 are in orthogonal subspaces of @xmath551,{\mathbb{r}}^{2n})$]@xmath552 it follows that @xmath553 it then suffices to show that @xmath554 suppose @xmath555,{\mathbb{r}}^2)$ ] , and write @xmath556 , and @xmath557 as @xmath558 . by definition ,
@xmath559 we compute @xmath560 & s^2 \cos^2 ( \pi t ) + 2s \cos ( \pi t ) \xi(t ) \\ & + 2s \sin ( \pi t ) \eta(t ) - 2s \sin ( \pi t ) \eta(t ) \end{aligned}\\ = & s^2 \cos^2 ( \pi t ) + \langle 2s e^+ , u \rangle - 2s \sin ( \pi t ) \eta(t)\\ = & s^2 ( 1 - \sin^2 ( \pi t ) ) + \langle 2s e^+ , u \rangle - 2s \sin ( \pi t ) \eta(t)\\ = & s^2 + \langle 2s e^+ , u \rangle - s \sin ( \pi t ) \left ( s \sin ( \pi t ) + 2 \eta(t ) \right).\end{aligned}\ ] ] this implies that @xmath561 , @xmath562 for @xmath563 , note that @xmath564 . since @xmath299 and @xmath565 when @xmath566 $ ] , this implies that @xmath567 as well , and , in particular , that @xmath568 . since @xmath299 , @xmath569 , we see that @xmath570 .
note that @xmath571 and @xmath572 are in orthogonal subspaces of @xmath289)$ ] , and therefore @xmath573 .
we now have @xmath574 by lemma [ lem : fourier decomposition ] , @xmath575 and @xmath576 so @xmath577 . therefore ,
writing @xmath578 , we see that @xmath579 integrating , we have @xmath580 which completes the proof .
[ lem : linking 1 ] there exists a @xmath581 such that for @xmath582 , @xmath583 first , recall that @xmath584 . since @xmath585 and @xmath586
we have that @xmath587 .
we now need to examine @xmath231 on the boundary regions , where either @xmath588 or @xmath589 .
we note that by the construction of @xmath329 above , there exists a constant @xmath590 such that @xmath591 therefore , @xmath592 we now estimate @xmath593 for @xmath594 . note that by lemma [ lem : fourier decomposition ] , @xmath595 .
lemma [ lem : integral inequality ] gives @xmath596 recalling that @xmath597 and @xmath598 , it follows that @xmath599 and thus there is a @xmath600 , such that @xmath601 .
[ lem : linking 2 ] there exists @xmath602 and @xmath603 such that @xmath604 the proof proceeds exactly as in @xcite . as they observe , this lemma follows from the sobolev inequality @xmath605 .
since @xmath329 vanishes at the origin , taylor s theorem and the fact that @xmath329 is quadratic at infinity implies that we may find a constant @xmath606 such that @xmath607 , and therefore @xmath608 for @xmath609 with @xmath610 sufficiently small , the result follows .
[ lem : linking 3 ] @xmath611 , for all @xmath612 .
the proof of this lemma proceeds as in @xcite , which we summarize here .
we use the leray - schauder degree to show the existence of an element in @xmath613 .
( see deimling @xcite , theorem 8.2 or zeidler @xcite , chapter 12 , for properties of the leray - schauder degree . )
let @xmath223 denote the space @xmath614 . using the expression in lemman [ lem : flowrepresentation ]
, we will rewrite the condition @xmath615 in the form @xmath616 for the operator @xmath617 defined by @xmath618 we remark that @xmath619 is continuous and maps bounded sets into relatively compact sets by lemma [ lem : flowrepresentation ] .
we now recall that , since @xmath620 , @xmath621 , for some @xmath622 , so the system of equations [ eq : condition 4 ] is equivalent to @xmath623 let @xmath624 denote the identity operator . by the leray - schauder degree theory ,
for any fixed @xmath534 , equation [ eq : l - sh ] has a solution @xmath620 if @xmath625 since , by lemmas [ lem : linking 1 ] and [ lem : linking 2 ] , @xmath626 for @xmath612 , there is no solution of equation [ eq : l - sh ] on the boundary @xmath627 .
therefore , since the leray - schauder degree is homotopy invariant , we have @xmath628 we see that @xmath629 , so @xmath630 .
we define @xmath631 \times x \to x^{+}$ ] by @xmath632 and we claim that @xmath633 for @xmath634 . to see this , note first that if @xmath620 solves @xmath635 then @xmath636 , so @xmath637 . therefore , @xmath638 , so @xmath639 if @xmath640 , which is true by hypothesis .
furthermore , since @xmath640 , @xmath641 , so by homotopy , @xmath642 this completes the proof .
we now proceed with the proof of proposition [ prop : existence of an orbit ] .
let @xmath602 be such that @xmath643 and @xmath644 satisfy the hypotheses of lemmas [ lem : linking 1 ] and [ lem : linking 2 ] .
let @xmath645 be the union @xmath646 and define @xmath647we wish to apply the minimax lemma to @xmath231 and @xmath648 .
we first check that @xmath648 is finite .
since , by lemmas [ lem : linking 1 ] , [ lem : linking 2 ] , and the hypothesis on @xmath602 , we have @xmath649 and @xmath650 , we have @xmath651 by lemma [ lem : b ] , @xmath231 maps bounded sets into bounded sets .
therefore , for each @xmath534 , @xmath652 combining the inequalitites [ eq : ineq 1 ] and [ eq : ineq 2 ] we see that for every @xmath534 , @xmath653 and therefore @xmath654 . by lemma [ lem : palais - smale ]
, @xmath231 satisfies the palais - smale condition , and by lemma [ lem : global flow ] , the equation @xmath655 generates a global flow , from which it follows that @xmath656 . by the minimax lemma , @xmath648 is a critical value .
there is therefore a point @xmath657 with @xmath658 and @xmath659 , which completes the proof .
theorems [ thm : upper bound ] and [ thm : relative non - squeezing ] now follow immediately .
in general , it is difficult to estimate the value of this capacity . in this section ,
we announce two energy - capacity inequalities for the capacity @xmath93 , which give some classes of examples of pairs @xmath35 for which it follows that the capacity is finite .
both results follow by arguments similar to those for energy - capacity inequalities given an action selector , as in @xcite .
our observations are that , first , the action - selector arguments apply for relative capacities , and , second , that the spectral norm in leclercq @xcite , as well as a variant of the action selector in @xcite , satisfy the axioms of an action selector .
the proofs of these two theorems are postponed to an upcoming paper @xcite .
we first define let @xmath660 be a @xmath40-dimensional symplectic manifold . given a coisotropic submanifold @xmath661 , and a @xmath40-dimensional symplectic submanifold @xmath662 with @xmath663 , denote by @xmath664 the infimum of the energy of any hamiltonian diffeomorphism which displaces @xmath79 from @xmath509 .
we have the following let @xmath5 be a symplectic manifold and @xmath665 a lagrangian submanifold .
suppose that @xmath666 , and that @xmath101 is a @xmath40-dimensional symplectic submanifold
. then @xmath667 .
similarly , we have let @xmath668 be a @xmath40-dimensional subset of @xmath20 for which @xmath669 .
the second author would like to thank the laboratoire jean leray of the universit de nantes and the dpartement de mathmatiques of the universit libre de bruxelles for their hospitality and the pleasant atmosphere during his visits to work on this project , and the institut mathmatiques de toulouse at the universit paul sabatier for the invitation to give a seminar talk on an early version of these results .
both authors are grateful to sobhan seyfaddini , rmi leclercq , vincent humilire , matthew strom borman , and emmy murphy for interesting and useful discussions .
the research of the first author was supported by the erc starting grant of frdric bourgeois stg-239781-contactmath and by the erc starting grant of vincent colin geodycon .
the second author was supported by the israeli science foundation grant 723/10 , erc 2012 advanced grant 20120216 , and short visit grants from the contact and symplectic topology research network of the european science foundation . | we introduce a notion of symplectic capacity relative to a coisotropic submanifold of a symplectic manifold , and we define a modification of the hofer - zehnder capacity and show it provides an example . as a corollary , we obtain a non - squeezing theorem for symplectic embeddings relative to coisotropic constraints . |
the universe is nowadays at big scales homogeneous and isotropic as measured in the cmbr by the cobe satellite @xcite , and must also has been having these properties since , at least , the era of nucleosynthesis @xcite . in order to explain the isotropy of the universe from theoretical anisotropic models , many authors have considered the bianchi models that can in principle evolve to a friedmann - robertson - walker ( frw ) cosmology .
it has been shown that some bianchi models in general relativity ( gr ) tend to their isotropic solutions , up some extent @xcite , and even that they can explain the level of anisotropy measured by cobe @xcite .
motived by these facts , we have been working in brans - dicke ( bd ) theory @xcite to investigate if bianchi universes are able to isotropize as the universe evolves , and if its evolution can be inflationary . in previous investigations
we have shown that anisotropic , bianchi type i , v , and ix models tend to isotropize as models evolve @xcite .
however , this may happen for some restrictive values of @xmath1 in the cases of bianchi type i and ix , and only bianchi type v model can accomplish an isotropization mechanism within bd current constraints @xcite on @xmath1 .
it has been also shown that the isotropization mechanism in the bianchi type v model can be inflationary , without the presence of any cosmological constant , when small values for the coupling constant @xmath1 are considered @xcite , as in the case of some induced gravity ( ig ) models @xcite .
we have recently shown , however , that the isotropization mechanism can be attained with sufficient amount of e - folds of inflation , only for negative values of @xmath1 . in the present report
we review and generalize some of the main results and present a new @xmath0 frw solution in bd theory .
this paper is organized as follows . in section [ bdsol ]
the bd bianchi type i , v and ix equations are presented . in section [ anisol ]
we review the main results on these models and present a new solution to curve @xmath0 frw cosmologies .
finally , conclusions are in section [ con ] .
@xmath13 \ , \left(h_{1}+h_{2}+h_{3}\right ) \frac{\psi^\prime}{\psi } \nonumber \\[2pt ] & & - ( 1-\nu)[1+\omega(1-\nu)/2 ] ( h_{1}+h_{2}+h_{3})^{2 } -\frac{\omega}{2 } \left ( \frac{\psi^\prime}{\psi}\right)^{2 } -\frac{r_{j}}{2 } a^{6 \nu } \nonumber \\[2pt ] & & = \,\ 8 \pi \ , \frac{\rho a^{3(1+\nu)}}{\psi } \ , \,\ , \end{aligned}\ ] ] @xmath14 \ ,
\rho a^{3(1+\nu ) } \,\ , \ ] ] where a column sum is given by @xmath15 , where @xmath16i , v or ix and @xmath17 / ( -2 a^6 ) \cr % & & & & \cr % & r_{ij}~ = & 0 & 2 / a_1 ^ 2 & [ a_2 ^ 4-a_3 ^
4-a_1 ^ 4 + 2 a_1 ^ 2 a_3 ^ 2 ] / ( -2 a^6 ) \cr % & & & & \cr % & & 0 & 2 / a_1 ^ 2 & [ a_3 ^ 4-a_1 ^ 4-a_2 ^ 4 + 2 a_1 ^ 2 a_2 ^ 2 ] / ( -2 a^6 ) \cr}\ ] ] for the bianchi type v model one has the additional constraint : @xmath18 , implying that @xmath19 and @xmath20 are inverse proportional functions , @xmath21 ; note that the mean hubble parameter , @xmath22 , is for this bianchi type model @xmath23 . additionally , the continuity equation yields : @xmath24 , @xmath25 being a dimensional constant depending on the fluid present .
the vacuum case is attained when @xmath26 .
the system of ordinary differential equations , eqs .
( [ a123]-[psi ] ) , can be once integrated to get ) is valid for bianchi models i , v and ix , as well for frw cosmologies . ] : @xmath27 \ , \psi^{\prime } + [ 2+(1-\nu)(1 + 3\nu)\omega]m_{\nu } \ ,
\psi + \nonumber \\ & & \frac{2}{3(1-\nu ) } \ , [ 2 - 3\nu + \frac{3}{2 } ( 1-\nu)^{2 } \omega ] \ , [ m_{\nu } ( 1 - 3\nu ) \ , \eta + \delta ] ^{2 } + ( 1-\nu ) ( h_{1}^{2}+ h_{2}^{2}+ h_{3}^{2 } ) = 0 \,\ , \end{aligned}\ ] ] where @xmath28 is an integration constant , @xmath29 , and the hubble rates are written as follows ( similar to the bianchi type i model deduced in ref .
@xcite ) : @xmath30 where the @xmath31 s are some unknown functions of @xmath32 that determine the anisotropic character of the solutions . if @xmath33 for @xmath34 , @xmath35 simultaneously , no anisotropy is present , which is the case of frw cosmologies .
furthermore , bianchi models obey the condition @xmath36 to demand consistency with eq .
( [ hi ] ) .
for the bianchi type v model one has additionally that @xmath37 , since @xmath38 as mentioned above .
equations ( [ hi ] ) and ( [ resth123 ] ) imply that the mean hubble parameter is determined by @xmath39 alone : @xmath40 \ , .\ ] ] in order to analyze the anisotropic character of the solutions , we consider the anisotropic shear , @xmath41 .
@xmath42 is a necessary condition to obtain a frw cosmology since it implies @xmath43 , cf .
if the sum of the squared differences of the hubble expansion rates tends to zero , it would mean that the anisotropic scale factors tend to a single function of time which is , certainly , the scale factor of a frw solution . the anisotropic shear becomes , using eqs .
( [ hi ] ) and ( [ resth123 ] ) , @xmath44 , or the dimensionless shear parameter @xcite , using eq .
( [ psih ] ) : @xmath45^{2 } } \,\ .\ ] ] if the above equations admit solutions such that @xmath46 as @xmath47 ( @xmath48 ) , then one has time asymptotic isotropization solutions , similar to solutions found for the bianchi models in gr @xcite .
in fact , one does not need to impose an asymptotic , infinity condition , but just that @xmath49 , where @xmath50 is yet some arbitrary value to warrant that @xmath51 can be bounded from above .
the problem to find solutions of bianchi and frw models in the bd theory has been reduced to solve the coupled system of equations ( [ psi1 ] ) , ( [ hi ] ) and ( [ a123 ] ) .
let us present in the following subsections the known and new solutions . for this bianchi model the known , the general solution is found in which the @xmath31 s are constants , then eq . ( [ psi1 ] )
is decoupled from eqs .
( [ a123 ] ) and ( [ hi ] ) , and the solution is @xmath52 , where the constants @xmath53 are reported elsewhere @xcite .
this model can be solved in a general way since the curvature sum column @xmath54 , then eq . ( [ psi ] ) can be directly integrated .
this solution represents a particular solution of eq .
( [ psi1 ] ) .
direct substitution of @xmath39 into eq .
( [ hi ] ) gives the hubble rates , and into eq .
( [ dlsig ] ) shows that solutions isotropize as time evolves , that is , @xmath46 as @xmath47 , see ref .
however , the isotropization mechanism is only possible for solutions such that @xmath55 @xmath56 \delta^{2 } \nonumber \\ & & + 3(1-\nu)^{2 } ( h_{1}^{2}+h_{2}^{2}+h_{3}^{2 } ) ] \end{aligned}\ ] ] is negative @xcite .
then , some restrictions on @xmath1 apply .
for instance , in dehnen s ig theory @xcite @xmath57 , then the isotropization mechanism is not possible in this bianchi model @xcite .
this bianchi model is more complicated because curvature terms are different than zero .
still , it is possible to find particular solutions of eq .
( [ psi1 ] ) , since for this model the @xmath31 s are constants too , and this equation is decoupled as is the case of bianchi type i model . the general solution of this model should be obtained through the general solution of eq .
( [ psi1 ] ) , yet unknown . we have found a particular solution that is again of the form @xmath58 , where the constants @xmath59 are reported in refs .
this solution is such that @xmath60 is negative for @xmath61 , so the solution tends to isotropic solution within bd theory constraints @xcite , @xmath62 , that is @xmath46 as @xmath47 . for this bianchi type model , dehnen s ig theory @xcite can achieve an isotropization mechanism @xcite .
an inflationary behavior may be observed in type v models , but to get enough e - foldings of inflation ( @xmath63 ) one must demand that @xmath64 @xcite in consistency with previous results @xcite .
bianchi type ix model is the most complicated to solve , since curvature terms involve quartic polynomials of the scale factors , see eq .
( [ curv ] ) . for this bianchi model it implies , by imposing the condition that @xmath31 s are constants , severe algebraic constraints on the scale factors , so it seems more likely that @xmath31 s are functions .
this explains why no totally anisotropic ( @xmath65 ) solution has been found yet . in ref .
@xcite we have analyzed the case when the polynomial solution for @xmath39 is valid . in this case , unfortunately , we could not found explicitly the values of the constants @xmath66 .
if this solution is valid , however , one has that @xmath67 , where @xmath68 are constants . accordingly , eq . ( [ dlsig ] ) indicates that the solution must tend , as time evolves , to the positive curvature frw solution , i.e. one has again that @xmath46 as @xmath47 .
however , a definitive answer will arrive by obtaining explicitly the values @xmath66 .
the only possible solutions for bianchi type i and v models imply that the @xmath31 s are constants , whereas for type ix they are unknown functions of @xmath32 .
an explanation of this fact resides in the property that bianchi type i and v models have curvature terms of frw type , whereas type ix has a very much complicated form , see eq .
( [ curv ] ) .
so the things , it seems that the most general solution of eq .
( [ psi1 ] ) with @xmath31 s constants would give general solutions for bianchi models i and v. the particular quadratic - polynomial solution of eq .
( [ psi1 ] ) represents in the case of bianchi type i model its the general solution , whereas for bianchi model v it is only a particular solution .
then , other particular solutions , possibly of non - polynomial nature , are expected to be found for eq .
( [ psi1 ] ) that will reveal new aspects of bianchi type v model .
finally , for bianchi type ix model the quadratic - polynomial can be a possible solution , not yet confirmed .
however , solutions with @xmath69 valid for the bianchi type ix model are almost impossible to find because of the complexities involved in the curvature terms .
it turns out that eq .
( [ psi1 ] ) is also valid for the frw models when the anisotropic parameters vanish , i.e. @xmath70 . solutions of this equation solve frw cosmologies in bd theory @xcite .
the known solutions for @xmath39 are quadratic polynomials in @xmath32 as well .
the general flat ( @xmath71 ) solution is a particular solution of eq .
( [ psi1 ] ) in which the coefficient of the quadratic polynomial term , @xmath72 , is equal to the corresponding coefficient ( @xmath73 ) of the bianchi i case . for curved ( @xmath74 )
frw cosmologies the known particular solution is such that the coefficient of @xmath72 of the quadratic polynomial term is equal to the corresponding coefficient ( @xmath75 ) in the bianchi v case . in this way , one can see a correspondence between anisotropic and isotropic solutions .
we have found a new solution of eq .
( [ psi1 ] ) that is valid for @xmath74 frw cosmologies .
the new solution is : @xmath76 \left[\kappa_{1 } + 2 \eta + \kappa_{2 } \ , \left[\kappa_{1 } + ( 1 - 3\nu ) \
, \eta \right]^{\frac{2}{1 - 3\nu } } \right ] \,\ , \ ] ] where @xmath77 are arbitrary integration constants .
this is the general solution of curved frw cosmologies when the following relationships are valid : @xmath78 \omega & = & \frac{- 18 \ , \nu}{(1 + 3\nu)^2 } \,\
.\end{aligned}\ ] ] though the latter relationship constrains the range of possible values of @xmath1 and @xmath10 , one can find values of physical interest , e.g. @xmath79 that has some interest in string effective theories , see for instance ref .
. moreover , when @xmath80 one obtains the gr limit @xmath81 .
finally , one gets the dust model ( @xmath82 ) in the limit when @xmath83 , like in dehnen s ig theory @xcite .
the hubble parameter is given by : @xmath84^{\frac{2}{1 - 3\nu } } } { \left [ \kappa_{1 } + ( 1 - 3\nu)\ , \eta \right ] \left[\kappa_{1 } + 2 \eta + \kappa_{2 } \
, \left[\kappa_{1 } + ( 1 - 3\nu ) \ , \eta \right]^{\frac{2}{1 - 3\nu}}\right ] } \,\ , \ ] ] from which one can find the scale factor : @xmath85^{\frac{1}{2(1 - 3\nu ) } } \frac{\left[\kappa_{1 } + 2 \eta + \kappa_{2 } \ , \left[\kappa_{1 } + ( 1 - 3\nu ) \ , \eta \right]^{\frac{2}{1 - 3\nu } } \right]^{\frac{1}{2(1 - 3\nu ) } } } { \left[\kappa_{1 } + ( 1 - 3\nu ) \ , \eta \right]^{\frac{3 \nu}{(1 - 3\nu)^{2 } } } } \,\ .\ ] ] the bd field ( @xmath86 ) is obtained through eqs .
( [ psifrw ] ) and ( [ afrw ] ) : @xmath87^{\frac{1 + 3\nu}{(1 - 3\nu)^{2 } } } } { \left[\kappa_{1 } + 2 \eta + \kappa_{2 } \ , \left[\kappa_{1 } + ( 1 - 3\nu ) \ , \eta \right]^{\frac{2}{1 - 3\nu } } \right]^{\frac{1 + 3\nu}{2(1 - 3\nu ) } } } \,\ .\ ] ] eqs .
( [ afrw ] ) and ( [ phifrw ] ) imply that the sign of @xmath88 is equal to the sign of @xmath89 for most values of @xmath10 . for open ( @xmath90 ) models
this implies that @xmath91 must be positive , and for closed ( @xmath92 ) models @xmath88 must be negative which allows the solutions to ( re)collapse : the value of @xmath91 determines the time of maximum expansion , so it is very related to the mass ( @xmath93 ) of the model . on the other hand , @xmath94 represents a @xmath32-time shift . because of the mathematical form of eq .
( [ hfrw ] ) it is not possible to have an inflationary era that lasts for a sufficient time period to solve the horizon and flatness problems . to show the behavior of the models we have plotted the above - given formulae for different values of @xmath1 , the curvature constant @xmath95 , and integration constants @xmath96 and @xmath88 . we have chosen in all figures that @xmath97 as initial condition . given a specific value for @xmath1 implies two possible values of @xmath10 consistent with eq .
( [ wnu ] ) . for @xmath79 , a value that makes bd theory to resemble string effective theories @xcite
, it corresponds @xmath98 ( that represents a quasi dust model ) and @xmath99 . in figures [ fig1 ] and [ fig2 ] we plotted the scale factor and bd field for @xmath100 .
figure [ fig1 ] shows an open model with @xmath101 , whereas figure [ fig2 ] is a closed model with @xmath102 and @xmath103 .
in closed models , the smaller @xmath88 is , the later in time they will recollapse .
however , for models with @xmath104 one can find an upper limit on @xmath88 such that models never recollapse ( this happens , for instance , when @xmath105 , making @xmath106 , and if @xmath107 ) ; the effect of negative pressure avoids recollapse .
one can compute the asymptotic limit for @xmath90 models , when @xmath108 .
there are two limit cases , when the quantity @xmath109 is greater or smaller than 1 , cf .
terms in the equations above . in the former case one
has that @xmath110^{\frac{1}{1 - 3\nu } } \nonumber \\[2pt ] \phi_{{\rm asimp1 } } & = & m_{\nu } \ , \kappa_2 \ , \left(\frac{1 - 3\nu}{1 + 3\nu}\right)^{2 } \left(-k \right)^{- \frac{3(1-\nu)}{2(1 - 3\nu ) } } \ , = \ , { \rm const . } \,\ , \end{aligned}\ ] ] whereas in the latter case one gets @xmath111^{\frac{1}{2(1 - 3\nu ) } } \left(1 - 3\nu \right)^{\frac{-3 \nu}{(1 - 3\nu)^{2 } } } \ , \eta^{\frac{1 - 9\nu}{2(1 - 3\nu)^{2 } } } \nonumber \\[2pt ] \phi_{{\rm asimp2 } } & = & 2 \ , m_{\nu } \ , \frac{(1 - 3\nu)^{3 + \frac{9\nu(1-\nu)}{(1 - 3\nu)^{2}}}}{(1 + 3\nu)^{2 } } \left(\frac{-\kappa_2}{2 k}\right)^{\frac{3(1-\nu)}{2(1 - 3\nu ) } } \ , \eta^{\frac{(1 + 3\nu)^{2}}{2(1 - 3\nu)^{2 } } } \,\
.\end{aligned}\ ] ] the plots of figure [ fig1 ] tends to eqs .
( [ limg ] ) ; one observes that @xmath112 tends to a constant value that can be fix to be @xmath113 through the right choice of the constants @xmath114 and @xmath88 .
( + 6.5,6.5)(0,1 ) = 5.0 in 1.5 cm ( 6.5,6.5)(0,1 ) 1.5 cm figures [ fig3 ] and [ fig4 ] show an open and closed model , respectively , with @xmath106 .
again , there are two possible values of @xmath10 , @xmath115 and @xmath116 . in these figures
we have chosen the value of @xmath117 . in figure [ fig3 ]
we have chosen @xmath101 , whereas in figure [ fig4 ] @xmath102 , @xmath118 .
the plots of figure [ fig3 ] do not follow eq .
( [ limg ] ) nor ( [ lims ] ) since @xmath109 is numerically very close to one , and the its asymptotic behavior lies somewhere in between . ( + 6.5,6.5)(0,1 ) = 5.0 in 1.5 cm ( 6.5,6.5)(0,1 ) 1.5 cm the relationship between the cosmic time @xmath119 and @xmath32 , @xmath3 , seems to be too complicated to be integrated in a closed form .
however , in the asymptotic limits given by eqs .
( [ limg ] ) and ( [ lims ] ) one obtains that @xmath120 t_{{\rm asimp2 } } & = & \frac{2 \left[\frac{-2k}{\kappa_{2}}\right]^{\frac{3 \nu}{2(1 - 3\nu ) } } ( 1 - 3\nu)^{2-\left(\frac{3\nu}{1 - 3\nu}\right)^{2 } } } { 2 - 9\nu(1+\nu ) } \ , \eta^{\frac{2 - 9\nu(1+\nu)}{2(1 - 3\nu)^{2 } } } \ , .\end{aligned}\ ] ] in both cases the functions are monotonic , and for @xmath100 and @xmath117 used in our plots , time grows as @xmath32 grows .
therefore , our time parametrization is appropriate . in this limit
, one can express our solutions in @xmath32 in terms of @xmath119 to get that : @xmath121 \phi_{{\rm asimp1 } } & = & m_{\nu } \
, \left(\frac{1 - 3\nu}{1 + 3\nu}\right)^{2 } \kappa_2 \ , \left(-k \right)^{- \frac{3(1-\nu)}{2(1 - 3\nu ) } } \ , = \ , { \rm const . } \end{aligned}\ ] ] and @xmath122^{\frac{1 - 3 \nu}{2 - 9\nu(1+\nu ) } } \left(1 - 3\nu\right)^{\frac{-2 + 3\nu(8 - 3\nu)(1 - 3\nu)}{2(1 - 3\nu)^{4 } } } \left(1 - 9\nu(1+\nu)/2\right)^{\frac{1 - 9 \nu}{2 - 9\nu(1+\nu ) } } \ , t^{\frac{1 - 9\nu}{2 - 9\nu(1+\nu ) } } \nonumber\\[2pt ] \phi_{{\rm asimp2 } } & = & \frac{2^{-\frac{(1 + 3\nu)(2 + 3\nu)}{2 - 9\nu(1+\nu)}}\ , m_{\nu}}{(1 + 3\nu)^{2 } } \left[\frac{-\kappa_{2}}{k}\right]^{\frac{3(1+\nu)(1 - 3 \nu)}{2 - 9\nu(1+\nu ) } } \left(1 - 3\nu\right)^{\frac{4 - 21\nu-27 \nu^{2}}{2 - 9\nu(1+\nu ) } } \nonumber\\ & & \left(2 - 9\nu(1+\nu)\right)^{\frac{(1 + 3\nu)^{2}}{2 - 9\nu(1+\nu ) } } \ , t^{\frac{(1 + 3\nu)^{2}}{2 - 9\nu(1+\nu ) } } \ , .\end{aligned}\ ] ] eq .
( [ solt1 ] ) is a particular known flat solution in bd theory @xcite or the @xmath90 vacuum solution of gr .
( [ solt2 ] ) is the flat space , nariai solution @xcite in which @xmath1 is given by eq .
( [ wnu ] ) .
we have presented a set of differential equations written in rescaled variables that let us integrate a general equation to get eq .
( [ psi1 ] ) , valid for bianchi type i , v and ix models , as well as for frw models .
this equation is coupled to eqs .
( [ a123 ] ) and ( [ hi ] ) , but for bianchi type i and v models the anisotropic parameters ( @xmath31 s ) are constants ( since their curvature terms are of frw type ) and eq .
( [ psi1 ] ) is decoupled .
this property allows one to find the general solution of bianchi type i model and a particular solution of type v ; both solutions are quadratic polynomials .
this solution let the models isotropize as time evolves , however , this can happens only for some parameter ( @xmath1 , @xmath10 ) range .
the polynomial solution may also be valid for bianchi type ix , but it is not proved yet . if it were , isotropization would be also guaranteed .
we have found a new solution of eq .
( [ psi1 ] ) valid for curved ( @xmath74 ) frw cosmologies , that is , a solution with @xmath33 .
this is the general solution subject to the constraint given by eq .
( [ wnu ] ) .
accordingly , we have analyzed two cases of physical interest : the case when @xmath79 , having some interest in string cosmology , that implies an equation of state of a quasi dust model ( @xmath123 ) , and the case when @xmath124 , consistent with current bd local experimental constraints @xcite , implying that @xmath125 .
although in a different manner , both the scale factor and @xmath112 grow as a function of the parametrized time @xmath32 in all figures presented here .
one can find specific values of the constants @xmath126 to have `` closed '' models without recollapse .
for this to happen @xmath109 must be less than 1 , i.e. , @xmath10 must be negative .
this is a known effect of negative pressures .
the new solution is non - inflationary and for asymptotic times is of power - law type for the scale factor . on the one hand , when the quantity @xmath109 is greater than 1 the bd field tends to a constant value , then bd s dynamics evolves similar to gr s .
one the other hand , when @xmath109 is smaller than 1 , both scale factor and bd field behave asymptotically with a power - law , and the solution is equal to nariai solution for flat space . further solutions of eq .
( [ psi1 ] ) are in order , which can be either within frw cosmologies or bianchi type v or ix models . in particular , the general solution with @xmath31
s constants should provide the general solution of both curved frw cosmologies and bianchi type v model .
finally , our results could be also of interest for @xmath127-theories , where the value of the coupling parameter in some early cosmological era could have been rather different than its value nowadays , @xmath62 .
dehnen , h. frommert , h. and ghaboussi , f. ( 1992 ) .
_ * 31 * 109 ; dehnen , h. and frommert , h. ( 1993 ) .
phys . _ * 32 * 135 ; cervantes - cota , j. l. and dehnen , h. ( 1995 ) . _ phys .
d _ * 51 * , 395 ; _ ibid . _ ( 1995 ) .
b _ * 442 * , 391 .
chauvet , p. , cervantes - cota , j. and nez - ypez , h. n. ( 1991 ) . in _ proceedings of the 7th latin american symposium on general relativity and gravitation , silarg vii _ ,
dolivo , j. c. , et al . , eds .
( world scientific , singapore ) , p. 487 .
morganstern , r. e. ( 1971 ) .
d _ * 4 * , 282 ; ruban , v. a. and finkelstein , a. m. ( 1976 ) .
_ astrofizika _ * 12 * , 371 ; lorenz - petzold , d. ( 1984 ) . _
space sci . _ * 98 * , 249 ; barrow , j. d. ( 1993 ) .
d _ * 47 * , 5329 . | using scaled variables we are able to integrate an equation valid for isotropic and anisotropic bianchi type i , v , ix models in brans - dicke ( bd ) theory .
we analyze known and new solutions for these models in relation with the possibility that anisotropic models asymptotically isotropize , and/or possess inflationary properties .
in particular , a new solution of curve ( @xmath0 ) friedmann - robertson - walker ( frw ) cosmologies in brans - dicke theory is analyzed .
1 cm key words : cosmology ; analytic solutions |
the nuclear inclination of active galactic nuclei ( agn ) is a fundamental parameter that still remains tricky to directly measure . in radio - quiet objects ,
proxies for measuring the orientation of agn usually rely on degenerate models and fitting procedures that sometimes give inconsistent results @xcite .
however , in the case of radio - loud agn , in which there is beamed emission from pc - scale jets and unbeamed emission from diffuse radio lobes , we have an attractive method since the ratio of the former to the latter ( the radio core dominance r ) should be statistically a function of inclination with respect to the line of sight @xcite .
operationally , it works well because on arcsec - scale maps , double radio sources generally show an unresolved and fairly isolated point source , usually with a flat synchrotron spectrum implying extreme compactness ( pc scales ) .
furthermore selection by the nearly - isotropic lobe emission assures that a sample has a random orientation distribution , filling all solid angles as seen from the nucleus .
this core dominance parameter is usually estimated in the radio because , in a general way , the jet and lobe radiation are linked physically . but ( for quasars only ) , the continuum optical flux density @xcite has also been used to normalize the beamed radio core flux density , instead of the flux density of the extended radio lobes . this core dominance parameter is called r@xmath3 . its usefulness has been confirmed by @xcite .
narrow line luminosities have also been tried @xcite .
very little is known about the reliability and precision of r as a statistical indicator of inclination .
one way to test this is to see to what extent it separates radio galaxies ( rg ) from quasars , since we know that among the high-@xmath4 3cr radio galaxies , all rg host hidden quasars , and that they generally have smaller inclinations .
we do nt know a priori whether or not a lot of noise will be added by a distribution of torus half - opening angles , `` holes '' in the torus , or dispersion in the radio core emission polar diagrams .
but if core dominance correlates very well with object type , then the most natural conclusion is that the sources are roughly generic in nature , core dominance indicates inclination well , and there are very few quasars at high - inclinations .
this is in fact what we will show .
the unified model in its simplest form , the `` straw person model '' of @xcite , seems to be correct to first order in this parameter space , as predicted by @xcite .
that is , the model shows very strong predictive power .
the same conclusion follows from a detailed study of the x - ray properties of the sample by @xcite .
high core dominance is known to be associated with low inclination and a flat ir sed ; it s also known that radio galaxies have lower core dominance than quasars ; we will show that limited optical polarization is also consistent with this idea when available .
the purpose of this paper is to note and exploit the fact that there is essentially zero overlap in r@xmath5 between rgs and quasars , using that fact and the special sample properties to derive a quantitative formula for converting core dominance to inclination , estimating that it is good to plus or minus 10 degrees . to obtain intelligible results ,
it is fundamental to select a sample were all rg have hidden quasars , and to select by an approximately isotropic luminosity ( the radio lobes luminosity here ) .
the high redshift ( i.e. @xmath0 1 ) radio - loud agn from the complete flux - limited , 178 mhz selected , `` 3crr '' catalog of @xcite meet this requirement very well . we know this for several reasons .
first , most these objects have been observed polarimetrically and all show substantial optical polarization , oriented roughly perpendicular to the radio axes @xcite .
they also show broad emission lines in polarized flux when that information is available . also , all of these radio galaxies for which data is available show strong mid - infrared reprocessing bumps @xcite and high ionization emission lines @xcite .
it is important to note that this sample comprises many of the most radio - luminous objects in the universe , and at lower radio luminosities many radio galaxies lack energetically significant hidden quasars @xcite ; the evidence for that statement from all wavebands is reviewed in detail in @xcite .
in the following subsections , we compare radio r to the quasar / rg optical classification and also to infrared colors and optical polarization . doing so , we aim to robustly constrain the reliability and precision of the radio core dominance parameter to build our own semi - empirical relation linking the nuclear inclination and r. we extracted the 3crr object types and redshifts @xmath4 from the on - line catalog and references therein .
the radio core dominance was estimated between the 178 mhz flux densities corrected to the flux scale used by @xcite and the flux density of the core at 5 ghz ( where it s better isolated ) on arcsecond scales .
additional r@xmath5 are extracted from @xcite , @xcite , @xcite , and @xcite .
infrared photometry was obtained with the spitzer space telescope and reported by @xcite .
the optical polarization data are summarized in @xcite .
our final , 100% completeness sample is listed in tab . [
table : data ] .
we plot the number distribution of the core dominance parameter for radio - galaxies and quasars in fig .
[ fig : histogram ] .
the top histogram shows the distribution of radio - galaxies ( in red ) and the bottom figure quasars ( in green ) .
the separation between the two radio - loud agn classes is almost perfect , with only a few objects possibly overlapping between 0.045 @xmath6 r@xmath5 @xmath6 0.079 : two quasars ( 3c 68.1 , 3c 318 ) and two radio - galaxies ( 3c 294 , 4c 13.66 ) .
3c 68.1 is a quasar with a very low core dominance ( r@xmath5 = 0.079 ) .
however , @xcite have shown that its spectral index is about 6 for the optical continuum , indicating heavy absorption .
3c 68.1 is also highly polarized , with a continuum polarization as high as 10% in the blue band ( 1900 rest frame ) , see @xcite .
this indicates that 3c 68.1 is a borderline object , and this is consistent with the core dominance being on the boundary between the two types .
3c 318 has an intermediate spectral type , being classified as a galaxy by @xcite .
broad h@xmath7 ( but not h@xmath8 ) is detected by @xcite .
this quasar shows very strong far - infrared ( fir ) fluxes ( 148 @xmath9 24 mjy at 60 @xmath10 m ) such as reported by @xcite , with fir fluxes likely to be dominated by a non - thermal beamed component @xcite .
recent observations using the herschel space observatory revealed that most of the infrared emission from 3c 318 originates from a pair of interacting galaxies , close in projection @xcite .
finally , note that the one radio galaxy plotted with r@xmath5 @xmath11 0.1 is an upper limit .
the overlap of quasars and radio - galaxies at 0.045 @xmath6 r@xmath5 @xmath6 0.079 is thus naturally explained by one misclassified source and one object whose infrared flux is polluted by a pair of interacting galaxies .
accounting for these peculiarities , the resulting association between radio r and optical type is perfect , perhaps fortuitously so .
we can draw two conclusions .
first , the nature of these powerful radio - loud agn is quite generic , i.e. the half - opening angle of the circumnuclear dusty region must be quite similar between radio - galaxies and quasars .
second , all radio - galaxies ( type-2 agn ) are seen at higher inclinations than the quasars ( type-1 agn ) ; there are no high - inclination quasars seen through `` holes '' in the torus .
these conclusions are in perfect agreement with the unified model of agn such as postulated by @xcite , @xcite , and @xcite .
measurements from @xmath12 , @xmath13 , @xmath14 and other infrared observatories showed that quasars have strong near - infrared ( nir ) and mir continua , bright silicate features and an emission bump at 2 5 @xmath10 m ( see , e.g. , @xcite ) .
their spectral energy distribution ( sed ) is quite generic , with much lower dispersion in their fluxes and ir colour compared with radio galaxies sed @xcite .
additionally , the 2 5 @xmath10 m radio galaxy extinctions derived by @xcite are higher than the ( wavelength - normalized ) values derived in the mir , indicating that the component responsible for the nir bump suffers more extinction than the mir emission .
this is consistent with the torus geometry .
hence , looking at the nir and mir fluxes it is possible to try to corroborate the association we found between r@xmath5 and optical type @xcite . by selecting our sample purely on unbeamed radio emission in the @xmath0 1 group , we are able to study the ir without biases . in the following ,
the spitzer 3.6 @xmath10 m and 24 @xmath10 m observed - frame fluxes are taken from @xcite , also listed in @xcite , and from the nasa / ipac extragalactic database .
[ fig : ir_fluxes ] shows the radio core dominance parameter versus nir ( top ) and mir ( bottom ) fluxes .
we see that the 3.6 @xmath10 m rg fluxes are , on average , a factor 6.3 lower than the quasar fluxes ( median : 6.8 ) , which is in good agreement with the quasar / radio - galaxy flux ratios ( 2.5 7.5 ) found by @xcite at 3.5 @xmath10 m .
the same conclusions apply for the 24 @xmath10 m fluxes against r@xmath5 , where the mir emission is found to be more isotropic . due to the dispersion of data points ,
the flux ratio is more complicated to estimate but the averaged value is 3.3 ( median : 3.4 ) . the flux ratio between nir and mir emission is shown in fig .
[ fig : ir_ratio ] .
the radio - galaxies are below the quasars on average , by a factor @xmath2 1.35 .
there is a fairly low dispersion among the quasars as their ir seds are generic ; in the case of radio - galaxies the distribution is less tight as rg show less uniformity than quasars @xcite . for a a given 24 @xmath10 m flux , the 3.6
@xmath10 m flux is much lower for the radio galaxies , as expected from the obscuration paradigm .
finally , it seems that we are also detecting an inclination - dependence of the nir / mir flux ratio within each class in the expected sense .
our conclusions that the @xmath0 1 3crr radio galaxies have several times weaker nir and mir emission compared with the matched quasars , and more varied seds , are not new .
they were stated explicitly by @xcite , @xcite , and references therein ; @xcite shows quotient spectra for radio galaxies over quasars . for lower redshift 3cr objects , @xcite reported these effects , with references to much earlier work .
the novel aspect here is again to exploit the near perfect separation of the classes by ir colors to constrain deviations from the simplest unified model .
the infrared fluxes have shown excellent agreement with the expectation from the unified model and strongly corroborate our finding that there is an almost perfect separation between the two optical classes of radio - loud agn based on r@xmath5 .
nevertheless , we decided to do a third check using near - ultraviolet , optical and near - ir continuum polarization measurements from the literature .
the unified model again makes strong predictions about the degree and orientation of agn polarization and we expect a clear dichotomy between type-1 and type-2 agn .
our sample is too small to demonstrate the predicted relationships by itself , but we can check for consistency . tab . [
table : data ] lists the only 13 polarimetric data reported for the 3crr objects in our sample .
these agn have been observed in u , b , v and/or r polarization filters , and were corrected for the small polarizations of the instrument and the galactic interstellar medium .
the polarization degree @xmath15 and the difference @xmath16 between the radio position angle pa and the optical polarization position angle @xmath17 are presented in fig .
[ fig : polarization ] top and bottom , respectively .
@xmath15 seems similar between rg and quasars , but radio galaxies are generally substantially diluted by starlight , while that of quasars is not @xcite , which is part of the reason why the polarization degree of the two groups seems similar .
however , there is a net difference between type-1 and type-2 agn : rg only show a polarization position angle approximately perpendicular to the projected axis of the radio structures ( @xmath16 = 90@xmath1 ) while quasars are characterized by a blend of perpendicular and parallel polarization angles
. the exact same dichotomy appears in radio - quiet agn @xcite and radio - loud agn at lower redshifts @xcite , where type-2s have perpendicular @xmath17 only .
the presence of perpendicular @xmath17 indicates that photons dominantly experience scattering in polar directions , while a parallel @xmath17 is the signature of equatorial scattering .
polar scattering dominated quasars tend to have higher degrees of polarization ( up to a couple of percents ) than equatorial scattering quasars , as already observed for their radio - quiet analogues @xcite .
a large part of the higher polarization in well - studied type-2 agn results from the fact that the direct light from the quasar is blocked ; otherwise it would dilute the observed polarization , often very strongly .
however , the relatively high polarization of the polar - scattered quasars indicates that our inclination - dependent view of the scattering medium also plays a role .
finally , we note that the incidence of high parallel polarization in the quasars is much higher than in other well - studied samples @xcite , although it s not clear why this should be the case .
since we are very confident that core dominance is a good qualitative inclination indicator , and that the sources are distributed uniformly in solid angle ( copernican principle ) , we can make a quantitative semi - empirical function for the average inclination @xmath18 as a function of core dominance .
it should be of interest to both observers and theorists @xcite .
to do so , we equally distributed the sources in solid angle and numerically fitted the relation using a polynomial regression .
we obtained the following semi - empirical relation between core dominance and agn inclination : @xmath19 or , if numerically inverted(r@xmath5 ) becomes less reliable for extremes values of @xmath18 . ] : @xmath20 with lr = @xmath21 .
the first formula , essentially the polar diagram for core emission , should be more useful to theorists , while the second version is directly useful to observers wishing to estimate the inclination for a given system .
[ tab : fit ] the parameters `` a '' and `` g '' are constants that represent the equivalent emission of an isotropic component . values for @xmath22 , @xmath23 , @xmath24 , @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 , @xmath32 , and @xmath33 , are given in tab . [
tab : fit ] , together with the goodness measures of the fitting function to the distribution of the sources in solid angle .
the typical uncertainty in @xmath18 can be no more than about @xmath9 10@xmath1 ; otherwise simulations show that the rg and quasars would overlap substantially .
the number distribution resulting from this analytical formula is shown in fig .
[ fig : theory ] ( top ) .
our new function properly reproduces the observed increase of quasar number at large r@xmath5 and predicts that the distribution should be almost uniform in @xmath18 at large inclinations .
note that our formula must not be used in highly core - dominant objects , since the distribution is not sampled well at small inclinations .
although we do nt have many sources , we can estimate the averaged half - opening angle of the circumnuclear region by estimating the ratios of type-1 to type-2 objects in out complete sample , which is 0.574 , corresponding to a critical inclination of 55@xmath1 ( 60@xmath1 if excluding the two peculiar quasars mentioned in sect .
[ main : histogram ] ) .
this result agrees with the values found by previous authors , including @xcite , @xcite , @xcite and many others , who looked at high - luminosity radio - loud agn and estimated that the half - opening angle of the equatorial region is @xmath2 45@xmath1 60@xmath1 . in the future
we may enlarge our sample , or explore other parameter space using the same methods .
we plotted r@xmath5 as a function of @xmath18 in fig .
[ fig : theory ] ( bottom ) , together with the analytical formula from @xcite , where it is assumed that jets are co - linear ( zero opening angle ) and moving with a single bulk velocity , so that r@xmath5(@xmath18 ) is just determined by special relativity . when compared with this naive model , our semi - empirical formula shows much stronger than predicted core fluxes at intermediate angles , so that the polar diagram is not as sharply peaked at @xmath18 = 0 as in the toy model .
taken at face value , the bulge in the polar emission diagram at @xmath34 40@xmath1 is consistent with the popular spine - and - sheath model @xcite , where there is a slow jacket to a fast very narrow jet .
the slower sheet ( layer ) is only modestly beamed and substantial broadening of the effective beaming cone is required .
this indicates a range of bulk speeds among objects or within individual jets , a range of direction of synchrotron - emitting plasma elements , optical depth effects , or other deviations from the single - zone jet @xcite .
however , our sample is small and better statistics are needed to evaluate the statistical significance of the bulge and , more generally , of the r(i ) formula .
we have demonstrated that the radio core dominance parameter r@xmath5 separates the radio - galaxies and quasars almost perfectly in the @xmath0 1 3crr sample . in other words , the agreement of r@xmath5 with optical type is not a coincidence , and probably both r@xmath5 and optical type are reliable orientation indicators .
since the 3crr catalog is a complete sample , it must fill the solid angle uniformly , except for small number statistics , just from the copernican principle .
therefore we were able to derive an empirical core dominance formula , where r@xmath5 is a function of inclination @xmath18 .
the essentially perfect separation of the optical types by radio core dominance and infrared ( and x - ray columns ) is most simply interpreted as meaning that there is only a small dispersion of opening angle , and no holes in the circumnuclear region which would let us see a type-1 agn at high inclination ( contrary to the hypothesis of @xcite ) . at lower luminosities ,
such predictions are less clear as the sample gets less clean .
it is possible to extend the catalog to lower redshifts but it would be necessary to exclude nonthermal rg .
weakly - accreting radio - galaxies mostly radiate through kinetic energy in the form of synchrotron jets and lack highly ionized line emission and strong ir reprocessing bumps , which indicates that they lack energetically significant hidden quasars ( @xcite and references therein ) .
it is then essential to remove them to have a correct sample . at higher redshifts
another problem may arise due to the fact that the ratio of lobe emission to jet power is sensitive to environment @xcite .
the density of the intergalactic medium ( igm ) scales with redshift as ( 1 + @xmath4)@xmath35 @xcite . for a @xmath4 = 5 quasar ,
the igm density is 27 times higher than compared to a @xmath4 = 1 quasar , and these should strongly affect the morphology , lobe flux ratios , and lobe distance . the authors would like to thank the anonymous referee for her / his useful and constructive comments on the manuscript .
we are grateful to belinda wilkes , peter barthel , patrick ogle , pece podigachoski , and alan marscher for their nice suggestions that helped to improve the quality of this paper . 99 antonucci , r. r. j. 1984 , apj , 278 , 499 antonucci , r. 1993 , ara&a , 31 , 473 antonucci , r. 2012 , astronomical and astrophysical transactions , 27 , 557 arshakian , t. g. 2005 , a&a , 436 , 817 barthel , p. d. 1989 , apj , 336 , 606 barthel , p.
d. , vestergaard , m. , & lonsdale , c. j. 2000 , a&a , 354 , 7 barthel , p. d. , & arnaud , k. a. 1996 , mnras , 283 , l45 blandford , r. d. , & knigl , a. 1979 , apj , 232 , 34 boksenberg , a. , carswell , r. f. , & oke , j. b. 1976 , apjl , 206 , l121 brotherton , m. s. , wills , b. j. , dey , a. , van breugel , w. , & antonucci , r. 1998 , apj , 501 , 110 cimatti , a. , dey , a. , breugel , w. v. , hurt , t. , & antonucci , r. 1997 , apj , 476 , 677 dey , a. 1999 , the most distant radio galaxies , 19 , proceedings of the colloquium , amsterdam , 15 - 17 october 1997 , royal netherlands academy of arts and sciences . edited by h. j. a. rttgering , p. n. best , and m. d. lehnert drouart , g. , de breuck , c. , vernet , j. , et al .
2012 , a&a , 548 , a45 elvis , m. , wilkes , b. j. , mcdowell , j. c. , et al .
1994 , apjs , 95 , 1 fan , j .- h . , yang , j .- h .
, pan , j. , & hua , t .- x .
2011 , research in astronomy and astrophysics , 11 , 1413 fanti , c. , fanti , r. , dallacasa , d. , et al . 1995 , a&a , 302 , 317 fanti , c. , fanti , r. , dallacasa , d. , et al . 2002 , a&a , 396 , 801 gelderman , r. , & whittle , m. 1994 , apjs , 91 , 491 haas , m. , willner , s. p. , heymann , f. , et al . 2008 , apj , 688 , 122 - 127 heckman , t. m. , odea , c. p. , baum , s. a. , & laurikainen , e. 1994 , apj , 428 , 65 he s , r. , barthel , p. d. , & hoekstra , h. 1995 , a&a , 303 , 8 hoekstra , h. , barthel , p. d. , & he s , r. 1997 , a&a , 319 , 757 hough , d. h. , zensus , j. a. , vermeulen , r. c. , et al . 1999 ,
apj , 511 , 84 hough , d. h. , vermeulen , r. c. , readhead , a. c. s. , et al .
2002 , aj , 123 , 1258 hnig , s. f. , leipski , c. , antonucci , r. , & haas , m. 2011 , apj , 736 , 26 laing , r. a. , riley , j. m. , & longair , m. s. 1983 , mnras , 204 , 151 lawrence , a. 1991 , mnras , 252 , 586 leipski , c. , haas , m. , willner , s. p. , et al .
2010 , apj , 717 , 766 ludke , e. , garrington , s. t. , spencer , r. e. , et al .
1998 , mnras , 299 , 467 macquart , j .-
, & koay , j. y. 2013 , apj , 776 , 125 marin , f. 2014 , mnras , 441 , 551 marin , f. 2016 , mnras , in print obied , g. , zakamska , n. l. , wylezalek , d. , & liu , g. 2016 , mnras , 456 , 2861 ogle , p. , whysong , d. , & antonucci , r. 2006 , apj , 647 , 161 orr , m. j. l. , & browne , i. w. a. 1982 , mnras , 200 , 1067 podigachoski , p. , barthel , p. , haas , m. , leipski , c. , & wilkes , b. 2015 , apjl , 806 , l11 podigachoski , p. , barthel , p. d. , peletier , r. f. , & steendam , s. 2016 , a&a , 585 , a142 roger , r. s. , costain , c. h. , & bridle , a. h. 1973 , aj , 78 , 1030 rawlings , s. , & saunders , r. 1991 , nature , 349 , 138 sanders , d. b. , phinney , e. s. , neugebauer , g. , soifer , b. t. , & matthews , k. 1989 , apj , 347 , 29 sazonov , s. , churazov , e. , & krivonos , r. 2015 , mnras , 454 , 1202 scarrott , s. m. , rolph , c. d. , & tadhunter , c. n. 1990 , mnras , 243 , 5p scheuer , p. a. g. , & readhead , a. c. s. 1979 , nature , 277 , 182 smith , j. e. , young , s. , robinson , a. , et al .
2002 , mnras , 335 , 773 sol , h. , pelletier , g. , & asseo , e. 1989 , mnras , 237 , 411 stockman , h. s. , angel , j. r. p. , & miley , g. k. 1979 , apjl , 227 , l55 tadhunter , c. 2005 , astronomical polarimetry : current status and future directions , 343 , 457 urry , c. m. , & padovani , p. 1995 , pasp , 107 , 803 wilkes , b. j. , hooper , e. j. , mcleod , k. k. , et al . 1999 ,
the universe as seen by iso , 427 , 845 .
p. cox & m. f. kessler .
esa - sp 427 wilkes , b. j. , kuraszkiewicz , j. , haas , m. , et al .
2013 , apj , 773 , 15 willott , c. j. , rawlings , s. , & jarvis , m. j. 2000 , mnras , 313 , 237 wills , b. j. , & brotherton , m. s. 1995 , apjl , 448 , l81 zirbel , e. l. , & baum , s. a. 1995 , apj , 448 , 521 | it is believed that , in radio - loud active galactic nuclei ( agn ) , the core radio flux density can be normalized to the flux density of the extended lobe emission to infer the orientation of a radio source .
however very little is known about the reliability and precision of this method , and we are unaware of any robust conversion recipe to infer the inclination from the core dominance . investigating whether or not the radio core dominance parameter r separates the quasars from the radio - galaxies in the @xmath0 1 3crr catalog
, we found excellent agreement of r with optical type , infrared flux ratios and optical polarization .
this indicates that probably both r and optical classification are very good orientation indicators , and the unified model is strongly predictive for these objects .
the relative number densities indicate half - opening angles close to 60@xmath1 , as expected from large surveys .
the separations of optical types according to radio core dominance as well as nir / mir ratios , which are essentially perfect , means that there can be only a small dispersion of torus half - opening angles .
also , even though torus dust is thought to be clumpy , there is an almost zero probability to see a type-1 source at high inclination . finally , using only the copernican principle , i , e , the assumption that solid angle is filled uniformly with source axis orientations , we estimated a semi - empirical relation between core dominance and agn inclination .
this makes it possible to use r to infer the inclination of a source to an accuracy of @xmath2 10 degrees or less , at least for this type of object . |
recent dramatic progress in laser technology has given rise to new projects ( e.g. , eli , xfel ) which may bring intensities of the order of 10@xmath0 - 10@xmath1 w@xmath2 within reach @xcite .
this opens up a wide range of possibilities for exploring novel regimes of ultraintense laser - matter interactions @xcite .
the main feature of such interactions is that the radiation reaction ( rr ) effects @xcite coming into play may strongly modify the conventional relativistic laser - plasma interactions .
first attempts to consider such ultra - relativistic interactions accounting for the radiation reaction force were recently made by using particle - in - cell ( pic ) simulations , particularly showing the qualitative difference between linear and circular polarization of the driving field @xcite . in this paper
we focus on fundamentally new features of self - consistent ultra - relativistic laser plasma interactions .
we pay particular attention to the circularly polarized lasers when electrodynamic description of the interaction can be simplified essentially , providing exact solutions for plasma - field structures .
for those , we derive fundamental definitions , such as nonlinear dielectric permittivity , ponderomotive and dissipative forces acting on an electron at such ultrahigh intensities . besides , based on these treatments an effective plasma converter of laser energy into gamma rays is also proposed .
to provide a self - consistent approach to ultra - relativistic laser plasma interaction by using macroscopic electrodynamic description ( see , for example , @xcite ) we should first clearly define the relationship between macroscopic fields , averaged over small physical volume , and local fields , acting on a particle situated at this point . in conventional electrodynamics of a plasma where the radiation reaction effect is not taken into account
these forces are equal ( see , e.g. , @xcite ) , and what is also important is that these fields at the given point are produced by all other particles except the one situated at this point . using this statement
we may incorporate the radiation reaction effect into macroscopic electrodynamics just by replacing in the single particle motion @xmath3 external fields @xmath4 , @xmath5 by macroscopic fields @xmath6 , @xmath7 , which are produced by all other particles and these fields ( as well including external fields ) are governed by the maxwell equations with the corresponding charges and currents ; usually the same procedure is used in pic modeling .
radiation reaction force can be taken in the lorentz - abraham - dirac ( lad ) form or landau - lifshitz one ; this was much discussed in the literature ( see , e.g. , @xcite ) .
detailed comparative analysis of single particle motion in a laser field was done in ref .
@xcite . for considering the ultra - relativistic laser plasma interaction
when the rr effects are of importance we employ the most simple approach based on relativistic hydrodynamic equations describing the motion of a cold electron fluid accounting for the rr force in the lorentz - abraham - dirac ( lad ) form : @xmath8\right\ } , \label{lad}\\ \frac{d n_e}{d t}+ n_e\nabla\cdot{\bf v}=0\label{cont } , \end{gathered}\ ] ] where , instead of electric and magnetic fields , we introduced vector @xmath9 and @xmath10 scalar potentials , @xmath11 , @xmath12 , @xmath13 is the full time derivative , @xmath14 , @xmath15 are the electron charge and mass , respectively , @xmath16 is the electron momentum , @xmath17 is the electron velocity , @xmath18 is the relativistic factor , @xmath19 is the speed of light , @xmath20 is the electron density .
this approach is also justified by the fact that , at extremely high intensities of interest rr acts as a cooling mechanism ( see , for example , @xcite ) essentially lowering electron temperature as compared to the case when rr is not included . adopting the coulomb gauge ( @xmath21 ) , @xmath9 describes the vortex electromagnetic field , and @xmath10 describes the electrostatic field due to charge separation in the plasma for which the maxwell equations reduce to @xmath22 here @xmath23 is the unperturbed plasma density . the fluid model , representing a significant simplification over a full kinetic treatment based on particle - in - cell simulations or the vlasov equation but retaining enough physics to be qualitatively and quantitatively useful@xcite , without the rr force was widely used in the conventional relativistic interactions , particularly to study fundamentally new properties of the interaction .
the three well known examples related to the case of circularly polarized electromagnetic wave are : ( i ) the propagation of intense plane wave due to the relativistic mass correction @xcite , ( ii ) exact solutions for plasma - field structures in the regime of relativistic self - induced transparency ( sit ) describing the penetration of incident radiation into overdense plasmas @xcite , and ( iii ) electromagnetic solitons of relativistically strong laser fields propagating in underdense and also in overdense plasmas @xcite .
the advantage of using circular polarization is that the ponderomotive force has no oscillating ( with double laser frequency ) component and therefore pushes electrons steadily without significant electron heating .
recently by taking this advantage several schemes for laser ion acceleration into gev energy range were proposed @xcite .
next we will also use this advantage but consider the case of importance of rr effects . paying particular attention to the one - dimensional problem and quasisteady model as well
, we assume that the longitudinal motion is frozen , so that in a circularly polarized field @xmath24 $ ] ( @xmath25 is the laser frequency ) the electron velocity can also be taken in the form @xmath26 $ ] @xcite , where @xmath27 , @xmath28 are the complex functions . substituting them into eq .
( [ lad ] ) for radiation reaction force we obtain @xmath29 where the parameter responsible for the rr effects is @xmath30 , which is approximately equal to 10@xmath31 for a 1 @xmath32 m wavelength .
it is worth noting that similar procedure can also be applied by considering the friction force in the form of landau - lifshitz @xcite instead of the lad form as in eq .
( [ lad ] ) , but for the latter algebraic operations are much simpler . and
as well , in spite of the fact that the lad force may lead to unphysical self - accelerated solutions for the single - electron interaction problem , in our case of a harmonic process no qualitatively new solutions are added . for convenience of notation let us introduce the following dimensionless variables and parameters : longitudinal coordinate ( along the propagation direction ) @xmath33 , vector @xmath34 and scalar @xmath35 potentials , @xmath36 and @xmath37 is the ratio of unperturbed plasma density @xmath23 over the critical one at given laser frequency , the ratio @xmath38 will be denoted as @xmath39 . in this case
( [ eq1 ] ) on the transverse projection gives @xmath40 whereas on the longitudinal one it arrive at @xmath41,\label{force}\ ] ] where the prime denotes the derivative with respect to @xmath42 .
it should be emphasized that equation ( [ force ] ) indicates that in the region where the electron density @xmath43 the force of a longitudinal field due to charge separation is compensated by a sum of two forces ; the first one is the ponderomotive force @xcite and the second one is a dissipative force due to the rr effect , which is generally caused by the imaginary part of dielectric permittivity .
the medium absorbs laser photons thus additionally acquiring an impulse along laser propagation direction . from eq .
( [ a1 ] ) we can also obtain an explicit expression of relativistic factor @xmath44 as a function of field amplitude @xmath45 . multiplying this equation by the complex conjugate term we arrive at the fourth order algebraic equation with respect to @xmath46
( see also @xcite ) : @xmath47 where we used @xmath48 .
this equation has two complex conjugate roots , one is negative and the positive root , suitable for @xmath46 only , is given below : @xmath49 where @xmath50^{-1/3}+\frac{[s+(s^2 + 54q^3)^{1/2}]^{1/3 } } { 2^{1/3}3\delta^2 } , \\
q=2^{1/3}\delta^2(3 + 4\mid a\mid ^2 ) , % \hspace{1 mm } s=27\delta^2[1-\delta^2(1+\mid a\mid ^2 ) ] .
\nonumber \end{gathered}\ ] ] for clarity , this function is depicted in fig .
[ gamma ] for both cases with and without the rr force clearly indicating lowering of the energy of the electron due to its radiation losses at @xmath51 . similar way we rewrite eqs .
( [ wave ] ) and ( [ puisson ] ) arriving at @xmath52a=0,\label{a}\\ \phi''=n_0(n_e-1).\label{phi}\end{gathered}\ ] ] thus the set of equations ( [ force ] ) , ( [ a ] ) , ( [ phi ] ) together with eq .
( [ gama ] ) represent the full set of self - consistent ordinary equations describing plasma - field structures that can be realised during ultra - relativistic laser plasma interaction .
it should be noted that ion dynamics is neglected here .
although the main reason is simplicity , at extreme field intensities in plasmas as high as 10@xmath0 w@xmath2 electron dynamics , as follows from simulations , is much faster than ion dynamics . summarizing the above consideration
we underline that in comparison with the conventional relativistic interaction ( @xmath53 ) , the set of equations ( [ force ] ) , ( [ gama ] ) - ( [ phi ] ) has two fundamentally new features .
the first one follows from eq .
( [ force ] ) describing the laser - plasma equilibrium when the quasistatic force due to charge separation ( left side ) compensates the action from the laser wave comprising the ponderomotive force ( the first term in the right - hand side ) and a new one , dissipative force arising from the laser energy absorption in the rr processes ( the second term ) . in the case of single electron motion the latter force is responsible for electron acceleration due to the rr effect ( see , e.g. , @xcite ) .
it should be noted that the right - hand side of eq .
( [ force ] ) is an exact expression for ponderomotive and dissipative forces arising from the laser fields acting onto the plasma .
the second feature is in the wave equation ( [ a ] ) in which there appeared an imaginary part , which indicates the importance of absorption processes due to the rr effects .
let us now apply these equations to particular problems of interest .
to begin with , we consider a very instructive example of a homogeneous plasma ( @xmath54 ) . in this case , considering equation ( [ a ] ) for the laser field only and assuming @xmath55 we can obtain the standard nonlinear dispersion relation @xmath56 ( where @xmath57 is the wavevector of a plane electromagnetic wave ) with the following expression for dielectric permittivity @xmath58 which contains real and imaginary parts . at @xmath59 , as is expected , the imaginary part goes to zero whereas the real one reduces to the conventional relativistic dielectric permittivity @xmath60 @xcite .
the qualitative contribution of the rr effects is in the imaginary part , i.e. , in the plasma conductivity @xmath61 that appeared due to laser energy absorption caused by the re - emission of photons . for estimates ,
the imaginary part is comparable with the plasma contribution to the real part ( @xmath62 ) at @xmath63 , i.e. , at @xmath64 . by introducing a single particle scattering cross - section through its definition as the energy absorbed in a volume unit , @xmath65 , being equal to the incident flux energy , @xmath66 , multiplying by @xmath67 , where @xmath20 is the number of scatters , i.e. , electrons
, we can easily obtain an exact expression for the relativistic tompson scattering cross - section @xmath68 : @xmath69 where @xmath70 ( @xmath71 is the electron radius ) is the tompson scattering cross - section . if in the @xmath72 range we obtain the well - known formula @xmath73 , whereas for a much higher amplitude @xmath74 the scattering cross - section decreases as @xmath75 .
the next also very instructive example is the interaction of electromagnetic waves with overdense plasmas , which are of interest from fundamental point of view and for applications .
the most striking effect is the so - called relativistic self - induced transparency ( rsit ) effect when an incident relativistically strong electromagnetic wave is able to penetrate deep into overdense plasmas , which are otherwise opaque for low intensity waves @xcite . to study this effect at ultra - relativistic intensities at which the rr effects should be taken into account
we will use eqs .
( [ force ] ) , ( [ gama ] ) - ( [ phi ] ) where we assume @xmath76
. then equation ( [ a ] ) for amplitude splits into two equations for the real functions @xmath77 and @xmath78 @xmath79u=0 , \label{u}\\ ( u^2\theta')'=\delta\gamma\frac{n_0n_e u^2}{1+\delta^2 \gamma^6 } , \label{flux}\end{gathered}\ ] ] which constitute together with eqs .
( [ force ] ) , ( [ gama ] ) and ( [ phi ] ) a self - consistent set of ordinary equations .
it is useful to note that eq .
( [ flux ] ) , expressing the electromagnetic energy conservation law , describes the energy flux decreasing along the propagation path due to radiation losses .
this set of equations can be solved numerically assuming for the case of semi - infinite plasma that at @xmath80 the system is unperturbed , i.e. , @xmath81 , and @xmath82 @xcite .
an example of the plasma - field structures which describe the plasma electron compression and formation of nonlinear skin - layer distribution for the laser field is shown in fig .
[ semi](a ) . an auxiliary curve ( brown dotted line ) , that is the difference of the forces @xmath83 acting on a probe electron from the laser and quasistatic field due to charge separation , help us to understand the physics behind the rsit effect .
in fact , at increasing incident intensities this difference of the forces acting on boundary electrons is positive and all electrons are pushed deeper into plasma .
however , the point where @xmath83 is equal to zero is moving to the boundary point . in the limiting case when this point exactly coincides with the boundary point corresponds to the threshold of the rsit effect @xcite .
indeed , at higher incident intensities exceeding this threshold , skin - layer solutions do not exist any longer .
moreover , boundary electrons are in unstable position , since small displacements from their position make the difference of the forces acting on them negative , which means that they will move towards the incident wave , thus allowing the wave to penetrate deeper into the plasma . in fig .
[ semi](b ) we present the rsit threshold as a function of plasma density with and without the rr effects that shows a decrease in the threshold when the friction force is taken into account .
physically this occurs due to arising of an additional , dissipative force [ second term in the right - hand side of eq .
( [ force ] ) ] giving the possibility to support such structures with lower incident intensities . as the rr effects at ultrahigh laser intensities
are directly connected with the generation of high energy photons , the very intriguing possibility is to use them in order to create very efficient converter of laser energy into gamma - rays . however , as shown in simulations @xcite and also clearly seen in fig .
[ semi](c ) at incident intensities high enough to generate high energy photons by a single electron , the laser field inside the plasma is much smaller than the incident one .
this happens due to the electrodynamic effect of laser interaction with plasma , i.e. , electron peaking by laser compression strongly reflects the incident wave resulting in appreciable decrease of electromagnetic fields in the plasma as shown in fig .
[ semi](c ) , thereby strongly suppressing the rr effects .
however , this obstacle might be overcome by using thin plasma layers ( thin foils ) and symmetrical irradiation from both sides .
first , we should optimize foil parameter so that the laser field should be able to penetrate deep into electron region and , second , symmetrical irradiation should arrest longitudinal electron motion . to do so , we will solve numerically eqs .
( [ force ] ) , ( [ gama ] ) , ( [ phi]),([u ] ) , and ( [ flux ] ) in the class of symmetrical solutions , i.e. , assuming that @xmath84 @xmath85 @xcite .
the typical plasma - field structure realized during the interaction is shown in fig .
[ foil](a ) that models the symmetrical irradiation of gold foil ( @xmath86 ) of 500 nm width .
we see that in spite of high incident intensity of @xmath87 w@xmath2 the field in the electron region is relatively small ( @xmath88 )
. strong reflection of the incident wave from the compressed electron layer ( compression factor is about 75 and the maximum electron density increases by several orders of magnitude reaching @xmath89 @xmath90 ) greatly suppresses the rr effects , thus gamma - rays are not generated in fact .
whereas by carefully choosing foil parameters , such as plasma density and width with respect to the given laser intensity , plasma - field structures can be favourably modified as depicted in fig . [ foil](b ) .
in this case , for example , for an incident intensity of @xmath91 w@xmath2 and the same foil parameter as in fig .
[ foil](a ) , being a single hump the electron layer experiences strong enough electromagnetic field for the rr effects to come into play .
in fact , electron layer compressed by a factor of @xmath92 , reaching a very narrow width of @xmath93 and density of @xmath94 @xmath90 , is able to convert 45% of incident energy into high energy photons of 45 mev .
w@xmath2 and ( b ) @xmath95 w@xmath2.,scaledwidth=35.0% ] based on the above results we may propose the following scheme of very efficient laser energy conversion into gamma - rays .
a thin foil target , preferably consisting of heavy ions , should be irradiated symmetrically from opposite directions by counter - propagated laser pulses with ultra - relativistic intensities of about @xmath96-@xmath97 w/@xmath2 or more . by properly choosing the foil thickness at the given incident intensity
, all compressed electrons will synchronically rotate in a narrow plane and produce synchrotron emission along instantaneous velocity .
the diagram of radiation , rotating within this plane too , can also be very narrow in proportion to @xmath98 .
such a very important issue of the gamma - ray source , which is essentially two - dimensional , can be used to collect high energy photons for applications by surrounding the gamma - ray radiated narrow electron layer by respective consumers as shown schematically in fig .
[ proposal](a ) . to show the efficiency of the method proposed
we have performed a parametric scan over a range of foil parameters and laser intensities and calculated the energy absorbed due to rr effects over the electron layer as follows from eq .
( [ a ] ) as @xmath99 , where @xmath100 is defined by eq .
( [ epsilon ] ) . in fig .
[ proposal](b ) we present the conversion efficiency defined as @xmath101 ( @xmath102 is the incident intensity ) as a function of the width of a gold foil that shows the maximum efficiency of 50% , which means that half of the laser energy can be converted into gamma - rays .
in summary , we have studied the fundamental issues of ultra - relativistic laser - plasma interactions when the rr effects can play a crucial role .
we paid particular attention to the case of circularly polarized lasers when electromagnetic forces acting along the propagation direction steadily shift electrons forwards , thus making the quasistationary approach suitable for analytical treatment .
it allows us to give fundamental definitions such as nonlinear dielectric permittivity , ponderomotive and dissipative forces acting on a plasma in exact form .
we have also determined plasma - field structures arising during ultra - relativistic interactions and shown that in optimal conditions of laser - foil interaction about 50% of laser energy may be converted into gamma - rays .
99 a. ringwald , phys .
b 5 * 10 * , 107 ( 2001 ) .
f. amiranoff , f. auge , h. backe , ph .
balcou , d .- b .
blaschke , d. bernard , j .-
brasile , s. bulanov , f. canova , j .-
chambaret , d. charalambidis , g. cheriaux , g. dollinger , f. druon , r. ferrand , n. forget , p. georges , f. gruener , d. habs , f. hannachi , k. hatsagortsyan , d. hulin , s. karsch , k. kompa , f. krausz , k. kushelnick , c. le blanc , k. ledingham , g. malka , v. malka , j. meyer - ter - vehn , p. mora , g .- a .
mourou , n. naumova , f. pegoraro , g. petite , a. piskarskas , m. pittman , f. ple , a. pukhov , m. roth , r. sauerbrey , a. rousse , d. schardt , j. schreiber , s. sebban , f. sette , r. schuetzhold , g. tsakiris , v. tikhonchuck , j. ullrich , h. videau , m. vrakking , o. willi , proposal for a european extreme light infrastructure ( eli ) , http://www.extreme-light-infrastructure.eu .
mourou , t. tajima , and s.v .
bulanov , rev .
78 * , 309 ( 2006 ) .
f. rohrlich , _ clasical charged particles _ ( addison - wesley , reading , ma , 1965 ) .
landau and e.m .
lifshitz , _ the classical theory of fields _ ( pergamon , new york , 1994 ) . a. di piazza ,
phys . * 83 * , 305 ( 2008 ) .
a. di piazza , k.z .
hatsagortsyan , c.h .
keitel , phys .
* 102 * , 254802 ( 2009 ) ; i.v .
sokolov , n.m .
naumova , j.a .
nees , g.a .
mourou , and v.p .
yanovsky , phys .
plasmas * 16 * , 093115 ( 2009 ) ; i.v .
sokolov , j.a .
nees , v.p .
yanovsky , n.m .
naumova , and g.a .
mourou , phys .
e * 81 * , 036412 ( 2010 ) .
a. gonoskov , a. bashinov , i. gonoskov , c. harvey , a. ilderton , a. kim , m. marklund , g. mourou , a. sergeev , arxiv:1306.5734 a. zhidkov , j. koga , a. sasaki , and m. uesaka , phys . rev . lett . * 88 * , 185002 ( 2002 ) . n. naumova , t. schlegel , v.t .
tikhonchuk , c. labaune , i.v .
sokolov , and g. mourou , phys .
* 102 * , 025002 ( 2009 ) .
m. tanburini , f. pegoraro , a. di piazza , c.h .
keitel and a. macchi , new j. phys .
* 12 * , 123005 ( 2010 ) ; m. tamburini , f. pegoraro , a. di piazza , c.h .
keitel , t.v .
liseykina , a. macchi , nucl .
methods phys .
a * 653 * , 181 ( 2011 ) .
ridgers , c.s .
brady , r. duclous , j.g .
kirk , k. bennett , t.d .
arber , a.p.l .
robinson , and a.r .
bell , phys .
* 108 * , 165006 ( 2012 ) ; c.s .
brady , c.p .
ridgers , t.d .
arber , a.r .
bell , and j.g .
kirk , phys .
109 * , 245006 ( 2012 ) ; t. nakamura , j.k .
koga , t.zh .
esirkepov , m. kando , g. korn , and s.v .
bulanov , phys .
108 * , 195001 ( 2012 ) .
j. d. jackson , classical electrodynamics , 3rd ed .
( wiley , new york , 1999 ) . b.b .
kadomtsev , zh .
* 33 * , 151 ( 1957 ) [ sov .
jetp * 6 * , 531 ( 1958 ) ] .
a. ilderton and g. torgrimsson , phys .
d * 88 * , 025021 ( 2013 ) .
bulanov , t.zh .
esirkepov , m. kando , j.k .
koga , and s.s .
bulanov , phys .
e * 84 * , 056605 ( 2011 ) .
hazeltine and s.m .
mahajan , phys .
e * 70 * , 046407 ( 2004 ) .
berezhiani , s.m.mahajan , and z. yoshida , phys .
e * 78 * 066403 ( 2008 ) .
berezhiani , r.d .
hazeltine and s.m .
mahajan , phys .
e * 69 * 056406 ( 2004 ) .
a.i . akhiezer and r.v .
polovin , sov .
jetp * 3 * , 696 ( 1956 ) .
sarachik and g.y .
schappert , phys .
d * 1 * , 2738 ( 1970 ) .
marburger and r.f .
tooper , phys .
lett . * 35 * , 1001 ( 1975 ) .
a. kim , f. cattani , d. anderson , m. lisak , jetplett .
* 72 * , 5 ( 2000 ) ; m. tushentsov , a. kim , f. cattani , d. anderson , and m. lisak , phys .
87 * , 275002 ( 2001 ) .
kozlov , a.g .
litvak , and e.v .
suvorov , sov .
jetp * 49 * , 75 ( 1979 ) .
kaw , a. sen , and t. katsouleas , phys .
* 68 * , 3172 ( 1992 ) ; d. farina and s.v .
bulanov , phys .
* 86 * , 5289 ( 2001 ) .
o. klimo , j. psikal , and j. limpouch , phys . rev .
beams * 11 * , 031301 ( 2008 ) ; a.p .
robinson , m. zepf , s. kar , r.g .
evans and c. bellei , new j. phys . *
10 * , 013021 ( 2008 ) ; x.q .
yan , c. lin , z.m .
sheng , z.y .
guo , b.c .
liu , y.r .
lu , j.x .
fang , and j.e .
chen , phys .
lett . * 100 * , 135003 ( 2008 ) ; t.v .
liseykina , m. borghesi , a. macchi and s. tuveri , plasma phys .
controlled fusion * 50 * , 124033 ( 2008 ) .
korzhimanov , a.a .
gonoskov , a.v .
kim , a.m. sergeev , jetp lett . * 86 * , 577 ( 2008 ) ; a.a .
gonoskov , a.v .
korzhimanov , v.i .
eremin , a.v .
kim , and a.m. sergeev , phys .
102 * 184801 ( 2009 ) .
zeldovich , phys .
usp . * 18 * , 79 ( 1975 ) ; s.v .
bulanov , t.zh .
esirkepov , j. koga , t. tajima , plasma phys . rep . * 30 * , 196 ( 2004 ) . s.s .
bulanov , t.zh .
esirkepov , a.g.r .
thomas , j.k .
koga , and s.v .
bulanov , phys .
lett . * 105 * , 220407 ( 2010 ) .
fradkin , phys .
* 42 * , 1209 ( 1979 ) .
eremin , a.v .
korzhimanov , and a.v .
kim , phys . of plasmas
* 17 * , 043102 ( 2010 ) .
korzhimanov , v.i .
eremin , a.v .
kim , m.r .
tushentsov , jetp * 105 * , 675 ( 2007 ) . | a simple electrodynamic model is developed to define plasma - field structures in self - consistent ultra - relativistic laser - plasma interactions when the radiation reaction effects come into play .
an exact analysis of a circularly polarized laser interacting with plasmas is presented .
we define fundamental notations such as nonlinear dielectric permittivity , ponderomotive and dissipative forces acting in a plasma .
plasma - field structures arising during the ultra - relativisitc interactions are also calculated .
based on these solutions we show that about 50% of laser energy can be converted into gamma - rays in the optimal conditions of laser - foil interaction . |
quantum monte carlo methods have proved to be very accurate and powerful tools for studying the structure of light nuclei with realistic two- and three - nucleon interactions . in a series of papers ,
we have calculated about 40 ground and low - lying excited state energies of different @xmath2 quantum numbers in @xmath4 = 610 nuclei with an accuracy of @xmath5 12% @xcite .
the first step is a variational monte carlo ( vmc ) calculation to find an optimal trial function , @xmath0 for a given state .
the @xmath1 , which is antisymmetric by explicit construction , is then used as input to a green s function monte carlo ( gfmc ) algorithm , which projects out the lowest - energy eigenstate by a propagation in imaginary time , @xmath6 .
the algorithm preserves the quantum numbers of @xmath1 , although it may introduce symmetric noise into the propagated @xmath7 . in principle , @xmath7 for large @xmath6 approaches the exact lowest - energy eigenstate with the specified quantum numbers .
vmc calculations of second and higher excited @xmath2 states have also been made , because a major step in the preparation of an optimal @xmath1 is a diagonalization in the small single - particle basis of different possible spatial - symmetry states @xcite .
gfmc calculations of second or higher @xmath2 states have not been attempted previously in nuclear physics , under the expectation that any small contamination of the excited trial state by the true first state would drive the calculated energy below its proper value . however , as is described in this paper , we have found that it is possible to obtain useful gfmc predictions of multiple states with the same quantum numbers . in this paper
we report such gfmc calculations of second and higher @xmath2 states in light ( @xmath4 = 68 ) p - shell nuclei .
vmc calculations of these states were reported previously @xcite , but we have made significant improvements in the trial functions since then .
details of the recent vmc work are discussed in sec .
ii . for the gfmc calculations , we find that the propagated energy , @xmath8 , for many states is stable and a useful excitation energy can be extracted .
gfmc calculations for ground states are reviewed in sec .
iii , where we also describe an improved algorithm for propagation with three - nucleon interactions that is both faster and more accurate than what we had used previously .
the gfmc algorithm for higher excited states is described in sec .
iv , along with various orthogonality tests .
numerical results are given in sec .
v for the argonne v@xmath3 ( av18 ) and simplified v@xmath9 ( av8@xmath10 ) two - nucleon ( @xmath11 ) interactions @xcite , and for av18 with either the urbana ix ( uix ) or illinois-2 ( il2 ) three - nucleon ( @xmath12 ) interactions @xcite added .
we find , consistent with our studies of first @xmath2 states , that the av18/il2 hamiltonian gives a good representation of the experimental spectrum ; the rms deviation from 36 experimental energies with @xmath13 is only 0.60 mev .
some concluding remarks are given in sec .
the vmc trial function , @xmath14 , for a given nucleus , is constructed from products of two- and three - body correlation operators acting on an antisymmetric single - particle state with the appropriate quantum numbers .
the correlation operators reflect the influence of the interactions at short distances , while appropriate boundary conditions are imposed at long range .
the @xmath1 contains variational parameters that are adjusted to minimize the energy expectation value , @xmath15 , which is evaluated by metropolis monte carlo integration .
a good variational trial function has the form @xmath16 \left [ { \mathcal s}\prod_{i < j}(1+u_{ij } ) \right ] |\psi_j\rangle \ .
\label{eq : psit}\ ] ] the @xmath17 and @xmath18 are non - commuting two- and three - nucleon correlation operators , @xmath19 indicates a symmetric sum over all possible orderings , and @xmath20 is the fully antisymmetric jastrow wave function . for s - shell nuclei
the latter has the simple form @xmath21 \left [ \prod_{i < j}f_c(r_{ij } ) \right ] |\phi_a(jmtt_{3})\rangle \ .
\label{eq : jastrow}\ ] ] here @xmath22 and @xmath23 are central two- and three - body correlation functions and @xmath24 is a slater determinant in spin - isospin space , e.g. , for the @xmath25-particle , @xmath26 .
the correlation operator @xmath17 includes spin , isospin , and tensor terms : @xmath27 where the @xmath28 \otimes [ 1,{\bf\tau}_{i}\cdot{\bf\tau}_{j}]$ ] are the same static operators that appear in the @xmath11 potential .
the functions @xmath29 and @xmath30 are generated by solving a set of six coupled differential equations with embedded variational parameters @xcite : two single - channel equations in @xmath31s and @xmath31p waves , and two coupled - channel equations in @xmath32s and @xmath32p waves .
the @xmath18 has the spin - isospin structure of the dominant parts of the @xmath12 interaction as suggested by perturbation theory . for p - shell nuclei ,
the jastrow wave function , @xmath20 , is necessarily more complicated .
it starts with a sum over independent - particle terms , @xmath24 , that have 4 nucleons in an @xmath25-like core and @xmath33 nucleons in p - shell orbitals .
we use @xmath34 coupling to obtain the desired @xmath35 value of a given state , as suggested by standard shell - model studies @xcite .
we also need to specify the spatial symmetry @xmath36 $ ] of the angular momentum coupling of the p - shell nucleons @xcite .
different possible @xmath37 $ ] combinations lead to multiple components in the jastrow wave function .
this independent - particle basis is again acted on by products of central pair and triplet correlation functions , but now they depend upon the shells ( s or p ) occupied by the particles and on the @xmath37 $ ] coupling : @xmath38 \big[\prod_{i < j \leq 4 } f_{ss}(r_{ij})\big ] \sum_{ls[n ] } \big ( \beta_{ls[n ] } \big[\prod_{k \leq 4<l \leq a } f^{ls[n]}_{sp}(r_{kl})\big ] \right .
\nonumber\\ & \times & \left .
\big[\prod_{4<l < m \leq a } f^{ls[n]}_{pp}(r_{lm})\big ] |\phi_{a}(ls[n]jmtt_{3})_{1234:5\ldots
a}\rangle \big ) \right\ } \ .
\label{eq : psip}\end{aligned}\ ] ] the operator @xmath39 indicates an antisymmetric sum over all possible partitions of the @xmath4 particles into 4 s - shell and @xmath33 p - shell ones .
the pair correlation for both particles within the s - shell , @xmath40 , is similar to the @xmath41 of the @xmath25-particle .
the pair correlations for both particles in the p - shell , @xmath42}_{pp}$ ] , and for mixed pairs , @xmath42}_{sp}$ ] , are similar to @xmath40 at short distance , but their long - range structure is adjusted to give appropriate clustering behavior , and they may vary with @xmath37 $ ] .
the single - particle wave functions @xmath24 are given by : @xmath43jmtt_{3})_{1234:5\ldots a}\rangle & = & |\phi_{4}(0 0 0 0)_{1234 } \big[\prod_{4 < l\leq a } \phi^{ls[n]}_{p}(r_{\alpha l})\big ] \\ & \times & \left\ { \big [ \prod_{4 < l\leq a } y_{1m_l}(\omega_{\alpha l } ) \big]_{lm_l[n ] } \big [ \prod_{4 < l\leq a } \chi_{l}(\frac{1}{2}m_s ) \big]_{sm_s } \right\}_{jm } \big [ \prod_{4
< l\leq a } \nu_{l}(\frac{1}{2}t_3 ) \big]_{tt_3}\rangle \nonumber \ .\end{aligned}\ ] ] the @xmath44}_{p}(r_{\alpha l})$ ] are p - wave solutions of a particle in an effective @xmath45 potential that has woods - saxon and coulomb parts .
they are functions of the distance between the center of mass of the @xmath25 core and nucleon @xmath46 , and may vary with @xmath37 $ ] .
the depth , width , and surface thickness of the single - particle potential are additional variational parameters of the trial function .
the overall wave function is translationally invariant , so there is no spurious center of mass motion .
the @xmath47}$ ] mixing parameters of eq .
( [ eq : psip ] ) are determined by a diagonalization procedure , in which matrix elements @xmath48 are evaluated using trial functions @xmath49 .
although the @xmath50jmtt_3)$ ] are orthogonal due to spatial symmetry , the pair and triplet correlations in @xmath1 make the different @xmath37 $ ] components nonorthogonal , so a generalized eigenvalue routine is necessary to carry out the diagonalization .
.@xmath47}$ ] components for @xmath51he states , listed in order of increasing excitation .
( l = s or d as appropriate . ) [ cols= " < , > , > " , ] [ table : vijk ] extensive breakdowns of the total energies were presented in ref .
the improved treatment of @xmath52 in the gfmc propagation has resulted in significant changes to the expectation values of @xmath52 and its components for the av18/il2 hamiltonian ; revised values are shown in table [ table : vijk ] .
the other contributions to the total energies shown in ref .
@xcite are not significantly changed .
the isovector and isotensor energies presented there have also not significantly changed in the current calculations .
tables [ table : energies-67 ] and [ table : energies-8 ] contain energies of a few @xmath53=1 states for @xmath51li and @xmath54be ; these have been computed perturbatively using the isovector and isotensor energies computed for @xmath51he and @xmath54li , respectively .
recently we have realized that there is a significant difficulty in extracting precise rms radii from gfmc calculations , especially for weakly bound systems .
for this reason we are deferring presenting updated values for the rms radii .
we have demonstrated that it is possible to use gfmc to compute the energies of multiple nuclear states with the same quantum numbers .
this substantially increases the number of nuclear level energies that can be compared to experimental values in the light p - shell region .
the av18/il2 hamiltonian presented in ref .
@xcite gives a good description of this increased set of energies .
we are currently extending these calculations to higher excited states in the @xmath4=9,10 nuclei .
we have also improved our treatment of the three - nucleon force in the gfmc calculations .
this has altered the detailed breakdown of @xmath12 contributions as shown in table [ table : vijk ] .
we expect a similar alteration of these terms in the @xmath4=9,10 nuclei , but the total energies should stay within 1 - 2% of our previously reported values @xcite .
we thank d. kurath , v. r. pandharipande , and k. varga for useful discussions .
the many - body calculations were performed on the parallel computers of the laboratory computing resource center , argonne national laboratory and the national energy research scientific computing center . the work of scp and rbw is supported by the u. s. department of energy , office of nuclear physics , under contract no .
w-31 - 109-eng-38 .
the work of jc is supported by the u. s. department of energy under contract no . | a variational monte carlo method is used to generate sets of orthogonal trial functions , @xmath0 , for given quantum numbers in various light p - shell nuclei .
these @xmath1 are then used as input to green s function monte carlo calculations of first , second , and higher excited @xmath2 states .
realistic two- and three - nucleon interactions are used .
we find that if the physical excited state is reasonably narrow , the gfmc energy converges to a stable result . with the combined argonne v@xmath3 two - nucleon and illinois-2 three - nucleon interactions , the results for many second and higher states in @xmath4 = 68 nuclei are close to the experimental values |
while it is accepted that electron correlation drives non - s - wave - pairing superconductivity , determination of detailed superconducting gap structure still remains a challenge for both theory and experiment . on a theoretical side , analyses starting at microscopic hamiltonians have succeeded in identifying gap structures of several materials . on an experimental side
, it is getting recognized that field - orientational dependence of specific heat or thermal conductivity carries a considerable information on gap structure . since the low - energy density of states changes according to the angle between gap nodes and magnetic field , one can derive gap structure by tracing a change of thermodynamic quantities as sweeping magnetic field in several directions .
actually , this method has been applied to a series of superconductors including sr@xmath7ruo@xmath8@xcite , cecoin@xmath9@xcite , @xmath10(et)@xmath7cu(ncs)@xmath7@xcite , yni@xmath7b@xmath7c@xcite , pros@xmath8sb@xmath11@xcite and upd@xmath7al@xmath12@xcite . however , their results left serious problems yet to be overcome .
for example , in the case of sr@xmath7ruo@xmath8 , thermal conductivity ( @xmath13 ) does not show any change with rotating magnetic field within the ab - plane , while specific heat ( @xmath14 ) shows clear four - fold oscillation . as to cecoin@xmath9 , @xmath13 and
@xmath14 show incompatible behavior . in yni@xmath7b@xmath7c , both @xmath13 and @xmath14
show a cusp - like singularity as sweeping magnetic field within the superconducting plane . in the previous paper
, we addressed the problem of sr@xmath7ruo@xmath8 in terms of the fine - structure of density of states@xcite . whereas , the latter two problems are left unsolved .
while it is proposed that the cusp - like singularity in yni@xmath7b@xmath7c is caused by point node in gap structure@xcite , their discussion is phenomenological in nature and unreliable , as we will show . the confusion as to the experiment is probably attributed to a lack of firm theoretical basis .
so far , the experimental data has been usually interpreted in terms of the `` doppler - shift criterion '' .
this criterion says _ larger density of states is obtained when a magnetic field is applied in antinodal rather than nodal directions .
_ however , this criterion is obtained on the basis of the phenomenological doppler - shift method that can be justified only at a limit of low - magnetic field .
hence , it is not clear whether this criterion is still valid in the experimentally accessible higher magnetic fields .
in fact , in the previous paper , we showed that there occurs a crossover from doppler - shift predominant region at low magnetic fields to core states predominant region at high magnetic fields@xcite . in the core state predominant region , larger density of states is obtained for field in nodal direction , in contrast to the `` doppler - shift criterion '' .
hence , in order to derive a correct nodal structure from the field - rotational experiments , it is highly desirable to establish a theoretical analysis applicable more widely . to this end , we present a more quantitative analyses of the crossover behavior by taking account of the field - angle dependence of order parameter .
in particular , we are interested in how the fermi surface or superconducting gap structures affect the crossover behavior . furthermore
, we would like to investigate field - orientational dependence of zero - energy density of states ( fodos ) in detail at the doppler - shift predominant region .
it is interesting to clarify how information on nodal structure is obtained from the shape of fodos .
the structure of this paper is as follows . in the next section ,
we summarize our method . in section [ isotropic ]
, we present a quantitative study of the crossover behavior for several models with different fermi surface and nodal structures . in section [ inplane ]
, we examine the effects of in - plane anisotropy of fermi surface to the crossover behavior . in section [ carbide ]
we address the problem of cusp - singularity in yni@xmath7b@xmath7c . and section [ conclusion ] is devoted to conclusions .
in this section , we introduce an approximate analytical solution of the quasiclassical eilenberger equation , which was originally proposed by pesch@xcite , and recently extended to unconventional superconductors by dahm et al.@xcite . with this method ,
one can obtain exact results at @xmath15 , and quantitatively reliable results to much lower fields .
all our analyses in this paper are based on this approximate analytical approach .
we start with the quasiclassical eilenberger equation@xcite .
this equation provides a convenient method to analyze inhomogeneous state of superconductors by decoupling slowly varying order parameter from fast oscillating fermi - particle degree of freedom . in this paper , we limit ourselves to the case of spin - singlet superconductors in clean limit . under these conditions ,
the quasiclassical eilenberger equations read @xmath16 @xmath17 with the normalization condition @xmath18 here , @xmath19 , @xmath20 and @xmath21 are normal and anomalous components of quasiclassical green function , @xmath22 is matsubara frequency , and @xmath23 denotes order parameter .
the order parameter @xmath24 is determined in a self - consistent way , by solving the following gap equation .
@xmath25 here , @xmath26 is zero - energy density of states ( zedos ) in the normal state , @xmath27 , a cut - off energy , @xmath28 , a paring potential . and
@xmath29 means taking average on the fermi surface , namely , @xmath30 .
free energy @xmath31 can also be expressed with the quasiclassical green function in a closed form@xcite .
@xmath32 with @xmath33 , free energy in the normal state , @xmath34 , external magnetic field , and @xmath35 , total magnetic field .
equations ( [ eilen1])-([free ] ) form a basis to study inhomogeneous superconductors .
next , we introduce an approximate analytical solution of eilenberger equation
. equations ( [ eilen1])-([eilennorm ] ) can be solved analytically only at @xmath15 .
however , by extending the analytical solution to lower fields , one can obtain quantitatively reliable results for density of states etc , well below @xmath36 . to obtain approximate analytical solution at arbitrary fields
, we have to make three assumptions which are exactly justified at @xmath15 .
first , we decouple the order parameter into the spatial and momentum parts @xmath37 and normalize @xmath38 and @xmath39 as @xmath40 @xmath41 here , @xmath42 means averaging in the real space . if @xmath38 is periodic , this is equivalent to taking average over a unit cell .
then , we assume the spatial variation @xmath38 is described by abrikosov vortex function , namely , @xmath43 .
\label{abrikosov}\end{aligned}\ ] ] @xmath44 can be obtained by solving linearized ginzburg - landau equation with different coherence lengths in x ( @xmath45 ) and in y ( @xmath46 ) directions . here
, we set @xmath47 and @xmath48 axes perpendicular to the magnetic field .
@xmath49 is a normalization constant to make @xmath50 .
area of unit cell , @xmath51 , can be related to the total magnetic field @xmath52 by @xmath53 @xmath54 denotes the anisotropy of superconducting coherence lengths , i.e. , @xmath55 .
one can estimate @xmath54 to be @xmath56 for single - band superconductors . here
we note that @xmath38 denotes an abrikosov vortex lattice which is spanned by two lattice vectors , @xmath57 and @xmath58 .
when @xmath59 , @xmath38 describes a square ( rectangular ) lattice , while when @xmath60 , @xmath38 describes a triangular lattice . however , the shape of the unit cell is unimportant in the subsequent analyses . as a second assumption , we set ginzburg - landau parameter @xmath13 to be much larger than 1 , hence @xmath61 .
this assumption is justified for a lot of unconventional superconductors including yni@xmath7b@xmath7c .
the third assumption is that we ignore spatial fluctuation of @xmath62 , namely , @xmath63 .
since at @xmath15 , the order parameter vanishes and translational symmetry is recovered , this assumption is also justified .
the validity of this assumption has been justified for lower fields@xcite . with these three assumptions , one can obtain an approximate analytical solution of eqs .
( [ eilen1])-([eilennorm ] ) at arbitrary magnetic fields by utilizing the operator techniques introduced in ref . .
we summarize the basic results below . for details ,
see , for example , refs . and . 1
. spatial average of @xmath62 @xmath64 where , @xmath65 @xmath66 here , effective fermi velocity @xmath67 is expressed as @xmath68 with the projection of fermi velocity on x ( @xmath69 ) and y ( @xmath70 ) axes .
dos @xmath71 @xmath72 in particular , the formula of zedos becomes quite simple .
@xmath73 3 .
gap equation + in order to obtain gap equation , we decouple @xmath74 as @xmath75 , then we obtain @xmath76 where @xmath77 is euler s constant .
the formula to determine @xmath36 can be obtained by setting @xmath78 in eq .
( [ peschgapeq ] ) .
@xmath79 4 .
free energy @xmath80 we calculate density of states by the following procedure . first , fermi velocity @xmath81 and nodal structure @xmath39 are determined by assuming a phenomenological model or starting at a microscopic hamiltonian .
then , at a given temperature @xmath82 and a magnetic field @xmath34 , one can calculate order parameter amplitude @xmath83 by solving eq .
( [ peschgapeq ] ) . by substituting @xmath83 into eq .
( [ pdos ] ) , one can obtain dos @xmath71 . when magnetic field is applied off a symmetry axis
, the ratio of coherence lengths @xmath54 deviates from 1 . in this case , we determine @xmath54 variationally by minimizing free energy eq .
( [ pfree ] ) with respect to @xmath54 .
this prescription is equivalent to the variational ansatz of abrikosov vortex lattice introduced in refs . and .
one can incorporate multi - band properties into this approximate analytical approach with a slight modification of [ 1]-[4]@xcite .
in multi - band systems , @xmath84 , @xmath81 , @xmath85 , @xmath86 , @xmath87 , @xmath24 , @xmath88 become band - dependent . to specify band index , we write these quantities as @xmath89 , @xmath90 , @xmath91 , @xmath92 , @xmath93 , @xmath94 , @xmath95 .
next , we decouple order parameter as @xmath96 here , we assume @xmath97 to be described by abrikosov lattice , @xmath44 .
furthermore , we assume that the anisotropy of coherence lengths @xmath98 are the same for all bands ( @xmath99 ) , and determine @xmath54 by minimizing free energy .
due to inter - band mixing effect by a pair - hopping term , the spatial variation of @xmath24 of each bands are not independent any longer .
then , one can take account of the inter - band mixing effect through minimization of @xmath54 . as we will discuss in section .
[ carbide ] , we are interested in a system where quasi - two - dimensional fermi surface and three - dimensional fermi surface co - exist . in such a system , superconducting coherence length may be isotropic due to the three - dimensional fermi surface , and lead to isotropic vortex structure . in this case , one may expect unusual phenomena characteristic of multi - band system , in the sense that quasi - two - dimensional fermi surface and isotropic vortex structure are incompatible in a single - band superconductor . here
, we note that some complication may arise in vortex structure due to multi - band effects .
for example , it is proposed that order parameter of different bands may have zero - points at different places .
we believe our results will not be influenced by the detail of vortex structure and ignore them . here ,
let us check how [ 1]-[4 ] are modified in a multi - band case .
[ 1]spatial average of @xmath62 also holds for multi - band systems , by making each quantity in eqs .
( [ pgreen])-([fermi ] ) band - dependent .
[ 2]dos @xmath71 and [ 4]free energy are also correct by making the quantities band - dependent , and summing over band indices in the right sides of eqs .
( [ pdos ] ) , ( [ pzedos ] ) , and ( [ pfree ] ) . only gap equation in [ 3 ] needs substantial modification . in a multi - band system
, we decouple pair potential @xmath100 as @xmath101 , then gap equation eq .
( [ peschgapeq ] ) becomes @xmath102
in this section , we present a quantitative study of the crossover behavior for the following three models .
below , we take @xmath103 , @xmath104 and @xmath105 as a right - handed orthonormal basis . 1 .
line nodes on spherical fermi surface @xmath106 @xmath107 2 .
line nodes on cylindrical fermi surface @xmath108 @xmath109 here , @xmath110 gives a small c - axis dispersion .
we set @xmath111 .
point nodes on spherical fermi surface @xmath106 @xmath112 these models are often used to approximate more complicated realistic materials . hence , it is important to study the basic properties of these models in detail . throughout our paper , we specify the direction of magnetic field with two parameters @xmath113 and @xmath114 as @xmath115 in principle , we fix @xmath116 and study the change of zedos with varying @xmath114 in this paper . in other words , we consider the ab - plane as a field - rotational plane . as to the three models above ,
@xmath117 and @xmath118 correspond to the cases where a magnetic field is applied in nodal direction and antinodal direction , respectively . here , we focus on zedos at low temperatures , because universal behaviors are expected in this region .
we note that some cares are necessary to discuss thermodynamic quantities at higher temperature , because dos at high energies depend on the magnetic field direction quite differently from zedos@xcite .
thus we fix @xmath119 and @xmath120 throughout this section . at this temperature , zedos well represents the thermodynamic properties of a system .
first , we consider zedos when magnetic field is applied in the nodal direction ( @xmath0 ) and in the antinodal direction ( @xmath1 ) . in the previous paper@xcite , we showed that the doppler - shift criterion is correct at low fields , while it is broken at high fields due to the quasiparticles propagating parallel to magnetic field . in the previous paper
, we used an assumption that @xmath83 does not depend on field direction . in the present paper , we calculate zedos with determining @xmath83 self - consistently .
in general , @xmath83 changes as field direction varies .
in particular , near @xmath36 , in - plane anisotropy of @xmath36 influences fodos through the field - angle dependence of @xmath83 .
perhaps , the anisotropy of @xmath36 continues to affect fodos to lower fields , and suppress the doppler - shift predominant region .
hence , it is of interest to estimate how the field - angle dependence of @xmath83 modifies the crossover behavior .
field dependence of zedos under magnetic field in the nodal and antinodal directions for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and ( c)point nodes on spherical fermi surface .
( inset ) field dependence of @xmath83 .
@xmath83 is normalized with t@xmath121 , and magnetic field is normalized with @xmath36 when a field is in the antinodal direction .
the agreement between ( a ) and the result of eilenberger approach ( fig . 4 of ref .
) is quite good.,title="fig:",scaledwidth=50.0% ] field dependence of zedos under magnetic field in the nodal and antinodal directions for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and ( c)point nodes on spherical fermi surface .
( inset ) field dependence of @xmath83 .
@xmath83 is normalized with t@xmath121 , and magnetic field is normalized with @xmath36 when a field is in the antinodal direction .
the agreement between ( a ) and the result of eilenberger approach ( fig . 4 of ref . ) is quite good.,title="fig:",scaledwidth=50.0% ] field dependence of zedos under magnetic field in the nodal and antinodal directions for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and ( c)point nodes on spherical fermi surface .
( inset ) field dependence of @xmath83 .
@xmath83 is normalized with t@xmath121 , and magnetic field is normalized with @xmath36 when a field is in the antinodal direction .
the agreement between ( a ) and the result of eilenberger approach ( fig . 4 of ref .
) is quite good.,title="fig:",scaledwidth=50.0% ] field dependence of @xmath122 and @xmath123 for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and ( c)point nodes on spherical fermi surface .
@xmath122 is normalized with t@xmath121 , and magnetic field is normalized with @xmath36 when a field is in the antinodal direction.,title="fig:",scaledwidth=50.0% ] field dependence of @xmath122 and @xmath123 for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and ( c)point nodes on spherical fermi surface .
@xmath122 is normalized with t@xmath121 , and magnetic field is normalized with @xmath36 when a field is in the antinodal direction.,title="fig:",scaledwidth=50.0% ] field dependence of @xmath122 and @xmath123 for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and ( c)point nodes on spherical fermi surface .
@xmath122 is normalized with t@xmath121 , and magnetic field is normalized with @xmath36 when a field is in the antinodal direction.,title="fig:",scaledwidth=50.0% ] in fig .
[ f_fodosamp ] , we show the fodos for the three models introduced above .
these show a similar crossover behavior with different crossover fields @xmath2 . for @xmath4 , @xmath124 ,
consistent with the `` doppler - shift criterion '' , while for @xmath6 , @xmath5 .
these results are consistent with our previous analyses@xcite .
since @xmath36 is smaller when field is in the nodal direction ( @xmath125 ) than in the antinodal direction ( @xmath126 ) in all the three cases , the anisotropy of @xmath36 generally supports the relation @xmath5 at high fields . in order to examine how the anisotropy of @xmath83 affects fodos at low fields , we plot @xmath127 in fig .
[ diff ] , together with the anisotropy of zedos , @xmath123 .
one can find that @xmath122 changes its sign near @xmath2 and becomes negative for @xmath128 .
this means that anisotropy of @xmath83 has the same sign with the relation @xmath3 in the doppler - shift predominant region . combining this with the effects of anisotropic @xmath36 at high magnetic fields
, it is shown that our previous result@xcite on the crossover behavior is not altered by taking account of the field - angle dependence of @xmath83 .
next , let us study individual cases .
first , let us compare the cases of line nodes on spherical fermi surface(fig .
[ diff](a ) ) and on cylindrical fermi surface(fig .
[ diff](b ) ) . in spite of the difference in dimensionality of fermi surfaces ,
both cases show similar behavior of @xmath0 and @xmath1 with very close @xmath2 .
in fact , the small c - axis dispersion of the cylindrical fermi surface plays a key role in this similarity . in eq .
( [ pzedos ] ) , information of fermi surface is included only in @xmath129 . in the case of cylindrical fermi surface , @xmath130 .
while @xmath110 is much smaller than 1 , one can not neglect the term @xmath131 . in the case of anisotropic fermi surface
, one has @xmath132 in the whole range of magnetic field , as a result of minimizing the free energy eq .
( [ free ] ) .
hence , the in - plane and the c - axis components of effective fermi velocity , @xmath133 and @xmath131 are scaled to be equivalent in magnitude .
next , we study the case of point nodes on spherical fermi surface ( fig .
[ diff](c ) ) . in this case
, one can find that the crossover field @xmath2 is smaller ( @xmath134 ) than the case of line nodes ( @xmath135 ) .
this difference in @xmath2 can be understood in terms of the shape of gap minima . to see this
, we note that the reversal of the doppler - shift criterion at high magnetic fields is caused by the quasiparticles propagating parallel to magnetic field@xcite . at high magnetic fields ,
contribution from such quasiparticles to zedos , @xmath136 , is smaller for @xmath126 than @xmath125 , resulting in the relation @xmath5 . here , for the point node case ,
@xmath137 is quadratic near gap minima ( for example , at @xmath138 ) , while it is linear in the line node cases .
therefore , the order parameter is smaller near gap minima in the point node case , leading to larger @xmath136 for @xmath125 . on the other hand
, there is not much difference in @xmath136 when @xmath126 . as a result , @xmath139 is larger in the point node case and continues to be positive to lower field .
one may suppose that the smaller @xmath2 can be attributed to the gap structure in the `` polar part''(@xmath140 ) .
since in the point node case , @xmath141 except for the `` equatorial part''(@xmath142 ) of the fermi surface , the effect of doppler - shift is weaker in the `` polar part '' , and may lead to the smaller @xmath2 .
however , this does not hold .
in fact , if we replace the gap function eq .
( [ gap_sph_p ] ) with @xmath143 , i.e. , a point node model which varies linearly near gap minima , we find that @xmath135(fig .
[ diff_sharp ] ) , comparable to the line node cases , although @xmath137 is larger than the quadratic point node model ( eq .
( [ gap_sph_p ] ) ) everywhere on the fermi surface .
field dependence of @xmath122 and @xmath123 for spherical fermi surface with point nodes .
gap function is given by @xmath144 .
, scaledwidth=50.0% ] in this subsection , we study not only the amplitude but also the whole field - orientational dependence of zedos . in particular
, the shape of minima is of our interest , since we will discuss the origin of cusp - like minima which has been attributed to a point node in the doppler - shift method@xcite . in fig .
[ f_fodos ] , we show our results for the three models .
these calculations are done at @xmath145 .
one can find similar fodos for the line node cases ( fig .
[ f_fodos](a),(b ) ) with a slight difference in amplitudes .
whereas in the point node case ( fig .
[ f_fodos](c ) ) , one obtains fodos with broader minima than the line node cases .
furthermore , in the point node case , tiny maxima appear when magnetic field is applied right in the nodal direction ( @xmath117 ) .
these features again reflect the quadratic variation of order parameter near the gap minima . for comparison ,
we show fodos for the gap function @xmath144 in fig .
[ sharp_p ] , where one can find narrower minima similar to the case of line nodes . in both cases , our calculation does not support the results of doppler - shift method .
field - orientational dependence of zedos for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and ( c)point nodes on spherical fermi surface.,title="fig:",scaledwidth=50.0% ] field - orientational dependence of zedos for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and ( c)point nodes on spherical fermi surface.,title="fig:",scaledwidth=50.0% ] field - orientational dependence of zedos for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and ( c)point nodes on spherical fermi surface.,title="fig:",scaledwidth=50.0% ] field - orientational dependence of zedos for ( a)spherical fermi surface with point nodes .
gap function is given by @xmath144.,scaledwidth=50.0% ] for the better understanding of the features in fodos , we plot contribution to fodos from several parts of the fermi surface in fig .
[ fermi_div ] . from these figures
, one can find that fodos is determined by the contribution from @xmath146 on the spherical fermi surface or @xmath147 on the cylindrical fermi surface .
in other words , most contribution comes from the part where effective fermi velocity @xmath67(eq .
( [ fermi ] ) ) is parallel to the field - rotational plane .
one can coherently understand the features of fodos and @xmath2 obtained in this section from this viewpoint .
first , the difference between point node and line node is unimportant , because these gap structures behave similarly at @xmath148 , where @xmath67 is parallel to the field - rotational plane .
the dimensionality of fermi surface is also unimportant .
while fermi velocity is almost parallel to the field - rotational plane on a cylindrical fermi surface , the effective fermi velocity is so only near @xmath147 , due to small @xmath54(@xmath149 ) . whereas it is important whether order parameter varies linearly or quadratically near gap minima , because this difference lies in @xmath146 . in this sense , @xmath2 _ and
fodos are determined by the local structure of order parameter on the part where _
@xmath150 _ is parallel to the field - rotational plane , rather than the global nodal structure of the entire fermi surface_. contribution to fodos from several parts of the fermi surface for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and(c)point nodes on spherical fermi surface . in these three cases ,
total fodos is given by averaging over @xmath151 ( @xmath152 ) or @xmath153 ( @xmath154).,title="fig:",scaledwidth=50.0% ] contribution to fodos from several parts of the fermi surface for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and(c)point nodes on spherical fermi surface .
in these three cases , total fodos is given by averaging over @xmath151 ( @xmath152 ) or @xmath153 ( @xmath154).,title="fig:",scaledwidth=50.0% ] contribution to fodos from several parts of the fermi surface for ( a)line nodes on spherical fermi surface , ( b)line nodes on cylindrical fermi surface , and(c)point nodes on spherical fermi surface .
in these three cases , total fodos is given by averaging over @xmath151 ( @xmath152 ) or @xmath153 ( @xmath154).,title="fig:",scaledwidth=50.0% ]
in this section , we examine the effects of in - plane anisotropy of fermi surface on fodos .
the anisotropy of fermi surface affects zedos through @xmath129 in eq .
( [ pzedos ] ) . if the distribution of @xmath129 changes according to the magnetic field direction , zedos also depends on the field direction , even without gap nodes . in this section ,
we focus on the shift of @xmath2 with increasing anisotropy of fermi surface . in order to take account of the in - plane anisotropy of fermi surface , we consider a simplified model shown in fig .
[ necylinder ] .
we make fermi surface by connecting the sets of wave numbers , i.e. by adding nested parts to a cylindrical fermi surface .
@xmath155 @xmath156 @xmath157 here , @xmath158 is a measure of anisotropy . when @xmath159 and @xmath160 , the linear parts occupy @xmath161 and @xmath162 of the fermi surface , respectively .
we assume the absolute value of in - plane components of fermi velocity @xmath163 are the same all over the fermi surface , and set a c - axis component @xmath164 .
we assume a d@xmath165-like pairing , i.e. @xmath166 i.e. , gap nodes are on the cylindrical parts of the fermi surface . a schematic picture of the fermi surface with in - plane anisotropy seen from the c - axis .
position of gap node is shown with filled circles , scaledwidth=50.0% ] when magnetic field is applied perpendicular to one of the nested parts , @xmath167 on the nested parts , leading to suppression of zedos .
whereas , when magnetic field is applied in the nodal direction , the nested parts does not influence zedos so much . hence , it is expected that the doppler - shift predominant region is suppressed as @xmath168 is increased . in fig .
[ nec_zedos ] , we show @xmath123 for several values of @xmath168
. one can find that @xmath2 is lowered as @xmath168 is increased . at @xmath169 , the doppler - shift predominant region
is suppressed to @xmath170 , i.e. , the almost entire field range at which experiments have been done .
this result suggests that many cares should be taken to apply the `` doppler - shift criterion '' when fermi surface is highly anisotropic .
at least one should check that maxima and minima of fodos are exchanged at high fields and low fields , in order to confirm that the system enters the doppler - shift predominant region at low magnetic fields .
it seems that recent experiments do not find any exchange of maxima and minima@xcite .
comparison of @xmath0 and @xmath1 for @xmath171 .
magnetic field is normalized with @xmath36 when a field is in the antinodal direction.,scaledwidth=50.0% ]
in this section , we study origin of the cusp - like singularity in fodos observed through specific heat@xcite and thermal conductivity@xcite in yni@xmath7b@xmath7c . in both cases , the cusp - like minima are observed when magnetic field is applied parallel to a or b - axis .
so far , the cusp - like minima are attributed to point nodes on the superconducting gap . in particular ,
s@xmath172g - wave gap structure has been proposed for yni@xmath7b@xmath7c@xcite . however , as we showed in section [ isotropic ] , rather broader minima of fodos are obtained from s@xmath172g - wave - like quadratic point nodes . here
, we would like to propose a new mechanism that leads to cusp - like minima in fodos .
in our theory , two conditions are required for the appearance of cusp - like minima .
the first condition is that a system has to have a quasi - two - dimensional nested fermi surface .
and the second condition is that the superconducting coherence lengths are isotropic . under these conditions ,
cusp - like minima appear when a magnetic field is applied perpendicular to the nested part of the fermi surface . here
, we briefly describe how cusp - like minima appear from the above conditions , in terms of eq .
( [ pzedos ] ) . provided that superconducting coherence lengths are isotropic , one obtains @xmath173 .
when the magnetic field is applied at an angle @xmath114 with the normal line of the nested part of the fermi surface , we have @xmath174 .
if the c - axis dispersion is weak enough ( @xmath175 ) , one can ignore the second term in the right - hand side , and the contribution to zedos from the nested part of the fermi surface @xmath176 is written as @xmath177 for small @xmath114 .
hence , fodos has cusp - like minima at @xmath178 , i.e. , when a magnetic field is applied perpendicular to the nested parts . the @xmath179 dependence in eq .
( [ cusporigin ] ) allows one a simple physical interpretation .
suppose a quasiclassical trajectory@xcite of nested fermi surface , which is running in the direction at an angle @xmath114 with magnetic field , and passing near vortex cores ( fig .
[ corestate ] ) . on this trajectory ,
phase of order parameter changes drastically near vortex cores , leading to formation of andreev bound states .
since the vortex cores are separated by the distance @xmath180 on the trajectory , the density of andreev bound states is proportional to @xmath181 , hence make a cusp - like singularity in fodos .
this brief discussion reveals that the cusp - like singularity is caused by the core states , which are not properly taken into account in the doppler - shift method .
( upper)a schematic picture of a trajectory running at an angle @xmath114 with magnetic field .
vortex cores are shown with dashed lines .
( lower)andreev bound states formed on the trajectory.,title="fig:",scaledwidth=50.0% ] ( upper)a schematic picture of a trajectory running at an angle @xmath114 with magnetic field .
vortex cores are shown with dashed lines .
( lower)andreev bound states formed on the trajectory.,title="fig:",scaledwidth=50.0% ] here , we discuss the basic properties of yni@xmath7b@xmath7c .
yni@xmath7b@xmath7c has three bands ( 17th,18th,19th ) crossing fermi energy@xcite .
the 17th band , which mainly consists of ni(4d ) orbits , forms an open fermi surface in the c - axis direction , hence has a quasi - two - dimensional nature@xcite .
it is reported that the corresponding fermi surface of luni@xmath7b@xmath7c has a nested part , with nesting vectors @xmath182@xcite .
kohn anomaly in phonon spectrum@xcite provides another evidence of the nesting . on the other hand , the 18th and the 19th bands form closed fermi surface in the c - axis direction ,
hence have three - dimensional tendency . as to the gap structure , existence of nodes
have been confirmed by specific heat@xcite and ultra - sonic attenuation@xcite etc . in order to discuss the features of yni@xmath7b@xmath7c , we study the following two - band model .
our model consists of ( @xmath183 ) a partly nested quasi - two - dimensional fermi surface as shown in fig .
[ necylinder ] , and ( @xmath184 ) a three - dimensional spherical fermi surface .
we suppose the normal - state density of states of these two fermi surfaces are the same .
we set @xmath185 for the first fermi surface .
then , about 4% of the entire fermi surfaces are nested , in agreement with ref . , where the proportion of the nested part is reported to be @xmath186% . as to the gap structure , we assume isotropic gap for the quasi - two - dimensional fermi surface , and the following two cases for the spherical fermi surface , with different nodal position : ( i)@xmath187 and ( ii ) @xmath188 . as to the interaction potential @xmath189 in eq .
( [ mpeschgapeq ] ) , we assume @xmath190 and @xmath191 . here , we assume spherical fermi surface has the larger superconducting instability , in order to make superconducting coherence lengths of this system isotropic@xcite .
( a)fodos of the two - band model .
gap function ( i ) is adopted for the nodal structure of the spherical fermi surface .
( b)contribution to fodos from each fermi surface.,title="fig:",scaledwidth=50.0% ] ( a)fodos of the two - band model . gap function ( i ) is adopted for the nodal structure of the spherical fermi surface .
( b)contribution to fodos from each fermi surface.,title="fig:",scaledwidth=50.0% ] ( a)fodos of the two - band model . gap function ( ii ) is adopted for the nodal structure of the spherical fermi surface .
( b)contribution to fodos from each fermi surface.,title="fig:",scaledwidth=50.0% ] ( a)fodos of the two - band model .
gap function ( ii ) is adopted for the nodal structure of the spherical fermi surface .
( b)contribution to fodos from each fermi surface.,title="fig:",scaledwidth=50.0% ] contribution to fodos from several @xmath153 of the quasi - two - dimensional fermi surface .
this figure should be compared with fig .
[ fermi_div](b),scaledwidth=50.0% ] in fig .
[ bolofodos1](a ) and fig .
[ bolofodos2](a ) , we give fodos of this model .
calculation is done at @xmath192 .
one can find a clear cusp - like minima in both cases .
furthermore , it is worth noting that fodos has minima for @xmath126(@xmath178 ) , irrespective of the nodal position . in fig .
[ bolofodos1](b ) and fig .
[ bolofodos2](b ) , we show contribution to fodos from each fermi surface .
one can find that the fodos arising from the quasi - two - dimensional fermi surface is larger and dominate the contribution from the spherical fermi surface .
it is surprising that only tiny anisotropy of fermi surface destroys the doppler - shift predominant region below @xmath192 .
it makes a contrast with the single band case in section [ inplane ] where as much as 40% of fermi surface has to be nested to suppress the doppler - shift predominant region to @xmath193 .
one can understand the origin of the enhancement of the contribution from the quasi - two - dimensional fermi surface , on the basis of the scenario in section [ isotropic ] . in section [ isotropic ] , we noted that largest contribution to fodos comes from the part of fermi surface where effective fermi velocity @xmath67 is nearly parallel to the field - rotational plane .
in the case of cylindrical fermi surface in section [ isotropic ] , such parts are limited to @xmath194 due to anisotropic coherence lengths @xmath195 . however , in the two - band model in this section , the superconducting coherence lengths are isotropic ( @xmath196 ) , hence @xmath67 is almost parallel to the field - rotational plane all over the quasi - two - dimensional fermi surface .
therefore , contribution to fodos come from all @xmath153 and becomes much larger than the contribution from the nodal structure on the spherical fermi surface .
we plot contribution to fodos from several @xmath153 in fig .
[ bolofodos_div ] . in this figure
, one can find that all @xmath153 contribute to fodos equally .
the cusp - like singularity shown here is basically owing to the co - existence of two - dimensional and three - dimensional fermi surfaces and to the nesting nature of two - dimensional fermi surface .
this example shows that many cares are needed for an analysis of experimental data in the multi - band system .
then , anisotropy of quasi - two - dimensional fermi surface is enhanced and invalidate the simple argument based on the doppler - shift methods .
we study the influence of superconducting gap and fermi surface structures on the field - orientational dependence of zedos .
first , we studied the crossover behavior of zedos , by taking account of the four - fold oscillation of @xmath83 , and found that there exists a crossover magnetic field @xmath2 so that @xmath3 for @xmath197 , while @xmath5 for @xmath6 , consistent with our previous analyses@xcite . next , we have investigated the roles of gap structure and nature of fermi surface . the crossover behavior and fodos at the doppler - shift predominant region
have been discussed .
we found that there are no significant differences in neither the value of @xmath2 nor the shape of fodos between the spherical fermi surface and the cylindrical fermi surface , in spite of the difference in the dimensionality of fermi surface .
we also found that when quadratic point nodes exist , @xmath2 is lowered than the line node cases and fodos has broader minima .
these results can be coherently understood by recognizing that contribution to fodos mostly comes from the part on the fermi surface where effective fermi velocity is nearly parallel to the field - rotational plane .
next , we examine the effects of in - plane anisotropy of fermi surface .
we showed that the doppler - shift predominant region may be suppressed to lower fields by the anisotropy of fermi surface .
this tendency becomes especially strong in a multi - band superconductor which has nearly isotropic coherence lengths .
therefore , some cares are necessary for analyzing the field - rotational experiments for yni@xmath7b@xmath7c and cecoin@xmath9 .
on the other hand , the field - rotational experiment and its interpretation based on the doppler - shift criterion are powerful tool for such material as sr@xmath7ruo@xmath8 , na@[email protected]@xmath7o@xcite and organic superconductors , which are well described by a single fermi surface or have plural fermi surfaces with the same dimensionality . finally , we propose a novel mechanism of the cusp - like minima found in yni@xmath7b@xmath7c , in terms of andreev bound states .
the cusp - like minima are attributed to nesting in quasi - two - dimensional fermi surface and the isotropy of superconducting coherence lengths .
we are grateful to y. matsuda , t. sakakibara , y. kato , k. izawa , h. kusunose and t. watanabe for fruitful discussions .
we would also like to thank k. kamata for providing her master thesis on the field - rotational experiments of yni@xmath7b@xmath7c .
k. izawa , h. takahashi , h. yamaguchi , yuji matsuda , m. suzuki , t. sasaki , t. fukase , y. yoshida , r. settai , and y. onuki , phys . rev .
lett * 86 * , 2653 ( 2001 ) .
m. a. tanatar , m. suzuki , s. nagai , z. q. mao , y. maeno , and t. ishiguro , phys .
lett * 86 * , 2649 ( 2001 ) .
k. deguchi , z. q. mao , h. yaguchi , and y. maeno , phys .
lett . * 92 * , 047002 ( 2004 ) .
k. deguchi , z. q. mao and y. maeno , j. phys .
* 73 * , 1313 ( 2004 ) .
k. izawa , h. yamaguchi , yuji matsuda , h. shishido , r. settai , and y. onuki , phys . rev .
lett * 87 * , 057002 ( 2001 ) .
h. aoki , t. sakakibara , h. shishido , r. settai , y. onuki , p. miranovi , and k. machida , j. phys . : condens .
matter * 16 * , no 3 , ( 2004 ) l13-l19 .
k. izawa , h. yamaguchi , t. sasaki , and yuji matsuda , phys .
lett * 88 * , 027002 ( 2002 ) .
k. izawa , k. kamata , y. nakajima , y. matsuda , t. watanabe , m. nohara , h. takagi , p. thalmeier , and k. maki phys .
lett * 89 * , 137006 ( 2002 ) .
tuson park , m. b. salamon , eun mi choi , heon jung kim , and sung - ik lee phys .
lett * 90 * , 177001 ( 2003 ) .
k. izawa , y. nakajima , j. goryo , y. matsuda , s. osaki , h. sugawara , h. sato , p. thalmeier , and k. maki phys . rev .
lett * 90 * , 117001 ( 2003 ) .
t. watanabe , k. izawa , y. kasahara , y. haga , y. onuki , p. thalmeier , k. maki and y. matsuda , cond - mat/0405211 m. udagawa , y. yanase , and m. ogata , cond - mat/0401206 .
p. thalmeier and k. maki , cond - mat/0210364 .
w. pesch , z. phys .
b * 21 * , 263 ( 1975 ) ; p. klimesch and w. pesch , j. low temp .
phys . * 32 * , 869 ( 1978 ) .
t. dahm , s. graser , c. iniotakis , and n. schopohl , phys .
b * 66 * , 144515 ( 2002 ) .
s. graser , t. dahm , and n. schopohl , phys .
b 69 , 014511 ( 2004 ) .
h. kusunose , cond - mat/0401077 .
l. tewordt and d. fay , phys .
b * 64 * , 024528 ( 2001 ) .
g. eilenberger , z. phys . , * 214 * , 195 ( 1968 ) .
a. i. larkin and y. n. ovchinikov , sov .
jetp * 46 * , 155 ( 1977 ) .
u. klein , j. low .
. phys . * 69 * , 1 ( 1987 ) .
n. schopohl , j. low temp .
* 41 * , 409 ( 1980 ) .
t. dahm and n. schopohl , phys .
lett * 91 * , 017001 ( 2003 ) .
p. miranovi , n. nakai , m. ichioka , and k. machida , phys .
b * 68 * , 052501 ( 2003 ) . j. i. lee , t. s. zhao , i. g. kim , b. i. min , and s. j. youn , phys .
b * 50 * , 403 ( 1994 ) .
d. j. singh , solid state commun .
* 49 * , 899 ( 1996 ) .
s. b. dugdale , _
et al_. , phys .
83 * , 4824 ( 1999 ) .
p. dervenagas , m. bullock , j. zarestky , p. canfield , b. k. cho , b. harmon , a. i. goldman , and c. stassis , phys .
b * 52 * , r9839 ( 1995 ) h. kawano , h. yoshizawa , h. takeya , and k. kadowaki , phys
lett . * 77 * , 4628 ( 1996 ) . c. stassis , m. bullock , j. zarestky , p. canfield , a. i. goldman , g. shirane and s. m. shapiro , phys .
b * 55 * , r8678 ( 1997 ) .
m. nohara , m. isshiki , f. sakai and h. takagi , j. phys .
jpn . , * 68 * , 1078 ( 1999 ) .
t. watanabe , m. nohara , t. hanaguri , and h. takagi , phys .
* 92 * , 147002 ( 2004 ) .
h. takagi , m. nohara and r. j. cava , physica b , * 237 - 238 * , 292 ( 1997 ) .
j. w. serene and d. rainer , phys .
* 101 * , 221 ( 1983 ) .
n. kopnin , _ theory of nonequilibrium superconductivity_. n. schopohl , cond - mat/9804064 .
k. takada , h. sakurai , e. takayama - muromachi , f. izumi , r. a. dilanian and t sasaki , nature * 422 * , 53 ( 2003 ) . | we have studied the field - orientational dependence of zero - energy density of states ( fodos ) for a series of systems with different fermi surface and superconducting gap structures . instead of phenomenological doppler - shift method , we use an approximate analytical solution of eilenberger equation together with self - consistent determination of order parameter and a variational treatment of vortex lattice .
first , we compare zero - energy density of states ( zedos ) when a magnetic field is applied in the nodal direction ( @xmath0 ) and in the antinodal direction ( @xmath1 ) , by taking account of the field - angle dependence of order parameter . as a result
, we found that there exists a crossover magnetic field @xmath2 so that @xmath3 for @xmath4 , while @xmath5 for @xmath6 , consistent with our previous analyses .
next , we showed that @xmath2 and the shape of fodos are determined by contribution from the small part of fermi surface where fermi velocity is parallel to field - rotational plane .
in particular , we found that @xmath2 is lowered and fodos has broader minima , when a superconducting gap has point nodes , in contrast to the result of the doppler - shift method .
we also studied the effects of in - plane anisotropy of fermi surface .
we found that in - plane anisotropy of quasi - two dimensional fermi surface sometimes becomes larger than the effects of doppler - shift and can destroy the doppler - shift predominant region .
in particular , this tendency is strong in a multi - band system where superconducting coherence lengths are isotropic .
finally , we addressed the problem of cusp - like singularity in yni@xmath7b@xmath7c and present a possible explanation of this phenomenon . |
we study four systems and their interactions : a unified system of coupled forward - backward stochastic partial differential equations ( fb - spdes ) with lvy jumps ; a unified system of fb - sdes , a special form of the fb - spdes , however , with skew reflections ; a stochastic differential game ( sdg ) problem with general number of players based on the fb - sdes ; and a system of queues and their associated reflecting diffusion approximations .
more precisely , there are four interconnected and streamlined aims involved in our discussions .
the first aim is to study the adapted 4-tuple strong solution @xmath1 to the unified system of coupled fb - spdes with lvy jumps with respect to time - position parameter @xmath2 , @xmath3 where , @xmath4 $ ] and @xmath5 $ ] is a stopping time with regard to a filtration defined later in the paper , @xmath6 or @xmath7 for a positive integer @xmath8 , and @xmath9 denotes the corresponding left limit at time point @xmath10 . in particular
, @xmath11 with a given @xmath12 is a connected domain , for examples , a @xmath13-dimensional box , a @xmath13-dimensional ball ( or a general manifold ) , a @xmath13-dimensional sphere ( or a general riemannian manifold ) , or the whole euclidean space @xmath14 of real numbers itself .
the f - spde in is with the given initial random vector - field @xmath15 , while the b - spde in has the known terminal random vector - field @xmath16 . in
, @xmath17 and @xmath18 are @xmath19-dimensional and @xmath20-dimensional random vector - field processes respectively , @xmath21 is a standard @xmath22-dimensional brownian motion , and @xmath23 is a @xmath8-dimensional centered lvy jump process ( or centered subordinator ) .
furthermore , the partial differential operators of @xmath19-dimensional vector @xmath24 , @xmath25-dimensional matrix @xmath26 , and @xmath27-dimensional matrix @xmath28 are functionals of @xmath17 , @xmath18 , @xmath29 , @xmath30 , and their partial derivatives of up to the @xmath31th order for @xmath32 .
so do the partial differential operators of @xmath20-dimensional vector @xmath33 , @xmath34-dimensional matrix @xmath35 , and @xmath36-dimensional matrix @xmath37 . more precisely
, for each @xmath38 , @xmath39 where the dot @xmath40 " in @xmath41 and its associated partial derivatives denotes the integration in terms of the so - called lvy measure .
however , if @xmath42 , the last line on the right - hand side of should be changed to the form , @xmath43 = 3.5 in note that our partial differential operators presented in can be general - nonlinear and high - order , e.g. , @xmath44 for a general nonlinear functional @xmath45 , where @xmath46 are nonnegative integers satisfying @xmath47 with @xmath32 . furthermore , the initial random vector - field @xmath15 , the terminal random vector - filed @xmath16 , and the 4-tuple solution process @xmath1 can be complex - valued . under an infinite sequence of generalized local linear growth and lipschitz conditions ,
we prove the existence and uniqueness of an adapted 4-tuple strong solution to the fb - spdes in a suitably constructed functional topological space .
the solution to the unified system in can be interpreted in a sample surface manner with time - position parameter @xmath48 ( see , e.g. , @xmath49 in figure [ samplesurface ] for such an example ) . the quite involved technical proof developed in this paper
is extended from our earlier work summarized in dai @xcite ( arxiv , 2011 ) for a unified b - spde .
more precisely , the newly unified system of coupled fb - spdes in covers many existing forward and/or backward sdes / spdes as special cases , where the partial differential operators are taken to be special forms . for examples , specific single - dimensional strongly nonlinear f - spde and b - spde driven solely by brownian
motions can be respectively derived for the purpose of optimal - utility based portfolio choice ( see , e.g , musiela and zariphopoulou @xcite ) . here
, the strong nonlinearity is in the sense addressed by lions and souganidis @xcite and pardoux @xcite .
furthermore , the single - dimensional stochastic hamilton - jacobi - bellman ( hjb ) equations are also examples of our unified system in , which are specific b - spdes ( see , e.g. , @xmath50ksendal _ et al . _
@xcite and references therein ) .
note that the proof of the well - posedness concerning solution to the b - spde derived in musiela and zariphopoulou @xcite and solution to the hjb equation derived in @xmath50ksendal _ et al . _
@xcite is covered by the study in dai @xcite ( arxiv , 2011 ) although the authors in both @xcite and @xmath50ksendal _ et al . _
@xcite claim it as an open problem .
the proof of the well - posedness about solution to the f - spde derived in musiela and zariphopoulou @xcite is covered by the even more unified discussion for the coupled fb - spdes in of this paper .
actually , partial motivations to enhance the unified b - spde in dai @xcite ( arxiv , 2011 ) to the coupled fb - spdes in are from the conference discussion @xcite during 45 minutes invited talk presented by zariphopoulou in icm 2014 , where the current author claimed that the well - posedness of solution to the f - spde in @xcite can be proved by the method developed in dai @xcite ( arxiv , 2011 ) .
besides these existing examples , our motivations to study the coupled fb - spdes in are also from optimal portfolio management in finance ( see , e.g. , dai @xcite ) , and multi - channel ( or multi - valued ) image regularization such as color images in computer vision and network applications ( see , e.g. , caselles _
et al . _
@xcite ) . in this part
, we also show the usages of our unified system in in heat diffusions and quantum hall / anomalous hall effects as two illustrative examples to support our first aim .
mathematically , we refine a stochastic dirichlet - poisson problem from heat diffusions and use stochastic schr@xmath51dinger equation as model for hall effects in quantum statistics .
it is worth to point out that the proving methodology developed in the current paper and its early version in dai @xcite ( arxiv , 2011 ) is aimed to provide a general theory and framework to show the well - posedness of a unified general system class of the coupled fb - spdes in .
however , some specific forms of the fb - spdes in ( either in forward manner or in backward manner ) may be solved by alternative techniques , e.g. , the author in his fields metal awarded work ( hairer @xcite and icm 2014 ) solves the kpz equation by rough path technology , and furthermore , the related rough path theory can deal with the lack of either temporal or spatial regularity ( see , e.g. , hairer @xcite and reference therein ) .
the second aim of the paper is to prove the well - posedness of an adapted 6-tuple weak solution @xmath52 with 2-tuple boundary regulator @xmath53 to the ( possible ) non - markovian system of coupled fb - sdes with lvy jumps and skew reflections under a given control rule @xmath54 , @xmath55 in , @xmath56 is a @xmath13-dimensional process governed by the f - sde with skew reflection matrix @xmath57 and @xmath18 is a @xmath20-dimensional process governed by the b - sde with skew reflection matrix @xmath58 .
furthermore , @xmath59 can increase only when @xmath56 is on a boundary @xmath60 with @xmath61 and @xmath62 can increase only when @xmath18 is on a boundary @xmath63 with @xmath64 , where @xmath65 and @xmath66 are two nonnegative integers .
both @xmath59 and @xmath62 are the regulating processes with possible jumps to push @xmath56 and @xmath18 back into the state spaces @xmath67 and @xmath68 respectively .
they are parts of the 6-tuple solution to and determined by solution pairs to the well - known skorohod problem ( see , e.g. , dai @xcite , dai and dai @xcite , or section [ gfbproof ] of the current paper for such a definition ) .
thus , we call them as skorohod regulators ( see , figure [ reflection ] for such an example ) . note that , comparing with the unified system in , the coefficients appeared in do not contain any partial derivative operator but
the fb - sdes themselves involve skew boundary reflections .
the proof for the well - posedness of an adapted 6-tuple weak solution to the fb - sdes is based on two general conditions .
the first one is a general completely-@xmath0 condition ( see , e.g. , dai @xcite , dai and dai @xcite , and figure [ reflection ] for an illustration ) .
the non - uniqueness of solution to an associated skorohod problem under this condition is one of the major difficulties in the proof .
= 3.5 in the second one is the generalized linear growth and lipschitz conditions , where the conventional growth and lipschitz constant is replaced by a possible unbounded but mean - squarely integrable adapted stochastic process ( see , e.g. , dai @xcite ) . in particular ,
if the completely-@xmath0 condition becomes more strict , e.g. , with additional requirements that the spectral radii in certain sense for both reflection matrices are strictly less than the unity , a unique adapted 6-tuple strong solution will be concerned .
concerning coupled fb - sdes , it motivates a hot research area ( see , e.g. , @xmath50ksendal _ et al . _
@xcite about the discussion of coupled fb - sdes with no boundary reflection , karatzas and li @xcite about the study of brownian motion driven b - sde with reflection , and references therein ) .
however , to our best knowledge , the coupled system in with double skew reflection matrices and the well - posedness study in terms of an adapted 6-tuple weak solution with lvy jumps and under a general completely-@xmath0 condition through the skorohod problem are new and for the first time in this area .
the third aim of the paper involves two folds .
on the one hand , we use the 4-tuple solution to the coupled fb - spdes in to obtain an adapted 6-tuple solution to the system in ; on the other hand , we use the obtained adapted 6-tuple solution to determine a pareto optimal nash equilibrium policy process to a non - zero - sum sdg problem in , which is newly formulated by the fb - sdes in . in this game , there are @xmath20-players and each player @xmath69 has his own value function @xmath70 subject to the system in under an admissible control policy @xmath54 .
every player @xmath71 chooses an optimal policy to maximize his own value function over an admissible policy set @xmath72 while the summation of all value functions is also maximized , i.e. , @xmath73 for each @xmath74 , where , @xmath75 note that the total value function @xmath76 does not have to be a constant ( e.g. , zero ) , or in other words , the game is not necessarily a zero - sum one . the contribution and literature review of the study associated with the game in - for the third aim can be summarized as follows .
one of the important solution methods for sde based optimal control is the dynamic programming . in general , this method is related to a special case of the unified system in ( or its earlier unified b - spde form in dai @xcite ( arxiv , 2011 ) ) , e.g. , the specific b - spde with @xmath77 ( called stochastic hjb equation ) in peng @xcite with no jumps and @xmath50ksendal _ et al . _
@xcite with jumps . here ,
we extend the discussions in peng @xcite and @xmath50ksendal _ et al . _
@xcite to a system of generalized coupled forward - backward oriented stochastic hjb equations with jumps corresponding to the case that @xmath78 .
more importantly , this system provides an effective way to resolve a non - zero - sum sdg problem with jumps and general number of @xmath20 players , which subjects to a non - markovian system of coupled fb - sdes with lvy jumps and skew reflections ( see , e.g. , figure [ gamei ] for such a game platform ( partially adapted from dai @xcite ) ) .
= 3.5 in by a solution to the fb - spdes in , we determine a solution to the fb - sdes in under a given control rule and then obtain a pareto optimal nash equilibrium policy process to the non - zero - sum sdg problem in .
note that , the concept and technique concerning the non - zero - sum sdg and pareto optimality used in this paper is refined and generalized from dai @xcite and karatzas and li @xcite .
the fourth aim of the paper involves three folds .
first , we study some queueing networks ( see , e.g. , figure [ queuei ] ) = 3.5 in whose dynamics ( e.g. , queue length process ) is governed by specific forms of the fb - sdes in .
these forms can be a lvy driven sde , a @xmath13-dimensional reflecting brownian motion ( rbm ) ( see , e.g. , dai @xcite , dai and dai @xcite , dai and jiang @xcite ) , or a reflecting diffusion with regime switching ( rdrs ) ( see , e.g. , dai @xcite ) .
the reflecting diffusion is the functional limit of a sequence of physical queueing processes under diffusive scaling , a general completely-@xmath0 boundary reflection constraint , and a well - known heavy traffic condition ( an analogous treatment as the one for infinite constant " in the kpz equation ( see , e.g. , hairer @xcite ) ) . in reality , the characteristics of lvy driven networks may be used to model or approximate more general batch - arrival and batch - service queueing networks .
second , we discuss how to use the queueing systems and their associated reflecting diffusion approximations to motivate the sdg problem . the criterion for the players in the game can be the queue length based performance optimization ones or queueing related cost / profit optimization ones .
third , we study the applications of the fb - spdes presented by in the queueing networks .
there are two types of equations involved .
one is the kolmogorov s equation or fokker - planck s formula oriented pdes / spdes , which are corresponding to the distributions of queueing length processes under given network control rules .
this type of equations are mainly used to estimate the performance measures of the queueing networks .
another type of equations are the hjb equation oriented pdes / spdes , which are mainly used to obtain optimal control rules over the set of admissible strategies for the queueing networks .
the remainder of the paper is organized as follows . in section
[ uniexist ] , we introduce suitable functional topological space and state conditions required for our main theorems to guarantee the well - posedness of an adapted 4-tuple strong solution to the unified system of coupled fb - spdes in . in section [ fbsdes ] , we study the unified system of coupled fb - sdes with lvy jumps and skew reflections in and present the well - posedness theorem . in particular , we establish the solution connection between the fb - spdes and the fb - sdes . in section [ queuesdg ] , we formally formulate the fb - sdes based sge problem in and determine the pareto optimal nash equilibrium policy process by a system of generalized stochastic hjb equations ( a particular form of the coupled fb - spdes ) .
related applications in queueing networks are also discussed .
finally , in sections [ uniextproof]-[gtheoremproof ] , we develop theory to prove our main theorems .
first of all , let @xmath79 be a fixed complete probability space .
then , we define a standard @xmath22-dimensional brownian motion @xmath80\}$ ] for a given @xmath81 with @xmath82 and a @xmath8-dimensional general lvy pure jump process ( or special subordinator ) @xmath83\}$ ] with @xmath84 on the space ( see , e.g. , applebaum @xcite , bertoin @xcite , and sato @xcite ) . note that the prime appeared in this paper is used to denote the corresponding transpose of a matrix or a vector .
furthermore , @xmath21 , @xmath85 , and their components are supposed to be independent of each other . for each @xmath86 ,
which is called a reversion rate vector in many applications , we let @xmath87 .
then , we denote a filtration by @xmath88 with @xmath89 for each @xmath90 $ ] , where @xmath91 is @xmath92-algebra independent of @xmath21 and @xmath85 .
in addition , let @xmath93 be the index function over the set @xmath94 and @xmath95 for each @xmath96 be a lvy measure .
then , we use @xmath97\times a)\equiv\sum_{0<s\leq t } i_{a}(l_{i}(s)-l_{i}(s^{-}))$ ] to denote a poisson random measure with a deterministic , time - homogeneous intensity measure @xmath98 .
thus , each @xmath99 can be represented by ( see , e.g. , theorem 13.4 and corollary 13.7 in pages 237 and 239 of kallenberg @xcite ) @xmath100}\int_{{\cal z}}z_{i}n_{i}(\lambda_{i}ds , dz_{i } ) , \;t\geq 0 .
\elabel{subordrep}\end{aligned}\ ] ] for convenience , we take the constant @xmath101 to be zero . in the subsequent two subsections ,
we first study the unified system in , which is real - valued with a closed position parametric domain .
then , we extend the discussion to a complex - valued system with an open position parametric domain .
we let @xmath11 be a closed position parametric domain and use @xmath102 for each @xmath103 and @xmath104 to denote the banach space of all functions @xmath45 having continuous derivatives up to the order @xmath31 with the uniform norm for each @xmath45 in this space , @xmath105 the @xmath106 in for each @xmath107 is the total number of the partial derivatives of the order @xmath108 @xmath109 with @xmath110 , @xmath111 , @xmath112 , and @xmath113 . here , we remark that , whenever the partial derivative on the boundary @xmath114 is concerned , it is defined in a one - side manner .
in addition , let @xmath115 where each @xmath116 corresponds to a @xmath13-tuple @xmath117 and a @xmath112 .
then , we use @xmath118 to denote the banach space @xmath119 where @xmath120 for some discrete function @xmath121 in terms of @xmath122 , which is fast decaying in @xmath108 . for convenience ,
we take @xmath123 with @xmath124^{c } , \nonumber\end{aligned}\ ] ] where the notation @xmath125 $ ] denotes the summation of the unity and the integer part of a real number .
next , let @xmath126,c^{\infty}(d;r^{l}))$ ] denote the set of all @xmath127-valued ( or called @xmath128-valued ) measurable random vector - field processes @xmath129 adapted to @xmath130\}$ ] for each @xmath131 , which are in @xmath118 for each fixed @xmath90 $ ] ) , such that @xmath132<\infty .
\elabel{adaptednormi}\end{aligned}\ ] ] in particular , let @xmath133 with @xmath104 denote the set of all @xmath127-valued random vector - fields @xmath134 that are @xmath135-measurable for each @xmath131 and satisfy @xmath136<\infty , \elabel{initialx}\end{aligned}\ ] ] where @xmath137 and @xmath138 .
in addition , let @xmath139\times { \cal z}^{h},$ ] @xmath140 be the set of all @xmath141-valued random vector - field processes denoted by @xmath142 @xmath143 @xmath144 @xmath145 , which are predictable for each @xmath131 and @xmath146 and are endowed with the norm @xmath147<\infty . \elabel{normsi}\end{aligned}\ ] ] thus , we can define @xmath148\times d ) & \equiv&\;\;\;l^{2}_{{\cal f}}([0,t],c^{\infty}(d , r^{r}))\elabel{bunisoi}\\ & & \times l^{2}_{{\cal f}}([0,t],c^{\infty}(d , r^{q } ) ) \nonumber\\ & & \times l^{2}_{{\cal f},p}([0,t],c^{\infty}(d , r^{q\times d } ) ) \nonumber\\ & & \times l^{2}_{p}([0,t]\times{\cal z}^{h},c^{\infty}(d , r^{q\times h } ) ) .
\nonumber\end{aligned}\ ] ] finally , let @xmath149 that is endowed with the norm @xmath150 for any @xmath151 and @xmath152 .
furthermore , define @xmath153 in the sequel , we let @xmath154 be the largest absolute value of entries ( or components ) of the given matrix ( or vector ) @xmath94 . furthermore , for each @xmath155 $ ] and @xmath146 , we let @xmath156 where @xmath157 for each @xmath96 .
then , we impose some conditions to guarantee the unique existence of an adapted 4-tuple strong solution to the unified system in . first , for each @xmath158 , every @xmath159 , and any @xmath160 with @xmath161 , we define @xmath162 then , we assume that the generalized local lipschitz condition is true almost surely ( a.s . ) , @xmath163 note that @xmath164 in with each @xmath122 is a nonnegative constant .
it depends on the domain @xmath67 and the differential order @xmath108 and may be unbounded as @xmath165 and @xmath166 .
@xmath167 denotes the @xmath71th order of partial derivative of @xmath168 in time variable @xmath169 .
@xmath170 denotes the @xmath171th order of partial derivative of @xmath172 in terms of a component of @xmath54 , @xmath173 , @xmath174 , or @xmath175 .
furthermore , for each @xmath176 , we suppose that @xmath177 where @xmath178 is the @xmath179th column of @xmath180 .
second , for each @xmath158 , every @xmath159 , and any @xmath181 , we suppose that the generalized local linear growth condition holds @xmath182 where , @xmath183
if @xmath184 and @xmath185 if @xmath186 . similarly , for each @xmath176 , we suppose that @xmath187 then , we can state our main theorem of this subsection as follows . [ bsdeyi ] suppose that @xmath188 and conditions in - are true
. furthermore , assume that each @xmath189 is @xmath190-adapted for every fixed @xmath131 , @xmath146 , and any given @xmath191 with @xmath192,c^{\infty}(d , r^{r})\right ) , \elabel{flipic}\\ & & { \cal j}(\cdot , x,0)\in l^{2}_{{\cal f}}\left([0,t],c^{\infty}(d , r^{r\times d})\right ) , \elabel{flipicadi}\\ & & { \cal i}(\cdot , x,0,\cdot)\in l^{2}_{{\cal f}}\left([0,t]\times r_{+}^{h},c^{\infty}(d , r^{r\times h})\right ) , \elabel{flipicadii}\\ & & \bar{{\cal l}}(\cdot , x,0)\in l^{2}_{{\cal f}}\left([0,t],c^{\infty}(d , r^{q})\right ) , \elabel{blipic}\\ & & \bar{{\cal j}}(\cdot , x,0)\in l^{2}_{{\cal f}}\left([0,t],c^{\infty}(d , r^{q\times d})\right ) , \elabel{blipicadi}\\ & & \bar{{\cal i}}(\cdot , x,0,\cdot)\in l^{2}_{{\cal f}}\left([0,t]\times r_{+}^{h},c^{\infty}(d , r^{q\times h})\right ) .
\elabel{blipicadii}\end{aligned}\ ] ] then , there exists a unique adapted 4-tuple strong solution to the system in , i.e. , @xmath193\times d ) , \elabel{buniso}\end{aligned}\ ] ] and @xmath194 is cdlg for each @xmath131 almost surely ( a.s . ) .
the proof of theorem [ bsdeyi ] is provided in section [ uniextproof ] . in this subsection
, we generalize the study in subsection [ boundedd ] to the case corresponding to an open ( or partially open ) position parametric domain @xmath67 ( e.g. , @xmath14 or @xmath195 ) .
more exactly , we assume that there exists a sequence of nondecreasing closed and connected sets @xmath196 such that @xmath197 furthermore , let @xmath198 with @xmath104 be the banach space endowed with the norm for each @xmath45 in the space @xmath199 where @xmath200 is the @xmath71-dimensional complex euclidean space and the norm @xmath201 in is interpreted in the corresponding complex - valued sense .
in addition , define @xmath202\times d)$ ] to be the corresponding space in if the terminal time @xmath203 is replaced by the stopping time @xmath204 in and the norm in is substituted by the one in . finally , we use the same way to interpret the spaces @xmath205 and @xmath206 .
then , we have the following theorem . [ infdn ]
suppose that @xmath207 and the system in satisfies the conditions in - over @xmath208 for each @xmath209 with associated ( local ) linear growth and lipshitz constant @xmath210 .
furthermore , assume that each @xmath211 is @xmath190-adapted for every fixed @xmath131 , @xmath146 , and any given @xmath181 with conditions in - being true .
then , the system in has a unique adapted 4-tuple strong solution @xmath212\times d ) , \elabel{obuniso}\end{aligned}\ ] ] and @xmath194 is cdlg for each @xmath131 a.s .
the proof of theorem [ infdn ] is provided in section [ uniextproof ] . in this subsection
, we present two illustrative examples related to heat diffusions and quantum hall / anomalous hall effects .
mathematically , the heat diffusions are modeled as a stochastic dirichlet - poisson problem and the hall effects are presented by a stochastic schr@xmath51dinger equation .
examples related to queueing systems and sdgs will be presented in section [ queuesdg ] after studying the system of coupled fb - sdes in .
0.4 cm * * heat diffusions : stochastic dirichlet - poisson problem * 0.4 cm for this example , we consider a special b - spde ( a specific backward part of the system in ) .
more precisely , the associated partial differential operators are given by @xmath213 where @xmath18 is single - dimensional ( i.e. , @xmath77 ) and @xmath214 is some integrable function .
then , we can obtain the corresponding b - spde with jumps as follows , @xmath215 when the terminal value @xmath216 is a boundary based condition over @xmath67 , we will call the related resolution problem _ a stochastic dirichlet - poisson problem with jumps _
, which is a randomized general form of the classical dirichlet - poisson problem ( see , e.g. , the definition of a classic case in @xmath50ksendal @xcite ) .
an explanation about how to use this problem to estimate the inner or surface temperature of certain material or an object ( e.g. , sun ) is displayed in figure [ dirichlet ] .
= 4.5 in physically , the randomized heat equation in is derived from a particle system just like its classic counterpart , which can be modeled by a diffusion process as follows , @xmath217 now , we use @xmath218 to denote the first exit time ( a stopping time ) of the process @xmath219 from the time - space domain @xmath220\times d$ ] , i.e. , @xmath221\times d\ } , \elabel{adifstop}\end{aligned}\ ] ] where the upper index @xmath222 means that @xmath219 starts from @xmath131 . then , we can impose a terminal - boundary condition with @xmath223 as follows , @xmath224 for all @xmath225\times d$ ] . in the case
that @xmath226 is a random variable independent of @xmath222 , the required smooth condition for the well - posedness of the b - spde in is satisfied . in general ,
@xmath226 can be approximated by sufficiently smooth function in @xmath222 as required . 0.4 cm * * hall effects : stochastic schr@xmath51dinger equation * 0.4 cm in quantum physics and statistical mechanics , some phenomena such as hall / anomalous hall effects ( see , e.g. , hall @xcite , karplus and luttinger @xcite , and the summarized description at wikipedia website ) are major concerns . by the definition of hall effect ,
the movements of quantum particles ( e.g. , electrons ) within a semiconduct / superconduct are along regular paths ( see figure [ hallano ] for such an example ) if the lorentz force generated by an external magnetic field with a perpendicular component is imposed .
= 3.5 in when this phenomenon happens , the collisions of quantum particles will be significantly reduced and the performance of the semiconduct / superconduct will be largely improved . however , in a real application , imposing an external magnetic field is frequently expensive .
thus , people try to develop some magnetic material based semiconduct / superconduct in order that the hall effect happens naturally ( see , e.g. , karplus and luttinger @xcite , chang _ et al . _
this phenomenon is called anomalous hall effect . besides observing the hall / anomalous hall effects by physical experiments ( see , e.g.
, hall @xcite and chang _ et al . _
@xcite ) , one can also analytically study and simulate these effects through a schr@xmath51dinger equation ( see , e.g. , thouless @xcite , chai @xcite , and the summarized descriptions about density functional theory and time - dependent density functional theory at wikipedia website ) .
the schr@xmath51dinger equation used in most existing studies is a form of the fokker - planck s formula ( see , e.g. , @xmath50ksendal @xcite ) . here , by taking a form of @xmath24 in the forward part of the system in
, we can unify these schr@xmath51dinger equations ( see , e.g. , bouard and debussche @xcite , thouless @xcite , chai @xcite ) into the _ generalized stochastic nonlinear schr@xmath51dinger equation _ with absorbing boundaries for each @xmath227 ) , @xmath228 where @xmath18 is single - dimensional ( i.e. , @xmath77 ) , @xmath179 is the imaginary number , and @xmath24 is a form of the operator , @xmath229 note that @xmath230 in is the potential , which may depend on external temperature and/or external magnetic field .
for example , the recent discovery about the anomalous hall effect ( see , e.g. , chang _ et al . _
@xcite ) is based on a lower temperature and without imposing external magnetic field .
furthermore , the related schr@xmath51dinger equation based studies can be found in chai @xcite , etc .
now , if the densities appeared in the hall / anomalous hall effects are the target stationary distributions ( i.e. , the terminal values @xmath231 in are given ) , we can take @xmath33 in the system of to be a form of the operator in . then , we can find the initial and transient distributions of quantum particles by the backward part of the system in . from physical viewpoint
, this study provides insights about how to characterize and manufacture the magnetic material based semiconductor / superconduct .
in this section , we suppose that the process @xmath56 governed by the forward sde in lives in a state space @xmath67 ( e.g. , a @xmath13-dimensional positive orthant or a @xmath13-dimensional rectangle ) . furthermore , let @xmath232 for @xmath61 be the @xmath179th boundary face of @xmath67 , where @xmath233 for @xmath234 , @xmath235 is some positive constant for @xmath236 , and @xmath237 is the inward unit normal vector on the boundary face @xmath60 . for convenience ,
we define @xmath238 .
in addition , let @xmath57 in be a @xmath239 matrix with @xmath240 , whose @xmath179th column denoted by @xmath13-dimensional vector @xmath241 is the reflection direction of @xmath56 on @xmath60 .
the process @xmath59 in is a nondecreasing predictable process with @xmath242 and boundary regulating property as explained in . in queueing system , this process is called boundary idle time or blocking process . similarly , we assume that @xmath18 takes value in a region @xmath68 with boundary face @xmath243 for @xmath244 , where @xmath245 is the inward unit normal vector on the boundary face @xmath63 . for convenience ,
we define @xmath246 . in finance , the given constant @xmath247
is called early exercise reward .
furthermore , @xmath58 in is supposed to be a @xmath248 matrix for a known @xmath249 .
in addition , @xmath250 in is a nondecreasing predictable process with @xmath251 and boundary regulating property as explained in . to guarantee the existence and uniqueness of an adapted 6-tuple weak solution to the coupled fb - sdes in
, we need to introduce the completely-@xmath0 condition on the reflection matrix @xmath57 ( and similarly on @xmath58 ) .
a @xmath252 square matrix @xmath57 is called completely-@xmath0 if and only if there is @xmath253 such that @xmath254 for each principal sub - matrix @xmath255 of @xmath57 , where the vector inequalities are to be interpreted componentwise .
furthermore , a @xmath239 matrix @xmath57 is called completely-@xmath0 if and only if each @xmath252 square sub - matrix of @xmath256 is completely - s . note
that the completely-@xmath0 condition on the reflection matrices guarantees that the coupled fb - sdes are of inward reflection on each boundary and corner of the orthant or the rectangle ( see , e.g. , figure [ reflection ] and dai @xcite ) .
furthermore , the reflection appeared here is called skew reflection that is a generalization of the conventional mirror ( or called symmetry ) reflection .
now , the coefficient functions given in are assumed to be @xmath190-predictable and are detailed as follows , @xmath257\times r^{p}\times r^{q}\times r^{q\times d}\times r^{q\times h}\times u\rightarrow r^{p},\nonumber\\ \sigma(t , x , u)\equiv\sigma(t , x , v,\bar{v},\tilde{v},u,\cdot)&:&[0,t]\times r^{p}\times r^{q}\times r^{q\times d}\times r^{q\times h}\times u\rightarrow r^{p\times d},\nonumber\\ \eta(t , x , u)\equiv\eta(t , x , v,\bar{v},\tilde{v},u , z,\cdot)&:&[0,t]\times r^{p}\times r^{q}\times r^{q\times d}\times r^{q\times h}\times u\times{\cal z}^{h}\rightarrow r^{p\times h},\nonumber\\ c(t , x , u)\equiv c(t , x , v,\bar{v},\tilde{v},u,\cdot)&:&[0,t]\times r^{p}\times r^{q}\times r^{q\times d}\times r^{q\times h}\times u\rightarrow r^{q } , \nonumber\\ \alpha(t , x , u)\equiv\sigma(t , x , v,\bar{v},\tilde{v},u,\cdot)&:&[0,t]\times r^{p}\times r^{q}\times r^{q\times d}\times r^{q\times h}\times u\rightarrow r^{q\times d},\nonumber\\ \zeta(t , x , u)\equiv\gamma(t , x , v,\bar{v},\tilde{v},u , z,\cdot)&:&[0,t]\times r^{p}\times r^{q}\times r^{q\times d}\times r^{q\times h}\times u\times{\cal z}^{h}\rightarrow r^{q\times h}. \nonumber\end{aligned}\ ] ] for @xmath258 , we suppose that @xmath259 furthermore , for each @xmath260 and @xmath146 , we suppose that @xmath261 where @xmath262 is the @xmath179th column of @xmath45 , and @xmath263 in addition , we assume that the terminal value @xmath264 satisfies the condition , @xmath265 finally , @xmath85 in - and is assumed to be a known non - negative stochastic process that is @xmath266-adapted and mean - squarely integrable , i.e. , @xmath267<\infty . \elabel{fbconiii}\end{aligned}\ ] ] [ fbexist ] under conditions - , the following two claims are true : 1 . if @xmath58 and @xmath57 satisfy the completely-@xmath0 condition , there exists a unique adapted 6-tuple weak solution to the system in when at least one of the forward and backward sdes has reflection boundary ; 2 .
furthermore , if each @xmath268 sub - principal matrix of @xmath269 and each @xmath252 sub - principal matrix of @xmath256 are invertible or if both of the sdes have no reflection boundaries , there is a unique adapted 6-tuple strong solution to the system in . due to the length ,
the proof of theorem [ fbexist ] is postponed to section [ gfbproof ] . in this subsection
, we consider a particular case of the coupled fb - spdes in but with an additional equation , which corresponds to the special forms of partial differential operators @xmath270 , @xmath271 , and @xmath272 .
more precisely , for each @xmath273 , we define @xmath274 where @xmath275 and @xmath276 for @xmath234 and @xmath277 are the @xmath278th entry and the @xmath279th column of @xmath280 respectively .
furthermore , @xmath281 and @xmath282 for @xmath69 and @xmath277 is the @xmath278th entry of @xmath283 .
in addition , @xmath284 for @xmath285 and @xmath286 for @xmath287 are some functions in @xmath169 and @xmath222 .
note that , the partial derivative @xmath288 should be interpreted according to chain rule since @xmath289 is also a function in @xmath222 through @xmath290 and @xmath291 , where @xmath292 finally , we define @xmath293 where , we assume that @xmath294 . then , we have the following definition
. [ admissibled ] @xmath72 is called the admissible set of adapted control processes if @xmath295 , @xmath296 together with @xmath297 satisfy the conditions as stated in theorem [ bsdeyi ] ( or theorem [ infdn ] ) .
[ gtheoremo ] let @xmath298 be the unique adapted 4-tuple strong solution to the @xmath299-dimensional coupled fb - spdes in , which corresponds to specific @xmath300 in - , terminal condition in , and a control process @xmath301 . if @xmath58 and @xmath57 satisfy the completely-@xmath0 condition , the following claims in are true : 1 .
there exists a unique adapted 6-tuple weak solution @xmath302 , @xmath303 to the system in when at least one of the sdes has reflection boundary , where @xmath304 for @xmath273 and @xmath277 , where @xmath305 2 .
there is a unique adapted 6-tuple strong solution to the system in when each @xmath268 sub - principal matrix of @xmath269 and each @xmath252 sub - principal matrix of @xmath256 are invertible or when both of the sdes have no reflection boundaries .
by theorem [ gtheoremo ] , we suppose that the @xmath306-tuple @xmath307 in is part of a solution @xmath308 to the non - markvian system of coupled fb - sdes with lvy jumps and skew reflections in
. then , let @xmath309 be the corresponding @xmath310-valued ( @xmath311 ) and @xmath190-adapted control process , whose @xmath71th component @xmath312 for each @xmath69 is the @xmath71th player s control policy .
furthermore , we assume that the utility function for each player @xmath69 is defined by @xmath313 thus , it follows from - , - , and that the value functions @xmath314 in are now well defined .
then , we can introduce the following concepts . by a non - zero - sum sdg to the system in , we mean that each player @xmath69 chooses an optimal policy to maximize his own value function expressed in . furthermore , the value functions @xmath314 do not have to add up to a constant ( e.g. , zero ) , or in other words , the sdg is not necessarily a zero - sum one .
@xmath315 is called a pareto optimal nash equilibrium policy process if , the process is also an optimal one to the sum of all the @xmath20 players value functions at time zero ; no player will profit by unilaterally changing his own policy when all the other players policies keep the same . mathematically , @xmath316 for each @xmath273 and any given admissible control policy @xmath54 , where @xmath317 @xmath318 together with @xmath319 are called satisfying the comparison principle in terms of @xmath54 if , for any two @xmath320 with @xmath161 and any two @xmath321-measurable @xmath322 with associated two solutions @xmath323 , respectively , of such that @xmath324 for all @xmath325\times d$ ] , we have @xmath326 [ gtheorem ] let @xmath298 be the unique adapted 4-tuple strong solution to the @xmath299-dimensional fb - spdes in , which corresponds to specific @xmath300 in - , terminal condition in , and a control process @xmath301 .
suppose that @xmath58 and @xmath57 satisfy the completely-@xmath0 condition .
if @xmath318 together with @xmath327 for suitably chosen @xmath328 and @xmath329 satisfy the comparison principle in terms of @xmath54 , the following two claims are true : 1 . there is a pareto optimal nash equilibrium point @xmath330 to the non - zero - sum sdg problem in when both of the sdes in have no reflection boundaries and if @xmath331 ; 2 .
there is an approximated pareto optimal nash equilibrium point @xmath330 to the non - zero - sum sdg problem in when at least one of the sdes in has reflection boundary and if @xmath328 , @xmath329 are taken to be infinitely smooth approximated functions of @xmath332(t , x ) and @xmath333 in @xmath222 .
queueing networks widely appear in real - world applications such as those in service , cloud computing , and communication systems .
they typically consist of arrival processes , service processes , and buffer storages with certain kind of service regime and network architecture ( see , e.g. , an example with @xmath13-job classes in figure [ queuei ] ) .
the major performance measure for this system is the _ queue length process _ denoted by @xmath334 , where @xmath335 is the number of @xmath179th class jobs stored in the @xmath179th buffer for each @xmath234 at time @xmath169 .
let @xmath336 be the initial queue length for the system .
then , the queueing dynamics of the system can be presented by @xmath337 where , the @xmath179th component @xmath338 of @xmath339 for each @xmath234 is the total number of jobs arrived to buffer @xmath179 by time @xmath169 , and the @xmath179th component @xmath340 of @xmath341 is the total number of jobs departed from buffer @xmath179 by time @xmath169 . in the following discussions , we use two generalized ways to characterize the arrival and departure processes .
first , we assume that each @xmath342 for @xmath234 is a time - inhomogeneous lvy process with intensity measure @xmath343 that is the job arrival rate to buffer @xmath179 at time @xmath169 and depends on the queue state at time @xmath169 .
similarly , we assume that each @xmath344 is also a time - inhomogeneous lvy process with intensity measure @xmath345 that is the assigned service rate to buffer @xmath179 at time @xmath169 .
furthermore , we assume that the routing proportion from buffer @xmath279 to buffer @xmath179 for jobs finishing service at buffer @xmath279 is @xmath346 .
then , by the f - sde in and the discussions in applebaum @xcite , the queue length process in for this case can be further expressed by @xmath347 where , @xmath348 , @xmath349 in for each @xmath285 is the skorohod regulator process and it can increase only at time @xmath169 when @xmath350 .
note that @xmath351 is a reflection matrix that may be time and queue state dependent , and the coefficients in may be discontinuous at the queue state @xmath352 .
however , since the system in is designed in a controllable manner , the service rate @xmath353 can always be set to be zero when @xmath352 , which implies that the reflection part in can be removed .
hence , the generalized lipschitz and linear growth conditions in - may be reasonably imposed to the system in .
thus , the system derived in can be well - posed .
furthermore , the optimal policies in terms of cost , profit , and system performance can be designed and analyzed ( see , e.g. , the related illustration in the coming subsection [ queuegame ] ) .
interested readers can also find some specific formulations of the queueing system in mandelbaum and massey @xcite , mandelbaum and pats @xcite , and konstantopoulos _ et al . _
@xcite , etc .
second , we assume that both the arrival and service processes are described by renewal processes , renewal reward processes , or doubly stochastic renewal processes . in this case , the driven processes for the queueing system do not have the nice statistical properties such as memoryless and stationary increment ones . thus , it is usually impossible to conduct exact analysis concerning the distribution of @xmath354 .
however , under certain conditions ( e.g. , the arrival rates close to the associated service rates ) , one can show that the corresponding sequence of diffusion - scaled queue length processes converges in distribution to a @xmath13-dimensional reflecting brownian motion ( rbm ) ( see , e.g. , dai @xcite , dai and dai @xcite , dai and jiang @xcite ) , or more generally , a reflecting diffusion with regime switching ( rdrs ) ( see , e.g. , dai @xcite ) .
in other words , we have that @xmath355 where @xmath356 " means converges in distribution " and @xmath357 is a rbm or a rdrs . to be simple ,
we consider the case that the limit @xmath357 in is a rbm living in the state space @xmath67 introduced in section [ fbsdes ] .
furthermore , let @xmath358 be a vector in @xmath14 and @xmath359 be a @xmath252 symmetric and positive definite matrix .
then , we can introduce the definition of a rbm ( see , e.g , dai @xcite ) as follows .
[ srbm ] a semimartingale rbm associated with the data @xmath360 that has initial distribution @xmath361 is a continuous , @xmath190-adapted , @xmath13-dimensional process @xmath362 defined on some filtered probability space @xmath363 such that under @xmath364 , @xmath365 where 1 .
@xmath56 has continuous paths in @xmath366 , @xmath364-a.s .
, 2 . under @xmath364 , @xmath362 is a @xmath13-dimensional brownian motion with drift vector @xmath358 and covariance matrix @xmath359 such that @xmath367 is a martingale and @xmath368 , 3 .
@xmath59 is a @xmath190-adapted , @xmath65-dimensional process such that @xmath364-a.s .
, for each @xmath61 , the @xmath179th component @xmath369 of @xmath59 satisfies 1 .
@xmath370 , 2 .
@xmath369 is continuous and non - decreasing , 3 .
@xmath369 can increase only when z is on the face @xmath60 , i.e. , as given in . from the physical viewpoint of queueing system (
see , e.g. , dai @xcite ) and the discussion in reiman and williams @xcite , the pushing process @xmath59 in definition [ srbm ] can be assumed to a.s .
satisfy @xmath371 now , assume that @xmath216 is the stationary distribution that we expect for the rbm @xmath56 .
for example , in reality , it is the given distribution of the long - run average queue lengths among different users or job classes .
theoretically , it can be computed by a method ( e.g. , the finite element method designed and implemented in dai _
et al . _
then , we can use a b - pde or a b - spde ( a special form of the system in ) to get the transition function at each time point to reach the targeted or limiting stationary distribution @xmath216 for the rbm @xmath56 for a given initial distribution ( e.g. , @xmath372 a.s . in many situations ) .
hence , the corresponding performance measures of the physical queueing system can be estimated .
more precisely , we have the following theorem and related remark . [ srbmth ]
suppose that the reflection matrix satisfies the completely-@xmath373 condition .
then , the transition function of the rbm @xmath56 over @xmath220 $ ] is determined by @xmath374 where @xmath18 is a @xmath375-dimensional function .
furthermore , @xmath24 is the following form of partial differential operator @xmath376 where @xmath377 is the gradient vector of @xmath18 in @xmath222 and @xmath378 is the indicator function over the set @xmath379 .
it follows from the completely-@xmath0 condition that the rbm @xmath56 is a strong markov process ( see , e.g. , dai and williams @xcite ) . then , by applying the it@xmath380 s formula ( see , e.g. , protter @xcite ) and fokker - planck s formula ( or called kolmogorov s forward / backward equations , see , e.g. , @xmath50ksendal @xcite ) , we know that the claim stated in the theorem is true . @xmath381 owing to the uncertainty error of measurement , @xmath216 could be random .
furthermore , the coefficients in may also be random , e.g. , for the case that the limit @xmath357 is a rdrs .
thus , a b - spde can be introduced .
furthermore , the indicator function @xmath382 can be approximated by a sufficient smooth function in order to apply theorem [ bsdeyi ] to the equation in , which is reasonable from the viewpoint of numerical computation . from the information system displayed in figures [ gamei]-[queuei ] ( presenting a parallel - server queueing system with @xmath383 )
, we can give an explanation about the decision process for such a game problem .
in this game , each player ( or called user in dai @xcite ) relates to a control process @xmath312 for @xmath69 over certain resource pool ( e.g. , called the transmission rate allocation process over a randomly evolving capacity region in dai @xcite ) . in the meanwhile , each player @xmath71 is assigned a surrogate utility function @xmath384 of his submitted bid ( called queue length in dai @xcite , or the approximated queue length rbm @xmath362 in definition [ srbm ] ) to the network , the price from the network to him , and the control policy at each time point by the central information administrative .
then , an optimal and/or fair control process can be determined by the utility functions of all players , queueing process , and the available resource constraint in a cooperative way ( see , e.g. , jones @xcite ) .
we justify the two theorems by first proving three lemmas in the following subsection .
[ martindecom ] assume that the conditions in theorem [ bsdeyi ] hold and take a quadruplet for each fixed @xmath131 and @xmath146 , @xmath385\times d).\elabel{threeu}\end{aligned}\ ] ] then , there exists another quadruplet @xmath386 such that @xmath387 where @xmath388 is a @xmath190-adapted cdlg process and @xmath389 is the corresponding predictable process . furthermore , for each @xmath131 , @xmath390<\infty , \elabel{maxnormo}\\ & & e\left[\int_{0}^{t}\|v^{2}(t , x)\|^{2}dt\right]<\infty , \elabel{maxnormi}\\ & & e\left[\int_{0}^{t}\|\bar{v}^{2}(t , x)\|^{2}dt\right]<\infty , \elabel{maxnormii}\\ & & e\left[\sum_{i=1}^{h}\int_{0}^{t}\int_{{\cal z } } \left\|\tilde{v}^{2}_{i}(t , x , z_{i})\right\|^{2 } \nu_{i}(dz_{i})dt\right]<\infty .
\elabel{maxnormiii}\end{aligned}\ ] ] for each fixed @xmath131 and a quadruplet as stated in , it follows from conditions - that @xmath391,c^{\infty}(d , r^{r } ) ) , \elabel{ladaptedi}\\ & & { \cal j}(\cdot , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\in l^{2}_{{\cal f}}([0,t],c^{\infty}(d , r^{r\times d } ) ) , \elabel{ladaptediai}\\ & & { \cal i}(\cdot , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\in l^{2}_{{\cal f}}([0,t]\times{\cal z}^{h},c^{\infty}(d , r^{r\times h } ) ) , \elabel{ladaptediaii}\\ & & \bar{{\cal l}}(\cdot , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\in l^{2}_{{\cal f}}([0,t],c^{\infty}(d , r^{q } ) ) , \elabel{fbladaptedi}\\ & & \bar{{\cal j}}(\cdot , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\in l^{2}_{{\cal f}}([0,t],c^{\infty}(d , r^{q\times d } ) ) , \elabel{fbladaptediai}\\ & & \bar{{\cal i}}(\cdot , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\in l^{2}_{{\cal f}}([0,t]\times{\cal z}^{h},c^{\infty}(d , r^{q\times h } ) ) .
\elabel{fbladaptediaii}\end{aligned}\ ] ] by considering @xmath24 , @xmath26 , and @xmath28 in - as new starting @xmath392 , @xmath393 , and @xmath394 , we can define @xmath395 by the forward iteration in .
furthermore , @xmath395 is a @xmath266-adapted cdlg process that is square - integrable for each @xmath131 in the sense of .
now , consider @xmath33 , @xmath35 , and @xmath396 in - as new starting @xmath397 , @xmath398 , and @xmath399 .
then , it follows from the martingale representation theorem ( see , e.g. , theorem 5.3.5 in page 266 of applebaum @xcite ) that there are unique predictable processes @xmath400 and @xmath401 such that @xmath402 \nonumber\\ & = & \hat{v}^{2}(0,x)+\int_{0}^{t}\bar{v}^{2}(s^{-},x)dw(s ) + \int_{0}^{t}\int_{{\cal z}}\tilde{v}^{2}(s^{-},x , z)\tilde{n}(\lambda ds , dz ) .
\nonumber\end{aligned}\ ] ] furthermore , @xmath403 and @xmath404 are square - integrable for each @xmath131 in the sense of - , and @xmath405 owing to the corollary in page 8 of protter @xcite , @xmath406 can be taken as a cdlg process .
now , define a process @xmath407 given by @xmath408 .
\nonumber\end{aligned}\ ] ] thus , it follows from - and simple calculation that @xmath409 is square - integrable in the sense of .
in addition , by - , we know that @xmath410 which implies that @xmath409 is a cdlg process .
hence , for a given quadruplet in , it follows from - and that the associated quadruplet ( @xmath411 , @xmath412 @xmath400 , @xmath401 ) satisfies the equation as stated in the lemma .
furthermore , we know that @xmath413 thus , we complete the proof of lemma [ martindecom ] . @xmath381 [ differentiablev ] under the conditions of theorem [ bsdeyi ] , consider a quadruplet as in for each fixed @xmath131 and @xmath146 .
define @xmath414 by .
then , @xmath415 , @xmath416 , @xmath417 , @xmath418 for each @xmath419 exists a.s . and satisfies @xmath420 where @xmath113 and @xmath110 with @xmath111 .
furthermore , @xmath421 , @xmath422 for each @xmath423 is a @xmath266-adapted cdlg process and @xmath424 , @xmath425 is the associated predictable processes .
all of them are squarely - integrable in the senses of - . without loss of generality
, we only consider the point @xmath131 , which is an interior one of @xmath67 .
otherwise , we can use the corresponding derivative in a one - side manner to replace the one in the following proof .
first , we show that the claim in the lemma is true for @xmath426
. to do so , for each given @xmath90,x\in d , z\in{\cal z}^{h}$ ] , and @xmath427 as in the lemma , let @xmath428 be defined by but each @xmath429 is replaced by its first - order partial derivative @xmath430 with respect to @xmath431 for @xmath111 if @xmath432 .
then , we can show that the quadruplet defined in for each @xmath71 is the required first - order partial derivative of @xmath1 in for the given @xmath433 .
in fact , considering an interior point @xmath222 of @xmath67 , we can take sufficiently small constant @xmath434 such that @xmath435 , where @xmath436 is the unit vector whose @xmath71th component is one and others are zero . without loss of generality ,
we assume that @xmath437 . then
, for each @xmath438 and @xmath432 with @xmath111 , we define @xmath439 furthermore , let @xmath440 and let @xmath441 for each @xmath442 .
now , let tr@xmath443 denote the trace of the matrix @xmath444 for a given matrix @xmath94 and let @xmath445 be the @xmath279th term in the summation of the trace .
furthermore , for each fixed @xmath90 $ ] , @xmath437 , and @xmath446 , define @xmath447 then , it follows from and the it@xmath380 s formula ( see , e.g. , theorem 1.14 and theorem 1.16 in pages 6 - 9 of @xmath50ksendal and sulem @xcite ) that @xmath448 if , in the last equality , we take @xmath449 note that @xmath450 in is a martingale of the form , @xmath451 thus , by the martingale property and , we know that @xmath452 \nonumber\\ & \leq&\hat{\gamma}e\left[\int_{0}^{t}\left\|\delta{\cal l}^{(1)}_{i_{l},\delta}(s , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\|^{2}e^{2\gamma s}ds\right ] \nonumber\\ & & + \hat{\gamma}\left[\int_{t}^{t}\left\|\delta\bar{{\cal l}}^{(1)}_{i_{l},\delta}(s , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\|^{2}e^{2\gamma s}ds\right ] . \nonumber\end{aligned}\ ] ] furthermore , by - and the burkholder - davis - gundy s inequality ( see , e.g. , theorem 48 in page 193 of protter @xcite ) , we have the following observation , @xmath453 \elabel{emdelta}\\ & & \leq \hat{\gamma}k_{1}e\left[\int_{0}^{t}\left\|\delta{\cal l}^{(1)}_{i_{l},\delta}(s , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\|^{2}e^{2\gamma s}ds\right ] \nonumber\\ & & \;\;\;+\hat{\gamma}k_{1}\left[\int_{t}^{t}\left\|\delta\bar{{\cal l}}^{(1)}_{i_{l},\delta}(s , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\|^{2}e^{2\gamma s}ds\right ] , \nonumber\end{aligned}\ ] ] where @xmath454 is some nonnegative constant depending only on @xmath455 , @xmath456 , @xmath203 , and @xmath22 .
note that , the detailed estimation procedure for the quantity on the right - hand side of is postponed to the same argument used for in the proof of lemma [ lemmathree ] since more exact calculations are required there .
next , for each fixed @xmath90 $ ] , @xmath131 , and @xmath457 , consider the random variable set @xmath458 , @xmath459\}$ ] .
it follows from lemma 1.3 in pages 6 - 7 of peskir and shiryaev @xcite that there is a countable subset @xmath460 $ ] such that @xmath461}z_{\delta}(t , x)=\sup_{\delta\in{\cal c}}z_{\delta}(t , x),\;\;\mbox{a.s .
} , \elabel{esssupeq}\end{aligned}\ ] ] where `` esssup '' denotes the essential supremum .
furthermore , take @xmath462 where @xmath463 for any two real numbers @xmath464 and @xmath465 .
obviously , @xmath466 the second inequality in implies that the set @xmath467 is upwards directed .
hence , for each @xmath90 $ ] , @xmath131 , @xmath457 , and the associated sequence of @xmath468 , it follows from that @xmath469 \elabel{difsuco}\\ & \leq&e\left[\mbox{esssup}_{\delta\in{\cal c } } \bar{z}_{\delta}(t , x)\right ] \nonumber\\ & = & \lim_{n\rightarrow\infty}e\left[\bar{z}_{\delta_{n}}(t , x)\right ] \nonumber\\ & = & \lim_{n\rightarrow\infty } e\left[\max_{\delta\in\{\delta_{1}, ... ,\delta_{n}\}}z_{\delta}(t , x)\right ] .
\nonumber\end{aligned}\ ] ] in addition , for each fixed @xmath470 , let @xmath471 thus , by the induction method in terms of @xmath472 and , we know that @xmath473 \elabel{difsucoi}\\ & \leq&\hat{\gamma}\lim_{n\rightarrow\infty } e\left[\int_{0}^{t}\max_{\delta\in\{\delta_{1}, ... ,\delta_{n}\ } } \left\|\delta{\cal l}^{(1)}_{i_{l},\delta}(s , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\|^{2}e^{2\gamma s}ds\right . \nonumber\\ & & \left.+\int_{t}^{t}\max_{\delta\in\{\delta_{1}, ... ,\delta_{n}\ } } \left\|\delta\bar{{\cal l}}^{(1)}_{i_{l},\delta}(s , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\|^{2}e^{2\gamma s}ds\right ] \nonumber\\ & & -\lim_{n\rightarrow\infty}e\left[\bar{m}_{\delta_{n}}(t , x)\right ] \nonumber\\ & \leq & ke\left[\int_{0}^{t}\mbox{esssup}_{0\leq\delta\leq\sigma } \left\|\delta{\cal l}^{(1)}_{i_{l},\delta}(s , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\|^{2}e^{2\gamma s}ds\right.\nonumber\\ & & \;\;\;\;\;\left.+\int_{t}^{t}\mbox{esssup}_{0\leq\delta\leq\sigma } \left\|\delta\bar{{\cal l}}^{(1)}_{i_{l},\delta}(s , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\|^{2}e^{2\gamma s}ds\right ] \nonumber\\ & & \;\;\;\;\;+\int_{0}^{t}\mbox{esssup}_{0\leq\delta\leq\sigma}\left\|\delta{\cal j}^{(1)}_{i_{l},\delta}(s^{-},x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\|^{2 } e^{2\gamma
s}ds \nonumber\\ & & \;\;\;\;\;+\left.\int_{0}^{t}\sum_{i=1}^{h}\int_{{\cal z } } \mbox{esssup}_{0\leq\delta\leq\sigma}\left\|\delta{\cal i}^{(1)}_{i , i_{l},\delta}(s^{-},x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1},z_{i})\right\|^{2}e^{2\gamma s}\lambda_{i}\nu_{i}(dz_{i})ds\right ] , \nonumber \ ] ] where @xmath474 is a nonnegative constant depending only on @xmath455 , @xmath22 , @xmath203 , and @xmath475 .
note that , in the second inequality , we have used the fact in and the following observation @xmath476\right| \leq e\left[\sup_{t\in[0,t]}\left\|m_{\delta_{n}}(t , x)\right\|\right ] + e\left[\sup_{t\in[0,t]}\left\|m_{\delta_{n-1}}(t , x)\right\|\right ] .
\elabel{expectc}\end{aligned}\ ] ] now , recall the condition that @xmath385\times d ) .
\nonumber\end{aligned}\ ] ] then , for each @xmath131 , @xmath146 , any @xmath423 , and any small number @xmath477 such that @xmath478 , we have that @xmath479 note that the related quantities on the right - hand side of are squarely integrable a.s . in term of the lebesgue measure and/or the lvy measure .
therefore , @xmath480 ( the integration of @xmath481 with respect to the lvy measure ) is also infinitely smooth in each @xmath131 due to the lebesgue s dominated convergence theorem .
thus , by the mean - value theorem , there exist some constants @xmath482 and @xmath483 , which depend on @xmath434 , such that @xmath484 a.s . for each @xmath485 .
due to , , and , the quantity on the left - hand side of for all @xmath434 is dominated by a squarely - integrable random variable in terms of the product measure @xmath486 . similarly , for @xmath487 and each @xmath488 , we a.s .
have that @xmath489 owing to , , and , the quantity on the left - hand side of for all @xmath434 is dominated by a squarely - integrable random variable in terms of the product measure @xmath490 .
therefore , it follows from - and the lebesgue s dominated convergence theorem that @xmath491 \elabel{difsuci}\\ & \leq & ke\left[\int_{0}^{t}\lim_{\sigma\rightarrow 0}\mbox{esssup}_{0\leq\delta\leq\sigma } \left\|\delta{\cal l}^{(1)}_{i_{l},\delta}(s , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\|^{2}e^{2\gamma s}ds\right . \nonumber\\ & & \;\;\;\;\;\;\;+\int_{t}^{t}\lim_{\sigma\rightarrow 0}\mbox{esssup}_{0\leq\delta\leq\sigma}\left\|\delta\bar{{\cal l}}^{(1)}_{i_{l},\delta}(s , x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\|^{2}e^{2\gamma s}ds \nonumber\\ & & \;\;\;\;\;\;\;+\int_{0}^{t}\lim_{\sigma\rightarrow 0}\mbox{esssup}_{0\leq\delta\leq\sigma}\left\|\delta{\cal j}^{(1)}_{i_{l},\delta}(s^{-},x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1})\right\| e^{2\gamma s}ds \nonumber\\ & & \;\;\;\;\;\;\;+\left.\int_{0}^{t}\sum_{i=1}^{h}\int_{{\cal z } } \lim_{\sigma\rightarrow 0}\mbox{esssup}_{0\leq\delta\leq\sigma}\left\|\delta{\cal i}^{(1)}_{i_{l},\delta}(s^{-},x , u^{1},v^{1},\bar{v}^{1},\tilde{v}^{1},z_{i})\right\|e^{2\gamma s}\lambda_{i}\nu_{i}(dz_{i})ds\right ] . \nonumber \ ] ] hence , by and the fatou s lemma , we know that , for any sequence @xmath492 satisfying @xmath493 along @xmath494 , there is a subsequence @xmath495 such that @xmath496 the convergence in implies that the first - order derivatives of @xmath17 and @xmath18 in terms of @xmath431 for each @xmath111 exists .
more exactly , they equal @xmath497 and @xmath498 a.s .
respectively for each @xmath90 $ ] and @xmath131 .
furthermore , they are @xmath266-adapted .
now , we prove the claim for @xmath29 .
in fact , it follows from the proof as in - that @xmath499 \nonumber \ ] ] is also bounded by the quantity on the right - hand side of .
thus , by and , we know that @xmath500 hence , the first - order derivative of @xmath29 in @xmath431 for each @xmath111 exists and equals @xmath501 a.s .
for every @xmath90 $ ] and @xmath131 .
furthermore , it is a @xmath190-predictable process .
similarly , we can get the conclusion for @xmath502 associated with each @xmath71 , @xmath169 , @xmath222 , and @xmath503 .
second , we suppose that @xmath504 @xmath505 @xmath506 corresponding to a given @xmath507 @xmath508 , @xmath509 @xmath510 @xmath511\times d)$ ] exists for any given @xmath512 .
then , we can show that @xmath513 exists for the given @xmath512 . in fact ,
consider any fixed nonnegative integer numbers @xmath514 satisfying @xmath515 for the given @xmath512 .
take @xmath516 , @xmath111 , and sufficiently small @xmath437 . then , let @xmath517 correspond to the @xmath518th - order partial derivative @xmath519 of @xmath429 via .
similarly , let @xmath520 be defined by , where @xmath429 are replaced by their @xmath108th - order partial derivatives @xmath521 corresponding to a given @xmath522 , @xmath523 , @xmath508 , @xmath524 , @xmath481 .
furthermore , let @xmath525 for each @xmath438 .
then , define @xmath526 for each @xmath442 .
thus , by the it@xmath380 s formula and repeating the procedure as used in the first step , we know that @xmath520 exist for the given @xmath512 and all @xmath111 .
therefore , the claim in is true .
third , by the induction method with respect to @xmath512 and the continuity of all partial derivatives in terms of @xmath131 , we know that the claims in the lemma are true . hence , we finish the proof of lemma [ differentiablev ] . @xmath381 0.4 cm to state and prove the next lemma , let @xmath527,c^{\infty}(d , r^{l}))$ ] with @xmath104 be the set of @xmath127-valued @xmath190-adapted and squarely integrable cdlg processes as in .
furthermore , for any given number sequence @xmath528 with @xmath529 , define @xmath530 $ ] to be the following banach space ( see , e.g. , the related explanation in yong and zhou @xcite , and situ @xcite ) @xmath531&\equiv&d^{2}_{{\cal f}}([0,t],c^{\infty}(d , r^{r } ) ) \elabel{combanach}\\ & & \times d^{2}_{{\cal f}}([0,t],c^{\infty}(d , r^{q } ) ) \nonumber\\ & & \times l^{2}_{{\cal f},p}([0,t],c^{\infty}(d , r^{q\times d}))\nonumber\\ & & \times l^{2}_{p}([0,t]\times r_{+}^{h},c^{\infty}(d , r^{q\times h } ) ) , \nonumber\end{aligned}\ ] ] which is endowed with the norm @xmath532 for any given @xmath533 $ ] , and @xmath534 \elabel{kcomnorm}\\ & & + e\left[\sup_{0\leq t\leq t}\left\|v(t)\right\|^{2}_{c^{c}(d , q)}e^{2\gamma_{c}t}\right ] \nonumber\\ & & + e\left[\int_{0}^{t } \left\|\bar{v}(t)\right\|^{2}_{c^{c}(d , qd)}e^{2\gamma_{c}t}dt\right ] \nonumber\\ & & + e\left[\int_{0}^{t}\left\|\tilde{v}(t ) \right\|^{2}_{\nu , c}e^{2\gamma_{c}t}dt\right ] .
\nonumber\end{aligned}\ ] ] then , we have the following lemma . [ lemmathree ] under the conditions of theorem [ bsdeyi ] , all the claims in the theorem are true . by , we can define the following map @xmath535 then , we show that @xmath536 forms a contraction mapping in @xmath537 $ ] .
in fact , consider @xmath538 \nonumber\end{aligned}\ ] ] for each @xmath539 , satisfying @xmath540 furthermore , define @xmath541 and take @xmath542 thus , it follows from and the similar argument as used in proving that , for a @xmath543 and each @xmath544 , @xmath545 where @xmath546 is some nonnegative constant depending only on @xmath455 . for the last inequality in , we have taken @xmath547 furthermore , @xmath548 appeared in is given by @xmath549 in addition , @xmath550 in is a martingale of the form , @xmath551 then , it follows from - and the martingale properties related to the it@xmath380 s stochastic integral that @xmath552 \nonumber\\ & \leq&\hat{\gamma}_{0}(t+1)k_{a,0}\left\|(\delta u^{i-1},v^{i-1},\delta\bar{v}^{i-1},\delta\tilde{v}^{i-1 } ) \right\|^{2}_{{\cal m}^{d}_{\gamma_{0},k}}. \nonumber\end{aligned}\ ] ] next , it follows from that @xmath553 \elabel{vitodii}\\ & \leq&2\sum_{j=1}^{d}e\left[\sup_{0\leq t\leq t}\left|\int_{0}^{t}\left(\delta u^{i}(s^{-},x)\right)'\right.\right.\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\ ; \left.\left.\delta{\cal j}_{j}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1 } ) e^{2\gamma_{0 } s}dw_{j}(s)\right|\right ] \nonumber\\ & & + 2\sum_{j=1}^{h}e\left[\sup_{0\leq t\leq t}\left|\int_{0}^{t } \int_{{\cal z}}\left(\delta u^{i}(s^{-},x)\right)'\right.\right.\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\left.\left.\delta{\cal i}_{j}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1},z_{j } ) e^{2\gamma_{0 } s}\tilde{n}(\lambda_{j}ds , dz_{j})\right|\right ] \nonumber\\ & & + 4\sum_{j=1}^{d}e\left[\sup_{0\leq t\leq t}\left|\int_{0}^{t}\left(\delta v^{i}(s^{-},x)\right)'\right.\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\left(\delta\bar{{\cal j}}_{j}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1 } ) \right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\left.\left.\left.+(\delta\bar{v}^{i-1})_{j } ( s^{-},x)-(\delta\bar{v}^{i})_{j } ( s^{-},x)\right)e^{2\gamma_{0 } s}dw_{j}(s)\right|\right ] \nonumber\\ & & + 4\sum_{j=1}^{h}e\left[\sup_{0\leq t\leq t}\left|\int_{0}^{t } \int_{{\cal z}}\left(\delta v^{i}(s^{-},x)\right)'\right.\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\left(\delta\bar{{\cal i}}_{j}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1},z_{j})\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\left.\left.\left.+(\delta\tilde{v}^{i-1})_{j}(s^{-},x , z_{j } ) -(\delta\tilde{v}^{i})_{j}(s^{-},x , z_{j})\right)e^{2\gamma_{0 } s}\tilde{n}(\lambda_{j}ds , dz_{j})\right|\right ] .
\nonumber\end{aligned}\ ] ] by the burkholder - davis - gundy s inequality ( see , e.g. , theorem 48 in page 193 of protter @xcite ) , the right - hand side of the inequality in is bounded by @xmath554 \nonumber\\ & & + \sum_{j=1}^{h}e\left[\left(\int_{0}^{t } \int_{{\cal z}}\left\|\delta u^{i}(s^{-},x)\right\|^{2}\right.\right . \nonumber\\ & & \;\;\;\;\;\left.\left.\left\|\delta{\cal i}_{j}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1},z_{j})\right\|^{2 } e^{4\gamma_{0 } s}\lambda_{j}\nu_{j}(dz_{j})ds\right)^{\frac{1}{2}}\right ] \nonumber\\ & & + \sum_{j=1}^{d}e\left[\left(\int_{0}^{t}\left\|\delta v^{i}(s^{-},x)\right\|^{2}\right.\right.\nonumber\\ & & \;\;\;\;\;\left\|(\delta\bar{{\cal j}}^{i})_{j}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1})\right .
\nonumber\\ & & \;\;\;\;\;\left.\left.\left.+(\delta\bar{v}^{i-1})_{j } ( s^{-},x)-(\delta\bar{v}^{i})_{j } ( s^{-},x)\right\|^{2}e^{4\gamma_{0 } s}ds\right)^{\frac{1}{2 } } \right]\nonumber\\ & & + \sum_{j=1}^{h}e\left[\left(\int_{0}^{t } \int_{{\cal z}}\left\|\delta v^{i}(s^{-},x)\right\|^{2}\right.\right .
\nonumber\\ & & \;\;\;\;\;\left\|\delta\bar{{\cal i}}_{j}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1},z_{j})\right .
\nonumber\\ & & \;\;\;\;\;\left.\left.\left.\left.+(\delta\tilde{v}^{i})_{j}(s^{-},x , z_{j } ) -(\delta\tilde{v}^{i})_{j}(s^{-},x , z_{j})\right\|^{2}e^{4\gamma_{0 } s}\lambda_{j}\nu_{j}(dz_{j})ds\right)^{\frac{1}{2}}\right]\right ) , \nonumber\end{aligned}\ ] ] where @xmath555 is some nonnegative constant depending only on @xmath455 and @xmath203 .
furthermore , it follows from the direct observation that the quantity in is bounded by @xmath556 \nonumber\\ & + & \left(e\left[\left(\sup_{0\leq t\leq t}\|\delta v^{i}(t , x)\|^{2}e^{2\gamma_{0 } t}\right)^{\frac{1}{2}}\right.\right . \nonumber\\ & & \left(\sum_{j=1}^{d}\left(\int_{0}^{t } \left\|\delta\bar{{\cal j}}_{j}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1})\right.\right.\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\ ; \left.\left.+(\delta\bar{v}^{i-1})_{j}(s^{-},x)-(\delta\bar{v}^{i})_{j}(s^{-},x)\right\|^{2}e^{2\gamma_{0 } s}ds\right)^{\frac{1}{2 } } \nonumber\\ & & \;\;+\sum_{j=1}^{h}\left(\int_{0}^{t } \int_{{\cal z}}\left\|\delta\bar{{\cal i}}_{j}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1},z_{j})\right.\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\left.\left.\left.\left.\left .
+ ( \delta\tilde{v}^{i-1})_{j}(s^{-},x , z_{j})-(\delta\tilde{v}^{i})_{j}(s^{-},x , z_{j})\right\|^{2 } e^{2\gamma_{0}s}\lambda_{j}\nu_{j}(dz_{j})ds\right)^{\frac{1}{2 } } \right)\right]\right ) . \nonumber\end{aligned}\ ] ] in addition , by the direct computation , we know that the quantity in is dominated by @xmath557 \elabel{iiivitodii}\\ & & + dk_{b,0}^{2}e\left[\int_{0}^{t}\mbox{tr}\left(\delta{\cal j}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1}\right ) e^{2\gamma_{0 } s}ds\right ] \nonumber\\ & & + k_{b,0}^{2}e\left[\sum_{j=1}^{h}\int_{0}^{t}\int_{{\cal z } } \mbox{tr}\left(\delta{\cal i}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1},z_{j})\right)_{j } \right . \nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\left.e^{2\gamma_{0 } s}\lambda_{j}\nu_{j}(dz_{j})ds\right ] \nonumber\\ & & + \frac{1}{2}e\left[\sup_{0\leq t\leq t}\|\delta v^{i}(t , x)\|^{2}e^{2\gamma_{0 } t}\right]\nonumber\\ & & + dk_{b,0}^{2}e\left[\left(\int_{0}^{t}\mbox{tr}\left(\delta\bar{{\cal j}}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1})\right.\right.\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\ ; \left.\left.\left.+\delta\bar{v}^{i-1}(s^{-},x)-\delta\bar{v}^{i}(s^{-},x ) \right)e^{2\gamma_{0 } s}ds\right)\right]\nonumber\\ & & + k_{b,0}^{2}e\left[\sum_{j=1}^{h}\int_{0}^{t}\int_{{\cal z } } \mbox{tr}\left(\delta\bar{{\cal i}}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1},z_{j})\right.\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\left.\left.+\delta\tilde{v}^{i-1}(s^{-},x , z)-\delta\tilde{v}^{i}(s^{-},x , z_{j})\right ) e^{2\gamma_{0 } s}\lambda_{j}\nu_{j}(dz_{j})ds\right ] .
\nonumber\end{aligned}\ ] ] due to , the quantity in is bounded by @xmath558 + e\left[\sup_{0\leq t\leq
t}\|\delta u^{i}(t)\|^{2}_{c^{0}(q)}e^{2\gamma_{0 } t}\right]\right ) \elabel{ivvitodii}\\ & & + \hat{\gamma}_{0}(t+1)dk_{a,0}k_{b,0}^{2}\left\|(\delta u^{i-1},\delta v^{i-1},\delta\bar{v}^{i-1},\delta\tilde{v}^{i-1 } ) \right\|^{2}_{{\cal m}^{d}_{\gamma_{0},k } } , \nonumber\end{aligned}\ ] ] where @xmath546 is some nonnegative constant depending only on @xmath203 , @xmath22 , and @xmath455 .
thus , it follows from and - that @xmath559+e\left[\sup_{0\leq t\leq t}\left\|\delta v^{i}(t)\right\|^{2}_{c^{0}(q)}e^{2\gamma_{0 } t}\right ] \elabel{firstineu}\\ & & \leq 2\left(1+dk_{b,0}^{2}\right)k_{a,0}\hat{\gamma}_{0}(t+1)\left\|(\delta u^{i-1},\delta v^{i-1},\delta\bar{v}^{i-1},\delta\tilde{v}^{i-1 } ) \right\|^{2}_{{\cal m}^{d}_{\gamma_{0},k}}. \nonumber\end{aligned}\ ] ] furthermore , it follows from and that , for @xmath560 , @xmath561 \elabel{secondineu}\\ & \leq & 2e\left[\int_{t}^{t}\mbox{tr}\left(\delta\bar{{\cal j}}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1 } , \tilde{v}^{i-1})\right.\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\ ; \left.\left.+\delta\bar{v}^{i-1}(s^{-},x)-\delta\bar{v}^{i}(s , x ) \right)e^{2\gamma_{0 } s}ds\right]\nonumber\\ & & + 2e\left[\int_{t}^{t}\mbox{tr}\left(\delta\bar{{\cal j}}(s^{-},x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1},\tilde{v}^{i-1 } ) \right.\right.\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\left.\left.+\delta\bar{v}^{i-1}(s^{-},x)\right)e^{2\gamma_{0 } s}ds\right ] \nonumber\\ & \leq&2\hat{\gamma}_{0}k_{c,0}\left(\left\|(\delta u^{i-1},\delta v^{i-1},\delta\bar{v}^{i-1},\delta\tilde{v}^{i-1 } ) \right\|^{2}_{{\cal m}^{d}_{\gamma_{0},k}}\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\ ; \left.+\left\|(\delta u^{i-2},\delta v^{i-2},\delta\bar{v}^{i-2},\delta\tilde{v}^{i-2 } ) \right\|^{2}_{{\cal m}^{d}_{\gamma_{0},k}}\right ) , \nonumber\end{aligned}\ ] ] where @xmath562 is some nonnegative constant depending only on @xmath455 and @xmath203 .
similarly , it follows from that @xmath563 \elabel{thirdvitodi}\\ & \leq&2e\left[\sum_{j=1}^{h}\int_{t}^{t}\int_{{\cal z } } \left(\mbox{tr}\left(\delta\bar{{\cal i}}(s , x , u^{i},v^{i},\bar{v}^{i},\bar{v}^{i},u^{i-1},v^{i-1 } , \bar{v}^{i-1},\tilde{v}^{i-1},z)\right.\right.\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\ ; \left.\left.\left.+\delta\tilde{v}^{i-1}(s^{-},x , z)-\delta\tilde{v}^{i}(s^{-},x , z ) \right)\right)_{j}e^{2\gamma_{0}s}\lambda_{j}ds\nu_{j}(dz_{j})\right ] \nonumber\\ & & + 2e\left[\sum_{j=1}^{h}\int_{t}^{t}\int_{{\cal z } } \left(\mbox{tr}\left(\delta\bar{{\cal i}}(s , x , u^{i},v^{i},\bar{v}^{i},\tilde{v}^{i},u^{i-1},v^{i-1},\bar{v}^{i-1 } , \tilde{v}^{i-1},z)\right.\right.\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\ ; \left.\left.\left.+\delta\tilde{v}^{i-1}(s^{-},x , z ) \right)\right)_{j}e^{2\gamma_{0}s}\lambda_{j}ds\nu_{j}(dz_{j})\right ] \nonumber\\ & \leq&2\hat{\gamma}_{0}k_{c,0}\left(\left\|(\delta u^{i-1},\delta v^{i-1},\delta\bar{v}^{i-1},\delta\tilde{v}^{i-1 } ) \right\|^{2}_{{\cal m}^{d}_{\gamma_{0},k}}\right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\ ; \left.+\left\|(\delta u^{i-2},\delta v^{i-2},\delta\bar{v}^{i-2},\delta\tilde{v}^{i-2 } ) \right\|^{2}_{{\cal m}^{d}_{\gamma_{0},k}}\right ) .
\nonumber\end{aligned}\ ] ] thus , by , - , and the fact that all functions and norms used in this paper are continuous in terms of @xmath222 , we have @xmath564 where @xmath565 is some nonnegative constant depending only on @xmath455 and @xmath203 . now , by lemma [ differentiablev ] and the similar construction as in , for each @xmath512 , we can define @xmath566 where @xmath567 then , it follows from the it@xmath380 s formula and the similar discussion for that @xmath568 where , for the last inequality of , we have taken the number sequence @xmath475 such that @xmath569 and @xmath570 for some @xmath437 such that @xmath571 is sufficiently small .
hence , we have @xmath572 since @xmath573 for @xmath574 , we have @xmath575 therefore , by , we know that @xmath576 thus , from , we see that @xmath577 with @xmath539 forms a cauchy sequence in @xmath530 $ ] , which implies that there is some @xmath1 such that @xmath578 .
\elabel{finalcon}\end{aligned}\ ] ] finally , by and the similar procedure as used for theorem 5.2.1 in pages 68 - 71 of @xmath50ksendal @xcite , we can complete the proof of lemma [ lemmathree ] .
@xmath381 by combining lemmas [ martindecom]-[lemmathree ] , we can reach a proof for theorem [ bsdeyi ] .
@xmath381 first , we consider a real - valued system corresponding to the case that @xmath579 , whose proof is along the line of the one for lemma [ lemmathree ] .
more precisely , for any given number sequence @xmath580 with @xmath581 , replace the norm for the banach space @xmath537 $ ] defined in by the one @xmath582 for any given @xmath1 in this space , where @xmath583 \nonumber\\ & & + e\left[\sup_{0\leq t\leq t}\left\|v(t)\right\|^{2}_{c^{c}(d_{c},q)}e^{2\gamma_{d_{c}}t}\right ] \nonumber\\ & & + e\left[\int_{0}^{t } \left\|\bar{v}(t)\right\|^{2}_{c^{c}(d_{c},qd)}e^{2\gamma_{d_{c}}t}dt\right ] \nonumber\\ & & + e\left[\int_{0}^{t}\left\|\tilde{v}(t ) \right\|^{2}_{\nu , c}e^{2\gamma_{d_{c}}t}dt\right ] .
\nonumber\end{aligned}\ ] ] then , it follows from the similar argument used for in the proof of lemma [ lemmathree ] that @xmath584\times d ) \nonumber\end{aligned}\ ] ] with @xmath585 , where @xmath433 is defined through in lemma [ martindecom ] . furthermore , over each @xmath586 with @xmath423 , we have that @xmath587 where @xmath434 is a constant that can be determined by suitably choosing a number sequence @xmath475 such that @xmath588 and @xmath589 ( note that @xmath590 may depend on both @xmath586 and @xmath108 for each @xmath423 ) .
thus , it follows from that the remaining justification for theorem [ infdn ] can be conducted along the line of proof for theorem [ bsdeyi ] .
second , we consider a real - valued system corresponding to the case that @xmath204 is a general stopping time . the proof for this case can be accomplished by extending the proof corresponding to @xmath579 via the techniques developed in dai @xcite for both forward and backward sdes , and the related discussions in yong and zhou @xcite .
third , by direct generalizing the discussion concerning the real - valued system to complex - valued system , we reach a proof for theorem [ infdn ]
to provide the proofs for theorem [ fbexist ] and theorem [ gtheorem ] , we first recall the skorohod problem and study some related properties .
let @xmath591,r^{b})$ ] with @xmath240 be the space of all functions @xmath592\rightarrow r^{b}$ ] that are right - continuous with left limits and are endowed with skorohod topology ( see , e.g. , billingsley @xcite , jacod and shiryaev @xcite ) .
then , we can introduce the skorohod problem as follows .
( the skorohod problem ) .
given @xmath593,r^{p})$ ] with @xmath594 , a @xmath595-regulation of z over [ 0,t ] is a pair @xmath596,d)\times d([0,t],r_{+}^{b})$ ] such that @xmath597,\nonumber\end{aligned}\ ] ] where , for each @xmath61 , 1 .
@xmath598 , 2 .
@xmath599 is nondecreasing , 3 .
@xmath599 can increase only at a time @xmath600 $ ] with @xmath601 .
furthermore , we define the modulus of continuity with respect to a function @xmath602,r^{b})$ ] and a real number @xmath437 by @xmath603 where the infimum takes over all the finite sets @xmath604 of points satisfying @xmath605 and @xmath606 for @xmath607 , and @xmath608)=\sup_{t_{1}\leq s\leq t\leq t_{2}}\|z(t)-z(s)\| . \elabel{osclationd}\end{aligned}\ ] ] then , we have the following lemma .
[ osclemma ] suppose that the reflection matrix @xmath57 in definition satisfies the completely-@xmath0 condition .
then , any @xmath595-regulation @xmath609 of @xmath593,r^{p})$ ] with @xmath610 satisfies the oscillation inequality over @xmath611 $ ] with @xmath612 $ ] @xmath613)&\leq&\kappa\mbox{osc}(z,[t_{1},t_{2 } ] ) , \elabel{oscinq}\\ \mbox{osc}(y,[t_{1},t_{2}])&\leq&\kappa\mbox{osc}(z,[t_{1},t_{2 } ] ) , \elabel{oscinqi}\end{aligned}\ ] ] where @xmath614 is some nonnegative constant depending only on the inward normal vector @xmath615 and the reflection matrix @xmath57 . for each @xmath616 $ ] ,
define @xmath617 since the reflection matrix @xmath57 satisfies the completely-@xmath0 condition , it is easy to check that the linear complementarity problem ( lcp ) @xmath618 is completely solvable ( see also theorem 2.1 in mandelbaum @xcite for the related discussion ) .
furthermore , we can conclude that @xmath619 for some nonnegative constant @xmath620 depending only on the inward normal vector @xmath615 and the reflection matrix @xmath57 .
then , the rest of the proof is the direct conclusion of the one for theorem 3.1 in dai @xcite or the one for theorem 4.2 in dai and dai @xcite .
@xmath381 [ complimentlemma ] assume that @xmath621 along @xmath472 in @xmath591,r^{p})\times d([0,t],r^{b})$ ] and @xmath622 is of bounded variation for each @xmath472 . furthermore , suppose that @xmath623 for all @xmath472 and each @xmath90 $ ] , where @xmath624,r^{b})$ ] is a @xmath65-dimensional bounded vector function .
then , for each @xmath90 $ ] , we have that @xmath625 it follows from the definition in pages 123 - 124 of billingsley @xcite or theorem 1.14 in page 328 of jacod and shiryaev @xcite that there is a sequence @xmath626 of continuous and strictly increasing functions mapping from @xmath220\rightarrow[0,t]$ ] with @xmath627 and @xmath628 such that @xmath629}\left|\gamma_{n}(t)-t\right|\rightarrow 0 , \elabel{changecm}\\ & & \sup_{t\in[0,t]}\left|(x^{n},y^{n})(\gamma_{n}(t))-(x , y)(t)\right| \rightarrow0 . \elabel{changecmi}\end{aligned}\ ] ] then , by the uniform convergence in - and the condition in , we know that @xmath630 where @xmath631 is the inverse function of @xmath632 for each @xmath472 . hence , we complete the proof of lemma [ complimentlemma ] .
@xmath381 we divide the proof of the theorem into four parts : part a ( existence , uniqueness ) , part b , part c , and part d , which correspond to different boundary reflection conditions .
0.3 cm * part a ( existence ) .
* we consider the case that @xmath633 appeared in - is a constant and both of the forward and the backward sdes have reflection boundaries . in this case , we need to prove the claim that there is an adapted weak solution @xmath52 to the system in . in fact
, for a positive integer @xmath65 , let @xmath634,r^{b})$ ] be the space of @xmath635-valued and @xmath266-adapted processes with sample paths in @xmath591,r^{b})$ ] .
furthermore , each @xmath636,r^{b})$ ] is square - integrable in the sense that @xmath637<\infty .
\elabel{adaptednormiy}\end{aligned}\ ] ] in addition , we use @xmath638,r^{b})$ ] to denote the corresponding predictable space .
then , for a given @xmath472 and a 4-tuple @xmath639 , r^{p})\times d^{2}_{{\cal f}}([0,t ] , r^{q})\times d^{2}_{{\cal f},p}([0,t ] , r^{q\times d } ) \elabel{initialp}\\ & & \times d^{2}_{{\cal f},p}([0,t]\times r^{h}_{+ } , r^{q\times h } ) \nonumber\end{aligned}\ ] ] with @xmath640 and @xmath641 , we have the following observation . by the study concerning the continuous dynamic complementarity problem ( dcp ) in bernard and el kharroubi @xcite
( see also the related discussions in mandelbaum @xcite , reiman and williams @xcite ) , theorem [ bsdeyi ] ( and its proof ) in the current paper , there is a 6-tuple @xmath642 , r^{p})\times d^{2}_{{\cal f}}([0,t ] , r^{b})\nonumber\\ & & \;\;\;\;\times d^{2}_{{\cal f}}([0,t ] , r^{q})\times d^{2}_{{\cal f},p}([0,t ] , r^{q\times d } ) \nonumber\\ & & \;\;\;\;\times d^{2}_{{\cal f},p}([0,t]\times{\cal z}^{h } , r^{q\times h})\times d^{2}_{{\cal f}}([0,t ] , r^{q\times\bar{b } } ) \nonumber\end{aligned}\ ] ] for each @xmath472 , satisfying the properties along each sample path : @xmath643 with @xmath644 and @xmath645 with @xmath646 where , @xmath647 furthermore , @xmath648 satisfies the property ( 3 ) in definition [ srbm ] . in other words , @xmath649 is a @xmath65-dimensional @xmath190-adapted process such that the @xmath179th component @xmath650 of @xmath649 for each @xmath61 @xmath364-a.s .
has the properties that @xmath651 , @xmath650 is non - decreasing , and @xmath650 can increase only when @xmath652 is on the boundary face @xmath60 , i.e. , @xmath653 similarly , @xmath654 also satisfies the property ( 3 ) in definition [ srbm ] .
more precisely , @xmath655 is a @xmath20-dimensional @xmath190-adapted process such that the @xmath179th component @xmath656 of @xmath655 for each @xmath244 @xmath364-a.s . has the properties that @xmath657 , @xmath656 is non - decreasing , and @xmath656 can increase only when @xmath658 is on the boundary face @xmath63 , i.e. , @xmath659 next , we prove that the following sequence of stochastic processes along @xmath472 , @xmath660 is relatively compact in the skorohod topology over the space @xmath661&\equiv & d^{2}_{{\cal f}}([0,t ] , r^{p})\times d^{2}_{{\cal f}}([0,t ] , r^{b } ) \elabel{sevenp}\\ & & \times d^{2}_{{\cal f}}([0,t ] , r^{q})\times d^{2}_{{\cal f},p}([0,t ] , r^{q\times d } ) \nonumber\\ & & \times d^{2}_{{\cal f},p}([0,t]\times{\cal z}^{h } , r^{q\times h})\times d^{2}_{{\cal f}}([0,t ] , r^{q\times\bar{b } } ) .
\nonumber\end{aligned}\ ] ] along the line of dai @xcite , dai and dai @xcite , and by corollary 7.4 in page 129 of ethier and kurtz @xcite , it suffices to prove the following two conditions to be true : first , for each @xmath662 and rational @xmath663 , there is a constant @xmath664 such that @xmath665 second , for each @xmath662 and @xmath666 , there is a constant @xmath437 such that @xmath667 to prove the two conditions stated in and , we first define the norm along each sample path @xmath668}&=&\sup_{a\leq t\leq b}\left\|f(t)\right\| \nonumber\end{aligned}\ ] ] for each @xmath669 and each @xmath670 $ ] .
then , we introduce the space for some constant @xmath446 that will be chosen and explained in the following proof , @xmath671&\equiv & d^{2}_{{\cal f}}([0,t ] , r^{p})\times d^{2}_{{\cal f}}([0,t ] , r^{q})\times d^{2}_{{\cal f},p}([0,t ] , r^{q\times d } ) \elabel{mspace}\\ & & \times d^{2}_{{\cal f},p}([0,t]\times{\cal z}^{h } , r^{q\times h } ) \nonumber\end{aligned}\ ] ] endowed with the norm @xmath672 } \elabel{fbnorm}\\ & \equiv&e\left[\sup_{t\in[0,t]}\left(\|x(t)\|^{2}+\|v(t)\|^{2}\right)e^{2\gamma t}\right]+e\left[\int_{0}^{t}\left\|\bar{v}(t)\right\|^{2}e^{2\gamma t}dt\right ] \nonumber\\ & & + e\left[\int_{0}^{t}\left\|\tilde{v}(t,\cdot)\right\|^{2}_{\nu } e^{2\gamma t}dt\right ] \nonumber\end{aligned}\ ] ] for each @xmath673 $ ] .
thus , by lemma [ osclemma ] , there is a positive constant @xmath674 such that @xmath675 ) \nonumber\\ & \leq & c_{1}\left(\left\|x(0)\right\|+\left\|z^{n}\right\|_{[0,t]}\right ) , \nonumber\end{aligned}\ ] ] and @xmath676 ) \nonumber\\ & \leq&\bar{c}_{1}\left(\left\|(v^{n}(t)\right\| + \left\|f^{n}(t)\right\| + \left\|u^{n}\right\|_{[0,t]}\right ) \nonumber\\ & \leq & \bar{c}_{2}\left(1+\left\|x^{n}(t)\right\| + \left\|u^{n-1}\right\|_{[0,t ] } + \left\|u^{n}\right\|_{[0,t]}\right ) , \nonumber\\ & \leq & c_{1}\left(1+\|x(0)\|+\left\|z^{n-1}\right\|_{[0,t ] } + \left\|u^{n-1}\right\|_{[0,t ] } + \left\|u^{n}\right\|_{[0,t]}\right ) , \nonumber\end{aligned}\ ] ] where , @xmath677 and @xmath678 are some nonnegative constants .
furthermore , we have taken @xmath679 , @xmath680 in the third equality of since the uniqueness for the martingale representation is in the sense of up to sets of measure zero in @xmath681 ( see , e.g. , theorem 4.3.4 in page 53 of @xmath50ksendal @xcite and theorem 5.3.5 in page 266 of applebaum @xcite ) .
thus , for each @xmath472 , the given linear growth constant @xmath682 in , and any constant @xmath683 , it follows from the markov s inequality that @xmath684}\right ] .
\elabel{ineqi}\end{aligned}\ ] ] furthermore , by lemma 4.2.8 in page 201 of applebaum @xcite ( or related theorem in page 20 of gihman and skorohod @xcite ) and the linear growth condition , we know that @xmath685}\right ] \elabel{ineqii}\end{aligned}\ ] ] for all nonnegative constant @xmath686 .
in addition , similar to the illustration of , we have that @xmath687}\right ] .
\elabel{ineqiu}\end{aligned}\ ] ] next , by the similar demonstration for and the linear growth condition , we know that @xmath688}\right ] . \elabel{ineqiiua}\end{aligned}\ ] ] furthermore , by the proof of theorem [ bsdeyi ] , we have that @xmath689}\right ] \nonumber\end{aligned}\ ] ] for some nonnegative constants @xmath690 and @xmath691 .
therefore , for each given @xmath662 , it follows from - , suitably chosen constants @xmath474 and @xmath692 , and the initial condition in that there is a nonnegative constant @xmath620 such that @xmath693 thus , the condition in is satisfied by the sequence of @xmath694 .
now , for any @xmath90 $ ] , it follows from the proof of proposition 18 for a bsde with jumps in dai @xcite and lemma [ osclemma ] that @xmath695 } \elabel{iineqiqo}\\ & \leq&k_{\gamma}\left(2l^{2}(t - t)+\left\|(x^{n-1},v^{n-1},\bar{v}^{n-1 } , \tilde{v}^{n-1})\right\|^{2}_{{\cal q}_{\gamma}[t , t]}\right ) \nonumber\\ & \leq&k_{\gamma}\left(2l^{2}(t - t)+e^{2\gamma t}e\left[\left\|v^{n-1}\right\|^{2}_{[t , t]}\right]+e^{2\gamma t } \int_{t}^{t}e\left[\left\|x^{n-1}\right\|^{2}_{[0,s]}\right]ds\right ) \nonumber\\ & & + k_{\gamma}\left\|(u^{n-1},\bar{v}^{n-1},\tilde{v}^{n-1})\right\|^{2}_{{\cal q}_{\gamma}[t , t ] } , \nonumber \ ] ] where , @xmath696 depending only on @xmath85 , @xmath203 , @xmath22 , and @xmath8 for some suitable chosen @xmath446 .
thus , by lemma [ osclemma ] , the it@xmath380 s isometry formula , and , we have that @xmath697}\right ] \elabel{iineqiq}\\ & \leq&\bar{k}_{1}\left(e\left[\|v^{n}(t)\|^{2}\right]+e\left[\|f^{n-1}(t)\|^{2}\right]+\kappa^{2 } e\left[\mbox{osc}(u^{n-1},[t , t])^{2}\right]\right ) \nonumber\\ & \leq&k_{1}\left(1+e\left[\|x^{n}\|^{2}_{[0,t]}\right]+\kappa^{2 } e\left[\mbox{osc}(u^{n-2},[0,t])^{2}\right]+\kappa^{2 } e\left[\mbox{osc}(u^{n-1},[t , t])^{2}\right]\right ) \nonumber\\ & \leq&k_{1}\left(1 + 24\kappa^{2}l^{2}t^{2 } + 24\kappa^{2}l^{2}(t - t)^{2}\right ) + k_{1}e\left[\|x^{n}\|^{2}_{[0,t]}\right ] \nonumber\\ & & + 24k_{1}\kappa^{2}l^{2}t\left(\int_{0}^{t}e\left[\left\|x^{n-2}\right\|^{2}_{[0,s]}\right]ds + e\left[\left\|v^{n-2}\right\|^{2}_{[0,t]}\right]\right ) \nonumber\\ & & + 24k_{1}\kappa^{2}l^{2}(t - t)\left(\int_{t}^{t}e\left[\left\|x^{n-1}\right\|^{2}_{[0,s]}\right]ds + e\left[\left\|v^{n-1}\right\|^{2}_{[t , t]}\right]\right ) \nonumber\\ & & + 24k_{1}\kappa^{2}l^{2}t\left\|(u^{n-2},\bar{v}^{n-2 } , \tilde{v}^{n-2})\right\|^{2}_{{\cal q}_{\gamma}[0,t ] } \nonumber\\ & & + 4k_{1}\kappa^{2}\left\|(u^{n-1},\bar{v}^{n-1 } , \tilde{v}^{n-1})\right\|^{2}_{{\cal q}_{\gamma}[0,t ] } \nonumber\\ & & + 24k_{1}\kappa^{2}l^{2}(t - t)\left\|(u^{n-1},\bar{v}^{n-1 } , \tilde{v}^{n-1})\right\|^{2}_{{\cal q}_{\gamma}[t , t ] } \nonumber\\ & & + 4k_{1}\kappa^{2}\left\|(u^{n},\bar{v}^{n } , \tilde{v}^{n})\right\|^{2}_{{\cal q}_{\gamma}[t , t ] } \nonumber\\ & \leq&k_{3}+k_{2}\left(\int_{0}^{t}e\left[\left\|x^{n-1}\right\|^{2}_{[0,s ] } \right]ds+\left\|(u^{n-2},\bar{v}^{n-2 } , \tilde{v}^{n-2})\right\|^{2}_{{\cal q}_{\gamma}[0,t]}\right . \nonumber\\ & & \left.+\left\|(u^{n-1},\bar{v}^{n-1 } , \tilde{v}^{n-1})\right\|^{2}_{{\cal q}_{\gamma}[0,t ] } + \left\|(u^{n},\bar{v}^{n } , \tilde{v}^{n})\right\|^{2}_{{\cal q}_{\gamma}[0,t]}\right ) , \nonumber\end{aligned}\ ] ] where , @xmath698 for @xmath699 are some nonnegative constants depending only on @xmath203 , @xmath85 , @xmath614 , and @xmath700 $ ] .
furthermore , for any @xmath90 $ ] , we have that @xmath701}\right ] & \leq&2e\left[\left\|x(0)\right\|^{2}\right]+2\kappa^{2 } e\left[\mbox{osc}(z^{n-1},[0,t])^{2}\right ] \elabel{iineqiw}\\ & \leq&2e\left[\left\|x(0)\right\|^{2}\right ] + 6\kappa^{2}l^{2}t^{2 } \nonumber\\ & & + 6\kappa^{2}l^{2}t\left(\int_{0}^{t } e\left[\left\|x^{n-1}\right\|_{[0,s]}^{2}\right]ds + e\left[\left\|v^{n-1}\right\|^{2}_{[0,t]}\right]\right ) \nonumber\\ & & + 6\kappa^{2}l^{2}t\left\|(u^{n-1},\bar{v}^{n-1 } , \tilde{v}^{n-1})\right\|^{2}_{{\cal q}_{\gamma}[0,t ] } \nonumber\\ & \leq&2e\left[\left\|x(0)\right\|^{2}\right ] + 12\kappa^{4}l^{2}t^{2 } + 6\kappa^{2}l^{2}te\left[\left\|v^{2}(t)\right\|\right ] \nonumber\\ & & + 6\kappa^{2}l^{2}t\int_{0}^{t } e\left[\left\|x^{n-1}\right\|_{[0,s]}^{2}\right]ds \nonumber\\ & & + 6\kappa^{2}l^{2}t\left(1 + 2\kappa^{2}\right)\left\|(u^{n-1},\bar{v}^{n-1 } , \tilde{v}^{n-1})\right\|^{2}_{{\cal q}_{\gamma}[0,t]}. \nonumber\end{aligned}\ ] ] therefore , for any @xmath662 and a constant @xmath437 , consider a finite set @xmath604 of points satisfying @xmath605 and @xmath702 with @xmath703 .
it follows from , - , and the similar explanation for that @xmath704 } + \left\|v^{n}\right\|^{2}_{[0,t]}\right]+\left\|(u^{n},\bar{v}^{n } , \tilde{v}^{n})\right\|^{2}_{{\cal q}_{\gamma}[0,t]}\right ) \nonumber\\ & \leq&\frac{3l^{2}\delta}{(\epsilon - l\delta)^{2 } } \left(a_{0}+\sum_{k=1}^{n}\frac{a^{k+1}_{1}t^{k+1}}{(k+1 ) ! } \left(1+k^{k}_{\gamma}\right)+a_{2}\sum_{k=1}^{n}k^{k}_{\gamma}\right ) , \nonumber\end{aligned}\ ] ] where @xmath705 , @xmath706 , and @xmath707 are some constants depending only on @xmath85 , @xmath203 , @xmath22 , and @xmath8 .
furthermore , by lemma 4.2.8 in page 201 of applebaum @xcite ( or related theorem in page 20 of gihman and skorohod @xcite ) and the linear growth condition , we know that @xmath708 + \delta e\left[\left\|v^{n}\right\|_{t}^{2}\right ] + e\left[\left\|(u^{n},\bar{v}^{n},\tilde{v}^{n})\right\|^{2}_{{\cal q}_{\gamma}[0,t]}\right]\right ) \nonumber\\ & \leq&\frac{\bar{\epsilon}}{\epsilon^{2}}+\frac{3l^{2}}{\bar{\epsilon}-3l^{2}\delta } \left(\delta\left(a_{0}+\sum_{k=1}^{n}\frac{a^{k+1}_{1}t^{k+1}}{(k+1 ) ! } \left(1+k^{k}_{\gamma}\right)+a_{2}\sum_{k=1}^{n}k^{k}_{\gamma}\right ) + a_{3}\sum_{k=1}^{n}k^{k}_{\gamma}\right ) \nonumber\end{aligned}\ ] ] for all nonnegative constant @xmath709 , where @xmath710 is some constant depending only on @xmath85 , @xmath203 , @xmath22 , and @xmath8 . similarly , there are some constants @xmath711 , @xmath712 , @xmath713 , and @xmath714 depending only on @xmath85 , @xmath203 , @xmath22 , and @xmath8 such that @xmath715 and @xmath716 hence , for each given @xmath662 , it follows from - and suitably chosen constants @xmath717 , @xmath434 , and @xmath475 that @xmath718 thus , the condition in is true for the sequence of @xmath694 .
hence , by , , and corollary 7.4 in page 129 of ethier and kurtz @xcite , this sequence is relatively compact .
therefore , there is a subsequence of @xmath694 that converges weakly to @xmath719 over the space @xmath720 $ ] . for convenience ,
we suppose that the subsequence is the sequence itself , i.e. , @xmath721 then , by the skorohod representation theorem ( see , e.g. , theorem 1.8 in page 102 of ethier and kurtz @xcite ) , we can assume that the convergence in is a.s . in the skorohod topology .
thus , by the claim ( a ) in theorem 1.14 ( or the claim ( a ) in proposition 2.1 ) of jacod and shiryaev @xcite and the facts that @xmath722 and @xmath649 is nondecreasing , we can conclude that @xmath242 and @xmath59 is nondecreasing .
furthermore , by lemma [ complimentlemma ] and @xmath723 similarly , we know that @xmath251 , @xmath62 is non - decreasing , and @xmath724 therefore , by the lipschitz condition in , we know that @xmath52 satisfies the fb - sdes in a.s .
thus , by the skorohod representation theorem again , it is a weak solution to the fb - sdes in
. 0.3 cm * part a ( uniqueness ) . *
assume that @xmath52 is a weak solution to the fb - sdes in . to prove its uniqueness , we introduce some additional notations .
let @xmath725 , @xmath726 , and define @xmath727 for each @xmath728 and each @xmath729 . in the sequel
, we call a set @xmath730 maximal " if @xmath731 , @xmath732 , and @xmath733 for any @xmath734 such that @xmath735 .
similarly , we can define the maximal set corresponding to a set @xmath736 .
furthermore , let @xmath737 and @xmath738 respectively denote the euclidean distance between @xmath222 and @xmath739 for a point @xmath131 and the euclidean distance between a point @xmath740 and @xmath741 .
then , it follows from lemma 3.2 in dai @xcite or lemma b.1 in dai and williams @xcite that there exist two constants @xmath742 and @xmath743 such that @xmath744 now , for each @xmath745 , @xmath730 , and @xmath736 ( including the empty set ) , we let @xmath746 where @xmath747 and @xmath748 .
thus , by lemmas 4.1 - 4.2 in dai and williams @xcite , we know that @xmath749 where , @xmath91 is the collection of subsets of @xmath750 consisting of all maximal sets in @xmath750 and @xmath751 is defined in the same way in terms of subsets of @xmath752 . for convenience , we order the sets in @xmath753 and @xmath751 .
then , we can define a sequence of 3-dimensional points @xmath754 with @xmath755 by induction .
in fact , since @xmath52 is a weak solution to the fb - sdes in , both @xmath756 and @xmath757 are defined .
thus , if @xmath758 is the first @xmath759 such that @xmath760 , we let @xmath761 furthermore , if @xmath762 has been defined on @xmath763 , we let @xmath764 be the first @xmath765 such that @xmath766 .
then , we can define @xmath767 on @xmath768 , we define @xmath769 , @xmath770 , and @xmath771 . due to the right - continuity of the sample paths of solution @xmath772 by the related property of lvy process driven stochastic integral ( see , e.g. , theorem 4.2.12 in page 204 of applebaum @xcite ) ,
@xmath773 is a nondecreasing sequence of @xmath266-stopping times , satisfying @xmath774 a.s . as @xmath775 .
hence , it suffices to prove the weak uniqueness of @xmath776 for each @xmath777 .
note that both @xmath778 and @xmath779 for each @xmath777 are subsets of cones .
thus , without loss of generality , we assume that both @xmath67 and @xmath68 are cones . therefore , we can prove the weak uniqueness by induction in terms of the numbers of boundary faces of @xmath67 and @xmath68 .
in fact , for the case that @xmath780 , it follows from the uniqueness of the skorohod mapping given by lemma 3.1 in dai @xcite or lemma 4.5 in dai and dai @xcite that the weak uniqueness is true .
now , we suppose that the weak uniqueness is true for the case that @xmath781 with @xmath782 and @xmath783 .
then , we can prove the case for @xmath784 . in this case
, we need to consider two folds indexed by two pairs of @xmath785 and @xmath786 . both of the folds can be proved by the similar discussion for theorem 5.4 in dai and williams @xcite .
therefore , we finish the proof of weak uniqueness .
0.3 cm * part b. * we consider the case that @xmath633 appeared in - is a constant and the spectral radii of @xmath58 and each @xmath252 sub - principal matrix of @xmath256 are strictly less than one . in this case , we need to prove that there is a unique strong adapted solution @xmath52 to the system of in . in fact , it follows from the discussions in reiman and harrison @xcite , dai @xcite , lemma 7.1 and theorem 7.2 in pages 164 - 165 of chen and yao @xcite that there exist two lipschitz continuous mappings @xmath787 and @xmath788 such that @xmath789 for each @xmath472 . then
, it follows from - , the related estimates in part a , and the conventional picard s iterative method , we can reach a proof for the claim in part b. 0.3 cm * part c. * we consider the case that @xmath633 appeared in - is a constant and both of the sdes have no reflection boundaries . in this case
, we need to prove that there is a unique strong adapted solution @xmath52 to the system of in .
in fact , by the related estimates in part a , this case can be proved by directly generalizing the conventional picard s iterative method .
actually , this case is a special one of theorem [ bsdeyi ] or theorem [ infdn ] .
0.3 cm * part d. * we consider the case that @xmath633 appeared in - is a general adapted and mean - squarely integrable stochastic process .
the proofs corresponding to the cases stated in part a , part b , and part c can be accomplished along the lines of proofs for lemma 4.1 in dai @xcite associated with a forward sde under random environment and proposition 18 in dai @xcite for a backward sde under random environment .
the key in the proofs is to introduce the following sequence of @xmath190-stopping times , i.e. , @xmath790 by the condition in , @xmath791 is nondecreasing and a.s
. tends to infinity as @xmath775 .
0.3 cm finally , by summarizing the cases presented in part a to part d , we finish the proof of theorem [ fbexist ] .
@xmath381 for a control process
@xmath792 , it follows from theorem [ bsdeyi ] that the @xmath299-dimensional fb - spdes in with the partial differential operators @xmath793 given by - and terminal condition in indeed admits a well - posed 4-tuple solution @xmath794 , @xmath49 , @xmath795 , @xmath796 .
thus , substituting @xmath797 into the system of coupled fb - sdes in , it follows from theorem [ fbexist ] that the claims in theorem [ gtheoremo ] are true . @xmath381
the proof of part 1 is the direct extension of the single - dimensional case ( i.e. , @xmath798 ) for the related optimal control problem in @xmath50ksendal _ et al .
_ @xcite .
the proof of part 2 can be done as follows . for each @xmath301 and @xmath331
, it follows from theorem [ gtheoremo ] that the regulator processes @xmath799 and @xmath800 exist . since they are nondecreasing with respect to time variable @xmath169 , the derivatives @xmath801 and @xmath333 exist a.e . in terms of time variable @xmath169 along
each sample path a.s . furthermore ,
if each @xmath268 sub - principal matrix of @xmath269 and each @xmath252 sub - principal matrix of @xmath256 are invertible , these derivatives are uniquely determined owing to the skorohod mapping
. nevertheless , if only the general completely-@xmath0 condition is imposed , these derivatives are weakly unique in a probability distribution sense .
in addition , it follows from proposition 7.1 in ethier and kurtz @xcite that these derivatives can be approximated by polynomials in terms of variable @xmath222 for each given @xmath169 , which are denoted by @xmath328 and @xmath329 .
then , the proof for the claim in part 2 follows from the one for the claim in part 1 .
hence , we reach a proof for theorem [ gtheorem ] . @xmath381 v. caselles , g. sapiro , and d. h. chung , vector median filters , morphology , and pde s : theoretical connections , proceedings of international conference on image processing , ieee cs press 4 ( 1999 ) , 177 - 184 . c. z. chang , j. zhang , x. feng , j. shen , z. zhang , m. guo , k. li , y. ou , p. wei , l. l. wang , z. q. ji , y. feng , s. ji , x. chen , j. jia , x. dai , z. fang , s. c. zhang , k. he , y. wang , l. lu , x. c. ma , and q. k. xue , experimental observation of the quantum anomalous hall effect in a magnetic topologic insulator , science 340(167 ) ( 2013 ) .
w. dai , brownian approximations for queueing networks with finite buffers : modeling , heavy traffic analysis and numerical implementations .
ph.d thesis , georgia institute of technology , 1996 . also published in umi dissertation services , a bell @xmath803 howell company , michigan , u.s.a . ,
w. dai , on the traveling neuron nets ( human brains ) controlled by a satellite communication system , proceedings of international conference on bioinformatics and biomedical engineering , pp . 1 - 4 , ieee computer society press , 2009 . w. dai , optimal rate scheduling via utility - maximization for @xmath805-user mimo markov fading wireless channels with cooperation , operations research 61(6 ) ( 2013 ) , 1450 - 1462 ( with 26 page online e - companion ( supplemental ) ) . w. dai , product - form solutions for integrated services packet networks and cloud computing systems , mathematical problems in engineering , volume 2014 ( regular issue ) , article i d 767651 , 16 pages , 2015 .
w. dai , mean - variance hedging based on an incomplete market with external risk factors of non - gaussian ou processes , mathematical problems in engineering , volume 2015 ( regular issue ) , article i d 625289 , 20 pages , 2015 .
i. karatzas and q. li , bsde approach to non - zero - sum stochastic differential games of control and stopping , in stochastic processes , finance and control , pp .
105 - 153 , world scientific publishers , 2012 .
x. shen , h. chen , j. g. dai , and w. dai , the finite element method for computing the stationary distribution of an srbm in a hypercube with applications to finite buffer queueing networks , queueing systems * 42 * ( 2002 ) , 33 - 62 . | we study four systems and their interactions .
first , we formulate a unified system of coupled forward - backward stochastic _ partial _ differential equations ( fb - spdes ) with lvy jumps , whose drift , diffusion , and jump coefficients may involve partial differential operators . a solution to the fb - spdes
is defined by a 4-tuple general dimensional random vector - field process evolving in time together with position parameters over a domain ( e.g. , a hyperbox or a manifold ) . under an infinite sequence of generalized local linear growth and lipschitz conditions , the well - posedness of an adapted 4-tuple strong solution
is proved over a suitably constructed topological space .
second , we consider a unified system of fb - sdes , a special form of the fb - spdes , however , with _
skew _ boundary reflections . under randomized linear growth and lipschitz conditions together with a general completely-@xmath0 condition on reflections , we prove the well - posedness of an adapted 6-tuple weak solution with _
boundary regulators _ to the fb - sdes by the skorohod problem and an oscillation inequality .
particularly , if the spectral radii in some sense for reflection matrices are strictly less than the unity , an adapted 6-tuple strong solution is concerned .
third , we formulate a stochastic differential game ( sdg ) with general number of players based on the fb - sdes . by a solution to the fb - spdes
, we get a solution to the fb - sdes under a given control rule and then obtain a pareto optimal nash equilibrium policy process to the sdg .
fourth , we study the applications of the fb - spdes / fb - sdes in queueing systems and quantum statistics while we use them to motivate the sdg . + * key words and phrases : * stochastic ( partial / ordinary ) differential equation , lvy jump , skew reflection , completely-@xmath0 condition , skorohod problem , oscillation inequality , stochastic differential game , pareto optimal nash equilibrium , queueing network + * mathematics subject classification 2000 : * 60h15 , 60h10 , 91a15 , 91a23 , 60k25 * unified systems of fb - spdes / fb - sdes with jumps / skew reflections and stochastic differential games * * wanyang dai * department of mathematics and state key laboratory of novel software technology + nanjing university , nanjing 210093 , china + email : [email protected] + date : 14 september 2015 0.1 in |
the strong cp problem reveals itself by the term of lagrangian permitting the cp - violation in an explicit form @xmath1 where _ a priori _ free quantity @xmath0 denotes the effective parameter including the bare contribution as well as the term caused by the matrix of quark masses @xmath2 , @xmath3 the smallness of @xmath0 following from the experimental observations , basically , by the constraints on the electric dipole moment of neutron , looks rather accidental in qcd , unless we do not involve some additional mechanisms .
the most elegant opportunity is the peccei
quinn symmetry @xcite introducing an axion field coupled to the cp - odd product of gluon strength tensors in ( [ bare ] ) . in that case
, the quantity @xmath0 gains the dynamical nature due to the axion @xmath4 , so that @xmath5 , where we have introduced the axion decay constant @xmath6 , and the problem is solved by an appropriate effective potential for @xmath0 , if the minimal energy is reached at @xmath7 .
since the bare and quark mass terms in @xmath0 lead to a global additive redefinition of axion field , we can easily see that the positive mass squared of axion in the effective action is a necessary condition in order to control the stability , while the position of stable point should be at @xmath8 to get the above solution of strong cp - problem .
the real value of axion mass depends on whether the axion is coupled to the quark fields or not .
the corresponding term of fermion contribution is given by @xmath9 where @xmath10 marks the quark flavor @xcite .
there is a possibility to introduce the coupling to the leptons , too , @xcite .
so , due to the fermion component the mixing of axion with the pseudo - goldstone bosons appearing under the spontaneous breaking of chiral symmetry results in the following general formula @xcite for the axion mass : @xmath11 where @xmath12 is the color anomaly of peccei
quinn current , which we put equal to 1 , @xmath13 denotes the quark condensate with @xmath14 and @xmath15 mev . ] , while @xmath16 is a dimensional quantity ordinary represented by the instanton - induced term in the effective potential .
therefore , the light quarks coupled to the axion essentially change the situation , since they suppress purely gluon contribution .
nevertheless , the consideration of gluon term in the quadratic @xmath0-action remains actual as the leading correction to the quark condensate , if the axion is coupled to the quarks , or as the main term , if the axion is coupled to the gluons , only . in the present study
we calculate the effective action quadratic in @xmath0 due to the contribution of the local gluon condensate
. this contribution can be competitive with the term induced by the instantons , since , first , as observed by the lattice simulations , the instantons do not dominate in the determination of vacuum properties in qcd because the vacuum is too warm , second , modelling the gluon condensate by the dilute instanton gas is problematic as the gas should consist of rather wide and dense instantons in order to reach the observed magnitude of the gluon condensate .
in addition , the distribution over the instanton sizes involves extra model - parameters .
thus , the gluon condensate contribution itself is of interest .
in the calculation of diagram shown in fig . [ fglu ] we take the leading term caused by the local gluon condensate , @xmath17 so that we can write down its contribution to the effective action in the explicit form @xmath18 where the combinatoric factor @xmath19 is determined by the permutations of gluon strength tensors and the evident relation @xmath20 .
( 100,45 ) ( 5,0)=85 ( 10,15)@xmath0 ( 80,15)@xmath0 ( 39,39)vacuum for the vacuum expectation value we evidently have the following general expression conserving the color and lorentz symmetries : @xmath21 then , we find the gluon condensate part of quadratic term in the effective potential for the global phase @xmath0 @xmath22 = \theta^2\ ; \left\langle \frac{\alpha_s^2}{8\pi^2 } g^{\mu\nu}_a g_{\mu\nu}^a \right\rangle,\label{glu}\ ] ] while the linear term due to the bare lagrangian in ( [ bare ] ) does not contribute into the vacuum energy , since the physical vacuum infers condition ( [ real ] ) . we can straightforwardly get the contribution given by the local quark condensate into the quadratic term of axion in the way analogous to the described calculation with the gluon condensate , so that @xmath23 = - a^2\;\sum_n \frac{|g_n|^2}{m_n}\,\langle \bar q_n q_n\rangle,\label{quark}\ ] ] which can be further transformed after the substitution for the universal coupling of the axion to the quarks @xcite @xmath24 expressing a consistency for the definition of axion decay constant through the calculation of axial anomaly and the absence of mass mixing between the axion and neutral pseudo - goldstone mesons @xmath25 and @xmath26 . here , @xmath2 denotes the diagonal mass matrix of light quarks ( @xmath27 ) .
then , we get @xmath23 = - \frac{1}{2 } \frac{a^2}{f_a^2}\ ; \langle \bar q q\rangle\,\frac{m_um_d}{m_u+m_d},\label{quark2}\ ] ] which reproduces the result of @xcite in the case of quark condensate dominance .
we stress that the mixing of the axion with the flavor - singlet meson leads to the quadratic axion potential tending to zero in the limit of massless light quark in agreement with the general formula given by ( [ chir ] ) despite the calculated term due to the local gluon condensate . indeed , comparing ( [ quark2 ] ) with ( [ glu ] ) after the substitution of @xmath28 we could observe that the contribution by the local gluon condensate dominates in the axion mass .
the picture is critically changed , if we take into account the neutral singlet of su(3 ) flavor group .
this singlet is coupled to the gluon anomaly , too , in the same manner as the axion with an analogous constant @xmath29 instead of @xmath6 , which produces the mixing of masses for the axion and singlet @xcite .
the resulting effect has been described in the introduction , and we return to the quark condensate dominance in the axion mass up to small corrections .
the description of above mechanism is given in @xcite in detail .
nevertheless , we stress that expression ( [ glu ] ) gives explicit result in contrast to rather indefinite instanton - induced potential .
let us discuss the renormalization group ( rg ) properties of results obtained .
first , we can easily see that the contribution by the quark condensate ( [ quark2 ] ) enters into the axion mass in the rg invariant form , since the ratio of quark masses and the product of quark mass with the condensate are invariant under the action of rg .
second , the gluon condensate term due to ( [ glu ] ) reveals the rg dependence with nonzero anomalous dimension in qcd .
however , this anomalous dimension is in the next @xmath30-order under study . therefore , we have no any contradiction with the qcd principles in the given accuracy over @xmath30
. we do not make some numerical estimates , since they would be rather qualitative under assumptions on the effective value of @xmath30 .
thus , we clarify the physical meaning of our calculations and results .
we have calculated the contribution by the local gluon condensate into the effective potential quadratic over the @xmath0 determining the cp - violating term of the gluon anomaly in qcd .
such the effective potential gives the contribution into the axion mass providing the solution of strong cp - problem in the peccei
quinn mechanism .
we expect the dominance of contribution caused by the local gluon condensate in comparison with the instanton - induced potential , which could be tested by the lattice simulations .
the axion plays an essential role in the connection of particle physics with the cosmology , in particular , as a dark matter candidate @xcite .
this fact makes the studies on its mass to be rather significant .
we see that the qcd dynamics in such the investigations is of importance .
we would like to point to some recent interesting papers devoted to the problem under discussion : solutions of strong cp - problem @xcite , on the yang
mills vacuum @xcite . *
* r. d. peccei and h. r. quinn , phys .
d * 16 * , 1791 ( 1977 ) .
j. e. kim , phys .
lett . * 43 * , 103 ( 1979 ) ; + m. a. shifman , a. i. vainshtein and v. i. zakharov , nucl .
b * 166 * , 493 ( 1980 ) .
m. dine , w. fischler and m. srednicki , phys .
b * 104 * , 199 ( 1981 ) ; + a. r. zhitnitsky , sov .
. phys . * 31 * , 260 ( 1980 ) [ yad . fiz .
* 31 * , 497 ( 1980 ) ] .
h. georgi , d. b. kaplan and l. randall , phys .
b * 169 * , 73 ( 1986 ) ; + k. choi , k. kang and j. e. kim , phys .
b * 181 * , 145 ( 1986 ) .
s. khalil and c. munoz , contemp .
phys . * 43 * , 51 ( 2002 ) [ arxiv : hep - ph/0110122 ] .
s. l. glashow , arxiv : hep - ph/0110178 .
z. berezhiani , l. gianfagna and m. giannotti , phys .
b * 500 * , 286 ( 2001 ) [ arxiv : hep - ph/0009290 ]
. v. chernyak , arxiv : hep - th/9808092 . | we calculate the contribution of local gluon condensate into the quadratic @xmath0-term of vacuum energy and discuss its application to the study of axion mass . |
quantum finite automata ( qfa ) , as theoretical models for quantum computers with finite memory , have been explored by many researchers .
so far , a variety of models of qfa have been introduced and explored to various degrees ( one can refer to a review article @xcite and references therein ) . among these qfa , there is a class of qfa that differ from others by consisting of two interactive components : a quantum component and a classical one .
we call them _ semi - quantum automata _ in this paper
. examples of semi - quantum automata are _ one - way qfa with control language _ ( cl-1qfa )
@xcite , _ one - way qfa together with classical states _
( 1qfac ) @xcite , and _ one - way finite automata with quantum and classical states _
( 1qcfa ) @xcite . here
`` one - way '' means that the automaton s tape head is required to move right on scanning each tape cell .
these semi - quantum automata have been proved to not only recognize all regular languages , but also show superiority over dfa with respect to descriptional power .
for example , 1qcfa , cl-1qfa and 1qfac were all shown to be much smaller than dfa in accepting some languages ( resolving some promise problems ) @xcite .
in addition , a lower bound on the size of 1qfac was given in @xcite , which stated that 1qfac can be at most exponentially more concise than dfa , and the bound was shown to be tight by giving some languages witnessing this exponential gap .
size lower bounds were also reported for cl-1qfa in @xcite and for 1qcfa in @xcite ( no detailed proof was given in @xcite for the bound of 1qcfa ) , but they were not proved to be tight . by the way , we mention that the result obtained in @xcite that 1qfca recognize only regular languages follows directly from @xcite , although a relatively complex procedure was used in @xcite to deduce this result . specially , one can see that complex technical treatments were used in @xcite to derive the bound for cl-1qfa and one may find that some key steps in @xcite were confused such that the proof there may have some flaws , which will be explained more clearly in section 4 .
it is also worth mentioning that the method used in @xcite is tailored for cl-1qfa and is not easy to adopt to other models .
therefore , it is natural to ask : is there a uniform and simple method giving lower bounds on the size of the above three semi - quantum automata ? this is possible , as 1qcfa , cl-1qfa and 1qfac have the similar structure as shown in @xcite , where they were described in a uniform way : a semi - quantum automaton can be seen as a two - component communication systems comprising a quantum component and a classical one , and they differ from each other mainly in the specific communication pattern : classical - quantum , or quantum - classical , or two - way .
it was also proved in @xcite that the three models can be simulated by the model of qfa with mixed states and trace - preserving quantum operations(referred as mo-1gqfa ) @xcite . in this paper , by using the above result
, we present a uniform method that gives a lower bound on the size of 1qcfa , cl-1qfa and 1qfac , and this lower bound shows that they can be at most exponentially more concise than dfa .
specifically , we first obtain a lower bound on the size of mo-1gqfa and then apply it to the three hybrid models by using the relationship between them and mo-1gqfa . compared with a recent work @xcite
, our method is much more concise and universal , and it can be applied to the three existing main models of semi - quantum automata .
in addition , our method may fix a potential mistake in @xcite that will be indicated later on .
throughout this paper , for matrix ( operator ) @xmath0 , @xmath1 and @xmath2 denote the conjugate and conjugate - transpose of @xmath0 , respectively , and @xmath3 and @xmath4 denote the trace and rank of @xmath0 , respectively . according to von neumann s formalism of quantum mechanics , a quantum system is associated with a hilbert space which is called the state space of the system . in this paper , we only consider finite dimensional spaces .
a ( mixed ) state of a quantum system is represented by a density operator on its state space . here a density operator @xmath5 on @xmath6 is a positive semi - definite linear operator such that @xmath7 .
when @xmath8 , that is , @xmath9 for some @xmath10 , then @xmath5 is called a pure state .
let @xmath11 and @xmath12 be the sets of linear operators and density operators on @xmath6 , respectively .
a trace - preserving quantum operation @xmath13 on state space @xmath6 is a linear map from @xmath14 to itself that has an _ operator - sum representation _ as @xmath15 with the completeness condition @xmath16 , where @xmath17 are called operation elements of @xmath13 . a general measurement is described by a collection @xmath18 of measurement operators , where the index @xmath19 refers to the potential measurement outcome , satisfying the condition @xmath20 if this measurement is performed on a state @xmath5 , then the classical outcome @xmath19 is obtained with the probability @xmath21 , and the post - measurement state is @xmath22 for the case that @xmath5 is a pure state @xmath23 , that is , @xmath24 , we have @xmath25 and the state @xmath23 `` collapses '' into the state @xmath26 a special case of general measurements is the projective measurement @xmath27 where @xmath28 s are orthogonal projectors .
@xmath29 has the singular value decomposition @xcite as follows : @xmath30 where @xmath31 , @xmath32 are called singular values of @xmath0 , and @xmath33 are two orthonormal sets . the trace norm of @xmath29 is defined as @xmath34 . by the singular value decomposition in ( [ svd ] ) , the trace norm can be characterized by singular values as @xmath35 note that if @xmath0 is positive semi - definite , then @xmath36 . for @xmath37 , the trace distance between them is @xmath38 the trace distance between two probability distributions @xmath39 and @xmath40 is @xmath41 recall results about the trace distance from @xcite as follows .
let @xmath5 and @xmath42 be two density operators .
then we have 1 .
@xmath43 for any trace - preserving quantum operation @xmath13 .
2 . @xmath44 where @xmath45 , @xmath46 and the maximization is over all povms @xmath47 .
[ lm - distance ] a linear mapping @xmath48 which maps a @xmath49 matrix to a @xmath50-dimensional column vector is defined as follows : @xmath51 in other words , @xmath52 is the vector obtained by taking the rows of @xmath0 , transposing them to form column vectors , and stacking those column vectors on top of one another to form a single vector .
for example , we have @xmath53 if we let @xmath54 be an @xmath55-dimensional column vector with the @xmath56th entry being 1 and else 0 s , then @xmath57 form a basis of @xmath58 .
therefore , the mapping @xmath59 can also be defined as follows : @xmath60 for any @xmath61 , it is easy to verify @xmath62 in this paper , the norm of @xmath63 is defined by @xmath64 . for @xmath29 , we observe the following relation between the two norms @xmath65 and @xmath66 .
let @xmath29 and @xmath67 .
then we have @xmath68[lm - norm ] _ proof .
_ suppose @xmath0 has the singular value decomposition @xmath69 .
then we have @xmath70 thus we have @xmath71 on the other hand , by the cauchy - schwarz inequality we have @xmath72 definitions of automata ----------------------- in the literature , there exist some hybrid models of qfa that differ from other qfa models by consisting of two interactive components : a quantum component and a classical one .
we call them _ semi - quantum automata _ in this paper .
as shown in @xcite , a semi - quantum automaton can be depicted in fig .
[ model ] , where an automaton comprises a quantum component , a classical component , a classical communication channel , and a classical tape head ( that is , the tape head is regulated by the classical component ) . on scanning an input symbol , the quantum and classical components interact to evolve into new states , during which communication may occur between them . in this paper
, we focus on automata with a one - way tape head , that is , after scanning an input symbol the model moves its tape head one cell right . as shown in @xcite ,
there are three models of semi - quantum automata fitting into fig .
[ model ] , with the essential difference being the specific communication pattern : * in cl-1qfa @xcite , only quantum - classical communication is allowed , that is , the quantum component sends its measurement result to the classical component , but no reverse communication is permitted . * in 1qfac @xcite , only classical - quantum communication is allowed , that is , the classical component sends its current state to the quantum component . * in 1qcfa @xcite , two - way communication is allowed : ( 1 ) first , the classical component sends its current state to the quantum component ; ( 2 ) second , the quantum component sends its measurement result to the classical component . in the following , we recall the detailed definitions of the existing models of semi - quantum automata . one of such models is called _ one - way qfa with control language _
( cl-1qfa)@xcite , defined as follows . a cl-1qfa is a 7-tuple @xmath73 where @xmath74 is a finite set of quantum basis states , @xmath75 is a finite alphabet , @xmath76 is a finite set of symbols ( measurement outcomes ) , @xmath77 is the initial quantum state , @xmath78 is a unitary operator for each @xmath79 , @xmath80 is a projective measurement given by a collection @xmath81 of projectors , and @xmath82 is a regular language ( called a control language).[df : cl - qfa ] in cl-1qfa @xmath83 , on scanning a symbol @xmath42 , a unitary operator @xmath78 followed by the projective measurement @xmath80 is performed on its current state . thus , given an input string @xmath84 , the computation produces a sequence @xmath85 of measurement results with a certain probability @xmath86 that is given by @xmath87 where we define the ordered product @xmath88 .
the input @xmath89 is said to be _ accepted _ if @xmath90 belongs to a fixed regular language @xmath91 .
thus the probability of @xmath92 accepting @xmath89 is @xmath93 recently , qiu et al @xcite proposed a new model named _
1qfa together with classical states _ ( 1qfac ) , defined as follows .
a 1qfac @xmath83 is defined by a 8-tuple @xmath94 where @xmath74 and @xmath95 are finite sets of quantum basis states and classical states , respectively , @xmath75 is a finite input alphabet , @xmath77 and @xmath96 are initial quantum and classical states , respectively , @xmath97 is a unitary operator on @xmath98 for each @xmath99 and @xmath42 , @xmath100 is a classical transition function , and for each @xmath99 , @xmath101 is a projective measurement given by projectors @xmath102 where the two outcomes @xmath103 and @xmath104 denote acceptance and rejection , respectively .
the machine starts with the initial states @xmath105 and @xmath106 .
on scanning an input symbol @xmath79 , @xmath97 is first applied to the current quantum state , where @xmath99 is the current classical state ; afterwards , the classical state @xmath99 changes to @xmath107 . finally ,
when the whole input string is finished , a measurement @xmath108 determined by the last classical state is performed on the last quantum state , and the input is accepted if the outcome @xmath103 is observed .
therefore , the probability of 1qfac @xmath83 accepting @xmath109 is given by @xmath110 where @xmath111 for @xmath112 .
ambainis and watrous @xcite proposed the model of _ two - way qfa with quantum and classical states _ ( 2qcfa ) .
as proved in @xcite , 2qcfa can recognize non - regular language @xmath113 in polynomial time and the palindrome language @xmath114 in exponential time , which shows the superiority of 2qcfa over their classical counterparts . in the following we recall 1qcfa @xcite , a one - way variant of 2qcfa .
note that in this paper the notion of 1qcfa is slightly more general than the one in @xcite .
the reason for why we adopt the current definition is that it has a more succinct form which simplifies some notations ( for example , in our version we need give only the set of general measurements , instead of two sets : unitary operators and projective measurements ) .
it is , however , worthwhile to emphasize that all results obtained in this paper hold surely for the model in @xcite .
a 1qcfa is specified by a 9-tuple @xmath115 where @xmath74 and @xmath95 are finite sets of quantum and classical states , respectively , @xmath75 is a finite input alphabet , @xmath76 is a finite set of symbols ( measurement outcomes ) , @xmath77 and @xmath96 are initial quantum and classical states , respectively , @xmath116 for each @xmath99 and @xmath42 is a general measurement on @xmath98 with outcome set @xmath76 , @xmath117 specifies the classical state transition , and @xmath118 denotes a set of accepting states.[df-1qcfa ] on scanning a symbol @xmath42 , at first the general measurement @xmath116 , determined by the current classical state @xmath99 and the scanned symbol @xmath42 , is performed on the current quantum state , producing some outcome @xmath119 ; then the classical state @xmath99 changes to @xmath120 by reading @xmath42 and @xmath121 . after scanning all input symbols , @xmath83 checks whether its classical state is in @xmath122 .
if yes , the input is accepted ; otherwise , rejected .
therefore , the probability of 1qcfa @xmath83 accepting @xmath109 is given by @xmath123 where : * @xmath124 is defined by @xmath125 * @xmath126 are measurement operators of @xmath127 .
* @xmath128 for @xmath112 . in fig.[model ] , let @xmath74 be the set of basis states of the quantum component and @xmath95 be the set of states of the classical component .
let @xmath129 and @xmath130 .
then we say that the semi - quantum automaton has @xmath131 quantum basisi states and @xmath132 classical states . for cl-1qfa
, @xmath95 denotes the set of states of the minimal dfa accepting the control language @xmath133 .
recall the model of mo-1gqfa which has mixed states and trace - preserving quantum operation as follows @xcite .
an mo-1gqfa @xmath83 is a five - tuple @xmath134 , where @xmath6 is a finite - dimensional hilbert space , @xmath75 is a finite input alphabet , @xmath135 , the initial state of @xmath83 , is a density operator on @xmath6 , @xmath136 corresponding to @xmath79 is a trace - preserving quantum operation acting on @xmath6 , @xmath137 is a projector on the subspace called accepting subspace of @xmath138 .
denote @xmath139 , then @xmath140 form a projective measurement on @xmath6 .
let @xmath141 .
then we call @xmath83 is an @xmath55-dimensional mo-1gqfa . on the input word @xmath142 , the above mo-1gqfa @xmath83 proceeds as follows : the quantum operations @xmath143 are performed on @xmath135 in succession , and then the projective measurement @xmath140 is performed on the final state , obtaining the accepting result with a certain probability .
thus , mo-1gqfa @xmath83 defined above induces a function @xmath144 $ ] as @xmath145 where @xmath146 stands for @xmath147 .
in fact , for every @xmath84 , @xmath148 represents the probability that @xmath83 accepts @xmath89 .
a dfa is a five - tuple @xmath149 where @xmath95 is a finite state set , @xmath75 is a finite alphabet , @xmath96 is the initial state , @xmath118 is the set of accepting states , and @xmath150 is the transition function : @xmath151 means that the current state @xmath99 changes to @xmath152 when scanning @xmath42 . furthermore , @xmath153 can be extended to @xmath154 by defining : i ) @xmath155 , and ii ) @xmath156 where @xmath84 and @xmath157 .
@xmath83 is said to accept @xmath84 , if @xmath158 .
some recent work ( e. g. @xcite ) shows that semi - quantum automata like cl-1qfa , 1qfac and 1qfac are much more concise than equivalent dfa . in order to see the limit of the descriptional power of these models , size
lower bounds were given in @xcite for these models .
however , it is worth mentioning that there were some potential flaws in the procedure to obtain lower bounds for cl-1qfa in @xcite .
indeed , one may find that in lemma 5 of @xcite the two states @xmath159 and @xmath160 were required to be product states , but when lemma 5 was used to get eq.(10 ) in page 110 , the states @xmath161 are not product states ( generally , they can be entangled states ) .
thus , the procedure to get eq.(10 ) may have a bug , whereas one can find that eq.(10 ) is a crucial step , without which the lower bound may not be obtained for cl-1qfa . in the following , we present another method which can not only fix the possible bug in @xcite , but also deal with uniformly lower bounds on the size of 1qcfa , cl-1qfa and 1qfac . before that
, we first recall some preliminary knowledge .
a language @xmath133 is said to be recognized by qfa @xmath83 with cut - point @xmath162 $ ] , if @xmath163 holds for all @xmath164 and @xmath165 holds for all @xmath166 .
the cut - point @xmath167 is said to be _ isolated _ whenever there exists @xmath168 $ ] such that @xmath169 , for all @xmath84 . in this case , we simply say that @xmath133 is recognized by @xmath83 with cut - point isolated by @xmath153 . two qfa @xmath170 and @xmath171 over @xmath75 are said to be equivalent , if @xmath172 holds for all @xmath173 .
below we first recall a result given in @xcite that is useful for obtaining our results in this paper . for any semi - quantum automaton ( including cl-1qfa , 1qfac and 1qcfa ) with @xmath131 quantum basis states and @xmath132 classical states
, there is a @xmath174-dimensional mo-1gqfa equivalent to it .
[ thm - simulation ] _ proof .
_ for readability , we present here the proof for cl-1qfa and one can refer to @xcite for other cases .
let cl-1qfa @xmath175 , with @xmath176 accepted by dfa @xmath177 .
we construct an mo-1gqfa @xmath178 from @xmath179 and @xmath180 as follows : * @xmath181 where @xmath182 denotes the hilbert space spanned by @xmath0 ; * @xmath183 ; * @xmath184 where @xmath185 is the identity operator on @xmath98 ; * for each @xmath79 , @xmath186 has operation elements @xmath187 where @xmath188 it is easy to verify that the collection of operators @xmath189 satisfies the completeness condition .
furthermore , for @xmath190 , we have @xmath191 where @xmath192 .
now let us check the behavior of @xmath193 on an input string .
suppose @xmath193 starts with the initial state @xmath135 and scans a symbol @xmath42 .
then the resulting state is @xmath194 where @xmath195 . in this way , after scanning a string @xmath109 , the final state is @xmath196 where @xmath197 and @xmath198 .
note that @xmath199 iff @xmath200 .
thus the probability of @xmath193 accepting @xmath89 is @xmath201 where @xmath202 is @xmath203 if @xmath99 is in @xmath122 and @xmath204 else .
note that the above probability is equal to the one of @xmath83 given in eq .
( [ f_cl ] ) .
therefore , we have completed the proof .
now we are in a position to give our main results .
suppose that @xmath133 is recognized by an @xmath55-dimensional mo-1gqfa with cut - point isolated by @xmath153 .
then @xmath133 can be recognized by a dfa with @xmath205 states satisfying @xmath206 _ proof .
_ assume that @xmath133 is recognized by mo-1gqfa @xmath207 with cut - point isolated by @xmath153 .
an equivalence relation `` @xmath208 '' on @xmath209 is defined by : @xmath210 if for all @xmath211 , @xmath212 iff @xmath213 .
then in terms of myhill - nerode theorem @xcite , it suffices to prove that the number of equivalence classes induced by `` @xmath208 '' is upper bounded by the right side of ( [ ubd ] ) .
let @xmath214 , i.e. , the state of @xmath92 after scanning the word @xmath89 . now
, suppose that @xmath215 , that is , there exists a string @xmath211 such that @xmath212 and @xmath216 .
then we have @xmath217 for some @xmath162 $ ] , where @xmath218 stands for @xmath219 .
denote @xmath220 then we have @xmath221 on the other hand , for any @xmath173 , we have @xmath222 , the first equality holds because @xmath223 is positive semi - definite , and the second equality holds because the operations used are trace - preserving . in summary
, we obtain the following two properties : * for any @xmath173 , @xmath224 lies in the unit sphere in @xmath225 . * for any two strings @xmath209 satisfying @xmath215 , we always have @xmath226 consists of @xmath205 equivalence classes , say @xmath227,[x_2],\cdots , [ x_d]$ ] . arbitrarily choose an element @xmath228 .
let@xmath229 i.e. , a sphere centered at @xmath228 with the radius @xmath230 . then all these spheres do not intersect pairwise except for their surface , and all of them are contained in a large sphere in @xmath225 centered at @xmath231 with the radius @xmath232 .
note that @xmath225 is an @xmath50-dimensional complex space and each element from it can be represented by an element of @xmath233 .
then the volume of a sphere of a radius @xmath104 in @xmath225 is @xmath234 where @xmath121 depends only on @xmath55 .
therefore , it holds that @xmath235 this completes the proof .
+ furthermore , we have let @xmath133 be a regular language whose minimal dfa has @xmath205 states .
then any @xmath55-dimensional mo-1gqfa recognizing @xmath133 with cut - point isolated by @xmath153 must satisfy @xmath236^{\frac{4}{9}}.\end{aligned}\ ] ] _ proof .
_ for @xmath237 and @xmath238 $ ] , we have @xmath239{n}n^2}= \left(\frac{2}{\delta}\right)^ { 2n^{\frac{9}{4}}},\end{aligned}\ ] ] where the second inequality holds because @xmath240 holds for any @xmath241 and @xmath238 $ ] . + therefore , based on theorem [ thm - simulation ] and the above theorem
, we obtain a lower bound on the size of semi - quantum automata as follows .
suppose a semi - quantum automaton ( including cl-1qfa , 1qfac and 1qcfa ) , with @xmath131 quantum basis states and @xmath132 classical states , recognizes a regular language whose minimal dfa has @xmath205 states , with cut - point isolated by @xmath153 .
then it holds that @xmath242^{\frac{4}{9}}.\end{aligned}\ ] ] not that recently ref .
@xcite dealt with the size lower bound of cl-1qfa .
compared with that , our work has the following advantages : ( i ) our method is much more concise , whereas complex technical treatments were used in @xcite ; and ( ii ) our method is universal , since it is applicable to the three existing models of semi - quantum automata , while the method in @xcite was tailored for cl-1qfa .
in addition , our bound is slightly more optimal than the one in @xcite , since the lower bound obtained in @xcite actually should be @xmath243^{\frac{4}{9}}$ ] , although it was claimed to be @xmath244^{\frac{4}{9}}$ ] ( the factor @xmath245 is from the fact the volume of a sphere of a radius @xmath104 in @xmath225 is @xmath234 , instead of @xmath246 , where @xmath121 depends only on @xmath55 ) .
we have presented a uniform method for obtaining the lower bound on the size of cl-1qfa , 1qfac and 1qcfa , and this bound shows that these automata can be at most exponentially smaller than dfa .
compared with a recent work @xcite , our method is much more concise and universal , and it is applicable to the three existing main models of semi - quantum automata . note that although our lower bound is universal , it is not necessarily optimal .
for instance , a better lower bound @xmath247 was obtained for 1qfac in @xcite .
thus , a natural open problem remains either to witness the optimality of our size lower bound for some specific model , or to improve it .
the authors are thankful to dr .
shenggen zheng for his useful comments . , _
quantum computing : 1-way quantum automata _ , in proceedings of the 9th international conference on developments in language theory , lecture notes in comput .
2710 , springer - verlag , berlin , 2003 , pp . 1 - 20 . | in the literature , there exist several interesting hybrid models of finite automata which have both quantum and classical states .
we call them semi - quantum automata . in this paper
, we compare the descriptional power of these models with that of dfa .
specifically , we present a uniform method that gives a lower bound on the size of the three existing main models of semi - quantum automata , and this bound shows that semi - quantum automata can be at most exponentially more concise than dfa . compared with a recent work ( bianchi , mereghetti , palano , theoret .
comput .
sci .
, 551(2014 ) , 102 - 115 ) , our method shows the following two advantages : ( i ) our method is much more concise ; and ( ii ) our method is universal , since it is applicable to the three existing main models of semi - quantum automata , instead of only a specific model . |
in this paper we present a reliable but compact approximation to the high temperature cooling coefficient of heavy elements in diffuse plasmas under non - equilibrium conditions .
such an approximation is essential for use with multi - dimensional hydrodynamics codes where the additional burden of following the details of ionization evolution severely restricts the available spatial resolution . with only a few exceptions , large hydrodynamic models that incorporate radiative cooling
characterize the cooling coefficient with a single parameter , the temperature , @xmath1 , where the total emissivity per unit volume is @xmath2 and @xmath3 and @xmath4 are the electron and hydrogen densities .
these cooling functions are determined by assuming either that the ionization state at a given temperature is characterized by collisional equilibrium , or that all gas follows a particular pre - calculated ionization history ( shapiro & moore 1976 ; edgar & chevalier 1983 ; sutherland & dopita 1993 ) . because the cooling of a plasma depends on the ionization history of the constituent ions , there
can actually be a large range in the value of @xmath5 at a given @xmath6 , depending upon the details of the ionization evolution .
we demonstrate , however , that cooling due to trace elements ( those heavier than helium ) can be approximated by following the evolution of just a single additional parameter , the mean charge on the trace ions , @xmath7 .
here , z is the element number , z is the ionic charge , @xmath8 is the ( linear ) abundance of element z relative to hydrogen , and @xmath9 is the concentration of a given ion ( the fraction of element @xmath10 with charge @xmath11 ) .
because the abundance is dominated by oxygen , @xmath12 ranges from zero to about nine .
our method generalizes from a cooling curve , @xmath1 , to a cooling plane @xmath13 .
it consequently requires another function that allows us to update the ionization level , @xmath14 . with the mean charge as our ionization level indicator
, @xmath15 is the difference between the mean ionization and recombination functions .
thus @xmath16 , where @xmath17 and @xmath18 , and where @xmath19 and @xmath20 are the ionization and recombination rate coefficients , respectively , at @xmath21 for stage @xmath11 of element @xmath10 .
the corresponding cooling coefficient is @xmath22 , where @xmath23 is the cooling coefficient per ion @xmath24 . to implement this approximation we search for a reasonable description of the nonequilibrium ionization concentrations , @xmath25 , representative of those found at @xmath21 and @xmath12
this assumes that under actual nonequilibrium conditions , the distribution over ion states will depend less on the details of the past history than on how far the present mean charge differs from the collisional equilibrium value at the current temperature . in this paper
, we consider the ionization distributions that arise in isothermal relaxation to equilibrium for each temperature @xmath6 , starting with nearly fully ionized or fully neutral gas .
the atomic data used to develop the manifolds comes from raymond & smith ( 1977 ) , with updates described in raymond & cox ( 1985 ) , corrections in oscillator strengths in the cooling transitions of li - like ions ( shull & slavin 1994 ) , and revised dielectronic recombination rates of romanik ( 1988 ) .
the abundances are taken from anders & grevesse ( 1989 ) .
the effects of charge exchange are not included ; errors introduced by neglecting this are smaller than the uncertainties in the atomic data and elemental abundances used to generate the cooling curve . [ secintro ]
we have performed the manifold of non - equilibrium isothermal evolutions described below to form the basis for our approximation , producing tables of the cooling coefficient of trace elements @xmath13 and our ionization evolution functions @xmath26 and @xmath27 .
gas is initialized with equilibrium ionization appropriate to some initial temperature @xmath28 .
its temperature is then suddenly changed to @xmath6 and held fixed as the ionization evolves to equilibrium . by choosing @xmath28 to be low ( e.g. , @xmath29 k ) , and performing the calculation for a dense set of temperatures @xmath6 between @xmath29 and @xmath30 k , the runs sample the full range of conditions possible in under - ionized gases . by repeating the whole set once again with very high @xmath28 ( e.g. @xmath30 k )
, conditions representing over - ionized gases are explored .
together these cases sample the full range of @xmath12 . a particular moment in an evolution
can be characterized by @xmath28 , @xmath6 , and the fluence ( @xmath31 ) .
the ionization state is known , allowing straightforward evaluation of the mean charge @xmath12 and then using as the index of the time evolution .
evolution at constant temperature makes it easy to acquire the tables of @xmath13 , @xmath26 , @xmath27 , and @xmath32 on a fixed grid of @xmath6 and .
[ top left panel ] , the mean ionization rate coefficient , @xmath33 [ top right panel ] , the mean recombination rate coefficient , @xmath34 [ bottom right panel ] , and the absolute value of the ionization evolution function , @xmath35 [ bottom left panel ] , as a function of electron temperature and mean charge @xmath36 . the dotted line in each panel shows the mean charge in collisional equilibrium as a function of gas temperature , @xmath37 .
the function @xmath35 passes through zero along this line . above the equlibrium curve ,
the gas is recombining and d ( , ) is negative ; below , it is ionizing and d ( , ) is positive .
these functions were calculated for a grid of isothermal evolutions .
gas was either started with ionization fractions corresponding to @xmath38 k or @xmath39 k. it then ionized up or recombined down to @xmath40 at fixed temperature @xmath6 .
the resultant rate coefficients were calculated versus @xmath12 for each @xmath6 for the range @xmath41k . ]
the cooling function shown in figure 1 shows a huge peak at @xmath42 k and @xmath43 due to collisional excitation of low stages of ionization .
it falls rapidly to the left due to boltzmann factors , and gradually to the right due to decreasing excitation cross - sections .
it falls at higher as fewer bound electrons with low enough energy are available for collisional excitation . above the equilibrium line
the gas is over - ionized and collisional excitation is difficult , particularly at lower temperatures . at a given @xmath12 ,
the mean ionization function , i , behaves much like the collisionally excited cooling , with vagaries at the transition from @xmath44 to 2 , and at @xmath45 where the mean ionization rate is sensitive to the relative proportions of helium- and lithium - like oxygen .
the structure of the mean recombination rate , r , is more gradual . in the upper left corner
it is dominated by radiative recombination and the gradual increase with @xmath12 and decrease with temperature are as expected .
deviations from that smooth pattern in the rest of the diagram are due to dielectronic recombination . in both i and r , there are distortions in the patterns just above the equilibrium line .
these arise because , in relaxing to equilibrium , a recombining plasma goes through a considerable compaction of its ionization distribution .
the principal feature of the rate function is that ionization toward equilibrium from below is much faster than recombination toward it from above . with increasing distance from equilibrium , the ionization rate increases dramatically , the recombination rate only gradually .
to test the accuracy of our approximation for a wide range of situations , a representative set of cases were examined in which single parcels of gas in collisional equilibrium at @xmath46 k were suddenly shock heated and then subjected to varying degrees of expansion .
the test situation was modeled on the sedov blast wave solution for an explosion of energy @xmath47 ergs into a homogeneous medium of particle density @xmath48 . by varying the preshock density and the assumed distance of the parcel from the explosion site
, we adjusted the post shock temperature and the timescales for depressurization versus radiative cooling .
these scenarios test the two behaviors in which drastic departures from equilibrium occur : the rapid ionization of gas passing through a shock front and the rapid adiabatic cooling of an over - ionized plasma .
each parcel s evolution can be fully characterized by @xmath48 and @xmath49 .
we have approximated @xmath50 , with @xmath51 being the time the parcel is shocked , and @xmath52 determining the rate of decompression .
the temperature evolution is then @xmath53 where @xmath54 is the particle density of hydrogen ( ionized and neutral ) .
we have carried out each simulation twice , first , solving the full ionization balance evolution exactly , then using our cooling approximation . in the latter ,
we solve the ionization evolution for hydrogen , helium , and ( four rate equations ) .
the cooling is the sum of that from hydrogen , helium , and our tabulated cooling function . in order to test the widest possible range of conditions , we chose three values of @xmath49 , with three different values of @xmath48 for each .
the values of density were spaced to alter the initial ratio of depressurization to radiative cooling from approximately 0.4 at the high density end , to 6 and then 4000 at the lower densities . for the densest runs ,
the structure is very similar to that of a steady state radiative shock . for the next lower density cases ,
depressurization has strong effects but cooling is still sufficiently rapid that the density eventually rises rather than falling with time . at the lowest densities , the density and temperature fall adiabatically and the ionization level is soon frozen in .
( dashed ) , @xmath55 ( dash - dot ) , and @xmath56 ( solid ) ; the central panel shows `` mixed '' evolution with @xmath57 ( dashed ) , @xmath58 ( dash - dot ) , and @xmath59 ( solid ) ; the right panel shows radiative cooling dominated evolution with @xmath60 ( dashed ) , @xmath61 ( dash - dot ) , and @xmath62 ( solid ) .
the time evolution for each curve is from right to left .
the cooling curve for collisional ionization equilibrium is also shown ( dotted ) .
the middle set of panels shows the ratio of the approximation for trace element cooling to the exact value for each of the above cases . for `` mixed '' and `` radiative '' evolution ,
the highest temperature cases are sufficiently similar that the curves overlie each other .
the lowest set of panels show the ionization evolution for these cases in the t- plane .
the solid curves show the exact evolution for the nine cases above , while the dashed lines show the slight departures that result by using the approximation we present .
the dotted line in each panel shows the mean charge in collisional equilibrium as a function of gas temperature , @xmath37 , as in figure 1 . ]
figure 2 shows the exact trace element cooling coefficient evolutions versus temperature , the ratio of our approximate results to the exact ones , and the evolutions of mean charge , both exact and approximate . in each case
, the cooling function begins very high just after the shock , when the gas is still briefly nearly neutral . at the two higher temperatures in the middle and right hand panels , the gas rapidly ionizes up and
the cooling function drops precipitously to very close to the equilibrium value .
this ion flash " is so rapid that the integrated cooling during this period is negligible , as evidenced by the nearly vertical tracks .
thereafter , these tracks follow the equilibrium curve down to just below @xmath63k , after which their recombination can not keep up with the cooling and the cooling rate falls below the equilibrium value . in the lowest temperature cases ,
significant cooling takes place before the cooling function drops through equilibrium , after which it plunges toward the nonequilibrium curve of the initially higher temperature cases .
the differences between the radiation dominated and mixed evolutions are almost negligible . in the expansion dominated cases , however , things are quite different .
the temperature drops so rapidly that the gas is soon over - ionized ; recombination is slower than the temperature drop , and the ionization level declines only gradually or is frozen in .
the middle row of panels shows the ratio of approximate to exact cooling coefficients . in the two right hand panels , the approximation is good to within about 10% except below @xmath64k where it is consistently high by about 40% , and in a peak at @xmath65k where it can be as much as 20 to 30% high . at this temperature ,
the approximate @xmath12 evolutions of the two higher temperature cases in the two right lower panels of figure 2 diverge by @xmath66 charge states from the exact results .
this is due in part to the excess cooling , but the recombination rate is almost certainly slightly too low as well . for the rapidly depressurizing cases , radiative cooling is inconsequential .
nevertheless , the cooling coefficient which differs substantially from the other cases , is still reasonably well fit , except for the intermediate temperature case for which @xmath12 gets frozen at a value of 5.5 . in that case , the value of @xmath12 is well represented by the approximation , but the approximate cooling coefficient is too high .
the freezing in of the ionization is extremely well matched , except very late in the highest initial temperature case .
we have also compared the time evolution of density , temperature , mean charge , and the cooling coefficients for the exact calculation , our approximation , and the case where the trace elements are assumed to be in collisional ionization equilibrium .
while the time evolution of temperature and density of the exact calculations and our approximation are in reasonably good agreement , the case with collisional ionization equilibrium shows significant differences in the time evolution of temperature . in addition
, the ionization equilibrium assumption provides a poor estimate of the true ionization level of the gas , especially for the expansion dominated cases .
we provide a means by which large hydrocodes can include an accurate approximation to the radiative cooling coefficient , one far more responsive to the vagaries of dynamical environments than any single function of temperature .
the method also follows the mean ionization level of the gas .
both the trace element cooling coefficient @xmath13 , and the rate of change of @xmath12 depend only on the identification of a manifold of representative ionization concentrations found at a particular temperature and mean charge .
the ionization concentrations we used came from the isothermal relaxations of both highly over - ionized and highly under - ionized gas . in the test cases presented ,
both the cooling coefficient and mean charge evolution are quite accurately approximated .
for all cases in which radiative cooling is significant , the approximation never errs by more than 30% .
the error is usually less than 10 to 15% , well within the accuracy of the true cooling coefficient , which is limited by uncertainties in atomic data and abundances .
the error in the cooling coefficient can be somewhat larger in examples with extreme amounts of depressurization , but never by enough to make the negligible radiative cooling appear to be significant .
our approximate cooling coefficients and charge evolution rates , however , showed patterns in their modest errors , patterns which we believe we understand and can eliminate with future work .
should a potential user wait for these improvements before beginning to implement this method ?
the current model provides an excellent approximation for the cooling coefficient . if one is interested only in dynamics , it is certainly sufficient .
implementation of the next generation requires only swapping one set of tables for another .
our future work will also examine the possibility that the ion distributions can be used to provide absorption and emission spectra in post processing .
it is possible to do this at the two parameter level , but higher order corrections will also be examined .
the latter are expected to be two to four times more complex , and will be pursued only if spectral accuracy requires it .
an additional consideration for gas with @xmath67 and @xmath68 is the effect of photoionization on the ionization balance and therefore the cooling .
our scheme can be easily modified to incorporate photoionization of hydrogen and helium .
however , the effects of photoionization of trace elements would require a table of photoionization correction factors for @xmath69 and @xmath15 .
this too will be considered in future refinements .
we would like to thank john raymond for compiling and providing much of the atomic data that went into this work , and for a very careful reading of the manuscript .
we would also like to thank nasa astrophysical theory grant nag5 - 8417 for financial support of this work . and
finally , we would like to acknowledge the valuable contributions of leo krzewina , angela klohs , andrew pawl , and tim freyer who all expended some effort on cracking this problem . | radiative cooling is an important ingredient in hydrodynamical models involving evolution of high temperature plasmas .
unfortunately , calculating an accurate cooling coefficient generally requires the solution of over a hundred differential equations to follow the ionization .
we discuss here a simple 2-parameter approximation for the cooling coefficient due to elements heavier than h and he , for the temperature range @xmath0k .
tests of the method show that it successfully tracks the ionization level in severe dynamical environments , and accurately approximates the non - equilibrium cooling coefficient of the trace elements , usually to within 10% in all cases for which cooling is actually important .
the error is large only when the temperature is dropping so rapidly due to expansion that radiative cooling is negligible , but even in this situation , the ionization level is followed sufficiently accurately .
the current approximation is fully implemented in publicly available fortran code .
a second paper will discuss general approaches to approximation methods of this type , other realizations which could be even more accurate , and the potential for extension to calculations of non - equilibrium spectra . |
it is well known that the hot intra - cluster medium ( icm ) in galaxy clusters radiates enormous amounts of energy at x - ray wavelengths ( @xmath7 ) .
taken at face value , this loss of energy from the icm implies that the gas should be cooling , losing pressure support and flowing inward .
early cooling flow models predicted cooling and star formation rates on the order of @xmath8 @xmath9 yr@xmath10 ( fabian 1994 ) . the lack of widespread detection of iron lines expected from cluster gas cooling below 1 - 2 kev in xmm - newton observations of cool core clusters ( peterson et al .
2003 ) contradicted this model and created the need to incorporate non - gravitational heating mechanisms to counterbalance the radiative energy loss .
while these observations have effectively ruled out accretion flows large enough to form an @xmath11 galaxy in 100 myr , they do not preclude that star formation is ongoing in cool core clusters at a much reduced rate .
the level of star formation still allowed is uncertain and it is important to investigate this through means independent of the x - ray - determined mass flow rate . of particular interest is the distinction between models in which heating completely balances cooling and those in which heating offsets most but not all of the radiative losses .
if significant star formation is occurring as a result of a heating imbalance it should be reflected in the optical properties of the brightest cluster galaxy ( bcg ) , which is often found deep within the cluster potential . a good way to test
this is to look for links between the bcg optical properties with the x - ray properties of their host clusters .
in fact , several case studies over the last two decades have found evidence for recent star formation in the central galaxies of cool core clusters ( johnstone et al .
1987 , mcnamara & oconnell 1989 , 1992 , allen et al .
1992 , donahue & voit 1997 , cardiel et al .
1998 , crawford et al .
1999 , edge 2001 , edge et al .
2002 , goto 2005 , mcnamara et al 2006 , wilman et al .
2006 , edwards et al .
furthermore , the strength of the recent star formation indicators is at least weakly correlated with their cooling - flow derived mass drop out rates ( johnstone et al .
1987 , heckman et al .
1989 , mcnamara 1997 ) . [ cols="<,<,<,<,<,<,<,<,<,^ " , ] for each bcg surface brightness profile a psf - convolved , circularly - symmetric
, single - component @xmath12 devaucouleurs model is fit with the following functional form : @xmath13 with @xmath14 and @xmath15 as the effective intensity and effective radius respectively .
the term @xmath16 is a constant background intensity term which is also included as a fit parameter . to determine the psf we measure the surface brightness profile of an unsaturated star close to the bcg .
the total integrated distribution is then normalized to unity in order to conserve total flux when applying the psf as a convolution response function in bcg modeling .
we fit a moffat profile ( moffat 1969 ) to the psf star with the following form : @xmath17^{-\beta}+i_{bgd } \label{eqn : moffat}\ ] ] where @xmath16 is a term added to take into account the local background level and @xmath18 is left as a free parameter .
estimates for the error in surface brightness for each radial bin are obtained by bootstrapping the pixel values in that bin .
bcg source pixels are randomly selected in each radial bin , allowing a given pixel to be selected multiple times .
this has the advantage that we do not need to make any apriori assumption of the functional form of the noise ( ie .
poisson noise ) , or to know the precise values of exposure times , gain or read - noise .
bootstrapping not only accounts for the random noise in the data , it also incorporates some of the contribution from non - uniform , systematic variations ( bright stars , variability in galactic dust absorption , crowding effects of nearby galaxy / group light , etc . ) across the cluster image .
we note , however , that bootstrapping estimates misrepresent the true error in radial bins that have only a few pixels available to randomly sample .
the bcg fitting procedure is restricted to data points at radii larger than twice the fwhm of the derived moffat profile psf ( @xmath19 ) . ignoring the central bins
helps to ensure that any small error in the psf determination will not seriously affect the bcg fit .
an upper radial cutoff to the fitting region is also applied .
the fit is initially restricted to those data points that lie at radii less than @xmath20 from the bcg centre .
this radius is chosen somewhat arbitrarily but it reflects an effort to isolate the bcg component within the overall cluster diffuse light profile .
the psf - convolved devaucouleurs + background model is then fit to the data lying within these two limits as described above .
a second iteration of the fit is then performed with the maximum fit radius set to 1.5@xmath21 and an extra term is added to @xmath22 .
the second fit iteration uses a @xmath22 calculated as follows : @xmath23 where the subscripts @xmath24 and @xmath25 refer to the first and second fit iteration parameters respectively . the extra background term in @xmath26 acts as a penalty to any major change in the background parameter and ensures that the background is primarily determined / constrained by the larger radius fit range .
the reduction in the fit range of the second fit iteration helps to further isolate the bcg signal with an upper radial cutoff that is motivated by the shape of the galaxy itself .
examples of the marginalized @xmath22 contours are shown in figure [ fig : chi2surf2259 ] for the fits to the abell 2259 bcg surface brightness profile .
it is immediately seen that @xmath14 and @xmath15 are the most degenerate parameter pair , highlighting a common trait of all of bcg surface brightness fits in the cccp sample . by leaving the background as a free parameter in the fitting routine , even though the image has been background subtracted , this technique shows how the remaining structural properties of the bcg are affected by a possible error in background . by inspection of the @xmath22 contours shown in figures [ fig : chi2surf2259 ]
it can be seen that a 3@xmath27 error in the background can affect the derived @xmath4 effective radius @xmath15 at the 3 - 4 % level for abell 2259 .
the marginalized @xmath22 surface diagrams are also useful for analyzing the stability of the fitting routine for a given bcg . in summary , these data indicate that our bcg fitting procedure determines the structural parameters with high precession ( within a few % ) and that the effects of error in the background subtraction are negligible .
figure [ fig : sb31_mega ] shows the surface brightness profiles in the @xmath3 and @xmath4 bands for all bcgs that were observed with megacam while figure [ fig : sb23_12k ] shows the surface brightness profiles of all bcgs that were observed with cfh12k .
the measured data are shown by the points , solid lines correspond to best fit psf - convolved @xmath12 models as described in the above section and dotted lines show similar fits without psf convolution . surface brightness data is colour coded by wavelength band with @xmath4 and @xmath28 data shown in red and @xmath3 and @xmath29 data shown in blue . in all cases the bcgs are brighter in the red filter than in the blue filter with a typical offset on the order of one magnitude at all radii .
vertical dashed lines show the moffat psf fwhm radius ( @xmath30 ) .
residuals are shown along the bottom of each plot . in each sub - panel
the dotted line corresponds to the difference between the psf - convolved and non - psf - convolved fits , and the solid lines with error bars show the difference between the psf - convolved models and the measured data .
note that the @xmath29-band data for ms0016 has been omitted because it was taken under non - photometric conditions .
see section [ sbprodiscuss ] for discussion .
figure [ fig : c31_mega ] shows the colour profiles in ( @xmath3-@xmath4 ) for the megacam bcgs , while figure [ fig : c22_12k ] shows the colour profiles in ( @xmath29-@xmath28 ) for the cfh12k bcgs . as before ,
the vertical dashed lines denote the @xmath30 of the psfs in @xmath4 or @xmath28 ( red ) and @xmath3 or @xmath29 ( blue ) .
points with error bars indicate the measured data and solid lines show the difference between the red and blue @xmath12 model fits to the surface brightness distribution .
note that even though this is not a direct fit to the colour profile the difference between the individual fits to the data frames match the colour data well in most cases .
see section [ colourprodiscuss ] for discussion .
the measurement of surface brightness profiles of bcgs requires a detailed understanding of the various sources of systematic effects that may contaminate the signal .
previous bcg and icl studies have employed a wide variety of tests that aim to quantify these sources of contamination ( gonzalez et al . 2005 , krick et al . 2005 , patel et al .
the following sections describe the tests that are used in this paper to investigate the contributions of systematics effects and the steps taken to minimize them .
the measurement of bcg surface brightness is complicated by the effects of crowding .
the shallow surface brightness profiles of other luminous galaxies in the field often overlap with that of the bcg . the method used to create the mask
therefore systematically affects the measured profile .
the most common mask generation method used in the literature rejects pixels around non - bcg galaxies that lie inside a particular isophote @xmath31 ( lin and mohr 2004 , feldmeier et al .
2004 , zibetti et al .
2005 , krick et al .
the advantage of this technique is that the methodology is clearly defined and reproducible , the disadvantage however , is that it does not take into account the profile shapes .
the mask generation method used here is not limited to a particular isophotal cutoff and faint shallow wings of bcg neighbors are masked extensively . in this section
we discuss the tests used to evaluate the influence of mask effects on surface brightness .
the test described here is designed to look for signal contamination by comparing the profiles in different angular sections of the bcg .
the image is divided along the bcg major and minor axes into four quadrants . under the assumption of zero contamination ,
the profiles measured in each quadrant should be identical .
the differences between the profiles in different quadrants is a measure of the asymmetric component of the contamination signal .
there are other causes of variation in the quadrant profiles such as asymmetric masking and position angle twists .
for instance , if an individual quadrant is heavily masked near the minor axis then the measured profile in that quadrant is more representative of the surface brightness along the major axis and is brighter than the ideal unmasked , zero contamination case . in this way ,
slight variation in the masking from one quadrant to another introduces scatter among the four quadrant profiles .
this test is applied to all bcgs and the presence of large scatter in the quadrant profiles is used to determine the radial limit to which the bcg surface brightness data are reliable in each of the images ( @xmath32 ) . the plots shown in figure [ fig : mask_incomp2.eps ] show the result of the masking incompleteness tests for the bcgs in the clusters abell 2261 and abell 2537 .
the profiles in the four quadrants are shown as different coloured curves .
dotted lines indicate the radial limit ( @xmath32 ) to which the the bcg profile is minimally affected by asymmetric mask incompleteness .
the profile deviations seen in the @xmath4 image of abell 2537 just inside @xmath32 are due to a position angle twist of the bcg .
this effect is symmetric in two of the four quadrants and anti - symmetric in the other two .
another major source of systematic error in bcg / icl studies comes about from large scale flatfielding error and large scale variations in the background .
feldmeier et al .
( 2002 ) and patel et al .
( 2006 ) present methods that can be used in order to quantify some of the properties of background variation on a variety of scales .
similar methods are used here in order to quantify the image background variability .
each image is portioned into a rectangular grid .
the background statistics are determined within each grid square using the same mask that is used to measure bcg surface brightness .
this process is repeated with variety of grid sizes in order to examine background variation on different scales .
the results of this procedure reveal the presence of structure in the background as a function of distance from the bcg centre and size of the grid . to quantify the effects of large - scale structure we examine the statistics of 10 independent surface brightness profiles that are centreed on various parts of the image background . from the analysis of several cluster images ,
we conclude that the variance in the background signal begins to dominate the surface brightness measurements at a typical limit of @xmath33 @xmath34 but that brighter than this limit , large - scale error does not significantly contribute to the bcg signal .
we test for the possibility that some of the bcg / icl might be removed in the background subtraction process due to its large extent . examining a range of mesh sizes used for background determination ,
we conclude that this is not an important effect for the results presented here ( 5% difference in @xmath15 in the most extreme case ) . | we present measurements of surface brightness and colour profiles for the brightest cluster galaxies ( bcgs ) in a sample of 48 x - ray luminous galaxy clusters .
these data were obtained as part of the canadian cluster comparison project ( cccp ) .
the kormendy relation of our bcgs is steeper than that of the local ellipticals , suggesting differences in the assembly history of these types of systems .
we also find that while most bcgs show monotonic colour gradients consistent with a decrease in metallicity with radius , 25% of the bcgs show colour profiles that turn bluer towards the centre ( blue - cores ) .
we interpret this bluing trend as evidence for recent star formation .
the excess blue light leads to a typical offset from the red sequence of @xmath0 to @xmath1 @xmath2 in ( @xmath3-@xmath4 ) , thus affecting optical cluster studies that may reject the bcg based on colour .
all of the blue - core bcgs are located within @xmath5kpc of the peak in the cluster x - ray emission .
furthermore , virtually all of the bcgs with recent star formation are in clusters that lie above the @xmath6 relation .
based on photometry alone , these findings suggest that central star formation is a ubiquitous feature of bcgs in dynamically relaxed cool - core clusters .
this implies that while agns and other heating mechanisms are effective at tempering cooling , they do not full compensate for the energy lost via radiation .
[ firstpage ] galaxies : clusters : general galaxies : elliptical and lenticular , cd galaxies : evolution cooling flows x - rays : galaxies : clusters |
the last decade has seen increasingly detailed development of the fractional equations describing the anomalous transport in physics , biology , chemistry [ 14 ] .
special attention has been paid to slow subdiffusive transport for which mean squared displacement is sublinear @xmath3 where @xmath0 is the anomalous exponent @xmath4 .
subdiffusion is experimentally observed for proteins and lipids on cell membranes @xcite , rna molecules in the cells @xcite , transport in spiny dendrites santa , etc .
the major feature of this process is the absence of characteristic microscopic time scale .
the theory of anomalous subdiffusion leads to fractional partial differential equations involving memory effects .
if we introduce the probability density function @xmath5 for finding the particle in the interval @xmath6 at time @xmath7 , then the subdiffusive transport of the particles under the influence of external time - independent force can be described by the fractional fokker - planck ( ffp ) equation @xmath8with@xmath9 the riemann - liouville derivative @xmath10 is defined as@xmath11and the anomalous exponent @xmath4 is assumed to be constant .
the central result of this paper is that the subdiffusive fractional equations with constant @xmath0 in a bounded domain @xmath1 $ ] are not structurally stable with respect to the non - homogeneous variations of parameter @xmath0 .
it turns out that the space variations of the anomalous exponent lead to a drastic change in asymptotic behavior of @xmath5 for large @xmath12 to show this high sensitivity to non - homogeneous perturbations , one can consider the following exponent @xmath13with constant @xmath0 and perturbation @xmath14 ( see fig .
the asymptotic long - time behavior of the density @xmath5 with ( [ per ] ) is quite different from that of the solution to ( [ fp ] ) with the constant value of @xmath0 .
it means that the standard subdiffusive equation with constant @xmath0 is not a robust model for subdiffusive transport in heterogeneous complex media . on the interval
@xmath15 $ ] . ]
[ fig1 ] now let us explain our main result .
the standard way to deal with the fractional equation like ( [ fp ] ) in the bounded domain @xmath1 $ ] is a method of separation of variables [ 1 ] .
let us consider the case of the reflecting boundaries at @xmath16 and @xmath17 when ( [ fp ] ) has a stationary solution @xmath18 satisfying @xmath19 we can write a partial solution of ( [ fp ] ) as @xmath20 then the time evolution is described by fractional relaxation equation @xmath21where @xmath22 is the separation constant .
the slow relaxation process from the initial distribution @xmath23 is described by @xmath24where @xmath25 and @xmath26 are the eigenfunctions of @xmath27here the operator @xmath28 is the adjoint of @xmath29 @xmath30 ( see details in @xcite ) .
the only difference between standard fokker - planck equation and ffp equation is the rate of relaxation of @xmath31 . in the anomalous case
the relaxation process is very slow and it is described by a mittag - leffler function @xmath32 with the power - law decay @xmath33 as @xmath34 the exponential decay @xmath35 is recovered for @xmath36 . in this paper we show that if we consider nonuniform perturbations of anomalous exponent as ( [ per ] ) , this relaxation picture is completely changed .
the method of separation of variables does not work for space dependent @xmath37 the asymptotic behavior of @xmath38 as @xmath39 is essentially different from that given by ( [ as ] ) .
it turns out that in the limit @xmath39 the probability density @xmath38 concentrates around the point @xmath40 , where the perturbation @xmath14 is located , while the stationary distribution @xmath18 is completely irrelevant ( see fig . 2 and fig .
the question is how to take into account the non - uniform distribution of the anomalous exponent @xmath0 .
we can not simply substitute the expression like ( [ per ] ) into ( [ fp ] ) .
so we need a fractional master equation with space dependent @xmath37 chechkin , gorenflo and sokolov were the first to derive the fractional diffusion equation with varying fractional exponent @xcite .
they studied a composite system with only two separate regions with different anomalous exponents and found interesting effects involving non - trivial average drift .
a similar phenomenon has been analyzed in terms of two equations with different exponent by korabel and barkai @xcite .
here we present an alternative derivation which is valid for a general space and time dependent jump densities . consider a ` space - jump ' random walk model in one space dimension .
the particle movement can be described as follows .
it waits for a random time ( residence time ) @xmath41 at each point @xmath40 in space before making a jump to another point .
the index @xmath40 indicates that the waiting time @xmath41 depends on a space coordinate @xmath40 .
it is convenient to define the hazard function @xcite as the escape rate of a walker from the point @xmath40 @xmath42next step is the introduction of the structured probability density function @xmath43 that the particle position @xmath44 at time @xmath7 is in the interval @xmath45 and its residence time @xmath41 at point @xmath40 is in the interval @xmath46 the advantage of the structured density @xmath47 is that a random walk can be considered as markovian .
this is a standard way to deal with non - markovian processes @xcite ( see also vr , yh , fedo ) .
this density @xmath43 obeys the balance equation @xmath48here we consider only the case when the residence time of random walker at @xmath49 is equal to zero , so the initial condition is @xmath50where @xmath23 is the density for the initial position @xmath51 .
the boundary condition at @xmath52 can be written as @xcite @xmath53where @xmath54 is the probability density for jumps @xmath55 from the point @xmath40 at time @xmath7 ( jumps are independent from the residence time ) .
our purpose now is to derive the fractional master equation for the probability density @xmath56it is convenient to introduce the integral escape rate @xmath57and integral arrival rate @xmath58as the density of particles with zero residence time .
the boundary condition ( [ arr ] ) can be rewritten as @xmath59differentiation of ( [ deng ] ) with respect to time and substitution of @xmath60 from ( [ basic ] ) together with ( [ ba0 ] ) gives @xmath61to close this equation we need to express the escape rate @xmath62 in terms of @xmath63 we solve ( [ basic ] ) by the method of characteristics@xmath64here we recognize the survival function @xcite@xmath65so the structural density @xmath47 can be rewritten as@xmath66the residence time pdf @xmath67 is related to @xmath68 as @xmath69the balance equation for @xmath38 can be found by substitution of ( [ s1 ] ) and the initial condition @xmath70 into ( [ deng ] ) @xmath71to obtain the equation for @xmath62 we substitute ( [ s1 ] ) and the initial condition into ( [ bau1 ] ) @xmath72using the laplace transform in ( [ ba1 ] ) and ( [ ba00 ] ) we eliminate @xmath73 and obtain @xcite @xmath74where @xmath75 is the memory kernel defined by its laplace transform @xmath76 let us consider the anomalous subdiffusive case with the survival probability @xcite @xmath77 , \ 0<\mu ( x)<1 , \label{ml}\]]where @xmath78 $ ] is the mittag - leffler function .
the laplace transforms of @xmath79 and @xmath80 are @xmath81the laplace transform of the memory kernel @xmath82 is@xmath83and the integral escape rate @xmath84 can be written as @xmath85substitution of this expression into ( [ master4 ] ) gives the fractional master equation @xmath86where @xmath87 is the riemann - liouville fractional derivative with varying order .
this equation can be used to derive the general fokker - planck equation @xcite .
if we assume that the anomalous exponent @xmath0 and time parameter @xmath88 are independent from coordinate @xmath89 , this equation can be rewritten in terms of caputo derivative@xmath90it should be noted that the fractional equation with the caputo derivative can not be served as a model for subdiffusion in heterogeneous media with varying in space anomalous exponent @xmath91
. master equation ( [ master8 ] ) can be a starting point for deriving nonlinear fractional equations .
if instead of @xmath92 we consider the mean density of particles @xmath93 and assume that jump pdf @xmath94 depends on @xmath95 then one can write @xmath96expansion of this equation in @xmath55 can give a variety of fractional non - linear pde s .
as an example , let us consider the case of the symmetrical kernel @xmath97 for which the first moment @xmath98 then ( [ master99 ] ) can be approximated by a non - linear fractional equation @xmath99with varying anomalous exponent @xmath91 and nonlinear fractional diffusion coefficient @xmath100 @xmath101 @xmath102 first , let us consider random walk on a lattice with the space size @xmath103 we denote the probability of a particle moving right and left from the point @xmath40 as @xmath104 and @xmath105 correspondingly ( @xmath106 ) .
then the jump pdf can be written as@xmath107the fractional master equation ( [ master8 ] ) takes the form @xmath108 in the limit of small @xmath109 and @xmath110 @xcite one can obtain from ( walk ) the ffp equation with varying anomalous exponent @xmath111with the finite values of the fractional diffusion coefficient @xmath112 and fractional drift @xmath113 @xmath114note that in order to keep the fractional drift @xmath115 finite as @xmath116 , we need to assume that @xmath117 if we put the reflecting barriers at @xmath16 and @xmath17 and consider constant exponent @xmath0 and diffusion @xmath118 then the ffp equation ( [ fpnew ] ) admits the stationary solution in the form of the gibbs - boltzmann distribution@xmath119 , \quad u(x)=-\frac{1}{d_{\mu } ^{{}}}\int^{x}v_{\mu } ( z)dz \label{sta}\]]with @xmath120 dx$ ] . if @xmath0 is constant , the fractional time derivative does not affect the gibbs - boltzmann distribution @xcite . but this result is structurally unstable with respect to any non - uniform variations of @xmath121 let us show now that the gibbs - boltzmann distribution ( [ sta ] ) is absolutely irrelevant for the long time behavior of the solution to the ffp equation ( [ fpnew ] ) with non - uniform distribution of @xmath91 ( [ per ] ) .
we divide the interval @xmath15 $ ] into @xmath122 discrete states . at each state @xmath123
, the probability of jumping in the neighborhood to the left or right is given respectively by @xmath124 and @xmath125 ( @xmath126 ) .
the fractional equation ( [ walk ] ) for @xmath127 can be rewritten as @xmath128subject to the conditions @xmath129 @xmath130 and @xmath131 @xmath132 .
note that the ffp equation ( [ fpnew ] ) is just a continuous approximation of eq .
( [ dis ] ) .
taking the laplace transform of ( [ dis ] ) and using @xmath133 , we obtain @xmath134if one @xmath135 is smaller than the others ( @xmath136 @xmath137 , one can find that @xmath138 and @xmath139 as @xmath140 it means that in the limit @xmath141 we obtain @xmath142this result in a continuous case can be rewritten as @xmath143 as @xmath141 where @xmath144 is the point on the interval @xmath1 $ ] at which @xmath91 takes its minimum value .
a similar result was obtained for a symmetrical random walk in @xcite in the context of chemotaxis ( anomalous aggregation ) .
note that shushin @xcite considered a two - state anomalous system with different anomalous exponent @xmath0 and found that in the long time limit the probability is located in the slower state ( see also @xcite ) .
to validate our results , we run monte carlo simulations with the following procedure .
random numbers with uniform distribution , @xmath145 and @xmath146 , are generated and then transformed into mittag - leffler distributed random numbers using the following inversion formula @xmath147 @xcite ( see for details @xcite ) .
we take @xmath148 and divide the interval @xmath149 $ ] into @xmath150 subintervals .
we use @xmath151 @xmath152 and @xmath153 this corresponds to @xmath154so the drift @xmath155 and the potential @xmath156 all the random walkers start in the same state @xmath157 , their number @xmath158 , @xmath159 for all @xmath123 , and the long time limit is set at @xmath160 first step is to compute the exact stationary pdf given by ( [ sta ] ) and see how well our monte carlo simulations work .
2 shows that the monte carlo simulations agree with the gibbs - boltzmann distribution . ) with @xmath161 for all @xmath123 .
gibbs - boltzmann distribution is represented by the line .
, title="fig : " ] [ fig2 ] ) when @xmath162 is subject to a perturbation .
the parameters are @xmath163 for all @xmath123 except @xmath164 for which @xmath165.,title="fig : " ] [ fig3 ] the next step is to show that the gibbs - boltzmann distribution ( [ sta ] ) is absolutely irrelevant as far as the long time behavior of non - uniform system is concerned . the anomalous exponent @xmath166 is assumed to be @xmath167 for all states except one , @xmath168 for which @xmath169 .
one can see from fig .
3 that in the long time limit the probability is concentrated at state @xmath164 . one can conclude that there is a complete break down in the predictions based on the ffp equation with uniform anomalous exponent .
if the system was structurally stable we would expect to see something more like fig . 2 again .
however , the outcome is completely dominated by the perturbation @xmath169 .
this result has a huge implication for modelling anomalous subdiffusive transport of proteins , porous media , etc . in reality
the environment in which anomalous transport takes place is never homogeneous .
several attempts have been made to take into account the random distribution of anomalous exponent ( see , for example , @xcite ) .
one can introduce pdf @xmath170 for a random @xmath0 and write down the distributed - order fractional fpe as @xmath171let us show that if we generate the random field @xmath91 along the space interval @xmath149,$ ] the asymptotic behavior of @xmath5 will be quite different from that of the average fractional equation ( [ order ] ) . of random anomalous exponent @xmath172 .
, title="fig : " ] [ fig4 ] along @xmath123 for @xmath173 .
, title="fig : " ] [ fig5 ] ) when @xmath162 is the random field represented in fig .
5.,title="fig : " ] [ fig6 ] fig .
4 shows the pdf @xmath170 which will be used to generate the discrete uncorrelated random field @xmath166 .
the probability is concentrated around the point @xmath174 such that @xmath175 .
this distribution is chosen so that extreme values are highly unlikely to occur , with a purpose to show that the extreme low values dominate the long time behavior .
5 shows one sample of random field @xmath91 on the interval @xmath149 $ ] which is subdivided into @xmath150 subintervals ( @xmath152 ) .
5 shows clearly that the values of @xmath166 fluctuate around the mean .
the value at @xmath176 has a very small probability , since@xmath177 .
it is a very unlikely event , yet one can see from fig .
6 the state @xmath178 completely dominates the long time outcome of ( [ dis ] ) .
this phenomenon can be interpreted as a ` black swan ' .
the distribution of @xmath5 is highly intermediate for large @xmath7 , so the average behavior described by ( [ order ] ) can be very misleading .
it has been found @xcite that the distribution of the anomalous exponent in eq.([order ] ) leads to ultra - slow kinetics , but the stationary distribution is still given by the gibbs - boltzmann distribution @xcite .
our results show that random space variation of the anomalous exponent leads to completely different behavior in the long time limit ( see fig .
it should be noted that anomalous diffusion is just an intermediate asymptotic .
when time tends to infinity we expect a cross - over from anomalous diffusion to normal diffusion , and then we will recover the gibbs - boltzmann distribution . the standard tool for studying a subdiffusion is a subordination technique @xcite with constant anomalous exponent .
it would be interesting to apply similar technique if possible to non - homogeneous case
. it would be also interesting to take into account chemical reactions together with non - uniform anomalous exponent @xcite .
we have demonstrated that when the anomalous exponent @xmath0 depends on the space variable @xmath40 , the gibbs - boltzmann distribution is not a long time limit of the fractional fokker - planck equation .
even very small variations of the exponent lead to a drastic change of @xmath5 in the limit @xmath39 .
we have derived the fractional master equation with space dependent anomalous exponent .
we analyzed asymptotic behavior of corresponding lattice model in a finite domain with @xmath122 states with different exponents .
we have found that in this situation the probabilities@xmath179 do not converge to the stationary distribution . to illustrate our ideas
, we ran monte carlo simulations which show a complete break down in the predictions based on the ffp equation with uniform anomalous exponent .
further , we have shown that the idea of taking into account the randomness of anomalous exponent @xmath0 by averaging the fractional equation with respect to the distribution @xmath180 is not applicable to a non - homogeneous finite domain .
monte carlo simulations show that for every random realization of @xmath91 the pdf @xmath5 is highly intermediate , so the average behavior can be misleading .
although it is possible in theory to have a completely homogeneous environment , in which @xmath0 is uniform , it is not useful in any real application like chemotaxis @xcite or morphogen gradient formation @xcite because any non - homogeneous variation destroys the predictions based on this model in the long time limit .
m. m. meerschaert , d. a. benson , h. p. scheffler , and b. baeumer , phys .
e 65 , 041103 ( 2002 ) ; m. magdziarz a. weron , and k. weron , phys . rev .
e 75 , 016708 ( 2007 ) ; a. weron , m. magdziarz , and k. weron , phys . rev .
e 77 , 036704 ( 2008 ) .
a. yadav and w. horsthemke , phys .
e * 74 * , 066118 ( 2006 ) ; y. nec and a. a. nepomnyashchy , j. phys .
a : math . theor . * 40 * 14687 ( 2007 ) ; s. fedotov , a. iomin , phys .
_ , _ * 98 * 118101 ( 2007 ) ; s. fedotov , phys . rev .
e * 81 * , 011117 ( 2010 ) . | we derive the fractional master equation with space dependent anomalous exponent .
we analyze the asymptotic behavior of corresponding lattice model both analytically and by monte carlo simulation .
we show that the subdiffusive fractional equations with constant anomalous exponent @xmath0 in a bounded domain @xmath1 $ ] are not structurally stable with respect to the non - homogeneous variations of parameter @xmath0 .
in particular , the gibbs - boltzmann distribution is no longer the stationary solution of the fractional fokker - planck equation whatever the space variation of the exponent might be .
we analyze the random distribution of @xmath2 in space and find that in the long time limit , the probability distribution is highly intermediate in space and the behavior is completely dominated by very unlikely events . |
broad - line type ic supernovae ( sne ic ) are a subclass of core collapse sne ic that have broad features in their spectra , indicating unusually high expansion velocities reaching close to 0.1@xmath8 at early times .
only a few candidates of this class are known .
some broad - line sne ic are associated with gamma ray bursts ( grbs ) ( galama et al .
1998 ; matheson et al . 2003 ; malesani et al . 2004 ) , or x - ray flash ( xrf ) ( pian et al . 2006 ; modjaz et al . 2006 ) , while some others do not show any clear evidence of being associated with a grb or an xrf ( kinugasa et al . 2002 ; foley et al . 2003 ; valenti et al .
2008 ) . the broad - line sne ic exhibit diversity in terms of the explosion energy , ejecta mass and mass of @xmath5ni produced during the explosion .
the photometric and spectral features of sn 1998bw are explained with @xmath9 ejected with a kinetic energy @xmath1010@xmath11 ergs , producing @xmath12 of @xmath5ni in the explosion ( iwamoto et al .
1998 ; nakamura et al . 2001
; maeda et al . 2006 , tanaka et al . 2007 ) .
on the other hand , modelling of nebular spectra of sn 2002ap indicates an ejected mass of @xmath13 with a kinetic energy @xmath14@xmath1510@xmath16 ergs and production of @xmath17 of @xmath5ni ( mazzali et al .
2007 ) , showing a large range in the physical parameters of broad - line sne ic . among these ,
sne with kinetic energy @xmath18 ergs are termed as " hypernovae ( hne )
( iwamoto et al . 1998 ) .
the broad - line sne that are not associated with grbs are found to have smaller values of ejecta mass , explosion energy and lower luminosity as compared to the grb - associated hne ( nomoto et al .
2007 ) .
sn 2007ru was discovered by donati & ciabattari on november 27.9 and independently by winslow & li with kait on november 30.15 in the spiral galaxy ugc 12381 .
there was no evidence of the supernova in the kait image taken on november 22.16 , down to a limiting magnitude of 18.9 ( donati & ciabattari , 2007 ) . based on a spectrum obtained on december 1 , sn 2007ru was classified as peculiar sn ic at premaximum phase ( chornock et al .
the caii h & k and caii near infrared ( nir ) triplet absorption troughs were found to be weak compared to other sne ic .
further , the oi line at 7774 indicated an expansion velocity of 19000 @xmath19 , similar to the expansion velocity seen in the broad - line sn ic sn 2006aj ( pian et al .
2006 ) . a search through the reported discoveries of grbs during october 15 to november 30
, 2007 does not show any possible association of a grb with this supernova .
the results of photometric and spectral monitoring of sn 2007ru until around 210 days after maximum , using the 2 m himalayan chandra telescope ( hct ) of the indian astronomical observatory ( iao ) , hanle , india , are presented in this paper .
sn 2007ru was observed in @xmath20 bands during 2007 december 2 ( jd 2454437.09 ) to 2008 july 03 ( jd 2454651.35 ) .
the observations were carried out with the himalayan faint object spectrograph camera ( hfosc ) mounted on the hct .
hfosc is equipped with 2k@xmath154k pixels ccd chip .
the central 2k@xmath152k region , with a plate scale of 0.296 arcsec pixel@xmath21 , covering a field of 10@xmath22@xmath1510@xmath22 , was used for imaging .
gain and readout noise of the ccd camera are 1.22 e@xmath21/adu and 4.87e@xmath21 , respectively .
further details on the hct and hfosc can be obtained from `` http://www.iiap.res.in/centers/iao '' .
photometric standard regions ( landolt 1992 ) were observed on 2007 december 25 and december 26 under photometric sky conditions to calibrate a sequence of secondary standards in the supernova field .
data reduction has been done in the standard manner using various tasks available within image reduction and analysis facility ( iraf ) .
aperture photometry was performed on the photometric standard stars and secondary standards , at an aperture radius determined using the aperture growth curve .
the secondary standard stars were then calibrated using the average colour terms and the photometric zero points determined on the individual night . a sequence of secondary standards calibrated in this way is marked in fig . [ fig : field ] and the @xmath20 magnitudes of the secondary standards averaged over the two nights are listed in table [ tab : std_star_mag ] .
the magnitudes of the supernova and secondary standards were measured using point spread function photometry , with a fitting radius equal to the fwhm of the stellar profile .
supernova magnitudes were calibrated differentially with respect to the local standards .
the supernova magnitudes in @xmath23 , @xmath2 , @xmath3 , @xmath24 and @xmath25 bands have been listed in table [ tab : sn_mag ] .
a series of spectra were taken during 2007 december 3 to 2008 june 12 , in the wavelength range ( 3500 - 7000 ) and ( 5200 - 9200 ) , with a resolution of @xmath0 7 .
the journal of spectroscopic observations is given in table [ tab : spec_log ] .
the spectrophotometric standard stars were observed on the same night to flux calibrate the supernova spectra .
spectroscopic data reduction was carried out using tasks available within iraf .
the spectra were bias subtracted , flat - fielded and the one dimensional spectra were extracted using the optimal extraction method .
the arc lamp spectra of fear and fene were used for wavelength calibration .
the instrumental response correction was effected using the spectroscopic standard spectrum .
the spectra in the two different regions were combined , scaled to a weighted mean , to give the final spectrum on a relative flux scale , which were then brought to an absolute flux scale using the @xmath20 magnitudes .
the supernova spectra were corrected for the host galaxy redshift of @xmath26 ( from ned ) and dereddened by the total reddening @xmath27 as estimated in section 5 .
the @xmath20 light curves ( lc ) of sn 2007ru are shown in fig . [
fig : light ] . the unfiltered discovery magnitudes and the pre - discovery limiting magnitude
are also included in the figure .
the lcs indicate that the maximum occurred earlier in the blue , similar to other broad - line sne ic .
the date of explosion can be constrained to @xmath28 6 days before discovery ( november 25 , jd 2454430@xmath13 ) , based on the non - detection on november 22 by kait and the subsequent discovery on november 27.9 . the rise time to @xmath2
maximum , which occurred on december 3 , is 5 - 11 ( 8@xmath13 ) days , indicating sn 2007ru is a fast rising sn ic with a rise time similar to broad - line sne 2002ap ( foley et al .
2003 ) , and 2006aj ( modjaz et al . 2006 ) , marginally faster than sn 2003jd ( valenti et al . 2008 ) but considerably faster than the grb - associated sne 1998bw and 2003dh ( galama et al .
1998 , matheson et al . 2003 ) .
the light curves of sn 2007ru are compared with those of broad - line sne sn 2002ap , sn 2003jd , grb 980425/sn 1998bw , normal sne ic sn 1994i and sn 2004aw in fig .
[ fig : fig_comp_lc ] .
a comparison of the decline in brightness 15 days after maximum light , @xmath29m@xmath30 in different bands shows that sn 2007ru has a decline similar to other broad - line supernovae .
in fact , the decline of sn 2007ru is faster than sn 1998bw but slower than sne 2003jd and 2006aj ( refer table [ tab : parameters ] ) .
the decline rates are estimated to be @xmath29m@xmath30(@xmath2)=1.57 , @xmath29m@xmath30(@xmath3)=0.92 , @xmath29m@xmath30(@xmath24)=0.69 and @xmath29m@xmath30(@xmath25)=0.50 . the light curves of sn 2007ru decline with decline rates of 0.021 mag day@xmath21 in @xmath3 , 0.028 mag day@xmath21 in @xmath24 and 0.030 mag day@xmath21 in @xmath25 bands during days @xmath31 .
these decline rates are comparable to the light curve decline rates of the broad - line sn 2003jd and marginally faster than sn 1998bw .
the decline rate in @xmath2 can not be estimated due to a sparse coverage during this period . during the late phases ( @xmath32 80 days after explosion ) , the @xmath3 and @xmath24 band light curves of sn 2007ru decline with decline rates of 0.0152 and 0.0116 mag day@xmath21 ,
respectively , slower than both sn 2003jd and sn 1998bw during the corresponding epochs ( fig .
[ fig : fig_comp_lc ] ) .
the decline rate of sn 2007ru during the late phases , is faster than the rate expected due to the radio active decay of @xmath5co into @xmath5fe .
this indicates inefficient trapping of @xmath33-rays by the ejecta , which suggests a low column density .
the peak absolute magnitudes were estimated using the apparent magnitude at maximum in different bands ( see 5 for reddening and distance estimate ) . from a comparison of the absolute magnitude , sn 2007ru appears to lie at the brighter end of the observed sne ic . with an absolute @xmath3 magnitude of @xmath34 ,
sn 2007ru is fainter than grb 031203/sn 2003lw [ @xmath35=@xmath36 ; malesani et al .
2004 ] , comparable in brightness with sn 1998bw [ @xmath35=@xmath37 , galama et al .
1998 ] , and brighter than xrf 060218/sn 2006aj [ @xmath35=@xmath38 ; modjaz et al . 2006 ] , broad - line sne 2002ap [ @xmath35=@xmath39 ; foley et al .
2003 , tomita et al . 2006 ] , 2003jd [ @xmath35=@xmath40 ; valenti et al . 2008 ] and normal sne ic 1994i [ @xmath35=@xmath41 ; richmond et al .
1996 , sauer et al .
2006 ] , and 2004aw [ @xmath35=@xmath42 ; taubenberger et al . 2006 ] .
the spectral evolution of sn 2007ru is presented in figure [ fig : spec1 ] . the first spectrum , obtained on 2007 december 3 , at @xmath43 @xmath44 days , where @xmath43 is days after explosion , ( the rise time is assumed to be 8 days , see 3 ) shows broad absorption features at @xmath45 due to siii ( and a possible contamination as discussed by valenti et al .
2008 ) , at @xmath46 due to oi and at @xmath47 due to weak caii nir triplet .
the spectrum at @xmath43 @xmath48 days shows well developed broad absorptions due to mgii @xmath494481 , blends of feii at @xmath0 4700 , siii @xmath496355 , oi @xmath497774 and caii nir triplet .
the continuum gets redder by @xmath43 @xmath48 days .
no significant evolution is seen in the spectrum taken at @xmath43 @xmath50 days . the last spectrum presented here , obtained @xmath51 days after explosion indicates the supernova to be in the nebular phase . comparing the spectrum and its evolution with that of the broad - line sn 2003jd and the normal spectrum ic supernova sn 2004aw ( fig.[fig : comp_spec1 ] and fig.[fig : comp_spec2 ] )
, it is seen that sn 2007ru is very similar to the broad - line sn ic , indicating high expansion velocities .
the spectral features continue to remain broad and similar to those of sn 2003jd even at @xmath43 @xmath50 days .
the high velocity of the supernova ejecta usually results in blending of the spectral lines , making direct measurement of the photospheric velocity difficult .
the photospheric velocity of sn 2007ru is estimated by fitting a gaussian profile to the minimum of the absorption trough of si ii 6355 line , in the redshift corrected spectra . in the first spectrum , @xmath08 days after the explosion , the absorption feature at 6200 consists of two components , similar to that seen in the premaximum spectra of the broad - line sn 2003jd ( valenti et al .
assuming the blue component is due to si ii and the red wing is due to possible contamination by other species , the photospheric velocity is estimated to be 20,000 @xmath19 .
this is consistent with the velocity estimate by chornock et al .
( 2007 ) using oi line in the spectrum .
[ fig : phot_vel ] the photospheric velocity of sn 2007ru , measured using si ii lines , and its evolution is compared with other sne ic in fig [ fig : phot_vel ] .
the photospheric velocity of sn 2007ru at @xmath0 8 days after explosion is lower than grb 980425/sn 1998bw ( patat et al .
2001 ) , but comparable to xrf 060218/sn 2006aj ( pian et al . 2006 ) and broad - line sn 2002ap ( foley et al .
however , at the later epochs , the photospheric velocity of sn 2007ru is comparable to those of sn 1998bw and sn 2003jd and higher than other sne ic in comparison . except for the early phase ( @xmath52
15 days after explosion ) the photospheric velocity evolution of sn 2007ru is very similar to that of the broad - line sn 2003jd .
the spectrum of sn 2007ru taken @xmath0 200 days after explosion ( refer fig [ fig : comp_spec_neb ] ) is dominated by the forbidden emission lines of [ oi ] @xmath49@xmath496300 , 6364 and [ caii ] @xmath49@xmath497291 , 7323 , possibly blended with [ oii ] @xmath49@xmath497320 , 7330 ( taubenberger et al .
these lines show a broad profile . due to poor signal - to - noise ratio of our spectrum
it is difficult to identify other spectral lines in the spectrum .
however , narrow lines due to h@xmath53 , [ nii ] @xmath496583 , [ sii ] @xmath546713 , 6731 , originating from the underlying hii region at the supernova location and superimposed on the spectrum of the supernova , are clearly seen in the spectrum , and are identified in fig [ fig : comp_spec_neb ] .
a comparison of the nebular spectrum of sn 2007ru is made with those of other sne ic in fig [ fig : comp_spec_neb ] .
the line profile of the [ oi ] @xmath49@xmath496300 , 6364 lines shows a sharp peak , very similar to that seen in sn 1998bw , sn 2004aw and sn 2007ru .
interestingly , despite the spectral similarity between sn 2007ru and sn 2003jd at early phase , the [ oi ] line profile in the nebular spectra is different : sn 2003jd shows a double peaked structure ( see 7 for implications ) .
the profile of [ caii ] @xmath49@xmath497291 , 7323/[oii ] @xmath49@xmath497320 , 7330 line is similar to the [ oi ] line profile , as seen in sn 2004aw ( taubenberger et al .
2006 ) , but different from the profiles seen in sne 1998bw and 2003jd , which show a flat topped profile .
the [ oi ] @xmath49@xmath496300 , 6364 and [ caii ] @xmath49@xmath497291 , 7323 lines show a blueshift of 2300@xmath1300 @xmath19 and 1200@xmath1200 @xmath55 , respectively .
this could be due to a kinematic offset ( maeda et al . 2007 ) , optical depth effect ( filippenko et al .
1994 ) , or extinction by the dust formed in the sn ejecta ( although there is no strong indication of the dust formation in the light curve ( 3 ) ) the velocities derived from full width at half maximum of [ oi ] and [ caii ] lines are found to be 14000@xmath12200 @xmath19 and 13500@xmath11300 @xmath55 , respectively .
these values are comparable to those seen in sn 1998bw and sn 2003jd .
the reddening corrected [ oi]/[caii ] flux ratio is found to be @xmath0 1.6 , which is again comparable to the ratios in sn 1998bw and sn 2003jd . thus , though the profile of the [ oi ] and [ caii ] lines in the the nebular spectrum of sn 2007ru and sn 2003jd differ , other properties like line width and [ oi]/[caii ] flux ratio are similar .
direct distance estimates to the host galaxy of sn 2007ru are not available in the literature .
the radial velocity of ugc 12381 , corrected for local group infall onto the virgo cluster is 4832 @xmath19 ( leda ) . for @xmath56 = 72 @xmath57 ,
the distance modulus to ugc 12381 is @xmath58 , where the error is estimated taking into account the errors in hi velocity measurement of the galaxy ( paturel et al .
2003 ) and the uncertainty in @xmath56 .
the naid absorption line is clearly seen in the spectrum with an average equivalent width of [email protected] .
based on the equivalent widths of naid absorption seen in several sne , turatto et al .
( 2003 ) find two distinct relations between naid equivalent width and the reddening @xmath59 .
using these relations , the observed naid equivalent width indicates @xmath59 values of [email protected] and [email protected] .
the galactic interstellar reddening in the direction of ugc 12381 is estimated to be 0.26 ( schlegel , finkbeiner & davis 1998 ) .
the naid absorption seen in the spectra of sn 2007ru is clearly from the milky way galaxy and no component due to the host galaxy is detected .
hence , an @xmath59 value of 0.27 is used for extinction correction .
the quasi bolometric lc of sn 2007ru is estimated using the @xmath20 magnitudes corrected for reddening with @xmath59=0.27 and the cardelli et al .
( 1989 ) extinction law .
the magnitudes were converted to the monochromatic flux , using zero points from bessell et al .
the fluxes were then spline interpolated and integrated from 3100 to 1.06@xmath60 m . since @xmath23 band observations are not available beyond 35 days since @xmath2 maximum
, the bolometric lc is estimated by integrating the @xmath61 fluxes only . the contribution of @xmath23 band to the bolometric flux at phases @xmath62 days
is estimated to be @xmath63 . in the later phases when
only @xmath3 , @xmath24 , @xmath25 or @xmath3 , @xmath24 magnitudes are available , the bolometric magnitudes are derived by applying a bolometric correction to the available magnitudes .
the bolometric corrections were estimated based on the last four points for which @xmath2 , @xmath3 , @xmath24 and @xmath25 measurements are available .
the quasi bolometric lc is shown in fig [ fig : light_bol ] . adding a conservative uncertainty of @xmath64 , the bolometric magnitude at maximum
is estimated as @xmath65 .
the quasi bolometric lcs , estimated in a similar manner , for sn 1998bw , sn 2002ap , sn 2004aw and sn 1994i , are also plotted in fig .
[ fig : light_bol ] . the quasi bolometric lc of sn 2007ru is brighter than the other well studied non - grb broad - line sn 2002ap and normal sne ic , and comparable to sn 1998bw .
the decline in the bolometric lc of sn 2007ru is considerably faster than sn 1998bw and comparable to sn 2002ap . using arnett s rule ( arnett 1982 )
, the mass of @xmath5ni required to power the quasi bolometric lc of sn 2007ru is estimated to be @xmath66 , whereas it is @xmath67 for sn 1998bw ( fig .
[ fig : light_bol ] ) .
it is to be noted here that the contribution due to nir bands is not included in the bolometric lc .
the nir contribution to bolometric flux for broad - line sne 2002ap and 1998bw is @xmath68 ( tomita et al .
2006 , valenti et al .
2008 ) , whereas for sn 1994i it is only @xmath0 10@xmath69 , while for sn 2004aw the nir contribution increases from @xmath0 31@xmath69 to @xmath0 45@xmath69 between + 10 and + 30 days ( taubenberger et al . 2006 ) . assuming an nir contribution to the bolometric flux similar to sne 2002ap and 1998bw ( @xmath0 30@xmath69 )
, the mass of @xmath5ni for sn 2007ru is estimated to be @xmath0 0.4 @xmath70 .
the total rate of energy production via @xmath5ni- @xmath5co chain estimated using the analytical formula by nadyozhin ( 1994 ) , for different values of mass of @xmath5ni synthesized during the explosion and are plotted with the quasi bolometric lc ( thin lines ) in fig [ fig : light_bol ] .
the plots indicate a good match of the energy production rate for @xmath66 of @xmath5ni with the initial decline of the quasi - bolometric light curve of sn 2007ru , in agreement with the estimate based on arnett s rule .
an attempt is made to estimate the metallicity of the region where the supernova exploded , based on the observed [ nii]/h@xmath53 flux ratio , from the underlying hii region , superimposed in the nebular spectrum of the supernova .
following pettini & pagel ( 2004 ) , the n2 index ( log[nii]/h@xmath53 ) is estimated to be @xmath71 . using this ,
an oxygen abundance of @xmath72 is derived .
another way of deriving oxygen abundance is using the @xmath73/h\alpha)$ ] diagnostic diagram ( kewley & dopita , 2002 ) , which requires an estimate of the ionization parameter @xmath74 or @xmath23 ( @xmath75 ; @xmath8 is the speed of light ) also .
the ionization parameter _
u _ can be estimated from the [ sii]/[siii ] ratio following diaz et al .
( 1991 ) . the observed flux of [ sii ] lines @xmath49@xmath496717 , 6731 and [ siii ] line @xmath499069 , seen in the nebular spectrum can be used to estimate the ionization parameter , however , our spectrum does not cover the [ siii ] @xmath499532 region . in the extragalactic hii regions ratio ( [ siii]@xmath499532/[siii]@xmath499069 ) is found to vary in the range 1.58 to 3.77 with an observed mean of [email protected] , against the theoretical value of 2.44 ( diaz et al . 1985 , vilchez et al .
1988 , kennicutt & garnett 1996 ) this indicates that the ionization parameter @xmath74 can vary in the range @xmath76 to 3@xmath77 .
using the log([nii]/h@xmath53 ) diagnostic diagram for the above estimated range of the ionization parameter , the oxygen abundance 12 + log(o / h ) is found to lie close to 8.8 , which is in good agreement with the independent estimate of 12 + log(o / h ) = 8.78 .
this indicates the oxygen abundance in region where the supernova sn 2007ru occurred is close to solar . in a recent study modjaz et al .
( 2008b ) have concluded that the broad - line supernovae associated with grbs are generally found in metal poor environments as compared to the broad - line supernovae without grbs .
they have shown that , in their sample , the oxygen abundance 12 + log(o / h)@xmath78 = 8.5 can be treated as the boundary between galaxies that have grbs associated supernovae and those without grbs .
the estimated oxygen abundance of @xmath0 8.8 at the location of sn 2007ru fits well in the range expected for a broad - line supernova without grb . the nuclear spectrum of the host galaxy of sn 2007ru ( obtained on 2008 october 29 ) is shown in fig .
[ fig : nuc_spec ] .
the nuclear spectrum shows strong hydrogen lines of the balmer series , permitted as well as forbidden lines of oxygen , lines due to helium and the calcium nir triplet ( refer fig . [
fig : nuc_spec ] ) .
the fwhm velocities of the lines indicate velocities of the order of 500 - 700 @xmath19 .
the hydrogen balmer lines show broad wings , with a noticable asymmetry in the blue wing .
such broad wings are not seen in the forbidden lines .
another interesting feature of the nuclear spectrum of the host galaxy is the presence of numerous feii lines seen at wavelengths 44004600 , 4924 , 5018 and 51005400 , similar to the spectra of active galactic nuclei ( veron - cetty et al .
2004 , veron - cetty et al .
2006 ) . using the diagnostic diagram ( ho , filippenko & sargent 1997 ) based on the line ratio @xmath79\lambda 5007/h\beta)$ ] versus @xmath80\lambda 6300/h\alpha)$ ] and @xmath79\lambda5007/h\beta)$ ] versus @xmath81\lambda\lambda6717,6731/h\alpha)$ ] the nucleus of ugc 12381 can be classified as belonging to the hii region class , with very weak [ sii ] and [ nii ] lines .
it thus appears that the host galaxy of sn 2007ru probably hosts a mild agn with a nuclear hii region , which needs further detailed study .
the optical spectra of sn 2007ru presented here show broad spectral features similar to that seen in grb - associated sn 1998bw , xrf - associated sn 2006aj and the broad - line sn 2003jd .
the expansion velocity of the ejecta of sn 2007ru is higher than that of the normal sn ic sn 1994i and sn 2006aj , but comparable to that of sn 1998bw and sn 2003jd .
the maximum luminosity of sn 2007ru is comparable to sn 1998bw .
the light curve reaches a peak in only @xmath82 days , which is remarkably faster than in sn 1998bw ( @xmath0 20 days ) .
the mass of @xmath5ni ejected in sn 2007ru is estimated as @xmath83 , which is similar to that estimated for sne 1998bw and 2003jd , slightly larger than that for sn 2004aw , and much larger than that for the broad - line sne 2002ap and 2006aj and the normal sn ic 1994i ( table [ tab : parameters ] ) . from the rise time of the light curve ( @xmath84 ) and the expansion velocity ( @xmath85 ) , we estimate mass of sn ejecta ( @xmath86 ) and the kinetic energy of the ejecta ( @xmath87 ) .
if the optical opacity is assumed to be constant , the timescale of the light curve is expressed as @xmath88 ( arnett 1982 ) .
the expansion velocity is given by @xmath89 .
the rise time of sn 2007ru ( @xmath90 days ) is comparable to that of the well studied sn 1994i , while the expansion velocity ( @xmath91 @xmath19 at maximum ) is about twice .
assuming @xmath92 and @xmath93 ergs for sn 1994i ( see table [ tab : parameters ] ) , and the observed expansion velocity of sn 2007ru , we estimate @xmath94 and @xmath95 ergs for sn 2007ru . a similar analysis for sn 1998bw , assuming the rise time and the velocity of sn 1998bw to be twice that of sn 1994i , leads to @xmath96 and @xmath97ergs , qualitatively consistent with the results of detailed modelling ( iwamoto et al . 1998 ; nakamura et al .
ratio ( @xmath98 ) , the higher end of the ejecta mass may be preferred ( @xmath99 , and @xmath100 erg ) .
since the explosion with a larger kinetic energy can produce a larger amount of @xmath5ni , the large @xmath5ni mass in sn 2007ru may also support this . ]
if we take sn 2003jd as a reference ( valenti et al . 2008 ) , @xmath101 and @xmath102 ergs are derived for sn 2007ru .
it should however be noted , that a detailed modelling is required to derive accurate values of @xmath103 and @xmath87 . and kinetic energy @xmath87 using the width of the light curve .
the width is defined as the time from peak of the bolometric light curve to the time when the luminosity is equal to the 1/_e _ times the peak luminosity , which is equivalent to a decline of 1.1 mag from peak .
the derived width is @xmath0 18 days for sn 2007ru . with this value
, we have estimated @xmath103 and @xmath87 in the same manner and derived larger values of @xmath103 and @xmath87 ( @xmath104 and @xmath105 erg ) .
our conclusion of large e / m is thus not affected . ]
sn 2007ru has a large kinetic energy while the ejecta mass is close to that of normal sne ic .
studies of sne ic ( e.g. , nomoto et al . 2007 ) have shown a trend , although weak , wherein sne having massive ejecta tend to have a larger kinetic energy and eject more @xmath5ni , connecting normal sne to grb - associated sne .
in contrast to this trend , sn 2007ru which resides at the higher energy end has a lower mass ejecta , leading to a higher @xmath106 .
sn 2007ru thus adds to the diversity of sne ic .
the spectroscopic properties of sn 2007ru at early phases are most similar to sn 2003jd .
also if we take the higher end of @xmath103 , the ejecta properties of sn 2007ru are also close to those of sn 2003jd ( although the estimated @xmath106 is higher for sn 2007ru ) .
however , the [ oi ] line profiles in the nebular spectra are dissimilar .
sn 2007ru shows a single peaked profile while sn 2003jd shows a double - peaked profile . in aspherical explosions , we would expect a single - peaked [ oi ] profile for the polar - viewed case , and a double - peaked profile for the side - viewed case ( mazzali et al .
2005 ; maeda et al . 2008 ; modjaz et al .
also , the polar - viewed aspherical explosion tends to show a brighter peak ( maeda et al .
2006 ) and faster velocity ( tanaka et al .
this matches with the properties of sn 2007ru .
thus , we suggest that sn 2007ru could be an aspherical explosion viewed from the polar direction . detailed multidimensional modelling is required to answer if the high @xmath106 derived for sn 2007ru results from the effect of asphericity .
the nebular spectrum of sn 2007ru shows narrow emission lines due to h@xmath53 , [ nii ] , [ sii ] and [ siii ] , arising from the underlying / neighbouring host galaxy hii region .
the flux ratios indicate an oxygen abundance of @xmath0 8.8 in the region of the supernova .
the nearly solar oxygen abundance at the location of the supernova matches well with earlier abundance studies for supernova host galaxies .
the nuclear spectrum of the host galaxy of sn 2007ru shows broad hydrogen balmer lines , with an asymmetric blue wing .
emission lines due to feii are also fairly prominent .
low ionization emission lines are also present . it appears that the galaxy hosts a mild agn with nuclear hii region .
we are thankful to the anonymous referee for valuable comments and t.p .
prabhu for useful discussions .
this work has been carried out under the insa ( indian national science academy ) - jsps ( japan society for promotion of science ) exchange programme .
this research was supported in part by the national science foundation under grant no .
phy05 - 51164 , world premier international research center initiative , mext , japan , and by the grant - in - aid for scientific research of the jsps ( 18104003 , 18540231 , 20540226 ) and mext ( 19047004 , 20040004 ) .
is supported through the jsps research fellowships for young scientists .
we thank all the observers of the 2-m hct ( operated by the indian institute of astrophysics ) , who kindly provided part of their observing time for the supernova observations . this work has made use of the nasa astrophysics data system and the nasa / ipac extragalactic database ( ned ) which is operated by jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
taubenberger , s. , et al .
2006 , mnras , 371 , 1459 tomita , h. , et al .
2006 , apj , 644 , 400 turatto m. , benetti s. , cappellaro e. , 2003 , in hillebrandt w. , leibundgut b. , eds , proceedings to the eso / mpa / mpe workshop `` from twilight to highlight : the physics of supernovae '' .
springer , berlin , p. 200
lccccc 1 & @xmath107 & @xmath108 & @xmath109 & @xmath110 & @xmath111 + 2 & @xmath112 & @xmath113 & @xmath114 & @xmath115 & @xmath116 + 3 & @xmath117 & @xmath118 & @xmath119 & @xmath120 & @xmath121 + 4 & @xmath122 & @xmath123 & @xmath124 & @xmath125 & @xmath126 + 5 & @xmath127 & @xmath128 & @xmath129 & @xmath130 & @xmath131 + 6 & @xmath132 & @xmath133 & @xmath134 & @xmath135 & @xmath136 + 7 & @xmath137 & @xmath138 & @xmath139 & @xmath140 & @xmath141 + 8 & @xmath142 & @xmath143 & @xmath144 & @xmath145 & @xmath146 + 9 & @xmath147 & @xmath148 & @xmath149 & @xmath150 & @xmath151 + 10 & @xmath152 & @xmath153 & @xmath154 & @xmath155 & @xmath156 + 11 & @xmath157 & @xmath158 & @xmath159 & @xmath160 & @xmath161 + 12 & @xmath162 & @xmath163 & @xmath164 & @xmath165 & @xmath166 + 13 & @xmath167 & @xmath168 & @xmath169 & @xmath170 & @xmath171 + 14 & @xmath172 & @xmath173 & @xmath174 & @xmath175 & @xmath176 + [ tab : std_star_mag ] lccccccc 02/12/2007 & 437.086 & -2 & @xmath177 & @xmath178 & @xmath179 & @xmath180 & @xmath181 + 04/12/2007 & 439.155 & 0 & @xmath182 & @xmath183 & @xmath184 & @xmath185 & @xmath186 + 10/12/2007 & 445.029 & 6 & @xmath187 & @xmath188 & @xmath189 & @xmath190 & @xmath191 + 11/12/2007 & 446.047 & 7 & @xmath192 & @xmath193 & @xmath194 & @xmath195 & @xmath196 + 12/12/2007 & 447.027 & 8 & @xmath197 & @xmath198 & @xmath199 & @xmath200 & @xmath201 + 18/12/2007 & 453.046 & 14 & & @xmath202 & @xmath203 & @xmath204 & @xmath205 + 19/12/2007 & 454.056 & 15 & & @xmath206 & @xmath207 & @xmath208 & @xmath209 + 22/12/2007 & 457.067 & 18 & & @xmath210 & @xmath211 & @xmath212 & @xmath213 + 24/12/2007 & 459.097 & 20 & & @xmath214 & @xmath215 & @xmath216 & @xmath217 + 25/12/2007 & 460.069 & 21 & & @xmath218 & @xmath219 & @xmath220 & @xmath221 + 26/12/2007 & 461.079 & 22 & @xmath222 & @xmath223 & @xmath224 & @xmath225 & @xmath226 + 30/12/2007 & 465.061 & 26 & @xmath227 & @xmath228 & @xmath229 & @xmath230 & @xmath231 + 31/12/2007 & 466.029 & 27 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 + 04/01/2008 & 470.045 & 31 & @xmath237 & @xmath238 & @xmath239 & @xmath240 & @xmath241 + 21/01/2008 & 487.041 & 48 & & & @xmath242 & @xmath243 & @xmath244 + 24/01/2008 & 490.063 & 51 & & @xmath245 & @xmath246 & @xmath247 & @xmath248 + 03/02/2008 & 500.055 & 61 & & @xmath249 & @xmath2504 & @xmath251 & @xmath252 + 29/02/2008 & 526.068 & 87 & & & @xmath253 & @xmath254 & @xmath255 + 03/05/2008 & 590.431 & 151 & & & @xmath256 & @xmath257 & @xmath258 + 06/06/2008 & 624.400 & 185 & & & @xmath259 & @xmath260 & + 17/06/2008 & 635.360 & 196 & & & @xmath261 & @xmath262 & + 03/07/2008 & 651.348 & 212 & & & @xmath263 & @xmath264 & + [ tab : sn_mag ] lcrc 03/12/07 & 438.21 & @xmath265 & 5200 - 9100 + 10/12/07 & 445.06 & @xmath266 & 3500 - 7000 ; 5200 - 9100 + 11/12/07 & 446.07 & @xmath267 & 3500 - 7000 ; 5200 - 9100 + 15/12/07 & 450.14 & @xmath268 & 3500 - 7000 ; 5200 - 9100 + 19/12/07 & 454.13 & @xmath269 & 3500 - 7000 ; 5200 - 9100 + 22/12/07 & 457.09 & @xmath270 & 3500 - 7000 + 04/01/08 & 470.09 & @xmath271 & 3500 - 7000 ; 5200 - 9100 + 03/02/08 & 500.08 & @xmath272 & 3500 - 7000 + 12/06/08 & 630.38 & @xmath273 & 5200 - 9100 + [ tab : spec_log ] sn 1994i & @xmath274 & 1.65 & 0.029 & 0.028 & 0.026 & 1 & 0.91 & 0.07(0.07 ) & @xmath275 & 1 , 2 , 3 + sn 2004aw & @xmath276 & 0.62 & 0.014 & 0.017 & 0.015 & 3.59.0 & 3.58.0 & 0.250.35(0.2 ) & @xmath275 & 4 + sn 2003jd & @xmath277 & 1.44 & 0.022 & 0.022 & 0.029 & 7@xmath278 & [email protected] & 0.36 & @xmath279 & 5 + sn 2002ap & @xmath280 & 0.87 & & & & 4 & 2.5 & 0.1(0.06 ) & @xmath281 & 6 , 7 , 8 + sn 2007ru & @xmath4 & 0.92 & 0.021 & 0.028 & 0.030 & 5@xmath282 & 1.3@xmath283 & 0.4 & @xmath284 & this work + sn 1998bw & @xmath285 & 0.75 & 0.020 & 0.022 & 0.022 & 30 & 10 & 0.40(0.5 ) & @xmath286 & 9 , 10 , 11 + sn 2006aj & @xmath287 & 1.14 & & & & 2 & 2 & 0.21 & @xmath275 & 12 , 13 + | photometric and spectral evolution of the type ic supernova sn 2007ru until around 210 days after maximum are presented .
the spectra show broad spectral features due to very high expansion velocity , normally seen in hypernovae . the photospheric velocity is higher than other normal type ic supernovae .
it is lower than sn 1998bw at @xmath0 8 days after the explosion , but is comparable at later epochs .
the light curve evolution of sn 2007ru indicates a fast rise time of 8@xmath13 days to @xmath2 band maximum and post - maximum decline more rapid than other broad - line type ic supernovae . with an absolute @xmath3 magnitude of @xmath4 ,
sn 2007ru is comparable in brightness with sn 1998bw and lies at the brighter end of the observed type ic supernovae .
the ejected mass of @xmath5ni is estimated to be @xmath6 . the fast rise and decline of the light curve and
the high expansion velocity suggest that sn 2007ru is an explosion with a high kinetic energy / ejecta mass ratio ( @xmath7 ) .
this adds to the diversity of type ic supernovae .
although the early phase spectra are most similar to those of broad - line sn 2003jd , the [ oi ] line profile in the nebular spectrum of sn 2007ru shows the singly - peaked profile , in contrast to the doubly - peaked profile in sn 2003jd .
the singly - peaked profile , together with the high luminosity and the high expansion velocity , may suggest that sn 2007ru could be an aspherical explosion viewed from the polar direction . estimated oxygen abundance 12 + log(o / h ) of @xmath08.8 indicates that sn 2007ru occurred in a region with nearly solar metallicity . |
the study of non - bps brane systems are important in string theory . some of the dynamics involved play crucial roles in the inflationary scenario in the brane world @xcite .
we shall start by calculating the @xmath0 potential per unit world volume @xmath4 , which is complex .
we first examine the potential @xmath5 between two @xmath3-branes at an angle @xmath6 and separation @xmath7 .
( the advantage of studying @xmath8 first is clear : the underlying physics is much easier to keep track , since the open string spectrum , in particular the tachyon mode , depends on both @xmath6 and @xmath7 . for @xmath9
, @xmath10 is the potential between the @xmath0 pair . )
the lightest open string mode has a mass that depends on @xmath6 and @xmath7 .
it is tachyonic for small @xmath7 .
the real part of the open string one - loop contribution simply yields the closed string exchange potential of the @xmath0 pair separated at a distance @xmath7 .
the large @xmath7 behavior is simply coulombic .
however , at small @xmath7 , this coulombic behavior is truncated . because the supersymmetry breaking is soft , a mass - level `` supersymmetric '' organization of the open string modes @xcite yields a finite effective potential , as shown in fig 1 . for @xmath4 , the open string spectrum level crossing and the unbounded growth in the `` soft '' supersymmetry breaking for massive modes
is qualitatively different from the small @xmath6 case @xcite .
we shall see that the scalar modes @xmath11 ( there are @xmath12 of them ) are also tachyonic when the @xmath3-brane is on top of the @xmath13-brane .
their tachyonic property is due to quantum effects ( radiative correction ) , a feature first seen in ref@xcite for small @xmath6 . in the open string one - loop channel ,
since only the tachyon mode contribution has an imaginary part at the one - loop level , the evaluation of @xmath14 is completely field theoretic @xcite .
a simple generalization from 4 spacetime dimensions to @xmath15 dimensions yields the imaginary part of @xmath4 , as shown in fig . 1 .
[ imagvww ] i m v(y ) = ( ) ^(p+1)/2 where ( @xmath16 is the open string regge slope ) , [ tachyonmass ] m_tachyon^2 = - its physical meaning has been discussed extensively @xcite . in the dual channel ,
that is , the closed string channel , we should obtain exactly the same result . for large separation @xmath7 , one obtains the well - known coulombic potential @xcite , due to the ns - ns and rr exchanges , which correspond to the attractive gravitational and rr forces . as we go to short distances , the massive closed string modes that are yukawa - suppressed start contributing to the potential @xmath4 . as @xmath17
, we see that ( naively ) the potential diverges .
this apparent divergence appears when the hagedorn - like degeneracy overcomes the yukawa suppression .
in fact , this happens precisely when the lightest open string mode becomes tachyonic .
a regularization ( that is , an analytic continuation of the integral ) renders the result finite , but with an imaginary part , as expected , precisely reproducing the result ( [ imagvww ] ) obtained in the open string channel .
this is clearly related to the decay of the @xmath0-brane pair . when applied to a non - bps @xmath3-brane
, this same regularization approach yields a finite @xmath18 closed strings @xmath19 .
the decay of an unstable non - bps d - brane to closed strings has been studied extensively@xcite . adapting the above calculation to this case
, we consider @xmath20 where @xmath21 is the closed string propagator and the sum over the closed string spectrum is implied .
this is equal to @xmath22 , which is finite with an imaginary part .
( the factor of @xmath23 is because the open string spectrum of the non - bps @xmath3-brane is half of those that stretched between the @xmath0-brane pair at @xmath24 ) . ) using optical theorem ( i.e. , perturbative unitarity ) , as shown in fig .
2 , we interpret this as the decay width @xmath25 of the non - bps @xmath3-brane to on - shell bosonic closed string modes : = v_p i m v(0 ) where @xmath26 is the @xmath3-brane world volume . as expected , @xmath25 is 0@xmath27 order in the string coupling .
here , the finite imaginary part appears due to the hagedorn - like degeneracy of the massive closed string modes , and the analytic continuation moves the closed string modes from off - shell to on - shell .
the decay first goes to very massive , non - relativistic on - shell bosonic closed string modes @xcite , with transverse momentum very small compared to the mass @xmath28 ( @xmath29 ) .
these non - relativistic massive modes then decay to relativistic light closed string modes , both bosonic and fermionic .
intuitively , this is quite reasonable , since the imaginary part of the effective potential should be related to some decay width .
the above decay width is obtained perturbatively so it is expected to go to perturbative modes , i.e. , closed and/or open string modes . in the decay of a non - bps @xmath3-brane , there is no brane left after the decay , so no open string modes can be present . since only closed string modes are present at the end of the decay , the decay width @xmath25 should be a measure of the decay to closed string modes .
( lower dimensional branes are solitonic , which are non - perturbative , so their production during the decay should not be included in @xmath25 . ) although @xmath30 is clearly related to the decay of the @xmath0 pair , it is harder to interpret @xmath31 as its decay to closed strings .
however , since closed string modes are the only perturbative modes left after the annihilation of the @xmath0 pair , its value may provide an estimate of the actual decay width to closed string modes .
this result has applications to brane inflation @xcite .
we shall first obtain the @xmath0 potential in the open string channel by extrapolating the potential for branes - at - angle .
next we see how the same finite complex @xmath0 potential emerges in the closed string channel .
we then relate its imagainary part at zero separation to the decay of a non - bps-@xmath3-brane to closed strings .
let us consider the potential @xmath33 per unit volume between a parallel @xmath0 pair separated by a distance @xmath7 , where the @xmath3-branes are bps with respect to each other .
let us write @xmath34 , where the @xmath35 is @xmath3-brane tension .
we shall consider @xmath36 . in the closed string channel , @xmath4
is given by @xcite v(y ) = -^_0()^-(p+1)/2 e^- ( ) ^416 _ m=1^ [ closedv ] where @xmath37 . at large @xmath7 , the above integral
is dominated by large @xmath38 ( which corresponds to long cylinder ) , so v(y ) - ( ( 7-p)/2 ) where @xmath39 . for @xmath40
, @xmath4 vanishes as @xmath41 .
@xmath42 for @xmath43 .
this is simply the attractive ns - ns ( gravitational ) plus rr interaction between the branes . at short distances ( small @xmath38 ) , this expression naively diverges ,
although @xmath4 is expected to approach a finite value even for @xmath24 .
let us first calculate @xmath4 in the open string channel , which yields a finite @xmath4 for all @xmath7 .
then we come back and show how to obtain the same result in the closed string channel . using the poisson resummation formula , the above @xmath4 for the @xmath0 system
may be rewritten as the open string one - loop amplitude : [ openv ] & & v(y ) = -^_0(8 ^2t)^-(p+1)/2 e^-t y^2/2 z(,t ) + & & z ( , t ) = _
m=1^ + & & = - 8 + 36 q^1/2 -128q + 402 q^3/2 - 1152 q^2 + 3064 q^5/2- 7680q^3 + & & + 18351 q^7/2- 42112 q^4 + 93300 q^9/2- 200448 q^5 + ... where @xmath44 .
this sum is over the open string modes . for large @xmath7
, this sum converges rapidly to give @xmath4 .
notice that as @xmath7 decreases , the mass of the lightest open string mode is given in ( [ tachyonmass ] ) , which becomes tachyonic for small values of @xmath7 . in the presence of the tachyon ,
the above sum apparently becomes ill - defined , that is , the integral diverges .
the tachyonic mode contribution to @xmath4 is [ openvt ] & & v(y ) = -^_0(8 ^2t)^-(p+1)/2 ( - 2 t ( y^2/4 ^ 2- 1/2 ) ) which is divergent .
this integral can be regularized by analytic continuation @xcite . to see how the divergence can be regularized using analytic continuation ,
consider a slightly different integral : _ ^x^- e^(a+ i)x = ( a ) ^ _
-a- i^ x^- e^-x this integral has a branch cut along the negative @xmath45 axis .
the @xmath46 prescription tells us to integrate under the branch cut .
the integral is now finite .
the divergence has been removed in favor of an imaginary part which does not depend on the value of @xmath47 .
the divergence is not an infinity of the theory , but rather an indication of the amplitude becoming complex .
the result of the analytic continuation is : [ imagin ] i m _
0^x^- e^ax= a^ for positive @xmath48 . using ( [ imagin ] ) , this gives , for @xmath49 , the imaginary part of @xmath4 : [ imav ] i m v(y ) = ( ) ^(p+1)/2 this @xmath50 is shown in fig .
since only the tachyon mode contributes to the imaginary part of @xmath4 , we can also evaluate @xmath50 using the standard quantum field theory method .
its contribution can be calculated via the coleman - weinberg effective potential @xcite where the one loop vacuum - to - vacuum amplitude for a point particle of mass m in @xmath15 dimensions is given by : z(m^2 ) = iv_p+1_0^ e^-(k^2+m^2)l/2 = _ 0^ inserting the open string tachyon mass @xmath51 we see that the above integral diverges . after a proper regularization using analytic continuation
, we get the imaginary part of the energy density im(e ) as ( [ imav ] ) . to calculate the real part of @xmath4 one has to include the whole tower of open string modes .
the result depends on how the oscillating terms are grouped .
the particular way of grouping the terms should be dictated by the soft supersymmetry breaking , as suggested by garcia - bellido , rabadan and zamora @xcite .
they applied it to the branes - at - small - angle case , where @xmath520 corresponds to two parallel bpd @xmath3-branes while @xmath9 corresponds to the @xmath0-brane pair . for the @xmath0 system ,
supersymmetry breaking becomes large and level - crossings happen .
we show that , despite these , the convergence remains intact .
when the two branes are parallel there is no potential between them because of the supersymmetry .
each mass level contains a set of the supermultiplets .
the contribution to the potential @xmath4 from the open string bosons is exactly cancelled by the contribution from the open string fermions , mass level by mass level .
we keep this mass level grouping .
the way we identify the grouping as the angle @xmath6 between the branes increases from zero to @xmath53 is by following the spectral flow , that is , the splitting of open string modes at each level due to the soft supersymmetric breaking .
this splitting leads to unequal contributions from the bosons and the fermions , resulting in a finite potential between the brane and the antibrane .
let us first review the branes - at - an - angle case . for simplicity , let us consider the case of two @xmath544-branes at an angle @xmath6 . the potential for two @xmath544-branes at an angle @xmath6 in the open string one - loop channel
is given by , v(y , ) = -v_4^_0(8 ^2 ^t)^-2 e^- y^2z(,t ) , + z(,t ) = for @xmath520 , the @xmath3-brane pair is supersymmetric and @xmath550 .
the splitting of states is shown in the figure 3 . to calculate
the splitting one observes that the @xmath56 sector zero point energy has an angle dependence given by @xmath57 .
the r sector zero point energy remains unaffected by the angle between the branes .
considering @xmath58 branes in the light cone gauge , we take the branes to make an angle @xmath6 in the @xmath59 plane , then the creation operators in these two dimensions are given by @xmath60 , where @xmath61 , and for the rest of the six dimensions the creation operators are @xmath62 , where @xmath63 .
similarly the creation operators in the @xmath64 sector are given by @xmath65 for the @xmath66 and the @xmath67 dimension and by @xmath68 for the other six dimensions , where @xmath69 .
note that since we have only six zero modes for the @xmath64 sector for nonzero @xmath6 , therefore , the fermions will be in @xmath70-dimensional representation of the dirac algebra .
the above consideration suggest the following grouping of the terms in the partition function : z(t)= = _ m=1^ the long distance behavior is determined by the @xmath71 limit of the partition function . in this limit , @xmath72 , and
the potential becomes : v(y , ) = - the short distance behavior is determined by the @xmath73 limit .
the open string spectrum splits into copies of the ( broken ) supermultiplet .
the susy breaking is due to the `` expectation value '' @xmath6 , which is a soft susy breaking , so that , for any @xmath7 , _ i(-1)^fm_i^2n=0 , n=1,2,3 where @xmath74 runs over the spectrum in each `` softly broken '' supermultiplet ( corresponding to a @xmath75 supermultiplet ) .
the lightest open string mode has mass m^2_tachyon= - in fig 3 , the open string modes are grouped into integer mass levels at @xmath520 , each of which contains a set of ( broken ) supermultiplets .
following the spectral flow and keeping this grouping for non - zero @xmath6 , @xmath76 can be written as : z ( , t ) & = & ( _ n=0^z^2n)_m=1^ + & = & ( _ n=0^z^2n ) [ 1 + q + q^2 + ... ] where @xmath77 and @xmath78 . here ,
for each mass level , the sum of the landau levels in @xmath79 .... after the integration over @xmath80 , we obtain v_1-loop= _ 0^ _ i(-1)^f_i e^-2^tm_i^2 + = _ i(-1)^f_im_i^4 ( 2^m_i^2 ) notice that @xmath81 is finite , since the quadratic divergence is absent for each multiplet .
now let us go back to the @xmath3-@xmath13 case .
the above expression diverges in this limit .
this divergence is well understood : it is due to the appearance of an additional dimension in the overlapping world volume of the @xmath3-branes . to remove this divergence
, we make the substitution : _
11(it/ , it ) il(8 ^2^t)^-1/2 e^-t(it)^-3 this substitution gives us the expected partition function for the brane - antibrane case : z ( , t ) z ( , t ) = ( q)^1/2 + = _ m=1^this gives the @xmath3-@xmath13 potential @xmath4 in the open string channel ( [ openv ] ) . to be specific ,
let us now focus on the @xmath82 case .
the long distance behavior is found by taking the @xmath71 limit of the partition function , with @xmath83 , we have v(y , ) = -2.56 10 ^ -4()^3/2 to see the short distance behavior , retain the mutiplet grouping structure at @xmath84 , we write the partition function as a function of @xmath85 and @xmath86 : & & z ( , t)= _ m=1^ + & & = [ 1 -4 q + 6 q^2 + ... ] we use this to evaluate the following integral which we shall need : v_1-loop= _ 0^ _ i(-1)^f_ie^-2tm_i^2 + = _ i(-1)^f_i ( 2^m_i^2)^5/2 where we have used the soft susy breaking condition to evaluate this integral .
now we can write down the potential terms for different string levels ( taking the relevant power of @xmath85 from the above expansion : order @xmath87 in @xmath88 : ( y^2- ^)^5/2 - 4(y^2)^5/2 + 6(y^2 + ^)^5/2 - 4(y^2 + 2^)^5/2 + ( y^2 + 3^)^5/2 order @xmath89 in @xmath88 : -4(1 + y^2- 2^)^5/2 + 24(1 + y^2- ^)^5/2 -60(1 + y^2)^5/2 + 80 + ( 1 + y^2 + ^)^5/2 -60(1 + y^2 + 2^)^5/2 + 24(1 + y^2 + 3^)^5/2 - 4(1 + y^2 + 4^)^5/2order @xmath90 in @xmath88 : ( 2 + y^2- 3^)^5/2 -10(2 + y^2- 2^)^5/2 + 40(2 + y^2- ^)^5/2 + -86(2 + y^2)^5/2 + 110(2 + y^2 + ^)^5/2 -86(2 + y^2 + 2^)^5/2 + 40(2 + y^2 + 3^)^5/2 + -10(2 + y^2 + 4^)^5/2 + ( 2 + y^2 + 5^)^5/2 and so on , where @xmath91 stands for @xmath92 and @xmath93 stands for @xmath94 .
this yields the potential in fig .
1 . the imaginary part comes only from the lightest open string mode when it becomes tachyonic .
the real part converges quite rapidly , we need to keep only up to @xmath95 to get within 1@xmath96 accuracy .
the behavior of the potential for short distances can be seen from fig .
an interesting result first observed in ref@xcite is the emergence of new tachyonic modes in the @xmath11-directions .
there is a dip in the potential close to the origin and the second derivative of the potential at the origin is negative .
this yields a tachyonic mass for @xmath11 as a function of @xmath6 .
m^2(y ) & = & 3.36 10 ^ -4 / + m^2(y ) & = & 4.51 10 ^ -5 = the appearance of the @xmath11 tachyons happens for generic @xmath97 . in the open string classical limit , the brane separation @xmath11 are @xmath2 moduli , so this tachyonic mass is a consequence of the one - loop open string contribution
; that is , radiative corrections analogous to the coleman - weinberg mechanism . this feature first appears in the branes at small @xmath6 system @xcite .
we see that it persists for all values of @xmath98 .
since this @xmath11 tachyon mass is much smaller than the lightest open string tachyonic mode , we expect the latter to dominate the brane dynamics at short distances .
the next step is to evaluate the same brane - antibrane potential @xmath4 from the closed - string perspective .
the @xmath3-brane is a solitonic object and it can emit / absorb closed string modes with arbitrary transverse momentum @xmath99 .
v(y ) = -_j ( 4 ^2 ) ^4-p ^_0ds ( -s ( ^2 + m^2_j)/4 + i * * ) [ closedvz ] + = -^_0()^-(p+1)/2 e^- y^2/(s^ ) ( ) ^4 z(s ) + z(s ) = 16 _ m=1^ = _ 0^a(n)w^n where @xmath100 and @xmath37 .
this is ( [ closedv ] ) , where [ acontour ] a(n)= dw = dz e^nz z(s ) we see that only bosonic closed string ( ns - ns and rr ) modes are included . long distance behavior is governed by the light ( @xmath101 ) closed string modes which dominate the s - channel contribution to the potential in the @xmath102 limit .
the result is finite and goes as @xmath103 .
next , we consider the small @xmath7 ( small @xmath38 and large @xmath104 ) behavior .
notice that @xmath105 grows monotonically with @xmath104 .
so , unlike the open - string calculation , where the coefficients in the expansion for @xmath106 oscillate in sign so that they can be grouped in a multiplet structure leading to a convergent sum , here the sum grows monotonically and will lead to a divergence if the degeneracy factor dominates the yukawa suppressing factor @xmath107 .
to get an idea of the behavior , we need the large @xmath104 behavior of @xmath105 . as @xmath108 ( see appendix ) : [ zas ] z(s ) ( ) ^4 ( ) and @xmath105 is obtained via the saddle - point approximation with the asymptotic form of @xmath106 ( [ zas ] ) in ( [ acontour ] ) , [ an ] a(n ) ( 2n)^-11/4 e^(8 ^ 2n)^1/2 ( 1 + o ( ) ) next , we perform the s - integration in @xmath4 to obtain , in the large @xmath104 approximation : v(y ) & ~ & - y^p-7 _
n a(n ) e^-2 y ( 1 + o(y ) + .. ) + & ~ & - y^p-7 _
n n^-11/4 \{4(1/-
y/2 ) } the mass of the closed string mode being exchanged is @xmath109 so that , for large values of @xmath7 , the yukawa suppression factor @xmath110 leads to a finite potential .
however , for @xmath111 , the exponent blows up for large n , so the potential @xmath4 apparently diverges .
that is , the exponentially large ( hagedorn - like ) number of massive closed string modes contributing to the potential overcomes the yukawa suppression . comparing with the open string tachyon mass formula ( [ tachyonmass ] ) , we see that the divergence appears exactly at the value of @xmath7 where the lightest open string mode becomes tachyonic . to emphasize this point ,
let us go back to @xmath4 ( [ closedv ] ) .
in the large @xmath104 approximation , we replace @xmath105 by its asymptotic value and replace the sum over @xmath104 by an integration over a continuous @xmath112 variable , where @xmath113 and @xmath114 : [ closedvn ] v(y ) & & -^_0 _ n ( 8 ^2t)^-(p+1)/2 t^4(2n)^-11/4 \{2 - t - } + & & -^_0 ^_0 ( 8 ^2t)^-(p+1)/2 t^4 u^-7/2 \{-(u - t)^2/t - 2 m^2_t t } where @xmath115 is precisely the mass squared of the lightest open string mode given in ( [ tachyonmass ] ) . the integral over @xmath112 is gaussian - like so it is convergent , while the integral over @xmath80 depends on @xmath7 . for small @xmath7 , @xmath116 is tachyonic and
the integral over @xmath80 diverges .
now we are ready to get @xmath14 in the closed string channel .
since the open string channel calculation of the potential is finite everywhere , we expect the same for the closed string calculation of @xmath4 .
the appearance of this divergence at small @xmath7 signals the appearance of an imaginary part of the potential .
we may use eq([imagin ] ) to obtain the imaginary part of @xmath4 in ( [ closedvn ] ) when @xmath117 .
this gives a crude value of @xmath14 .
it turns out that the correction ( i.e. , subleading ) terms in @xmath105 are important in the evaluation of @xmath14 , and the convergence seems slow .
in fact , it is easier to evaluate @xmath14 simply by going back to @xmath106 . putting the asymptotic ( @xmath118 ) form of @xmath106 ( [ zas ] ) into @xmath4 ( [ closedv ] ) and using ( [ imagin ] ) ,
we obtain precisely the @xmath119 given in ( [ imav ] ) .
as we shall see , corrections to the @xmath118 form of @xmath106 do not contribute to @xmath119 . while the imaginary part seen in the open string channel
is due to the single tachyonic mode , the imaginary part in the closed string channel is due to the cumulative effect of the hagedorn - like degeneracy of the massive closed string states .
these two effects are dual to each other .
this point will become even more explicit below .
we still have to see how the real part of the potential @xmath120 emerges in the closed string channel .
this is carried out by by an exact correspondence between the open string calculation and the closed string calculation . to demonstrate this exact correspondence
, we shall start from the closed string channel and show that taking into account all orders of correction to the asymptotic expression for @xmath106 leads back to the open string channel expression .
furthermore , the asymptotic expression for @xmath106 ( [ zas ] ) corresponds precisely to the open string tachyon mode .
this shall also answer the question as to what happens to the supersymmetric grouping of the terms present in the open string partition function ( that we make use of through the soft susy breaking condition to find a finite sum in the open string channel ) when we go to the closed string channel .
we follow the calculation of the asymptotic form for @xmath106 making use of the hardy - ramanujan formula | appendix| but this time by keeping track of the correction terms that would arise in the @xmath121 region .
the result is : z(s ) = 16 e^2 ^ 2/s()^4 e^ where @xmath122 contains the correction terms , = -de^-2 ^ 2/s+ e^-4 ^ 2/s- e^-6 ^ 2/s+e^-8 ^ 2/s + ..... expanding the @xmath123 term for @xmath124 , we get : z(s ) = ( ) ^4e^ -8 + 36e^--128e^+ 402e^-6- ... substituting this into the closed string channel @xmath4 ( [ closedv ] ) , and making a change of variable @xmath125 , we recover the potential @xmath4 in the open string channel ( [ openv ] ) : v(y ) = -_0^(8 ^ 2t)^-5/2 e^ e^t -8 + 36e^-t-128e^-2t+ 402e^-3 t- ... one clearly sees that the imaginary part of the potential using the asymptotic form of @xmath106 in the closed string channel is exactly equivalent to the tachyon contribution in the open string channel .
the massless and massive level contributions to the potential in the open string channel are exactly equal to the contribution coming from the corrections to this asymptotic form
. they do not contribute to @xmath119 .
one could go on and start calculating the real part of the potential in the closed string channel by taking the term by term contribution coming from the subleading terms to @xmath106 , grouping them as suggested by the soft susy breaking in the open string channel .
the decay rate per unit world volume of a non bps @xmath3-brane to closed strings can be written as the square of the amplitude @xmath126 . via the optical theorem
, we expect this to be given by @xmath127 , as shown in fig .
@xmath21 is the closed string propagator given by @xmath128 , @xmath129 and @xmath130 being the virasoro generators , and @xmath131 .
let us consider an unstable non - bps-@xmath3-brane in type ii string theory
. for a non - bps-@xmath3-brane , both ends of an open string end in the same brane , so @xmath24 . for the same reason ,
the open string spectrum in a non - bps-@xmath3-brane has only one real tachyon mode , instead of a complex tachyon field , and it has only half of the spectrum of open strings that stretch between the branes in the @xmath3-@xmath13 system at @xmath24 .
so the analysis of the one - loop open string effective potential in this case is identical to that for @xmath22 , and we obtain @xmath132 as given in ( [ imav ] ) with @xmath133 . performing the @xmath38-integration in ( [ closedvz ] ) at @xmath24 , we obtain [ vprop ] < dp| |dp > = = -_j where the sum is over closed string modes .
recall that a(dp dp ) _ j _ j i(k^2-m_j^2 ) |f_j|^2 [ factor ] where , ignoring momenta @xmath134 parallel to the brane , @xmath135 .
we see that @xmath136 is a constant independent of the closed string mode @xmath137 , and that the imaginary part comes only from on - shell closed string modes . before analytic continuation , @xmath22 is real and infinite , with @xmath138 . after analytic continuation , @xmath22 becomes finite with an imaginary part .
a comparison of ( [ vprop ] ) , ( [ closedvz ] ) and ( [ factor ] ) suggests that the analytic continuation moves the propagators on - shell , that is , @xmath139 is due to the sum over the on - shell poles .
since the divergence comes from large @xmath104 , it further suggests that most of @xmath139 comes from massive on - mass - shell closed string modes . for the massive closed string states , the poles get so dense for large @xmath104 that they are well represented by a branch cut .
our prescription for regularizing the divergence and isolating the imaginary part is equivalent to applying the @xmath140 prescription of the analytic continuation to all asymptotically massive closed string modes and moving them on - shell .
so @xmath141 should be interpreted as the decay width of a non - bps @xmath3-brane into on - mass - shell closed string modes .
we may see this a little more clearly by separating the regularization procedure into 2 steps : first go on - shell and then perform the analytic continuation . ignoring @xmath134 , we may restore the @xmath142 integral and rewrite ( [ vprop ] ) as ~^dndk_^9-pdk_0 ( k^2 - 4n/ ) ( k_0 ) + dk_^9-p_0^ dk_0 where the sum over @xmath137 is replaced by an integration over @xmath100 , with measure @xmath105 . carrying out the @xmath104-integration with the delta function keeps only the on - mass - shell states . using the asymptotic form of @xmath105 ( [ an ] ) , we find that @xmath143 .
this integral is divergent .
after the analytic continuation of this expression , the integral becomes finite and obtains an imaginary part . presumably , the inclusion of subleading terms of @xmath105 will reproduce the correct @xmath144 . in fig .
2 , only bosonic closed string modes are involved .
they are the ns - ns and rr modes . in ( [ closedvz ] )
( with @xmath24 ) , we see that @xmath145 has a gaussian distribution , with @xmath146 . from ( [ zas ] ) and ( [ acontour ] )
, we have @xmath147 . putting them together
, we see that , for massive states , so the momenta of the massive closed string modes transverse to the brane are negligible .
conservation of momenta tangential to the brane implies that they are negligible too .
since it is the large @xmath104 behavior that brings in the imaginary part , the decay is to very massive non - relativistic closed string modes , as pointed out in ref @xcite .
these massive closed string modes will then decay to light ( both bosonic and fermionic ) closed string modes , which are expected to be relativistic .
apriori , the tachyon couples to other open string modes and the decay of a non bps @xmath3-brane corresponds to the rolling of the tachyon @xcite , so one naively expects some energy to go to open string modes .
however , no open string mode exists after the complete decay ( and the disappearance ) of the brane .
this issue is resolved in ref@xcite : one expects the ends of an open string mode to have a flux tube ( @xmath148 ) between them , so that the flux tube together with the open string form a closed string . as a result , only closed string modes are produced .
the estimate of @xmath25 is for the tachyon at or close to the top of the potential ( at @xmath149 ) .
the time scale @xmath150 of the tachyon rolling is around @xmath151 . in the compactified case
where the world volume @xmath152 , @xmath150 is comparable to or larger than the inverse of @xmath25 . in this case , the above estimate of @xmath25 should be valid . for @xmath150
much smaller than @xmath153 , we may expect the decay to start with the tachyon rolling , which goes to tachyon matter @xcite , which then decays to relativistic closed strings . in this case , the above estimate may not be applicable .
@xmath3-branes in the bosonic string theory may also be considered : = _ 0^ds e^s _ m=1^ ( 1 - e^-ms)^-24 + _ 0^ ds e^s z_b(s ) for @xmath118 , @xmath154 so the degeneracy @xmath155 .
this yields for @xmath114 and @xmath156 ( ignoring powers of @xmath80 and @xmath157 ) , ~_0^ dt dv ( -2 ( v - t)^2/t + 2 t ) completely analogous to ( [ closedvn ] ) , the last term in the exponent is identified as @xmath158 with the tachyon mass @xmath159 .
this is the origin of the apparent divergence in the integration over @xmath80 . using the @xmath108 form of @xmath160 in @xmath20 ,
we obtain , after analytic continuation , the expression ( [ imagvww ] ) for @xmath161 , with @xmath162 and @xmath163 .
following the earlier analysis , we see that subleading terms of @xmath160 do not contribute to @xmath164 .
it is suggestive to take @xmath165 ( [ imav ] ) to be the decay width of the @xmath3-@xmath13-brane pair to closed strings , since perturbatively , only closed strings are available after their annihilation
. it will be important to understand this issue better .
for branes at an angle @xmath6 , the branes recombine as @xmath166 .
perturbatively , we expect the decay to release energy to both open and closed strings .
it is interesting to compare this result to that in quantum field theory . in quantum field theory ,
the propagator of a single field at the tree level has only @xmath47 finction as its imaginary part .
here the imaginary part appears classically due to the hagedorn - like degeneracy of the closed string spectrum .
however , this is a perturbative quantum effect in the open string channel .
we thank nick jones , finn larsen , hong liu and raul rabadan for valuable discussions .
this material is based upon work supported by the national science foundation under grant no .
let @xmath167 be given by : ( s ) = _ m=1^ = _ 0^ ( n)w^n where @xmath168 . using complex variables , we can write : @xmath169 , where ( n)= dw where the contour is a circle around the origin . to find the asymptotic form of @xmath170 , we write : ( s)= d f(2s ) - 2d f(s ) where @xmath171 can be expressed in the mellin representation : f(s ) = _ n=1^(1 - e^-ns ) = - _ re(z)=cdz(z)(1+z)(z)s^-z , where @xmath172 , and @xmath173 is the riemann zeta - function .
the integrand has a first - order pole at @xmath174 and a second - order pole at @xmath175 .
shifting the line of integration from @xmath176 to @xmath177 we arrive at : f(s ) = - - ( ) - _ re(z ) = c^dz(z)(1+z)(z)s^-z . letting @xmath178 and using the identities : & & ( z)= 2^z^z-1()(1-z ) ( 1-z ) , + & & ( z)(1-z)= , one obtains : f(s ) = - - ( ) + _ re(z)=c^dz(z ) ( 1+z)(z)()^-z , where @xmath179 .
finally we move the path of integration to the right side of the first - order pole at @xmath180 and get the hardy - ramanujan formula : f(s ) = - ( ) + + f ( ) in the limit @xmath181 , and one obtains the asymptotic behavior of @xmath182 as : ( s ) ( ) ^ e^ using this result and carrying out a saddle point evaluation of the contour integral , we get : [ a2 ] ( n)()^(d+1)/4n^-(d+3)/4 e^ ( 1 - + ... ) for general @xmath38 , ( s)= + ( ) -de^-2 ^ 2/s+ e^-4 ^ 2/s- e^-6 ^ 2/s+e^-8 ^ 2/s + ..... this leads to the following asymptotic expansion for @xmath182 : [ a3 ] ( s)= ( ) ^ e^ g. dvali and s .- h .
h. tye , _ phys .
lett . _ * b450 * ( 1999 ) 72 , hep - ph/9812483 ; + c. p. burgess , m. majumdar , d. nolte , f. quevedo , g. rajesh and r .- j .
zhang , _ jhep _ * 07 * ( 2001 ) 047 , hep - th/0105204 ; + n. jones , h. stoica , and s .- h .
h. tye , _ jhep _ * 07 * ( 2002 ) 051 , hep - th/0203163 ; + s. sarangi and s .- h .
h. tye , _ phys .
lett . _ * b536 * ( 2002 ) 185 , hep - th/0204074 .
m. kleban , a. e. lawrence and s. shenker , hep - th/0012081 ; + g. gibbons , k. hori and p.
yi , nucl . phys .
b * 596 * , 136 ( 2001 ) , hep - th/0009061 ; + m. gutperle and a. strominger , jhep * 0204 * , 018 ( 2002 ) , hep - th/0202210 ; + a. strominger , hep - th/0209090 ; + a. maloney , a. strominger and x. yin , hep - th/0302146 ; + t. okuda and s. sugimoto , nucl .
b * 647 * , 101 ( 2002 ) , hep - th/0208196 ; + f. larsen , a. naqvi and s. terashima , jhep * 0302 * , 039 ( 2003 ) , hep - th/0212248 ; + n. r. constable and f. larsen , hep - th/0305177 ; + b. chen , m. li and f. l. lin , jhep * 0211 * , 050 ( 2002 ) , hep - th/0209222 ; + n. moeller and b. zwiebach , jhep * 0210 * , 034 ( 2002 ) , hep - th/0207107 ; + m. fujita and h. hata , hep - th/0304163 ; + s. j. rey and s. sugimoto , hep - th/0303133 ; + j. l. karczmarek , h. liu , j. maldacena and a. strominger , hep - th/0306132 . | we calculate the potential for @xmath0 pair and show that the coincident @xmath0 system has @xmath1 tachyonic modes , with @xmath2 of them due to radiative corrections .
we propose that the decay width of an unstable non - bps-@xmath3-brane to closed strings is given by the imaginary part of the one - loop contribution to the effective potential of the open string tachyon mode . |
the teacup agn ( sdss j143029.88 + 133912.0 , 2masx j14302986 + 1339117 ) was originally discovered by massimo mezzoprete20topic=7539.msg62340 ] in 2007 as part of the sloan digital sky survey galaxy zoo project @xcite .
one of the campaigns during the galaxy zoo project was to find agn hosting extended emission - line regions ( eelrs ) , gaseous regions that are agn - ionized at distances of @xmath110 kpc .
spectroscopically confirmed as an eelr by @xcite , the teacup possesses a redshift of z = 0.086 to yield a distance of 344 mpc ( 1@xmath2 @xmath11.7 kpc for @xmath3 km s@xmath4 mpc@xmath4 ) .
the `` teacup '' nickname was given to describe its eelr , a spectacular `` handle '' or loop of ionized gas extending out to @xmath115 kpc northeast from the nucleus of the galaxy , a feature best seen via hst imaging ( keel et al 2013 ; also posted online ) .
eelrs offer a unique way to study agn , by allowing us to study ionization interactions between the central agn engine and exterior regions of the galaxy .
we use multiple sources of data available to us , the first being images and spectra from the sloan digital sky survey . in figure 1
, we show the sdss image from the g , r and i bandpasses combined ; in the northeast corner the ionized handle is visible as a purple loop , which can also be seen in the contour map .
the continuum emission shows a somewhat disturbed morphology , possibly from a merger or some other interaction .
figure 2 shows a spectrum of the optical center of the galaxy taken from sdss with a spectral resolving power of r @xmath5 @xmath0/@xmath6 @xmath7 2000 and a 3@xmath2 diameter fiber ; strong emission lines characteristic of agn - ionized gas are present .
figure 3 presents one of our spectra taken from lowell observatory at 5@xmath2 east from center , intersecting the loop structured eelr . to explain the color of this feature in the sdss image , g , r and i filter response curves are overplotted .
note that the g filter has been color coded to blue , r to green and i to red .
interestingly , [ o iii ] @xmath05007 falls in between the g and r bandpass filters . because the rest of the emission lines fall inside the ranges of both g and i filters while redshifted @xmath05007 , the strongest emission line , is clipped out by the bandpasses , the ionized handle of gas in the ne appears purple in figure 1 , as does ionized gas in the sw and other portions of the galaxy .
the spectrum is typical of a type 2 agn ( khachikian & weedman 1971 ) , with strong emission lines covering a broad range in ionization , a weak continuum with several stellar absorption features , and no evidence for broad ( fwhm @xmath8 1000 km s@xmath4 ) emission lines . in this paper
, we present results from long - slit spectroscopic and imaging observations from lowell , kpno , lick , sdss and vla observatories .
the data provide an exciting opportunity to study the long - term variability of agn on time scales of @xmath9 10@xmath10 years . in @xmath11[observations ] we describe our methods for data acquisition and reduction . in @xmath11[analysis ]
we explain our analysis of the data , yielding dereddend line ratios .
our observational results are presented in @xmath11[obs results ] , including kinematics of the teacup agn as well as first radio data from vla . in @xmath11[models ] we describe our process in creating photoionization models for the teacup , and review the implications of these models . in @xmath11[discussion ] we discuss our major findings and conclusions .
we obtained our primary spectroscopic observations on clear nights at lowell observatory s perkins 1.8-m reflector near flagstaff , arizona .
we obtained two dimensional long - slit spectra with the deveny `` blue '' spectrograph with a 300 line mm@xmath4 grating , a slit of 2@xmath2 width positioned north - south , and a spectral resolution of @xmath6 @xmath7 3.0 , with r ranging from 1400 - 2500 , between 4300 @xmath12 7600 .
a total of nine long - slit observations were obtained at six parallel positions relative to the galactic center to cover the 12@xmath2@xmath1312@xmath2 field containing the teacup agn .
four observations were made at optical center ; for these observations , we averaged the line ratios , and the standard deviation between these four data sets yielded our errors .
three observations were taken moving eastward of the optical center at offset positions of 2@xmath2 , 4@xmath2 and 5@xmath2 in order to obtain spectra across the teacup s ionized `` handle '' , with the last two positions overlapping by 1@xmath2 .
two more observations were made at 2@xmath2 and 4@xmath2 west of the optical center . because of differences in grating positioning in the deveny spectrograph only data from slit positions 5@xmath2 east of nucleus and nuclear center ( 0@xmath2 )
extend into the blue enough to capture [ o ii ] @xmath03727 . additionally , we obtained long - slit spectra from the 2.1-m telescope under clear conditions at kitt peak national observatory and 3.0-m shane telescope at lick observatory . at kpno
we used the goldcam spectrograph with the `` 26new '' grating which is a 600 line mm@xmath4 grating with @xmath6 @xmath7 3.3 resolution .
this 2@xmath2 wide long - slit spectrometer was given a position angle of 37 to capture the top side of the teacup s
handle feature .
data collected from the lick observatory was obtained with the kast double spectrograph , which was set to a 2@xmath2 slit width and position angle of 95 covering the bottom side of the ionized
handle . to achieve the correct wavelength coverage we used the 600/7500 and 600/4310 grisms for the red and blue channels respectively .
figure 4 shows the vertical 2@xmath2 slit positions from lowell divided into 31 2@xmath2@xmath132@xmath2 extraction bins .
overplotted are the kpno slit ( position angle 37 ) and the slit from lick observatory ( position angle of 95 ) ; see also keel et al .
lastly , data from the vla first survey @xcite were analyzed in order to gain a radio luminosity and classification of this object .
table 1 gives a full account of all optical observations including that from sdss .
we reduced our optical data two - dimensionally using iraf , as well as the noao and ctio packages , resulting in wavelength calibrated spectral images in flux units .
two dimensional reduction allowed us to extract multiple spectra in the cross - dispersion direction for each offset slit position in the dispersion direction . while the width of the extraction bin is constrained to 2@xmath2 by the slit width , we had the ability to sample our data in the cross - dispersion direction in either 1@xmath2 or 2@xmath2 bins . extracting spectra in this manner from each of our six slit positions yielded a total of 62 or 31 individual spectra in 2@xmath2@xmath131@xmath2 ( 3.4 [email protected] kpc ) or 2@xmath2@xmath132@xmath2 ( 3.4 [email protected] kpc ) bins , respectively .
spectra sampled in 2@xmath2@xmath131@xmath2 bins have a higher spatial resolution but a lower signal - noise ratio than the data sampled in 2@xmath2@xmath132@xmath2 bins .
therefore , our signal - noise requirements dictated when we could use data binned in 2@xmath2@xmath131@xmath2 or data in 2@xmath2@xmath132@xmath2 bins . to ensure the quality of our data
, we required that the fluxes of the important emission lines for each type of analysis have signal - noise ratios @xmath14 3@xmath15 . in the case of our kinematic analysis ,
the emission line of interest is [ o iii ] @xmath05007 , because it is the strongest emission - line in our spectra .
the strength of [ o iii ] @xmath05007 allowed us to use the 2@xmath2@xmath131@xmath2 bins to obtain better spatial resolution for our kinematic measurements .
for photoionization modeling , the lines of interest were the blended [ s ii ] @xmath16 6716 , 6731 doublet , to obtain accurate densities .
this feature is much weaker than the [ o iii ] @xmath05007 , so we used 2@xmath2@xmath132@xmath2 bins . in the case of the 2@xmath2@xmath132@xmath2 bins , three low signal to noise bins from the 4@xmath2 west slit were discarded leaving 28 bins with acceptable signal to noise for the majority of our analysis .
the seeing for the long slit spectroscopy was between 1 `` and 2 '' ( fwhm of the point spread function ) , so some of the kinematics measurements could have been oversampled by up to a factor of two in the spatial direction .
as the spectrum from sdss has a larger aperture , greater wavelength coverage , and better signal to noise than the rest of our observations , we matched the absolute flux as a function of wavelength from sdss to that of our central extracted bin from lowell , in order to correct for atmospheric ( including refraction ) losses .
we applied this normalization to the rest of the data for each night before analysis of emission line fluxes for each bin .
the series of bins allowed us to create a grid of spatially resolved spectra covering the entire galaxy .
we analyzed the reduced data utilizing interactive data language ( idl ) .
for each spectral bin we measured emission line fluxes for every non - blended line with a linear interpolation procedure that integrates under each line and above a baseline which connects the adjacent continua .
for the blended lines of [ n ii ] @xmath0@xmath06548 , 6583 and h@xmath17 , and the [ s ii ] @xmath0@xmath06716 , 6731 doublet , we took a template from the [ o iii ] @xmath05007 profile and reproduced it at the positions of the blended components simultaneously . by scaling the template up and down it was possible to find a scaling factor for each line relative to [ o iii ] . to determine good fits between the templates and blended lines , we subtracted the scaled components to look at the residuals , rescaling until the residuals were close to the established continuum ensuring good fits ( crenshaw & peterson 1986 ) .
all continuum - subtracted fluxes were then divided by their corresponding h@xmath18 fluxes in order to gain normalized spectral ratios .
reddening , e(b@xmath12v ) , for each 2@xmath2@xmath132@xmath2 bin was found by taking h@xmath17/h@xmath18 = 2.9 , appropriate for recombination ( osterbrock & ferland 2006 ) . to calculate the reddening across the entire spectral range for each emission line and
correct each spectrum , we used a standard galactic reddening curve described by @xcite .
as shown in @xcite , reddening curves do not differ much in this region of the spectrum and any of a number of curves will provide a reasonable correction to observed nlr line ratios .
the values of all measured h@xmath18 fluxes and e(b@xmath12v ) values are listed by position in arcseconds relative to the nucleus in table 2 .
extracted positions listed in each table are given a cartesian coordinate , where x corresponds to east ( positive ) and west ( negative ) positions relative to the nuclear center , and y corresponds to the north ( positive ) and south ( negative ) directions . following this convention ,
the nuclear center is labeled as ( 0@xmath2,0@xmath2 ) .
the resultant normalized and dereddened line ratios can be found in tables 3 and 4 .
we attribute the uncertainties in our flux measurements to a combination of photon noise ( propagated through the data reduction process ) , measurement errors ( including placement of the continuum ) , and errors in the reddening correction ( propagated from the lines used to determine the reddening ) . to constrain our errors in measurement
, we performed the same flux measurement procedure three times on each line with different reasonable continuum placements and found the standard deviation .
the leading source of error in our reddening correction is from our errors associated with the measured h@xmath18 fluxes , as the h@xmath17/h@xmath18 ratio was employed to determine e(b@xmath12v ) .
these sources of error were all summed in quadrature , resulting in our final dereddened line ratio errors . dereddened normalized fluxes were employed in several diagnostic tests using `` bpt '' diagrams as shown in figure 5 ( baldwin et al .
according to these diagrams adopted from from @xcite , our emission line ratios indicate that the teacup agn is indeed a seyfert 2 undergoing photoionization from a central ionizing source .
there is no evidence from these diagrams for a strong contribution from starbursts to the ionizing radiation .
because of the relative strength of the [ o iii ] @xmath05007 emission line we were able to employ our 2@xmath2@xmath131@xmath2 bins to map out [ o iii ] flux and kinematics , as well as include the bins located at 4@xmath2 west . to probe large scale motions of the ionized gas in this galaxy
, kinematics were determined by fitting gaussians to the [ o iii ] @xmath05007 lines ( see fischer et al . 2010 ) and determining their central wavelengths for all 62 bins spanning a 20.4 kpc @xmath13 28.9 kpc field .
these wavelengths were converted to velocity offsets from the rest frame of the galaxy , defined in this case by the velocity of our centrally extracted bin .
our data do not have high enough signal to noise to get a galactic velocity from stellar absorption lines .
the data from kpno and lick observatories provided us with additional 2-d spectra which allowed us to investigate the kinematics of the ionized loop of gas protruding from the nucleus of the object .
the fwhm([o iii ] ) was also measured for each lowell bin ; the instrumental fwhm was subtracted off these values in quadrature to obtain the intrinsic fwhm .
having mapped out the teacup with spatially resolved spectra , we were able to investigate the emission line reddening , fluxes , and kinematics in a grid over our object .
figure 6 shows a map of measured e(b@xmath12v ) reddening values per location on the left hand side .
the regions around ( 0@xmath2,2@xmath2 ) and ( @xmath122@xmath2,4@xmath2 ) show the highest levels of reddening .
the reddening map suggests a possible dust lane running southeast to northwest at pa @xmath7 @xmath1245 .
figure 6 also shows [ o iii ] @xmath05007 flux for each 2@xmath2@xmath132@xmath2 bin , measured in terms of ergs s@xmath4 @xmath19 .
the position ( 0@xmath2,0@xmath2 ) shows the largest [ o iii ] flux of 9.7@xmath1310@xmath20 ergs s@xmath4 @xmath19 at the location of the nucleus .
our [ o iii ] velocity map , presented on the left of figure 7 , shows the ne region of the ionized handle at position ( 2@xmath2,9@xmath2 ) receding with a maximum redshifted velocity of 150 km s@xmath4 relative to the center of the galaxy , while the position ( @xmath124@xmath2,@xmath122@xmath2 ) has the largest blueshifted velocity of 340 km s@xmath4 .
the velocity pattern shows a clear kinematic axis running se to nw with a low range in radial velocities , indicative of rotational motion .
considering the region sampled hosts a majority of the galaxy , the observed kinematics must be due primarily to galactic rotation .
this places the axis of rotation at pa @xmath7 @xmath1250 .
the right box in figure 7 shows measured fwhm values in terms of km s@xmath4 , for the [ o iii ] @xmath05007 emission line . in general , the fwhm increases toward the central nucleus , which is also supporting evidence for galactic rotation .
however , the dust lane runs along the same direction as the rotational axis , which is certainly not typical for normal disk galaxies , suggesting disturbed kinematics , possibly due to a merger . additional kinematic results from both kpno 2.1-m and
lick shane 3-m telescopes agree with this interpretation .
as these observations were made at different slit position angles , we created a procedure that utilized our grid of lowell observed velocities to simulate an artificial slit at the kpno / lick observation position angles and then extracted the pseudo - slit kinematics .
figure 8 displays radial velocity curves from kpno at 37 and lick at 95 as well as an overplot of our extracted velocity curves from lowell . in general , the velocities from the different data sets agree extremely well , and the kpno and lick observations allow us to trace the radial velocities out to greater distances .
the radial velocity curve from the lick data turns over from a redshift to a blueshift at around 22@xmath2 northeast of the nucleus .
this suggests once again that there are other physical phenomena affecting the kinematics at large distances in addition to mere galactic rotation .
this turnover is not seen in the velocity map compiled from lowell , because at 22@xmath2 away from the center , it is outside the scope of our lowell observations .
a slight bump in the radial velocity curves at the location of the teacup handle suggests an additional kinematic component at this location , possibly outflow , even though the magnitude is relatively low ( @xmath150 km s@xmath4 ) .
we also investigate the radio properties of the teacup .
figure 9 shows a 20 cm radio image from the first survey . on the right of figure 9
is the sdss g - band image .
the radio emission has a flux of [email protected] mjy ( 4.2@xmath13@xmath22 w hz@xmath4 ) and extends along the direction of the
the radio image was deconvolved into two gaussian components using the routine jmfit in aips
. the stronger component , with a flux of [email protected] mjy corresponds to the nucleus .
the second component , at 6.1@xmath2 from the nucleus , along pa@xmath23 , has a flux of [email protected] mjy .
there is a correspondence between the direction of the radio emission and the ionized gas in the handle .
the radio luminosity of the teacup is only 4.2@xmath1310@xmath22 watts hz@xmath4 , which is only slightly above the radio - loud dividing line of 1@xmath1310@xmath22 watts hz@xmath4 ( best et al . 2005 ) , making the teacup a radio - intermediate agn .
to gain physical insight , we generated photoionization models for the teacup agn with cloudy version 8.00 ( see ferland et al .
1998 ) , which replicates the observed line ratios by modeling a central ionizing source acting upon a gaseous region , in this case modeled with a slab geometry at each position ( kraemer et al . 1994 ) . to satisfy signal to noise requirements , we modeled each spectrum extracted from the 2@xmath2 by 2@xmath2 regions ( excluding 4@xmath2 west data because of poor signal to noise ) . in order to accurately model the photoionizing conditions of the teacup we first had to derive an expression for the spectral energy distribution ( sed ) .
our initial strategy followed from @xcite , looking at the observed ratio of he ii @xmath04686 relative to h@xmath18 .
we determined the ionizing uv power law index @xmath17 from @xmath24 where r is the ratio of photons ( @xmath25 ) which are ionizing h@xmath26 and he@xmath27 .
r can be determined from relative densities ( @xmath28 ) , intensities ( @xmath29 ) , and volume emissivities ( @xmath30 ) of the two lines in question , along with the dereddened ratio : @xmath31 using this method for each extracted spectrum we arrived at an initial average index of @xmath17 = @xmath121.7 . however , upon inspection of our initial models with this spectral index we found that we were predicting 30@xmath32 more he ii / h@xmath18 than we were observing .
for this reason we decreased @xmath17 to @xmath122.0 which resulted in our models matching he ii / h@xmath18 to within @xmath2110 % .
we described our final sed as a piecewise function , similar to that used by our group for other agn ( kraemer & crenshaw 2000 ) : @xmath33 @xmath34 @xmath35 previous work by @xcite has demonstrated solar abundances are usually sufficient to accurately model the physical conditions in seyfert 2 narrow - line regions .
based on this experience solar abundances were employed with and without dust depletion and we found no appreciable difference . because there is no direct evidence for dust depletions or elemental peculiarities we assumed solar elemental abundances @xcite without dust depletion .
the numerical values relative to hydrogen for these abundances are : he = 0.10 , c = 2.45@xmath1310@xmath36 , n = 6.03@xmath1310@xmath37 , o = 4.57@xmath1310@xmath36 , ne = 6.92@xmath1310@xmath37 , na = 2.09@xmath1310@xmath38 , mg = 3.39@xmath1310@xmath37 , al = 3.09@xmath1310@xmath38 , si = 3.24@xmath1310@xmath37 , s = 1.38@xmath1310@xmath37 , ar = 2.51@xmath1310@xmath38 , ca = 2.24@xmath1310@xmath38 , fe = 2.82@xmath1310@xmath37 , and ni = 1.78@xmath1310@xmath38 . by using the ratio of [ s ii ] @xmath16 6716 to 6731 to estimate the density @xmath39 ( osterbrock & ferland 2006 ) , our last free parameter is the unitless ionization parameter , u , described by @xcite : @xmath40 where @xmath41 is the number of ionizing photons emitted by the source per second , r is the distance from the central source to the ionized cloud and @xmath39 is the hydrogen density of the gas .
we assumed that r is approximately the projected distance from each extracted bin to the ionizing source , in this case simplified to be bin ( 0@xmath2,0@xmath2 ) .
our initial models consisted of only a single component , with sed and abundances set as described above . for each extracted position
, @xmath39 was set from the ratio of [ s ii ] , leaving u and column depth @xmath42 as free parameters .
values of u were then picked based on specific line intensities such as [ o iii ] @xmath05007 , [ n ii ] @xmath06584 and [ s ii ] @xmath06731 given by @xcite .
initial values for @xmath42 were chosen to be @xmath110@xmath43 @xmath19 , similar to values previously used by kraemer et al .
our initial values of u and @xmath42 were then varied until our models fit our data to within a factor of two .
single component models were sufficient to replicate most of the observed emission lines .
however these models were unable to match the observed intensity of [ o ii ] @xmath03727 within an acceptable error defined as the ratio of modeled line intensity divided by dereddened observed line intensity , with a good fit having a ratio @xmath9 2 and a perfect fit with a ratio = 1 . as our best single component models underpredicted the [ o ii ] intensity by a factor of three or greater , we included a second component into our cloudy models .
our two component models consisted of a low ionization component ( log(u ) @xmath9 @xmath123.0 ) and a high ionization component ( log(u ) @xmath8 @xmath123.0 ) .
the low ionization component contributes the majority of the emission lines including [ o ii ] @xmath03727 , [ o i ] @xmath06300 , [ n ii ] @xmath166548 , 6584 and [ s ii ] @xmath0@xmath0 6716 , 6731 , while the high ionization component contributes the majority of the [ o iii ] @xmath05007 emission .
blending the line ratios of both the high and low ionization components yields our composite model line intensities , which we compared to our observed dereddened values .
because our spectra are spatially resolved , we had to assure that both high and low ionization components for each model were co - located with the same distance r from the source . from equation 6 , we were able to constrain each co - located set with the condition that log(u@xmath44 ) + log(@xmath45 ) = log(u@xmath46 ) + log(@xmath47 ) .
however , our two component models still underpredicted [ o ii ] @xmath03727 by a factor of @xmath13 or more in most cases , forcing us to consider other possible ionizing mechanisms , that might boost [ o ii ] .
while our bpt diagrams suggest this object is photoionized , we decided to explore the possibilities of shock ionization . to consider shock ionization we matched both our observed fwhm values and observed densities to the shock velocities and hydrogen densities used in radiative shock models , generated by allen et al .
( 2008 ) , comparing our data only to allen s models with solar abundance with the magnetic parameter ranging from b = [email protected] @xmath48 g . comparing these results to those from photoionization
, the photoionization models still yielded an overall better fit to our observed spectra , as the shock models underpredicted [ o ii ] @xmath03727 even more than the photoionization models ( factor of @xmath111 versus a factor of @xmath13 ) .
shock models fit [ o iii ] @xmath04363 marginally better , with a factor of @xmath12 too low , versus our cloudy models which were a factor of @xmath12.6 too low . for [ o iii ] @xmath05007 our average photoionization models matched within a factor of @xmath11 ( almost a perfect fit ) ; the shock models matched with an average factor of @xmath11.5 .
lastly for the [ s ii ] @xmath0@xmath06716 , 6731 lines , we modeled with cloudy to an average factor of @xmath11 , while the shock models were off by an average factor of @xmath14 .
considering the mismatches associated with the shock models , we abandoned the idea of shock ionization being the primary mode of ionization .
we consider the possibility of heating by cosmic rays .
another way to explain the excess [ o ii ] emission is radio plasma within the agn , which can act as a source of cosmic rays to heat the gas , as previously shown by ferland & mushotzky ( 1984 ) and @xcite . in order to determine the density of cosmic rays in the handle region we used our first radio measurements , described in the previous section .
the 20 cm flux in the nuclear region corresponds to 20.5 mjy , while the flux in the handle is 5.9 mjy . since the radio observations were done with a resolution of @xmath15@xmath49 , we are not able to use them to measure the size of the emitting region
. instead , we use the sloan @xmath50 band image ( figure 9 ) , assuming that the radio emission is confined inside the handle .
we find that the handle has a radius of 1.7@xmath49 , which corresponds to 2.83@xmath49 if one assumes a gaussian emitting region .
using the minimum energy equations from miley ( 1980 ) we can calculate the magnetic field inside this region and the pressure and density of relativistic particles ( @xmath51 ) .
we assume a spectral index @xmath52 , a ratio of energy in heavy particles to the energy in electrons @xmath53 and a filling factor @xmath54 .
we find b@xmath55 g , p@xmath56 dyne @xmath19 and n@xmath57 @xmath58 .
these values have a small dependence on the choice of @xmath17 , @xmath59 and @xmath60 . using the scale factor @xmath61 ( ferland & mushotzky 1984 ) , which is defined as the ratio of total heating to coulomb heating
, we calculate a density of cosmic rays of @xmath62 @xmath58 .
this value can be considered an underestimate , considering the choice of @xmath59 and @xmath60 values , so we adopted log(@xmath51)= @xmath122.5 @xmath58 .
returning to our two component cloudy models , the expected cosmic ray density of @xmath51 = 10@xmath63 @xmath58 was added to the parameters in our low ionization component . here
we make the assumption that the low ionization regions of gas have been compressed by radio plasma jets , which results in the low ionization component having a higher density than the high ionization component .
this assumption also implies that the low and high ionization components are displaced from one another with respect to the radio plasma , although they are assumed to be at roughly the same radial distances with respect to the agn .
the resulting models were able to match the observed strong intensities of [ o ii ] @xmath03727 .
however adding the cosmic rays also added a collisional ionization factor to the gas , boosting the intensities of [ n ii ] @xmath0@xmath06548 , 6584 and [ s ii ] @xmath0@xmath0 6716 , 6731 to a factor of 2 or more higher than observed . this unwanted increase in ionization due to collisions
can be moderated by applying a scaling law defined by @xcite which describes how the cosmic ray density at the ionized face of the cloud ( @xmath51 ) decreases as it propagates through the cloud .
this is described by a power law @xmath51 @xmath64 ( r@xmath65)@xmath66 , where r@xmath65 is the depth of the cloud approximated by r@xmath65 @xmath110@xmath13(n@xmath67/n@xmath67 ) and we started with power law @xmath17 = -2.5 .
we then varied values for r@xmath65 and @xmath17 until proper fits were achieved .
our condition for an acceptable fit between our cloudy models and observed spectra was to match the observed lines to within a factor of two or less .
this was achieved via two component cloudy models .
the observed strong intensity of [ o ii ] @xmath03727 proved to be a difficult feature to replicate .
however , by considering the addition of cosmic rays as a result of radio plasma within the low ionization gas we were able to increase the levels of [ o ii ] by a factor @xmath68 3 , making it possible to match the observed [ o ii ] intensities as well as the other lines .
modeled values are presented in tables 3 @xmath69 4 in parentheses under observed line intensities .
overall our modeled line intensities fit our observed data within a factor of two or less .
however , the modeled values for [ n ii ] @xmath166548 , 6583 and [ s ii ] @xmath166716 , 6731 are often a factor of two or stronger than what is observed . our modeled values for strong features such as [ o iii ] @xmath05007 are typically matched to within 90@xmath32 of the observed intensity or better .
table 5 shows the position for each spectrum , projected distance to the ionizing source ( r ) , the high ionization component values for ionization parameter ( u ) , hydrogen density ( @xmath39 ) , column depth ( @xmath42 ) , cloud depth ( r@xmath65 ) , as well as the low ionization parameter values for u , @xmath39 , @xmath42 , r@xmath65 , cosmic ray density ( @xmath51 ) , cosmic ray scaling law ( @xmath17 ) , and r@xmath65 .
the projected distance ( r ) of the central position is a geometrical approximation as there is no information on where the ionized gas is concentrated within the central resolution element . at ( 0@xmath2,0@xmath2 ) , r = 1.02 kpc , which is the average distance from the edge of a 3.4 [email protected] kpc square to its center . errors associated with model input parameters u
, @xmath42 , r@xmath65 and @xmath51 were constrained iteratively by incrementally varying their values after achieving a best fit model .
once a model no longer matched our criteria for an acceptable fit , the difference between the incremented parameter values and the best fit values yielded our errors .
errors for @xmath39 came directly from our errors in measurement for [ s ii ] @xmath0@xmath0 6716 , 6731 .
the uncertainty in projected distance r is equal to the half width of each bin , as each model s location has been approximated to the central position of each bin .
once we had acceptable models for each spatially resolved spectrum , we were able to determine @xmath41 for each 2@xmath2@xmath132@xmath2 bin from equation 6 . by integrating our piecewise sed for energies greater than 1 ev , we can derive the ratio of @xmath41 and bolometric luminosity ( @xmath70 ) .
the ratio of @xmath41/@xmath70 = 6.5@xmath1310@xmath71 photons erg@xmath72 , allowed us to convert from our calculated @xmath41 to @xmath70 seen by the gas at each position .
table 5 shows the log of the derived bolometric luminosity ( @xmath70 ) for each modeled position .
the errors in @xmath70 are propagated uncertainties in ionization parameter u , radial distance r , and hydrogen density ( @xmath39 ) . for the central region ( 0@xmath2,0@xmath2 )
, r is by far the leading source of error in @xmath70 due to the large fractional uncertainty in r for this position .
the contribution of uncertainty in @xmath70 decreases quickly for regions at greater distances of r. figure 10 gives a visual representation of @xmath70 for each position , which shows that the lowest bolometric luminosities reside at the nuclear center and surrounding regions , while the regions further out and close to the handle seem to be experiencing much higher levels of luminosity .
we found the lowest bolometric luminosity at the nuclear center ( 0@xmath2,0@xmath2 ) to be 2.3@xmath1310@xmath73 erg s@xmath4 , more than two orders of magnitude less luminous than that of region ( 2@xmath2,8@xmath2 ) , the furthest region sampled from the source , at which @xmath70 = 2.2@xmath1310@xmath74 erg s@xmath72 , @xmath114.1 kpc away from its ionizing source .
this drop in luminosity , by a factor greater than 90 over a time frame of @xmath14@xmath1310@xmath75 years , fits nicely within the time frame of 0.2@xmath122@xmath1310@xmath10 years reported by @xcite for agn possessing eelrs such as hanny s voorwerp / ic 2497 .
figure 11 plots projected distance from the nucleus in kpc versus the log of our bolometric luminosities .
note , due to the way our data were sampled , we have multiple extracted bins that share the same radial distance from the nuclear center .
therefore , luminosities at the same radial distance have been offset by @xmath210.1@xmath76 in position to see our error bars .
there is no significant differences between east , west , and center points compared to fit .
luminosity increasing with radius as shown in figures 10 and 11 suggest that the luminosity of the teacup s central agn is decreasing dramatically with time .
although @xmath70 is a function of u , r and @xmath39 , it is worth pointing out that small deviations of these variables within our observational constraints will still yield the same result of @xmath70 increasing with radial distance away from the teacup s nucleus .
we find no evidence from our observed spectra that ionization parameter drops off from the central agn within the observed region .
furthermore , the addition of cosmic rays to our models has no effect on the values used for ionization parameter . thus adding cosmic rays
does not change the increase in @xmath70 with radius as calculated from our models .
we have obtained long - slit spectra of the teacup agn in the optical range of 3500@xmath129000 resulting in spatially resolved spectra .
our kinematic results from the measured [ o iii ] @xmath05007 centroids suggest primarily galactic rotation , with the major axis of rotation at pa @xmath7 @xmath1250 .
the area in the ne corner of the galaxy where the eelr is located is red - shifted , while the sw area is blue - shifted , relative to the galactic nucleus .
however the morphology of the galaxy , the dust lane along the rotation axis , and the turnover of the radial velocity curves suggest a more complex situation , possibly due to a merger .
radial velocity plots from kpno and lick spectra show a slight bump in velocity over the area of the teacup s eelr , suggesting there could be something more than galactic rotation acting on the ionized handle .
one possibility could be that the eelr is a slow moving outflow .
the magnitude of observed velocities is quite low compared to what is normally observed in nearby agn outflows , ( crenshaw et al .
2000 ) , but this would be an outflow on much larger scales , up to 15 kpc from the nucleus . from our analysis of bpt diagrams we were able to diagnose the teacup as a type 2 agn undergoing photoionization .
using cloudy we created photoionization models for a grid of 28 spatially resolved spectra gathered from lowell observatory .
each spectrum was replicated with composite models consisting of high and low ionization components .
our models match our observed emission lines quite well ( within a factor of two or less ) .
to achieve acceptable fits to the [ o ii ] emission , we had to include cosmic rays from the radio plasma in our models .
the cosmic ray density was approximated using the radio luminosity of this object . from our cloudy models we were able to calculate values of bolometric luminosity for each position , and find that @xmath70 increases radially from the nuclear center ( 0@xmath2,0@xmath2 ) .
we interpret this finding as strong evidence that the central agn engine decreased in luminosity by a factor @xmath190 over a period @xmath146,000 years .
@xcite demonstrated luminosity fading on timescales between 0.2@xmath122@xmath1310@xmath10 years for multiple galaxy zoo objects with eelrs , including hanny s voorwerp and the teacup agn this conclusion was based on the strong ionizing flux needed to create the eelr , the lack of a bright agn , and the lack of strong ir flux that would indicate an obscured agn .
keel et al .
concluded that the nuclear ir luminosities are not enough to produce the levels of ionization observed in these object s eelrs , indicating the agn output has decreased with time .
@xcite then bolstered this claim with a detailed study of the eelr hanny s voorwerp and its host galaxy ic2497 . using space telescope imaging spectrograph ( stis ) data they found that hanny s voorwerp exhibits higher levels of ionization in the optical regime than the central agn of ic2497 . using a combination of x - ray and ir data , keel et al .
are able to derive a spectral energy distribution for the agn .
once again they find that the agn within ic2497 does not possess enough energy to account for observed levels of ionization in hanny s voorwerp .
@xcite show that the ionizing luminosity of ic2497 s agn has dropped by a factor of @xmath68 100 in the last ( 1 - 2)@xmath1310@xmath10 years . because we have measured luminosities at multiple distances from the ionizing nucleus
, we can extend the analysis to study changes in luminosity over radius and thus time .
our results show a continuous drop in luminosity for the teacup agn within a timescale consistent with previous results by keel et al .
thus , we have additional evidence showing results that agree with previous findings that are independent of ir luminosity arguments .
the teacup agn presents two additional interesting characteristics : one is the disturbed morphological loop of the northeast region .
the second interesting phenomenon is its extremely strong [ o ii ] @xmath03727 emission .
one possible explanation for both is that the loop was created by expanding radio - jet plasma , consistent with the direction of the current jet , and cosmic rays from the radio plasma also boost the [ o ii ] intensities to what we observe .
the galaxy 3c 48 shows both similar morphology and [ o ii ] @xmath03727/h@xmath18 ratios within its narrow - line region @xcite .
further studies of these objects might offer new perspectives on the interaction between radio - jet plasma and ionized gas in agn .
the eelr around the ionized handle experienced the greatest luminosity , with the location ( 2@xmath2,8@xmath2 ) being greater than 90 times more luminous than the center at @xmath146,000 light years away .
considering the change in luminosity across the distance between the nucleus and the ne corner at ( 2@xmath2,8@xmath2 ) , the time frame for the drop in luminosity fits in with previous studies ( keel et al .
2012 ) . from our results ,
the teacup s nucleus is dimming at @xmath14.7@xmath1310@xmath77 erg s@xmath4 per year , so that 46,000 years ago it was luminous enough to qualify it as a type 2 quasar , while currently the nucleus is in the luminosity regime of a seyfert 2 . from our data presented in figure 11
we see luminosity is decreasing approximately exponentially . at the current rate
at which the teacup is dimming , it will fall out of the realm of seyfert luminosity at @xmath110@xmath78 erg s@xmath4 , in another @xmath1470 years . to further probe the underlying physics behind this change in luminosity , as well the spectacular ionized handle of gas extending out in the north - east corner ,
further data should be collected .
analysis of high resolution imaging from hst will yield a better look at the extended structure of the handle .
x - ray observations would allow a probe of the central agn , and radio observations would provide the opportunity to look for possible jets , which could be driving the handle .
ks gratefully acknowledges support from swiss national science foundation grant pp00p2_138979/1 .
jg would like to acknowledge ks for the naming of the teacup agn , as well as thank dmc and sbk for their mentorship .
2005 may 13 & sdss & survey 2.5-m & @xmath79 & center & 2880 & 3800 - 9250 + 2009 april 18 & lowell & perkins 1.8-m & @xmath79 & center & 1200 & 3860 - 7650 + 2009 april 18 & lowell & perkins 1.8-m & @xmath79 & 5@xmath2 east & 1200 & 3860 - 7650 + 2010 may 20 & lowell & perkins 1.8-m & @xmath79 & center & 1200 & 3870 - 7660 + 2010 june 15 & kpno & 2.1-m & @xmath80 & center & 2700 & 3270 - 5710 + 2010 july 16 & lick & shane 3.0-m & @xmath81 & center & 1800 & 3500 - 5600 + 2010 july 16 & lick & shane 3.0-m & @xmath81 & center & 1800 & 5540 - 8340 + 2011 may 3 & lowell & perkins 1.8-m & @xmath79 & center & 1800 & 3930 - 7730 + 2011 may 3 & lowell & perkins 1.8-m & @xmath79 & 2@xmath2 west & 1800 & 3930 - 7730 + 2011 may 3 & lowell & perkins 1.8-m & @xmath79 & 4@xmath2 west & 1800 & 3930 - 7730 + 2011 may 5 & lowell & perkins 1.8-m & @xmath79 & center & 1800 & 3930 - 7730 + 2011 may 5 & lowell & perkins 1.8-m & @xmath79 & 2@xmath2 east & 1800 & 3930 - 7730 + 2011 may 5 & lowell & perkins 1.8-m & @xmath79 & 4@xmath2 east & 1800 & 3930 - 7730 + lcccc ( @xmath82 ) & 0.17 @xmath21 0.14 & 1.21@xmath1310@xmath83 ( @xmath21 0.21 ) + ( @xmath84 ) & 0.00 @xmath21 0.37 & 1.78@xmath1310@xmath83 ( @xmath21 0.23 ) + ( @xmath85 ) & 0.00 @xmath21 0.31 & 1.54@xmath1310@xmath83 ( @xmath21 0.42 ) + ( @xmath86 ) & 0.11 @xmath21 0.47 & 0.57@xmath1310@xmath83 ( @xmath21 0.16 ) + ( @xmath87 ) & 0.16 @xmath21 0.14 & 0.67@xmath1310@xmath83 ( @xmath21 0.09 ) + ( @xmath88 ) & 0.21 @xmath21 0.16 & 1.40@xmath1310@xmath83 ( @xmath21 0.19 ) + ( @xmath89 ) & 0.21 @xmath21 0.19 & 1.63@xmath1310@xmath83 ( @xmath21 0.16 ) + ( @xmath90 ) & 0.40 @xmath21 0.15 & 1.50@xmath1310@xmath83 ( @xmath21 0.16 ) + ( @xmath91 ) & 0.35 @xmath21 0.24 & 0.90@xmath1310@xmath83 ( @xmath21 0.16 ) + ( @xmath92 ) & 0.27 @xmath21 0.18 & 0.38@xmath1310@xmath83 ( @xmath21 0.13 ) + ( @xmath93 ) & 0.14 @xmath21 0.19 & 0.36@xmath1310@xmath83 ( @xmath21 0.12 ) + ( @xmath94 ) & 0.03 @xmath21 0.18 & 1.15@xmath1310@xmath83 ( @xmath21 0.18 ) + ( @xmath95 ) & 0.22 @xmath21 0.14 & 2.47@xmath1310@xmath83 ( @xmath21 0.30 ) + ( @xmath96 ) & 0.35 @xmath21 0.10 & 4.40@xmath1310@xmath83 ( @xmath21 0.44 ) + ( @xmath97 ) & 0.53 @xmath21 0.08 & 3.97@xmath1310@xmath83 ( @xmath21 0.30 ) + ( @xmath98 ) & 0.60 @xmath21 0.15 & 2.03@xmath1310@xmath83 ( @xmath21 0.31 ) + ( @xmath99 ) & 0.43 @xmath21 0.29 & 0.76@xmath1310@xmath83 ( @xmath21 0.18 ) + ( @xmath100 ) & 0.04 @xmath21 0.29 & 1.11@xmath1310@xmath83 ( @xmath21 0.33 ) + ( @xmath101 ) & 0.43 @xmath21 0.20 & 2.08@xmath1310@xmath83 ( @xmath21 1.21 ) + ( @xmath102 ) & 0.62 @xmath21 0.11 & 6.82@xmath1310@xmath83 ( @xmath21 1.84 ) + ( @xmath103 ) & 0.53 @xmath21 0.10 & 12.72@xmath1310@xmath83 ( @xmath21 3.02 ) + ( @xmath104 ) & 0.41 @xmath21 0.13 & 4.76@xmath1310@xmath83 ( @xmath21 0.44 ) + ( @xmath105 ) & 0.26 @xmath21 0.20 & 0.90@xmath1310@xmath83 ( @xmath21 0.27 ) + ( @xmath106 ) & 0.70 @xmath21 0.17 & 1.21@xmath1310@xmath83 ( @xmath21 0.21 ) + ( @xmath107 ) & 0.40 @xmath21 0.09 & 6.08@xmath1310@xmath83 ( @xmath21 0.62 ) + ( @xmath108 ) & 0.26 @xmath21 0.07 & 7.91@xmath1310@xmath83 ( @xmath21 0.52 ) + ( @xmath109 ) & 0.43 @xmath21 0.15 & 2.94@xmath1310@xmath83 ( @xmath21 0.43 ) + ( @xmath110 ) & 0.66 @xmath21 0.22 & 0.57@xmath1310@xmath83 ( @xmath21 0.11 ) + ( @xmath82 ) & @xmath111 & @xmath112 & & @xmath113 & & @xmath114 & & + & ( @xmath115 ) & ( @xmath116 ) & ( @xmath117 ) & ( @xmath118 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath121 ) & ( @xmath122 ) + ( @xmath84 ) & @xmath123 & @xmath124 & @xmath125 & @xmath126 & @xmath127 & @xmath128 & @xmath129 & @xmath130 + & ( @xmath115 ) & ( @xmath116 ) & ( @xmath117 ) & ( @xmath118 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath121 ) & ( @xmath122 ) + ( @xmath85 ) & @xmath131 & @xmath132 & @xmath133 & @xmath134 & & @xmath135 & & + & ( @xmath136 ) & ( @xmath137 ) & ( @xmath138 ) & ( @xmath139 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath140 ) & ( @xmath141 ) + ( @xmath86 ) & @xmath142 & @xmath143 & & @xmath144 & & @xmath145 & & + & ( @xmath146 ) & ( @xmath147 ) & ( @xmath117 ) & ( @xmath148 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath149 ) & ( @xmath150 ) + ( @xmath87 ) & & @xmath151 & & @xmath152 & @xmath153 & @xmath154 & & + & ( @xmath155 ) & ( @xmath116 ) & ( @xmath117 ) & ( @xmath118 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath121 ) & ( @xmath122 ) + ( @xmath88 ) & & @xmath156 & & @xmath157 & @xmath158 & @xmath159 & @xmath160 & + & ( @xmath161 ) & ( @xmath162 ) & ( @xmath163 ) & ( @xmath164 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath165 ) & ( @xmath141 ) + ( @xmath89 ) & & @xmath166 & & @xmath167 & @xmath168 & @xmath169 & @xmath170 & + & ( @xmath171 ) & ( @xmath172 ) & ( @xmath163 ) & ( @xmath173 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath149 ) & ( @xmath117 ) + ( @xmath90 ) & & @xmath174 & & @xmath175 & @xmath168 & @xmath176 & @xmath177 & @xmath178 + & ( @xmath171 ) & ( @xmath172 ) & ( @xmath163 ) & ( @xmath173 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath149 ) & ( @xmath117 ) + ( @xmath91 ) & & @xmath179 & & & & @xmath180 & & + & ( @xmath181 ) & ( @xmath182 ) & ( @xmath183 ) & ( @xmath150 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath149 ) & ( @xmath184 ) + ( @xmath92 ) & & @xmath185 & & & & & & + & ( @xmath186 ) & ( @xmath187 ) & ( @xmath138 ) & ( @xmath188 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath189 ) & ( @xmath141 ) + ( @xmath93 ) & & @xmath190 & & & & & & + & ( @xmath181 ) & ( @xmath191 ) & ( @xmath183 ) & ( @xmath139 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath149 ) & ( @xmath120 ) + ( @xmath94 ) & & @xmath192 & & @xmath135 & & @xmath193 & & @xmath194 + & ( @xmath195 ) & ( @xmath191 ) & ( @xmath163 ) & ( @xmath150 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath140 ) & ( @xmath196 ) + ( @xmath95 ) & & @xmath197 & @xmath198 & @xmath199 & @xmath200 & @xmath201 & @xmath202 & @xmath203 + & ( @xmath204 ) & ( @xmath205 ) & ( @xmath163 ) & ( @xmath148 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath206 ) & ( @xmath207 ) + ( @xmath96 ) & & @xmath208 & & @xmath153 & @xmath209 & @xmath210 & @xmath211 & @xmath212 + & ( @xmath213 ) & ( @xmath214 ) & ( @xmath215 ) & ( @xmath118 ) & ( @xmath216 ) & ( @xmath217 ) & ( @xmath183 ) & ( @xmath215 ) + ( @xmath97 ) & & @xmath218 & & @xmath219 & @xmath130 & @xmath220 & @xmath221 & @xmath222 + & ( @xmath171 ) & ( @xmath172 ) & ( @xmath163 ) & ( @xmath173 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath149 ) & ( @xmath117 ) + ( @xmath98 ) & & @xmath223 & & @xmath224 & @xmath224 & & & @xmath225 + & ( @xmath171 ) & ( @xmath172 ) & ( @xmath163 ) & ( @xmath173 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath149 ) & ( @xmath117 ) + ( @xmath99 ) & & @xmath226 & & & & & & + & ( @xmath227 ) & ( @xmath228 ) & ( @xmath229 ) & ( @xmath148 ) & ( @xmath216 ) & ( @xmath217 ) & ( @xmath141 ) & ( @xmath230 ) + ( @xmath100 ) & & @xmath231 & & @xmath232 & & @xmath233 & & + & ( @xmath234 ) & ( @xmath235 ) & ( @xmath229 ) & ( @xmath173 ) & ( @xmath216 ) & ( @xmath120 ) & ( @xmath165 ) & ( @xmath164 ) + ( @xmath101 ) & @xmath236 & @xmath237 & & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 + & ( @xmath243 ) & ( @xmath244 ) & ( @xmath229 ) & ( @xmath245 ) & ( @xmath119 ) & ( @xmath217 ) & ( @xmath246 ) & ( @xmath230 ) + ( @xmath102 ) & @xmath247 & @xmath248 & @xmath249 & @xmath250 & @xmath219 & @xmath251 & @xmath221 & + & ( @xmath252 ) & ( @xmath253 ) & ( @xmath117 ) & ( @xmath254 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath255 ) & ( @xmath122 ) + ( @xmath103 ) & @xmath256 & @xmath257 & @xmath258 & @xmath249 & @xmath259 & @xmath260 & @xmath261 & + & ( @xmath262 ) & ( @xmath162 ) & ( @xmath117 ) & ( @xmath164 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath263 ) & ( @xmath206 ) + ( @xmath104 ) & @xmath264 & @xmath265 & @xmath266 & @xmath200 & @xmath170 & @xmath267 & @xmath241 & + & ( @xmath227 ) & ( @xmath228 ) & ( @xmath229 ) & ( @xmath148 ) & ( @xmath216 ) & ( @xmath217 ) & ( @xmath141 ) & ( @xmath230 ) + ( @xmath105 ) & @xmath268 & @xmath269 & @xmath270 & & & @xmath271 & & + & ( @xmath227 ) & ( @xmath228 ) & ( @xmath229 ) & ( @xmath148 ) & ( @xmath216 ) & ( @xmath217 ) & ( @xmath141 ) & ( @xmath230 ) + ( @xmath106 ) & & @xmath272 & & & & & & + & ( @xmath273 ) & ( @xmath274 ) & ( @xmath117 ) & ( @xmath139 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath149 ) & ( @xmath148 ) + ( @xmath107 ) & & @xmath275 & & @xmath241 & @xmath202 & @xmath220 & & + & ( @xmath276 ) & ( @xmath214 ) & ( @xmath117 ) & ( @xmath164 ) & ( @xmath119 ) & ( @xmath120 ) & ( @xmath263 ) & ( @xmath148 ) + ( @xmath108 ) & & @xmath277 & & @xmath278 & @xmath279 & @xmath280 & & + & ( @xmath281 ) & ( @xmath282 ) & ( @xmath229 ) & ( @xmath230 ) & ( @xmath216 ) & ( @xmath120 ) & ( @xmath229 ) & ( @xmath283 ) + ( @xmath109 ) & & @xmath284 & & @xmath198 & & @xmath285 & & + & ( @xmath286 ) & ( @xmath235 ) & ( @xmath229 ) & ( @xmath173 ) & ( @xmath216 ) & ( @xmath217 ) & ( @xmath263 ) & ( @xmath230 ) + ( @xmath110 ) & & @xmath287 & & & & & & + & ( @xmath288 ) & ( @xmath289 ) & ( @xmath255 ) & ( @xmath230 ) & ( @xmath216 ) & ( @xmath120 ) & ( @xmath183 ) & ( @xmath283 ) + ( @xmath82 ) & @xmath290 & @xmath291 & & @xmath292 & & @xmath293 & @xmath294 & @xmath295 & @xmath296 + & ( @xmath297 ) & ( @xmath298 ) & ( @xmath263 ) & ( @xmath138 ) & ( @xmath121 ) & ( @xmath299 ) & ( @xmath300 ) & ( @xmath301 ) & ( @xmath302 ) + ( @xmath84 ) & @xmath303 & @xmath304 & & @xmath305 & & @xmath306 & @xmath307 & @xmath308 & @xmath309 + & ( @xmath297 ) & ( @xmath298 ) & ( @xmath263 ) & ( @xmath138 ) & ( @xmath121 ) & ( @xmath299 ) & ( @xmath300 ) & ( @xmath301 ) & ( @xmath302 ) + ( @xmath85 ) & @xmath310 & @xmath311 & & @xmath312 & & @xmath313 & @xmath314 & @xmath315 & @xmath316 + & ( @xmath317 ) & ( @xmath318 ) & ( @xmath229 ) & ( @xmath138 ) & ( @xmath121 ) & ( @xmath299 ) & ( @xmath319 ) & ( @xmath320 ) & ( @xmath321 ) + ( @xmath86 ) & @xmath322 & @xmath323 & & @xmath324 & & @xmath325 & @xmath326 & & + & ( @xmath327 ) & ( @xmath328 ) & ( @xmath165 ) & ( @xmath329 ) & ( @xmath183 ) & ( @xmath330 ) & ( @xmath331 ) & ( @xmath332 ) & ( @xmath333 ) + ( @xmath87 ) & @xmath334 & @xmath335 & & & & @xmath336 & @xmath337 & @xmath338 & @xmath338 + & ( @xmath339 ) & ( @xmath340 ) & ( @xmath263 ) & ( @xmath341 ) & ( @xmath189 ) & ( @xmath342 ) & ( @xmath343 ) & ( @xmath116 ) & ( @xmath344 ) + ( @xmath88 ) & @xmath345 & @xmath346 & & @xmath347 & & @xmath348 & @xmath349 & @xmath350 & @xmath351 + & ( @xmath352 ) & ( @xmath353 ) & ( @xmath255 ) & ( @xmath354 ) & ( @xmath246 ) & ( @xmath283 ) & ( @xmath355 ) & ( @xmath356 ) & ( @xmath357 ) + ( @xmath89 ) & @xmath358 & @xmath359 & & @xmath360 & & @xmath361 & @xmath362 & @xmath363 & @xmath363 + & ( @xmath364 ) & ( @xmath365 ) & ( @xmath255 ) & ( @xmath173 ) & ( @xmath255 ) & ( @xmath283 ) & ( @xmath366 ) & ( @xmath116 ) & ( @xmath116 ) + ( @xmath90 ) & @xmath367 & @xmath368 & & @xmath135 & & @xmath369 & @xmath124 & @xmath370 & @xmath371 + & ( @xmath364 ) & ( @xmath365 ) & ( @xmath255 ) & ( @xmath173 ) & ( @xmath255 ) & ( @xmath283 ) & ( @xmath366 ) & ( @xmath116 ) & ( @xmath116 ) + ( @xmath91 ) & @xmath372 & @xmath373 & & @xmath374 & & @xmath375 & @xmath376 & @xmath377 & @xmath378 + & ( @xmath379 ) & ( @xmath380 ) & ( @xmath263 ) & ( @xmath381 ) & ( @xmath263 ) & ( @xmath254 ) & ( @xmath382 ) & ( @xmath137 ) & ( @xmath282 ) + ( @xmath92 ) & @xmath383 & @xmath384 & & & & @xmath385 & @xmath386 & & + & ( @xmath387 ) & ( @xmath388 ) & ( @xmath229 ) & ( @xmath381 ) & ( @xmath263 ) & ( @xmath254 ) & ( @xmath389 ) & ( @xmath137 ) & ( @xmath357 ) + ( @xmath93 ) & @xmath390 & @xmath391 & & & & @xmath392 & @xmath393 & @xmath394 & @xmath395 + & ( @xmath327 ) & ( @xmath396 ) & ( @xmath263 ) & ( @xmath381 ) & ( @xmath263 ) & ( @xmath254 ) & ( @xmath382 ) & ( @xmath137 ) & ( @xmath282 ) + ( @xmath94 ) & @xmath397 & @xmath398 & & @xmath399 & & @xmath361 & @xmath400 & @xmath401 & @xmath402 + & ( @xmath403 ) & ( @xmath404 ) & ( @xmath405 ) & ( @xmath120 ) & ( @xmath215 ) & ( @xmath120 ) & ( @xmath406 ) & ( @xmath289 ) & ( @xmath407 ) + ( @xmath95 ) & @xmath334 & @xmath408 & @xmath160 & @xmath409 & @xmath410 & @xmath347 & @xmath294 & @xmath411 & @xmath412 + & ( @xmath413 ) & ( @xmath414 ) & ( @xmath255 ) & ( @xmath138 ) & ( @xmath121 ) & ( @xmath299 ) & ( @xmath319 ) & ( @xmath320 ) & ( @xmath321 ) + ( @xmath96 ) & @xmath415 & @xmath416 & @xmath177 & @xmath417 & @xmath418 & @xmath419 & @xmath420 & @xmath421 & @xmath422 + & ( @xmath423 ) & ( @xmath424 ) & ( @xmath121 ) & ( @xmath164 ) & ( @xmath229 ) & ( @xmath164 ) & ( @xmath425 ) & ( @xmath426 ) & ( @xmath214 ) + ( @xmath97 ) & @xmath427 & @xmath428 & @xmath160 & @xmath429 & @xmath430 & @xmath429 & @xmath431 & @xmath432 & @xmath433 + & ( @xmath364 ) & ( @xmath365 ) & ( @xmath255 ) & ( @xmath173 ) & ( @xmath255 ) & ( @xmath283 ) & ( @xmath366 ) & ( @xmath116 ) & ( @xmath116 ) + ( @xmath98 ) & @xmath434 & @xmath435 & & @xmath436 & & @xmath437 & @xmath438 & @xmath439 & @xmath360 + & ( @xmath364 ) & ( @xmath365 ) & ( @xmath255 ) & ( @xmath173 ) & ( @xmath255 ) & ( @xmath283 ) & ( @xmath366 ) & ( @xmath116 ) & ( @xmath116 ) + ( @xmath99 ) & @xmath440 & @xmath441 & & @xmath442 & & @xmath443 & @xmath444 & @xmath445 & @xmath446 + & ( @xmath447 ) & ( @xmath448 ) & ( @xmath206 ) & ( @xmath449 ) & ( @xmath246 ) & ( @xmath139 ) & ( @xmath426 ) & ( @xmath450 ) & ( @xmath282 ) + ( @xmath100 ) & @xmath451 & @xmath452 & & @xmath453 & & @xmath454 & @xmath455 & @xmath445 & @xmath456 + & ( @xmath457 ) & ( @xmath458 ) & ( @xmath206 ) & ( @xmath245 ) & ( @xmath229 ) & ( @xmath164 ) & ( @xmath459 ) & ( @xmath450 ) & ( @xmath460 ) + ( @xmath101 ) & @xmath461 & @xmath462 & @xmath430 & @xmath463 & & @xmath232 & @xmath464 & @xmath465 & @xmath466 + & ( @xmath467 ) & ( @xmath468 ) & ( @xmath206 ) & ( @xmath122 ) & ( @xmath141 ) & ( @xmath118 ) & ( @xmath469 ) & ( @xmath470 ) & ( @xmath357 ) + ( @xmath102 ) & @xmath471 & @xmath472 & @xmath160 & @xmath113 & & @xmath473 & @xmath474 & @xmath475 & @xmath233 + & ( @xmath476 ) & ( @xmath477 ) & ( @xmath263 ) & ( @xmath478 ) & ( @xmath263 ) & ( @xmath354 ) & ( @xmath479 ) & ( @xmath357 ) & ( @xmath480 ) + ( @xmath103 ) & @xmath481 & @xmath482 & @xmath202 & @xmath483 & & @xmath484 & @xmath485 & @xmath486 & @xmath419 + & ( @xmath487 ) & ( @xmath488 ) & ( @xmath165 ) & ( @xmath478 ) & ( @xmath263 ) & ( @xmath217 ) & ( @xmath489 ) & ( @xmath357 ) & ( @xmath344 ) + ( @xmath104 ) & @xmath490 & @xmath491 & @xmath261 & @xmath185 & & @xmath185 & @xmath492 & @xmath493 & @xmath494 + & ( @xmath447 ) & ( @xmath448 ) & ( @xmath206 ) & ( @xmath449 ) & ( @xmath246 ) & ( @xmath139 ) & ( @xmath426 ) & ( @xmath450 ) & ( @xmath282 ) + ( @xmath105 ) & @xmath495 & @xmath496 & & @xmath497 & & @xmath498 & @xmath499 & @xmath500 & @xmath501 + & ( @xmath447 ) & ( @xmath502 ) & ( @xmath206 ) & ( @xmath449 ) & ( @xmath246 ) & ( @xmath139 ) & ( @xmath426 ) & ( @xmath450 ) & ( @xmath282 ) + ( @xmath106 ) & @xmath503 & @xmath504 & & @xmath505 & & @xmath506 & @xmath507 & @xmath508 & @xmath305 + & ( @xmath509 ) & ( @xmath510 ) & ( @xmath165 ) & ( @xmath381 ) & ( @xmath263 ) & ( @xmath254 ) & ( @xmath382 ) & ( @xmath511 ) & ( @xmath282 ) + ( @xmath107 ) & @xmath512 & @xmath513 & & @xmath514 & & @xmath157 & @xmath515 & @xmath516 & @xmath432 + & ( @xmath352 ) & ( @xmath517 ) & ( @xmath165 ) & ( @xmath230 ) & ( @xmath255 ) & ( @xmath217 ) & ( @xmath389 ) & ( @xmath244 ) & ( @xmath289 ) + ( @xmath108 ) & @xmath427 & @xmath518 & & @xmath519 & & @xmath520 & @xmath521 & @xmath522 & @xmath433 + & ( @xmath523 ) & ( @xmath524 ) & ( @xmath206 ) & ( @xmath525 ) & ( @xmath141 ) & ( @xmath164 ) & ( @xmath425 ) & ( @xmath450 ) & ( @xmath162 ) + ( @xmath109 ) & @xmath526 & @xmath527 & & @xmath528 & & @xmath529 & @xmath530 & @xmath531 & @xmath532 + & ( @xmath533 ) & ( @xmath534 ) & ( @xmath206 ) & ( @xmath120 ) & ( @xmath215 ) & ( @xmath283 ) & ( @xmath535 ) & ( @xmath426 ) & ( @xmath302 ) + ( @xmath110 ) & @xmath536 & @xmath537 & & @xmath538 & & @xmath539 & @xmath540 & @xmath541 & @xmath501 + & ( @xmath542 ) & ( @xmath543 ) & ( @xmath189 ) & ( @xmath544 ) & ( @xmath246 ) & ( @xmath150 ) & ( @xmath470 ) & ( @xmath253 ) & ( @xmath545 ) + ( @xmath82 ) & 10.88 & -2.4 & 1.8@xmath546 & 19.6 & -3.4 & 2.8@xmath546 & 19.7 & -2.5 & -3.5 & 18.2 & 46.22@xmath547 + ( @xmath84 ) & 9.18 & -2.4 & 1.8@xmath548 & 19.6 & -3.4 & 2.8@xmath548 & 19.7 & -2.5 & -3.5 & 18.2 & 46.07@xmath549 + ( @xmath85 ) & 8.50 & -2.3 & 1.6@xmath550 & 20.5 & -3.4 & 2.7@xmath550 & 19.7 & -2.5 & -4.0 & 17.7 & 45.90@xmath551 + ( @xmath86 ) & 9.18 & -2.4 & 1.7@xmath552 & 19.8 & -3.5 & 2.8@xmath552 & 19.7 & -2.5 & -4.0 & 17.7 & 45.97@xmath553 + ( @xmath87 ) & 12.24 & -2.4 & 1.8@xmath554 & 19.6 & -3.4 & 2.8@xmath554 & 19.6 & -2.5 & -3.5 & 17.7 & 46.32@xmath555 + ( @xmath88 ) & 9.69 & -2.4 & 2.1@xmath556 & 20.4 & -3.1 & 2.8@xmath556 & 20.2 & -2.5 & -3.5 & 18.5 & 46.42@xmath557 + ( @xmath89 ) & 7.65 & -2.4 & 2.1@xmath558 & 20.2 & -3.2 & 2.9@xmath558 & 20.1 & -2.5 & -5.0 & 18.2 & 46.21@xmath559 + ( @xmath90 ) & 6.80 & -2.4 & 2.1@xmath558 & 20.2 & -3.2 & 2.9@xmath558 & 20.1 & -2.5 & -5.0 & 18.2 & 46.11@xmath560 + ( @xmath91 ) & 7.65 & -2.3 & 2.1@xmath561 & 20.5 & -3.4 & 3.2@xmath561 & 19.8 & -2.4 & -3.5 & 17.6 & 46.31@xmath562 + ( @xmath92 ) & 9.69 & -2.3 & 2.1@xmath552 & 20.5 & -3.4 & 3.2@xmath552 & 19.8 & -2.5 & -3.5 & 17.6 & 46.52@xmath563 + ( @xmath93 ) & 14.11 & -2.3 & 1.6@xmath564 & 20.5 & -3.4 & 2.7@xmath564 & 19.7 & -2.5 & -4.0 & 17.7 & 46.34@xmath565 + ( @xmath94 ) & 10.20 & -2.3 & 1.6@xmath566 & 20.5 & -3.4 & 2.7@xmath566 & 19.7 & -2.5 & -4.0 & 17.7 & 46.10@xmath567 + ( @xmath95 ) & 7.65 & -2.4 & 1.8@xmath568 & 19.6 & -3.4 & 2.8@xmath568 & 19.7 & -2.5 & -3.5 & 18.2 & 45.91@xmath569 + ( @xmath96 ) & 4.76 & -2.4 & 2.1@xmath570 & 20.2 & -3.2 & 2.9@xmath570 & 20.1 & -2.5 & -5.0 & 18.2 & 45.80@xmath571 + ( @xmath97 ) & 3.40 & -2.4 & 2.1@xmath572 & 20.2 & -3.2 & 2.9@xmath572 & 20.1 & -2.5 & -5.0 & 18.2 & 45.51@xmath573 + ( @xmath98 ) & 4.76 & -2.4 & 2.1@xmath574 & 20.2 & -3.2 & 2.9@xmath574 & 20.1 & -2.5 & -5.0 & 18.2 & 45.80@xmath575 + ( @xmath99 ) & 7.65 & -2.4 & 2.1@xmath576 & 20.4 & -3.1 & 2.8@xmath576 & 20.2 & -2.5 & -3.5 & 18.5 & 46.21@xmath577 + ( @xmath100 ) & 10.20 & -2.3 & 1.6@xmath578 & 20.5 & -3.4 & 2.7@xmath578 & 19.7 & -2.5 & -4.0 & 17.7 & 46.06@xmath579 + ( @xmath101 ) & 6.80 & -2.4 & 1.7@xmath580 & 20.0 & -3.4 & 2.7@xmath580 & 19.7 & -2.5 & -3.5 & 18.0 & 45.71@xmath581 + ( @xmath102 ) & 3.40 & -2.4 & 1.7@xmath582 & 19.6 & -3.4 & 2.7@xmath582 & 19.7 & -2.5 & -3.5 & 18.0 & 45.11@xmath573 + ( @xmath103 ) & 1.02 & -2.4 & 1.7@xmath572 & 19.6 & -3.2 & 2.8@xmath572 & 20.0 & -2.5 & -2.5 & 18.4 & 44.36@xmath583 + ( @xmath104 ) & 3.40 & -2.4 & 1.7@xmath570 & 19.6 & -3.4 & 2.7@xmath570 & 19.8 & -2.6 & -2.5 & 18.0 & 45.11@xmath584 + ( @xmath105 ) & 6.80 & -2.4 & 1.7@xmath585 & 19.6 & -3.4 & 2.7@xmath585 & 19.7 & -2.5 & -3.5 & 18.0 & 45.71@xmath581 + ( @xmath106 ) & 7.65 & -2.4 & 2.1@xmath568 & 20.2 & -3.2 & 2.9@xmath568 & 20.1 & -2.5 & -5.0 & 18.2 & 46.21@xmath586 + ( @xmath107 ) & 4.76 & -2.4 & 2.1@xmath587 & 19.6 & -3.2 & 2.9@xmath587 & 20.0 & -2.4 & -2.0 & 18.1 & 45.40@xmath588 + ( @xmath108 ) & 3.40 & -2.4 & 2.1@xmath589 & 20.2 & -3.2 & 2.9@xmath589 & 20.1 & -2.5 & -5.0 & 18.2 & 45.51@xmath590 + ( @xmath109 ) & 4.76 & -2.4 & 2.1@xmath552 & 20.5 & -3.2 & 2.9@xmath552 & 20.1 & -2.5 & -5.0 & 18.2 & 45.10@xmath591 + ( @xmath110 ) & 7.65 & -2.4 & 2.1@xmath552 & 20.2 & -3.2 & 2.9@xmath552 & 20.1 & -2.5 & -5.0 & 18.2 & 45.21@xmath592 + | the sloan digital sky survey ( sdss ) galaxy zoo project has revealed a number of spectacular galaxies possessing extended emission - line regions ( eelrs ) , the most famous being hanny s voorwerp galaxy .
we present another eelr object discovered in the sdss endeavor : the teacup active galactic nucleus ( agn ) , nicknamed for its eelr , which has a `` handle '' like structure protruding 15 kpc into the northeast quadrant of the galaxy .
we analyze physical conditions of this galaxy with long - slit ground based spectroscopy from lowell , lick , and kpno observatories . with the lowell 1.8-m perkin s telescope we took multiple observations at different offset positions , allowing us to recover spatially
resolved spectra across the galaxy .
line diagnostics indicate the ionized gas is photoionized primarily by the agn .
additionally we are able to derive the hydrogen density from the [ s ii ] @xmath06716/@xmath06731 ratio .
we generated two - component photoionization models for each spatially resolved lowell spectrum .
these models allow us to calculate the agn bolometric luminosity seen by the gas at different radii from the nuclear center of the teacup .
our results show a drop in bolometric luminosity by more than two orders of magnitude from the eelr to the nucleus , suggesting that the agn has decreased in luminosity by this amount in a continuous fashion over 46,000 years , supporting the case for a dying agn in this galaxy independent of any ir based evidence .
we demonstrate that spatially resolved photoionization modeling could be applied to eelrs to investigate long time scale variability . |
according to birman @xcite , markov @xcite originally stated his braid equivalence theorem using _ three _ braid moves ; later there was another brief announcement of an improved version of markov s theorem by weinberg @xcite consisting of the _ two _ well - known braid - equivalence moves : conjugation in the braid groups and the ` stabilizing ' or ` markov ' moves ( @xmath1moves ) .
our first main result is a _ one - move _ markov theorem which we now state .
an _ @xmath2move _ on a braid consists of cutting one arc of the braid open and splicing into the broken strand new strands to the top and bottom , both either _ under _ or _ over _ the rest of the braid : @xmath3 @xmath2moves and isotopy generate an equivalence relation on braids called _ @xmath2equivalence_. in 4 we prove that @xmath2equivalence classes of braids are in bijective correspondence with isotopy classes of oriented links in @xmath0 , where the bijection is induced by ` closing ' the braid to form a link . as a consequence
, @xmath2equivalence is the same as the usual markov equivalence and thus the classical markov theorem is sharpened .
the proofs are based on a canonical process for turning a combinatorial oriented link diagram in the plane ( with a little extra structure ) into an open braid .
our braiding as well as the @xmath2moves are based on the building blocks of combinatorial isotopy , the _ triangle moves _ or _
@xmath4moves_. this makes the proof conceptually very simple .
our braiding operation is essentially the same as the operation using a ` saw - tooth ' given by j. birman in @xcite .
we use the point at infinity as the reference point for braiding and so a saw - tooth becomes a pair of vertical lines that meet at infinity .
the change of reference point to infinity makes the proof of markov theorem easier because there are very few ways that saw - teeth can obstruct each other .
moreover , the local nature of our proof allows us to formulate and prove first the relative version of markov theorem , another new result stating that , if two isotopic links contain the same braided portion then any two corresponding braids which both contain that braided portion differ by a sequence of @xmath2moves that do not affect the braided portion .
we then prove a second relative version of the theorem for links which contain a common closed braid and deduce an analogue of markov theorem for knot isotopy in complements of knots / links , which we finally extend to an analogue of markov theorem for links in arbitrary closed 3manifolds .
more precisely , let @xmath1 be a closed connected orientable 3manifold .
we can assume that @xmath1 is given by surgery on the closure of a braid @xmath5 in @xmath0 , thus @xmath1 is represented by @xmath5 .
furthermore , any oriented link in @xmath1 can be represented as the closure of a further braid @xmath6 such that @xmath7 is also a braid and we shall call it _ mixed braid_. then the mixed braid equivalence in @xmath0 that reflects isotopy in @xmath1 is generated firstly by @xmath2moves , as above , but performed only on strings of @xmath6 and secondly by _ braid band moves _ or _ b.b .
moves_. these are moves that reflect the sliding of a part of a link across the 2disc bounded by the specified longitude of a surgery component of the closure of @xmath5 .
i.e. suppose a string of @xmath6 is adjacent to one of @xmath5 , then a slide of the first over the second ( with a half - twisted band ) replaces @xmath6 by a braid with one or more extra strings parallel to the strings of @xmath5 which form the appropriate component in the closure : @xmath8 most results are based on material in @xcite worked at warwick university under the advice of the second author . however , they are sharpened versions and some parts of the proofs are improved considerably .
* acknowledgements * the second author would like to acknowledge a conversation with dale rolfsen which was crucial in formulating the one - move markov theorem .
the first author wishes to acknowledge discussions with jozef przytycki , cameron gordon and david yetter on the set up and some technical details , in warsaw 95 , and also to thank the e.u . for financial support and the sfb 170 in gttingen for hospitality and financial support
a knot is a special case of a link , so from now on we shall be referring to both knots and links as ` links ' . throughout the first part of this paper we shall be working in a combinatorial setting . in particular
, we consider oriented _ links / geometric braids _ in 3space each of the components / strands of which is made of a finite number of straight arcs endowed with matching orientation ; also their _ diagrams _ , that is , regular projections on the plane ( with only finitely many crossings ) , where in addition no vertex of the link / braid should be mapped onto a double point . a _
geometric braid diagram _ or simply a _ braid _ has a top - to - bottom direction and , in addition , we require that no two crossings are on the same horizontal level .
if we slice up in general position ( i.e. without cutting through crossings ) a braid diagram on @xmath9 strands , it may be seen as a word on the ( well - known ) basic crossings @xmath10 and @xmath11 for @xmath12 ( we refer the reader to figure 9 for an example ) .
the set of braids on @xmath9 strands modulo isotopy gives rise to the braid group @xmath13 with presentation : @xmath14 the operation in the group is _ concatenation _ ( we place one braid on top of the other ) , and the identity element is the braid on @xmath9 strands that does no braiding .
note that the elements of @xmath13 are also called ` braids ' , but in here there will be no ambiguity , since we shall be working with geometric braid diagrams .
there are two combinatorial moves on diagrams which we shall consider .
* * @xmath4move : * an arc is replaced by two arcs forming a triangle ( and its inverse ) , respecting orientation and crossings . `
respecting crossings ' means that , if we lift the diagram to an embedding in 3space , then the @xmath4move lifts to an elementary isotopy ( see figure 3 for examples ) . * * subdivision : * a vertex is introduced / deleted in an arc of the diagram .
subdivision moves may be viewed as special cases of @xmath4moves .
we shall call the equivalence relation generated by these two moves a _ combinatorial isotopy _ or just an _ isotopy_. it is a classical result of combinatorial topology that this notion of isotopy is equivalent to the standard definition of combinatorial ( or pl ) isotopy of the embedding obtained by lifting to 3space , and this in turn is equivalent to the notion of isotopy in the smooth category
. moreover , reidemeister @xcite ( and alexander , briggs @xcite ) proved that a @xmath4move can break into a finite sequence of planar @xmath4moves and the three local @xmath4moves illustrated below ( known as ` reidemeister moves ' ) with their obvious symmetries ( for a detailed account see @xcite ) .
@xmath15 let @xmath16 be a link diagram / braid and @xmath17 a point of an arc of @xmath16 such that @xmath17 is not vertically aligned with any of the crossings or ( other ) vertices of @xmath16 ( note that @xmath17 itself may be a vertex ) .
then we can perform the following operation : cut the arc at @xmath17 , bend the two resulting smaller arcs apart slightly by a small isotopy and introduce two new vertical arcs to new top and bottom end - points in the same vertical line as @xmath17 .
the new arcs are both oriented downwards and they run either both _ under _ or both _ over _ all other arcs of the diagram .
thus there are two types of @xmath2moves , an _ under @xmath2move _ or _
@xmath18move _ and an _ over @xmath2move _ or _
@xmath19move_. ( recall figure 1 above for an example of the two braid @xmath2moves . ) below we illustrate @xmath19moves applied to an edge @xmath20 of an oriented link diagram such that @xmath20 moves via isotopy from a downward arc to an upward arc through the horizontal position .
the thick circle will represent ` the rest of the diagram ' , while the region inside the circle shall be called ` the magnified region ' . note that , if we join freely ( using the dotted arcs ) the two new arcs we obtain a link diagram isotopic to the one on which the @xmath2move was applied . @xmath21 definition 2.1 could be given even more generally so as to accommodate the possibility of the new vertical strands running _ upwards _
( figure 4b ) .
although , for the purposes of this paper we shall only consider @xmath2moves with the new vertical arcs oriented _
downwards_. using a small braid isotopy , a braid @xmath2move can be equivalently seen with a crossing ( positive or negative ) formed : @xmath22 this gives the following algebraic expression for an @xmath19move and an @xmath18move respectively .
@xmath23 @xmath24 where @xmath25 , @xmath26 are elements of @xmath13 and @xmath27 , @xmath28 are obtained from @xmath25 , @xmath26 by replacing each @xmath29 by @xmath30 for @xmath31 .
@xmath2moves and isotopy generate an equivalence relation in the set of braids , which we shall call _ @xmath2equivalence_. let now @xmath5 be a braid .
the _ closure _ of @xmath5 is the oriented link diagram @xmath32 obtained by joining each top end - point to the corresponding bottom end - point by almost vertical arcs as illustrated in figure 6 , where @xmath5 is contained in a ` box ' .
( note that we draw some smooth arcs for convenience . )
@xmath33 our aim is to prove the following theorem .
@xmath34 induces a bijection between the set of @xmath2equivalence classes of braids and the set of isotopy types of ( oriented ) link diagrams .
we shall define an inverse bijection to @xmath34 by means of a canonical _ braiding process _ which turns an oriented link diagram ( with a little extra structure ) into a braid .
note that we only work with oriented diagrams , so in the sequel we shall drop the adjective .
let @xmath16 be a link diagram with no horizontal arcs , and consider the arcs in @xmath16 which slope upwards with respect to their orientations ; call these arcs _
opposite arcs_. in order to obtain a braid from that diagram we want : 1 ) to keep the arcs that go downwards .
\2 ) to eliminate the opposite arcs and produce instead braid strands .
if we run along an opposite arc we are likely to meet a succession of overcrossings and undercrossings .
we subdivide ( marking with points ) every opposite arc into smaller if necessary
pieces , each containing crossings of only _
one _ type ; i.e. we may have : @xmath35 we call the resulting pieces _ up arcs _ , and we label every up arc with an _ ` o'/`u ' _ according as it is the _ over / under _ arc of a crossing ( or some crossings ) .
if it is a _
free up arc _ (
and therefore it contains no crossings ) , then we have a choice whether to label it _ ` o ' _ or _ `
u'_. the idea is to eliminate the opposite arcs by eliminating their up
arcs one by one and create braid strands instead .
let now @xmath36 be the top vertices of the up
arcs ; fix attention on one particular top vertex @xmath37 and suppose that @xmath17 is the top vertex of the up
arc @xmath38 . associated to @xmath38 is the _ sliding triangle _
@xmath39 , which is a special case of a triangle needed for a @xmath4move ; it is right - angled with hypotenuse @xmath38 and with the right angle lying below the up arc .
note that , if @xmath38 is itself vertical , then @xmath39 degenerates into the arc @xmath38 .
we say that a sliding triangle is of type _ over _ or _ under _ according to the label of the up arc it is associated with .
( this implies that there may be triangles of the same type lying one on top of the other . )
the germ of our braiding process is this .
suppose for definiteness that @xmath38 is of type _
over_. then perform an @xmath19move at @xmath17 followed by a @xmath4move across the sliding triangle @xmath39 ( see figure 8) . by general position
the resulting diagram will be regular and @xmath40 may be assumed to slope slightly downwards .
if @xmath38 were _ under _ then the @xmath19move would be replaced by an @xmath18move .
note that the effect of these two moves has been to replace the up
arc @xmath38 by three arcs none of which are up arcs , and therefore we now have fewer up arcs . if we repeat this process for each up arc in turn , then the result will be a braid .
@xmath41 in the following example we illustrate the braiding process applied to a particular link diagram ( we do not draw the sliding triangles ) .
the numbering of the strands indicates the order in which we eliminated the up arcs . obviously , for different orders we may obtain different braids .
@xmath42 it is clear from the above that different choices when doing the braiding ( e.g. when choosing labels for the free up arcs or when choosing an order for layering sliding triangles of the same type ) as well as slight isotopy changes on the diagram level may result in important changes in the braid picture . for the proof of theorem 2.3 we would like to control such changes as much as possible
for this we shall need a more rigorous setting .
sliding triangles are said to be _ adjacent _ if the corresponding up
arcs have a common vertex ( then the sliding triangles will have a common corner ) .
* triangle condition * _ non - adjacent sliding triangles are only allowed to meet if they are of opposite types ( i.e. one over and the other under ) . _
@xmath43 given a link diagram @xmath16 , there is a subdivision @xmath44 of @xmath16 such that ( for appropriate choices of under / over for free up
arcs ) the triangle condition is satisfied . let @xmath45 minimum distance between any two crossings of @xmath16 .
let @xmath46 be such that any circle of radius @xmath47 centred at a crossing point does not intersect with any other arc of the diagram .
let @xmath48 be the minimum distance between any two points in @xmath16 further than @xmath47 away from any crossing point .
now let @xmath49min\{@xmath48,@xmath47 } and @xmath44 be a subdivision of the diagram such that the length of every up
arc is less than @xmath50 .
then the triangle condition is satisfied , provided we make the right choices ( under / over ) for sliding triangles of free up
arcs near crossings ( see picture above ) .
\(1 ) the triangle condition implies that the eliminating moves do not interfere with each other , so there are no pairs of sliding triangles that need layering and therefore it does not matter in what order we eliminate the up arcs .
in fact we can eliminate all of them _ simultaneously_. ( 2 )
the choice of @xmath50 in the last proof is far smaller than necessary .
in fact we only need to consider the subdiagram consisting of up arcs and crossings on up arcs .
( 3 ) notice that the choice of under / over for some of the free up
arcs may be forced by the labelling of other up arcs ( see for example the top _ over _ up arc in figure 10 ) .
we can if we wish further subdivide the diagram ( here the under up arc ) to make such pairs disjoint .
( 4 ) if a diagram satisfies the triangle condition , then so does any subdivision ( where the smaller triangles may need to retain the same labels ) . a _ generic diagram _ is a link diagram with subdividing points and sliding triangles put in general position with respect to the height function , such that the following conditions hold : 1 )
there are no horizontal arcs , \2 ) no two disjoint subdividing points are in vertical alignment , where by ` disjoint ' we mean subdividing points that do not share a common edge .
\3 ) any two non - adjacent sliding triangles satisfy the triangle condition and if they intersect , this should be along a common interior ( and not a single point ) . conditions 1 and 3 are related to the braiding , whilst condition 2 ensures that no pair of strands in the resulting braid will be in the same vertical line .
generic @xmath4move _ is a @xmath4move between generic diagrams . in the sequel , by ` @xmath4moves '
we shall always refer to the local planar @xmath4moves together with the reidemeister moves .
an isotopy between generic link diagrams can be realized using only generic @xmath4moves .
if after some @xmath4move during the isotopy appears a horizontal arc or vertical alignment of vertices we remove it by replacing one of the participating vertices by a point arbitrarily close to it , so that the new point will not cause such a singularity or violation of condition 3 in _ all _ the diagrams of the isotopy chain .
this is possible by a general position argument .
if two non - adjacent sliding triangles touch so that condition 3 is violated ( figure 11a ) we argue as above . also , by definition of regular isotopy , two non - adjacent triangles can not touch on two subdividing points or on a point with a hypotenuse or along their hypotenuses .
therefore , the remaining possibilities are the ones illustrated in figure 11b .
@xmath51 as before , by general position we can replace one of the participating vertices by a point arbitrarily close to it , so that the new point will not violate conditions 1 , 2 and 3 in all the diagrams of the isotopy chain .
finally , if two triangles happen to intersect after a @xmath4move and they are of opposite type , then the move is generic .
if they are of the same type and the intersection is not essential ( see figure 12 ) we simply subdivide further and we carry the subdivision through in the whole chain of the isotopic diagrams , taking care that the subdividing point will not violate condition 2 in the whole chain .
( note that , by remark 3.2 ( 4 ) , condition 3 will not be violated . )
@xmath52 if the intersection is essential ( for example in figure 13 a free up arc labelled ` u ' moves by a @xmath4move over another arc ) , then we introduce an appropriate extra subdividing point so as to create a smaller free up arc
to which we attach the opposite label .
this is always possible by lemma 3.1 .
@xmath53 it is clear from lemma 3.5 that generic link diagrams are dense in the space of all diagrams ; indeed , a non - generic link diagram may be seen as a middle stage of the isotopy between generic diagrams .
thus w.l.o.g .
we shall assume from now on that all diagrams are generic ; also , by virtue of lemma 3.5 that all isotopy moves are also generic .
we are now ready to give a rigorous braiding process .
namely , take a link diagram and eliminate one by one the up arcs in the way described above . by remark 3.2
( 1 ) the order of the eliminating moves is now irrelevant . any ( oriented ) link diagram is isotopic to the closure of a braid .
the braiding process comprises @xmath2moves and then isotopies .
but the effect of an @xmath2move after closure is by definition an isotopy .
by the local nature of the @xmath4moves we may assume that for a given link diagram we have done the braiding for all up
arcs except for the ones that we are interested in every time ; these will be lying in the magnified region placed inside the braid . by remark 3.2 ( 1 ) this choice does not affect the final braid .
now , two braids that differ by a finite sequence of @xmath2moves have isotopic closures .
therefore the function @xmath34 from @xmath2equivalence classes of braids to isotopy types of link diagrams is well - defined . to show that @xmath34 is a bijection we shall use our braiding process to define an inverse function @xmath54 .
namely , for a diagram @xmath16 let @xmath55 be the braid resulting from the braiding algorithm applied to it .
we have to show that @xmath54 is a well - defined function from link - diagram types to @xmath2equivalence classes of braids , therefore we have to check that @xmath55 does not depend up to @xmath2equivalence on the choices made before the braiding and on @xmath4moves between link diagrams .
the choices made before the braiding consist of the subdividing points we choose and the labelling we may have to choose for some free up
finally , we have to show that @xmath34 and @xmath54 are mutually inverse .
this is easy : closing the result of the braiding process yields a link isotopic to the original one , therefore @xmath56 .
moreover , applying the braiding process to the closure of a braid yields precisely the same braid back again ( from the way we defined closure ) , so @xmath57 .
the proof relies entirely on remark 3.2 ( 1 ) and on the following two lemmas .
if we add on an up arc , @xmath58 , an extra subdividing point @xmath17 and label the two new up arcs , @xmath25 and @xmath26 , the same as @xmath58 , the corresponding braids are @xmath2equivalent .
for definiteness we assume that @xmath58 is labelled with an ` o ' .
we complete the braiding of the original diagram by eliminating @xmath58 ( see picture below ) .
then , on the new horizontal piece of string , we take an arbitrarily small neighbourhood @xmath59 around @xmath60 , the projection of @xmath17 ( see picture below ) . by general position @xmath59 slopes slightly downwards .
we then perform an _ over _ l - move at @xmath60 . finally ,
sliding an appropriate piece of string using braid planar isotopy , we obtain the braid that would result from the original diagram with the subdividing point @xmath17 included ( see figure 14 ) .
( note that new vertical strands coming from the braiding that may run over or under @xmath58 do not affect the result . )
@xmath61 when we meet a free up arc , which we have the choice of labelling ` u ' or ` o ' , the resulting braid does not depend up to @xmath2equivalence on this choice .
first , we shall assume for simplicity that the sliding triangle of the up arc does not lie over or under any other arcs of the original diagram .
also we assume for definiteness that the up
arc is originally labelled ` o ' .
we complete the braiding by eliminating it .
then on the new almost - horizontal piece of string we take an arbitrarily small neighbourhood @xmath59 of a point @xmath60 that is a projection of an arbitrarily small neighbourhood @xmath62 on the original up arc , and such that there is no other vertical line between the vertical line of @xmath17 and the one of @xmath60 ( figure 15 ) .
we then perform an @xmath18move at @xmath60 .
@xmath63 the fact that the original up arc is free and small enough implies that only vertical strands can pass over or under its sliding triangle . therefore
as @xmath62 is arbitrarily small there is no arc crossing @xmath64 so as to force it be an _ under _ arc .
also , by braid planar isotopy we shift @xmath65 slightly _ higher _ , so as to come to the position where we can undo an @xmath19move ( see figure 16 ) .
we undo it , so the final braid up to a small braid planar isotopy can be seen as the braid that we would have obtained from the original diagram with the free up
arc labelled with ` u ' instead of ` o ' .
@xmath66 notice that , if the original up arc were an ` u ' we would perform an @xmath19move at @xmath60 . to complete the proof of the lemma
we assume that the sliding triangle of our up
arc lies over or under other arcs of the original diagram . in this case
we subdivide it ( using lemma 4.1 ) into arcs small enough to ensure that all the sliding triangles are clear ; we give all new arcs the labelling of the original one .
then we change the labelling of each up arc using the above and , using lemma 4.1 again , we eliminate all the new subdividing points ( figure 17 ) . @xmath67 if we have a chain of overlapping sliding triangles of free up
arcs so that we have a free choice of labelling for the whole chain then , by lemmas 4.1 and 4.2 , this choice does not affect up to @xmath2equivalence the final braid .
if by adding a subdividing point on an up arc we have a choice for relabelling the resulting new up arcs so that the triangle condition is still satisfied then , by lemmas 4.1 and 4.2 , the resulting braids are @xmath2equivalent .
given any two subdivisions , @xmath68 and @xmath69 , of a diagram which will satisfy the triangle condition with appropriate labellings , the resulting braids are @xmath2equivalent . by remark 3.2
( 4 ) this can be easily seen if we consider the subdivision @xmath70 and apply the lemmas above .
corollary 4.5 proves independence of subdivision and labelling for diagrams , and thus we are done with the static part of the proof . using corollary 4.5
we have also proved independence of @xmath4moves related to condition 3 of definition 3.3 ( recall proof of lemma 3.5 ) .
it remains to consider the effect of the rest of the @xmath4moves .
we shall check first the planar @xmath4moves .
these will include the examination of @xmath4moves related to conditions 1 and 2 of definition 3.3 .
now , a @xmath4move can be regarded as a continuous family of diagrams ( see figure 18 ) . from the above we
shall assume that the triangle condition is not violated .
also , by symmetry , we only have to check the @xmath4moves that take place in the first quarter of the plane , above the edge @xmath64 , and w.l.o.g .
@xmath64 is an up arc .
therefore , the @xmath4moves we want to check can split into moves that take place above @xmath64 within the vertical zone defined by @xmath65 and @xmath5 and moves that take place outside this zone .
all resulting braids will be compared with the braid obtained by subdividing @xmath64 at an appropriate point @xmath17 .
i.e. we have : _ ( i ) inside subdivision _ consider the continuous family of triangles with edge @xmath64 and all the new vertices lying on the vertical line of an interior point @xmath17 of @xmath64 ( figure 18 ) .
a new vertex can not aligne horizontally with @xmath5 .
so , if @xmath71 is a new edge before the horizontal position from below , we can see @xmath2equivalence of the corresponding braids by introducing an extra vertex in @xmath64 at @xmath17 and applying lemma 4.1 ( see left - hand side of figure 19 ) .
@xmath72 if @xmath73 has passed above the horizontal position and therefore it becomes a down arc ( cf .
definition 3.3 , condition 1 ) , we introduce an @xmath2move at the point @xmath5 and we show that again we obtain the same braid as if we had originally subdivided @xmath64 at @xmath17 ( see right - hand side of figure 19 ) .
the type of the @xmath2move is over / under according as @xmath64 is an over / under up arc .
@xmath74 note that , if @xmath64 were originally a down arc , then the new edges would pass from down arcs to opposite arcs through the horizontal position . in this case
, according to the way we do the braiding and to the triangle condition , we may need to subdivide further the new opposite arcs and apply lemmas 4.1 and 4.2 . before continuing with the proof
we introduce some extra notation .
we shall denote by @xmath75 the braid obtained by completing the braiding on the arcs @xmath76 and @xmath77 .
suppose now we subdivide @xmath64 at a different point , @xmath78 say .
then , by lemma 4.1 we have @xmath79 , since they both differ by an @xmath2move from @xmath80 . in figure 20
the dotted line indicates a possible braid strand between the vertical lines of @xmath17 and @xmath78 ( cf .
definition 3.3 , condition 2 ) .
@xmath81 _ ( ii ) outside ` subdivision ' _ consider now the continuous family of triangles with edge @xmath64 and all the new vertices lying on the vertical line of a point @xmath17 that lies outside the vertical zone defined by @xmath65 and @xmath5 ( see figure 21 ) . by symmetry
we only need to examine the @xmath4moves taking place on the side of @xmath5 , say .
assume first that @xmath17 is close enough to @xmath5 so that no braid strand passes in between the vertical lines of @xmath17 and @xmath5 .
then obviously @xmath82 ( braid isotopy ) .
reasoning as above , if @xmath60 is a new vertex such that @xmath71 is below the horizontal position , then @xmath83 differs from @xmath75 by an @xmath2move introduced at @xmath5 .
also , if @xmath84 is a new vertex such that @xmath85 is below the horizontal , then @xmath86 differs from @xmath83 by an @xmath2move introduced on @xmath87 at the point @xmath84 .
@xmath88 assume now that the @xmath4move introduces a new vertex @xmath78 so that between the vertical lines of @xmath17 and @xmath78 there is a braid strand ( cf .
definition 3.3 , condition 2 ) . in this case
we see the @xmath2equivalence by introducing and deleting two @xmath19moves ( figure 22 ) . @xmath89 we shall now check the reidemeister moves ( recall figure 3 ) . to check the first two moves we follow similar reasoning as above , but we shall demonstrate it here for completeness .
for the first one , we illustrate below that the braid obtained after the performance of the move is equivalent to the one obtained before the move , up to braid isotopy and one @xmath19move : @xmath90 for the second move we complete the braiding of the left - hand side diagram and we notice that , using braid isotopy , we can undo an @xmath18move .
we then obtain the braid that we would obtain from the right - hand side diagram with label ` u ' for the up arc .
@xmath91 we shall now check the third type of reidemeister moves or ` triple point moves ' . if all three arcs are down arcs then the move is a braid isotopy .
if either of the outer arcs is an up arc then the move is invisible in the braid .
it remains to check the case where both outer arcs are down arcs and the middle arc is an up arc .
but then the following trick changes the situation to the one where all three arcs are down arcs .
@xmath92 we have now proved that the function @xmath54 from isotopy types of link diagrams to @xmath2equivalence classes of braids is well - defined , and the proof of theorem 2.3 is now completed .
@xmath93 the classical markov braid theorem states that : _ isotopy classes of ( oriented ) link diagrams are in 1 - 1 correspondence with certain equivalence classes in the set of all braids , the equivalence being given by the following two algebraically formulated moves between braids in @xmath94 : _ * _ conjugation : _ if @xmath95 then @xmath96 . * _ markov moves or @xmath1moves : _ if @xmath97 then @xmath98 and @xmath99 .
it is clear from remark 2.2 that an @xmath1move is a special case of an @xmath2move .
also , it follows from the proof of theorem 2.3 that conjugation can be realized by a ( finite ) sequence of @xmath2moves .
indeed , let @xmath58 and @xmath100 be two conjugate braids .
then their closures @xmath101 and @xmath102 are isotopic diagrams and so they differ by a sequence of @xmath9 , say , @xmath4moves ; so , for each stage of isotopy we obtain the following sequence of link diagrams @xmath103 which we turn into the braids @xmath104 , say , using our algorithm . from the above ,
every consecutive pair of these braids differs by a finite sequence of @xmath2moves .
now , the way we defined the closure of a braid ( recall figure 6 ) guarantees that @xmath105 is @xmath100 and @xmath13 is @xmath58 .
thus , conjugation is only a redundant ` auxilliary ' move used to bring the @xmath1moves inside the braid box , and therefore theorem 2.3 indeed sharpens the classical result . as a concrete example
, we illustrate below the main instances of the isotopy sequence from the closure of a braid that is conjugate to @xmath58 by an elementary crossing , up to @xmath101 .
@xmath106 \(1 ) the proof of theorem 2.3 may be clearly used as an alternative proof of the classical theorem .
indeed , whenever in the proof appears an @xmath2move we use conjugation to bring it to the position of an @xmath1move .
the same reasoning holds for all markov - type theorems that follow in the sequel .
( 2 ) it should be stressed that theorem 2.3 sharpens the classical result only if we work with open braids .
if we work with closed braids we do not need conjugation in the braid equivalence .
we should also stress that vertical alignment of top vertices is the most important case to check in the proof of theorem 2.3 because it corresponds precisely to conjugation by a generator @xmath10 in @xmath13 . as a consequence of the proof of theorem 2.3 we can prove a relative version of the result . by a _ braided portion _ of a link diagram we mean a finite number of arcs in the link which are all oriented downwards ( so that the braid resulting from our braiding process will contain this braided portion ) .
let @xmath107 , @xmath108 be oriented link diagrams which both contain a common braided portion @xmath5 .
suppose that there is an isotopy of @xmath107 to @xmath108 which finishes with a homeomorphism fixed on @xmath5 .
suppose further that @xmath109 and @xmath110 are braids obtained from our braiding process applied to @xmath107 and @xmath108 respectively .
then @xmath109 and @xmath110 are @xmath2equivalent by moves that do not affect the common braided portion @xmath5 .
assume first that the isotopy from @xmath107 to @xmath108 keeps @xmath5 fixed ; the result then follows immediately from the proof of theorem 2.3 because the braided portion does not participate in the proof .
we shall use some standard pl topology to reduce the general case to this special case .
let @xmath62 denote a relative regular neighbourhood of @xmath5 relative to the ends of the arcs ( i.e. @xmath62 comprises a number @xmath111 say of 3balls each of which contains an arc of @xmath5 as an unknotted subarc ) .
think momentarily of these balls as small balls centred at points @xmath112 .
the isotopy restricted to @xmath113 determines a loop in the configuration space of @xmath111 points in @xmath114 . but
this configuration space is well - known to be simply - connected and therefore we may assume that the isotopy is fixed on @xmath113 . by the regular neighbourhood theorem we may now assume that the isotopy fixes @xmath62 setwise .
now restrict attention to one of the balls .
the isotopy on this ball determines a loop in the space of the pl homeomorphisms of the 3ball . but
@xmath115 of this space is generated by a rotation through @xmath116 about some axis .
we may suppose that this axis is the corresponding unknotted arc of @xmath5 and hence that the isotopy fixes this subarc pointwise .
thus we may assume that the whole of @xmath5 is fixed pointwise .
the result now follows from the special case .
in this section we use the methods of theorem 2.3 to prove a second relative version ( relative to a fixed closed subbraid ) and deduce an analogue of markov s theorem for isotopy of oriented links in knot / link complements in @xmath0 .
this is then used for extending the results to closed connected orientable 3manifolds .
let @xmath117 be the complement of the oriented knot @xmath118 in @xmath0 .
by ` knot complement ' we refer to complements of both knots and links .
using alexander s theorem and the definition of ambient isotopy , @xmath117 is homeomorphic to @xmath119 , where @xmath120 is isotopic to @xmath118 and it is the closure of some braid @xmath5 .
let now @xmath2 be an oriented link in @xmath1 .
if we fix @xmath120 pointwise on its projection plane we may represent @xmath2 unambiguously by the _
mixed link _ @xmath121 in @xmath0 , that is , a link in @xmath0 consisting of the _ fixed part _
@xmath120 and the _ standard part _
@xmath2 that links with @xmath120 ( for an example see figure 27a ) .
let now @xmath107 , @xmath108 be two oriented links in @xmath1 .
it follows from standard results of pl topology that @xmath107 and @xmath108 are isotopic in @xmath1 if and only if the mixed links @xmath124 and @xmath125 are isotopic in @xmath0 by an ambient isotopy which keeps @xmath120 pointwise fixed .
see for example @xcite ; chapter 4 . in terms of mixed diagrams
this isotopy will involve the fixed part of the mixed links only in the moves illustrated in figure 27b .
these shall be called ` extended reidemeister moves ' .
therefore reidemeister s theorem generalizes as follows in terms of mixed link diagrams . two ( oriented ) links in @xmath1 are isotopic if and only if any two corresponding mixed link diagrams ( in @xmath0 ) differ by a finite sequence of the extented reidemeister moves together with the planar @xmath4moves and the reidemeister moves for the standard parts of the mixed links .
we now wish to find an analogue of markov s theorem for links in @xmath1 .
since @xmath120 must remain fixed we need a braiding process for mixed link diagrams that maps @xmath120 to @xmath5 ( and _ not _ an @xmath2equivalent braid or a conjugate of @xmath5 ) .
such braids shall be called _ mixed braids _ ( see figure 28c for an example ) .
also the subbraid of a mixed braid that complements @xmath5 shall be called _ the permutation braid_. the fixed part @xmath120 may be viewed in @xmath0 as the braid @xmath5 union an arc , @xmath129 say , at infinity .
this arc is the identification of the two horizontal arcs containing the endpoints of @xmath5 and thus it realizes the closure of @xmath5 .
( in figure 28a the braid @xmath5 is drawn curved but this is just for the purpose of picturing it . )
let @xmath130 be a small regular neighbourhood of @xmath129 and let @xmath2 be a link in @xmath1 . by general position @xmath2 misses @xmath130 and therefore it can be isotoped into the complement of @xmath130 in @xmath0 . by expanding @xmath130 we can view its complement as the cylinder @xmath131 that contains @xmath5 ( figure 28b ) .
we then apply our braiding in @xmath132 .
this will leave @xmath5 untouched but it will braid @xmath2 in @xmath132 .
finally we cut @xmath120 open along @xmath129 in order to obtain a mixed braid ( figure 28c ) .
an alternative more ` rigid ' proof modifying the braiding of section 3 is given in @xcite,@xcite have counter - clockwise orientation ; then we cut the braid open along a line that starts from the central point and cuts through the closing side of @xmath120 . ] . two oriented links in @xmath0 which contain a common subbraid @xmath5 are isotopic rel @xmath5 if and only if any two corresponding mixed braids @xmath127 and @xmath134 are @xmath2equivalent by moves that do not touch the common subbraid @xmath5 . further any oriented links in @xmath126 are isotopic in @xmath126 if and only if any two corresponding mixed braids @xmath127 and @xmath135 in @xmath0 are @xmath2equivalent by moves that do not touch the common subbraid @xmath5 . consider an isotopy of a link @xmath2 in the complement of @xmath136 . by a general position argument
we may assume that this isotopy only crosses the arc @xmath129 a finite number of times and that these crossings are clear vertical cuts ( figure 29a ) .
the neighbourhood @xmath130 misses also the isotopy apart from those clear crossings . as in the proof above , we expand @xmath130 and thus the isotopy now takes place inside @xmath132 apart from the crossings with @xmath129 .
applying now the proof of theorem 2.3 , the isotopies within @xmath132 can be converted into a sequence of @xmath2moves that do not affect @xmath5 .
therefore it only remains to check @xmath2equivalence at a crossing of @xmath2 with @xmath129 .
but after expanding @xmath130 such a crossing looks like exchanging a braid strand , @xmath137 say , that runs along the front of the mixed braid for one with the same endpoints that runs along the back of the braid ( figure 29b ) . if @xmath137 is an up arc it can be made into a free up arc and so it dissappears after the braiding .
if @xmath137 is a down arc we introduce an @xmath18move at its top part so that between the endpoints of the new strand and the endpoints of @xmath137 there is no other braid strand .
we then undo an @xmath19move as depicted below , and the proof is concluded .
note that it is necessary to use the version of an @xmath2move with a crossing ( recall figure 5 ) .
let @xmath1 be a closed connected orientable ( c.c.o . )
it is well known @xcite , @xcite , @xcite that @xmath1 can be obtained by surgery on a framed link in @xmath0 with integral framings and w.l.o.g .
this link is the closure @xmath136 of a braid @xmath5 .
we shall refer to @xmath136 as the _ surgery link _ and we shall write @xmath140 . moreover by the proof given in @xcite it can be assumed that all the components of the surgery link are unknotted and furthermore that they form the closure of a pure braid .
thus we may assume that @xmath5 is a pure braid .
so @xmath1 may be represented in @xmath0 by the framed @xmath136 and if we fix @xmath136 pointwise we have that links / braids in @xmath1 can be unambiguously represented by mixed links / braids in @xmath0 exactly as discussed in 5.1 .
thus theorem 5.3 holds also in this case . only here
, the braid @xmath5 is framed and we refer to it as the _ surgery braid_. now a link @xmath2 in @xmath1 may be seen as a link in @xmath126 with the extra freedom to slide across the 2discs bounded in @xmath1 by the specified longitudes of the components of @xmath136 .
let @xmath141 be the oriented boundary of a ribbon and let @xmath142 and @xmath143 be two oriented mixed links , so that @xmath143 is the band connected sum ( over @xmath141 ) of a component , @xmath144 , of @xmath107 and the specified ( from the framing ) longitude of a surgery component of @xmath136 .
this is a non isotopy move in @xmath0 that reflects isotopy between @xmath107 and @xmath108 in @xmath1 and we shall call it _ band move_. a band move can be thus split in two steps : firstly , one of the small edges of @xmath141 is glued to a part of @xmath144 so that the orientation of the band agrees with the orientation of @xmath144 .
the other small edge of @xmath141 , which we shall call _ little band _ ( in ambiguity with the notion of a band ) , approaches a surgery component of @xmath136 in an arbitrary way .
secondly , the little band is replaced by a string running in parallel with the specified longitude of the surgery component in such a way that the orientation of the string agrees with the orientation of @xmath141 and the resulting link is @xmath143 .
since we consider oriented links in @xmath1 there are two types , @xmath58 and @xmath6 say , of band moves according as in the second step the orientation of the string replacing the little band agrees ( type @xmath58 ) or disagrees ( type @xmath6 ) with the orientation of the surgery component and implicitely of its specified longitude .
( see figure 31 , where @xmath145 is an integral framing of the surgery component . )
\(1 ) let @xmath142 and @xmath143 represent the isotopic links @xmath107 and @xmath108 in @xmath1 .
if in the isotopy sequence there is no band move involved on the mixed link level , then @xmath107 and @xmath108 are also isotopic in @xmath126 . in particular , the first step of a band move reflects isotopy in @xmath126 .
so , from now on whenever we say ` band move ' we will always be referring to the realization of the second step of a band move . ( 2 ) a band move , that is , the second step of the move described in definition 5.6 , takes place in an arbitrarily thin tubular neighbourhood of the component of the surgery link that contains no other part of the mixed link ; and since @xmath136 is pointwise fixed it follows that this tubular neighbourhood projects the same for any diagram of @xmath143 .
so by ` band move ' we may unambiguously refer to both the move in the 3space and its projection .
the discussion can be summarized by saying that link isotopy in @xmath1 can be regarded as isotopy in @xmath149 together with a finite ( by transversality ) number of band moves .
therefore theorem 5.2 extends to : two ( oriented ) links in @xmath1 are isotopic if and only if any two corresponding mixed link diagrams differ by a finite sequence of the band moves , the extented reidemeister moves and also by planar @xmath4moves and the reidemeister moves for the standard parts of the mixed links .
note now that neither of the two types of band moves can appear as a move between braids ; so in order to state our extension of markov s theorem we modify the band move of type @xmath58 appropriately by twisting the little band before performing the move .
so we have the following . a _ braid band move _ or , abbreviated , a _
b.b.move _ is a move between mixed braids that reflects isotopy in @xmath140 and is described by the following picture ( where the middle stage is only indicative ) .
let @xmath107 , @xmath108 be two oriented links in @xmath1 and let @xmath151 , @xmath152 be two corresponding mixed braids in @xmath0 .
then @xmath107 is isotopic to @xmath108 if and only if @xmath151 is equivalent to @xmath152 by the braid band moves and under @xmath2equivalence that does not affect @xmath5 .
let @xmath153 and @xmath154 be two mixed link diagrams of the mixed links representing @xmath107 and @xmath108 . by theorems 5.8 and 5.5 we only have to check that
if @xmath155 and @xmath156 differ by a band move then @xmath151 and @xmath152 differ by b.b.moves and @xmath2equivalence that does not affect @xmath5 . in @xmath157 the little band would be like depending on the orientation . if the little band is an opposite arc , w.l.o.g .
we may assume that it satisfies the triangle condition .
the algorithm we use ensures that we may assume that @xmath155 and @xmath159 are braided everywhere except for the little band in @xmath155 ( if it is an opposite arc ) and its replacement after the performance of the band move .
this happens because the band move takes place arbitrarily close to the surgery strand ; so we can produce such a zone locally in the braid ( figure 35 ) , and consequently a b.b.move can not create problems with the triangle condition
. moreover the new strand from the band move ( as far as other crossings are concerned ) behaves in the same way as the surgery string itself .
so , whenever we meet other opposite arcs we label them in the same way that we would do if the new strand were missing . * proof .
* we start with the left part of move _ ( i ) _ and we twist the little band ( isotopy in @xmath126 ) using a _ negative _ crossing . then we perform a b.b.move with a _ positive _ crossing and we end up with the right part of move _ ( i ) _ ( see figure 38 ) . *
* we start with the right part of move _ ( ii)_. in the front of the otherwise braided link
we do a twist of the new string using a _ negative _ crossing ( see figure 40 ) .
then we consider the little twisted arc as a little band and we perform another b.b.move around the same surgery component .
this second band move takes place closer to the surgery component than the first one .
now , the shaded region in the picture below is formed by two similar sets of opposite twists of the same string around the surgery string .
so it bounds a disc ( together with the little band that is missing ) , the circumference of which is _ not _ linked with the surgery component ; but this is isotopic in @xmath126 to the left part of move _ ( ii)_. i.e. move _ ( ii ) _ is a finite sequence of @xmath2moves and b.b.moves .
note that in the pictures above we have included another string of the mixed braid that links with the surgery component .
clearly this does not affect the proof .
the proof of theorem 5.10 is now concluded .
\(1 ) the proof holds even if the surgery braid is not a pure braid . in this case a b.b.move is modified so that the replacement of the little band links only with one of the strings of the same surgery component and runs in parallel to all remaining strings of the surgery component .
\(2 ) an analogue of reidemeister s theorem for c.c.o .
3manifolds is given in @xcite and an analogue of markov s theorem for c.c.o .
3manifolds is given in @xcite , both using the intrinsic structure of the manifold .
( 3 ) theorems 5.5 and 5.7 above are described only geometrically . as shown in @xcite , @xcite
if our manifold is the complement of the unlink or a connected sum of lens spaces then in the set of mixed braids we have well - defined group structures , while in the other 3manifolds we have coset structures , which enable us to give the above theorems an algebraic description .
in particular , if our space is a solid torus ( i.e. the complement of the unknot ) or a lens space @xmath169 the related braid groups are the artin groups of type @xmath170 ( for details and jones - type knot invariants we refer the reader to @xcite , @xcite , @xcite ) .
s.l . : 16 mill lane , dpmms , university of cambridge , cambridge cb2 1sb , u.k . ,
e - mail [email protected] and sfb 170 , ` geometrie und analysis ' , bunsenstrasse 3 - 5 , 37073 gttingen , germany . c.p.r . : mathematics institute , university of warwick , coventry cv4 7al , u.k .
, e - mail [email protected] | in this paper we first give a one - move version of markov s braid theorem for knot isotopy in @xmath0 that sharpens the classical theorem .
then a relative version of markov s theorem concerning a fixed braided portion in the knot .
we also prove an analogue of markov s theorem for knot isotopy in knot complements .
finally we extend this last result to prove a markov theorem for links in an arbitrary orientable 3manifold . 6.4 in 8.5 in
[ section ] [ th]lemma [ th]corollary [ th]proposition [ th]criterion [ th]algorithm [ th]definition [ th]conjecture [ th]example [ th]problem [ th]remark [ th]remarks [ th]note [ th]claim [ th]summary [ th]case [ th]acknowledgment = cmcsc10 = cmti8 = cmbx8 = " 1218 | + = # 1by#2(#3 ) to # 2 ' '' '' width # 1 height 0pt depth 0pt |
we begin by introducing the quantities studied below .
a necessary ingredient in engineering a fractional chern insulator is a flat , topologically non - trivial band , defined as follows .
let @xmath1 be a tight - binding orbital localized at position @xmath2 ; the fourier transform of the @xmath3th basis orbital ( where @xmath3 ranges from @xmath4 to @xmath5 ) is @xmath6 is a crystal momentum restricted to the first brillouin zone ( bz ) and @xmath7 is the number of unit cells in the system , which are indexed by lattice vectors @xmath8 .
eigenstates of the tight - binding hamiltonian are bloch functions @xmath9 indexes the bands . at a fixed @xmath10 ,
the tight - binding hamiltonian is an @xmath11 matrix with entries @xmath12 and band energies @xmath13 . in general , for @xmath14 , neither @xmath15 nor @xmath16 will have the periodicity of the reciprocal lattice , and this property is not needed to define the berry curvature via eq .
( [ eq_curvature_def ] ) below . imposing this periodicity by hand has led to demonstrably incorrect calculations in previous literature .
we clarify this point with a discussion in supplementary note 1 and illustrate the consequences of incorrect computations in supplementary figs . 1 , 2 .
nontrivial topological order in a band @xmath17 is indicated by a non - vanishing value of the ( first ) chern number @xmath18 where @xmath19 is the area of the momentum - space brillouin zone , @xmath20 denotes the average over the bz , normalized so @xmath21 , and the berry curvature@xcite of the band @xmath17 is defined as @xmath22 the leading - order condition ( in terms of a long - wavelength expansion ) for the existence of an isomorphism between the band - projected density operators and the gmp algebra found in ref .
is that the berry curvature should be constant as a function of @xmath10 . in the results we report here
, we quantify fluctuations of berry curvature over the bz by their root - mean - square ( rms ) value , @xmath23 we normalize @xmath24 in the same way as the chern number , so that ( [ eq_sigmac_def ] ) is dimensionless as well as insensitive to the scales over which deviations from the mean curvature occur . the higher - order conditions obtained ref .
involve the pull - back of the fubini - study metric on hilbert space@xcite , which we refer to below as the quantum metric . in terms of bloch functions ,
it is given by @xmath25 .
\nonumber\end{aligned}\ ] ] the next - to - leading order condition of ref .
is that the quantum metric also be constant over the bz .
we adopt @xmath26 as the appropriate generalization of rms fluctuation to tensor quantities .
the final constraint on the band geometry is that @xmath27 it was shown in ref . that the left - hand side of ( [ eq - detineq ] ) is always nonnegative ; the condition that it vanishes is equivalent to the condition that @xmath28 and @xmath29 are the real and imaginary components of a khler metric @xmath30 .
this means that , unlike the first two conditions , the metric determinant inequality @xmath31 measures deviations from lowest landau level physics , specifically .
analogous conditions may be derived for higher landau levels .
a stronger condition can be obtained by considering the trace of the quantum metric instead .
it was additionally shown in ref . that @xmath32 in supplementary note 2 we show that if this inequality is saturated , the quantum metric is isotropic and @xmath33 must vanish .
hence the condition @xmath34 is equivalent to requiring that the algebra of band - projected density operators be identical to the gmp algebra , while @xmath35 merely requires that they be isomorphic . the purpose of the present work is to investigate the degree to which the above criteria are satisfied in several fci models known to exhibit fqh - like phases@xcite . in this section ,
we outline procedures common to all models studied .
the stability of an fci phase is trivially influenced by the dispersion of the occupied band .
fortunately , the dispersion of a band is independent of its berry curvature and quantum metric : the former only depends on the hamiltonian s spectrum while the latter depend only on its eigenvectors .
this allows us to eliminate any dispersion - related confounding effects by energetically flattening the bands of each lattice model , which is equivalent to smearing nearest - neighbor hopping terms over an exponentially - localized area@xcite .
differences between chern bands and landau levels also enter in the form of the hamiltonian s interaction term .
unlike the energetic considerations , this dependence is still poorly understood , so we have limited the scope of the present paper to on - site repulsive interactions only ( which necessitates bosonic statistics ) , since this is the lattice interaction which most closely matches the isotropy present in the continuum . for each lattice model considered , we therefore investigate the bosonic laughlin state at filling fraction @xmath36 ( stabilized by a two - body delta - function interaction ) and the bosonic moore - read state at @xmath37 ( stabilized by a three - body delta - function interaction ) .
these states have completely different topological orders ; furthermore the laughlin state is known to be more robust in general than the moore - read state , so examining both provides a useful probe of the sensitivity of band - geometric arguments .
the band geometry hypothesis predicts that the most important factor will be berry curvature fluctuations .
low curvature fluctuations were heuristically identified as a desirable criterion in the earliest fci literature@xcite , which has been well established by subsequent work ( see in particular refs . ) . in the present work we therefore focus on the sub - leading conditions , namely the influence the quantum metric has on the gap .
fluctuations of the quantum metric are predicted to be the next most relevant quantity , but in the models examined this was found to have a high degree of linear correlation with the berry curvature fluctuations ( see supplementary fig .
there is no a priori reason this should be the case : other metric - derived quantities were found to be largely independent of berry curvature .
in addition , we found that the trace inequality ( [ eq - trineq ] ) was far more correlated with the gap than the determinant inequality ( [ eq - detineq ] ) for all models examined , despite corresponding to a stronger condition on the algebra of density operators .
these findings go beyond the scheme laid out in ref . .
due to space constraints , we present data on the dependence of the gap on the determinant condition for low values of @xmath24 in supplementary fig . 4 for the kagom lattice model and supplementary fig . 5 for the ruby lattice model .
the first chern insulator model was introduced by haldane@xcite , who considered a tight - binding model on the honeycomb lattice with nearest- and next - nearest - neighbor hoppings ( fig .
[ fig - haldane]a ) and a peierls phase due to non - uniform threading of magnetic flux through each hexagon .
the single - particle hamiltonian for the haldane model is @xmath38 where the pauli sigma matrices act on the band index , @xmath39 , @xmath40 and @xmath41 is interpreted cyclically mod 3 .
the lower band has a nonzero chern number when @xmath42 .
this model has been extensively studied numerically , and the addition of short - ranged repulsive interactions has been shown to yield both the bosonic@xcite and fermionic@xcite laughlin states at appropriate filling fractions . for purposes of comparison with refs . and
, we consider the model at @xmath43 .
the energy spectrum has band crossings for these parameters , meaning that the bands can not be flattened by local operators and the model analyzed in those references is not adiabatically connected to ( [ eq - h - haldane ] ) . in practice
, however , one is most interested in the @xmath44 subspace ; the hamiltonian then depends only on the combination @xmath45 , and an increase in @xmath46 which removes the crossing may then be compensated by a shift in @xmath47 which leaves the hamiltonian ( [ eq - h - haldane ] ) unchanged up to a scale .
the momentum dependence of the berry curvature and quantum metric is shown in fig .
[ fig - haldane]b for parameters which minimize @xmath24 and maximize the gaps for the laughlin state of bosons and fermions ; these values are listed in supplementary table 1 .
we see that the distribution of berry curvature minimizing @xmath24 interpolates between that which maximizes the gap for the bosonic and fermionic laughlin states , and likewise the value of @xmath47 minimizing @xmath24 lies between the values minimizing the gaps ( fig .
[ fig - haldane]e , f ) .
the band geometry argument does nt distinguish the statistics of the underlying particles , due to the fact that the projected density operators are bilinear in particle operators and bosonic in either case .
the @xmath48 parameter space may be sampled exhaustively ; many - body gaps for the bosonic and fermionic laughlin states are shown over the full topologically non - trivial region of parameter space in figs .
[ fig - haldane]c , [ fig - haldane]d .
the largest gaps and most uniform band geometry both occur for @xmath44 .
are shown in red ; basis elements are shown by differently colored / numbered sites .
hopping elements are shown with black and blue edges ; arrowheads indicate the chirality convention for complex hoppings .
( * b * ) band geometry over the reciprocal lattice unit cell spanned by @xmath49 , for parameter values maximizing the gap and minimizing @xmath50 .
axes of ellipses are proportional to the eigenvectors of the quantum metric @xmath51 , and ellipse color is given by the relative deviation of berry curvature @xmath52 from its brillouin zone - averaged value .
( * c * ) gap @xmath0 as a function of @xmath53 for @xmath54 bosons at @xmath55 with an on - site repulsion .
( * d * ) gap as a function of @xmath53 for @xmath54 fermions at @xmath56 with nearest - neighbor repulsion .
note that this plot differs from fig .
8 of ref . because we exclude ( light gray ) parameters failing to meet energetic and entanglement - based criteria for laughlin - type order .
( * e * ) berry curvature fluctuations @xmath50 ( left scale ) and metric fluctuations @xmath57 and average trace inequality @xmath58 ( right scale ) as functions of @xmath47 at @xmath44 .
( * f * ) reproduction of the gap data from panels ( * c * ) , ( * d * ) along @xmath44 . ] the other band - geometric criteria are highly correlated with the curvature fluctuation @xmath24 and yield little new information for this model ( fig .
[ fig - haldane]e ) , beyond being close to the location of the maximum gaps ( fig .
[ fig - haldane]f ) .
we prove in supplementary note 3 that the remaining band - geometric criterion , the determinant condition ( [ eq - detineq ] ) , is necessarily saturated for any two - band model , but the trace condition ( [ eq - trineq ] ) remains nontrivial here .
although the haldane model at fractional filling exhibits a robust laughlin state , its berry curvature remains highly nonuniform even in the best case ( fig .
[ fig - haldane]b ) .
we would like to be able to compare this model with the kagom and ruby lattice models , in which more uniform curvature may be achieved .
in addition , since we re interested in the sub - leading effects the quantum metric has on the gap , we want to examine band configurations for which the metric is more independent of the berry curvature than in fig .
[ fig - haldane]e .
this may be accomplished by adding a third - nearest - neighbor hopping term , with independent coupling @xmath59 , to the haldane model hamiltonian ( [ eq - h - haldane ] ) . in the previous section
, we saw that a sub - lattice chemical potential @xmath60 always reduces the gap of an fci phase , so we set @xmath44 below .
the remaining couplings in the new hamiltonian may be parameterized by @xmath61 and @xmath62 , so the interesting region of parameter space is still two - dimensional .
the momentum dependence of the berry curvature and quantum metric is shown in fig .
[ fig - haldane - t3]a for parameters which minimize @xmath24 and maximize the gaps for the laughlin and moore - read states ; these values are given in supplementary table 2 .
comparison with fig .
[ fig - haldane]b shows that curvature fluctuations have been reduced ; furthermore , the minimum of @xmath24 ( shown in fig . [ fig - haldane - t3]b ) occurs at different parameter values than the minimum of the trace condition ( shown in fig .
[ fig - haldane - t3]c ) .
gaps for the laughlin and moore - read states are shown in figs .
[ fig - haldane - t3]d and [ fig - haldane - t3]e , respectively ; the maximum gaps in both cases occur at lower values of @xmath63 than the minimum value of @xmath50 , which one may attribute to the influence of @xmath64 . , for parameter values maximizing the gaps and minimizing @xmath50 .
axes of ellipses are proportional to the eigenvectors of the quantum metric @xmath65 .
ellipse color is given by the relative deviation of berry curvature @xmath52 from its brillouin zone - averaged value .
( * b * ) berry curvature fluctuations @xmath50 and ( * c * ) brillouin zone - averaged trace inequality @xmath58 as a function of couplings @xmath66 , where we set @xmath67 and @xmath68 in this and remaining panels without loss of generality .
( * d * ) gap @xmath0 as a function of couplings for the bosonic laughlin state of @xmath54 bosons at @xmath55 .
( * e * ) gap as a function of couplings for the bosonic moore - read state of @xmath69 bosons at @xmath70 . ] a chern insulator defined on the kagom lattice was introduced by tang , mei and wen @xcite ( fig .
[ fig - kagome]a ) .
this model is attractive for our purposes since it has three bands , while remaining structurally similar to the haldane model . defining a complex hopping matrix element for the relative embedding of the sublattices in the unit cell as @xmath71 where @xmath72 is the unit matrix whose @xmath73th entry is equal to @xmath4 , the momentum space hamiltonian for
the kagom lattice model is @xmath74 \nonumber \\ { } & \qquad + \text{h.c . } \end{aligned}\ ] ] where @xmath75 is an abbreviation for the hermitian conjugate , @xmath40 , and @xmath76 is interpreted cyclically mod 3 .
the relative offsets @xmath77 are as depicted by the numbered sites in fig .
[ fig - kagome]a .
the momentum dependence of the berry curvature and quantum metric is shown in fig .
[ fig - kagome]b for parameters minimizing @xmath24 and maximizing the gaps for the laughlin and moore - read states ; these values are listed in supplementary table 3 .
are shown in red ; basis elements are shown by differently colored / numbered sites .
hopping elements are shown with black and blue edges ; arrowheads indicate the chirality convention for complex hoppings .
( * b * ) band geometry over the reciprocal lattice unit cell spanned by @xmath49 , for parameter values minimizing @xmath50 and maximizing gaps , respectively .
axes of ellipses are proportional to the eigenvectors of the quantum metric @xmath51 .
ellipse color is given by the relative deviation of berry curvature @xmath52 from its brillouin zone - averaged value . ] .
( * b * ) the same data as a one - dimensional submanifold in band geometry space , which we parameterize in terms of berry curvature fluctuations @xmath24 and the averaged trace condition @xmath58 .
the upper set ( larger gap @xmath0 ) of points are gaps for the @xmath54 bosonic laughlin state at @xmath55 , while the lower are for the bosonic moore - read state at @xmath70 .
( * c * ) gap for the kagom lattice model with both nearest - neighbor and next - nearest - neighbor couplings , as a function of band - geometric parameters @xmath78 for the laughlin state of @xmath54 bosons at @xmath55 .
( * d * ) gap as a function of @xmath78 for the moore - read state of @xmath69 bosons at @xmath70 , for the same coupling values . ] . *
( * a * ) gaps for the bosonic laughlin state of @xmath54 bosons at @xmath55 , as a function of the brillouin zone average of the trace condition @xmath58 .
we only plot gaps for parameter values which have berry curvature fluctuations @xmath24 less than twice its minimum value . ( * b * ) gap of the bosonic laughlin state vs. @xmath58 for parameter values randomly chosen on isosurfaces of constant @xmath50 in the space of couplings .
the parameter space sampling procedure used to obtain these sets of points is described in the methods .
( * c * ) , ( * d * ) the same , for the bosonic moore - read state of @xmath69 bosons at @xmath70 .
note that the same sets of model parameters are used in each column . ]
we first consider the model with nearest - neighbor hoppings only ( figs .
[ fig - kagome - gaps]a , b ) . because band geometry is determined by single - particle quantities ,
the values of @xmath79 for a given value of the effective nearest - neighbor ( nn ) coupling @xmath80 are identical for the laughlin and moore - read states . despite having different topological orders , gaps for both states decline monotonically
as one proceeds from the region of minimum @xmath50 and @xmath64 , and this trend holds over the entire phase ( i.e. , up to the closure of the gap . ) to establish that the above trends were not coincidental , we studied the entire @xmath81 phase containing the nn - only point @xmath82 ; @xmath83 for the bosonic laughlin ( fig .
[ fig - kagome - gaps]c ) and moore - read ( fig .
[ fig - kagome - gaps]d ) states .
the gap s sensitivity to band geometry is most apparent for the more fragile moore - read state : the state is only stable in a small region , with the largest gaps ( white points ) attained at parameters with the lowest values of @xmath24 and @xmath64 found .
the fact that the region of stability is an arc , rather than a vertical line , demonstrates that @xmath64 describes independent , non - negligible factors influencing the stability of this state .
these phenomena are less evident in the laughlin state ( fig .
[ fig - kagome - gaps]c ) , which remains stable over a wide range of parameter values . because quantum metric - dependent quantities enter at a higher order than berry curvature fluctuations in the band geometry analysis
, they should have a subdominant effect on the gap . in fig .
[ fig - kagome - shells ] we control for the effects of large curvature fluctuations by restricting attention to parameter values giving near - minimal values of @xmath50 . including all such parameters
yields a one - way relationship for the laughlin and moore - read states ( figs .
[ fig - kagome - shells]a , c ) , in the sense that large gaps are obtained only at low values of @xmath64 , but small gaps can be obtained at any value of @xmath64 .
further detail is evident if we take parameter values chosen to give the same value of @xmath24 ( see the methods for a description of the sampling procedure used . )
results for sets of points chosen to have four different values of @xmath24 are shown in figs .
[ fig - kagome - shells]b , d for the laughlin and moore - read states . removing the variation in @xmath50 reveals a full - fledged negative correlation between @xmath64 and the gaps :
the trend is approximately linear for models having curvature fluctuations near the minimum and becomes less so as curvature fluctuations are allowed to increase .
hu , kargarian and fiete @xcite described a chern insulator model on the ruby lattice ( fig .
[ fig - ruby]a ) . in the limit of total spin polarization
, they showed that hopping parameters could be chosen such that the lowest band had @xmath84 and a bandgap to bandwidth ratio of @xmath85 .
the hamiltonian for this model is @xmath86 where @xmath87 is defined in eq .
( [ eq_hopping_def ] ) . here
@xmath40 and the index on @xmath88 is interpreted cyclically mod 3 , but the indices on @xmath87 are interpreted cyclically mod 6 .
the relative offsets @xmath77 are as depicted by the numbered sites in fig .
[ fig - ruby]a .
we considered the @xmath84 phase containing the flat - band point found in ref . , with @xmath89 , @xmath90 , and @xmath91 .
momentum dependence of the berry curvature and quantum metric is shown in fig .
[ fig - ruby]b for parameters which minimize @xmath24 and maximize the gaps for the laughlin and moore - read states ; these values are listed in supplementary table 4 . remarkably , the complexity of this hamiltonian works in our favor : one can find parameter values which greatly reduce the fluctuations in band geometry relative to the kagom lattice model , which means that the ruby lattice model may be tuned to produce a much closer approximation to lowest landau level physics . as a consequence ,
trends identified in the kagom lattice model are manifest here with a much higher degree of correlation
[ fig - ruby]c , d show that the gaps of the laughlin and moore - read states are strongly correlated with band geometry as measured by both @xmath24 and @xmath64 . in both cases
the gap can be seen to decrease with increasing @xmath64 , even for the same values of @xmath24 ; in particular , the moore - read state is only stable in the lower right half of the plot area . restricting our attention to parameters yielding small fluctuations in berry curvature , in fig .
[ fig - ruby - shells ] we display data analogous to that presented for the kagom lattice model in fig .
[ fig - kagome - shells ] . again
, if we only impose an upper bound on @xmath24 we see that it s not possible to have a large gap for large values of @xmath64 ( figs .
[ fig - ruby - shells]a , c ) , while upon restricting to specific values of @xmath24 ( figs . [ fig - ruby - shells]b , d ) we see that the negative correlation between @xmath64 and the gaps is even tighter , becoming nearly linear for low @xmath24 and gaining more scatter as @xmath24 is increased .
are shown in red ; basis elements are shown by differently colored / numbered sites .
hopping elements are shown with black ( @xmath92 ) , blue ( @xmath93 ) and brown ( @xmath94 ) edges ; arrowheads indicate the chirality convention for complex hoppings .
( * b * ) band geometry over the reciprocal lattice unit cell spanned by @xmath49 , for parameter values minimizing @xmath50 and maximizing gaps , respectively .
axes of ellipses are proportional to the eigenvectors of the quantum metric @xmath51 .
ellipse color is given by the relative deviation of berry curvature @xmath52 from its brillouin zone - averaged value .
( * c * ) gap @xmath0 as a function of the average berry curvature fluctuation @xmath24 and averaged trace condition @xmath58 for the laughlin state of @xmath54 bosons at @xmath55 .
( * d * ) gap as a function of @xmath95 for the moore - read state of @xmath69 bosons at @xmath70 . ] . *
( * a * ) gaps for the bosonic laughlin state of @xmath54 bosons at @xmath55 , as a function of the brillouin zone average of the trace condition @xmath58 .
we only plot gaps for parameter values which have berry curvature fluctuations @xmath24 less than twice its minimum value . ( * b * ) gap of the bosonic laughlin state vs. @xmath58 for parameter values randomly chosen on isosurfaces of constant @xmath50 in the space of couplings .
( * c * ) , ( * d * ) the same , for the bosonic moore - read state of @xmath69 bosons at @xmath70 .
note that the same sets of model parameters are used in each column . ] in this section we describe two approaches to quantifying the degree of correlation between the band geometry and the many - body gaps described above . in particular , the fact that fqh - type states are destabilized by fluctuations in berry curvature is readily apparent and was anticipated in the first work on fcis ; in the present work we are interested in possible additional dependence on conditions derived from the quantum metric , so we seek to measure correlation between the gap and the trace condition @xmath64 .
this is not fully straightforward , due to correlation of @xmath64 and @xmath50 with each other , as evidenced by the fact that parameter values are not uniformly distributed in figs .
[ fig - kagome - gaps]c and [ fig - ruby]c .
one approach is to compare data for parameters yielding the same value of @xmath50 , some of which is shown in figs .
[ fig - kagome - shells]b , d and [ fig - ruby - shells]b , d ( see the methods for the procedure used to sample from isosurfaces of constant @xmath50 ) .
we do not have quantitative predictions for the functional dependence of the gap on any band - geometric quantity , so to avoid introducing assumptions we use spearman s @xmath96 as a nonparametric measure of correlation .
this is defined as the linear ( pearson ) correlation coefficient between the rankings of the data points when rank - ordered by @xmath0 and by @xmath64 , and takes values ranging from @xmath97 for any monotonically increasing function to @xmath98 for any monotonically decreasing function . for the correlation between gap and averaged trace condition @xmath58 on separate samples of @xmath99 points sampled from isosurfaces of constant root - mean - square berry curvature fluctuation @xmath50 .
a value of @xmath100 corresponds to perfectly monotonic anti - correlation .
we show results for laughlin ( dark blue ) and moore - read ( light blue ) states in ( * a * ) the kagom lattice model and ( * b * ) the ruby lattice model .
error bars show the bootstrapped standard error . ] in fig .
[ fig - correls ] we plot the results of this test for the laughlin and moore - read data in the kagom and ruby lattice models .
this score was found to be within @xmath101 of the linear correlation between @xmath64 and @xmath0 for each isosurface , which indicates the robustness of our conclusion and implies that all values of @xmath24 considered lie in a weak - fluctuation regime .
the general trend evident in fig .
[ fig - correls ] is that @xmath0 is highly negatively correlated with @xmath64 when curvature fluctuations are constrained to be near their minimum value .
as curvature fluctuations are allowed to increase , this relationship becomes less exact , but converges to an asymptotic value well above 0 for both the laughlin and moore - read states .
this confirms the qualitative picture evident in figs .
[ fig - kagome - shells]b , d and [ fig - ruby - shells]b , d. alternatively , we can analyze the larger set of data having unrestricted values of @xmath24 , at the cost of assuming linear relationships between all variables ; the similarity between the results for spearman s and pearson s @xmath96 mentioned above suggest that this is justified .
one can then compute the partial correlation , denoted here by @xmath102 , as the degree of linear correlation between the residuals of @xmath0 and @xmath64 , after first subtracting the best - fit linear dependence of each on @xmath50 ; similarly , this score ranges from @xmath97 to @xmath98 . for the augmented haldane , kagom and ruby lattice models , respectively , we find @xmath103 and @xmath104 for the laughlin state and @xmath105 and @xmath106 for the moore - read state .
sample sizes were , respectively , @xmath107 and @xmath108 .
the fact that these values are lower than those obtained by the isosurface method describes the non - negligible correlation between @xmath50 and @xmath64 .
all correlation scores quoted above and shown in fig . [ fig - correls ] are statistically significant at the @xmath109 level ( at most ) .
this is the main result of this section : gaps for both states are independently sensitive to variations in the trace condition , beyond the correlations induced in the latter via curvature fluctuations .
this confirms that band geometry plays a significant role in realistic fci models .
the band geometry hypothesis claims that the most stable fqh - like phases are obtained when the band geometry is tuned to be as close to that of a landau level as possible .
we have shown above that this holds for several different lattice models as their couplings are varied , but comparison of the models shows an apparent contradiction : the gaps reported above are smallest for the ruby lattice model , despite the fact that this model can be made to approximate landau level physics more closely than the other models studied here .
the resolution of this apparent contradiction lies in the fact that the models considered have different numbers of tight - binding sites per unit cell .
this factor enters into the interaction term of the hamiltonian , and hence the gap : because we have flattened the dispersion of the kinetic term , the interaction strength is the only energy scale in the problem . in supplementary note
4 we give a scaling argument that the strength of a delta - function interaction in the continuum should be multiplied by a factor of @xmath5 when discretized to an on - site repulsion in a model with @xmath5 bands .
this means that , when comparing the gaps of bosonic laughlin states in different models , the appropriate quantity to compare is @xmath110 .
for the three - body delta - function interaction which stabilizes the bosonic moore - read state , the same considerations yield a scaling factor of @xmath111 . in fig .
[ fig - crossmodel ] we compare the gaps for the laughlin state for the augmented haldane , kagom and ruby lattice models when scaled by this factor , as a function of the band - geometric parameters @xmath79 .
the domain of each plot is restricted to be the common overlap of the three models in @xmath79 space . because of this restriction , we were unable to perform a meaningful analysis for the moore - read state , as it is unstable in most of this region ( compare figs .
[ fig - kagome - gaps]d , [ fig - ruby]d ) .
despite the fact that the scaling argument is exact only in the large-@xmath5 limit and the models we compare have @xmath112 and @xmath113 , we find roughly similar behavior across all three models , both in terms of the magnitude of the scaled gap and of its dependence on @xmath50 and @xmath64 .
in this work we have presented numerical results which systematically map out the robustness of topologically ordered fqh - like phases in fci systems with realistic , short - ranged hamiltonians and band geometry which is less than perfectly uniform .
we presented quantitative evidence that the band - geometric quantities identified in ref .
remain strongly correlated with the size of the gap even when its conditions on band geometry are not met exactly .
this leads us to propose a geometric stability hypothesis for fqh - like phenomena in fcis : in spite of the fact that the gmp algebra is not perfectly reproduced , we conjecture that an approximate version of the single - mode approximation correctly describes the low - energy physics of these fci models . as a practical corollary , we predict that single - particle hamiltonians with more uniform band geometry specifically , as measured by the hierarchy of three criteria will produce more stable fqh - like states .
the validity of the single - mode approximation in fcis has been investigated by a number of other authors using approaches complementary to that taken here@xcite . within the context of their hamiltonian approach to the fqhe
, murthy and shankar showed that composite fermion degrees of freedom could be chosen which reproduce an exact version of the gmp algebra@xcite . more directly , in ref .
, the lowest - lying neutral excitation of the kagom and ruby lattice models was found to be well described by the magnetoroton mode of the corresponding fqh state on a torus , using a phenomenological mapping between the fqh and fci hilbert spaces described in ref . .
a related mapping was originally proposed by qi@xcite , but the image of fqh pseudopotential interactions under this mapping is not well localized and strongly anisotropic@xcite , making the relationship to physical fci hamiltonians unclear . comparing the results from the kagom and ruby lattice models , it appears easier to engineer uniform geometry in more complicated hamiltonians , both in the sense of having more tunable couplings and in the sense of having more bands .
the latter property is expected to hold on general grounds , as noted in refs . .
increasing the size of the unit cell reduces the effectiveness of a fixed - strength repulsion , however , so an optimal choice would balance these two factors .
this has immediate relevance to experimental design : laboratory hamiltonians are necessarily more complicated than those in idealized theoretical models ( e.g. , the proposal in ref . involves an eight - dimensional parameterization of the applied electric field used to obtain a synthetic gauge potential . ) performing many - body simulations on a representative set of parameters in such a large space is prohibitively time - consuming ; the geometric stability hypothesis can be used to reduce this to a manageable subspace .
in addition , band geometry may , by definition , be tuned independently of energetic considerations such as the bandwidth . a pressing direction for future work is to further develop the band geometry hypothesis by investigating its validity in less straightforward scenarios : stable fci states where the berry curvature is not particularly uniform
have also been proposed ; furthermore , the stability of the state may also depend on the filling fraction and the particular state sought to be stabilized@xcite . among other aspects which would be interesting to clarify are the role of bosonic versus fermionic particle statistics , nn- and longer - ranged inter - particle interactions , anisotropic interactions induced by the lattice structure , etc .
in particular , the distribution of geometric quantities differs from band to band ; this can be selected by fully filling a number of bands in a fermionic system , which could lead to new phenomena .
one could also consider the wide range of more elaborate fci models in the literature , possessing , e.g. , chern numbers @xmath114 , non - abelian berry curvature arising from multiple degenerate bands , etc .
we direct the reader to the recent reviews@xcite for a more extensive discussion and bibliography .
we study the dependence of the many - body gap on band - geometric quantities in the haldane model @xcite and models proposed for the kagom @xcite and ruby @xcite lattices .
the parameter space for the haldane model hamiltonian is small enough that the vector of parameters @xmath115 which maximize the gap @xmath0 may be found from exhaustive sampling .
the parameter spaces for the other two models are higher - dimensional and non - compact , so this strategy will not work . according to the geometric stability hypothesis ,
the berry curvature fluctuations @xmath50 should be inversely correlated with the gap , meaning that it may be employed as a proxy for the latter which requires far less effort to compute . in supplementary note 5
we derive an expression for the derivative of @xmath50 with respect to any parameter appearing in the single - particle hamiltonian , allowing steepest - descent methods to be employed .
of course , we do not expect the vector of parameters @xmath116 minimizing @xmath50 to precisely coincide with those maximizing the gap , but if the single mode approximation is applicable , @xmath116 will be a viable initial guess for @xmath115 , to be refined as described below .
indeed , a state with a robust topological gap in the presence of large curvature fluctuations would be of immediate interest as an fci phase not describable in terms of an fqh universality class . in the neighborhood of @xmath116 , a surface of constant @xmath50 in parameter space will be approximated by an ellipsoid given by the hessian of @xmath117 at @xmath116 , which may be calculated by the method in supplementary note 5 .
we sample points uniformly from the surface of the ellipsoid defined by the hessian by the well - known method of projecting vectors of parameters @xmath118 sampled from the corresponding multinormal distribution .
we then perturb this ellipsoid by shifting the vectors @xmath118 radially along the rays connecting each with @xmath116 until we find parameters @xmath119 such that @xmath120 is equal to the target value .
these isosurfaces of different parameters with the same value of @xmath50 permit us to study the sub - leading effects of the quantum metric on the gap predicted by the geometric stability hypothesis : in fig .
[ fig - kag - samples ] we depict the isosurfaces found for the kagom lattice model which were used to generate the data shown in figs .
[ fig - kagome - shells]b , d. finally , we locate the parameters giving the maximum gap by fitting a quadratic form @xmath121 to the gaps from the isosurface data and sampling new parameters from a gaussian distribution centered on @xmath122 with covariance matrix @xmath123 .
we have verified that deviations of the actual @xmath115 found from this data from the fitted value @xmath122 are negligible compared to the scales set by @xmath123 , hence this search does not need to be iterated further .
of the kagom lattice model are marked by the large , central black dot .
the large red dot identifies parameter values which yielded the maximum gap for the laughlin state of @xmath54 bosons at @xmath55 .
concentric shaded surfaces are isosurfaces upon which @xmath24 takes a constant value equal to @xmath124 and @xmath125 times its minimum value ( respectively ) .
blue contours on the walls of the box are sections through these surfaces at the minimum @xmath50 point . small grey points
identify random samples taken from the @xmath126 isosurface , to illustrate uniformity . ]
unless explicitly identified otherwise , all data were obtained from exact diagonalization of the many - body hamiltonian for @xmath54 bosons interacting with a two - body on - site repulsion at a filling fraction ( ratio of the number of particles to the number of unit cells ) of @xmath55 on a periodic lattice of @xmath127 unit cells ( for the laughlin - like fci state ) or @xmath69 bosons at a filling fraction of @xmath70 on a lattice of @xmath128 unit cells , interacting with a three - body on - site repulsion only ( for the moore - read - like fci state ) . in the main text we also presented additional data obtained for the fermionic laughlin - like state in the haldane model with @xmath54 particles at @xmath56 on a @xmath129 lattice with two - body nearest - neighbor repulsion .
the fact that the bosonic moore - read state is realized in a filled landau level means that , relative to the laughlin state , we are able to simulate more particles with a many - body hilbert space of roughly the same size . flattening the spectrum of the single - particle hamiltonian
removes the only other energy scale from the problem , so the strength of the repulsive interaction sets the units of the many - body gap @xmath0 . because we do simulations for thousands of parameter values ,
a detailed finite - size scaling analysis of each is beyond the scope of this paper .
we have , however , made spot checks by doing simulations of @xmath130 to @xmath131 particles at select parameter values .
representative data is shown in supplementary figure 6 . from this analysis ,
we have concluded that the system sizes described above strike an acceptable balance between accurately approximating the thermodynamic ( @xmath132 ) limit and making the volume of simulations manageable on the computational resources available to us .
we employ several criteria to identify topological order in the results of numerical simulations , employing information from both the many - body energy spectrum and the many - body wavefunctions themselves .
we first require that the spectrum has the correct number of degenerate ground states in the proper momentum sectors ( two states at @xmath133 for bosonic laughlin - type order , and three states at @xmath134 and @xmath135 for bosonic moore - read - type order ) . in a finite - size simulation
these ground states will only be approximately degenerate ; we require that the gap to the lowest - energy excited state ( out of all momentum sectors ; i.e. the indirect gap ) be at least as large as the spread in ground state energies .
we ascertain whether a ground state wavefunction has the topological order of the laughlin or moore - read state through properties of its entanglement spectrum @xcite : we first verify that the spectrum of the reduced density matrix obtained by tracing out four bosons is gapped .
there is no current quantitative theoretical interpretation of the magnitude of the entanglement gap ; instead , the discriminative power of this criterion comes from requiring that the number of eigenvalues below the gap in each momentum sector obeys counting rules dictated by the topological order of continuum fqh state @xcite . for the lattice geometry used in our studies , laughlin - type order in a system of @xmath54 bosons at @xmath136
is identified by the @xmath137 counting rule ( in the terminology of the previous references ) , which requires @xmath138 states below the gap at @xmath133 , @xmath139 states below the gap in sectors @xmath140 , and @xmath141 , and @xmath142 states below the gap in all other momentum sectors .
similarly , the counting rule for moore - read - type order in @xmath69 bosons at @xmath143 requires @xmath144 states below the gap at @xmath133 and @xmath135 , and @xmath145 states below the gap in all other momentum sectors .
simulations that fail any of the above tests are assigned an energy gap of zero .
10 url # 1`#1`urlprefix[2]#2 [ 2][]#2 .
_ _ * * , ( ) .
_ et al . _ .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . , &
_ _ * * , ( ) . , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . , &
. _ _ * * , ( ) .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al .
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) .
_ _ * * , ( ) . , &
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , &
_ _ * * , ( ) . , &
_ _ * * , ( ) . , & .
_ _ * * , ( ) . .
_ _ * * , ( ) . .
_ _ * * , ( ) . .
_ _ * * , ( ) . &
_ _ * * , ( ) . , , &
_ _ * * , ( ) .
_ * * , ( ) . , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . .
_ _ * * , ( ) . & . &
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . & .
_ _ * * , ( ) .
the authors thank abhishek roy for preliminary code written in the initial stage of the project and useful discussions and emil bergholtz for useful comments on a draft of this manuscript .
r. r. acknowledges support from the sloan foundation .
g. m. acknowledges support from the leverhulme trust under grant no .
ecf-2011 - 565 , from the newton trust of the university of cambridge , and from the royal society under grant uf120157 .
this work used computational and storage services associated with the hoffman2 shared cluster provided by ucla institute for digital research and education s research technology group .
part of our numerical work was performed using the darwin supercomputer of the university of cambridge high performance computing service funded by strategic research infrastructure funding from the higher education funding council for england and funding from the science and technology facilities council .
, for models treated in the main text
. plots in each column use the same color scale and set of contours , as indicated .
( * a * ) ( * c * ) berry curvature distribution for the haldane , kagom lattice and ruby lattice models , respectively , at parameter values which minimize the root - mean - square curvature fluctuation @xmath50 over the unit cell .
curvature fluctuations of the ruby lattice model at its @xmath146 point ( * c * ) are too small to be visible on the common scale .
( * d * ) ( * f * ) berry curvature distribution for the transformed versions of these models used in supplementary ref . , evaluated the parameter values minimizing the values of @xmath50 as computed with the transformed hamiltonians . ] and root - mean - square berry curvature fluctuation @xmath50 .
data shown are for the laughlin state of @xmath54 fermions in the kagom lattice model with nearest - neighbor repulsion on a @xmath129 lattice .
the single - particle parameters used are @xmath147 and @xmath148 , with the color scale corresponding to @xmath149 .
this model and parameters are chosen to match those used in supplementary ref .
( compare fig .
26 of that reference ) ; the thick black line corresponds to values of @xmath24 computed using the transformed hamiltonian of that reference , while colored circles denote values of @xmath24 computed using the hamiltonian in the main text .
the two sets of data have the same gaps ; only the computed values of @xmath24 differ . ] ) is plotted against the rms fluctuation of the berry curvature @xmath24 , for all parameter values used in the text for the ( * a * ) the kagom lattice model and ( * b * ) the ruby lattice model . despite sampling a large volume of the parameter space of both models , the fluctuations in both quantities show a large degree of linear correlation . ] . *
( * a * ) gaps for the bosonic laughlin state of @xmath54 bosons at @xmath55 , as a function of the brillouin zone average of the determinant condition @xmath150 .
we only plot gaps for parameter values which have berry curvature fluctuations @xmath24 less than twice its minimum value . ( * b * ) gap of the bosonic laughlin state vs. @xmath150 for coupling values randomly chosen on isosurfaces of constant @xmath50 in the space of couplings .
the parameter space sampling procedure used to obtain these sets of points is described in the main text ( see methods ) .
( * c * ) , ( * d * ) the same , for the bosonic moore - read state of @xmath69 bosons at @xmath70 .
note that the same sets of couplings are used in each column . ] . *
( * a * ) gaps for the bosonic laughlin state of @xmath54 bosons at @xmath55 , as a function of the brillouin zone average of the determinant condition @xmath150 .
we only plot gaps for parameter values which have berry curvature fluctuations @xmath24 less than twice its minimum value .
( * b * ) gap of the bosonic laughlin state vs. @xmath150 for coupling values randomly chosen on isosurfaces of constant @xmath50 in the space of couplings . the parameter space sampling procedure used to obtain these sets of points
is described in the main text ( see methods ) .
( * c * ) , ( * d * ) the same , for the bosonic moore - read state of @xmath69 bosons at @xmath70 .
note that the same sets of couplings are used in each column . ]
particles , which represents a compromise between the need to accurately estimate the gap in thermodynamic limit and the need to exhaustively sample the parameter space of single - particle hamiltonians using the computational resources available to us . here
we plot the many - body gap @xmath0 for the bosonic laughlin state in the kagom lattice model with nearest - neighbor hopping @xmath148 , variable @xmath149 , and next - nearest - neighbor hopping @xmath151 .
red , black , green and blue points denote data for systems with @xmath130 , @xmath152 , @xmath153 and @xmath154 particles , respectively : the behavior changes significantly between system sizes @xmath130 and @xmath54 , but the gap data for @xmath69 and @xmath131 closely mirrors that of @xmath54 , especially near the maximum of the gap .
inset : finite size scaling of the gap as a function of @xmath155 for selected values of @xmath149 . near the maximal gap ,
the thermodynamic extrapolation ( @xmath132 ) for the gap differs from the finite size value at @xmath54 by about @xmath153% . ]
cxxx & & & + @xmath47 & 0.225 & 0.131 & 0.354 + @xmath60 & 0.0 & 0.0 & 0.0 + @xmath50 & 0.99326 & 1.24755 & 1.12760 + @xmath57 & 2.85734 & 3.97585 & 3.25586 + @xmath150 & 0 & 0 & 0 + @xmath58 & 0.46705 & 0.84597 & 0.99358 + @xmath0 & & 0.02554 & 0.17243 as in the main text , we take the hopping amplitudes to be @xmath156 , with the remaining couplings @xmath47 , @xmath60 as free variables .
columns list information for the following values of @xmath48 of interest : those which minimize berry curvature fluctuations @xmath50 ( obtained using the steepest - descent procedure described in supplementary note [ sec - gradients ] ) and those maximizing the many - body gap @xmath0 for the laughlin state for @xmath54 fermions at @xmath56 and @xmath54 bosons at @xmath55 .
the latter two parameter values were found using our parameter space sampling procedure ( see methods ) .
for these parameters , we compute the root - mean - square average over the brillouin zone of fluctuations in the berry curvature ( @xmath24 ) and quantum metric ( @xmath157 ) , as well as the brillouin zone averages of the determinant ( @xmath150 ) and trace ( @xmath58 ) conditions .
cxxx & & & + @xmath158 & 1.0 & 0.55 & 0.75 + @xmath159 & -1.0 & -0.65 & -0.65 + @xmath50 & 0.14699 & 0.43579 & 0.34909 + @xmath57 & 3.00134 & 1.70829 & 2.29776 + @xmath150 & 0 & 0 & 0 + @xmath58 & 3.44154 & 0.75821 & 1.46273 + @xmath0 & & 0.19153 & 0.12614 as in the main text , we fix @xmath160 , @xmath161 , and @xmath44 , allowing @xmath158 , @xmath159 to vary .
columns list information for the following values of @xmath162 of interest : those which minimize berry curvature fluctuations @xmath50 ( obtained using the steepest - descent procedure described in supplementary note [ sec - gradients ] ) and those maximizing the many - body gap @xmath0 for the laughlin state of @xmath54 bosons at @xmath55 and for the moore - read state of of @xmath69 bosons at @xmath70 .
the latter two parameter values were found using our parameter space sampling procedure ( see methods ) .
for these parameters , we compute the root - mean - square average over the brillouin zone of fluctuations in the berry curvature ( @xmath24 ) and quantum metric ( @xmath157 ) , as well as the brillouin zone averages of the determinant ( @xmath150 ) and trace ( @xmath58 ) conditions .
cxxx & & & + @xmath149 & 0.745 & 0.725 & 0.934 + @xmath63 & -0.361 & -0.361 & -0.168 + @xmath163 & 0.078 & 0.055 & -0.129 + @xmath50 & 0.22178 & 0.23231 & 0.34967 + @xmath57 & 1.31746 & 1.25972 & 1.25434 + @xmath150 & 0.09608 & 0.09857 & 0.41665 + @xmath58 & 0.79612 & 0.70181 & 0.42712 + @xmath0 & & 0.17614 & 0.08246 we fix an energy scale by setting @xmath164 and letting the remaining model parameters ( @xmath165 , @xmath158 , @xmath166 ) vary .
columns list information for the following parameter values of interest : those which minimize berry curvature fluctuations @xmath50 ( obtained using the steepest - descent procedure described in supplementary note [ sec - gradients ] ) and those maximizing the many - body gap @xmath0 for the laughlin state of @xmath54 bosons at @xmath55 and for the moore - read state of of @xmath69 bosons at @xmath70 .
the latter two parameter values were found using our parameter space sampling procedure ( see methods ) .
for these parameters , we compute the root - mean - square average over the brillouin zone of fluctuations in the berry curvature ( @xmath24 ) and quantum metric ( @xmath157 ) , as well as the brillouin zone averages of the determinant ( @xmath150 ) and trace ( @xmath58 ) conditions .
cxxx & & & + @xmath167 & 1.857 & 1.886 & 1.672 + @xmath168 & 0.685 & -0.569 & -1.261 + @xmath169 & 2.079 & 2.536 & 2.789 + @xmath170 & -2.118 & -2.022 & -2.007 + @xmath50 & 0.13398 & 0.14347 & 0.17980 + @xmath57 & 0.47669 & 0.21574 & 0.38617 + @xmath150 & 0.17740 & 0.25569 & 0.30456 + @xmath58 & 0.17168 & 0.08262 & 0.11139 + @xmath0 & & 0.10173 & 0.03132 we fix an energy scale by setting @xmath171 and letting the remaining parameters ( @xmath172 , @xmath93 , @xmath94 ) vary .
columns list information for the following parameter values of interest : those which minimize berry curvature fluctuations @xmath50 ( obtained using the steepest - descent procedure described in supplementary note [ sec - gradients ] ) and those maximizing the many - body gap @xmath0 for the laughlin state of @xmath54 bosons at @xmath55 and for the moore - read state of of @xmath69 bosons at @xmath70 .
the latter two parameter values were found using our parameter space sampling procedure ( see methods ) .
for these parameters , we compute the root - mean - square average over the brillouin zone of fluctuations in the berry curvature ( @xmath24 ) and quantum metric ( @xmath157 ) , as well as the brillouin zone averages of the determinant ( @xmath150 ) and trace ( @xmath58 ) conditions .
here we discuss issues arising in the definition of berry curvature for bloch bands which are absent from the general formalism of geometric phases .
the issues identified here are not new@xcite ; we discuss them here in order to demonstrate that the values the berry curvature and quantum metric take at a specific crystal momentum @xmath10 are unambiguously defined and , in principle , measurable .
berry curvature arises in physical applications at the as the commutator of band - projected position operators ; as examples , we cite the semiclassical approximation to the orbital magnetization@xcite which contributes to anomalous thermoelectric transport , and the intrinsic contribution to the anomalous hall conductance@xcite @xmath173 other applications are reviewed in supplementary ref . .
we mention these observables here because they depend on the berry curvature through forms _ other _ than its brillouin zone ( bz ) average ( the chern number ) , meaning that the distribution of curvature within the bz is an experimentally meaningful quantity .
there is also a recent experimental proposal to measure the berry curvature directly@xcite .
textbook discussions of berry s phase are usually framed in the context of adiabatic evolution of a quantum state tracing out a closed cycle in some parameter manifold . in chern insulator applications ,
the parameter manifold is the brillouin zone , and the instantaneous eigenfunctions at a parameter @xmath10 are the spatially periodic part of the bloch functions @xmath16 .
the berry connection is @xmath174 , which has curvature given by @xmath175 the first chern number is defined as the surface integral of the berry curvature @xmath176 and is topologically quantized to integer values due to the fact that the brillouin zone is a compact manifold ( a torus ) .
eigenstates of the band hamiltonian @xmath177 are only defined up to an overall phase @xmath178 is any smooth function satisfying @xmath179 .
this is the gauge symmetry of the band hamiltonian , and it is the only such symmetry in the absence of energy degeneracies ( assumed for simplicity throughout this section ) . as with the @xmath180 gauge symmetry of electromagnetism , gauge transformations of the form ( [ eq_bloch_gauge ] )
alter the berry connection but leave the berry curvature ( analogue of the magnetic field ) and quantum metric unchanged , as can be seen from their explicitly gauge - invariant forms ( [ eq_b_curvature ] ) , ( [ eq_q_metric ] ) .
the reader will note that the band hamiltonians used in this article do _ not _ have the periodicity of the reciprocal lattice ; nor do their eigenfunctions @xmath181 , and so neither can immediately be viewed as functions on the bz torus defined by identifying the points @xmath10 and @xmath182 for any reciprocal lattice vector @xmath183 .
we first explain that the formalism of band geometry is unchanged in this situation , and then argue that this choice of basis is the canonically correct one , in the sense of corresponding to the observable quantities mentioned above .
because the full hamiltonian is periodic in real space , bloch s theorem implies @xmath184 is unchanged under @xmath185 ; in other words , there exists a unitary matrix @xmath186 such that @xmath187 for _ all _ @xmath188 .
because @xmath189 is independent of @xmath10 , it drops out of the expressions for the berry curvature and quantum metric , which are therefore periodic in @xmath10 .
one could also obtain manifestly periodic bloch functions by performing momentum - dependent phase shifts @xmath190 , with different offsets @xmath191 for each sublattice @xmath3 , so that the transformed bloch functions are invariant under @xmath192 : the transformed curvature and metric are then periodic as well . the resulting hamiltonian is , of course , gauge - inequivalent to the original one , which can be seen from the fact that the curvature itself changes : under transformations @xmath193 for @xmath194 , the berry curvature at @xmath10 changes by @xmath195 because this is a sum of total derivatives , the surface integrals of @xmath196 and @xmath197 yield the same chern number .
the difference itself , however , only vanishes when @xmath191 is the same for all @xmath3 , which is the gauge transformation ( [ eq_bloch_gauge ] ) .
phase shifts of the form ( [ eq_not_gauge ] ) were employed in recent publications@xcite to obtain band hamiltonians that were periodic in @xmath10 .
this was described as a `` gauge transformation '' in these references , but as we ve noted , the only gauge symmetry of the hamiltonian is with respect to _ bands _
( i.e. @xmath180 rotations in the eigenbasis @xmath198 ) . for the transformations made in supplementary ref . , the difference in curvature fluctuations is substantial , as shown in supplementary fig .
[ fig - zoo ] for the haldane , kagom lattice and ruby lattice models .
note that each panel of this figure shows the curvature distribution for parameters which minimize @xmath24 as computed with each panel s respective hamiltonian ; i.e. they depict the closest one may get to uniform band geometry in the parameter space of the hamiltonian considered . for completeness
, we note that unlike these single - particle properties , the many - body gap is invariant under the generalized phase shift ( [ eq_not_gauge ] ) , because single - particle density operators @xmath199 are left unchanged by the transformation . as an example , in supplementary fig .
[ fig - zoo26 ] we reproduce the results shown in fig .
26 of supplementary ref . ,
along with the fluctuations in the canonically defined berry curvature for the same system .
the latter are lower than in the transformed hamiltonian used in that reference , meaning that the negative correlation between @xmath50 and the gap is stronger than depicted there : introducing phase shifts by hand distorts the berry curvature distribution to a degree which significantly affects the conclusions one may draw from that data .
we have shown that band geometry may be defined for non - periodic band hamiltonians ; we now argue that the non - periodic basis used in this paper is in fact the one measured by any ( direct or indirect ) experimental probe , and hence is the only one which should be regarded as physical .
the feature possessed by bloch bands which is absent from the general theory of berry phases is the fact that the parameter space in question is defined via the fourier transform of the kinematic setting of the physical system . because this is a _
global _ transform , in doing the fourier sum latexmath:[\[\label{eq_tb_fourier } we have already implicitly chosen a basis for the band hilbert space at each @xmath10 .
for example , the the transformations made in supplementary refs .
correspond to defining modified tight - binding states on the reciprocal lattice via @xmath201 where @xmath202 ; this is manifestly invariant under @xmath203 , but in doing so the position operator is no longer consistently defined for orbitals with different offsets @xmath77 : the transformation is equivalent to shifting the orbitals of the crystal basis to the origins of various lattice cells . as an obviously apparent consequence ,
the crystal symmetry of the transformed berry curvatures in the second row of supplementary fig .
[ fig - zoo ] are unphysically broken . because experiments probe position - space quantities ,
we conclude that the basis defined by and used in this paper is the canonical , physically relevant quantity .
[ sec - ineqs ] the inequalities @xmath204 were proved in supplementary ref . . assuming the quantum metric is nondegenerate , it may be factored as @xmath205 where the scalar @xmath206 and the second factor is a unimodular matrix . using the standard inequality between arithmetic and geometric means , @xmath207 by unimodularity @xmath208 , so @xmath209 with equality if and only if @xmath28 is proportional to the identity matrix . because the berry curvature is algebraically independent of the components of the quantum metric
, we can conclude that saturation of trace inequality ( [ eq_trineq ] ) implies saturation of the determinant inequality ( [ eq_detineq ] ) . in physical terms , a constant curvature and metric that saturate the determinant inequality imply the closure of a modified girvin - macdonald - platzman ( gmp ) algebra ; saturation of the trace inequality then corresponds to the stronger condition that the algebra of band projected density operators is not only isomorphic to the gmp algebra , but identical to it .
by adding suitably chosen exponentially - localized couplings , any two - band hamiltonian may be brought to a band - flattened form parameterized by a unit 3-vector @xmath210 , where @xmath211 are the three pauli matrices .
the eigenvectors of the original and flattened hamiltonians are identical ( by construction ) and can be obtained analytically : @xmath212 using this expression gives the berry curvature and quantum metric of the lower band as @xmath213 since @xmath214 . using this and standard identities relating multiple dot and cross products , it follows that @xmath215 the determinant condition @xmath216 is necessarily satisfied for any two - band model .
as our results with the kagom and ruby lattice models show , this ceases to be the case for models having more than two bands .
fractional chern insulators ( fcis ) exhibit states which may be thought of as discretizations of continuum fractional quantum hall ( fqh ) states , in that they have identical topological and long - wavelength properties ( see , e.g. , supplementary ref . ) . to that end , consider a continuum position - space wavefunction @xmath217 which is discretized to set of values @xmath218 in a tight - binding model with @xmath5 sites per unit cell .
the normalization conditions on the corresponding bloch functions are @xmath219 for any @xmath10 , where the integral is taken over a unit cell of the lattice . in the limit of large @xmath5 , we may assume that @xmath220 at @xmath221 is proportional to the discretized value @xmath222 .
approximating the continuum normalization integral by a sum introduces a factor of @xmath223 from the integration measure , which is compensated by the scaling @xmath224 now consider the matrix elements of the delta - function interaction employed in the main text to stabilize the bosonic laughlin state . in the continuum , these are @xmath225 times a momentum - conserving @xmath226-function . comparing this with an on - site interaction in the discretized model @xmath227 and using shows that the discretized interaction strength should be scaled as @xmath228 this means that , given gaps which have been obtained for the fci laughlin state in two lattice models @xmath229 , the quantities which should be compared are @xmath230 and @xmath231 ( assuming that the single - particle dispersion has been flattened ; i.e. that the gap is only set by the interaction term . )
similar considerations show that for the three - body delta - function interaction used to stabilize the moore - read state , the leading scaling should be @xmath232 .
we note that the above argument is not as simple as it may appear : we re implicitly assuming @xmath217 has the character of a low - lying eigenstate more specifically , that it has support over almost all tight - binding orbitals .
this is the case when @xmath222 is a randomly chosen vector from the @xmath5-dimensional band hilbert space , but one can easily construct counterexamples violating this assumption : for example , let lattice model @xmath233 ( with @xmath234 ) have a block - diagonal kinetic term , the first block of which is identical to the kinetic term of model @xmath235 and the second block of which describes additional , trivial `` spectator '' orbitals with very high occupation energies .
the low - lying bands @xmath236 will have vanishing amplitude on these spectator orbitals , and will numerically be identical to the corresponding eigenfunctions @xmath237 of model @xmath235 , with zeros appended . in this scenario ,
the eigenfunction scaling assumption is violated , and the interaction strength ( and gap size ) would _ not _ scale with the number of spectator orbitals .
the scaling argument can be placed on a more rigorous footing in the context of the hofstadter model@xcite in the limit of small flux per plaquette @xmath238 , which offers a sequence of lattice models ( with @xmath239 bands ) which converges to the continuum fqhe in the @xmath132 limit . in this case , the scalings , are numerically observed to be obeyed to high accuracy@xcite . for the case of the fci models studied in this work , we know
the ground states of different models lie in the same universality class as the continuum fqhe state ( as can be determined by topological order , done here by entanglement spectrum counting ) , but we do not expect , e.g. , the ruby lattice model ground state to be an interpolation of the kagom lattice model ground state . the fact that we observe the scaled gaps to be so close to each other provides an _ a posteriori _ confirmation of this argument s validity even for the relatively small values of @xmath240 .
the berry curvature and quantum metric are defined in terms of the derivatives of a wavefunction over some parameter manifold . for both numerical and theoretical purposes ,
it becomes convenient to recast these expressions in terms of derivatives of the hamiltonian itself , rather than its eigenfunctions , since ordinarily only the former is known analytically .
this , of course , is not a new observation @xcite . in order to avoid overall phase ambiguities and the necessity of using multiple charts to cover the brillouin zone ( bz ) in topologically non - trivial situations
, it s preferable to work with the occupied band projector @xmath241 , instead of the eigenfunction @xmath242 itself . in these terms
, the feynman - hellman theorems are @xmath243 where the projected resolvent operator @xmath244 is @xmath245 equations ( [ eq_de_def ] ) and ( [ eq_dp_def ] ) hold for any parameter @xmath246 upon which the hamiltonian smoothly depends , and they are valid for arbitrary values of @xmath246 and hence may be further differentiated . assuming no additional degeneracies , @xmath247 may be inverted in the subspace orthogonal to @xmath248 , and @xmath249 the relations ( [ eq_de_def ] ) , ( [ eq_dp_def ] ) and ( [ eq_dr_def ] ) then form a closed system and may be iterated to any order to develop the taylor expansion of band - geometric quantities .
for example , the berry curvature and quantum metric for a single occupied chern band are@xcite @xmath250 the derivative of the berry curvature with respect to a coupling @xmath246 follows from a straightforward computation as @xmath251 \right\}.\end{aligned}\ ] ] other quantities such as the hessian @xmath252 may be calculated in a similar manner , although the algebra rapidly becomes tedious .
these may , in turn , be used to find the variation in brillouin zone - averaged quantities with respect to couplings in the hamiltonian ; for example , denoting by @xmath253 the brillouin zone average,@xmath254 here we ve made use of the fact that @xmath255 while we remain in the same topological phase , which can easily be checked during numerical computations . using ( [ eq_curvature_gradient ] ) in ( [ eq_rms_curvature_gradient ] ) then permits us to rapidly perform a steepest - descent minimization of berry curvature fluctuations , even in a high - dimensional parameter space .
10 url # 1`#1`urlprefix[2]#2 [ 2][]#2 | the fractional quantum hall ( fqh ) effect illustrates the range of novel phenomena which can arise in a topologically ordered state in the presence of strong interactions .
the possibility of realizing fqh - like phases in models with strong lattice effects has attracted intense interest as a more experimentally accessible venue for fqh phenomena which calls for more theoretical attention . here
we investigate the physical relevance of previously derived geometric conditions which quantify deviations from the landau level physics of the fqhe .
we conduct extensive numerical many - body simulations on several lattice models , obtaining new theoretical results in the process , and find remarkable correlation between these conditions and the many - body gap .
these results indicate which physical factors are most relevant for the stability of fqh - like phases , a paradigm we refer to as the geometric stability hypothesis , and provide easily implementable guidelines for obtaining robust fqh - like phases in numerical or real - world experiments . the fractional quantum hall effect ( fqhe )
provides a spectacular manifestation of the breakdown of the spin - statistics relation in two dimensions : one obtains quantum number fractionalization@xcite and , potentially , non - abelian statistics@xcite which can form the substrate for topologically robust quantum computing@xcite .
progress has been hampered by the considerable experimental difficulties involved in realizing the fqhe in the usual setting of a semiconductor heterostructure , but a flurry of interest in the field was set off by the recent insight@xcite that these exotic phases of matter may also arise in topologically nontrivial insulators with partially filled flat bands , or fractional chern insulators@xcite ( fcis ) .
the attractiveness of fcis stems from the fact that the bandgap @xmath0 may be set without the use of a large external magnetic field , the strength of which is one of the limiting factors in the semiconductor fqhe .
there are currently a range of experimental proposals for realizing fcis in cold atom systems@xcite , transition metal oxides@xcite and elsewhere ; a successful experimental implementation in ultracold fermions was recently announced in ref . .
moreover , fcis raise theoretical questions independent of their experimental interest .
the majority of theoretical work on the fqhe over its thirty - year history has focused on the influence of interactions on the landau level hamiltonian , which occupies a unique , highly symmetric point in the space of single - particle hamiltonians .
fci phenomena constitute a non - trivial and poorly - understood generalization of the fqhe in which lattice effects are non - negligible ; a generic fci does not have the fqhe as a continuum limit , and examples of lattice effects without any continuum analog have already been noted@xcite .
a generalization of our theoretical understanding of the fqhe to cover the case of fcis is hence both nontrivial and experimentally relevant .
one possible approach to the stability of fcis is via the single - mode approximation used by girvin , macdonald and platzman@xcite ( gmp ) , who made the ansatz that the most relevant excitations which destabilize an fqh ground state are neutral magnetoroton modes generated by the action of electron density operators projected to the lowest landau level .
these operators do not commute with each other , due to the projection , but gmp found that the set of operators remains a closed algebra under commutation .
intuitively , one expects that the form of this algebra plays a crucial role in the stability of the fqhe phases by limiting the set of possible destabilizing interactions . in a generic fci , however , the analogous set of band - projected density operators is not a closed algebra , nor do the projected densities span the space of single - particle operators@xcite : there is no canonical mapping between a general lattice fci and the continuum fqhe . in previous work , one of us@xcite
derived sufficient conditions for the band - projected density operators in an fci to satisfy a closed algebra isomorphic to that present in the fqhe , which justified and elaborated upon a heuristic criterion used in previous fci literature .
quantities describing the geometry of the chern band ( its embedding in hilbert space ) enter this analysis in a natural way as coefficients of terms which must necessarily vanish in order to obtain a closed algebra ; remarkably , only three conditions need to be placed on the band s geometry for the isomorphism to be present to all orders in a long - wavelength expansion .
heuristically , one might expect that reproducing the density operator algebra would then suffice to reproduce the full physics of the fqhe , but this argument has not been fully tested in the literature . in the present work , we report the results of extensive numerical simulations which demonstrate that quantitative measures based on the band - geometric conditions of ref .
are robustly correlated with the many - body gap in realizations of fqh - type phases in different fci lattice models .
in addition to numerical data , we obtain several theoretical results , such as a scaling relation between the gap of an fqh - like state and the number of bands in an fci model , which is essential for comparing different models .
we find that the berry curvature was computed incorrectly in a number of prior references ; in supplementary note 1 , we discuss why this quantity is defined unambiguously
. the remarkably high degree of correlation we find between band geometry and the many - body gap leads us to propose a geometric stability hypothesis : that the algebra of band - projected density operators governs fqh - type phenomena in fcis , even when the isomorphism does nt hold exactly , and that the single - particle conditions investigated here are accurate qualitative estimators of the stability of an fqh - like state .
this frames the theoretical problem of generalizing results on the fqhe to cover fci physics by distilling the effects of the lattice into a small number of quantities measuring the relevant deviations of an fci from lowest landau level behavior .
our results are also of use in experimental design , as they provide a computationally inexpensive means to estimate which choices of fci model parameters are most likely to yield a fqh - like state with the largest possible gap ; a naive analysis of the scales involved has suggested this may be on the order of room temperature@xcite . from the opposite point of view , our results also indicate which areas of parameter space should be searched to find possible fci states which do not correspond to fqh universality classes . |
according to the factorization theorem , the cross sections for dis reactions and some classes of hadron - hadron collisions can be expressed as convolution of two parts : a short - distance perturbative and a long - distance nonperturbative ones .
the perturbative part , describing partonic cross sections at sufficiently high scale of the momentum transfer @xmath0 can be calculated within perturbative chromodynamics ( pqcd ) .
the non - perturbative part contains universal , process independent parton distribution functions @xmath1 ( pdf ) and fragmentation functions @xmath2 ( ff ) , which can be measured experimentally .
the evolution of these functions with the interaction scale @xmath3 is again described with the use of the perturbative qcd methods .
the standard dglap approach @xcite enables one to calculate parton densities which characterize the internal nucleon structure at a given scale @xmath3 when these densities are known for a certain input scale @xmath4 .
we have shown that also the truncated mellin moments of the pdfs , @xmath5 , satisfy the dglap evolution equations and can be an additional tool in the qcd analysis of structure functions . the major advantage of the tmma is a possibility to adapt theoretical qcd analysis to the experimentally available region of the bjorken-@xmath6 variable . in this way , one can avoid the problem of dealing with the unphysical region @xmath7 corresponding to the infinite energy of interaction .
a number of important issues in particle physics , e.g. , solving of the ` nucleon spin puzzle ' , quark - hadron duality or higher twist contributions to the structure functions refers directly to moments . note that tmm , contrary to standard moments , may be directly extracted from the accurate ( jlab ) data by appropriate binning ( keeping @xmath3 fixed ) .
these issues initiate a large number of experimental projects and theoretical studies as well .
below we present the generalization of dglap evolution equations within tmma and discuss some applications in spin physics .
the main finding of the tmma is that the generalized truncated ( cut ) moments ( cmm ) obtained by multiple integrations as well as multiple differentiations of the original parton distribution also satisfy the dglap equations with the simply transformed evolution kernel @xcite .
a similar generalized evolution equation , with the correspondingly modified coefficient functions , can also be obtained for structure functions . in table
[ table 1 ] , we summarize the generalized cmm together with the correspondingly transformed dglap evolution kernels .
ll|l & generalized cmm & dglap kernel @xmath8 + 1 . &
@xmath1 & @xmath9 + & & + 2 . & @xmath10 & @xmath11 + & & + 3 . & @xmath12 & @xmath11 + & & + 4 . & @xmath13 & @xmath14 + & & + 5 . & @xmath15 & @xmath16 + & & + 6.&@xmath17 $ ] & @xmath18 +
below we present examples of applications of tmma to analysis of the bjorken sum rule @xcite and fragmentation functions . for any normalized function @xmath19 , @xmath20 , one can construct generalized cmm @xmath21 as a mellin convolution with the function @xmath22 , which obeys the dglap evolution equation with the rescaled kernel @xcite : @xmath23 @xmath24 the special case of @xmath25 is suitable for the generalized bjorken sum rule ( bsr ) : @xmath26 @xmath27 has the same evolution kernel as @xmath28 and the generalized bsr is equal to the ordinary bsr : @xmath29 the corresponding cut first moments of @xmath27 go to the bsr limit as the cut point @xmath30 goes to zero .
this allows one to study behaviour of the generalized cut moments near @xmath31 and estimate the value of the bsr from the cut integrals @xmath32 at @xmath33 .
the attempts for the case @xmath34 are shown in fig .
1 and can be tested experimentally .
contributions to the bjorken sum rule obtained within tmma .
comparison to compass data . ]
contributions to the bjorken sum rule obtained within tmma .
comparison to compass data . ]
we have also calculated contributions to the bsr itself and compared them to the experimental data . in fig .
2 we compare tmma predictions for the contributions to the bsr to recent compass data .
one can see that simple input parametrization 7 , where @xmath35 can satisfactorily reproduce the experimental data .
this relation between @xmath36 and @xmath37 , together with the positivity constraint can provide knowledge on the small-@xmath6 behaviour of the polarized structure function @xmath28 . from our analysis
the favoured small-@xmath6 behaviour of @xmath28 is @xmath38 , where @xmath39 .
table [ table 2 ] contains the truncated contributions to the bjorken sum rule in the experimentally available @xmath6-region , @xmath40 obtained for different input parametrizations @xmath41 our predictions are compared with the hermes @xcite and compass @xcite data .
llll|ll n&input & @xmath6 -range & @xmath42 $ ] & @xmath43 & exp @xmath43 + 1.&@xmath44 & & & 0.161 & hermes + 2.&@xmath45 & 0.021 - 0.9 & @xmath46 & 0.149 & @xmath47 + 3.&@xmath48 & & & 0.131 & + 4.&@xmath44 & & & 0.177 & compass + 5.&@xmath45 & 0.004 - 0.7 & @xmath49 & 0.173 & @xmath50 + 6.&@xmath48 & & & 0.163 & + 7.&@xmath51 & & & 0.168- & compass + 8.&@xmath52 & 0.0025 - 0.7 & @xmath49 & 0.170 & @xmath53 + finally , it is worthy to mention that tmma can be very useful in analysis of the hadron fragmentation functions as in the small-@xmath6 region behaviour of ff is known very poorly . in this way , one can restrict the analysis to well determined models for @xmath54 .
ff also obey the corresponding dglap evolution and their cmm can provide new insight into the hadron structure .
tmm of the kaon ff , @xmath55 ( thin , thick ) .
inputs : kns @xcite , fs @xcite ] tmm of the kaon ff , @xmath55 ( thin , thick ) .
inputs : kns @xcite , fs @xcite ] in figs .
3,4 we present evolution of the truncated moments of ff contributing to the quark charge conservation , @xmath56 where @xmath57 denotes a fragmentation function of the hadron @xmath58 from a parton @xmath59 and @xmath60 are the charges of the hadron @xmath58 and parton @xmath59 .
we use different parametrization of the ff at the initial scale @xcite .
the excess of the obtained moments for pions , providing the main contributions to sum rule , over the charge conservation values at @xmath61 may be considered as a support of inapplicability of independent fragmentation picture at this region .
this fact is also exhibited in the large differences in the presented predictions in this region , depending on the input parametrization , for both pions and kaons . therefore tmm approach can be a natural tool also in the study of the fragmentation functions , requiring further investigations .
we are grateful to a.v .
efremov to stimulating discussions .
o.t . is indebted to s. kumano for useful discussions .
this work is supported by the bogoliubov - infeld program , grant no .
01 - 3 - 1113 - 2014/2018 and by the russian foundation for fundamental research , grant no .
14 - 01 - 00647a .
99 gribov v n and lipatov l n 1972 _ sov .
phys . _ * 15 * 438 , _ ibid .
_ 675 yu . l. dokshitzer yu l 1977 _ sov .
. jetp _ * 46 * 641 altarelli g and parisi g 1977 _ nucl .
_ b * 126 * 298 kotlorz d and kotlorz a 2007 _ phys .
_ b * 644 * 284 kotlorz d and kotlorz a 2011 _ acta phys . pol .
_ b * 42 * 1231 kotlorz d and mikhailov s v 2014 _ jhep _ * 06 * 065 j. d. bjorken j d 1966 _ phys .
rev . _ * 148 * 1467 , _ ibid .
_ 1970 d * 1 * ( 1970 ) 1376 strzik - kotlorz d , mikhailov s v and teryaev o v 2014 _ proc . of science , baldin ishepp xxii _ * 033 * airapetian a _ et al . _
[ hermes collaboration ] 2005 _ phys .
_ d * 71 * 012003 , _ ibid .
_ 2007 d * 75 * 012007 alexakhin v yu _
[ compass collaboration ] 2007 _ phys .
_ b * 647 * 8 , alekseev m _
[ compass collaboration ] 2008 , _ phys .
_ b * 660 * 458 , adolph c _ et al . _ [ compass collaboration ] 2015 [ arxiv:1503.08935][hep - ex ] hirai m , kumano s , nagai t -h and sudoh k 2007 _ phys .
_ d * 75 * 094009 efremov a v and radyushkin a v 1980 _ rivista del nuovo cimento _ * 3 * 1 , sec
. 1.6 de florian d , sassot r and stratmann m 2007 _ phys . rev . _
d * 75 * 114010 | we present progress in development of the truncated mellin moments approach ( tmma ) .
we show our recent results on the generalization of dglap evolution equations and discuss some their applications in spin physics . |
recently three - dimensional topological insulators ( tis ) with magnetically ordered surface states had attracted much attention , both theoretically @xcite and experimentally.@xcite the topological surface states , usually gapless and featuring a low - energy dirac dispersion protected by time - reversal ( tr ) symmetry , generally become gapped when a ferromagnetically ordered layered material is grown over it . in this case
the tr symmetry is broken by the proximity effect to the magnetic material .
interesting effects on the magnetization dynamics arise when the topological surface becomes ferromagnetic.@xcite quantum fluctuations of the surface dirac fermions induce an additional berry phase , which modifies the landau - lifshitz ( ll ) equation for the magnetization dynamics.@xcite furthermore , when an electric field is present , either external or originating from coulomb interactions , a topological magnetoelectric torque also appears in the ll equation.@xcite actually both effects arise from an induced chern - simons ( cs ) term @xcite upon integrating out the dirac fermions.@xcite it is important to notice that most of the theoretical calculations in tis were performed so far only at zero temperature , although some aspects like the shift of the curie temperature in the ferromagnetic insulator / ti heterostructure were discussed.@xcite the disregard of the temperature effects till now is partially connected to the fact that finite temperature is not expected to modify the topological contribution to the effective electromagnetic lagrangian of a ti in the bulk,@xcite @xmath4 where @xmath5 is the fine - structure constant and @xmath6 is the so - called axion field @xcite .
while both the dielectric constant , @xmath7 , and magnetic permeability , @xmath8 , receive finite temperature corrections , @xmath6 remains temperature - independent.@xcite this somewhat confirms the naive expectation for the term being topological in origin .
this , however , is not the full story .
for example , differently from the chiral anomaly leading to the axion term in eq .
( [ eq : l - em - ti ] ) , the parity anomaly in @xmath9 dimensions does indeed receive finite temperature corrections.@xcite therefore , integrating out the dirac fermions in 2 + 1 dimensions and at finite temperature generates additional temperature - dependent cs term .
another important finite temperature aspect relates to the electromagnetic response and the many - body screening of the coulomb interaction . in graphene , where the chemical potential is typically zero
, there is no screening of the long - range coulomb interaction at zero temperature , although a renormalization of the dielectric constant occurs.@xcite however , at finite temperature a thermal screening should occur , corresponding to a situation reminiscent of massless qed in @xmath10 dimensions.@xcite moreover , in contrast to graphene , most three - dimensional tis feature surface dirac fermions at a non - vanishing chemical potential , and we may expect that the screening in tis @xcite and in doped graphene @xcite at zero temperature behaves differently . in recent experiments @xcite thin films of eus / bi@xmath0se@xmath1 heterostructure were grown , and it has been shown that the topological surface becomes ferromagnetic due to a proximity - induced symmetry breaking mechanism . the experiments also confirmed the theoretical expectations that the surface dirac fermions become gapped.@xcite many measurements in such a heterostructure are made at temperatures close to the curie temperature . therefore , it is of paramount importance to study theoretically quantum field theory models of dirac fermions at finite temperature and chemical potential . as we have mentioned above ,
there are finite temperature corrections to the generated cs term .
furthermore , since the cs term provides a correction to the berry phase of the proximity - induced ferromagnetism,@xcite the precessional behavior of the magnetization will be modified at finite temperature and in turn affect the spin - wave excitations .
a finite chemical potential influences the physics on the surface dramatically even at zero temperature .
for example , it is known that when the chemical potential is larger than the gap , the hall conductivity is not quantized any longer and is given by @xcite @xmath11 .
here @xmath12 represents the zero temperature gap .
this suggests that in such a metallic regime the hall conductivity can not always be identified with the coefficient of a _ local _ fluctuation - generated cs term , i.e. , the topological mass which is induced by quantum fluctuations rather than by an external response . since the relevant situation for the magnetization dynamics is when the chemical potential is inside the gap , it is important to ask whether the hall conductivity remains quantized in this case when the temperature is finite . in this paper
we will derive an analytical expression for hall conductivity at finite temperature and chemical potential in precisely this regime .
we will show that a plateau persists at finite temperature when the chemical potential is inside the gap .
the plan of the paper is as follows . in section
ii we introduce the model and discuss some simple thermally induced screening effects following from the vacuum polarization on the ti surface . in section
iii we derive the low - energy cs term at finite temperature and chemical potential .
we discuss the physical consequences of this result for the magnetization dynamics in section iv , and in section v we summarize the main results .
our starting point is the following hamiltonian for the surface dirac fermions , @xmath13 , where @xmath14^t$ ] .
the @xmath15 matrix @xmath16 reads , @xmath17 where @xmath18 is a scalar potential including contributions both from external electric field and coulomb interaction , and the magnetization @xmath19 satisfies the constraint @xmath20 .
the lagrangian @xmath21 can be after a rescaling @xmath22 conveniently written in qed - like form,@xcite @xmath23 in the above equation the dirac slash notation is used , @xmath24 , with @xmath25 , @xmath26 , and @xmath27 , and @xmath28 , @xmath29 , and the vector field @xmath30 . in order to study thermal effects on the magnetization dynamics and the screening of the coulomb interaction , we will perform the calculations in the imaginary time formalism , setting as usual @xmath31 with @xmath32 $ ] , and integrating out the fermions with anti - periodic boundary conditions , @xmath33 . thus , the amplitude @xmath34 appearing in the functional integral becomes @xmath35 in the imaginary time formalism , with the lagrangian in euclidean spacetime given by , @xmath36 where now the dirac slash features the euclidean dirac matrices @xmath37 and a chemical potential @xmath3 was included ( note that @xmath38 ) . integrating out the fermions yields the effective action in the form , @xmath39 , with @xmath40 where @xmath41 is the number of dirac fermion species and @xmath42 is the ( infinite ) volume .
@xmath43 is the magnetic action , which has the general form in the imaginary time formalism , @xmath44,\ ] ] featuring a berry gauge potential satisfying @xmath45 , representing a magnetic monopole in the magnetization space .
the magnetic hamiltonian density @xmath46 may contain several contributions , the most important ones being the coupling to external fields and the exchange terms . in tis
the number @xmath41 of fermion species is odd . in the present context
this is essential , otherwise no parity and tr symmetry breaking via the generation of a cs term would occur dynamically.@xcite assuming that the system orders along the @xmath47-axis , we can write @xmath48 and expand eq . ( [ eq : sf ] ) up to quadratic order in the fields , @xmath49,\end{aligned}\ ] ] where @xmath50|_{n_z=\langle n_z\rangle}$ ] , is the vacuum polarization tensor at finite temperature encompassing screening effects in the coulomb potential and the transverse magnetic susceptibility , and @xmath51|_{n_z=\langle n_z\rangle}$ ] , is the longitudinal magnetic susceptibility . the fermionic propagator has the form , @xmath52 , or in momentum space , @xmath53/[m^2+{\bf p}^2-(i\omega_n+\mu)^2]$ ] , where @xmath54 , @xmath55 , and @xmath56 is the usual fermionic matsubara frequency .
the units are such that @xmath57 .
the calculation of the vacuum polarization in @xmath10 dimensions and zero temperature is well known @xcite and has been reviewed by us in detail recently in ref . .
due to current conservation , it fulfills @xmath58 .
periodicity in the matsubara time allows one to choose a rest frame for the heat bath given by the vector @xmath59.@xcite therefore , in this case we can write the vacuum polarization in momentum space in the form , @xmath60 where @xmath61 and @xmath62 are both transverse in 2 + 1 dimensions , with @xmath61 being transverse and @xmath62 longitudinal in two spatial dimensions . thus , we have @xmath63 , @xmath64 , and @xmath65 , where latin indices refer to spatial dimensions . at the same time , @xmath66,\end{aligned}\ ] ] where @xmath67 and @xmath68 , @xmath69 . at finite temperatures ,
an interesting aspect of the vacuum polarization in relativistic - like fermionic systems is the generation of a thermal mass for the vector field along the time direction.@xcite in other words , the coulomb potential acquires a thermal gap or thermal screening . in the case of a ti surface , for example , the coulomb interaction @xmath70 in momentum space
is given by @xmath71 , similarly to interacting graphene.@xcite the vacuum polarization screens this coulomb interaction , and we have , @xmath72^{-1}+\pi_{00}(0,{\bf q})\}$ ] , allowing us to define the so called _ electric mass_,@xcite @xmath73 .
an explicit calculation assuming @xmath74 yields , @xmath75 -\frac{{\bf q}^2 + 2m^2}{\sqrt{{\bf q}^2 + 4m^2}}\frac{\sinh\left(\frac{\sqrt{{\bf q}^2 + 4m^2}}{2t}\right)}{\cosh(\mu / t ) + \cosh\left(\frac{\sqrt{{\bf q}^2 + 4m^2}}{2t}\right)}\right\},\ ] ] so that the electric mass is given by , @xmath76 \right.\nonumber\\ & -&\left.\frac{|m|\sinh(|m|/t)}{\cosh(|m|/t)+\cosh(\mu / t)}\right\}.\end{aligned}\ ] ] at zero temperature , @xmath77\theta(\mu-|m_0|)$ ] , where @xmath78 .
thus , the electric mass vanishes for @xmath79 . using the above results
, we obtain that in the long wavelength limit the effective coulomb interaction becomes , @xmath80^{-1}+\pi_{00}(0,0)}=\frac{2\pi e^2}{\epsilon(|{\bf q}|+s)},\ ] ] where @xmath81 , featuring in this way a screening of the thomas - fermi type . at zero temperature , @xmath82 . in real space
we have , @xmath83\},\ ] ] where @xmath84 is a bessel function of second kind and @xmath85 is a struve function .
the above results imply that no screening would occur at zero temperature if @xmath86 , i.e. , the screening takes place only when the chemical potential is located above the gap .
this is a reasonable result , since this regime corresponds to a metallic state . for temperatures above the curie temperature , @xmath87 , we have , @xmath88.\ ] ] the above equation is also valid when no ferromagnetic material is in contact with the ti surface . in this case
( [ eq : s - large - t ] ) is valid for all @xmath89 . interestingly , since we are dealing with interacting dirac fermions , we have that in absence of ferromagnetism eq .
( [ eq : s - large - t ] ) would also apply to doped graphene @xcite if one replaces @xmath90 to account for the even number of dirac cones .
next , we consider the finite temperature behavior of the cs term , which will yield a finite temperature correction to the berry phase on the ti surface .
the finite temperature behavior of the cs term is generally a highly nontrivial matter,@xcite depending on the type of background field configuration involved .
however , in our case we are mainly interested in the low - energy behavior , so that the calculation simplifies considerably .
the calculation parallels the zero temperature one,@xcite and thus @xmath91 , where @xmath92 when the chemical potential is inside the gap , i.e. , @xmath93 , the surface state is insulating and we can take @xmath94 in the low - energy regime , since in this case @xmath95 yields the dominant energy scale in the problem and there is no fermi momentum .
the effective action is local in this case , following straightforwardly from a derivative expansion . on the other hand , in the metallic regime where @xmath96 , the effective action is inherently non - local . in this case
there is a fermi surface singularity with a fermi momentum @xmath97 .
generally , the coefficient of the cs term , the so called topological mass , yields a measure of the hall conductivity , when scaled in appropriate unities .
we will now show that in the insulating regime , the topological mass can be evaluated analytically at finite temperature and chemical potential . in this regime
the cs contribution to the effective action is given approximately by , @xmath98 where in the above expression we have returned to real time and have defined @xmath99 .
expressing @xmath100 in units of conductivity via @xmath101 and using that @xmath102/(8\pi|m|)$ ] ( see appendix a ) , we obtain , @xmath103},\ ] ] where we have reintroduced the planck constant .
the above equation provides an analytical expression for the hall conductivity for the case where the chemical potential is inside the gap , i.e. , in the insulating regime .
however , it differs from the actual hall conductivity @xmath104 when @xmath105 . in order to see this ,
let us first compare eq .
( [ eq : sigma ] ) to the hall conductivity at zero temperature , where it can be calculated exactly for a nonzero chemical potential . introducing the unit vector @xmath106 associated to the mean - field hamiltonian @xmath107 , we can write an expression for @xmath108 in terms of the topological charge density , @xmath109.\ ] ] as @xmath110,\ ] ] where @xmath111 $ ] are fermi - dirac distributions . thus , it is straightforward to compute @xmath104 at @xmath112 , @xcite @xmath113,\ ] ] and we see that the hall conductivity at @xmath112 is non - quantized and non - universal in the metallic regime ( @xmath114 ) .
the topological mass , on the other hand , vanishes in the limit @xmath115 when @xmath116 , @xmath117 thus , when the system is in the metallic phase , the topological mass obtained from the low - energy regime of the cs action does not agree with the hall conductivity .
this means that non - local corrections have to be considered in order to make the local effective action approach to agree with the expression obtained from linear response .
further insight on this point is obtained by considering how the topological mass relates to the actual hall conductivity and see how it deviates from it in the metallic regime .
first , we note that we can write , @xmath118 where @xmath119,\ ] ] the quantity @xmath120 yields the deviation of the coefficient of the local contribution to the cs term from the hall conductivity when the chemical is above the gap .
we see that for @xmath121 the contribution @xmath120 is responsible for canceling the non - quantized contribution at @xmath112 , since , @xmath122 precisely canceling the last term in eq .
( [ eq : sigma_xy-0 ] ) .
thus , we see that the chemical potential acts as a cutoff setting the limit of validity of the local effective action . in fig .
[ fig : sigmat](a ) we show @xmath123 for a typical mean - field theory dependence with the temperature for @xmath95 . in order to plot @xmath124 and @xmath104
we have used values for @xmath125 and curie temperature estimated from the experiment by wei _
@xcite performed on bi@xmath0se@xmath1/eus samples . there
the estimated value for @xmath125 at the interface is considerably larger ( @xmath126 mev @xcite ) than the one obtained in _ ab initio _
calculations for bi@xmath0se@xmath1/mnse,@xcite which features an antiferromagnet material ( mnse ) rather than a ferromagnetic one .
the numerical calculation of the hall conductivity is shown in fig .
[ fig : sigmat](b ) .
we note a plateau similar to the one arising for @xmath124 , which indicates that quantization nearly holds for finite @xmath2 and @xmath3 in a region of the @xmath127 plane determined by the inequality @xmath128 .
however , there is also a region in the @xmath127 plane where the hall conductivity is non - quantized while the topological mass @xmath124 vanishes .
note that sharp and large plateau regions up to @xmath129 [ fig . 1
, panels ( a ) and ( b ) ] occur only if the gap is larger than @xmath87 , which is precisely the case in the experiment from ref . .
in panels ( c ) and ( d ) of fig .
1 we also illustrate the opposite situation , taking for example , @xmath130 . observe that the plateau region here is not as sharp and loses its approximate quantization long before @xmath87 is reached .
let us now derive the ll equation in the insulating regime where @xmath93
. in this case the cs action ( [ eq : cs - action ] ) contributes to the ll equation in two different ways .
this can be see by decomposing eq .
( [ eq : cs - action ] ) into two parts , @xmath131 , where @xmath132 is the correction to the berry phase , and @xmath133 is the magnetoelectric contribution .
therefore , these two contributions lead to the ll equation , @xmath134\partial_t{\bf n}= { \bf n}\times{\bf h}_{\rm eff}-\alpha({\bf n}\times\partial_t{\bf n})+{\bf t}_{\rm tme},\ ] ] where @xmath135 is the magnetoelectric torque .
note that within the one - loop accuracy of our calculations @xmath136 . thus , due to the magnetoelectric torque
, the spin - wave excitation will be gapped in general .
for instance , if @xmath137 with @xmath138 , we have , @xmath139 where in the absence of an external magnetic field , @xmath140 as @xmath141 .
we have discussed the screening of the coulomb interaction on the surface of a ti at finite temperature and chemical potential . in the presence of proximity - induced ferromagnetism we found a peculiar behavior where the screening length depends on the magnetization via the electronic gap .
this implies in particular that no screening occurs at zero temperature if the chemical potential is smaller than the proximity - induced fermionic gap .
we have also calculated the coefficient of the cs term ( the topological mass ) analytically for finite temperature and chemical potential in the regime @xmath93 .
this topological mass yields the hall conductivity in this regime and directly influence the magnetization dynamics at the surface , modifying the spin - wave velocity and inducing an magnetoelectric gap in the spin - wave excitation spectrum .
interestingly , in the insulating regime the topological mass remains nearly quantized even at finite temperature and chemical potential .
we would like to thank the sfb tr 12 program of the deutsche forschungsgemeinschaft ( dfg ) for the financial support .
the work of ie is also supported by the ministry of education and science of the russian federation in the framework of increase competitiveness program of nust `` misis '' ( nr .
k2 - 2014 - 015 ) in this appendix we will discuss the integral ( [ eq : integral ] ) in more detail .
let us first consider the case where @xmath142 and @xmath143 simultaneously .
in this case it is very simple to perform the momentum integral to obtain , @xmath144,\end{aligned}\ ] ] which after some straightforward simplifications lead to eq .
( [ eq : sigma ] ) .
now , let us set @xmath143 and leave the bosonic matsubara frequency @xmath145 . if we now perform the fermionic matsubara sum , we obtain , @xmath146},\ ] ] where @xmath147 .
if we now take the limit @xmath148 , we obtain the result needed to compute the hall conductivity . clearly , @xmath149 is not the same as @xmath150 in eq .
( [ eq : i00 ] ) .
thus , the matsubara sum and momentum integral do not commute with the limit @xmath148 .
we note that when @xmath148 before performing the matsubara sum , the poles coalesce , such that an additional contribution arises , which makes @xmath124 to differ from the hall conductivity by the term @xmath120 [ recall eq .
( [ eq : spin - conduc ] ) ] .
s. v. eremeev , v. n. menshov , v. v. tugushev , p. m. echenique , and e. v. chulkov , phys . rev .
b * 88 * , 144430 ( 2013 ) ; v. n. menshov , v. v. tugushev , s. v. eremeev , p. m. echenique , and e. v. chulkov , phys . rev .
b * 88 * , 224401 ( 2013 ) .
y. s. hor , p. roushan , h. beidenkopf , j. seo , d. qu , j. g. checkelsky , l. a. wray , d. hsieh , y. xia , s .- y .
xu , d. qian,, m. z. hasan , n. p. ong , a. yazdani , and r. j. cava , phys .
b * 81 * , 195203 ( 2010 ) . | motivated by recent experiments on eus / bi@xmath0se@xmath1 heterostructures , we study the temperature dependent screening effects on the surface of a three - dimensional topological insulator proximate to a ferromagnetically ordered system . in general , we find that besides the chemical potential and temperature , the screening energy scale also depends on the proximity - induced electronic gap in an essential way . in particular , at zero temperature the screening energy vanishes if the chemical potential is smaller than the proximity - induced electronic gap .
we show that at finite temperature , @xmath2 , and/or chemical potential , @xmath3 , the chern - simons ( topological ) mass , which is generated by quantum fluctuations arising from the proximity - effect , can be calculated analytically in the insulating regime . in this case
the topological mass yields the hall conductivity associated to edge states .
we show that when the chemical potential is inside the gap the topological mass remains nearly quantized at finite temperature . |
atomic spin squeezing @xcite of a large ensemble of spin systems has potential applications in precision measurements @xcite and particle entanglement @xcite .
the measurement accuracy of atomic clocks for time and frequency metrology is limited by the quantum mechanical uncertainity of spin operators @xcite .
kitagawa and ueda @xcite proposed the generation of spin - squeezed states , which redistributes the uncertainity unevenly between the two components of the total angular momentum and the component with the reduced uncertainity can be utilized for precision measurements with applications in atomic clocks @xcite and quantum information processing @xcite.atomic spin squeezing requires nonlinear interactions between spins which can be realized by one - axis twisting and two - axis counter - twisting @xcite .
quantum entanglement / correlations among the spins is responsible for spin squeezing . in this context
it has been found in various models that quantum fisher information ( qfi ) together with spin squeezing(ss ) are connected to quantum entanglement @xcite . in the context of bose - einstein condensates ,
spin - squeezed states can be experimentally achieved @xcite .
the dicke model describes a large number of two - level atoms interacting with a single optical mode . above a certain critical light - matter coupling strength ,
leads to a quantum phase transition from a normal phase to a superradiant phase @xcite . the dicke model has been realized experimentally in a bec confined in an optical cavity @xcite . in this work
, we study the spin squeezing and quantum fisher information of the jaynes - cummings dicke model realizable in a two - componet atomic bec in a cavity , which has twisting and coupling in two angular momentum components @xcite .
approximate analytical expressions of spin squeezing and quantum fisher information are derived under the frozen spin approximation @xcite .
we show for the first time that both maximum ss and quantum entanglement exists in the superradiant phase near the critical point .
spin squeezing dynamics and chaos of the dicke model has been studied earlier @xcite while a recent study explored the influence of rabi oscillation energy on quantum fisher information and phase sensitivity of bec in a double - well potential @xcite .
exact numerical analysis of spin squeezing in different atomic systems has been performed earlier @xcite .
we consider a rb atomic bec optically trapped in an optical cavity as described in ref . @xcite . under the two - mode approximation , only the atomic states , @xmath0 ( @xmath1 ) ground state @xmath2 and @xmath3 ,
metastable state @xmath4 are considered in this model .
these two states @xmath2 and @xmath5 are coupled by an ancillary excited state @xmath6 ( @xmath7 ) , where @xmath8 and @xmath9 states coupled by the quantized optical cavity field strength @xmath10 and an external classical field strength @xmath11 respectively .
an effective two - level configuration can be obtained after adiabatically eliminating the ancillary excited state @xmath12 under the assumption of large one - photon detuning . following the work of @xcite , one obtains an effective jc - dicke hamiltonian for @xmath13 number of atoms @xmath14 where @xmath15 , @xmath16 , @xmath17 , @xmath18 and @xmath19 @xmath20 @xmath21 .
@xmath22 is the internal energy levels for atomic state @xmath23 , @xmath24 \phi_{l}(\vec{r})$ ] ( @xmath25 ) is the trapped frequency for the states @xmath26 and @xmath27 with @xmath28 being the trapped potential , m is the atomic mass and @xmath29 is the corresponding bec wavefunction .
the cavity mode annihilation ( creation ) operator is defined as @xmath30(@xmath31 ) with frequency @xmath32 and @xmath33 is the reduced effective coupling strength for the two - photon raman process . here
@xmath34 is the large single - photon detuning .
also @xmath35 is the frequency of the classical optical field . @xmath36 and @xmath37 with @xmath38 and @xmath39 , the intraspecies and interspecies @xmath40wave scattering lengths , respectively .
now taking @xmath41 , we rewrite the hamiltonian of eqn.(1 ) as @xmath42 in the bad cavity limit , the decay rate of cavity photons ( @xmath43 ) is much larger than any other rates of the system and the cavity field always follows the atomic dynamics adiabatically . under this assumption
, we can adiabatically eliminate the cavity mode and obtain a hamiltonian for the atomic degrees of freedom alone .
@xmath44,\ ] ] @xmath45 @xmath46
in this section , we will work in the heisenberg picture and study in detail the spin squeezing and quantum fisher information in the hamiltonian of eqn.(3 ) .
in particular , we will explore the dependence of ss and qfi on the atom - photon coupling strength .
following the criteria of kitagawa and ueda of spin squeezing @xcite , we introduce the squeezing parameter @xmath47 where @xmath48 is the minimal spin fluctuation in a plane perpendicular to the mean spin direction . when @xmath49 , the spin systems are uncorrelated while @xmath50 signifies spin squeezing and a signature of entanglement . now in order to study qfi in our system , we follow @xcite who in order to characterize multi - particle entanglement in an @xmath13 qubit state had introduced the following quantity @xmath51 ( termed as rmqfi ) , @xmath52},\ ] ] where , @xmath53 $ ] is the quantum fisher information and @xmath54 is an arbitrary direction . here , @xmath55 implies multi - particle entanglement . following @xcite , the expression for @xmath56
can be written for our system as , @xmath57 where , @xmath58 is the maximum variance in a plane perpendicular to the mean spin direction .
the hamiltonian ( 1 ) demonstrates a second - order superradiant quantum phase transition at the critical point @xmath59 . in order to determine the critical point @xmath59 , we start with the heisenberg equations of moton of the angular momentum operators @xmath60 , @xmath61 , @xmath62 and the steady state values of the cavity field operators @xmath63
. @xmath64 @xmath65 @xmath66 @xmath67 a steady state analysis for @xmath68 ( corresponds to the normal phase i.e. all atoms are in the lower state ) of the above mean field equations of motion ( for large @xmath13 ) yields the critical value as @xmath69 starting from the hamiltonian ( 3 ) , we now examine the generation and coherent control of spin squeezing and quantum fisher information .
we prepare the system to start from the lowest eigenstate of @xmath62 , @xmath70 . in order to investigate the quantum spin squeezing and quantum fisher information ,
we examine the heisenberg equations of motion of @xmath71 and @xmath72 derived from hamiltonian ( 3 ) , @xmath73 @xmath74 if @xmath75 , @xmath76 , then the external field forced the total spin to freeze along the negative @xmath77 direction i.e. @xmath78 .
this approximation is termed as frozen spin approximation @xcite which permits harmonic solutions of @xmath71 and @xmath72 , @xmath79 @xmath80 where , @xmath81 @xmath82 @xmath83 @xmath84 as evident from the above expressions , the analysis is valid in the superradiant phase i.e @xmath85 , since for @xmath86 , @xmath32 is not real .
when the system starts from @xmath87 , the only non - vanishing spin component is @xmath88 because @xmath89 @xmath90 @xmath91 @xmath90 @xmath92 at all times .
now with the help of @xmath71 and @xmath72 , we can investigate quantum spin squeezing and quantum fisher information .
if @xmath93 @xmath90 @xmath94 @xmath95 , i.e. there is no correlation between @xmath71 and @xmath72 , the increase or reduced spin fluctuations occur either in the @xmath96 or @xmath97 direction .
however in our case , @xmath98 consequently since @xmath99+@xmath94 @xmath100 @xmath92 , the reduced spin fluctuations neither occurs along @xmath96 nor along @xmath97 direction .
now in order to determine the increased and reduced spin fluctuation directions , we introduce the spin component , @xmath101 where the unit vector @xmath102 , with @xmath103 as the angle between @xmath96 axis and @xmath104 .
since @xmath89 @xmath90 @xmath105 @xmath90 @xmath92 , the fluctuation of the spin component @xmath106 reads , @xmath107 the optimally squeezed angle @xmath108 is obtained via minimizing @xmath109 with respect to @xmath103 .
this yields , @xmath110 .
the squeezing parameter @xmath111 and rmqfi @xmath51 are determined as , @xmath112^{1/2},\ ] ] @xmath113^{1/2}.\ ] ] [ cols="^,^ " , ] fig.1 shows the plot of spin squeezing parameter @xmath111 ( left plot ) and the quantum fisher information @xmath51 ( right plot ) as a function of @xmath114 for two values of @xmath115 ( solid line ) and @xmath116(dashed line ) .
clearly for @xmath115 , both @xmath111 and @xmath51 has minimum values compared to that for @xmath116 .
this clearly demonstrates that near the critical point in the superradiant phase maximum spin squeezing and maximum quantum entanglement occurs . the close connection between spin squeezing and entanglement has been demonstrated earlier @xcite . as a function of times , periodically maximum squeezing and entanglement reoccurs .
the fig.1 also illustrates that initially at @xmath117 , there is no spin squeezing but as the dynamics evolves , spin squeezing occurs .
the atom - photon coupling then emerges as coherent handle to control the spin squeezing and quantum entanglement.in a recent experiment , bell correlations was reported between pseudo spins of about 480 atoms in a spin - squeezed bec @xcite .
bell correlations also suggests entanglement .
the present work is highly relevant to quantum information processing and precision spectroscopy .
in conclusion , we have investigated the spin squeezing and quantum fisher information dynamics in a nonlinear system composed of a bose - einstein condensate in an optical cavity .
the nonlinearity emerges from the twisting and coupling of the @xmath71 and @xmath72 components of the angular momentum.the system represents a variant of the dicke model which undergoes a dicke quantum phase transition from the normal phase to a superradiant phase above a critical atom - photon coupling strength.it is found that near the critical point in the superradiant phase maximum spin squeezing can be achieved . at the same point where @xmath111 is minimum , @xmath51 is also found to be minimum which signifies that near the critical point in the superradiant phase , maximum quantum entanglement exists .
a. bhattacherjee acknowledges financial support from the university grants commission , new delhi under the ugc - faculty recharge programme .
deepti sharma acknowledges financial assistance from jawaharlal nehru university . c. gross , j. phys .
b : at . mol .
. phys . * 45 * , 103001 ( 2012 ) .
j. ma , x. wang , c.p .
sun and f. nori , phys .
rep . * 509 * , 89 ( 2011 ) .
d. j. wineland , j. j. bollinger , w. m. itano , f. l. moore and d. j. heinzen , phys . rev .
a * 46 * , r6797 ( 1992 ) .
d. j. wineland , j.j .
bollinger , w. m. itano and d. j. heinzen , phys .
a * 50 * , 67 ( 1994 ) .
e. s. polzik , nature * 453 * , 45 ( 2008 ) .
a. d. cronin , j. schmiedmayer , d. e. pritchard , rev .
modern phys .
* 81 * , 1051 ( 2009 ) .
a. sorensen , l. m. duan , j. i. cirac , p. zoller , nature * 409 * , 63 ( 2001 ) .
n. bigelow , nature * 409 * , 27 ( 2001 ) .
o. guehne , g. toth , phys . rep . * 474 * , 1 ( 2009 ) . m. kitagawa and m. ueda , phys .
a * 47 * , 5138 ( 1993 ) .
q. a. turchette et al . , phys .
lett . * 81 * , 3631 ( 1998 ) . v. meyer et al .
lett . * 86 * , 5870 ( 2001 ) .
d. leibfried et al .
, science * 304 * , 1476 ( 2004 ) .
x. wang , b. c. sanders , phys .
a * 68 * , 012101 ( 2003 ) .
j. k. korbicz , j. i. cirac , m. lewenstein , phys .
lett . * 95 * , 120502 ( 2005 ) ; j. k. korbicz et al .
a * 74 * , 052319 ( 2006 ) .
s. yi , h. pu , phys . rev .
a * 73 * , 023602 ( 2006 ) .
g - j hu and x - x hu , int .
j. theor phys . , * 53 * , 533 ( 2014 ) . y - h jiang and s - s li , int . j. theor .
phys . , * 52 * , 2826 ( 2013 ) .
s - s li , h - g yi and r - h chen , int . j. theor . phys . , * 52 * , 1175 ( 2013 ) . c. orzel , a. k. tuchman , m.l .
fenselau , m. yasuda and m. a. kasevich , science * 291 * 2386 ( 2001 ) .
m. f. riedel , nature * 464 * 1170 ( 2010 ) .
j. esteveeve , c. gross , a. weller , s. giovanazzi and m. k. oberthaler , nature * 455 * 1216 ( 2008 ) .
w. muessel et al . , arxiv:1507.02930 .
c. emary and t. brandes , phys .
e , * 67 * , 066203 ( 2003 ) .
f. brennecke , t. donner , s. ritter , t. bourdel , m. k hl and t. esslinger , nature , * 450 * , 268 ( 2007 ) .
y. li , p. zhang and z. d. wang , eur .
j. d , * 58 * , 379 ( 2010 ) .
s. c. li , l - b .
fu and j. liu , phys .
a , * 84 * , 053610 ( 2011 ) . c. k. law , h. t. ng , and p. t. leung , phys .
a , * 63 * , 055601 ( 2001 ) .
a. b. bhattacherjee , v. ranjan and m. mohan , int .
b * 17 * , 2579 ( 2003 ) .
l. song , d. yan , j. ma and x. wang , phys . rev .
a. , * 79 * , 046220 ( 2009 ) .
zhu , g - q .
huang and c - l .
luo , int .
j. theor phys . ,
* 55 * , 715 ( 2016 ) .
g. r. jin and s. w. kim , phys .
a , * 76 * , 043621 ( 2007 ) .
g. r. jin and c. k.law , phys .
a , * 78 * , 063620 ( 2008 ) .
w. wang and y. w. lu , j. phys .
b : at . mol .
* 43 * , 045301 ( 2010 ) .
g.chen , x. wang , j.q . liang and z.d .
wang , phys .
a , * 78 * , 023634 ( 2008 ) . l. pezze and a. smerzi , phys .
* 102 * , 100401 ( 2009 ) .
liu , l - h .
zhang and c - j .
. phys . * 49 * , 2463 ( 2010 ) .
r. schmied at al . ,
science , * 352 * , 441 ( 2016 ) . | we investigate spin squeezing ( ss ) and the quantum fisher information ( qfi ) for the jaynes - cummings dicke ( jc - dicke ) model in a two component atomic bose - einstein condensate ( bec ) inside an optical cavity .
analytical expressions for spin squeezing and the reciprocal of the quantum fisher information per particle ( rmqfi ) are derived using the frozen spin approximation .
it is shown that in the superradiant phase near the critical point , maximum squeezing and maximum quantum entanglement occurs .
the present study is relevant to quantum information processing and precision spectroscopy .
* keywords . *
spin squeezing , quantum fisher information , jaynes - cummings dicke model |
a nonplanar graph @xmath0 is called _ almost - planar _ if for every edge @xmath1 of @xmath0 , at least one of @xmath2 and @xmath3 is planar .
the following is an equivalent definition . for any specified set @xmath4 of graphs ,
let us call a graph @xmath4-_free _ if it does not contain any graph in @xmath4 as a minor . using this terminology
, we can say that a graph @xmath0 is almost - planar if and only if @xmath0 is not @xmath5-free but for every edge @xmath1 of @xmath0 , at least one of @xmath2 and @xmath3 is @xmath5-free .
the second definition is a special case of a concept in matroid theory .
let @xmath4 be a specified set of matroids .
@xmath4-free matroids _ be defined analogous to @xmath4-free graphs .
then a matroid @xmath6 is called @xmath4-_fragile _ if @xmath6 is not @xmath4-free but for every element @xmath1 of @xmath6 , at least one of @xmath7 and @xmath8 is @xmath4-free . in recent years
, fragility has received a lot of attention due to its connection with representability .
there are characterizations of @xmath4-fragile matroids for several choices of @xmath4 . among these ,
the first is a characterization of @xmath5-fragile 3-connected graphic matroids .
this result was obtained by gubser in his dissertation @xcite , which was done before the term @xmath4-fragile " was introduced . in the following discussion , we will use the term almost - planar graph " instead of @xmath5-fragile graph " . to state the characterization of gubser
we need a few definitions .
let @xmath9 be an integer .
a _ wheel _ of size @xmath10 , denoted by @xmath11 , is the graph obtained from a cycle of length @xmath10 by adding a vertex and joining it to all vertices of the cycle .
double wheel _ of size @xmath10 , denoted by @xmath12 , is the graph obtained from a cycle of length @xmath10 by adding two adjacent vertices and joining them to all vertices of the cycle .
a _ mbius ladder _ of length @xmath10 , denoted by @xmath13 , is obtained from a cycle of length @xmath14 by joining opposite pairs of vertices on the cycle .
finally , let @xmath15 denote the set of all graphs constructed by identifying three triangles from three wheels .
in other words , each graph @xmath16 admits a partition @xmath17 of its vertex set such that @xmath18 $ ] is a triangle , @xmath19 $ ] is a wheel @xmath20 , and @xmath0 has no edges other than those in these three wheels .
[ thm : main ] let @xmath0 be a 3-connected nonplanar graph
. then the following are equivalent .
+ ( i ) @xmath0 is almost - planar ; + ( ii ) @xmath0 is a minor of a double wheel , a mbius ladder , or a graph in @xmath15 ; + ( iii ) @xmath0 is @xmath21-free , where @xmath22 shown in figure [ fig : exminors ] . , @xmath23 , @xmath24 , @xmath25 , and @xmath26 we remark that terminology @xmath27 is inherited from @xcite .
we also remark that our formulation of this theorem is slightly different from that given in @xcite .
first , we modified the definition of graphs in @xmath15 .
we make the change because our description better reveals the structure of these graphs .
second , set @xmath21 given in @xcite consists of eight graphs , instead of five graphs .
these eight include the five listed in figure [ fig : exminors ] and another three graphs , which were denoted by @xmath28 , @xmath29 , and @xmath30 .
since each of these three extra graphs contains @xmath31 as a minor , they are removed from our statement . by seymour
s splitter theorem @xcite , every almost - planar graph can be generated from @xmath32 and @xmath33 by repeatedly adding edges and splitting vertices , although not every graph constructed this way has to be almost - planar .
therefore , to prove that every almost - planar graph must be one of the three types listed in ( ii ) , one only needs to show that any extension of any graph of the three types either results in another graph of the three types or contains a member of @xmath21 as a minor .
this is how the proof went in @xcite .
the author divided the analysis into fifteen cases , depending on which graph is extended and how the graph is extended . since the case checking is lengthy , the published proof @xcite includes only two of the fifteen cases .
the main purpose of this paper is to provide a short proof of theorem [ thm : main ] . at the end of the paper
, we will also discuss the structure of almost - planar graphs that are not 3-connected .
this discussion corrects a few flaws appearing in @xcite .
we close this section by making two more remarks .
notice that graphs in @xmath15 can be naturally divided into three groups , depending on how the three hubs are distributed on the common triangle .
when none of the hubs are identified , the resulting graph is in fact a minor of a mbius ladder , as illustrated in figure [ fig : mobius ] .
what this means is that we could define @xmath15 differently so that these graphs are not included in @xmath15 .
we choose to leave them in @xmath15 since it makes the definition more clean .
this is another difference between our formulation of theorem [ thm : main ] and the formulation given in @xcite .
are minors of a mbius ladder ] next , for each @xmath34 , let @xmath35 be obtained from @xmath0 by adding a new vertex @xmath36 and a new edge @xmath1 between @xmath36 and @xmath0 .
then @xmath35 is not almost - planar since both @xmath37 and @xmath38 are nonplanar .
notice that @xmath23 contains @xmath39 , while @xmath40 and @xmath31 contain @xmath41 .
hence @xmath40 , @xmath23 , and @xmath31 are not minimally non - almost - planar nonplanar graphs . however , by theorem [ thm : main ] , within the universe of 3-connected graphs , these three graphs are minimal and in fact , members of @xmath21 are precisely the only five minimal graphs . in the last section of this paper ,
we determine all minimal graphs without imposing any connectivity .
in this section we present a few lemmas that will be used in our main proof .
let @xmath42 be an integer .
a _ @xmath43-separation _ of a graph @xmath0 is an unordered pair @xmath44 of proper induced subgraphs of @xmath0 such that @xmath45 and @xmath46 . the following is an immediate corollary of ( 2.4 ) of @xcite .
let @xmath47 be a cubic vertex of a nonplanar graph @xmath0 .
suppose @xmath0 has no @xmath43-separation @xmath44 with @xmath48 and such that some @xmath49 contains @xmath47 and all its three neighbors .
then @xmath0 has a subgraph @xmath50 such that @xmath50 is a subdivision of @xmath33 and @xmath47 is cubic in @xmath50 .
[ lem : deg3 ] a 3-connected graph @xmath0 with @xmath51 is called _ internally @xmath52-connected _ if for every 3-separation @xmath53 of @xmath0 , exactly one of @xmath54 is @xmath55 .
notice that no cubic vertex belongs to a triangle in an internally 4-connected graph .
we will use this property repeatedly .
let @xmath56 be the graph obtained from a cycle of length @xmath10 by joining all pairs of vertices of distance two on the cycle .
let @xmath0 be an internally 4-connected nonplanar minor of a mbius ladder
. then @xmath57 or @xmath58 for some integer @xmath9 .
[ lem : minorm ] let @xmath59 be the smallest integer such that @xmath0 is a minor of @xmath60 .
let @xmath60 consist of a hamilton cycle @xmath61 and pairwise crossing chords .
the minimality of @xmath43 implies that no chord is deleted in obtaining @xmath0 .
moreover , since @xmath0 is nonplanar , no chord is contracted and no edge of @xmath61 is deleted .
so @xmath0 is obtained by contracting some edges of @xmath61 .
since no cubic vertex of @xmath0 belongs to a triangle , it follows that either @xmath62 or @xmath63 is odd and @xmath64 .
we remark that the lemma holds even if we drop the nonplanarity assumption .
the next is a theorem of maharry and robertson @xcite on @xmath65-free graphs .
an _ alternating double wheel _ of length @xmath14 ( @xmath66 ) , denoted by @xmath67 , is obtained from a cycle @xmath68 by adding two new adjacent vertices @xmath69 such that @xmath70 is adjacent to @xmath71 for all @xmath72 and @xmath73 .
if an internally @xmath52-connected graph @xmath0 is @xmath65-free then at least one of the following holds . + ( i ) @xmath0 is planar ; + ( ii ) @xmath74 and @xmath75 is edgeless , for some distinct @xmath76 ; + ( iii ) @xmath77 or @xmath67 for some @xmath9 ; + ( iv ) @xmath0 is the line graph of @xmath33 ; + ( v ) @xmath0 has fewer than eight vertices . [
lem : m4free ] for any two vertices @xmath78 of a path @xmath79 , let @xmath80 $ ] denote the subpath of @xmath79 between @xmath78 .
let @xmath81 be a graph .
if @xmath0 is a subdivision of @xmath81 , then _ branch vertices _ of @xmath0 are vertices of @xmath0 that correspond to vertices of @xmath81 , and _ arcs _ of @xmath0 are paths of @xmath0 that correspond to edges of @xmath81 .
a _ triad addition _ of @xmath81 is obtained by adding a new vertex @xmath36 to @xmath81 and joining @xmath36 to three distinct vertices of @xmath81 .
[ lem : t - add ] let @xmath81 be a simple graph with @xmath82 . if a @xmath83-connected graph @xmath0 has a subgraph @xmath50 such that @xmath50 is a subdivision of @xmath81 and @xmath84 , then @xmath0 contains a triad addition of @xmath81 as a minor .
let @xmath85 and let @xmath86 be three paths of @xmath0 from @xmath36 to @xmath50 such that they are disjoint except for @xmath36 .
let @xmath87 ( @xmath88 ) be the other end of @xmath89 .
we first prove that @xmath90 can be chosen so that @xmath91 are not contained in a single arc of @xmath50 .
suppose @xmath50 has an arc @xmath92 containing @xmath91 .
let @xmath78 be the two ends of @xmath92 and let @xmath93 be the order that they appear in @xmath92 .
suppose @xmath90 are chosen such that @xmath94\cup a[y , v_3]|$ ] is minimized .
since @xmath0 is 3-connected , @xmath95 has a path @xmath96 from @xmath97){\backslash}\{v_1,v_3\}$ ] to @xmath98)$ ] . by the minimality of @xmath94\cup a[y , v_3]|$ ] ,
the other end of @xmath96 must belong to @xmath99 .
it follows that @xmath100 contains a subdivision of @xmath81 and three required paths .
now we assume that @xmath91 are not contained in a single arc .
it is easy to see that this property can be preserved when we contract @xmath50 to @xmath81 . at the end
we obtain a triad addition of @xmath81 as a minor .
before proving theorem [ thm : main ] we establish three lemmas , which are the main parts of our proof .
let @xmath44 be a 3-separation of @xmath0 .
for @xmath72 , let @xmath101 be obtained from @xmath49 by adding all missing edges between vertices of @xmath102 ; let @xmath103 be obtained from @xmath49 by adding a new vertex @xmath87 and joining @xmath87 to all three vertices of @xmath102
. in the rest of this paper , we use @xmath104 to refer to the graph @xmath24 and @xmath105 to refer to the graph @xmath25 since these names better reflect the fact that @xmath24 and @xmath25 are formed from @xmath33 and @xmath32 , respectively , by adding a handle edge .
let @xmath0 be @xmath83-connected and @xmath106-free . if @xmath44 is a @xmath83-separation of @xmath0 such that @xmath107 is nonplanar , then @xmath108 is a wheel .
[ lem : wheel ] let @xmath109 and let @xmath47 be the cubic vertex added to @xmath110 in the formation of @xmath107 . since @xmath0 is 3-connected and @xmath107 is nonplanar , by lemma [ lem : deg3 ] , @xmath107 contains a subdivision of @xmath33 such that @xmath47 is cubic in the subdivision .
it follows that any cycle of @xmath111 disjoint from @xmath112 results in a @xmath104 minor in @xmath0 . thus @xmath113 is a forest .
let @xmath114 .
since @xmath0 is 3-connected , @xmath111 contains three induced paths @xmath115 from @xmath116 to each of @xmath91 , respectively , such that the three paths are disjoint except for @xmath116 .
if @xmath111 has a vertex outside @xmath117 , then @xmath0 has a @xmath118 minor .
so @xmath119 . if @xmath86 are all single edges , then @xmath120 is a wheel @xmath121 .
otherwise , suppose without loss of generality that @xmath122 .
let @xmath123 be any internal vertex of @xmath124 .
since @xmath124 is an induced path and @xmath0 is 3-connected , @xmath123 must be adjacent to some vertices outside @xmath124 . since @xmath125 is a tree , the only possible neighbors of @xmath123 outside @xmath124 are @xmath126 and @xmath127 . if @xmath123 is adjacent to both @xmath128 then @xmath0 has a @xmath41 minor . so @xmath123 is adjacent to exactly one of @xmath128 . if @xmath129 are distinct internal vertices of @xmath124 such that both @xmath130 and @xmath131 are edges of @xmath0 , then @xmath0 again has a @xmath41 minor .
therefore , all internal vertices of @xmath124 are cubic and they have a common neighbor outside @xmath124 , which we may assume to be @xmath126 . since @xmath0 is @xmath41-free
, we deduce that @xmath132 and no internal vertex of @xmath133 is adjacent to @xmath134 .
so all vertices of @xmath135 are cubic and adjacent to @xmath126 .
therefore , @xmath111 consists of path @xmath136 , all edges from @xmath126 to @xmath137 , and possibly edges between @xmath91 , which implies that @xmath108 is a wheel .
let @xmath0 be connected and @xmath138-free .
if @xmath0 contains an @xmath65-minor then @xmath0 is a minor of @xmath13 for some @xmath139 .
[ lem : hasm4 ] since @xmath65 is cubic , it is a topological minor of @xmath0 , so @xmath0 has a subgraph @xmath81 isomorphic to a subdivision of @xmath65 . since @xmath0 is @xmath41-free , @xmath81 must be a spanning subgraph and no chord of @xmath65 is subdivided .
it follows that the rim of @xmath81 is a hamilton cycle @xmath61 of @xmath0 .
let @xmath140 be the eight cubic vertices of @xmath81 , listed in the order they appear on @xmath61 .
we denote the path of @xmath61 from @xmath87 to @xmath141 by @xmath142 , where @xmath143 .
( in this proof the indices are taken modulo 8 . )
let @xmath1 be any chord of @xmath61 .
we prove that there exists @xmath144 such that the two ends of @xmath1 belong to @xmath142 and @xmath145 , respectively .
we first observe that no @xmath142 contains both ends of @xmath1 because then @xmath146 contains a @xmath41 minor .
next , assume one end of @xmath1 is an internal vertex of some @xmath142 .
if the other end of @xmath1 does not belong to @xmath145 , it is straightforward to verify that @xmath146 can be contracted to @xmath147 , for some @xmath148 , and thus @xmath0 contains a @xmath41 minor .
so both ends of @xmath1 are cubic vertices of @xmath81 , implying @xmath149 ( @xmath150 ) , which satisfies the requirement .
to finish the proof , we only need to show that any two non - adjacent chords @xmath151 must cross . by what we proved in the last paragraph , this is clear if the four ends of @xmath151 are not contained in @xmath152 for any @xmath144 .
if all ends of @xmath151 are contained in @xmath152 for some @xmath144 and if @xmath151 do not cross each other , then @xmath153 can be contracted to @xmath154 , which contains a @xmath105 minor .
therefore , @xmath151 must cross and that completes our proof .
let @xmath0 be internally @xmath52-connected , nonplanar , and @xmath155-free
. then @xmath0 is @xmath13 or @xmath156 or @xmath12 or @xmath67 for some @xmath9 .
[ lem : int4conn ] if @xmath0 has an @xmath65-minor then the result follows immediately from lemmas [ lem : hasm4 ] and [ lem : minorm ] .
so we assume that @xmath0 is @xmath65-free .
since @xmath0 is nonplanar , one of ( ii)-(v ) of lemma [ lem : m4free ] must hold .
if ( iii ) holds then we are done .
case ( iv ) does not hold since @xmath157 contains a @xmath104 minor , as shown in figure [ fig : lk33 ]
. is a minor of @xmath157 , and @xmath41 is a minor of cube@xmath158 suppose ( ii ) holds .
let @xmath159 .
note that each @xmath70 has degree either 3 or 4 .
if two or more of them are of degree 4 then @xmath41 is a minor of @xmath0 .
if every @xmath70 is cubic , since no two of them have the same set of neighbors , it follows that @xmath160 and @xmath0 is the cube , which does not happen since @xmath0 is nonplanar .
so we assume that @xmath161 has degree 4 , while every other @xmath70 is cubic . as a result , @xmath160 or 5 . if @xmath160 then @xmath0 is @xmath162 .
if @xmath163 then @xmath164 is the cube and so @xmath0 has a @xmath41 minor , as shown in figure [ fig : lk33 ] .
it remains to consider case ( v ) .
if @xmath165 , since @xmath0 is nonplanar , we have @xmath166 . if @xmath167 , then @xmath0 contains a @xmath33 subgraph , as every 3-connected nonplanar graph with more than five vertices contains a @xmath33 minor .
since no cubic vertex of @xmath0 belongs to a triangle , it follows that either @xmath168 , or @xmath0 contains at least two edges from each color class of @xmath33 .
if @xmath0 contains no other edges then @xmath169 ; if @xmath0 contains at least one other edge then @xmath0 contains @xmath39 .
if @xmath170 , we may assume that @xmath0 contains a subgraph obtained from @xmath33 by subdividing an edge exactly once .
let @xmath171 and @xmath172 be the two color classes of @xmath33 and let @xmath36 be the vertex subdividing edge @xmath173 . since @xmath36 is not a cubic vertex that belongs to a triangle , we deduce that @xmath36 has degree at least 4 . if @xmath36 is adjacent to both @xmath174 or both @xmath175 , then @xmath0 contains @xmath41 .
so the degree of @xmath36 is 4 , and we may assume that @xmath36 is adjacent to @xmath176 and @xmath177 .
let @xmath81 be the subgraph of @xmath0 obtained from the subdivision of @xmath33 by adding edges @xmath178 and @xmath179 .
now @xmath180 and @xmath181 are in triangles , so they must have degree at least 4 in @xmath0 . if @xmath0 contains the edge @xmath182 , then @xmath0 contains @xmath41 .
so @xmath180 must be adjacent to @xmath176 or @xmath183 , and @xmath181 must be adjacent to @xmath177 or @xmath184 .
it follows that @xmath183 and @xmath184 are in triangles and thus they must also have degree at least 4 in @xmath0 .
notice that @xmath185 , which contains @xmath105 .
so @xmath0 is obtained from @xmath81 either by adding edges @xmath186 and @xmath187 , which leads to @xmath188 , or by adding paths @xmath189 and @xmath190 , which leads to @xmath191 .
now we are ready to prove theorem [ thm : main ] .
[ proof of theorem [ thm : main ] ] we prove implications ( ii ) @xmath192 ( i ) @xmath192 ( iii ) @xmath192 ( iv ) @xmath192 ( ii ) , where ( iv ) is the statement \(iv ) _ @xmath0 is @xmath155-free .
_ \(ii ) @xmath193 ( i ) : note that , if @xmath194 is the class of all planar and almost - planar graphs , then @xmath194 is closed under taking minors .
therefore , we only need to verify that each mbius ladder , each double wheel , and each graph in @xmath15 is almost - planar . in @xmath195
deleting a rim edge or contracting a chord results in a planar graph , so @xmath13 is almost - planar . in @xmath12
, deleting the axle or a rim edge or contracting a spoke results in a planar graph , so @xmath12 is almost - planar . in any @xmath16 , contracting a spoke or an edge in the common triangle or deleting a rim edge results in a planar graph , so @xmath0 is almost planar .
\(i ) @xmath193 ( iii ) : since @xmath194 is closed under taking minors , we only need to show that each @xmath196 is not in @xmath194 . to do this , we only need to find an edge @xmath1 such that both @xmath197 and @xmath198 are nonplanar . in @xmath40
every edge satisfies the requirement . in @xmath23 and @xmath31 ,
any edge incident with the top cubic vertex ( of the drawing shown in figure [ fig : exminors ] ) has the required property . in @xmath104 and @xmath105 ( @xmath24 and @xmath29 , respectively )
, the handle edge meets the requirement .
\(iii ) @xmath192 ( iv ) : since @xmath199 , we only need to show that @xmath0 is @xmath200-free .
suppose otherwise .
then @xmath0 has a vertex @xmath36 such that @xmath201 is nonplanar , which implies that @xmath0 has a subgraph @xmath50 such that @xmath84 and @xmath50 is a subdivision of @xmath32 or @xmath33 . by lemma [ lem : t - add ]
, @xmath0 contains a triad addition of @xmath32 or @xmath33 as a minor .
however , @xmath23 is the only triad addition of @xmath32 , and @xmath202 are the only triad additions of @xmath33 .
thus ( iii ) is violated no matter what is the outcome , which proves ( iv ) .
\(iv ) @xmath193 ( ii ) : assume @xmath0 is @xmath155-free
. we will show @xmath0 is a minor of a double wheel , a mbius ladder , or a graph in @xmath15 . recall that a _ 3-sum _ of two graphs @xmath54 is obtained by identifying a triangle of @xmath110 with a triangle of @xmath111 , and then deleting 0 , 1 , 2 , or 3 of the three identified edges .
if @xmath0 has three vertices whose deletion results in more than two components , since @xmath0 is @xmath41-free , there should be exactly three components and thus @xmath0 can be expressed as a 3-sum of three graphs @xmath203 over a common triangle . for each @xmath88 , by applying lemma [ lem : wheel ] to the 3-separation of @xmath0 determined by @xmath204 , where @xmath205 , we conclude that @xmath49 is a wheel , which implies that @xmath16 . in the rest of this proof
we assume that any three vertices separate @xmath0 into at most two components , and we call such a graph _ strongly connected_. we claim that @xmath0 has an internally 4-connected nonplanar minor @xmath50 such that @xmath0 is obtained by 3-summing wheels to distinct triangles of @xmath50 .
suppose the claim is false .
we consider a counterexample @xmath0 with @xmath206 as small as possible .
since internally 4-connected graphs are not counterexamples , @xmath0 must have a 3-separation @xmath44 such that neither part is @xmath55 . since @xmath0 is nonplanar , at least one of @xmath107 and @xmath207 is nonplanar . without loss of generality ,
suppose @xmath107 is nonplanar .
let us choose such a 3-separation with @xmath110 as small as possible .
since @xmath208 , @xmath209 is a 3-connected minor of @xmath0 .
notice that @xmath209 is also strongly connected , because any bad 3-cut of @xmath209 would lead to a bad 3-cut of @xmath0 .
moreover , @xmath209 is nonplanar , because otherwise , since @xmath107 is nonplanar , the common triangle @xmath210 of @xmath209 and @xmath108 would not be a facial triangle of @xmath209 , which implies that @xmath211 has more than two components , contradicting the strong connectivity of @xmath0 .
therefore , @xmath209 satisfies all the assumptions of our claim and thus it is not a counterexample , as it is smaller than @xmath0 .
so @xmath209 is obtained from an internally 4-connected nonplanar minor @xmath50 by 3-summing wheels to different triangles of @xmath50 . by the minimality of @xmath110 ,
triangle @xmath210 must be contained in @xmath50 and thus @xmath0 is not a counterexample because @xmath0 is also obtained from @xmath50 by 3-summing wheels to distinct triangles of @xmath50 , since @xmath108 is a wheel by lemma [ lem : wheel ] .
this contradiction proves our claim . by lemma [ lem : int4conn ] , we know @xmath212 or @xmath156 or @xmath12 or @xmath67 for some @xmath9 .
note that @xmath156 is a minor of @xmath213 and @xmath67 is a minor of @xmath214 so the theorem holds if @xmath215 .
next , suppose at least one wheel is added to @xmath50 . since @xmath13 and @xmath67 do not have any triangles , we only need to consider all possible ways to add wheels to triangles of @xmath156 and @xmath12 .
suppose @xmath216 ( @xmath9 ) .
observe that the 3-sum of @xmath50 and a wheel is a graph with an @xmath65 minor . by lemma [ lem : hasm4 ]
, @xmath0 is a minor of a mbius ladder and thus the theorem holds .
suppose @xmath217 ( @xmath139 ) is obtained by adding two adjacent vertices @xmath161 and @xmath218 to a cycle @xmath219 and joining each of @xmath69 to all @xmath87 ( indices are taken modulo @xmath10 in this paragraph ) . let @xmath81 be a 3-sum of @xmath50 and a wheel @xmath220 over a triangle @xmath125 . if @xmath125 is @xmath221 then @xmath81 contains a @xmath41 minor since @xmath222 contains a @xmath33 minor .
there is only one other type of triangle in @xmath50 , so suppose @xmath125 is @xmath223 .
if @xmath224 is an edge of @xmath81 then @xmath81 contains a @xmath105 minor , which can be seen by considering the graph formed by cycle @xmath61 , cycle @xmath225 , edges @xmath226 , and three paths from @xmath227 to @xmath125 .
the same configuration also shows that if @xmath228 then the hub of @xmath220 is not identified with @xmath134 or @xmath126 .
so the effect of 3-summing @xmath220 to @xmath50 is to subdivide an edge @xmath229 and to join all the subdividing vertices with some @xmath230 .
many wheels can be added to @xmath12 in the same fashion . thus 3-summing wheels
to @xmath12 either creates a @xmath41 or @xmath105 minor or results in a minor of a double wheel .
finally , let @xmath231 ( note @xmath232 ) .
we assume that @xmath0 is @xmath65-free because otherwise the theorem holds by lemma [ lem : hasm4 ] .
let @xmath233 be the set of _ summing triangles _
, which are triangles of @xmath50 over which a wheel is 3-summed to @xmath50 in obtaining @xmath0 .
to finish off this last case we first determine @xmath233 and then determine how wheels are 3-summed to each member of @xmath233 .
let @xmath234 and let @xmath235 and @xmath236 .
we claim that , up to a permutation , either @xmath237 or @xmath238 . first we observe that turning triangles 123 , 124 , 125 into triads in @xmath50 results in @xmath104 , and turning 123 , 134 , 145 into triads results in @xmath65 .
therefore , \{123 , 124 , 125 } and \{123 , 134 , 145 } are not subsets of @xmath233 because otherwise @xmath0 contains @xmath104 or @xmath65 as a minor .
similarly \{123 , 124 , 134 , 234@xmath239 because otherwise @xmath0 contains a @xmath41 minor ( deleting any of the triads still leaves a nonplanar graph ) .
now it is straightforward to verify that either some vertex @xmath144 belongs to at least three summing triangles , in which case @xmath237 ( up to a permutation ) , or every vertex @xmath144 belongs to at most two summing triangles , in which case @xmath238 ( up to a permutation ) , and thus the claim is proved .
suppose a wheel @xmath220 is 3-summed to @xmath50 over @xmath240 . if @xmath241 then at least one of the three edges of @xmath210 is not in @xmath0 , because otherwise @xmath0 contains a @xmath39 minor . if @xmath242 and if the hub of @xmath220 is identified with @xmath47 then @xmath243 because otherwise @xmath0 contains a @xmath105 minor ( figure [ fig : k5a](i ) ) .
contains @xmath105 , @xmath41 , or @xmath104 ] we call @xmath47 a _ hub _ of @xmath210 if either @xmath242 and the hub of @xmath220 is identified with @xmath47 , or @xmath241 and @xmath47 meets all edges of @xmath0 between @xmath244 .
the above observations imply that every summing triangle has a hub , and if @xmath47 is a hub then @xmath243 . in the following we make two observations .
let @xmath245 .
\(a ) _ if @xmath246 then either @xmath247 or @xmath248 is a hub of @xmath210 , where we assume @xmath249 . _ + ( b ) _ if @xmath250 then either @xmath247 or @xmath248 is a hub of @xmath210 , where we assume @xmath251 . _
to prove these two statements we assume that neither @xmath247 nor @xmath248 is a hub of @xmath210 .
then either @xmath242 and the hub of @xmath220 is identified with @xmath47 , or @xmath241 and both @xmath252 are edges of @xmath0 . in both cases ( a ) and ( b ) , if @xmath241 then @xmath0 contains a @xmath41 minor ( figure [ fig : k5a](ii - iii ) ) , and if @xmath242 then @xmath0 contains a @xmath104 minor ( figure [ fig : k5a](iv - v ) ) .
these contradictions prove both ( a ) and ( b ) .
. then none of the edges 23 , 24 , 34 are in @xmath0 , because otherwise @xmath0 contains a @xmath41 minor .
this observation and ( a ) imply that 1 is a hub for all three summing triangles . as a result , @xmath253 is a cycle and thus @xmath0 is a minor of some @xmath12 .
suppose @xmath238 .
we assume without loss of generality that @xmath254 .
we first note that 14 is not an edge of @xmath0 because otherwise @xmath0 contains a @xmath104 minor . this observation and ( a ) and ( b ) imply that 5 is a hub of 145 , and hubs of 123 , 234 are in @xmath255 . if 2 ( or 3 ) is a hub for both 123 and 234 , then @xmath0 is a minor of some @xmath12 ( figure [ fig : k5](i ) ) ; if 2 is a hub of 123 and 3 is a hub of 234 , then @xmath0 is a minor of a mbius ladder ( figure [ fig : k5](ii ) ) .
now the proof of theorem [ thm : main ] is complete .
is a minor of @xmath12 or a mbius ladder ]
in the published version @xcite of @xcite , graphs of low connectivity are also considered .
some of the statements in @xcite are not accurate . its second main theorem ( theorem 2.2 )
states : _ a graph is neither planar nor almost - planar if and only if it has a @xmath256-minor . _ as we pointed out in the introduction , @xmath41 is a counterexample to this statement . in this section
we prove a corrected version of this theorem . for any graph @xmath0 ,
let @xmath257 be obtained from @xmath0 by adding two adjacent new vertices , and let @xmath258 be obtained from @xmath0 by deleting all its isolated vertices ( assuming that @xmath259 ) .
let @xmath260 be the set of edges @xmath1 of @xmath0 such that @xmath2 is planar .
[ thm : lowconn ] the following are equivalent for any nonplanar graph @xmath0 .
+ ( i ) @xmath0 is almost - planar ; + ( ii ) @xmath258 is obtained from a 3-connected almost - planar @xmath81 by subdividing edges in @xmath261 ; + ( iii ) @xmath0 is @xmath262-free , where @xmath263 .
one possible sharpening of ( ii ) is to describe @xmath261 explicitly .
if @xmath81 is a mbius ladder or a double wheel or a graph in @xmath15 , then it is easy to determine @xmath261 . for each nonplanar minor @xmath264 of @xmath81
, one could also describe @xmath265 because the structure of @xmath81 is simple enough .
however , we choose not to include it in the paper since such a description is tedious and its derivation is straightforward . statement ( ii ) is also touched in @xcite , but the treatment is not rigorous . + + in particular , both corollaries are not if and only if " type of statements , and ( 2.4 ) can not be turned into such a statement .
our ( ii ) corrects both problems .
\(ii ) @xmath193 ( i ) : since adding isolated vertices to an almost - planar graph results in an almost - planar graph , we only need to show that @xmath258 is almost - planar .
this is clear because for each edge @xmath1 of @xmath258 , @xmath266 is planar if @xmath1 is an edge obtained by subdividing an edge of @xmath261 , while @xmath267 is planar if @xmath1 belongs to @xmath268 .
\(i ) @xmath193 ( iii ) : as in the proof of ( i ) @xmath193 ( iii ) of theorem [ thm : main ] , we only need to find , for each @xmath269 , an edge @xmath1 of @xmath81 such that both @xmath197 and @xmath198 are nonplanar .
the case @xmath270 or @xmath104 was settled in the proof of theorem [ thm : main ] .
if @xmath271 , @xmath41 , @xmath272 , or @xmath273 , then the edge outside @xmath32 or @xmath33 satisfies the requirement .
\(iii ) @xmath192 ( ii ) : suppose the implication does not hold .
we choose a counterexample @xmath0 with @xmath206 as small as possible .
we first prove that @xmath0 is connected .
if @xmath0 is disconnected , since @xmath0 is nonplanar , one component of @xmath0 contains a @xmath274 or @xmath33 minor .
if another component of @xmath0 contains an edge , then @xmath0 contains @xmath272 or @xmath273 as a minor .
so @xmath0 must have an isolated vertex @xmath36 . by the minimality of @xmath0 , @xmath201 satisfies ( ii ) , which implies @xmath0 satisfies ( ii ) .
this contradiction proves that @xmath0 is connected . if @xmath0 has a 1-separation @xmath44 , since @xmath0 is nonplanar , at least one @xmath49 contains @xmath32 or @xmath33 as a minor .
hence @xmath0 contains @xmath275 or @xmath118 as a minor .
this contradiction proves that @xmath0 is 2-connected .
since 3-connected @xmath276-free nonplanar graphs are almost - planar , as shown in the proof of theorem [ thm : main ] ( implications ( iv ) @xmath192 ( ii ) @xmath192 ( i ) ) , @xmath0 can not be 3-connected and thus @xmath0 has a 2-separation @xmath44 .
let @xmath277 and , for @xmath72 , let @xmath278 or @xmath279 in case @xmath280 is not an edge of @xmath0 .
since @xmath0 is nonplanar , we may assume without loss of generality that @xmath281 is nonplanar .
note that @xmath110 is planar because otherwise @xmath0 would contain a @xmath39 or @xmath41 minor .
it follows that @xmath282 , since @xmath281 is nonplanar .
let @xmath79 be an induced path of @xmath111 between @xmath78 .
if @xmath111 has a vertex @xmath36 outside @xmath79 , then @xmath201 is nonplanar , because @xmath283 , a subdivision of @xmath281 , is nonplanar .
this implies that @xmath0 contains a @xmath39 or @xmath41 minor , which is impossible .
therefore , @xmath111 must equal @xmath79 and thus @xmath0 is obtained from @xmath281 by subdividing edge @xmath280 . by the minimality of @xmath0 ,
@xmath281 must satisfy ( ii ) .
let @xmath281 be obtained from a 3-connected almost - planar graph @xmath81 by subdividing edges in @xmath261 .
each edge of @xmath268 is not subdivided in the formation of @xmath281 and deleting such an edge in @xmath281 leaves a nonplanar graph . therefore , @xmath280 is not such an edge since @xmath284 is planar .
it follows that there is an edge @xmath285 such that @xmath280 belongs to a path ( of @xmath281 ) obtained by subdividing @xmath1 .
now it is clear that @xmath0 is obtained by repeatedly subdividing @xmath1 , and thus @xmath0 satisfies ( ii ) , which contradicts the assumption that @xmath0 is a counterexample .
this contradiction completes our proof .
john maharry and neil robertson .
the structure of graphs not topologically containing the wagner graph .
the webpage of john maharry , associate professor of mathematics , ohio state university , accessed 11/23/15 , https://u.osu.edu / maharry.1/. | a nonplanar graph @xmath0 is called _ almost - planar _ if for every edge @xmath1 of @xmath0 , at least one of @xmath2 and @xmath3 is planar . in 1990 , gubser characterized 3-connected almost - planar graphs in his dissertation .
however , his proof is so long that only a small portion of it was published .
the main purpose of this paper is to provide a short proof of this result .
we also discuss the structure of almost - planar graphs that are not 3-connected . |
quantum chromodynamics ( qcd ) is the modern theory of the strong nuclear force .
the physical degrees of freedom are quarks and gluons .
the latter are the quanta of gauge fields , in some ways analogous to photons .
the crucial difference is that gluons couple directly to each other , whereas photons do not .
a consequence of the self - coupling is that the force between quarks does not vanish at large distances ( as the coulomb force @xmath0 does ) but becomes a constant @xmath1 .
it would , thus , require a vast amount energy to separate quarks out to a macroscopic distance .
indeed , long before the separation becomes large enough to measure , the energy stored in the gluon field `` sparks '' into quark - antiquark pairs .
this phenomenon of qcd explains why freely propagating quarks are not observed in nature , and it is called confinement . qcd is a quantum field theory , which means that from the outset one must deal with mathematical objects that are unfamiliar even to many physicists .
the most widely used theoretical tool for quantum field theories is relativistic perturbation theory , as developed for quantum electrodynamics ( qed ) by feynman , schwinger , tomonaga and others sixty years ago @xcite .
perturbation theory provides the key connection between the mathematical theory with experiments in qed @xcite and the glashow - weinberg - salam theory merging qed with the weak nuclear force @xcite .
a textbook example of perturbative quantum field theory is to calculate how virtual pairs induce a distance dependence on the coupling in quantum gauge theories . in qed ,
one considers the fine structure constant , @xmath2 where @xmath3 is the charge of the electron . in qcd
, one has the strong coupling @xmath4 , related to the gauge coupling @xmath5 by analogy with equation ( [ eq : alpha ] ) .
when virtual pairs are taken into account , @xmath6 and @xmath4 are not constant but depend on distance ; they are said to `` run . ''
the running , depicted in figure [ fig : run ] , is completely different in the two theories .
( a ) ( b ) in qed @xmath6 becomes constant at distances @xmath7 ( the electron s compton wavelength ) , but at short distances it grows . in qcd @xmath4 grows at long distance ; perturbative running no longer makes sense once @xmath8 is in the confining regime . at short distances or , equivalently , high energies
the qcd coupling is small enough that perturbation theory can again be used .
for example , the cross section for an electron - positron pair to annihilate and produce a quark - antiquark pair can be reliably computed as a power series in @xmath4 .
the quark and antiquark each fragment into a jet of hadrons mostly pions , but some protons and neutrons , too and
general properties such as the energy flow and angular distributions of these jets can be traced back to the quark and antiquark . in this way
, the experiments can measure quark properties , and these are found to agree with theoretical calculations @xcite .
qcd and the electroweak theory form the foundation of the standard model of elementary particles .
they are so well established that many particle physicists consider the chromodynamic and electroweak gauge symmetries , and the associated quantum - number assignments of the quarks and leptons , to be laws of nature .
we can not conceive of a more comprehensive theory that does not encompass them .
but there is more to the standard model .
we know that something must spontaneously break the electroweak symmetry and that something ( perhaps the same thing ) must generate masses for the quarks and leptons .
the standard model contains interactions that _ model _ these phenomena .
the model interactions are consistent with observations yet also incomplete .
it is hoped that insights obtained with the large hadron collider ( lhc ) , to commence operations at cern later this year , will enable particle physicists to solve many of the puzzles raised by the standard model .
the purpose of this paper is to cover qcd , particularly in the strong - coupling regime where perturbation theory is insufficient .
it is worth bearing in mind , however , that there is a close connection between many of the results discussed below and some of the big questions in particle physics .
a simple example is the masses of the quarks . to determine the quark masses ( which because of confinement can not be measured )
, we have to calculate the relation between quark masses and measurable quantities , such as hadron masses .
quark masses are interesting for many reasons , the most unsettling of which is as follows .
the mass of the top quark is larger than that of the bottom quark ; see figure [ fig : mqineq ] .
similarly , the mass of the charmed quark is larger than that of the strange quark . if this simple pattern held with the up and down quarks the ones that make up proton and neutrons
then the up quark would be more massive than the down .
but then protons would decay to neutrons , positrons , and neutrinos ( @xmath9 ) .
the positrons would find electrons and annihilate to photons .
this universe would consist of neutron stars surrounded by a swarm photons and neutrinos , and nothing else .
our universe is not at all like this , because neutrons are more massive than protons .
the allowed decay reaction is @xmath10 , leading to an abundance of protons and electrons , and making possible chemistry and biochemistry .
with qcd we trace the neutron - proton mass difference back to the down - up quark masses and then to the deeper origin of quark masses .
the pattern of quark masses is necessarily not simple , and the search for simple explanation of a messy pattern occupies many particle theorists . below we give further examples where qcd is essential for working out the mysteries of particle physics .
qcd is also the cornerstone of modern nuclear physics .
the simplest nucleus is nothing but the proton , a bound state of two up quarks and one down quark .
the neutron is a bound state of two down quarks and one up quark .
the confining force is very , very strong . by comparison ,
the force traditionally called the strong nuclear force is a residue of the fundamental chromodynamic force .
this is similar to van der waals forces among molecules , which are electromagnetic forces between neutral objects with structure leading to a distribution of electric charge .
some basic problems are , therefore , to understand nucleon structure directly from qcd , to study the excitation spectrum of nucleons ( and their cousins with strangeness and other flavors ) , and to see how few - nucleon ( or , more generally , few - hadron ) systems interact .
another active area of research in nuclear qcd is the properties of quark matter at high temperature and density .
this regime is important to cosmology , because an epoch of the early universe consisted of hot quark matter , and to astrophysics , because , for example , neutron stars are composed of dense quark matter . in the laboratory ,
nuclear physicists create hot , dense quark matter by colliding heavy ions , at present with brookhaven s rhic , soon with ( the heavy - ion mode of ) cern s lhc , and someday with gsi s fair . at strong coupling it is necessary to go beyond perturbation theory . in a relativistic quantum field theory , such as qcd , this is not easy .
an important obstacle is that a quantum field associates one ( or perhaps a few ) degrees of freedom with every point in space - time .
hence , the number of degrees of freedom is uncountably infinite .
furthermore , the fluctuations at short distances , if added up naively , contribute an ultraviolet divergence to anything of interest .
these fluctuations are treated by a set of ideas known as _
_ @xcite .
in short , one must find a tool that ( correctly ) handles both strongly coupled fields and the renormalization of short - distance fluctuations .
this tool is lattice field theory or , for gauge theories like qcd , lattice gauge theory .
the idea , which goes back at least as far as heisenberg , is to replace continuous space ( or , in practice , space - time ) with a discrete grid , or lattice .
then the number of degrees of freedom is countable .
if the lattice is of finite extent , it occupies a finite physical volume , and , moreover , the number of degrees of freedom is finite . in 1974 , kenneth wilson showed how to maintain gauge symmetry with a discrete lattice @xcite . in this and subsequent work
, he and others elaborated how the lattice provides a mathematically rigorous definition of quantum field theory via functional integrals @xcite .
the functional integral for quantum fields on a finite lattice is a familiar object , a multi - dimensional integral , that can be evaluated on a computer . before downloading some publicly available code @xcite and trying to run it on a laptop ,
let us pause to contemplate how many variables of integration are reasonable ( for physics ) .
one wants to have the spacing between sites on the lattice small enough to resolve hadron structure , and one wants the volume large enough that the boundary conditions do not modify the structure .
if one supposes that the ratio of these two lengths is 32 , then the dimensions of the integrals of interest are @xmath11 because gluons come in 8 colors with 4 polarization states each .
the factor 128 presumes a space - time lattice with a time extent long ( in relativistic units ) compared to the spatial extent ; this is common practice . to cope with integrals of such large dimension ,
the only practical technique is monte carlo integration with importance sampling .
the weight guiding the importance sampling is @xmath12 , where @xmath13 is the classical action for the random sample of gluon variables . in functional integral formalisms for quantum mechanics ,
the expression for the action is the defining equation of the physical system .
the development of algorithms to generate these samples is a vibrant subject , covered in part by blint jo s poster here at the scidac conference .
this paper will not dwell on algorithms , but it is important to mention one more complication .
quarks are fermions and , as such , must satisfy the pauli exclusion principle . in the functional integral formalism , this is handled by introducing anticommuting grassmann variables for fermions .
the integration is no longer riemann ( or lebesgue ) integration but a formal procedure called berezin integration . to make a long story short , in lattice qcd we always can and do carry out the berezin integration by hand .
the outcome is that the weight for the importance sampling becomes @xmath14 , where @xmath15 is a sparse matrix with space - time indices .
the matrix @xmath15 is @xmath16 , where @xmath17 the quark field represents @xmath18 flavors with 3 colors and 2 spin states of quark and antiquark for every flavor and color .
the matrix @xmath15 is sparse because it is a lattice version of a differential operator , the dirac operator ( generalized to qcd ) .
incorporating @xmath19 into the weight is computationally demanding , increasing by two or three orders of magnitude the amount of floating - point operations needed to generate a new configuration of gluons .
the lattice is a helpful device mathematically and computationally , but it is not physical . to obtain a real result of the chromodynamics of continuum spacetime , one must work out the integrals for a sequence of lattices , and take the limit @xmath20 in a way that respects renormalization .
taking @xmath21 smaller and smaller increases the computing burden as onerously as @xmath22 .
the exponent @xmath23 depends on the algorithm and is typically around 1 or 2 . the 4 in
the exponent is the key reason why numerical lattice qcd requires the most advanced computing facilities available .
no algorithm can reduce it , because it is the dimension of space - time .
most lattice - qcd calculations are carried out in two steps .
the first is to generate an ensemble of gluon fields , with a certain lattice spacing and quark masses .
this requires the computationally super - demanding @xmath19 . to generate an ensemble of useful size ,
one needs several hundred , perhaps even a few thousand , samples of the gluon field .
computers of the highest available capability are needed for this step . in recent years
, lattice gauge theorists have carried out this step with the special - purpose computer qcdoc , large clusters of pcs , and , more recently , leadership - class machines constructed under the auspices of the doe s office of advanced scientific computing research ( ascr ) .
these ensembles of gluon fields are valuable : within the u.s .
they are generated by and for particle and nuclear physicists and shared under various agreements .
file - sharing is facilitated by formats and software developed by the international lattice data grid @xcite .
the second step is to mine these ensembles for physics .
a simple example is to compute the quantum mechanical amplitude for a proton to propagate from one point to another , and study the behavior of the amplitude as the separation varies : this is what one does to compute the proton mass , in fact .
the key ingredients are a few rows of the inverse of the matrix @xmath15 , introduced above . with the same ensemble
one can study protons , pions , and more uncommon hadrons such as the charmed strange pseudoscalar meson ( or @xmath24 for short ) .
different rows of @xmath25 ( i.e. , different flavors , colors , spins , and space - time separations ) are combined in various ways to obtain the quantum numbers of the hadrons in the problem at hand .
because each problem is different , and there are so many of them , we must attack this problem with the highest capacity computers .
the job mix here is heterogeneous , with many users whose job streams range from many large jobs to extremely many medium - sized jobs .
the challenge is to develop systems that scale well ( for the large jobs ) , while maintaining the flexibility needed to handle the heterogeneity .
the usqcd collaboration has designed clusters of pcs that deliver this capacity in a cost - effective way @xcite .
the remainder of this paper focuses on physics .
it is organized according to four principal themes of lattice gauge theory : * determination of fundamental parameters of the standard model of particle physics * study of nucleon structure , the hadron spectrum , and hadron interactions * simulation of the thermodynamic properties of qcd at nonzero temperature and density * lattice gauge theories beyond qcd , such as those appearing in models of electroweak symmetry breaking roughly speaking , the first and last are particle physics , and the second and third are nuclear physics .
but the lines are blurry : for example , some of the simplest calculations of proton structure ( the second item ) may play a role more accurate calculations of @xmath26 scattering cross sections at the lhc . the way that scientific goals drive the computing needs of lattice gauge theory
are discussed further , along the lines of these themes , in public whitepapers of the usqcd collaboration @xcite .
in quantum electrodynamics , the fundamental parameters are the masses of electrically charged particles and the fine structure constant @xmath6 .
similarly , in quantum _ _
chromo__dynamics , the fundamental parameters are the masses of colored particles ( the quarks ) and the strong coupling @xmath4 . just as @xmath6 can be determined in many ways @xcite , @xmath4 can be determined at high energies from jet cross sections with perturbative qcd , and at low energies from the hadron spectrum with lattice qcd .
two recent results are ( evaluated at distance @xmath27 ) @xmath28 where both error bars reflect experimental and theoretical uncertainties .
the first result is determined at high energies , around 100 gev , in the regime where quarks can be treated as nearly free .
the second is determined at low energies , of a few gev and lower , where quarks are confined .
both are linked soundly to the defining equations of qcd .
the agreement demonstrates the richness and broad validity of qcd and should make any enthusiastic scientist say `` wow ! ''
a straightforward application of lattice qcd is to work out how hadron masses depend on quark masses .
quark masses are interesting for several reasons . as fundamental constants of nature , they are interesting in their own right .
as discussed in the introduction , the pattern of quark masses is puzzling and , hence , a motivation to search for extensions of the standard model that would contain a simple explanation .
three flavors of quarks top , bottom , and charm
have masses large enough to influence some high - energy scattering and decay processes .
their masses can be determined with perturbative qcd as well as lattice qcd .
( the results agree @xcite . )
the other three quarks up , down , and strange have masses so small that the full bound - state problem must be considered .
lattice qcd , therefore , provides the best information .
some recent results are the following @xcite : @xmath29 one remarkable aspect of these results is that the strange quark s mass is somewhat smaller than had been thought on the basis of less reliable techniques for attacking strongly - coupled qcd .
this point is , perhaps , of interest mostly to particle physicists , but a further one is of broader interest .
the up and down masses the constituents of protons and neutrons and , thus , everyday matter are very small , within an order of magnitude of the electron mass ( @xmath30 ) . only about 12% of the proton mass ( @xmath31 ) or neutron mass ( @xmath32 )
is accounted for by the quark masses .
consequently , most of their masses and the mass of everyday matter arises from confinement , namely , the binding energy of the gluons and the ( relativistic ) kinetic energy of bound quarks .
this is a stunning insight , to be revisited in section [ sec : bsm ] .
the caption of figure [ fig : mqineq ] states that the pairings of quarks in each doublet stem from the electroweak interactions .
that is not the full story .
there is no reason for the eigenstates of mass to be the eigenstates of the interaction term with @xmath33 gauge bosons .
the two are related by a unitary transformation , @xmath34 , namely , @xmath35 here the initials ckm stand for cabibbo - kobayashi - maskawa ; cabibbo introduced equations ( [ eq : ckm ] ) for two generations , and kobayashi and maskawa were the first to study the ramifications for three generations . the left - hand side of the first equation denotes the electroweak basis .
the labels on the elements of @xmath34 are the weak isopartner and the mass eigenstate , because the transition amplitude for @xmath36 , for example , is proportional to @xmath37 .
the study of physical processes related to @xmath34 is called flavor physics , because the central issues are to understand why there are several flavors of quarks and what lends them their separate identities . as a unitary matrix
, @xmath34 contains complex entries .
some of the phases can be absorbed into unobservable phases of the quark wave functions .
for only two generations , one physical parameter remains , and it is real . for three generations , as in nature , three physical parameters are real and one phase remains .
the complex couplings imparted by the phase lead to physical processes that proceed at different rates for particles and the corresponding antiparticles .
such reactions are needed to explain the abundance of matter and the dearth of antimatter in the universe . among particle physicists ,
this phenomenon is called @xmath38 violation .
it is clearly intriguing to know whether the @xmath38 violation of the ckm matrix is enough to account for the matter - antimatter asymmetry of the universe . according to present measurements and theoretical understanding ,
it is insufficient , so one would like to know what other interactions violate @xmath38 . in the standard model
, there is a scalar boson called the higgs boson .
section [ sec : bsm ] will explain more about it .
for now let us note that both the quark masses and the ckm matrix have their origin in matrices of couplings between quark fields and the higgs field .
therefore , the puzzles of @xmath38 violation and the pattern of the quark masses are tied together by the standard model . to make progress in flavor physics , lattice - qcd calculations
are needed to interpret the experimental measurements . in a schematic form
the rate @xmath39 is given by @xmath40 where the `` known factor '' consists of well - measured physical constants and numerical factors like @xmath41 .
the paradigm is to measure as many flavor - changing processes as possible , and use the standard - model formulae to over - determine the ckm matrix .
if all the determinations are consistent , then one can set limits on non - standard sources of flavor and @xmath38 violation .
to do so , the qcd factor must be computed , and often the only way to do so is with lattice qcd .
there are dozens of relevant measurements , but because the ckm matrix is unitary , the elements of the ckm matrix must satisfy constraints .
a particular vivid one is the unitarity triangle that stems from the orthogonality of columns of a unitary matrix : @xmath42 which traces out a triangle in the complex plane .
the triangle and the present uncertainties from various constraints is shown in figure [ fig : ut ] , dividing each side by @xmath43 . and @xmath44 is omitted , because better measurements are needed ; they are expected from lhc@xmath45.,title="fig : " ] + and @xmath44 is omitted , because better measurements are needed ; they are expected from lhc@xmath45.,title="fig : " ] except for the wedge labelled @xmath46 the uncertainties can be reduced with lattice calculations of the qcd factor .
lattice calculations for a wide range of observables relevant to the ckm matrix have appeared over the past few years , and further improvements are under way .
these include calculations of properties of strange kaons , charmed @xmath47 and @xmath24 mesons @xcite , and beautiful @xmath48 and @xmath49 mesons . for recent reviews , including references to work in asia , australia , and europe ,
see @xcite .
the past several months witnessed an unexpected development in flavor physics , in the field of leptonic decays of pseudoscalar mesons .
( pseudoscalar bosons have spin 0 and negative parity ; the most common examples are the pion and kaon . )
one can consider the measured decay rates to determine a qcd matrix element called the decay constant , and denoted @xmath50 ( for the pion ) , @xmath51 ( for the kaon ) , and so on . table [ tab : f ] shows the results of some scidac - supported calculations of decay constants and the corresponding measurements .
.decay constants @xmath52 of pseudoscalar mesons ( here @xmath53 ) .
values in mev .
experimental determinations assume a suitable element of the ckm matrix element consistent with all flavor data . [
cols="^,^,^,^,^,^,^",options="header " , ] the calculations agree very well with the measurements , except in the case of the charmed strange pseudoscalar meson @xmath24 .
this set of circumstances is odd , because the computation is easiest for the @xmath24 .
algorithms for the propagators of light quarks slow down as the quark mass @xmath54 is decreased . as a consequence , properties of hadrons containing up or down quarks
are reached via a ( controlled ) extrapolation in @xmath54 .
this introduces a source of uncertainty that is absent for the @xmath24 , which , as the lightest bound state of a charmed quark and a strange antiquark , can be computed by working directly at the physical quark masses .
there is no obvious explanation for this discrepancy within the standard model .
unless there is a real blunder somewhere , the discrepancy points to a non - standard particle mediating the decays @xmath55 and @xmath56 , thereby changing the interpretation of the `` measured '' @xmath57 .
it turns out to be rather easy to devise models of such interactions @xcite .
given the agreement of cleo s new measurement of @xmath58 @xcite , new @xmath59 bosons and charged higgs bosons seem unlikely , making leptoquarks
particles with baryon and lepton number are the most plausible agents of decay .
let us turn now to nuclear physics , where the importance of lattice qcd is difficult to overstate .
many nuclear laboratories , such as jefferson laboratory in the u.s . , have turned their attention to the structure of the proton ( the simplest nucleus ! ) @xcite , the excitation spectrum of baryons @xcite , and processes that may produce glueballs . on the theoretical side ,
the previous decade witnessed dramatic growth in the application of effective field theories to understand nuclear interactions @xcite .
these techniques connect a wealth of nuclear and hadronic data to the qcd lagrangian .
lattice qcd is an irreplaceable tool to interpret the new and upcoming experiments at jlab and to compute the so - called low - energy constants of the effective theory @xcite . a topic at the interface between nuclear and particle physics concerns the distributions of partons ( a generic term for quarks and gluons ) inside the proton . in high - energy collisions of @xmath60 ( at the tevatron ) or @xmath26 ( at the lhc ) , the transverse motion of partons can be neglected .
to compute cross sections ( to produce , say , a higgs boson ) , the key property needed is the distribution of the longitudinal momentum . from a nuclear perspective , this is also one of the most basic aspects of proton structure .
usually one studies parton distributions @xmath61 as a function of the fraction @xmath62 of longitudinal momentum carried by a parton , @xmath63 .
fits to scattering data can determine @xmath64 only over a range , @xmath65 , limited by kinematics and statistics .
lattice qcd can compute moments of @xmath61 : @xmath66 . clearly
the two kinds of information are complementary . the experimental range can be extended with _ ad hoc _ functions possessing the right asymptotic forms at the endpoints , and in this way an ansatz - guided experimental determination of @xmath67 is possible .
one finds good agreement with lattice calculations @xcite . to explore nucleon structure further
, the next step is to probe the transverse structure and compute the so - called generalized parton distributions @xcite .
the axial charge of the nucleon is another topic of keen interest , because it appears in the description of @xmath68 decay .
it is also related to the relative spin polarization of up and down quarks in the proton .
some recent results with lattice qcd are @xmath69 showing good agreement with experiment ( @xmath70 ) .
work is under way to reduce the theoretical uncertainty . in all fields of physics ,
spectroscopy is a time - honored approach to learn about the dynamics of the underlying phenomena of interest . in nuclear and hadronic physics , therefore , an important area of investigation is the excitation spectrum of hadrons , particularly baryons .
excited states present additional challenges for lattice qcd , but the tools needed are available and have been demonstrated to work @xcite .
this subject extends to include topics such as the intriguing roper resonance @xcite and electromagnetic transitions of the form @xmath71 , where @xmath72 is an excited nucleon @xcite . because the gluons couple to each other , it is conceivable that there are hadrons consisting essentially of gluons , with no valence quarks .
( virtual quark - antiquark pairs appear in all hadrons . )
the most persuasive evidence that glueballs do indeed exist come from lattice - qcd calculations . in an approximation omitting virtual quark pairs , the evidence is clear that stable gluon - only bound states arise @xcite . there are two challenges with glueballs .
one , obviously , is to extend the lattice calculations to full qcd with virtual quark - antiquark pairs .
the other is to identify them unambiguously in the laboratory @xcite .
experiments with @xmath73 collisions are expected to produce such states and are planned at the 12 gev upgrade of cebaf , the accelerator at jefferson lab . to plan this program
it is essential to know the photocouplings of the mesons ( exotic or not ) .
it is possible to compute these photocouplings in lattice qcd , as has been demonstrated for charmonium ( @xmath74 mesons ) @xcite .
an extension of this work to light mesons , including exotics and glueballs , is under way .
a long - term goal of nuclear physics is to compute many - body interactions from qcd .
in addition to providing a better understanding of the nucleus , these calculations are relevant to mesonic atoms , namely , those containing a positively and negatively charged meson , and to strangeness in neutron stars .
although in some cases lattice qcd results can be compared to measurements , the more interesting situation is where this can not be done .
the aim here is to gain confidence in the reliability of these difficult calculations , so that one can use them ( with robust uncertainty estimates ) in applications where it is otherwise impossible to acquire the needed information any other way .
an interesting example is the lattice calculation of the @xmath75 scattering lengths , of which there are two , @xmath76 and @xmath77 , depending on whether the @xmath75 system has isospin @xmath78 or @xmath79 .
a recent lattice - qcd calculation finds @xcite @xmath80 in convenient dimensionless units .
these are in rough , but not spectacular , agreement with other determinations using chiral perturbation theory or the roy - steiner equations .
for example , an analysis based on the roy - steiner method @xcite finds @xmath81 employing an extrapolation from high energies down to the threshold ( where the scattering length is defined ) ; see @xcite for more discussion .
a more direct experimental determination is proposed by the dirac collaboration , which would form @xmath82 atoms @xcite . in this sense , eqs .
( [ eq : a1/2 ] ) and ( [ eq : a3/2 ] ) represent something exciting : a theoretical prediction awaiting definitive experimental confirmation .
multimeson systems are proving to be a fruitful area of research , with further work on @xmath83 @xcite , @xmath84 @xcite , and even states with as many as twelve pions @xcite .
once methods have been vetted in mesonic systems , the next step is to consider nucleon - nucleon @xcite and nucleon - hyperon @xcite systems .
this is more difficult , because the signal - to - noise ratio in the monte carlo estimate of baryonic correlation functions is worse . here
the methods of effective field theories can be used to extend feasible lattice - qcd calculations to a wider range of nuclear phenomena .
the hyperon - nucleon interaction is of interest , because there is speculation that the large fermi energy may make it energetically favorable for some nucleons to transmute to hyperons @xcite .
the discussion in the introduction of the computational scope of numerical lattice qcd mentions that , but does not explain why , the spacetime lattice usually has a time extent much longer than the spatial dimension , for example , @xmath85 .
the reason is that the functional integral actually describes a grand canonical ensemble ( in the sense of thermodynamics ) at temperature @xmath86 , @xmath87 where @xmath88 is the number of sites in the time direction , @xmath21 is the lattice spacing , and @xmath89 is boltzmann s constant . in the applications to particle and nuclear physics discussed in sections [ sec : sm ] and [ sec : kern ] ,
one wants @xmath90 and , hence , takes @xmath88 large enough for the boltzmann suppression @xmath91 to eliminate thermal effects , where @xmath92 is a hadron mass .
there are , however , many areas of research where a nonzero temperature and a nonzero chemical potential arise . for nonzero temperature ,
one simply makes @xmath88 smaller , typically @xmath93 in current work .
such lattices take somewhat less memory , but not less computing because thermodynamics obviously requires calculations at several temperatures .
it is conceptually easy to introduce a nonzero chemical potential into lattice gauge theory , but the conceptually clean approach is computationally beyond present resources . for small chemical potential ,
several approaches are available and efficient @xcite .
the physical phenomena of interest are summarized by the phase diagram in figure [ fig : phase - diagram ] .
the key feature is a phase transition from hadronic matter to a substance called the quark - gluon plasma .
the transition is expected to be first order and end at a critical point at chemical potential @xmath94 . between the critical point and the @xmath95 axis
the transition is a rapid crossover .
lattice - qcd calculations , as well as measurements of heavy - ion collisions at rhic and lhc , probe the crossover , which is relevant to the cooling of the early universe .
neutron stars appear at high baryonic density , namely , high @xmath96 . at even higher densities , several phases with color superconductivity
are thought to exist .
apart from heavy - ion collisions ( so far at small @xmath96 ) much of this phase diagram is unexplored , and lattice gauge theory offers the only _ ab initio _ , quantitative , theoretical tool to see if it really reflects qcd .
the central concept in qcd thermodynamics is the equation of state ( eos ) , namely , the relationship between the pressure @xmath97 and energy density @xmath98 as the temperature @xmath86 and chemical potential @xmath96 are varied .
the eos reveals the relevant degrees of freedom . in qcd , are they hadrons ? or quarks and gluons ? at low temperatures , hadrons are clearly present , and phenomenological models of hadron gases can describe qcd thermodynamics . at very high temperatures ,
perturbation theory in @xmath99 can be used .
but near the transition , nonperturbative methods are needed .
the first step is to find the transition temperature , which has now been done with 2 + 1 flavors of realistic sea quarks @xcite .
the different physics below , near , and above the critical temperature @xmath100 also has ramifications for controling systematic errors .
discretization effects from the lattice , for example , depend on the basic degrees of freedom .
( for a crossover it is , strictly speaking , incorrect to speak of a critical temperature .
but the critical point in the phase diagram is close enough to the @xmath95 axis that thermodynamic variables change rapidly at @xmath100 . )
the next step is to map out the thermal contribution to the trace of the energy - momentum tensor , @xmath101 .
from the temperature dependence of @xmath102 other thermodynamic variables , such as the entropy density , can be worked out .
this also has now been done with 2 + 1 flavors of realistic sea quarks @xcite , including at small @xmath96 @xcite .
as the system is heated , even to @xmath103 , the behavior of @xmath102 shows that the system does not turn rapidly into weakly interacting quarks and gluons . to make contact with heavy - ion collisions ,
the last step would be to take the output of lattice qcd as input to hydrodynamic models . with these models one would then compute experimental observables , such as elliptic flow and quarkonium suppression . with current computing resources , the eos has been pinned down with 2025% accuracy . to make a big impact on hydrodynamic modeling , however
, it will be necessary to reduce the error to 5% .
this is conceptually straightforward to achieve , but it will be computationally challenging .
part of the strategy is to reduce discretization effects , which scale as @xmath104 ( @xmath21 is the lattice spacing ) , while the net cost scales as @xmath105 .
as mentioned above , particle physics awaits a new era when the lhc starts producing physics results . commencing later this year
, the lhc will collide two beams of 7 tev ( tev = teraelectronvolt ) protons .
( amusingly , some particle physicists now call this energy scale the terascale , deliberately alluding to terascale computing
. it would be unkind to tell them that in computing we are moving on to the petascale . ) in particular , the lhc will address the central problem of particle physics , which is the origin of the @xmath33 and @xmath106 boson masses .
the fact that these masses do not vanish demonstrates that some mechanism breaks the electroweak gauge symmetry spontaneously .
`` spontaneous symmetry breaking '' means simply that , although the equations of motion are symmetric , the solution chosen is not symmetric .
for example , in the standard model a scalar field breaks electroweak symmetry , via a technique invented by higgs .
interactions between the higgs field and the fermions also generate ( in the standard model ) the quark and charged lepton masses and the ckm matrix
. therefore , flavor physics will remain interesting in the lhc era , and the physics program discussed in section [ sec : sm ] will need to be continued and extended . in this section , however , i would like to consider applications of lattice gauge theory to understanding the mechanism of electroweak symmetry breaking . since we suspect that the standard higgs sector is far from the full story , it is possible ( some would say likely ) that strongly coupled fields are involved .
for example , instead of being an elementary field , the higgs boson ( and its siblings , the longitudinal polarization states of the @xmath33 and @xmath106 bosons ) could be composite .
the constituents would then form a whole spectrum of states , just as quarks bind into many different hadrons . even if the higgs is elementary at the terascale ,
other agents of electroweak symmetry breaking could be strongly coupled . numerical work on theoretical bounds on the higgs mass are a long - standing topic in lattice field theory @xcite .
the issue is that the higgs mass is proportional to its self - coupling , which runs at short distances in a way similar to @xmath6 in qed ; see figure [ fig : run ] .
the perturbative running becomes invalid once the coupling is too large , and nonperturbative methods are required .
a subtle relation arises between the higgs mass and a scale where new phenomena imply that the standard model breaks down . to make a long story short , one can in this way bound the higgs mass from above .
a similar analysis applied to the yukawa coupling between the higgs field and the top quark leads to a lower bound on the higgs mass , as discussed in a recent comprehensive analysis @xcite .
a popular class of non - standard models is to replace the higgs sector with a strongly interacting sector that is , in some respects at least , analogous to qcd .
in fact , absent a stronger source of symmetry breaking , the pions of qcd would break the electroweak symmetry , leading to @xmath33 and @xmath106 masses around 100 mev .
what one needs , then , is a new quantum number , called technicolor , and a gauge force to bind the corresponding particles , called techniquarks , together .
these dynamics are posited to take place at terascale energies and , by analogy with qcd , generate vector boson masses around 100 gev , as observed .
the earliest models of technicolor followed the example of qcd too closely and have been ruled out by experimental measurements .
one way around the experimental constraints is to assume that the coupling evolves more slowly than in qcd .
this is called `` walking '' ( i.e. , slower than running ) technicolor . for a wide range of scales below the confining scale ,
the coupling is too strong for perturbation theory to be relied on , so nonperturbative work with lattice gauge theory is called for . in the past year , the walking hypothesis has indeed been tested , by examining the scaling behavior of the technipion mass , the vector meson mass , and the technipion decay constant @xcite .
even more recently , the evolution of @xmath4 in ( a tev - scale version of ) qcd with several flavors has been computed @xcite , to see how many flavors are needed to having a walking coupling , and how a walking coupling changes the physics . the most popular way to relieve the theoretical problems of the standard model is to introduce supersymmetry , a symmetry that transforms fermions into bosons , and vice versa . in these models , every known particle ( quarks , leptons , gauge bosons )
comes with another state known as a superpartner . in the minimal supersymmetric model ,
the higgs sector contains three neutral bosons , a charged pair , and their superpartners .
no superpartner has been observed , so at the terascale and below supersymmetry must be broken somehow .
usually supersymmetric models are conceived of as weakly coupled , but , since they are gauge theories , strong coupling is not to be dismissed out of hand .
it is therefore of great interest to formulate and simulate supersymmetric gauge theories on a spacetime lattice .
this has been an extremely fruitful theoretical field during the past several years @xcite , and now first numerical simulations are being carried out @xcite .
this rapid development yields hope that nonperturbative studies of supersymmetric models will be possible , including studies of supersymmetry breaking .
every now and then one hears the question , `` how much computing would it really take to solve qcd on a computer ? ''
the questioner is ( usually ) well meaning , but most experts would agree that the question is not well conceived .
the calculations for some of the simplest hadronic quantities are finally becoming mature , with analyses that forthrightly address all sources of uncertainty .
this class of observable , sometimes called `` gold - plated '' @xcite , consists of masses of particles that are stable ( with respect to the strong interactions ) , not too close to thresholds , and of hadronic transition matrix elements with one or zero of these stable hadrons in the final and initial states .
many , but not all , aspects flavor physics ( section [ sec : sm ] ) , and some aspects of hadronic structure ( section [ sec : kern ] ) , lie in this class , and the first solid round of serious computation is nearing its end .
but this is the end of the beginning .
in these nearly mature fields , it turns out that precision is important , at least for the foreseeable future .
for example , to gain persuasive evidence of new phenomena in flavor physics , 525% accuracy is not enough . a better target ( given the stakes and the quality of the corresponding experiments ) is 13% .
if the history of kaon physics is a guide , it will eventually be important to go further .
happily , the need for and clear path to precision has persuaded the particle - physics community that the resources needed for lattice flavor physics are worth the cost .
even a quick glance at section [ sec : kern ] reveals that the large majority of interesting problems in nuclear physics require much more computing resources than gold - plated observables . by definition
the excited hadrons are not stable under the strong interaction
. they will be harder to pin down , and it is possible that we will not know how hard until the work currently under way reaches some of its milestones .
the calculations of scattering lengths , while impressive , must be repeated with several values of the physical volume to ensure that the mathematical formalism used to extract them works as it should .
given the importance of nuclear lattice qcd in guiding nuclear experimentation , one would hope for full support from experimenters for the needed computing resources . the `` how much ? ''
question should be rephrased .
it is really a class of questions , `` how much computing would it take to compute my favorite observable ? '' of course , the answer depends on your observable and its wider context .
it should be clear , however , that physics with either precision and complexity leads to questions whose answers will need one or two orders of magnitude more computing than what is available now .
another reason the `` how much ? ''
question is ill - conceived is that it often overlooks the heterogeneous job mix needed to attack many problems in lattice qcd .
it is not enough to generate huge ensembles of lattice gauge fields , with ever smaller lattice spacing and ever smaller quark masses , on the world s highest - capability computer .
this step is necessary , creating a gold mine of information .
it is not enough , however , merely to open a mine ; the gold must be extracted and , here , we need a large capacity of computing that is flexible in every way ( while having significant capability of its own ) . we do not know how broad the future of lattice gauge theory will be . if the lhc discovers a strongly interacting sector beyond the standard model , it is likely that chiral fermions play a role .
this is simply because , at the terascale , we already know that left - handed fermions and right - handed fermions are different fields . but chiral fermions possess some conceptual and computational challenges not found in qcd .
a strongly coupled terascale will bring many more particle theorists to lattice gauge theory , and this larger community will clamor for computing that is up to the task . this work has been supported in part by the united states national science foundation and the office of science of the u.s .
department of energy ( doe ) .
software development and hardware prototyping within the usqcd collaboration are supported by scidac .
fermilab is operated by fermi research alliance , llc , under contract de - ac02 - 07ch11359 with the us doe . | we survey results in lattice quantum chromodynamics from groups in the usqcd collaboration .
the main focus is on physics , but many aspects of the discussion are aimed at an audience of computational physicists . |
in recent years there have been several papers studying the effect that flattening a domain has on the eigenvalues of the laplace operator @xcite ; see also the books @xcite and the references therein for similar problems with boundary conditions other than dirichlet . in these papers the main objective has been the derivation of the asymptotics of these eigenvalues in terms of a scalar parameter measuring how thin the domain becomes in one direction , as this parameter approaches zero .
as far as we are aware , almost if not all such existing examples in the literature are concerned with domains in euclidean space where the limiting problem degenerates to a domain of zero measure and therefore eigenvalues approach infinity .
a slightly different set of problems which has been considered consists of domains which are perturbations of singular sets such as thin tubular neighbourhoods of graphs , i.e. , domains which locally are like thin tubes
see @xcite , for instance , and also @xcite for a review . as in the papers cited above , again the limiting domains have zero measure and the spectrum behaves in quite a different way from the model considered here .
in this paper we study a situation which , although different from that described in the first paragraph , has in common with it the process by which the limiting domain is approached .
more precisely , consider the case of a given domain @xmath3 in @xmath0 satisfying certain restrictions which for the purpose here may be stated roughly as being bounded from above and below by the graphs of two functions see section [ formulation ] for a precise formulation .
the domain @xmath3 is then flattened towards a domain @xmath4 in @xmath5 via a ( continuous ) one - parameter family of domains @xmath6 .
these domains are obtained as the functions mentioned above are multiplied by the parameter @xmath7 .
the problem that shall concern us here is the study of the evolution of the eigenvalues of the laplace - beltrami operator on the one - parameter family of compact hypersurfaces @xmath8 which are the boundaries of the domains @xmath6 described above , as @xmath7 approaches zero .
one of the differences in this instance is that while the domain @xmath9 has zero @xmath10measure as stated above , @xmath11 retains positive @xmath12measure , developing instead a singularity on the boundary of the domain @xmath4 ( when considered as a domain in @xmath5 ) .
we thus expect these eigenvalues to remain finite as the parameter @xmath7 approaches zero , and to converge to a limiting spectral problem on the double sided flat hypersurface .
this is indeed the case , and the relevant spectral problems turn out to be the dirichlet and neumann problems on the domain @xmath4 , with the two next asymptotic terms after that being of orders @xmath13 and @xmath14 .
these results have been announced in @xcite . with a cross - section at the edge ] in order to understand the origin of the @xmath1 term in the expansion , it turns out that it is sufficient to consider the case where @xmath15 equals one , that is when the boundary is basically @xmath16 . because of this , it is not necessary to take into consideration the geometric intrincacies of the problem which appear in higher dimensions and it is possible to obtain the full description of eigenvalues in terms of elliptic integrals .
more precisely , for an ellipse of radii @xmath17 and @xmath7 we have that the eigenvalues are given by @xmath18 for @xmath19 and where @xmath20 is the complete elliptic integral of the second type yielding one quarter of the perimeter of the ellipse for @xmath21 . combining the above with the asymptotic expansion for @xmath22 yields @xmath23 in some sense ,
the purpose of the analysis that we shall carry out in what follows is to show that the above result may actually be extended to higher dimensions .
it should be noted here that this expansion depends on the relation between the different variables at the endpoints of the segment , which in this case is of the form @xmath24 .
clearly different relations between the leading powers will lead to different expansions . more generally , the issue is that the points of the boundary of @xmath3 where there is a tangent in the direction along which the domain is being flattened will play a special role . throughout the paper
we assume this set of points to be contained in a hyperplane orthogonal to the scaling direction , and that this tangency is simple . in the vicinity of these points
we take the cross - section of our surface as indicated in fig .
[ crosssect ] which , with the assumptions made , will be similar to the one - dimensional ellipse described above .
our results then state that in the higher - dimensional case the asymptotics for the eigenvalues still behave in a similar fashion and thus the logarithmic terms appearing above persist in this more general setting . apart from the intrinsic interest of the behaviour of the spectrum close to double sided flat domains , we point out that such manifolds have appeared in the literature in connection with eigenvalues as maximizers of the invariant eigenvalues among all surfaces isometric to surfaces of revolution in @xmath25 @xcite and for hypersurfaces of revolution diffeomorphic to a sphere and isometrically embedded in @xmath0 @xcite .
in fact , it is shown in those papers that these optimal singular _ double flat disks _ maximize the whole invariant spectrum and not just a specific eigenvalue .
another source of interest for such asymptotic expansions lies with the fact that , in some cases , they turn out to be fairly good approximations for low eigenvalues also for values of the parameter @xmath7 away from zero see @xcite .
we remark in passing that another problem for which it is conjectured that the optimal shape is given by a double sided flat disk is alexandrov s conjecture relating the area and diameter of surfaces of non negative curvature .
the structure of the paper is as follows . in the next section
we give a precise formulation of the problem under consideration and state our main results , namely , the nature of the limiting problem and the relation of the limit and approximating operators .
this includes the form of the asymptotic expansion and the expressions for the first three coefficients and an application to the case of the surface of an ellipsoid .
section [ prelim ] is then devoted to several preliminaries and auxiliary material used in sections [ unifresconv ] and [ sec4 ] , where the proofs of the main results are presented .
let @xmath26 , @xmath27 be cartesian coordinates in @xmath28 and @xmath0 respectively , @xmath29 , @xmath30 be a bounded domain in @xmath5 with infinitely smooth boundary .
let also @xmath31 denote two arbitrary functions and define the manifold @xmath32 where @xmath7 is a small positive parameter .
we assume @xmath33 to be infinitely differentiable and to have no self - intersections . to ensure this
, we make the following assumptions on @xmath34 , the first of which ensures the absence of self - intersections , 1 .
[ a1 ] the relations @xmath35 hold true . to state the second assumption we need to introduce some additional notation .
let @xmath36 , @xmath37 , be the inward normal to @xmath38 , and denote by @xmath39 the distance to a point measured in the direction of @xmath40 .
consider equations @xmath41 our second assumption concerns the solvability of these equations with respect to @xmath39 and implies the smoothness of @xmath33 in a neighbourhood of @xmath38 : 1 .
[ a2 ] there exists @xmath42 such that for all @xmath43 $ ] , @xmath37 , equations ( [ 1.2 ] ) have a unique solution given by @xmath44\times\p\om),\ ] ] such that @xmath45 we observe that assumptions @xmath46 and @xmath47 imply that @xmath48 the main object of our study is the laplace - beltrami operator @xmath49 on @xmath33 .
we introduce it rigorously as the self - adjoint operator associated with a symmetric lower - semibounded sesquilinear form @xmath50:=(\nabla u,\nabla v)_{l_2(\operatorname{\mathcal s}_\e)}\quad\text{on}\quad \h^1(\operatorname{\mathcal s}_\e).\ ] ] we recall that on an arbitrary manifold with metric tensor @xmath51 this may be written in local coordinates @xmath52 as @xmath53 where @xmath54 are the entries of the inverse to the metric tensor .
if in our case we take @xmath55 as local coordinates on @xmath33 , then on each side @xmath56 the operator @xmath49 may be written in the form @xmath57 where @xmath58 is the @xmath59 identity matrix and @xmath60 is the matrix with entries @xmath61 .
on the boundary @xmath38 the coefficients of such operator have singularities , and this is why in a neighbourhood of @xmath38 it is more convenient to employ the coordinates @xmath62 , where @xmath63 are some local coordinates on @xmath38 .
we do not give here the expression of the operator @xmath49 in such coordinates , as it requires the introduction of additional ( cumbersome ) notation .
the purpose of the present paper is to describe the asymptotic behavior of the resolvent and the spectrum of @xmath64 as @xmath65 . in this limit
, the hypersurface @xmath33 collapses to a flat two - sided domain @xmath66 , where @xmath67 are two copies of @xmath30 understood as the _ upper _ and _ lower _ sides of @xmath68 . because of this , it is natural to expect that the limiting operator for @xmath64 as @xmath65 is the laplacian on @xmath68 , i.e. , that on @xmath67 subject to certain boundary conditions .
indeed , this is true , and it is our first main result .
namely , we introduce the space @xmath69 as consisting of the vectors @xmath70 , where the functions @xmath71 are defined on @xmath67 and @xmath72 .
we can natural identify @xmath69 with @xmath73 . in the same way we introduce the sobolev spaces
@xmath74 assuming that for each @xmath75 the functions @xmath76 satisfy the boundary conditions @xmath77 the meaning of these boundary conditions is that the functions @xmath71 should be `` glued smoothly '' while moving from @xmath78 to @xmath79 via @xmath80 .
we observe that @xmath81 is embedded into @xmath82 , but does not coincide .
it is also clear that for any @xmath83 the function @xmath84 belongs to @xmath85 .
similarly , if @xmath86 , @xmath87 , respectively , @xmath86 , @xmath88 , then @xmath89 , respectively , @xmath90 .
let @xmath91 be the self - adjoint operator in @xmath69 associated with the closed symmetric lower - semibounded sesquilinear form @xmath92:=(\nabla \bs{u},\nabla \bs{v})_{l_2(\bs\om ) } \quad\text{on}\quad \h^1(\bs{\om}).\ ] ] by @xmath93 we denote the domain of an operator , the symbol @xmath94 indicates the norm of an operator acting from the hilbert space @xmath95 to a hilbert space @xmath96
. given any vector @xmath70 defined on @xmath68 , by @xmath97 we denote the function on @xmath33 being @xmath98 on @xmath99 and @xmath100 on @xmath101 . and vice versa , given any function @xmath102 defined on @xmath33 , by @xmath103 we denote the vector @xmath70 , where @xmath104 , @xmath105 , @xmath106 .
[ th1.1 ] for each @xmath107 there exists @xmath108 such that the estimate @xmath109 holds true .
[ rm1.1 ] the statement of this theorem includes the fact that the operator @xmath110 is well - defined as a bounded one from @xmath111 into @xmath112 . in view of the embedding of @xmath85 into @xmath113 , and the compact embedding of the latter into @xmath114 , the operator @xmath49 has a compact resolvent .
hence , it has a pure discrete spectrum accumulating only at infinity .
the same is true for the dirichlet and neumann laplacians @xmath115 and @xmath116 on @xmath30 .
recall that @xmath115 is the friedrichs extension in @xmath117 of @xmath118 from @xmath119 , and @xmath116 is the self - adjoint operator in @xmath117 associated with the sesquilinear form @xmath120 on @xmath121 . in
what follows @xmath122 denotes the discrete spectrum of an operator .
our next result follows from theorem [ th1.1 ] and ( * ? ? ?
viii.23 , viii.24 ) .
[ th2.2 ] the eigenvalues of @xmath49 converge to those of @xmath91 as @xmath7 goes to zero . in particular , if @xmath123 , then @xmath124 for @xmath7 small enough . for each @xmath125-multiple eigenvalue @xmath126 there
exist exactly @xmath125 eigenvalues ( counting multiplicities ) of @xmath49 converging to @xmath127 as @xmath65 .
let @xmath128 be the projector on the eigenspace associated with @xmath127 , @xmath129 be the total projector associated with the eigenvalues of @xmath49 converging to @xmath127 .
then the convergence @xmath130 holds true .
let now @xmath127 be an eigenvalue of @xmath91 with multiplicity @xmath125 and @xmath131 be associated eigenfunctions orthonormalized in @xmath69 .
it will be shown in the next section in lemma [ lm1.2 ] that the asymptotics @xmath132 hold true , where @xmath133 by @xmath134 we denote the laplace - beltrami operator on @xmath38 , where the metric @xmath135 on @xmath38 is induced by the euclidean one in @xmath136 .
for any smooth functions @xmath137 on @xmath38 , we shall denote the pointwise scalar product of its gradients by @xmath138 .
let @xmath139 employing the coefficients of the asymptotics ( [ 2.15 ] ) , we introduce two real symmetric matrices @xmath140 , @xmath141 with entries @xmath142 and @xmath143 \\ & -\int\limits_{\p\om}\frac{1 + 4\ln 2+\ln a_2}{4a_2 } \big(\psi_i^{(1)}\psi_j^{(1)}+\l\psi_i^{(0 ) } \psi_j^{(0)}-\nabla\psi_i^{(0)}\cdot \nabla\psi_j^{(0 ) } \big)\di s , \end{aligned}\label{2.11b}\ ] ] where @xmath144 it will be shown in sec .
4 that the matrix @xmath141 is well - defined . by the theorem on simultaneous diagonalization of two quadratic forms , in what follows the eigenfunctions @xmath145
are supposed to be orthonormalized in @xmath146 and the matrix @xmath147 to be diagonal .
the eigenfunctions @xmath145 chosen in this way depend on @xmath7 , but it is clear that the norms @xmath148 are bounded uniformly in @xmath7 for all @xmath149 , @xmath150 .
[ th2.3 ] let @xmath127 be an @xmath125-multiple eigenvalue of @xmath91 and @xmath145 , @xmath150 , be the associated eigenfunctions of @xmath91 chosen as described above .
then there exist exactly @xmath125 eigenvalues @xmath151 , @xmath152 ( counting multiplicity ) of @xmath49 converging to @xmath127 .
these eigenvalues satisfy the asymptotic expansions @xmath153 where @xmath154 are the eigenvalues of the matrix @xmath147 , and @xmath155 is any constant in @xmath156 .
the eigenvalues @xmath157 are holomorphic in @xmath158 and converge to the eigenvalues of @xmath140 as @xmath159 .
in addition to the asymptotic expansions for the eigenvalues @xmath160 given in this theorem , we also obtain the asymptotics for the total projector associated with these eigenvalues . however , to formulate this result we have to introduce additional notation and it is thus more convenient to postpone its statement which will them be made at the end of sec . [ sec4 ] see theorem [ th4.7 ] .
let us describe briefly the main ideas employed in the proof of the main results .
the proof of the uniform resolvent convergence in theorem [ th1.1 ] is based on the analysis of the quadratic forms associated with the perturbed and the limiting operators and on the accurate estimates of the functions in certain weighted sobolev spaces .
the proof of the first theorem uses essentially the method of matching asymptotic expansions @xcite for formal construction of the asymptotics for the eigenfunctions associated with @xmath151 .
these asymptotics are constructed as a combination of outer and inner expansions .
the former depends on @xmath55 and its coefficients have singularities at @xmath38 . in the vicinity of @xmath38
we introduce a special rescaled variable @xmath161 as @xmath162 and @xmath163 as @xmath164 .
this variable then describes the slope of @xmath33 in the vicinity of @xmath7 see also equations ( [ 4.54 ] ) giving the parametrization of @xmath33 in the vicinity of @xmath38 .
after rewriting the eigenvalue equation in the variables @xmath165 , where @xmath63 are local coordinates on @xmath38 , its leading term is in fact the laplace - beltrami operator on the ellipse giving rise to the logarithmic terms in the asymptotics for both the eigenvalues and the eigenfunctions . despite the fact that we are only presenting the leading terms of the asymptotics for @xmath151 and for the associated total projector in theorems [ th2.3 ] and [ th4.7 ] , respectively , our approach also allows us to construct the complete asymptotic expansions if required .
although this would need to be checked in a way similar to what was done here for the first few terms , the ansatzes ( [ 4.1 ] ) and ( [ 4.116 ] ) suggest that the complete asymptotic expansion for the eigenvalues should be @xmath166 where @xmath167 are functions holomorphic in @xmath158 .
these higher - order terms would then still reflect the behaviour observed in the ellipse example given in the introduction .
although the above formulas for @xmath168 and ( specially ) @xmath169 may look quite cumbersome at a first glance , they will actually simplify when computed for particular cases as some of the terms involved will vanish depending on whether we are considering dirichlet or neumann boundary conditions on @xmath170 .
we note that a similar effect was already present when computing the coefficients in the expansions obtained in @xcite .
this is particularly clear in the second of these papers dealing with dimensions higher than two , where the general expression is quite complicated and needs to be computed specifically in each case .
when this is done for general ellipsoids in any dimension , for instance , it yields a much simpler one - line expression .
we shall illustrate this by considering a thin ellipsoidal surface . to this end
take @xmath30 to be the unit disk centred at the origin with @xmath171 in this instance the limiting eigenvalues may be found via separation of variables and they will be of the form @xmath172 , where @xmath173 are the zeroes of the bessel function @xmath174 and its derivative @xmath175 , corresponding to eigenfunctions satisfying dirichlet and neumann boundary conditions on @xmath38 , respectively .
the following examples illustrating both cases are taken from @xcite , where the details may be found .
we consider the case of dirichlet boundary conditions first , i.e. , @xmath176 substituting these formulas and ( [ ex2 ] ) into ( [ 2.11a ] ) and ( [ 2.11b ] ) , we then obtain @xmath177 and @xmath178 the asymptotics ( [ 2.16 ] ) thus become @xmath179 and , for a particular eigenvalue , the remaining integral may be computed numerically .
we illustrate this by considering the case corresponding to the first dirichlet eigenvalue on the disk which yields @xmath180 as an example of limiting multiple eigenvalue we consider the first nontrivial neumann eigenvalue of the disk . in two dimensions
this is a double eigenvalue with associated ( normalized ) eigenfunctions given by @xmath181 where @xmath182 is the polar angle corresponding to @xmath183 and @xmath184 is the first nontrivial zero of @xmath185 . the eigenfunctions in @xmath69
are then given by @xmath186 , @xmath187 , from which we have @xmath188 and @xmath189 @xmath190 proceeding as before , we @xmath191 for the next term we now obtain @xmath192\di r -\l\ln 2.\end{aligned}\ ] ] for @xmath187 and @xmath193 for @xmath194 .
> from this , and again computing the relevant integrals numerically , we obtain @xmath195 due to the radial symmetry of @xmath30 , it is clear that these two eigenvalues should coincide , and the associate eigenfunctions converge to @xmath196 and @xmath197 .
in this section we discuss two parameterizations of the surface @xmath33 and prove three auxiliary lemmas which will be used in the next sections for proving theorems [ th1.1 ] , [ th2.3 ] .
the first parametrization is that used in the definition of @xmath33 in ( [ 1.0 ] ) , i.e. , each point on @xmath33 is described as @xmath198 , @xmath199 , where the sign corresponds to the upper or lower part of @xmath33 .
let us first calculate the metrics on @xmath33 in terms of the variables @xmath55 .
the tangential vectors to @xmath33 at the point @xmath105 , @xmath106 are @xmath200 where `` @xmath17 '' stands on @xmath201-th position .
thus , the metric tensor has the form @xmath202 it easy to see that @xmath203 where @xmath204 is treated as a column vector , and `` @xmath205 '' denotes transposition .
[ lm3.2 ] the matrix @xmath206 has two eigenvalues , the @xmath207-multiple eigenvalue @xmath17 , and the simple eigenvalue @xmath208 .
the identity @xmath209 holds true . from ( [ 3.3 ] ) we may write the eigenvalue problem for the matrix @xmath210 as @xmath211 and @xmath212 where @xmath213 . we thus see that any vector orthogonal to @xmath214 is an eigenvector for the above equation with eigenvalue @xmath215 equal to one .
this yields an eigenvalue of multiplicity @xmath216 if @xmath214 is not zero , and @xmath15 in case @xmath214 vanishes . in the former case
, we easily see that @xmath214 is also an eigenvector , now with eigenvalue @xmath217 , which will have multiplicity one .
the determinant of @xmath210 is thus @xmath218 , yielding the volume element to be @xmath219 as desired . in what follows
we shall make use of the differential expression for the operator @xmath49 , namely , its expansion w.r.t .
the expression itself is given by ( [ 2.3a ] ) , while using ( [ 3.3 ] ) allows us to expand some of the terms in this expression in powers of @xmath7 , @xmath220 where the plus and minus signs correspond to the upper and lower parts of @xmath221 , respectively .
we substitute these formulas into ( [ 2.3a ] ) and get @xmath222 the disadvantage of the parametrization by the variables @xmath55 is that the functions @xmath34 are not smooth in a vicinity of @xmath38 and their derivatives blow - up at the boundary @xmath38 .
we shall show it below while introducing the second parametrization .
the main idea of the second parametrization is to use special coordinates in a vicinity of @xmath38 so that they involve smooth functions only ; this parametrization is purely local and will be used only in a vicinity of @xmath38 .
it is natural to expect the existence of such coordinates since the surface @xmath33 is infinitely differentiable . in a neighborhood of @xmath38
we introduce new coordinates @xmath62 , where @xmath223 are local coordinates on @xmath38 corresponding to a @xmath224-atlas , and @xmath39 , we remind , is the distance to a point measured in the direction of the inward normal @xmath225 to @xmath38 .
let @xmath226 be the vector - function describing @xmath38 .
we have @xmath227 where @xmath228 and the other vectors in the definition of @xmath229 are treated as columns .
the vectors @xmath230 are tangential to @xmath229 and linear independent , while @xmath228 is orthogonal to @xmath38 .
thus , the matrix @xmath229 is invertible for all sufficiently small @xmath39 and all @xmath231 .
the inequalities @xmath232,\ ] ] are valid , where @xmath233 , @xmath234 are positive constants independent of @xmath62 .
it follows from these estimates and ( [ 3.8 ] ) that the matrix @xmath235 is infinitely differentiable in the neighbourhood @xmath236 of @xmath38 .
consider now equations ( [ 1.2 ] ) . by assumption ( [ a2 ] ) they have the smooth solution @xmath237 and , for small @xmath238 , the function @xmath239 behaves as @xmath240 hence , @xmath241,\label{3.12}\end{aligned}\ ] ] where @xmath242 , @xmath243 are positive constants independent of @xmath62 .
as we see from the last estimates , the functions @xmath34 are not smooth at the point @xmath244 , i.e. , at @xmath38 .
we employ once again assumption ( [ a2 ] ) and pass from equations @xmath245 to @xmath246 it follows from ( [ 1.4 ] ) that the function @xmath247 can be represented as @xmath248 , where @xmath249\times\p\om)$ ] and @xmath250 for sufficiently small @xmath251 .
we introduce a new variable @xmath252 . > from assumption ( [ a2 ] ) we conclude that @xmath253\times\p\om)\ ] ] for a fixed small constant @xmath254 , and the taylor series for @xmath239 and @xmath255 read as follows , @xmath256 where @xmath257 .
we define a rescaled variable @xmath258 .
the final form of the second parametrization for @xmath33 is as follows , @xmath259,\ ] ] where @xmath260 and @xmath254 is a fixed sufficiently small number .
we observe that by the definition of @xmath261 @xmath262 as in ( [ 5.1 ] ) , we shall also employ the expansion in @xmath7 of the differential expression for @xmath49 corresponding to the second parametrization .
we find first the tangential vectors to @xmath221 corresponding to the parametrization ( [ 4.54 ] ) , @xmath263 it is clear that the vectors @xmath264 , @xmath265 belong to the tangential plane and are orthogonal to @xmath40 . employing this fact and ( [ 4.15 ] ) , we calculate the metric tensor , @xmath266 by weingarten equations we see that @xmath267 where @xmath268 @xmath135 is the metric tensor of @xmath38 associated with the coordinates @xmath63 , @xmath269 is the second fundamental form of @xmath38 corresponding to the orientation defined by @xmath40 .
hence , the metric tensor @xmath270 of @xmath221 associated with the parametrization ( [ 4.54 ] ) reads as follows , @xmath271 by direct calculations we check that @xmath272 the quantities in ( [ 4.17 ] ) are well - defined provided @xmath254 is sufficiently small . indeed , by ( [ 4.9 ] )
@xmath273 that implies the existence of @xmath274 and @xmath275 . in what follows
we assume that @xmath254 is chosen in such a way . by @xmath276 , @xmath277 , we denote the principal curvatures of @xmath38 , and @xmath278 .
we note that @xmath279 is the mean curvature of @xmath38 and let @xmath280 [ lm4.1 ] the identities @xmath281 hold true . the identities ( [ 4.36 ] )
follow directly from the definition of @xmath282 , @xmath283 , and @xmath284 .
we make linear transformations in ( [ 4.17 ] ) to calculate the determinant of @xmath270 , @xmath285 that proves ( [ 4.28a ] ) .
it is easy to see that @xmath286 in view of ( [ 4.18 ] ) we get @xmath287 we substitute the obtained formula and ( [ 4.52 ] ) into ( [ 4.27 ] ) and arrive at ( [ 4.28b ] ) . employing ( [ 4.18 ] ) , ( [ 4.36 ] ) , by direct calculations we check @xmath288 hence , by ( [ 4.28a ] ) , ( [ 4.28b ] ) and the definition of @xmath275 @xmath289 where @xmath290 are some functions . in particular , @xmath291 while the function @xmath292 , @xmath293 satisfy the uniform in @xmath294 and @xmath295 estimates @xmath296 the obtained formulas , lemma [ lm4.1 ] , and ( [ 4.17 ] ) allow us to write the expansion for @xmath297 , @xmath298 taking into account ( [ 4.28a ] ) , ( [ 4.28b ] ) , we write the operator @xmath64 in terms of the variables @xmath299 , where @xmath300 , @xmath301 and @xmath302 are the entries of the inverse matrix ( [ 4.17 ] ) .
it follows from the last formula and ( [ 4.17 ] ) that @xmath303 we employ the obtained equation , ( [ 4.37a ] ) , ( [ 4.37 ] ) and ( [ 4.38 ] ) , and expand the coefficients of @xmath64 in powers of @xmath7 leading us to the identities @xmath304 we proceed to the auxiliary lemmas which will be used for proving theorem [ th2.3 ] .
[ lm4.2 ] in a vicinity of @xmath38 the identities @xmath305 hold true , where @xmath306 it follows from ( [ 3.8 ] ) and the weingarten formulas that @xmath307 where @xmath308 are the entries of the matrix @xmath309 , and all vectors are treated as rows .
a straightforward direct calculation allows us to check that the inverse matrix @xmath310 reads as follows , @xmath311 where , as before , @xmath312 indicates matrix transposition , and @xmath313 are the entries of the matrix @xmath314 .
let @xmath315 be any two functions with the corresponding supports located in a neighbourhood of @xmath38 , where the coordinates @xmath62 are well - defined .
we integrate by parts , @xmath316 hence , @xmath317 in view of ( [ 4.30 ] ) we have @xmath318 the obtained formulas and ( [ 4.34 ] ) imply the statement of the lemma .
we recall that the set @xmath319 was introduced in ( [ 3.12a ] ) .
[ lm4.3 ] let the functions @xmath320 satisfy the differentiable asymptotics @xmath321 uniformly in @xmath322 , where @xmath323 , and @xmath324 are some functions .
suppose the condition @xmath325 \\ & -\int\limits_{\p\om}(f_{-2}^+ - f_{-2}^- ) \psi_i^{(1)}\di s + \int\limits_{\p\om } ( f_{-2}^+ + f_{-2}^- ) \psi_i^{(0 ) } k \di s \\ & + 2\int\limits_{\p\om } \big(v^{(0 ) } \psi_i^{(1)}-v^{(1)}\psi_i^{(0)}\big)\di s=0,\quad i=1,\ldots , m , \end{aligned}\ ] ] holds true .
then there exist the unique solutions @xmath326 to the equations @xmath327 these solutions satisfy differentiable asymptotics @xmath328 uniformly in @xmath322 , where @xmath329 are some functions , and the condition @xmath330 holds true . let @xmath331 be the cut - off function introduced in the proof of lemma [ lm3.3 ] .
we introduce the functions @xmath332 employing lemma [ lm4.2 ] , one can check that @xmath333 where @xmath334 .
we construct the solutions to ( [ 4.80 ] ) as @xmath335 substituting this identity and ( [ 4.82 ] ) into ( [ 4.80 ] ) , we obtain the equations for @xmath336 , @xmath337 and by ( [ 4.78 ] ) we have @xmath338 .
hence , we can rewrite these equations as @xmath339 since @xmath127 is a discrete eigenvalue of @xmath340 , the solvability condition of the last equation is @xmath341 which can be rewritten as @xmath342 or , equivalently , @xmath343 integrating by parts and taking into account ( [ 4.82 ] ) , ( [ 4.84 ] ) , we get @xmath344 here we have used that the normal derivative on @xmath345 is that w.r.t . to @xmath39 up to the sign .
we parameterize the points of @xmath345 by those on @xmath38 via the relation @xmath346 . in view of (
[ 3.8 ] ) and ( [ 4.23 ] ) we have @xmath347 taking this formula into account , we continue the calculations , @xmath348 we substitute the last identities into ( [ 4.86 ] ) and arrive at ( [ 4.79 ] ) .
thus , the condition ( [ 4.79 ] ) imply the existence of solutions to ( [ 4.80 ] ) .
the functions @xmath349 satisfy ( [ 1.6 ] ) in the sense of traces .
denote @xmath350 the solution to ( [ 4.85 ] ) is defined up to a linear combination of the eigenfunctions . in view of the belongings
@xmath351 we can choose the mentioned linear combination of the eigenfunctions so that the condition ( [ 4.81a ] ) is satisfied .
then the solution to ( [ 4.85 ] ) is unique and the same is obviously true for ( [ 4.80 ] ) . to prove the asymptotics ( [ 4.81 ] )
it is sufficient to study the smoothness of @xmath336 at @xmath38 . by standard smoothness improving theorems
we conclude that @xmath352 .
moreover , given any @xmath353 , it is easy to construct the function @xmath354 similar to @xmath355 such that @xmath356 where @xmath357 , and @xmath358 , @xmath359 .
then , proceeding as above , we can construct the solutions to ( [ 4.80 ] ) as @xmath360 , where @xmath361 solves the equation @xmath362 it is clear that @xmath363 belongs to @xmath364 , where @xmath365 as @xmath359 .
hence , by the smoothness improving theorems @xmath366 , @xmath367 , @xmath359 . choosing @xmath368 large enough , we arrive at the asymptotics ( [ 4.81 ] ) .
[ lm4.4 ] for all @xmath369 in a small vicinity of @xmath38 the identities @xmath370 hold true .
let @xmath369 be two arbitrary functions with supports in a small vicinity @xmath371 , where @xmath372 is a small fixed number .
we choose @xmath372 so that in this vicinity the coordinates @xmath62 are well - defined . taking ( [ 3.3 ] ) and
( [ 3.8 ] ) into account , we pass to the variables @xmath62 and integrate by parts to obtain @xmath373 which proves ( [ 4.43 ] ) .
the identity ( [ 4.44 ] ) follows from ( [ 3.8 ] ) and ( [ 4.29a ] ) , @xmath374
in this section we prove theorem [ th1.1 ] . we begin with two auxiliary lemmas .
[ lm1.1 ] the identity @xmath375 holds true and for each @xmath376 the operator @xmath91 acts as @xmath377 . for each @xmath378
the estimate @xmath379 holds for some constant @xmath380 , where @xmath381 denotes the imaginary part of @xmath215 .
the first part follows from the definitions and the considerations above for the space @xmath382 .
the second part of the statement follows from the fact that the operator @xmath340 is self
adjoint with compact resolvent .
the description of the spectrum of @xmath91 as being made up of the union of the dirichlet and neumann spectra , is given in the following lemma , together with some properties which will be useful in the sequel .
[ lm1.2 ] the spectrum of @xmath91 coincides with the union of spectra of @xmath115 and @xmath116 counting multiplicities .
namely , if @xmath127 is an @xmath383-multiple eigenvalue of @xmath115 with the associated eigenfunctions @xmath384 , @xmath385 , and is an @xmath386-multiple eigenvalue of @xmath116 with the associated eigenfunctions @xmath387 , @xmath388 , then @xmath127 is @xmath389-multiple eigenvalue of @xmath91 with the associated eigenfunctions @xmath390 and @xmath391 . for any eigenfunction @xmath392 of @xmath91
we have @xmath393 and the asymptotics @xmath394 where @xmath395 and @xmath396 for small positive @xmath39 .
clearly if @xmath397 is an eigenvalue of @xmath115 with eigenfunction @xmath102 , then @xmath397 is an eigenvalue of @xmath91 with eigenfunction @xmath398 .
similarly , an eigenvalue of @xmath116 with eigenfunction @xmath214 will also be an eigenvalue of @xmath91 with eigenfunction @xmath399 .
assume now that @xmath400 is an eigenfunction of @xmath91 and consider the functions @xmath401 and @xmath402 .
then , provided they do not vanish identically , both @xmath403 and @xmath404 will be eigenfunctions of @xmath115 and @xmath116 , respectively . in case @xmath403 vanishes
identically , then @xmath405 and @xmath102 will be an eigenfuntion of @xmath116 , while if @xmath404 vanishes @xmath406 and this will be an eigenfunction of @xmath115 .
the remaining part of the lemma follows from standard arguments . by @xmath407
we indicate the subspace of @xmath69 consisting of the functions @xmath408 with the finite norm @xmath409 as consisting of @xmath410 with the finite norm @xmath411 where @xmath412 .
[ lm3.1 ] the spaces @xmath111 and @xmath407 are isomorphic and the isomorphism is the operator @xmath413 . if @xmath414 , then @xmath415 , and if @xmath416 , then @xmath417 .
the inequality @xmath418 holds true , where @xmath419 , @xmath70 . the fact that @xmath420 is a bijection between the two spaces follows directly from its definition . regarding the inequalities we have @xmath421 where we have used the knowledge of the eigenvalues of @xmath210 and the fact that @xmath422 .
denote @xmath423 .
we recall that the set @xmath319 was introduced in ( [ 3.12a ] ) , and in what follows @xmath424 is @xmath319 considered as a two - sided domain .
[ lm3.3 ] if @xmath425 , respectively , @xmath426 , then @xmath427 , respectively , @xmath428 .
the inequalities @xmath429 hold true , where @xmath380 are positive constants independent of @xmath7 and @xmath408 .
let @xmath410 , then @xmath430 , and for almost all @xmath37 the function @xmath431 belongs to @xmath432 .
let @xmath433 be an infinitely differentiable cut - off function vanishing as @xmath434 and being one as @xmath435 .
then @xmath436 for @xmath437 $ ] , and @xmath438,\ ] ] where @xmath380 is a positive constant independent of @xmath295 and @xmath71 .
we multiply the last inequality by @xmath439 , integrate over @xmath38 , and take into account ( [ 3.9 ] ) to obtain @xmath440 where @xmath380 is a positive constant independent of @xmath37 , and @xmath71 .
the above estimate , inequality ( [ 3.12 ] ) , the definition ( [ 3.4 ] ) of @xmath439 and the smoothness of @xmath34 imply @xmath441 , \\ & \int\limits_{\om^\d } |u_\pm|^2 j_\e^\pm\di x'\leqslant c(\d ) \|u_\pm\|_{l_2(\om^\d)}^2 , \\ & \int\limits_{\om_\d }
|u_\pm|^2 j_\e^\pm\di x ' = \int\limits_{0}^{\d}\di \tau \int\limits_{\p\om } |u_\pm|^2 j_\e^\pm \\ & \hphantom{\int\limits_{\om^\d } |u_\pm|^2 j_\e^\pm\di x ' = } \leqslant c\|u_\pm\|_{\h^1(\om)}^2 \int\limits_{0}^{\d}\sqrt{1+c_4\e^2\tau^{-1}}\di\tau , \end{aligned}\ ] ] where the constants @xmath380 and @xmath442 are independent of @xmath7 and @xmath71 , and @xmath380 is independent of @xmath443 . taking @xmath444 , we see that @xmath427 and thus the estimate ( [ 3.23a ] ) holds .
if we now take @xmath445 in ( [ 3.14 ] ) instead and use the identity @xmath446 we obtain ( [ 3.23c ] ) . let us prove ( [ 3.23e ] ) .
we integrate by parts as follows , @xmath447 by the embedding of @xmath448 into @xmath449 we have the estimate @xmath450 where the constants @xmath380 are independent of @xmath7 and @xmath408 .
these two last estimates together with ( [ 3.7 ] ) yield ( [ 3.23e ] ) . to prove the second part of the lemma related to the case @xmath451 it is sufficient to note that since @xmath452 , by the first part of the lemma these functions belong to @xmath453 , and the estimates ( [ 3.23a ] ) , ( [ 3.23c ] ) are valid for @xmath408 replaced by @xmath454 .
this completes the proof .
let @xmath455 , then @xmath456 .
denote @xmath457 , @xmath458 . by the definition of @xmath64 and @xmath340 we have @xmath459-z(u^{(\e)},\vp)_{l_2(\operatorname{\mathcal s}_\e)}=(f,\vp)_{l_2(\operatorname{\mathcal s}_\e ) } & & \text{for each}\quad\vp\in \h^1(\operatorname{\mathcal s}_\e ) , \label{3.16 } \\ & \ho[\bs{u}^{(0)},\bs{\vp}]-z(\bs{u}^{(0)},\bs{\vp})_{l_2(\bs{\om } ) } = ( \bs{f},\bs{\vp})_{l_2(\bs{\om } ) } & & \text{for each}\quad\vp\in \h^1(\bs{\om}).\label{3.17}\end{aligned}\ ] ] since @xmath460 , by lemmas [ lm3.2 ] and [ lm3.3 ] @xmath461 .
hence , @xmath462 and this can be used as a test function in ( [ 3.16 ] ) , @xmath463-z(u^{(\e)},v^{(\e)})_{l_2(s_\e)}=(f , v^{(\e)})_{l_2(s_\e)}.\ ] ] the identity @xmath464 yields @xmath465 we parameterize @xmath221 as @xmath466 , @xmath198 , and use the definition of the scalar product @xmath467 .
it implies @xmath468 where @xmath469 and @xmath470 are the entries of the inverse matrix @xmath471 .
we substitute the last formula into ( [ 3.19 ] ) and then sum it with ( [ 3.17 ] ) , where we take @xmath472 , @xmath473 let us estimate @xmath474 which we shall write as @xmath475 and @xmath476 .
as @xmath477 , by ( [ 3.12 ] ) we have @xmath478 hereinafter by @xmath380 we indicate non - essential positive constants independent of @xmath7 , @xmath479 , @xmath480 , and @xmath481 .
hence , by lemmas [ lm3.2 ] , [ lm3.3 ] and schwarz s inequality @xmath482 and therefore @xmath483 to estimate @xmath484 we employ ( [ 3.23a ] ) , ( [ 3.23c ] ) , ( [ 3.23e ] ) .
we begin with the first term in @xmath484 applying again schwarz s inequality and ( [ 3.23e ] ) to obtain @xmath485 employing ( [ 3.7 ] ) , ( [ 3.23a ] ) and ( [ 3.23e ] ) in the same way we get two more estimates , @xmath486 since @xmath487 by schwarz s inequality we have @xmath488 here we have used the inequality @xmath489 which follows from lemma [ lm3.2 ] . using ( [ 3.23d ] )
we get @xmath490 which with ( [ 3.25 ] ) and ( [ 3.26 ] ) yield @xmath491 together with ( [ 2.7a ] ) , ( [ 3.20 ] ) , ( [ 3.21 ] ) , ( [ 3.22 ] ) it follows that @xmath492 since @xmath493 we arrive at ( [ 1.8 ] ) , completing the proof .
the proof above uses the estimates from lemma [ lm3.3 ] which include a measure of the boundary behaviour by means of the weight function @xmath494 . a different approach which
may also be used to prove convergence of the resolvent in similar situations is based on inequalities of hardy type instead , possibly allowing for a better control of the behaviour near the boundary see @xcite for an illustration of this principle . in the proof of theorem [ th2.3 ] in the next section we shall use the following auxiliary lemma which is convenient to prove in this section .
[ lm3.4 ] let @xmath127 be a @xmath125-multiple eigenvalue of @xmath91 , and @xmath160 , @xmath150 , be the eigenvalues of @xmath49 taken counting multiplicity and converging to @xmath127 , and @xmath495 be the associated eigenfunctions orthonormalized in @xmath496 . for @xmath215 close to @xmath127 the representation @xmath497 holds true , where the operator @xmath498 is bounded uniformly in @xmath7 and @xmath215 .
the range of @xmath499 is orthogonal to all @xmath495 , @xmath150 .
we choose a fixed @xmath443 so that the disk latexmath:[$b_\d(\l):=\{z : except @xmath127 and @xmath501 then , by theorem [ th2.2 ] , for sufficiently small @xmath7 this disk contains the eigenvalues @xmath160 , @xmath150 , and no other eigenvalues of @xmath49 , and @xmath502 denote by @xmath503 the orthogonal complement to @xmath495 , @xmath150 , in @xmath496 . by (
3.5 , eqs .
( 3.21 ) ) the representation ( 3.29 ) holds true , where @xmath499 is the part of the resolvent @xmath504 acting in @xmath503 and @xmath505 for @xmath506 , where we have used ( [ 3.31 ] ) .
hence , the range of @xmath499 is orthogonal to @xmath495 , @xmath150 .
it is easy to check that the function @xmath507 , @xmath508 solves the equation @xmath509 hence , by the definition of @xmath49 and ( [ 3.32 ] ) @xmath510 where the constant @xmath442 is independent of @xmath7 and @xmath481 .
the last estimate and ( [ 3.32 ] ) complete the proof .
in this section we give the proof of theorem [ th2.3 ] which will be divided into two parts .
we first build the asymptotic expansions formally , where the core of the formal construction is the method of matching asymptotic expansions @xcite .
the second part is devoted to the justification of the asymptotics , i.e. , obtaining estimates for the error terms .
the formal construction consists of determining the outer and inner expansions on the base of the perturbed eigenvalue problem and the matching of these expansions .
the outer expansion is used to approximate the perturbed eigenfunctions outside a small neighborhood of @xmath38 .
it is constructed in terms of the variables @xmath55 using the first parametrization of @xmath33 given in the previous sections . in a vicinity of @xmath38
the perturbed eigenfunctions are approximated by the inner expansion which is based on the second parametrization of @xmath33 and is constructed in terms of the variables @xmath165 . by theorem [ th2.2 ] there
exist exactly @xmath125 eigenvalues of @xmath49 converging to @xmath127 counting multiplicities .
we denote these eigenvalues by @xmath151 , @xmath152 , while the symbols @xmath511 will denote the associated eigenfunctions .
we construct the asymptotics for @xmath151 as @xmath512 hereinafter terms like @xmath513 are understood as @xmath514 . in accordance with the method of matching asymptotic expansions we form the asymptotics for @xmath511 as the sum of outer and inner expansions .
the outer expansion is built as @xmath515 where @xmath516 , @xmath517 , and the eigenfunctions @xmath518 are chosen as described before the statement of theorem [ th2.3 ] in sec .
we also recall that these functions depend on @xmath7 in the case where @xmath127 is a multiple eigenvalue .
we substitute the identities ( [ 4.1 ] ) , ( [ 4.2 ] ) , and ( [ 5.1 ] ) into the eigenvalue equation @xmath519 and take into account the eigenvalue equations for @xmath145 .
it implies the equations for @xmath520 , namely , @xmath521 the functions @xmath522 are infinitely differentiable in @xmath523 , and thus @xmath524 as @xmath525 , where by the definition of the domain of @xmath91 @xmath526 the functions @xmath527 depend on @xmath7 only if @xmath127 is a multiple eigenvalue , since the same is true for the functions @xmath518 . in view of the identity ( [ 4.55 ] ) we rewrite ( [ 4.7 ] ) as @xmath528 in accordance with the method of matching asymptotic expansions the identities ( [ 4.2 ] ) , ( [ 5.5a ] ) yield that the inner expansion for the eigenfunctions @xmath511 should read as follows , @xmath529 where the coefficients must satisfy the following asymptotics as @xmath530 @xmath531 these asymptotics mean that the first term of the outer expansion is matched with the inner expansion .
we substitute ( [ 4.1 ] ) , ( [ 4.12 ] ) , ( [ 4.32 ] ) , ( [ 4.41 ] ) into the eigenvalue equation ( [ 4.5 ] ) and equate the coefficients of @xmath532 .
this implies the equation for @xmath533 , @xmath534 the solution to the last equation satisfying ( [ 4.13 ] ) is obviously as follows , @xmath535 we then substitute this identity and ( [ 4.1 ] ) , ( [ 4.12 ] ) , ( [ 4.32 ] ) , ( [ 4.39a ] ) , ( [ 4.40a ] ) , ( [ 4.32 ] ) into ( [ 4.5 ] ) and equate the coefficients at @xmath536 , @xmath537 , leading us to the equations for @xmath538 , @xmath539 , @xmath540 were we have used that @xmath541 due to ( [ 4.39a ] ) , ( [ 4.40a ] ) , ( [ 4.45 ] ) .
the only solution to ( [ 4.46a ] ) satisfying ( [ 4.18a ] ) is independent of @xmath294 , @xmath542 where @xmath543 is an unknown function to be determined
. the equation ( [ 4.46 ] ) can be solved , and the solution satisfying ( [ 4.14 ] ) is @xmath544 where @xmath545 is an unknown function to be determined . in view of ( [ 4.46b ] ) , ( [ 4.47 ] ) ,
( [ 4.39a ] ) , ( [ 4.40a ] ) and ( [ 4.46 ] ) , equation ( [ 4.46c ] ) may be written as @xmath546 employing the formulas ( [ 4.41 ] ) , ( [ 4.47 ] ) and ( [ 4.48a ] ) , we solve the last equation , @xmath547 where @xmath548 and @xmath549 are unknown functions to be determined .
we substitute ( [ 4.46b ] ) , ( [ 4.47 ] ) , ( [ 4.48a ] ) , ( [ 4.49a ] ) , ( [ 4.39a ] ) , ( [ 4.40a ] ) , ( [ 4.41b ] ) , ( [ 4.27a ] ) and ( [ 4.41 ] ) into equation ( [ 4.46d ] ) and then solve it to obtain @xmath550 where @xmath551 , @xmath552 and @xmath553 and @xmath554 are unknown functions to be determined . to determine the coefficient @xmath555 in the outer expansion and the functions @xmath556 in the inner one
, we should match the constructed functions @xmath538 with the outer expansion . in order to do it
, we must find the asymptotics for the functions @xmath538 as @xmath530 .
we observe that the functions @xmath557 satisfy the identities @xmath558 uniformly in @xmath559 , with @xmath560 any fixed constant .
taking these asymptotics into account , we write the asymptotics for @xmath538 as @xmath530 and then pass to the variables @xmath561 , @xmath562 where @xmath563 taking into account the obtained formulas and ( [ 4.2 ] ) , in accordance with the method of matching asymptotic expansions we conclude that @xmath564 while the solutions to the equation ( [ 4.6b ] ) should satisfy the asymptotics @xmath565 moreover , the identity @xmath566 should hold .
we substitute ( [ 4.23 ] ) and ( [ 4.7 ] ) into the eigenvalue equation for @xmath567 and equate the coefficient of @xmath568 .
it leads us to identity ( [ 4.67 ] ) .
we proceed to the problem ( [ 4.6b ] ) , ( [ 4.35b ] ) .
to study its solvability we shall make use of one more auxiliary lemma .
recall that the matrices @xmath569 and @xmath570 are defined in ( [ 3.8 ] ) and ( [ 4.29a ] ) , respectively .
[ lm4.5 ] the functions @xmath571 introduced in ( [ 4.6b ] ) satisfy the hypothesis of lemma [ lm4.3 ] . in particular , the asymptotics ( [ 4.78 ] ) holds true with @xmath572 we begin with an obvious identity @xmath573 which follows from the definition of @xmath571 in ( [ 4.6b ] ) . to prove the lemma , we shall pass to the variables @xmath62 in the obtained identity .
it follows from ( [ 4.8 ] ) , ( [ 4.55 ] ) and the definition of @xmath221 that @xmath574 hence , by ( [ 4.9a ] ) , ( [ 4.52 ] ) @xmath575 thus , employing ( [ 3.8 ] ) and ( [ 4.91 ] ) , we conclude that the functions @xmath576 satisfy the hypothesis of lemma [ lm4.3 ] and in particular the asymptotics ( [ 4.78 ] ) holds true .
it remains to prove the identities ( [ 4.90 ] ) .
it follows from ( [ 4.44 ] ) that @xmath577 ) into the obtained identity and arrive at the asymptotics for @xmath578 , @xmath579 employing these formulas and ( [ 3.8 ] ) , ( [ 4.29a ] ) , ( [ 4.7 ] ) and ( [ 4.44 ] ) we rewrite the second term in the right hand side of ( [ 4.91 ] ) as follows , @xmath580 where @xmath581 are some functions , and , in particular , @xmath582 to obtain the same asymptotics for the first term in the right hand side of ( [ 4.91 ] ) , we employ first ( [ 4.43 ] ) , @xmath583 it follows from the equations ( [ 4.23 ] ) , ( [ 4.29a ] ) , ( [ 4.92 ] ) that @xmath584 where @xmath585 are some functions , @xmath586 are some @xmath15-dimensional vector - functions , and @xmath587 and @xmath588 .
we substitute the last identities into ( [ 4.100 ] ) , which yields @xmath589 the last identity , ( [ 4.98 ] ) , ( [ 4.99 ] ) , ( [ 4.91 ] ) imply the formulas ( [ 4.90 ] ) . taking into account ( [ 4.7 ] ) , we apply lemma [ lm4.5 ] to problem ( [ 4.6b ] ) .
it implies that the right hand side of ( [ 4.6b ] ) satisfies the hypothesis of lemma [ lm4.3 ] with the first four coefficients given by ( [ 4.90 ] ) .
given some functions @xmath590 , suppose the solvability condition ( [ 4.79 ] ) holds true .
then by ( [ 4.81 ] ) , ( [ 4.67 ] ) , ( [ 4.90 ] ) there exists the unique solution to ( [ 4.6b ] ) with the asymptotics @xmath591 where @xmath592 are some functions satisfying ( [ 4.81a ] ) .
we compare the last asymptotics with ( [ 4.56a ] ) , ( [ 4.56b ] ) , ( [ 4.35b ] ) , take into consideration the identity ( [ 4.67 ] ) and arrive at the formulas for @xmath593 , @xmath594 , @xmath595 and @xmath596 , @xmath597 in what follows the functions @xmath593 , @xmath594 , @xmath595 and @xmath596 are supposed to be chosen in accordance with the above given formulas .
bearing these formulas , ( [ 4.67 ] ) and ( [ 4.90 ] ) in mind , we write the solvability conditions ( [ 4.79 ] ) for the equation ( [ 4.6b ] ) , @xmath598 \\ & + \int\limits_{\p\om}\frac{b_1 ^ 2}{2 \ln\e}(2\ln 2-\ln b_1 + 1 ) \psi_i^{(0 ) } ( \d_{\p\om}+ \l)\psi_k^{(0)}\di s \\ & + \int\limits_{\p\om } \frac{b_1}{\ln\e } ( 2\ln 2-\ln b_1 - 1 ) \psi_i^{(0 ) } \nabla b_1\cdot\nabla \psi_k^{(0)}\di s \\ & + \int\limits_{\p\om}\frac{b_1 ^ 2}{2\ln\e}(2\ln 2- \ln b_1 ) \psi_k^{(1 ) } \psi_i^{(1)}\di s \\ &
-\int\limits_{\p\om}\frac{b_1 ^ 2}{2 } \left(\psi_k^{(1 ) } \psi_i^{(1)}+
\psi_i^{(0 ) } \bigg ( \d_{\p\om } + \frac{2}{b_1}\nabla b_1\cdot \nabla+ \l\bigg)\psi_k^{(0 ) } \right)\di s \\ & + \mu_k\d_{ik}=0,\quad i , k=1,\ldots , m .
\end{aligned}\ ] ] let us simplify the obtained identity .
we first rewrite the formulas ( [ 4.6b ] ) of @xmath571 in a more convenient form employing the eigenvalue equation for @xmath567 and the definition of the matrix @xmath60 , @xmath599 employing this representation , we integrate by parts to obtain @xmath600 applying ( [ 4.44 ] ) , we have @xmath601 in a vicinity of @xmath38 .
hence , by ( [ 4.7 ] ) , ( [ 4.92 ] ) and ( [ 4.93 ] ) , @xmath602 substituting the last identity into ( [ 4.20 ] ) and using ( [ 4.21 ] ) and ( [ 4.67 ] ) , we get @xmath603 we integrate by parts once again , this time over @xmath38 , we have @xmath604 substituting two the last identities into ( [ 4.89 ] ) yields @xmath605 \\ & + \int\limits_{\p\om}\frac{b_1 ^ 2}{4\ln\e}(1 + 4\ln 2 - 2\ln b_1 ) \big(\psi_i^{(1)}\psi_k^{(1)}+ \psi_i^{(0 ) } ( \d_{\p\om}+ \l)\psi_k^{(0 ) } \big)\di s \\ & + \int\limits_{\p\om } \frac{b_1}{\ln\e } ( 2\ln 2-\ln b_1 ) \psi_i^{(0 ) } \nabla b_1\cdot\nabla \psi_k^{(0)}\di s \\ & -\int\limits_{\p\om}\frac{b_1 ^ 2}{2 } \left(\psi_k^{(1 ) } \psi_i^{(1)}+ \psi_i^{(0 ) } \bigg ( \d_{\p\om } + \frac{2}{b_1}\nabla b_1\cdot \nabla+ \l\bigg)\psi_k^{(0 ) } \right)\di s \\ & + \mu_k\d_{ik}=0 , \end{aligned}\ ] ] as @xmath606 .
it follows from ( [ 4.25 ] ) , ( [ 4.94 ] ) and ( [ 4.7 ] ) that @xmath607 hence , the limit in ( [ 4.51 ] ) is finite . to calculate the boundary integrals in ( [ 4.51 ] )
we integrate by parts as follows @xmath608 due to this identity , ( [ 4.103 ] ) , the definition of @xmath609 in ( [ 4.52 ] ) and the definitions ( [ 2.11a ] ) and ( [ 2.11b ] ) of the matrices @xmath140 and @xmath141 , respectively , we can rewrite ( [ 4.51 ] ) in the final form @xmath610 since the matrix on the right hand side of the last identity is diagonal , we conclude that the solvability condition for the problem ( [ 4.6b ] ) , ( [ 4.35b ] ) is satisfied provided @xmath154 are the eigenvalues of the matrix @xmath611 .
it follows from ( * ? ? ?
6.1 ) that the eigenvalues of this matrix are holomorphic in @xmath158 and converge to those of @xmath140 as @xmath159 . in view of the choice of @xmath612 the problems ( [ 4.6b ] ) , ( [ 4.95 ] ) are solvable .
we observe that each of the functions @xmath613 is defined up to a linear combination of the eigenfunctions @xmath522 .
the exact values of the coefficients of these linear combinations can be determined while constructing the next terms in the asymptotic expansions for @xmath151 and @xmath511 .
the formal constructing of the asymptotic expansions is complete . in order to justify the obtained asymptotics
, one has to construct additional terms .
this is a general and standard situation for singularly perturbed problems . in our case one
should construct the terms of the order up to @xmath614 in the outer expansion for the eigenfunctions and for the eigenvalues , and the terms of order up to @xmath615 in the inner expansion for the eigenfunctions .
the asymptotics with the additional terms read as follows , @xmath616 where @xmath617 , @xmath618 , @xmath619 , and we used that @xmath620 by ( [ 4.46b ] ) , ( [ 4.56c ] ) .
the equations for @xmath621 are @xmath622 the functions @xmath621 should satisfy the asymptotics @xmath623 the equations for the functions @xmath624 , @xmath625 are obtained in the same way as those for @xmath538 , @xmath626 , from @xmath627 where the operators @xmath628 , @xmath629 are the next terms in the expansion ( [ 4.32 ] ) . it can be shown that the problem for @xmath621 is solvable for some @xmath630 . the equations for @xmath624 and @xmath625 can be solved explicitly .
the arbitrary coefficients @xmath631 , @xmath632 , @xmath633 , @xmath634 appearing in @xmath624 , @xmath625 can be determined while matching the inner and outer expansions .
we now introduce the partial sums @xmath635 and define the final approximation for the eigenfunctions as @xmath636 where @xmath637 is a fixed constant , and @xmath638 is the cut - off function introduced in the proof of lemma [ lm3.3 ] .
[ lm4.6 ] the function @xmath639 satisfies the convergence @xmath640 and the equation @xmath641 where for the right hand side the uniform in @xmath7 estimate @xmath642 holds true .
the relations @xmath643 are valid .
the proof of this lemma is not very difficult and is based on lengthy and rather technical , but straightforward calculations . because of this , and in order not to overload the text with long technical formulas we shall skip these here .
it follows from lemma [ lm3.4 ] and equation ( [ 4.107 ] ) that @xmath644 and , by ( [ 4.108 ] ) , @xmath645 where the constant @xmath380 is independent of @xmath7 .
we calculate the scalar products of the functions @xmath646 in @xmath496 taking into consideration ( [ 4.110 ] ) and the properties of the operator @xmath647 described in lemma [ lm3.4 ] : @xmath648 the identities obtained and ( [ 4.111 ] ) , ( [ 4.106 ] ) , ( [ 4.109 ] ) yield @xmath649 in particular , as @xmath650 it implies latexmath:[\[\label{4.113 } @xmath7 .
we introduce the matrix @xmath652 and rewrite ( [ 4.115 ] ) as @xmath653 , @xmath65 , where @xmath312 denotes matrix transposition .
thus , @xmath654 as @xmath65
. therefore , for each sufficiently small @xmath7 there exists a permutation @xmath655 such that @xmath656 for a given @xmath7 we rearrange the eigenvalues @xmath160 and @xmath646 so that @xmath657 that by ( [ 4.113 ] ) , ( [ 4.114 ] ) it yields @xmath658 in view of the definition of @xmath659 , ( [ 4.108 ] ) , and the normalization of @xmath495 it follows @xmath660 choosing @xmath661 , we arrive at the asymptotics ( [ 2.16 ] ) .
denote now @xmath662 by direct calculations one can check that @xmath663 this identity and ( [ 4.111 ] ) imply @xmath664 since the right hand sides of these identities are linear independent , the functions @xmath665 form a basis spanned over the eigenfunctions @xmath495 , @xmath150 .
hence , we arrive at [ th4.7 ] let @xmath129 be the total projector associated with the eigenvalues @xmath160 , @xmath150 , @xmath666 be the projector on the space spanned over @xmath667 , @xmath150
. then @xmath668 where @xmath155 is any constant in @xmath156 .
both authors were partially supported by fct s projects ptdc/ mat/ 101007/2008 and pest - oe / mat / ui0208/2011 .
d.b . was partially supported by rfbr and by federal task program ( contract 02.740.11.0612 ) .
d. borisov and p. freitas , singular asymptotic expansions for dirichlet eigenvalues and eigenfunctions of the laplacian on thin planar domains , ann .
h. poincar anal .
non linaire * 26 * ( 2009 ) 547 - 560 . | we study the asymptotic behaviour of the eigenvalues of the laplace beltrami operator on a compact hypersurface in @xmath0 as it is _ flattened _ into a singular double sided flat hypersurface .
we show that the limit spectral problem corresponds to the dirichlet and neumann problems on one side of this flat ( euclidean ) limit , and derive an explicit three - term asymptotic expansion for the eigenvalues where the remaining two terms are of orders @xmath1 and @xmath2 . |
quantum information processing as a growing exciting field has attracted researchers from different disciplines .
it utilizes the laws of quantum mechanical operations to perform exponentially speedy computations . in an open system
, one might wonder how to perform such computations in the presence of decoherence and noise that disturb quantum states storing quantum information .
ultimately , the goals of quantum error - correcting codes are to protect quantum states and to allow recovery of quantum information processed in computational operations of a quantum computer .
henceforth , one seeks to design good quantum codes that can be efficiently utilized for these goals . a well - known approach to derive quantum error - correcting codes from self - orthogonal ( or dual - containing ) classical codes
is called stabilizer codes , which were introduced a decade ago .
the stabilizer codes inherit some properties of clifford group theory , i.e. , they are stabilized by abelian finite groups . in the seminal paper by calderbank _ at .
@xcite , various methods of stabilizer code constructions are given , along with their propagation rules and tables of upper bounds on their parameters . in a similar tactic , we also present subsystem code structures by establishing several methods to derive them easily from classical codes .
subsystem codes inherit their name from the fact that the quantum codes are decomposed into two systems as explained in section [ sec : background ] .
the classes of subsystem codes that we will derive are superior because they can be encoded and decoded using linear shirt - register operations .
in addition , some of these classes turned out to be optimal and mds codes .
subsystem codes as we prefer to call them were mentioned in the unpublished work by knill @xcite , in which he attempted to generalize the theory of quantum error - correcting codes into subsystem codes .
such codes with their stabilizer formalism were reintroduced recently @xcite .
an @xmath0 subsystem code is a @xmath1-dimensional subspace @xmath2 of @xmath3 that is decomposed into a tensor product @xmath4 of a @xmath5-dimensional vector space @xmath6 and an @xmath7-dimensional vector space @xmath8 such that all errors of weight less than @xmath9 can be detected by @xmath6 .
the vector spaces @xmath6 and @xmath8 are respectively called the subsystem @xmath6 and the co - subsystem @xmath8 . for some background on subsystem codes see the next section .
this paper is structured as follows . in section
[ sec : background ] , we present a brief background on subsystem code structures and present the euclidean and hermitian constructions . in section [ sec : cyclicsubsys ] , we derive cyclic subsystem codes and provide two generic methods of their constructions from classical cyclic codes . consequently in section [ sec : dimensions ] , we construct families of subsystem bch and rs codes from classical bch and rs over @xmath10 and @xmath11 defined using their defining sets . in sections [ sec :
mdssubsys],[sec : extendshortensubsys],[sec : combinesubsys ] , we establish various methods of subsystem code constructions by extending and shortening the code lengths and combining pairs of known codes , in addition , tables of upper bounds on subsystem code parameters are given . finally , the paper is concluded with a discussion and future research directions in section [ sec : conclusion ]
. _ notation .
_ if @xmath12 is a set , then @xmath13 denotes the cardinality of the set @xmath12 .
let @xmath14 be a power of a prime integer @xmath15 .
we denote by @xmath10 the finite field with @xmath14 elements .
we use the notation @xmath16 to denote the concatenation of two vectors @xmath17 and @xmath18 in @xmath19 .
the symplectic weight of @xmath20 is defined as @xmath21 we define @xmath22 for any nonempty subset @xmath23 of @xmath24 . the trace - symplectic product of two vectors @xmath25 and @xmath26 in @xmath24 is defined as @xmath27 where @xmath28 denotes the dot product and @xmath29 denotes the trace from @xmath10 to the subfield @xmath30 .
the trace - symplectic dual of a code @xmath31 is defined as @xmath32 we define the euclidean inner product @xmath33 and the euclidean dual of @xmath34 as @xmath35 we also define the hermitian inner product for vectors @xmath36 in @xmath37 as @xmath38 and the hermitian dual of @xmath39 as @xmath40
in this section we give a quick overview of subsystem codes .
we assume that the reader is familiar the theory of stabilizer codes over finite fields , see @xcite and the references therein .
let @xmath10 denote a finite field with @xmath14 elements of characteristic @xmath15 .
let @xmath41 be a fixed orthonormal basis of @xmath42 with respect to the standard hermitian inner product , called the computational basis . for @xmath43 , we define the unitary operators @xmath44 and @xmath45 on @xmath42 by @xmath46 where @xmath47 is a primitive @xmath15th root of unity and @xmath48 is the trace operation from @xmath10 to @xmath30 .
the set @xmath49 forms an orthogonal basis of the operators acting on @xmath42 with respect to the trace inner product , called the error basis .
the state space of @xmath50 quantum digits ( or qudits ) is given by @xmath51 .
an error basis @xmath52 on @xmath3 is obtained by tensoring @xmath50 operators in @xmath53 ; more explicitly , @xmath54 where @xmath55 for @xmath56 and @xmath57 .
the set @xmath53 is not closed under multiplication , whence it is not a group .
the group @xmath58 generated by @xmath52 is given by @xmath59 and @xmath58 is called the error group of @xmath3 .
the error group is an extraspecial @xmath15-group .
the weight of an error in @xmath58 is given by the number of nonidentity tensor components ; hence , the weight of @xmath60 is given by the symplectic weight @xmath61 .
an @xmath0 subsystem code is a subspace @xmath4 of @xmath3 that is decomposed into a tensor product of two vector spaces @xmath6 and @xmath8 of dimension @xmath62 and @xmath63 such that all errors in @xmath58 of weight less than @xmath9 can be detected by @xmath6 .
we call @xmath6 the subsystem and @xmath8 the co - subsystem . the information is exclusively encoded in the subsystem @xmath6 .
this yields the attractive feature that errors affecting co - subsystem @xmath8 alone can be ignored . a particularly fruitful way to construct subsystem codes proceeds by choosing a normal subgroup @xmath64 of the error group @xmath58 , and
this choice determines the dimensions of subsystem and co - subsystem as well as the error detection and correction capabilities of the subsystem code , see @xcite .
one can relate the normal subgroup @xmath64 to a classical code , namely @xmath64 modulo the intersection of @xmath64 with the center @xmath65 of @xmath58 yields the classical code @xmath66 .
this generalizes the familiar case of stabilizer codes , where @xmath64 is an abelian normal subgroup .
it is remarkable that in the case of subsystem codes _ any _ classical additive code @xmath67 can occur .
it is most convenient that one can also start with any classical additive code and obtain a subsystem code , as is detailed in the following theorem from @xcite : [ th : oqecfq ] let @xmath68 be a classical additive subcode of @xmath24 such that @xmath69 and let @xmath70 denote its subcode @xmath71 . if @xmath72 and @xmath73 , then there exists a subsystem code @xmath74 such that @xmath75 , @xmath76 .
the minimum distance of subsystem @xmath6 is given by @xmath77 if @xmath78 ;
@xmath79 if @xmath80 .
thus , the subsystem @xmath6 can detect all errors in @xmath53 of weight less than @xmath9 , and can correct all errors in @xmath53 of weight @xmath81 .
see ( * ? ? ?
* theorem 5 ) . a subsystem code that is derived with the help of the previous theorem
is called a clifford subsystem code .
we will assume throughout this paper that all subsystem codes are clifford subsystem codes .
in particular , this means that the existence of an @xmath0 subsystem code implies the existence of an additive code @xmath82 with subcode @xmath71 such that @xmath83 , @xmath84 , and @xmath85 . a subsystem code derived from an additive classical code @xmath68
is called pure to @xmath86 if there is no element of symplectic weight less than @xmath86 in @xmath68 .
a subsystem code is called pure if it is pure to the minimum distance @xmath9 .
we require that an @xmath87 subsystem code must be pure .
we also use the bracket notation @xmath88_q$ ] to write the parameters of an @xmath89 subsystem code in simpler form .
some authors say that an @xmath88_q$ ] subsystem code has @xmath90 gauge qudits , but this terminology is slightly confusing , as the co - subsystem typically does not correspond to a state space of @xmath90 qudits except perhaps in trivial cases .
we will avoid this misleading terminology .
an @xmath91 subsystem code is also an @xmath92 stabilizer code and vice versa .
subsystem codes can be constructed from the classical codes over @xmath10 and @xmath11 .
we recall the euclidean and hermitian constructions from @xcite , which are easy consequences of the previous theorem .
[ lem : css - euclidean - subsys ] if @xmath68 is a @xmath93-dimensional @xmath10-linear code of length @xmath50 that has a @xmath94-dimensional subcode @xmath95 and @xmath96 , then there exists an @xmath97_q\ ] ] subsystem code .
[ lem : css - hermitina - subsys ] if @xmath68 is a @xmath93-dimensional @xmath11-linear code of length @xmath50 that has a @xmath94-dimensional subcode @xmath98 and @xmath96 , then there exists an @xmath99_q\ ] ] subsystem code .
in this section we shall derive subsystem codes from classical cyclic codes .
we first recall some definitions before embarking on the construction of subsystem codes . for further details concerning cyclic codes see for instance @xcite and @xcite .
let @xmath50 be a positive integer and @xmath10 a finite field with @xmath14 elements such that @xmath100 . recall that a linear code @xmath101 is called _ cyclic _ if and only if @xmath102 in @xmath68 implies that @xmath103 in @xmath68 . for @xmath104 in @xmath105 $ ] , we write @xmath106 to denote the principal ideal generated by @xmath104 in @xmath105 $ ]
. let @xmath107 denote the vector space isomorphism @xmath108 / ( x^n-1)$ ] given by @xmath109 a cyclic code @xmath110 is mapped to a principal ideal @xmath111 of the ring @xmath112 . for a cyclic code @xmath68 , the unique monic polynomial @xmath104 in @xmath105 $ ] of the least degree such that @xmath113 is called the _ generator polynomial _ of @xmath68 .
if @xmath101 is a cyclic code with generator polynomial @xmath104 , then @xmath114 since @xmath100 , there exists a primitive @xmath115 root of unity @xmath116 over @xmath10 ; that is , @xmath117 $ ] is the splitting field of the polynomial @xmath118 over @xmath10 .
let us henceforth fix this primitive @xmath115 primitive root of unity @xmath116 .
since the generator polynomial @xmath104 of a cyclic code @xmath101 is of minimal degree , it follows that @xmath104 divides the polynomial @xmath118 in @xmath105 $ ] .
therefore , the generator polynomial @xmath104 of a cyclic code @xmath101 can be uniquely specified in terms of a subset @xmath119 of @xmath120 such that @xmath121 the set @xmath119 is called the _ defining set _ of the cyclic code @xmath68 ( with respect to the primitive @xmath115 root of unity @xmath116 ) . since @xmath104 is a polynomial in @xmath105 $ ] , a defining set is the union of cyclotomic cosets @xmath122 , where @xmath123 the following lemma recalls some well - known and easily proved facts about defining sets ( see e.g. @xcite ) .
[ lem : definingsets ] let @xmath124 be a cyclic code of length @xmath50 over @xmath10 with defining set a @xmath125 for @xmath126 .
let @xmath127 and @xmath128 for some integer @xmath129
. then @xmath130 has defining set @xmath131 .
@xmath132 has defining set @xmath133 .
@xmath134 if and only if @xmath135 .
@xmath136 has defining set @xmath137 .
@xmath138 has defining set @xmath139 provided that @xmath140 for some positive integer @xmath90 .
_ notation .
_ throughout this section , we denote by @xmath64 the set @xmath141
. the cyclotomic coset of @xmath17 will be denoted by @xmath122 .
if @xmath119 is a defining set of a cyclic code of length @xmath50 and @xmath129 is an integer , then we denote henceforth by @xmath142 the set @xmath143 as in the previous lemma .
we use a superscript , since this notation will be frequently used in set differences , and arguably @xmath144 is more readable than @xmath145 .
now , we shall give a general construction for subsystem cyclic codes .
we say that a code @xmath68 is self - orthogonal if and only if @xmath146 .
we show that if a classical cyclic code is self - orthogonal , then one can easily construct cyclic subsystem codes .
[ lem : cyclic - subsysi ] let @xmath70 be a @xmath147-dimensional self - orthogonal cyclic code of length @xmath50 over @xmath10 .
let @xmath148 and @xmath149 respectively denote the defining sets of @xmath70 and @xmath150 . if @xmath119 is a subset of @xmath151 that is the union of cyclotomic cosets , then one can define a cyclic code @xmath68 of length @xmath50 over @xmath10 by the defining set @xmath152 .
if @xmath153 is in the range @xmath154 , and @xmath155 , then there exists a subsystem code with parameters @xmath156_q$ ] .
since @xmath70 is a self - orthogonal cyclic code , we have @xmath157 , whence @xmath158 by lemma [ lem : definingsets ] iii ) .
observe that if @xmath159 is an element of the set @xmath160 , then @xmath161 is an element of @xmath12 as well .
in particular , @xmath162 is a subset of @xmath163 . by definition ,
the cyclic code @xmath68 has the defining set @xmath164 ; thus , the dual code @xmath165 has the defining set @xmath166 furthermore , we have @xmath167 therefore , @xmath168 by lemma [ lem : definingsets ] i ) . since @xmath169 and @xmath170 , we have @xmath171 and @xmath172 .
thus , by lemma [ lem : css - euclidean - subsys ] there exists an @xmath10-linear subsystem code with parameters @xmath173_q$ ] , where @xmath174 , @xmath175 , @xmath176 , as claimed
. we can also derive subsystem codes from cyclic codes over @xmath11 by using cyclic codes that are self - orthogonal with respect to the hermitian inner product .
[ lem : cyclic - subsysii ] let @xmath70 be a cyclic code of length @xmath50 over @xmath11 such that @xmath177 .
let @xmath148 and @xmath178 respectively be the defining set of @xmath70 and @xmath179
. if @xmath119 is a subset of @xmath180 that is the union of cyclotomic cosets , then one can define a cyclic code @xmath68 of length @xmath50 over @xmath11 with defining set @xmath181 .
if @xmath169 and @xmath182 with @xmath183 , and @xmath184 , then there exists an @xmath185_q$ ] subsystem code . since @xmath186 , their defining sets satisfy @xmath187 by lemma [ lem : definingsets ] iii ) . if @xmath159 is an element of @xmath188 , then one easily verifies that @xmath189 is an element of @xmath188 .
let @xmath127 . since the cyclic code @xmath68 has the defining set @xmath181 , its dual code @xmath190 has the defining set @xmath191 we notice that @xmath192 thus , @xmath193 by lemma [ lem : definingsets ] i ) . since @xmath169 and @xmath182 , we have @xmath194 and @xmath195 .
thus , by lemma [ lem : css - hermitina - subsys ] there exists an @xmath173_q$ ] subsystem code with @xmath196 , @xmath197 , @xmath198 , as claimed .
we include an example to illustrate the construction given in the previous proposition .
consider the narrow - sense bch code @xmath179 of length @xmath199 over @xmath200 with designed distance 5 .
the defining set @xmath178 of @xmath179 is given by @xmath201 , where the cyclotomic cosets of @xmath202 and @xmath203 are given by @xmath204 if @xmath205 , then the defining set of the dual code @xmath70 is given by @xmath206 therefore , @xmath207 , @xmath208 and @xmath209 .
if we choose @xmath210 , then @xmath211 , whence the defining set @xmath212 of the code @xmath68 is given by @xmath213 it follows that @xmath214 and @xmath215 .
therefore , the construction of the previous proposition yields a bch subsystem code with parameters @xmath216_2.$ ] the general principle behind the previous example yields the following simple recipe for the construction of subsystem codes : choose a cyclic code ( such as a bch or reed - solomon code ) with known lower bound @xmath217 on the minimum distance that contains its ( hermitian ) dual code , and use proposition [ lem : cyclic - subsysi ] ( or proposition [ lem : cyclic - subsysii ] ) to derive subsystem codes .
this approach allows one to control the minimum distance @xmath9 of the subsystem code , since @xmath218 is guaranteed .
another advantage is that one can exploit the cyclic structure in encoding and decoding algorithms .
for example , if we start with primitive , narrow - sense bch codes , then proposition [ lem : cyclic - subsysi ] yields the following family of subsystem codes : consider a primitive , narrow - sense bch code of length @xmath219 with @xmath220 over @xmath10 with designed distance @xmath217 in the range @xmath221.\ ] ] if @xmath119 is a subset of @xmath222 that is a union of cyclotomic cosets and @xmath170 with @xmath154 , where @xmath223 , then there exists an @xmath224_q\ ] ] subsystem code . by
* theorem 2 ) , a primitive , narrow - sense bch code @xmath150 with designed distance @xmath217 in the range ( [ eq : ddistrange ] ) satisfies @xmath225 . by (
* theorem 7 ) , the dimension of @xmath150 is given by @xmath226 , whence @xmath227 .
let @xmath148 and @xmath149 respectively denote the defining sets of @xmath70 and @xmath150 .
it follows from the definitions that @xmath228 and that @xmath119 is a subset of @xmath229 if @xmath230 denotes the defining set of a cyclic code @xmath68 , then @xmath231 . by proposition [ lem : cyclic - subsysi ]
, there exists an @xmath232_q$ ] subsystem code , which proves the claim .
.subsystem bch codes that are derived using the euclidean construction [ cols="<,<,^ " , ] + @xmath233 punctured code + @xmath234 extended code
in this section we show how one can trade the dimensions of subsystem and co - subsystem to obtain new codes from a given subsystem or stabilizer code .
the results are obtained by exploiting the symplectic geometry of the space .
a remarkable consequence is that nearly any stabilizer code yields a series of subsystem codes .
our first result shows that one can decrease the dimension of the subsystem and increase at the same time the dimension of the co - subsystem while keeping or increasing the minimum distance of the subsystem code .
[ th : shrinkk ] let @xmath14 be a power of a prime @xmath15 . if there exists an @xmath0 subsystem code with @xmath235 that is pure to @xmath86 , then there exists an @xmath236 subsystem code that is pure to @xmath237 .
if a pure @xmath238 subsystem code exists , then there exists a @xmath239 subsystem code . by definition
, an @xmath0 clifford subsystem code is associated with a classical additive code @xmath240 and its subcode @xmath71 such that @xmath72 , @xmath73 , @xmath241 , @xmath242 , and @xmath85 if @xmath243 , otherwise @xmath79 if @xmath80 .
we have @xmath244 for some positive integer @xmath245 .
since @xmath5 and @xmath7 are positive integers , we have @xmath246 and @xmath247 for some integers @xmath248 , and @xmath249 .
there exists an @xmath30-basis of @xmath68 of the form @xmath250 that can be extended to a symplectic basis @xmath251 of @xmath24 , that is , @xmath252 , @xmath253 , @xmath254 for all @xmath255 , see ( * ? ? ?
* theorem 8.10.1 ) .
define an additive code @xmath256 it follows that @xmath257 and @xmath258 by definition , the code @xmath68 is a subset of @xmath259 .
the subsystem code defined by @xmath259 has the parameters @xmath260 , where @xmath261 and @xmath262 .
for the claims concerning minimum distance and purity , we distinguish two cases : if @xmath263 , then @xmath235 and @xmath264 .
since by hypothesis @xmath265 and @xmath266 , and @xmath267 by construction , we have @xmath268 ; thus , the subsystem code is pure to @xmath237 .
if @xmath269 , then @xmath270 , that is , @xmath271 ; it follows from the assumed purity that @xmath272 .
this proves the claim
. for @xmath10-linear subsystem codes there exists a variation of the previous theorem which asserts that one can construct the resulting subsystem code such that it is again @xmath10-linear .
[ th : fqshrinkk ] let @xmath14 be a power of a prime @xmath15 . if there exists an @xmath10-linear @xmath88_q$ ] subsystem code with @xmath273 that is pure to @xmath86 , then there exists an @xmath10-linear @xmath274_q$ ] subsystem code that is pure to @xmath237 .
if a pure @xmath10-linear @xmath275_q$ ] subsystem code exists , then there exists an @xmath10-linear @xmath276_q$ ] subsystem code .
the proof is analogous to the proof of the previous theorem , except that @xmath10-bases are used instead of @xmath30-bases .
there exists a partial converse of theorem [ th : shrinkk ] , namely if the subsystem code is pure , then it is possible to increase the dimension of the subsystem and decrease the dimension of the co - subsystem while maintaining the same minimum distance .
[ th : shrinkr ] let @xmath14 be a power of a prime @xmath15 . if there exists a pure @xmath0 subsystem code with @xmath277 , then there exists a pure @xmath278 subsystem code .
suppose that the @xmath0 clifford subsystem code is associated with a classical additive code @xmath279 let @xmath280 .
we have @xmath281 , @xmath282 , hence @xmath283 and @xmath284 .
furthermore , @xmath79 .
the code @xmath285 has the subcode @xmath71 . since @xmath286 ,
the parameters of the clifford subsystem code associated with @xmath68 are @xmath287 .
since @xmath288 , the minimum distance @xmath86 satisfies @xmath289 on the other hand , @xmath290 , whence @xmath291 .
furthermore , the resulting code is pure since @xmath292 .
replacing @xmath30-bases by @xmath10-bases in the proof of the previous theorem yields the following variation of the previous theorem for @xmath10-linear subsystem codes .
[ th : fqshrinkr ] let @xmath14 be a power of a prime @xmath15 . if there exists a pure @xmath10-linear @xmath88_q$ ] subsystem code with @xmath293 , then there exists a pure @xmath10-linear @xmath294_q$ ] subsystem code .
the purity hypothesis in theorems [ th : shrinkr ] and [ th : fqshrinkr ] is essential , as the next remark shows .
the bacon - shor code is an impure @xmath295_2 $ ] subsystem code .
however , there does not exist any @xmath296_2 $ ] stabilizer code .
thus , in general one can not omit the purity assumption from theorems [ th : shrinkr ] and [ th : fqshrinkr ] .
an @xmath297_q$ ] stabilizer code can also be regarded as an @xmath298_q$ ] subsystem code .
we record this important special case of the previous theorems in the next corollary .
[ cor : generic ] if there exists an ( @xmath10-linear ) @xmath297_q$ ] stabilizer code that is pure to @xmath86 , then there exists for all @xmath90 in the range @xmath299 an ( @xmath10-linear ) @xmath300_q$ ] subsystem code that is pure to @xmath237 .
if a pure ( @xmath10-linear ) @xmath88_q$ ] subsystem code exists , then a pure ( @xmath10-linear ) @xmath301_q$ ] stabilizer code exists .
we have shown in @xcite that a ( primitive or non - primitive ) narrow sense bch code of length @xmath50 over @xmath10 contains its dual code if the designed distance @xmath217 is in the range @xmath302).\ ] ] for simplicity , we will proceed our work for primitive narrow sense bch codes , however , the generalization for non - primitive bch codes is a straightforward .
[ lem : bchexistfq ] if @xmath14 is power of a prime , @xmath245 is a positive integer , and @xmath303 $ ] .
then there exists a subsystem bch code with parameters @xmath304_q$ ] where @xmath305 . we know that if @xmath306 $ ] , then there exists a stabilizer code with parameters @xmath307_q$ ] .
let r be an integer in the range @xmath308 . from (
* theorem 2 ) , then there must exist a subsystem bch code with parameters @xmath309_q$ ] .
we can also construct subsystem bch codes from stabilizer codes using the hermitian constructions .
[ lem : bchexistfq2 ] if @xmath14 is a power of a prime , @xmath245 is a positive integer , and @xmath217 is an integer in the range @xmath310}-1 -(q^2 - 2)[m \textup { even}]$ ] , then there exists a subsystem code @xmath2 with parameters @xmath311_q\ ] ] that is pure up to @xmath217 , where @xmath312 . if @xmath313}-1 -(q^2 - 2)[m \textup { even}]$ ] , then exists a classical bch code with parameters @xmath314_q$ ] which contains its dual code . from
* theorem 2 ) , then there must exist a subsystem code with the given parameters .
recall that an @xmath88_q$ ] subsystem code derived from an @xmath10-linear classical code @xmath315 satisfies the singleton bound @xmath316 , see ( * ? ? ?
* theorem 3.6 ) .
a subsystem code attaining the singleton bound with equality is called an mds subsystem code .
an important consequence of the previous theorems is the following simple observation which yields an easy construction of subsystem codes that are optimal among the @xmath10-linear clifford subsystem codes .
[ th : puremds ] if there exists an @xmath10-linear @xmath297_q$ ] mds stabilizer code , then there exists a pure @xmath10-linear @xmath317_q$ ] mds subsystem code for all @xmath90 in the range @xmath318
. an mds stabilizer code must be pure , see ( * ? ? ?
* theorem 2 ) or ( * ? ? ?
* corollary 60 ) . by corollary [ cor : generic ] ,
a pure @xmath10-linear @xmath297_q$ ] stabilizer code implies the existence of an @xmath10-linear @xmath319_q$ ] subsystem code that is pure to @xmath9 for any @xmath90 in the range @xmath318 .
since the stabilizer code is mds , we have @xmath320 . by
the singleton bound , the parameters of the resulting @xmath10-linear @xmath321_q$ ] subsystem codes must satisfy @xmath322 , which shows that the minimum distance @xmath323 , as claimed .
we conjecture that @xmath10-linear mds subsystem codes are actually optimal among all subsystem codes , but a proof that the singleton bound holds for general subsystem codes remains elusive . in the next corollary
, we give a few examples of mds subsystem codes that can be obtained from theorem [ th : puremds ] .
these are the first families of mds subsystem codes ( though sporadic examples of mds subsystem codes have been established before , see e.g. @xcite ) .
a. an @xmath10-linear pure @xmath324_q$ ] mds subsystem code exists for all @xmath50 , @xmath9 , and @xmath90 such that @xmath325 , @xmath326 , and @xmath327 . b. an @xmath10-linear pure @xmath328_q$ ] mds subsystem code exists for all @xmath329 and @xmath90 such that @xmath330 and @xmath331 .
c. an @xmath10-linear pure @xmath332_q$ ] mds subsystem code exists for all @xmath217 and @xmath90 such that @xmath333 and @xmath334 .
d. an @xmath10-linear pure @xmath335_q$ ] mds subsystem code exists for all @xmath333 and @xmath336 .
e. an @xmath10-linear pure @xmath337_q$ ] mds subsystem code exists for all @xmath217 and @xmath90 in the range @xmath338 and @xmath339 . f. an @xmath10-linear pure @xmath340_q$ ] mds subsystem code exists for all @xmath217 and @xmath341 in the range @xmath338 and @xmath342 . by (
* theorem 14 ) , there exist @xmath10-linear @xmath343_q$ ] stabilizer codes for all @xmath50 and @xmath9 such that @xmath325 and @xmath326 .
the claim follows from theorem [ th : puremds ] .
+ by ( * ? ? ?
* theorem 5 ) , there exist a @xmath344_q$ ] stabilizer code . in this case ,
the code is derived from an @xmath11-linear code @xmath67 of length @xmath50 over @xmath11 such that @xmath345 .
the claim follows from lemma [ l : hermitian - linear ] and theorem [ th : puremds ] .
+ @xmath346 , there exist @xmath10-linear stabilizer codes with parameters @xmath347_q$ ] and @xmath348_q$ ] for @xmath333 , see ( * ? ? ?
* theorem 9 ) .
theorem [ th : puremds ] yields the claim .
+ @xmath346 , there exist @xmath10-linear stabilizer codes with parameters @xmath349_q$ ] and @xmath350_q$ ] .
for @xmath338 by ( * ? ? ?
* theorem 10 ) .
the claim follows from theorem [ th : puremds ] .
the existence of the codes in i ) are merely established by a non - constructive gilbert - varshamov type counting argument . however , the result is interesting , as it asserts that there exist for example @xmath351_q$ ] subsystem codes for all prime powers @xmath352 , @xmath353_q$ ] subsystem codes for all prime powers @xmath352 , and other short subsystem codes that one should compare with a @xmath354_q$ ] stabilizer code . if the syndrome calculation is simpler , then such subsystem codes could be of practical value . the subsystem codes given in ii)-vi ) of the previous corollary are constructively established . the subsystem codes in ii )
are derived from reed - muller codes , and in iii)-vi ) from reed - solomon codes .
there exists an overlap between the parameters given in ii ) and in iv ) , but we list here both , since each code construction has its own merits . by theorem
[ th : fqshrinkr ] , pure mds subsystem codes can always be derived from mds stabilizer codes , see table [ table : optimalmds ] .
therefore , one can derive in fact all possible parameter sets of pure mds subsystem codes with the help of theorem [ th : puremds ] . in the case of stabilizer codes
, all mds codes must be pure .
for subsystem codes this is not true , as the @xmath295_2 $ ] subsystem code shows .
finding such impure @xmath10-linear @xmath88_q$ ] mds subsystem codes with @xmath355 is a particularly interesting challenge . &
+ @xmath356_3 $ ] & @xmath357_{3 ^ 2}$ ] + @xmath358_3$]&@xmath359_{3 ^ 2}$ ] + @xmath360_3$]&@xmath361_{3 ^ 2}$ ] + @xmath362_3$]&@xmath363_{3 ^ 2}^{\dag } , \delta=3 $ ] + @xmath364_3$]&@xmath365_{3 ^ 2}^{\dag } , \delta=6 $ ] + @xmath366_4 $ ] & @xmath367_{4 ^ 2}$ ] + @xmath368_4$]&@xmath369_{4 ^ 2}$ ] + @xmath370_4$]&@xmath371_{4 ^ 2}$ ] + @xmath372_4 $ ] & @xmath373_{4 ^ 2}^{\dag},\delta= 4 $ ] + @xmath374_5 $ ] & @xmath375_{5 ^ 2}$ ] + @xmath376_5 $ ] & @xmath377_{5 ^ 2}$ ] + @xmath378_5 $ ] & @xmath379_{5 ^ 2}$ ] + @xmath380_5 $ ] & @xmath381_{5 ^ 2}$ ] + @xmath382_5 $ ] & @xmath383_{5 ^ 2}$ ] + @xmath384_5$]&@xmath385_{5 ^ 2}^ { * } , \delta=5 $ ] + @xmath386_5$]&@xmath387_{5 ^ 2}^ { * } , \delta=20 $ ] + @xmath388_7 $ ] & @xmath389_{7 ^ 2}$ ] + + * punctured code + @xmath390 extended code recall that a pure subsystem code is called perfect if and only if it attains the hamming bound with equality .
we conclude this section with the following consequence of theorem [ th : puremds ] : if there exists an @xmath10-linear pure @xmath297_q$ ] stabilizer code that is perfect , then there exists a pure @xmath10-linear @xmath317_q$ ] perfect subsystem code for all @xmath90 in the range @xmath391 .
in section [ sec : dimensions ] , we showed how one can derive new subsystem codes from known ones by modifying the dimension of the subsystem and co - subsystem . in this section ,
we derive new subsystem codes from known ones by extending and shortening the length of the code .
[ lemma_n+1k ] if there exists an @xmath0 clifford subsystem code with @xmath392 , then there exists an @xmath393 subsystem code that is pure to 1 .
we first note that for any additive subcode @xmath394 , we can define an additive code @xmath395 by @xmath396 we have @xmath397 . furthermore , if @xmath398 , then @xmath399 is contained in @xmath400 for all @xmath116 in @xmath10 , whence @xmath401 . by comparing cardinalities
we find that equality must hold ; in other words , we have @xmath402 by theorem [ th : oqecfq ] , there are two additive codes @xmath68 and @xmath70 associated with an @xmath0 clifford subsystem code such that @xmath403 and @xmath404 we can derive from the code @xmath68 two new additive codes of length @xmath405 over @xmath10 , namely @xmath406 and @xmath407 . the codes @xmath406 and @xmath408 determine a @xmath409 clifford subsystem code .
since @xmath410 we have @xmath411 .
furthermore , we have @xmath412 .
it follows from theorem [ th : oqecfq ] that @xmath413 , @xmath414 , @xmath415 .
since @xmath406 contains a vector @xmath416 of weight @xmath202 , the resulting subsystem code is pure to 1 .
if there exists an @xmath88_q$ ] subsystem code with @xmath417 and @xmath418 , then there exists an @xmath419_q$ ] subsystem code that is pure to 1
. we can also shorten the length of a subsystem code in a simple way as shown in the following theorem .
[ lem : n-1k+1rule ] if a pure @xmath0 subsystem code exists , then there exists a pure @xmath420 subsystem code . by ( * ?
* lemma 10 ) , the existence of a pure clifford subsystem code with parameters @xmath0 implies the existence of a pure @xmath421 stabilizer code .
it follows from ( * ? ? ?
* lemma 70 ) that there exist a pure @xmath422 stabilizer code , which can be regarded as a pure @xmath423 subsystem code .
thus , there exists a pure @xmath420 subsystem code by theorem [ th : shrinkr ] , which proves the claim . in bracket
notation , the previous theorem states that the existence of a pure @xmath88_q$ ] subsystem code implies the existence of a pure @xmath424_q$ ] subsystem code .
in this section , we show how one can obtain a new subsystem code by combining two given subsystem codes in various ways .
[ thm : twocodes_n1k1r1d1n2k2r2d2 ] if there exists a pure @xmath425_2 $ ] subsystem code and a pure @xmath426_2 $ ] subsystem code such that @xmath427 , then there exist subsystem codes with parameters @xmath428_2\ ] ] for all @xmath90 in the range @xmath429 , where the minimum distance @xmath430 .
since there exist pure @xmath425_2 $ ] and @xmath426_2 $ ] subsystem codes with @xmath427 , it follows from theorem [ th : shrinkr ] that there exist stabilizer codes with the parameters @xmath431_2 $ ] and @xmath432_2 $ ] such that @xmath427 .
therefore , there exists an @xmath433_2 $ ] stabilizer code with minimum distance @xmath434 by ( * ? ? ?
* theorem 8) .
it follows from theorem [ th : shrinkk ] that there exists @xmath435_2 $ ] subsystem codes for all @xmath90 in the range @xmath429 .
[ lem : twocodes_nk1r1s1k2r2d2 ] let @xmath436 and @xmath437 be two pure subsystem codes with parameters @xmath438_q$ ] and @xmath439_q$ ] , respectively .
if @xmath440 , then there exists pure subsystem codes with parameters @xmath441_q\ ] ] for all @xmath90 in the range @xmath442 , where the minimum distance @xmath443 . by assumption
, there exists a pure @xmath444_q$ ] subsystem code , which implies the existence of a pure @xmath445_q$ ] stabilizer code by theorem [ th : shrinkr ] , where @xmath446 . by (
* lemma 74 ) , there exists a pure stabilizer code with parameters @xmath447_q$ ] such that @xmath448 . by theorem
[ th : shrinkk ] , there exist a pure subsystem code with parameters @xmath449_q$ ] for all @xmath90 in the range @xmath442 , which proves the claim . further analysis of propagation rules of subsystem code constructions , tables of upper and lower bounds , and short subsystem codes are presented in @xcite .
subsystem codes are among the most versatile tools in quantum error - correction , since they allow one to combine the passive error - correction found in decoherence free subspaces and noiseless subsystems with the active error - control methods of quantum error - correcting codes . in this paper
we demonstrate several methods of subsystem code constructions over binary and nonbinary fields .
the subclass of clifford subsystem codes that was studied in this paper is of particular interest because of the close connection to classical error - correcting codes . as theorem
[ th : oqecfq ] shows , one can derive from each additive code over @xmath10 an clifford subsystem code .
this offers more flexibility than the slightly rigid framework of stabilizer codes .
we showed that any @xmath10-linear mds stabilizer code yields a series of pure @xmath10-linear mds subsystem codes .
these codes are known to be optimal among the @xmath10-linear clifford subsystem codes .
we conjecture that the singleton bound holds in general for subsystem codes .
there is quite some evidence for this fact , as pure clifford subsystem codes and @xmath10-linear clifford subsystem codes are known to obey this bound .
we have established a number of subsystem code constructions .
in particular , we have shown how one can derive subsystem codes from stabilizer codes . in combination with the propagation rules
that we have derived , one can easily create tables with the best known subsystem codes .
further propagation rules and examples of such tables are given in @xcite , and will appear in an expanded version of this paper .
this research was supported by nsf grant ccf-0622201 and nsf career award ccf-0347310 .
part of this paper is appeared in proceedings of 2008 ieee international symposium on information theory , isit08 , toronto , ca , july 2008 .
s. a. aly , a. klappenecker , and p. k. sarvepalli .
primitive quatnum bch codes over finite fields . in _ proc .
2006 ieee international symposium on information theory , seattle , usa _ , pages 1114 1118 , july 2006 .
let @xmath455 be a basis of @xmath456 .
then @xmath457 is an element @xmath458 of @xmath10 ; hence , @xmath459 . let @xmath460 it follows from this definition that @xmath461 and that @xmath462 . furthermore ,
if @xmath463 and @xmath464 are elements of @xmath67 with @xmath465 in @xmath19 , then @xmath466 on the right hand side , all terms but the last are in @xmath10 ; hence we must have @xmath467 , which shows that @xmath468 , whence @xmath451 .
expanding @xmath469 in the basis @xmath470 yields a code @xmath471 , and we must have equality by a dimension argument .
since the basis expansion is isometric , it follows that @xmath472 the @xmath10-linearity of @xmath68 is a direct consequence of the definition of @xmath68 . | subsystem codes protect quantum information by encoding it in a tensor factor of a subspace of the physical state space .
subsystem codes generalize all major quantum error protection schemes , and therefore are especially versatile .
this paper introduces numerous constructions of subsystem codes .
it is shown how one can derive subsystem codes from classical cyclic codes .
methods to trade the dimensions of subsystem and co - subystem are introduced that maintain or improve the minimum distance . as a consequence ,
many optimal subsystem codes are obtained .
furthermore , it is shown how given subsystem codes can be extended , shortened , or combined to yield new subsystem codes .
these subsystem code constructions are used to derive tables of upper and lower bounds on the subsystem code parameters . |
we thank the nci national facility and eresearch sa for generous grants of supercomputing time which have enabled this project .
this research is supported by the australian research council .
n. isgur , g. karl , phys .
b72 ( 1977 ) 109 . n. isgur , g. karl , phys .
d19 ( 1979 ) 2653 . z .-
burkert , z .- j .
li , phys .
d46 ( 1992 ) 7074 . c. e. carlson , n. c. mukhopadhyay , phys .
67 ( 1991 ) 37453748 .
p. a. m. guichon , phys .
b164 ( 1985 ) 361 .
o. krehl , c .
hanhart , s .
krewald , j. speth , phys .
c62 ( 2000 ) 025207 .
, d. b. leinweber , phys .
d51 ( 1995 ) 63836393 . . | we present results for the first positive parity excited state of the nucleon , namely , the roper resonance ( @xmath0=1440 mev ) from a variational analysis technique . the analysis is performed for pion masses as low as 224 mev in quenched qcd with the flic fermion action .
a wide variety of smeared - smeared correlation functions are used to construct correlation matrices .
this is done in order to find a suitable basis of operators for the variational analysis such that eigenstates of the qcd hamiltonian may be isolated .
a lower lying roper state is observed that approaches the physical roper state . to the best of our knowledge , the first time this state has been identified at light quark masses using a variational approach .
roper resonance , roper state , positive parity , excited state 11.15.ha , 12.38.gc , 12.38.-t , 13.75.gx one of the long - standing puzzles in hadron spectroscopy has been the low mass of the first positive parity , @xmath1 , excitation of the nucleon , known as the roper resonance @xmath2(1440 mev ) . in constituent or valence quark models with harmonic oscillator potentials
, the lowest - lying odd parity state naturally occurs below the @xmath3 state ( with principal quantum number @xmath4 ) @xcite whereas , in nature the roper resonance is almost 100 mev below the @xmath5(1535 mev ) state .
similar difficulties in the level orderings appear for the @xmath6 and @xmath7 resonances , which have led to the speculation that the roper resonance may be more appropriately viewed as a hybrid baryon state with explicitly excited gluon field configurations @xcite , or as a breathing mode of the ground state @xcite or states which can be described in terms of meson - baryon dynamics alone @xcite .
the first detailed analysis of the positive parity excitation of the nucleon was performed in ref .
@xcite using wilson fermions and an operator product expansion spectral ansatz .
since then several attempts have been made to address these issues in the lattice framework @xcite , but in many cases no potential identification of the roper state has been made @xcite .
recently however , in the analysis of @xcite , a low - lying roper state has been identified using bayesian techniques . + here
, we use a ` variational method ' @xcite , which is based on a correlation matrix analysis and has been used quite extensively in refs . @xcite .
though the ground state mass of the nucleon has been described successfully , an unambiguous determination of the roper state with this method has not been achieved in the past , though significant amounts of research have been carried out in ref .
@xcite , by the cssm lattice collaboration @xcite , the bgr @xcite collaboration and in refs .
@xcite .
+ in this letter , we present evidence of a low - lying roper state for the first time using a variational analysis .
the observed state displays chiral curvature and approaches the physical mass of the roper state .
the standard nucleon interpolating field @xmath8 is considered in this analysis .
various sweeps of gaussian smearing @xcite are used to construct a smeared - smeared correlation function basis from which we obtain the correlation matrices .
+ the two point correlation function matrix for @xmath9 can be written as @xmath10 where , dirac indices are implicit . here , @xmath11 and @xmath12 are the couplings of interpolators @xmath13 and @xmath14 at the sink and source , respectively .
@xmath15 enumerates the energy eigenstates with mass @xmath16 .
+ since the only @xmath17 dependence comes from the exponential term , one can seek a linear superposition of interpolators , @xmath18 , such that ( more detail can be found in refs .
@xcite ) , @xmath19 for sufficiently large @xmath17 and @xmath20 , see refs . @xcite and @xcite . multiplying the above equation by @xmath21^{-1}$ ] from the left leads to an eigenvalue equation , @xmath22_{ij}\ , u^{\alpha}_{j } & = c^{\alpha}\ , u^{\alpha}_{i } , \label{eq : right_evalue_eq}\end{aligned}\ ] ] where @xmath23 is the eigenvalue . similar to eq.([eq : right_evalue_eq ] ) , one can also solve the left eigenvalue equation to recover the @xmath24 eigenvector , @xmath25_{ij } & = c^{\alpha}v^{\alpha}_{j}. \label{eq : left_evalue_eq}\end{aligned}\ ] ] the vectors @xmath26 and @xmath27 diagonalize the correlation matrix at time @xmath17 and @xmath20 making the projected correlation matrix , @xmath28 the parity projected , eigenstate projected correlator , @xmath29 is then analyzed using standard techniques to obtain masses of different states .
our analysis is exploratory , seeking to develop techniques to access the roper state in lattice gauge theory .
our lattice ensemble consists of 200 quenched configurations with a lattice volume of @xmath30 .
gauge field configurations are generated by using the dbw2 gauge action @xcite and an @xmath31-improved flic fermion action @xcite is used to generate quark propagators .
this action has excellent scaling properties and provides near continuum results at finite lattice spacing @xcite .
the lattice spacing is @xmath32 fm , as determined by the static quark potential , with the scale set using the sommer scale , @xmath33 fm @xcite . in the irrelevant operators of the fermion action we apply four sweeps of stout - link smearing to the gauge links to reduce the coupling with the high frequency modes of the theory @xcite providing @xmath34 improvement @xcite .
we use the same method as in ref .
@xcite to determine fixed boundary effects , and the effects are significant only after time slice 25 in the present analysis . various sweeps ( 1 , 3 , 7 , 12 , 16 , 26 , 35 , 48 sweeps corresponding to rms radii , in lattice units , of 0.6897 , 1.0459 , 1.5831 , 2.0639 , 2.3792 , 3.0284 , 3.5237 , 4.1868 ) of gauge invariant gaussian smearing @xcite are applied symmetrically at the source ( at @xmath35 ) and at the sink .
this is to ensure a variety of overlaps of the interpolators with the lower - lying states .
the analysis is performed on ten different quark masses corresponding to pion masses @xmath36 gev .
error analysis is performed using a second - order single elimination jackknife method , where the @xmath37 is obtained via a covariance matrix analysis method .
our fitting method is discussed extensively in ref .
@xcite .
+ the nucleon interpolator we consider is the local scalar - diquark interpolator having a non - relativistic reduction @xcite ,
@xmath38 we consider several @xmath39 matrices .
each matrix is constructed with different sets of correlation functions , each set element corresponding to a different numbers of sweeps of gauge invariant gaussian smearing at the source and sink of the @xmath40 correlators .
this provides a large basis of operators with varieties of overlap among energy states .
@xmath41{thesis_ma_new_n.x1x1.3.12.26.35_4sts_q1.eps } & \includegraphics [ height=0.48\textwidth , angle=90]{thesis_ei_new_n.x1x1.3.12.26.35_4sts_q1.eps } \end{array}$ ] we consider seven combinations \{1=(1,7,16,35 ) , 2=(3,7,16,35 ) , 3=(1,12,26,48 ) , 4=(3,12,26,35 ) , 5=(3,12,26,48 ) , 6=(12,16,26,35 ) , 7=(7,16,35,48 ) } of @xmath39 matrices . in ref .
@xcite it was shown that one can not isolate a low - lying excited eigenstate using a single fixed - size source smearing .
the superposition of states manifested itself as a smearing dependence of the effective mass . in this letter
we exploit this sensitivity to isolate the energy eigenstates . in fig.[fig :
mass_and_eig_for_3.12.26.35_q1 ] , we show the mass from the projected correlation functions and from the eigenvalues ( as shown in ref .
@xcite ) for the fourth combination ( 3,12,26,35 ) of @xmath39 matrices .
we note that similar results in mass from the projected correlation functions and the eigenvalues are observed in this analysis as in ref . @xcite .
though the mass of the first excited state from projected correlation functions show little change with variational parameters , see fig.[fig : mass_and_eig_for_3.12.26.35_q1 ] , the second and third excited states start a little below which indicates @xmath17 and @xmath20 are not sufficiently large .
with larger euclidean times fewer states will contribute significantly to the correlators .
the robust aspect of fitting projected correlators is manifest in fig.[fig : mass_and_eig_for_3.12.26.35_q1 ] , and reflects the stability of the eigenvectors against changes in @xmath17 and @xmath20 .
in contrast , the mass from the eigenvalue analysis shows significant dependence on the variational parameters . the same method as described in ref .
@xcite is applied in this letter to extract the mass from the projected correlation functions . in fig.[fig : mass_for_all_combinations_q1 ] , masses extracted from all the combinations of @xmath39 matrices ( from @xmath42 to @xmath43 ) are shown for the pion mass of 797 mev . some dependence of the excited states on smearing count is also observed here as in ref .
@xcite for a few of the interpolator basis smearing sets .
however the ground and first excited states are robust against changes in the interpolator basis , providing evidence that an energy eigenstate has been isolated .
it should be noted that the highest excited state ( the third excited state ) is influenced more by the level of smearing than the lower excited states .
this is to be expected as this state must accommodate all remaining spectral strength .
the @xmath44 combination in fig.[fig : mass_for_all_combinations_q1 ] provides heavier excited states as this basis begins with a low number of smearing sweeps ( a sweep count of 1 ) and also contains another low smearing set of 7 sweeps .
the second and third excited states , and more importantly , the first excited state sits a bit high in comparison with the other bases .
hence , extracting masses with this basis is not as reliable as other sets .
the @xmath45 combination also contains elements with a small smearing sweep count ( 3 and 7 ) , hence this basis also provides heavier excited states and shows some systematic drift in the second excited state . however , this basis has reduced contamination from the excited states when compared with the first basis .
the @xmath46 combination also starts at the low smearing count , so the mass from this basis for the third excited state is a little high .
we can observe at this point that including basis elements with a low smearing count will increase the masses of excited states ( for instance , consecutive low numbers of smearing sweeps 1,7 and 3,7 , respectively ) .
this is because the correlation functions with these low sweep counts have a large overlap with several heavier excited states in their sub - leading exponential .
we also observe that the inclusion of basis elements with a high level of smearing ( for instance , a sweep count of 48 ) results in larger statistical errors in the analysis .
the @xmath47 combination starts from a moderate smearing sweep count ( 12 ) and also contains elements with consecutive smearing sweep counts which provides less diversity in the basis .
similarly , the @xmath48 combination contains consecutive large smearing sweep counts of 35 and 48 and operators for these levels of smearing are very similar challenging the isolation of single energy eigenstates .
the @xmath49 and @xmath50 combinations are well spread over the given range of smearing sweeps and start from a sweep count of 3 .
they do nt include successive lower smearing sweep counts .
the @xmath51 combination contains the basis element with a sweep count of 48 but has only slightly larger statistical errors than the @xmath49 basis choice .
both these bases provide diversity .
it is observed that both combinations provide consistent results for the states .
nonetheless , it should be noted that the @xmath46 to @xmath52 combinations all provide very consistent results for the lower three ( ground , first and second ) energy states , shown in fig.[fig : mass_for_all_combinations_q1 ] .
while the @xmath45 and the @xmath53 excited states display some evidence of eigenstate mixing , the ground and the @xmath42 excited states are robust in this analysis as they agree within one standard deviation for all the combinations of @xmath39 matrices .
hence , an analysis is performed to calculate the systematic errors associated with the choice of basis over the preferred four combinations ( from @xmath46 to @xmath52 ) with @xmath54 , where , @xmath55 is the number of bases , in this case is equal to 4 .
it should be noted that for the ground and first excited states the systematic errors associated with choice of basis are very small in comparison with their statistical errors ( see table [ table : table.mpi.mavg.err - sum - over - qudrature.4comb.4states ] ) . in fig.[fig
: paper_m.avg_4com.x1x1.4x4.sqrt_erravg_4combs_errbasis_4combs.4states.allq.new ] , the masses from projected correlation functions are shown for all pion masses averaged over four combinations of correlation matrices ( from @xmath46 to @xmath52 ) .
masses are averaged over these four bases and errors ( average statistical errors over these four bases and systematic errors associated with basis choices ) are combined in quadrature , @xmath56 .
eigenstate mixing may be affecting the results for the second and the third excited states .
the ground and the @xmath44 excited states systematically approach the physical mass of the nucleon and roper state .
this is the first time evidence of a low - lying roper state has appeared from variational analysis near the chiral regime . in fig.[fig :
paper_m.avg_4com.x1x1.4x4.sqrt_erravg_4combs_errbasis_4combs.2states.allq.new ] , the ground and the first excited states of fig.[fig : paper_m.avg_4com.x1x1.4x4.sqrt_erravg_4combs_errbasis_4combs.4states.allq.new ] are shown in larger scale for clarity .
the non - interacting two particle p - wave @xmath57 is shown by the dashed line .
it is interesting to note that the observed lattice roper state sits lower than the p - wave @xmath57 , indicative of attractive @xmath58 interactions producing a resonance at physical quark masses .
the very consistent euclidean - time fit window for all the bases and for all the quark masses , noticeably the lower time of the fit as shown in table [ table : mass_roperstates_4combs ] , assist to identify the robust nature of the extracted roper state which comes at the same euclidean - time from the heaviest to the lightest quarks masses .
there are quenched artifacts that will have significant influence on our results should we progress to lighter quark masses .
not only are @xmath59 couplings different in quenched qcd , but there are also contributions from unphysical , degenerate p - wave @xmath60 two - particle states @xcite .
future calculations approaching the decay threshold should be done in full dynamical - fermion qcd . in conclusion , through the use of a variety of smeared - smeared correlation functions in constructing the correlation matrix , the first positive parity excited state of the nucleon @xmath0 , the roper state , has been observed for the first time using the variational analysis . while the @xmath61 correlation matrix analysis of standard interpolators is insufficient to isolate these excited energy eigenstates of qcd @xcite , using the new correlation matrix construction with smeared - smeared correlators
enables us to extract the otherwise elusive roper state . |
the intracluster medium ( icm ) has long been thought to play a vital role in galaxy evolution in clusters ( gunn & gott 1972 ) .
ram pressure and turbulence / viscous stripping by the icm can efficiently remove cold galactic gas and may be responsible for transforming disk galaxies with active star formation into red s0 galaxies ( e.g. , quilis et al . 2000 ) .
stripping of the galactic ism in clusters has been extensively examined in simulations ( e.g. , schulz & struck 2001 , ss01 hereafter ; roediger & hensler 2005 , rh05 hereafter ) , which show that stripping has significant impacts on the properties of galaxies . in the process of stripping ,
the galactic ism is removed from the disk and the outer galactic disk is truncated , while the inner disc is compressed , accompanied by formation of numerous flocculent arms ( ss01 ; rh05 ) .
the galactic sfr is also modified .
compression of disk ism may trigger prodigious star formation ( e.g. , ss01 ; bekki & couch 2003 ; rh05 ) in the first @xmath6 yr of interaction , even though star formation activity will eventually be suppressed as the galactic ism is depleted .
this initial starburst may explain the blue `` butcher - oemler galaxies '' in @xmath70.3 clusters ( butcher & oemler 1978 ) .
the further evolution of the stripped ism from the disk is not clear because of uncertainties in the transport coefficients ( conductivity and viscosity ) .
it has been suggested that some stripped ism may be able to cool to form stars because the main ism heating source , stellar uv radiation , is much weaker in intracluster space ( e.g. , ss01 ) .
therefore , ism stripping in clusters is important for both galaxy evolution and star formation in clusters .
observational evidence of stripping in cluster late - type galaxies has only begun to accumulate in recent years .
tails behind late - type galaxies have been observed in hi , h@xmath0 and x - rays ( e.g. , gavazzi et al .
2001 ; wang et al . 2004
; oosterloo & van gorkom 2005 ; sun & vikhlinin 2005 ; yagi et al .
recently we found a long x - ray tail in a rich cluster a3627 , associated with a late - type galaxy ( sun et al .
2006 , s06 hereafter ) .
the narrowness and length of the tail make it the most dramatic x - ray tail of a late - type galaxy to date .
it is also the only known x - ray tail from a late - type galaxy in a rich cluster .
there are only two other known x - ray tails of late - type galaxies in clusters ( c153 in a2125 , wang et al .
2004 ; ugc 6697 in a1367 , sun & vikhlinin 2005 ) .
both are in @xmath8 3 kev clusters that are more distant than a3627 ( note @xmath9 for a2125 ) .
we have also examined all and data for @xmath103 kev clusters at [email protected] ( 62 clusters in total ; for the work on a subsample at [email protected] , see sun et al .
no other significant stripped tails of late - type galaxies have been detected , which implies that s current stage is of short duration .
a3627 is a nearby rich cluster ( @xmath12=0.01625 , @xmath13 = 925 km / s and @xmath14=6 kev ) rivaling coma and perseus in mass and galaxy content ( kraan - korteweg et al .
1996 ; woudt et al . 2007 ) . is a blue emission - line galaxy ( woudt et al .
2004 ) that is only @xmath8 200 kpc from the cluster s x - ray peak on the plane of the sky .
its radial velocity ( 4630 km s@xmath3 ) is close to that of a3627 ( 4871@xmath1554 km s@xmath3 , woudt et al .
s06 found a long narrow x - ray tail behind , in both and data ( also shown in fig .
the tail extends to at least 70 kpc from the galaxy with a length - to - width ratio of @xmath8 10 .
the x - ray tail is luminous ( @xmath8 10@xmath16 ergs s@xmath3 ) , with an x - ray gas mass of @xmath17 m@xmath5 .
s06 interpret the tail as the stripped ism of mixed with the hot icm , while this blue galaxy is being converted into a gas - poor galaxy .
the data also reveal three hard x - ray point sources ( @xmath18 ergs s@xmath3 ) along the tail ( fig .
1 ) , and the possibility of all of them being background agn is @xmath11 0.1% .
thus , we suggested that some of them may be ultra - luminous x - ray sources ( ulxs ) born from active star formation in the tail .
nevertheless , the optical properties of this interesting galaxy are poorly known in the literature ( e.g. , woudt et al .
2004 ; 2007 ) . in this paper
, we present the results from our h@xmath0 and optical broad - band imaging observations of with soar , and spectroscopic observations with the ctio 1.5 m telescope .
we adopt a cluster redshift of 0.01625 for a3627 ( woudt et al . 2007 ) , which is a little larger than what we used in s06 ( 0.0157 from kraan - korteweg et al .
assuming h@xmath19 = 73 km s@xmath3 mpc@xmath3 , @xmath20@xmath21=0.24 , and @xmath22=0.76 , the luminosity distance is 67.6 mpc , and 1@xmath23=0.317 kpc .
was observed with the 4.1 m southern observatory for astrophysical research ( soar ) telescope on cerro pachon on aug .
28 , 2006 ( ut ) and mar .
15 , 2007 ( ut ) .
both nights were clear and photometric .
the observations were made with the soar optical imager ( soi ) , which covers a 5.26@xmath24 square field of view with two ccds ( 2.6@xmath24 width each , separated by an 8.1@xmath23 gap ) . in the 2006 run ( @xmath25 seeing , 1.23 - 1.31 airmass ) ,
two 20-minute exposures were collected with the h@xmath0 filter and five 20-second exposures were taken in the @xmath26 band .
however , the telescope guider happened to be south of the galaxy in two h@xmath0 exposures , blocking any emission 26@xmath23 south of the galactic nucleus . in the 2007 run ( @xmath27 seeing , 1.17 - 1.29 airmass ) , we took four 20-minute exposures in the h@xmath0 band , two 10-minute exposures in another narrow band close to the h@xmath0 band ( we call it the h@xmath1 band hereafter ) , seven 90-second exposures in the @xmath28 band and four 30-second exposures in the @xmath26 band .
the h@xmath0 filter used is the ctio filter i d 6649 - 76 ( @xmath29 = 6650 , fwhm=77 ) as the redshifted h@xmath0 line of is centered at 6664 . the h@xmath1 filter used is the ctio filter i d 6520 - 76 ( @xmath29 = 6530 , fwhm=71 ) .
the bandwidths of the two narrow - band filters overlap little ( at 6590 , the transmission of the 6649 - 76 filter is 4.8% , while that of the 6520 - 76 filter is 2.6% ) .
both narrow - band filters are 2 inches @xmath30 2 inches , so the non - vignetted field is only @xmath31 .
dithers of 15@xmath32 were made between exposures . during the second half of 2006 ,
the triplet lens of the focal reducer in soi became partially unbounded , which caused the photometry to be uncertain to @xmath8 0.2 mag from the field center to the corners .
the lens of soi was repaired in the beginning of 2007 so the photometry accuracy was recovered before our 2007 run . in this work ,
we use only the 2007 data in the analysis .
nevertheless , we examined the emission - line objects ( see the definition and discussion in @xmath333 ) in the 2006 data ( which had better seeing ) .
all 29 emission - line objects selected from the 2007 data ( @xmath333 ) are confirmed in the 2006 data , although several of them are close to the guider in the 2006 data so their fluxes are reduced .
each image was reduced using the standard procedures with the iraf mscred package .
the pixels were binned 2@xmath34 , for a scale of 0.154@xmath23 per pixel .
the fringing on the @xmath26 band images was subtracted with the standard fringe frames .
dome flats were used .
the us naval observatory ( usno ) a2.0 catalog was used for the wcs alignment , which was done with wcstools .
spectrophotometric standards are ltt 4364 and eg 274 ( hamuy et al .
ten standard stars in two fields ( pg 1047 and ru 149 ) on the @xmath35 and @xmath26 bands were also observed at airmass of 1.1 - 1.8 .
is near the galactic plane ( @xmath36=325.25 deg , @xmath37=-6.97 deg ) so the galactic extinction is substantial .
the galactic extinction at the @xmath28 , @xmath26 , h@xmath0 and h@xmath1 bands is 0.89 mag , 0.40 mag , 0.554 mag and 0.568 mag respectively , derived from the galactic extinction law by cardelli , clayton & mathis ( 1989 ) for e(b - v ) = 0.207 mag ( from ned ) .
as is not far from the galactic center , the uncertainties of galactic extinction may not be small , but our conclusions of this paper are hardly affected .
spectroscopic observations of were conducted with the r - c spectrograph on the 1.5 m telescope in ctio , operated by the small and moderate aperture research telescope system ( smarts ) consortium .
two service observations were made on june 20 and 21 , 2007 , with the 47/ib and 26/ia grating setups respectively . the 26/ia setup covers a wavelength range of 3660 - 5440 with a dispersion of 1.5 per pixel .
the spectral resolution is @xmath8 4.3 .
the 47/ib setup covers a wavelength range of 5652 - 6972 with a dispersion of 1.1 per pixel .
the spectral resolution is @xmath8 3.1 .
the slit is always aligned east - west with a length of @xmath38 and a width of 3@xmath23 .
the pixel scale on the ccd is 1.3@xmath23 per pixel and the best instrumental spatial seeing for the r - c spectrograph is only @xmath39 . for the 26/ia observation ( 1.22 - 1.37 airmass ) ,
the slit was set across s center .
five exposures ( 20 minutes each ) were made in the 26/ia run and we combined them .
the 47/ib run on june 20 , 2007 ( 1.60 - 1.91 airmass ) was composed of three exposures ( 25 minutes each ) .
the slit was moved by several arcsec in the north - south direction between exposures by mistake , although it was requested in a fixed position the same as the one for the 26/ia observations .
thus , we analyzed the three exposures separately .
the pattern of the stellar spectra on the ccd ( through the long slit ) allows us to recover the slit positions .
spectroscopic standards were observed at each night , ltt 4364 for the 26/ia run and feige 110 for the 47/ib run .
as these observations of the standards used a 2@xmath23 slit , absolute flux calibration is impossible , but the analysis of the emission line ratios should not be affected .
with data in two narrow bands ( h@xmath0 and h@xmath1 ) and two broad bands ( @xmath28 and @xmath26 ) , we derived color - color and color - magnitude relations for sources in the field covered by all data ( 3.42@xmath40 around shown as the dotted box in fig .
sextractor 2.5.0 was used to measure photometry and color .
the h@xmath0 and h@xmath1 magnitudes are ab magnitudes defined in hamuy et al .
( 1994 ) and are calibrated with the data of eg 274 and ltt 4364 .
galactic extinction has been corrected in all bands . in the h@xmath0 - h@xmath1 vs. h@xmath41 map ( fig .
2 ) , over 97% of 1708 sources in the field are galactic stars with little excess or decrement emission in the h@xmath0 filter .
the h@xmath0 - h@xmath1 histogram peaks at @xmath8 -0.02 mag and the typical 1 @xmath42 dispersion is @xmath8 0.15 mag .
the h@xmath41 histogram peaks at @xmath8 0.46 mag and is skewed to higher color values .
the 1 @xmath42 dispersion on the lower color side is @xmath8 0.22 mag . besides these stars , there are about 30 sources that are especially bright in the h@xmath0 band and occupy a different region in the color - color map . in this work
, we selected 29 emission - line objects with the following conservative criteria : h@xmath0 - h@xmath43 mag and h@xmath44 mag .
these emission - line objects have equivalent widths ( ew ) of 84 - 758 (we adopted the definition of ew in gavazzi et al .
2006 ) . their positions , as shown in fig .
1b , are all downstream of the galaxy and in or around the x - ray tail .
if we relax the criteria to h@xmath0 - h@xmath45 mag and h@xmath46 mag , six additional sources are selected and they are also in or around the x - ray tail ( fig .
1b ) . the h@xmath0 mag vs. h@xmath41 relation is also shown in fig .
2 . after correcting the atmospheric and galactic extinction , the fluxes of these emission - line objects range from 1.5@xmath47 ergs s@xmath3 @xmath48 ( no correction for intrinsic extinction ) .
we also derived the @xmath49 color for sources in the field .
as some emission - line objects are not detected in the @xmath28 and @xmath26 bands , @xmath49 color can be constrained only for 17 emission - line objects ( fig .
the emission - line objects are generally bluer than the galaxy ( without correction for intrinsic extinction ) and galactic stars .
the @xmath49 colors of galactic stars span from m type ( 3.6 - 5.1 mag ) to a type ( 0 - 0.4 mag ) . as the galactic reddening ( 0.49 mag )
was applied on all sources , the actual separation between @xmath49 colors of emission - line objects and galactic stars will be wider if zero or smaller galactic extinction applies on stars and some amount of intrinsic extinction is applied for emission - line objects . therefore , these emission - line objects are blue and a small amount of intrinsic extinction can drive their colors to those of pure o / b star clusters . what is the nature of these emission - line objects
? they can not be galactic h@xmath0 emitters as the required velocity ( 2600 km / s - 5300 km / s ) is too high .
the other emission line close to the h@xmath0 filter band is [ sii ] 6716 ( but generally with small ew ) .
however , a velocity of @xmath11 -1600 km / s is required to put the [ sii ] line in the h@xmath0 filter band , which is also too large for galactic objects .
thus , these emission - line objects are extragalactic .
besides being h@xmath0 emitters in a3627 , these objects could also be [ oiii ] @xmath505007 emitters at @xmath51 , h@xmath52 emitters at @xmath53 , [ oii ] @xmath503727 emitters at @xmath54 and ly@xmath0 emitters at @xmath55 .
we have examined the expected number density of these background objects in the following two ways .
first , we can examine the h@xmath1 frame in the same way as we did for the h@xmath0 frame to select emission - line objects in the h@xmath1 band ( or sources with large h@xmath0 - h@xmath1 color ) .
both narrow - band filters have very similar transmission and band width .
the difference in their central wavelength is small .
thus , we expect that the number of background emission - line objects detected in the two narrow bands should be similar .
h@xmath1 - h@xmath0 vs. h@xmath56 and h@xmath1 vs. h@xmath56 maps ( shown in fig .
2 ) are derived for sources detected in the h@xmath1 band .
the same criteria used to select emission - line objects in the h@xmath0 band were applied ( with just the change from h@xmath0 magnitude to h@xmath1 magnitude ) .
no emission - line objects have been found in the h@xmath1 band , even though the loose criteria are applied ( h@xmath1 - h@xmath57 -0.5 mag and h@xmath58 0.25 mag ) .
in fact , the h@xmath1 exposure is deep enough to detect the faintest emission - line object in the h@xmath0 band , if a similar object is present in the h@xmath1 frame .
this single self - test already demonstrates that the expected number of background emission - line objects is very small .
second , we examined the number density of emission - line objects from the large area ly@xmath0 survey at @xmath59 ( malhotra & rhoads 2002 ) as they used the narrow - band filters with similar central wavelength and bandwidth ( @xmath8 80 ) to ours .
their covered field is 0.72 deg@xmath60 and the achieved sensitivity is about 5 times deeper than ours . in three non - overlapping h@xmath0 filters ( the central one has the same central wavelength as our h@xmath0 filter )
, 157 ly@xmath0 candidates ( with ew @xmath61 and without @xmath28 band detection ) were detected .
the size of our field is 3.42@xmath40 .
thus , the expected number of emission - line objects selected in malhotra & rhoads ( 2002 ) is @xmath8 0.02 in our field ( adjusted to our flux limit by assuming a power index of -1.6 for the luminosity function , malhotra & rhoads 2002 ) .
cortese et al . ( 2004 ) examined @xmath8 2.5 deg@xmath60 h@xmath0 frames of the virgo cluster ( @xmath62 , @xmath63 = 95 ) obtained at the isaac newton telescope .
six emission - line objects were detected ( for a h@xmath0 flux limit of about 6 times shallower than the faintest emission - line object in our sample ) and their follow - up spectroscopic observations showed that five of them were [ oiii ] @xmath505007 emitters at @xmath64 . if we take their number as the number density of @xmath51 [ oiii ] @xmath505007 emitters in the high luminosity and apply a luminosity function with a power index of -1.4 ( e.g. , pascual et al . 2001 ) , the expected number of [ oiii ] emitters in our whole field is @xmath8 0.09 . for our line flux limit ( @xmath65 ergs s@xmath3 @xmath48 ) , the combined number density of the h@xmath52 and the [ oii ] @xmath503727 emitters is at most comparable to that of [ oiii ] @xmath505007 emitters ( e.g. , pascual et al .
2001 ) , and these two lines generally do nt have high ew . therefore , the expected number of background emission - line sources is much smaller than one in our 3.42@xmath40 field and it is very likely that _ all 29 emission - line objects are h@xmath0 emitters associated with a3627_. in fact , this conclusion is even much stronger when we consider the spatial concentration of emission - line objects immediately downstream of and around the x - ray tail . _ these 29 + 6 sources are not randomly distributed in the field and relative to the galaxy . _ twenty - five emission - line objects cluster in a 40@xmath66 box region immediately downstream of , while the area of our total field is 20 times larger . in @xmath334.1
, we further show that seven emission - line objects are caught in one or more ctio 1.5 m slit spectra .
the spectra strongly support the argument of hii regions ( @xmath334.1 ) , at least for these seven emission - line objects . as h@xmath0 emitters in a3627
, they are too luminous to be intracluster planetary nebulae .
the h@xmath0 luminosity of the faintest emission - line object in our sample is @xmath67 7.2 times larger ( without intrinsic extinction ) than the [ oiii ] @xmath505007 luminosity of the most luminous intracluster planetary nebulae in the virgo cluster ( ciardullo et al .
1998 ) and the h@xmath0 line of a planetary nebula is generally several times fainter than the [ oiii ] line .
thus , we conclude that they are hii regions most likely associated with and its tail in a3627 .
hereafter , we simply refer these emission - line objects as hii regions .
the properties of these hii regions and the embedded star clusters can be estimated .
the intrinsic extinction is unknown .
gerhard et al . ( 2002 ) and cortese et al . ( 2004 ) measured @xmath8 1 mag intrinsic extinction for two isolated hii regions in the virgo cluster .
we simply adopt this value for all hii regions .
the [ nii ] @xmath506548 and @xmath506584 lines are in the h@xmath0 filter band .
the [ nii ] @xmath506584 line is close to the wing ( transmission there is 65% of the value in the center ) .
gerhard et al .
( 2002 ) and cortese et al .
( 2004 ) measured h@xmath0 / ( h@xmath0 + [ nii ] @xmath506548 + [ nii ] @xmath506584 ) @xmath8 0.81 for two isolated hii regions in the virgo cluster .
this fraction is assumed in our analysis .
the derived h@xmath0 luminosities of these hii regions are plotted in fig .
3 versus their distance to s nucleus .
the cumulative luminosity function of these hii regions , n ( @xmath67 log @xmath68 , has a slope of [email protected] at @xmath69 ergs s@xmath3 . with the scaling relation derived by kennicutt ( 1998 ) , sfr ( m@xmath5/yr ) = @xmath70 ergs s@xmath3 ) ,
the sfr in these hii regions ranges from 0.0008 to 0.17 m@xmath5/yr , with a total sfr of 0.59 m@xmath5/yr for 29 hii regions selected .
assuming an electron temperature of 10@xmath71 k and case b of nebular theory , the number of ionizing photons @xmath72(h ) ranges from 9.5@xmath73 to 2.1@xmath74 s@xmath3 .
we can apply the starburst99 model ( leitherer et al .
1999 ) to estimate the age and the total mass of the starbursts in these hii regions .
the age of the starburst can be estimated from the ratio @xmath72(h ) / @xmath75 , or the ratio @xmath72(h ) / @xmath76 ( e.g. , gerhard et al .
2002 ) , or the h@xmath0 equivalent width ew ( h@xmath0 ) .
the first two estimates may only provide lower limits on the age as the intrinsic extinction of the hii regions in the @xmath28 and @xmath26 bands is unknown .
ew ( h@xmath0 ) is not affected by intrinsic extinction , although the uncertainties are generally large and there are only lower limits in many cases .
we selected eight representative hii regions ( marked elo1 to 8 in fig .
1c and 1d ) to demonstrate the range of the properties of these 29 objects .
elo1 is the brightest and the only resolved one ( with a radius of @xmath8 0.6 kpc ) .
elo2 and 3 are the next two most luminous ones .
elo4 is the one with the lowest ew ( h@xmath0 ) .
elo5 is only @xmath77 offset from the brightest hard x - ray point source in the tail ( p1 in s06 ) .
elo6 is the faintest emission - line object , while elo7 and 8 are most distant from s nucleus in projection ( 30 - 39 kpc ) .
a metallicity of 0.4 solar is assumed , which is similar to those of two hii regions studied by gerhard et al .
( 2002 ) and cortese et al .
we assume instantaneous star formation with a salpeter imf , @xmath78 and @xmath79 .
the results are listed in table 1 .
the age estimated from @xmath72(h ) / @xmath75 is always the smallest , which implies intrinsic extinction in hii regions .
the age is always less than @xmath8 7 myr , which is typical for bright hii regions ( e.g. , gerhard et al .
2002 ; mendes de oliveira et al . 2004 ) .
the estimated age from ew ( h@xmath0 ) is not sensitive to the assumed metallicity for ew ( h@xmath0 ) @xmath67 400 hii regions . even for low ew ( h@xmath0 ) regions ,
the uncertainty of age is not very big .
for example , the predicted age of elo4 from ew ( h@xmath0 ) is 6.7 , 6.3 , 6.5 , 8.3 , and 11.5 myr for assumed metallicity of 2 , 1 , 0.4 , 0.2 , 0.05 solar respectively .
it is clear that these hii regions are all young ( @xmath80 10 myr ) . the derived total mass ( @xmath81 m@xmath5 @xmath82 m@xmath5 ) is sensitive to the age of the starburst , as shown for two hii regions with similar h@xmath0 luminosities ( elo6 and 8 , table 1 ) .
the older the hii region is , the more massive it is . nevertheless , the brightest 5 - 10 hii regions host star clusters with total mass of @xmath83 m@xmath5 , or super star clusters ( 10@xmath84 - 10@xmath85 m@xmath5 , oconnell 2004 ) .
the total mass in the 29 putative hii regions is about 10@xmath4 m@xmath5 , with at least a factor of four uncertainty because of the unknown extinction and uncertain age .
the h@xmath0 emission of hii regions fades rapidly for a single starburst and will be much fainter after the initial 10 myr .
the radial velocity difference between and a3627 is only 214 km / s , which implies that s infalling is almost in the plane of the sky .
if we assume a velocity of equal to the velocity dispersion of the cluster ( 1600 km / s ) , travels only @xmath8 16 kpc in 10 myr .
it is clear that there are no very luminous hii regions beyond 15 kpc from the nucleus ( fig .
we also examined the h@xmath0 ew of the hii regions with offset from the galaxy .
there is no clear relation found as 13 sources have only lower limits . within 8 kpc from the nucleus , the h@xmath0 ew span a big range from 77 to 614 . in any case , the three hii regions at 29 - 39 kpc projected from the nucleus must have had star formation happening in the last @xmath11 10 myr , at least 10 - 15 myr after the gas was removed from .
thus , the star formation in these hii regions does not always start immediately after they are displaced from the galactic disk , if star formation in these hii regions is indeed a single burst .
it is likely that more hii regions downstream of are yet to form and many hii regions may have already faded in or around the tail .
we discuss the population of the hii regions more in @xmath336 .
the spectroscopic properties of the central emission nebula of are examined with the ctio 1.5 m spectra .
the slit positions are recovered from the stellar spectra on slits , and they are shown in fig .
the # 1 slit of the 47/ib observations misses the central emission nebula of the galaxy .
the stellar spectrum at this offset position is too faint to be studied with this single exposure , while a nearby star ( close to source `` a '' in fig .
4 , see also fig .
1 ) is several times brighter .
the # 3 slit of the 47/ib observations covers the h@xmath0 peak of and the resulting spectrum is the brightest .
the spectra of s central part are shown in fig . 5 , from the combined 26/ia exposures and the # 3 47/ib exposure .
the measured velocity with rvsao is 4667@xmath15135 km / s from the 26/ia spectrum and 4640@xmath1520 km / s from the # 3 47/ib spectrum , which is consistent with 4630@xmath1558 km / s measured by woudt , kraan - korteweg & fairall ( 1999 ) . as shown in fig . 5 , many line ratios can be determined .
however , the 26/ia and 47/ib spectra were taken at different nights and at different positions .
it is also impossible to do absolute flux calibration for these spectra .
therefore , we restrict our line ratio analysis only to lines in the same wavelength ranges ( 26/ia or 47/ib ) , which makes the usual method to constrain the intrinsic reddening with the h@xmath0 / h@xmath52 ratio impossible to be applied . instead , we use the h@xmath86 / h@xmath52 ratio to constrain the intrinsic reddening .
the theoretical value for pure recombination is taken from osterbrock ( 1989 ) , 0.468 for 10@xmath71 k gas ( case b ) at @xmath87 @xmath88 .
the line ratio measured is 0.35 , after the correction for the galactic reddening . using the galactic extinction curve from cardelli et al .
( 1989 ) and assuming @xmath89 = 3.1 , we derive @xmath90 = 1.7 mag for the intrinsic extinction along s core .
the corresponding intrinsic extinction for the h@xmath0 line is 1.4 mag , which is consistent with the typical extinction of the h@xmath0 line for nearby spirals ( 0.5 - 1.8 mag , e.g. , kennicutt 1983 ) and the likely near edge - on orientation of the galaxy ( see @xmath335 ) .
therefore , this amount of intrinsic extinction has been applied to the line ratio analysis in this paper .
the gas properties can be determined with the emission line ratios .
we derived various line ratios from the combined 26/ia spectrum and the # 3 47/ib spectrum as their emission comes from similar regions of the galaxy ( fig .
we caution that the flux calibration at the blue end of the spectrum is more vulnerable to the flat field correction so the error of the [ oii ] flux is bigger than those of other strong emission lines .
the [ sii ] @xmath506716 / [ sii ] @xmath506731 ratio is @xmath8 1.38 , which is typical for hii regions and is comparable to the low - density limit of 1.35 ( van zee et al . 1998 ) . the strong [ oii ] line compared to the [ oiii ] lines suggests a low ionization parameter of the gas .
we derived : log([oi ] @xmath506300 / h@xmath91 -1.65 , log([oiii ] @xmath505007 / [ oii ] ) = -1.39 , log([nii ] @xmath506584 / h@xmath0 ) = -0.42 , log([sii ] @xmath926716 , 6731 / h@xmath0 ) = 0.39 and log([oiii ] @xmath505007 / h@xmath52 ) = -0.51 .
all these ratios indicate that the central emission nebula of resembles a typical giant hii region ( fig . 4 and 5 of kewley et al .
the low [ oi ] @xmath506300 / h@xmath0 ratio implies that any central agn , if existed , is very weak , which is consistent with the x - ray non - detection of the nucleus ( a 3@xmath42 limit of 5@xmath93 ergs s@xmath3 in the 0.5 - 10 kev band ) .
the gas metallicity can be estimated from several emission line ratios ( e.g. , kewley & dopita 2002 ) , but these line ratios also depend on the ionization parameter . the ionization parameter can be estimated from the [ oiii ] @xmath505007 / [ oii ] ratio , 0.041 , which implies an ionization parameter ( @xmath94 ) of @xmath95 cm / s for gas metallicity of @xmath11 2.0 solar ( kewley & dopita 2002 ) .
we then estimate the gas metallicity from the following line ratios for @xmath96 cm / s ( kewley & dopita 2002 ) : 1 .
log([nii ] @xmath506584 / [ sii ] @xmath926716 , 6731 ) = 0.015 , which implies log(o / h)+12 = 8.75 - 8.95 . 2 .
log([oii ] @xmath503727 + [ oiii ] @xmath924959 , 5007 / h@xmath52 ) ( or r@xmath97 ) = 0.90 , which implies log(o / h)+12 = 8.05 - 8.20 or @xmath8 8.65 .
log([nii ] @xmath506584 / h@xmath0 ) = -0.44 , which implies log(o / h)+12 @xmath8 8.73 or 9.35 . 4 .
log([nii ] @xmath506584 / [ oii ] ) = -0.84 , which implies log(o / h)+12 = 8.62 - 8.75 .
as the line fluxes of [ nii ] and [ oii ] come from different observations , their ratio is estimated by multiplying these three line ratios , [ nii ] @xmath506584 / h@xmath0 , h@xmath0 / h@xmath52 and h@xmath52 / [ oii ] . for the h@xmath0 / h@xmath52 ratio ,
the theoretical value of 2.86 is used for case b recombination at 10@xmath71 k and @xmath98 @xmath88 ( osterbrock 1989 ) . in spite of uncertainties ,
we find that log(o / h)+12 @xmath99 8.7 ( 0.6 solar ) is consistent with all estimates . as shown in fig . 4 , seven emission - line objects are also in one or more slit spectra . a large part of elo1 s emission is in the 26/ia spectra
however , the best instrumental spatial seeing for the spectrograph is @xmath39 ( or 2.7 pixels ) . since both elo1 and the h@xmath0 core of are extended ( fig . 1 ) and
their peaks are only 4.6@xmath23 away , their lines will be blended in the spectra .
nevertheless , we indeed observe extension of the h@xmath52 , [ oii ] @xmath503727 and [ oiii ] @xmath505007 emission towards the direction of elo1 ( or upward on the ccd plane , fig .
the scale of the extension is consistent with the position of elo1 .
the other six emission - line objects only appear as a single line in each spectrum , because their continua are very weak .
although a single line detection does not determine their redshifts , the following arguments support that at least the bright ones among them are hii regions .
first , if we stack all three 2d spectra shown in fig .
4 , we detect significant [ nii ] @xmath926548 , 6584 emission at @xmath100 .
second , the centroids of these lines are only @xmath11 3 ( or 137 km / s ) higher than the h@xmath0 line of the galaxy .
third , if the detected line is [ oiii ] @xmath505007 or h@xmath52 at higher redshifts , we expect significant h@xmath52 or [ oiii ] 5007 line also in the 47/ib spectra from the usual [ oiii ] 5007 / h@xmath52 ratio ( kewley et al .
however , they are not detected in either a single exposure or the stacked exposure .
good quality spectra of these emission - line objects would require a big telescope .
the photometric properties of are also studied . as shown in fig .
1 , the galaxy may have a distorted inner disk or bar within 1 kpc radius .
a dust lane is also significantly detected in all four bands at 0.5 kpc to the south of the nucleus .
however , detail around the nucleus can only be obtained with imaging .
the morphological type of ( e.g. , from sb to sd ) is also unclear .
downstream in the @xmath28 and @xmath26 band images , some substructures are detected to @xmath8 6 kpc from the nucleus , including two blue streams ( fig .
interestingly , these structures all have nearby emission - line objects .
arm - like features are also detected at the north ( @xmath101 from the nucleus ) and the south ( @xmath102 from the nucleus ) .
to quantitatively examine the optical light distribution of , we measured surface brightness profiles along the minor axis and the major axis ( fig .
although we estimated the intrinsic reddening around the galactic center in @xmath334.1 , the intrinsic reddening outside the h@xmath0 emission nebula is unknown and may be smaller .
thus , we elect not to correct for the intrinsic extinction for these profiles .
the h@xmath0 emission is very asymmetric along the minor axis because of the hii regions and the h@xmath0 tail downstream of the galaxy .
the net h@xmath0 emission is truncated sharply upstream at @xmath103 kpc from the nucleus .
the h@xmath0 emission is much more symmetric along the major axis , and the net h@xmath0 emission is truncated sharply at @xmath104 kpc from the nucleus . in the optical , the galaxy is composed of at least two components ( fig .
the bright inner component extends to @xmath8 1.9 kpc in radius along the major axis and to @xmath8 1.2 kpc in radius along the minor axis .
almost all h@xmath0 emission is within the inner component . within the central 1.5 kpc radius ,
the @xmath28 and @xmath26 band light distributions have multiple peaks .
the outer component can be fitted with an exponential profile and the derived scale height is @xmath105 along the minor axis and @xmath106 along the major axis ( see the caption of fig .
the light distribution is a little more extended in the north , compared to that in the south ( fig .
we also measured the light profiles in elliptical annuli centered on the h@xmath0 peak ( see fig .
again , the light profiles at @xmath107 ( @xmath108 : the semi - major axis ) can be fitted with an exponential profile with a scale height of @xmath109 , while fits with the de vaucouleurs 1/4 law overestimate the emission at @xmath110 .
although the measured light profiles ( without correction on the intrinsic reddening ) imply little change of the @xmath49 color with @xmath108 , the known intrinsic reddening in the central h@xmath0 nebula implies that the @xmath49 color is 1.2 mag bluer there .
this blue and bright core may be a central bulge in formation .
the orientation of s disk can be estimated from the classical hubble formula .
assuming an axis ratio of @xmath8 2 and a morphological type from sb to sd , the estimated angle between the line of sight and the disk plane is 26 - 29 deg .
thus , the putative disk is viewed close to edge - on .
we also derived the total @xmath28 and @xmath26 band magnitudes of with the light profiles derived in elliptical annuli ( fig .
the total @xmath28 band magnitude measured within a 20@xmath111 ellipse is 14.31 @xmath15 0.08 mag ( or @xmath8 2/3 @xmath112 in the @xmath28 band ) , after correcting the atmospheric and galactic extinction .
this value can be compared with the @xmath28 magnitude listed on hyperleda , 14.05 @xmath15 0.23 mag from the eso survey .
our images are deeper and allow much better masking of galactic stars as some of them are very close to the nucleus ( fig .
the total @xmath26 band magnitude measured in the same aperture is 13.20 @xmath15 0.07 mag , after correcting the atmospheric and galactic extinction . the half - light size of the galaxy is also estimated from the light profiles measured in elliptical annuli : 14.5@xmath23 ( or 4.45 kpc ) semi - major axis in the @xmath28 band and 14.0@xmath23 ( or 4.30 kpc ) semi - major axis in the @xmath26 band .
the 2mass @xmath113 total magnitude is 12.163 mag , after correction for the galactic extinction ( 0.076 mag ) . the intrinsic reddening measured for the h@xmath0 emission nebula would imply a corrected @xmath114 color of @xmath8 0 , although the real averaged value should be higher as the regions outside of the h@xmath0 emission nebula may have smaller intrinsic extinction . in any case , the @xmath114 color of is much bluer than that of a typical late - type galaxy ( @xmath8 2 - 3.5 , jarrett 2000 ) .
the galaxy is way off the red - sequence of galaxies in clusters and is in the so - called `` blue cloud '' in the cluster color - magnitude relation .
the @xmath113 band absolute magnitude is -21.97 mag .
bell et al .
( 2003 ) determined the local @xmath113 band luminosity function and found an absolute magnitude of -23.97 mag for an @xmath112 galaxy ( with early - type and late - type galaxies mixed , adjusted to the cosmology we used ) .
therefore , is a 0.16 @xmath112 galaxy in the @xmath113 band .
the total stellar mass of can also be estimated .
the mass - to - light ratios in the @xmath26 and @xmath113 bands can be estimated from the relations derived in bell & de jong ( 2001 ) : @xmath115 = 0.59 @xmath116 , @xmath117 = 0.35 @xmath118 ( using the function for the @xmath119 band ) for @xmath120 = 1.1 mag .
the resulting stellar mass of is 4.9@xmath121 m@xmath5 from @xmath76 and 4.6@xmath121 m@xmath5 from @xmath122 .
the total stellar mass will be smaller if a correction for the intrinsic reddening is made .
for example , for an intrinsic reddening of 0.5 mag for the @xmath120 color , the total stellar mass of will be one third smaller .
bell et al .
( 2003 ) also determined the @xmath123 of the local stellar mass function , @xmath124 m@xmath5 . therefore , though is not a small galaxy , it is blue and has a low stellar mass ( @xmath125 ) .
the net h@xmath0 emission of was also studied .
the continuum emission in the h@xmath0 image is determined from the h@xmath1 frame and the scaling factor was determined to account for all emission of stars in the h@xmath0 image ( or h@xmath0 - h@xmath1 = 0 , fig .
the net h@xmath0 image is shown in fig .
1c and 1d .
the net h@xmath0 flux is also measured .
the calibration sources ltt 4364 and eg 274 have nearly flat spectra within the h@xmath0 filter , while the net narrow - band image of is composed of three narrow lines .
we can correct for this spectral difference and subtract the [ nii ] flux with the ctio spectrum ( @xmath334.1 ) and the transmission curve of the h@xmath0 filter .
the h@xmath0 and the [ nii ] @xmath926548 , 6584 lines can be approximated by three gaussians with @xmath42 of @xmath8 3 .
their flux ratios are known from the ctio data .
thus , the conversion factor from the count rate to flux can be derived .
the flux is first measured in an elliptical aperture with 3@xmath23 ( semi - minor axis ) @xmath126 ( semi - major axis ) centered on the h@xmath0 peak ( fig .
6 ) , which encloses almost all emission of the central nebula .
the h@xmath0 ew is 50 , which can be compared with the h@xmath0 ew derived from the ctio spectra ( 26 from the slit position # 2 , 63 from the slit position # 3 ) . with an intrinsic extinction of 1.4 mag ( @xmath334.1 ) ,
the flux is 2.7@xmath127 ergs s@xmath3 @xmath48 and the luminosity is 1.5@xmath128 ergs s@xmath3 .
the total h@xmath0 luminosity of , measured in a boxy aperture ( 20@xmath129 , shown in fig .
1d , including hii regions elo1 and several others ) , is 2.5@xmath128 ergs s@xmath3 ( still assuming 1.4 mag intrinsic extinction ) . with the scaling relation by kennicutt ( 1998 ) ,
the current sfr in the galaxy is @xmath8 2 m@xmath5/yr .
the h@xmath0 tail behind is significant in the first 1@xmath24 from the nucleus , even without continuum subtraction ( fig .
1 ) . nevertheless , the high density of the foreground stars makes the detection of faint , diffuse h@xmath0 emission not easy in this region . in our analysis ,
bright stars are masked and the rescaled h@xmath1 image is subtracted .
the net h@xmath0 image shows a tail to at least 2.2@xmath24 ( or 40 kpc ) from the galaxy .
we notice that the end of the x - ray tail is close to the brightest star in the field , which makes detection of faint , diffuse h@xmath0 emission very difficult there .
the width of the h@xmath0 tail ranges from 3 - 4 kpc .
the h@xmath0 tail aligns with the x - ray tail very well and both of them are brighter in the first 1@xmath24 from the nucleus ( the `` head '' region of the x - ray tail discussed in s06 ) than in regions beyond .
the total h@xmath0 flux in the tail is measured in an 106@xmath130 box covered the whole tail .
this box is adjacent to the 20@xmath129 aperture used in @xmath334.2 to measure the total h@xmath0 flux in .
emission - line objects and stars in the tail are masked and the h@xmath0 fluxes in their positions are recovered from their immediate surroundings . without correction for intrinsic extinction ,
the total h@xmath0 flux in the tail is 4.4@xmath131 ergs s@xmath3 @xmath48 and the total luminosity is 2.4@xmath132 ergs s@xmath3 ( with the same conversion factor used for the h@xmath0 emission of the galaxy in @xmath334.2 ) .
the h@xmath0 surface brightness in the tail ranges from 7.4@xmath133 ergs s@xmath3 @xmath48 arcsec@xmath134 near the galaxy to 1.6@xmath133 ergs s@xmath3 @xmath48 arcsec@xmath134 in the faint region , with an average of @xmath8 2.8@xmath133 ergs s@xmath3 @xmath48 arcsec@xmath134 .
this average surface brightness is similar to that of d100 s h@xmath0 tail in coma cluster ( 0.5 - 4 @xmath133 ergs s@xmath3 @xmath48 arcsec@xmath134 , yagi et al .
2007 ) and 5 - 10 times higher than those of h@xmath0 tails behind two irregular galaxies in a1367 ( gavazzi et al . 2001 ) . in coma and a1367 ,
the surface density of foreground stars is a lot smaller , which permits detection of faint h@xmath0 features .
we follow yagi et al .
( 2007 ) to estimate the mass of the h@xmath0 tail . assuming a cylinder with a 3.5 kpc diameter and a 40 kpc length ( note the tail is somewhat twisted ) , the rms electron density in the tail is @xmath8 0.045 @xmath88 .
the total mass is then 5@xmath135 m@xmath5 for a filling factor of unity .
this amount of mass for 10@xmath71 k gas in the tail is similar to the x - ray gas mass of the tail in the same portion , @xmath8 6@xmath135 m@xmath5 from s06 .
however , both the 10@xmath71 k gas and the 10@xmath4 k gas in the tail can be very clumpy .
if the filling factor for the h@xmath0 emitting gas is 0.05 , the total mass of the h@xmath0 tail reduces to 10@xmath85 m@xmath5 . nevertheless , the stripped ism in the tail accounts for a significant portion of the original ism in the galaxy . the tail may also have cooler gas component . was undetected in the hi observations with atca by vollmer et al .
however , the limit , @xmath17 m@xmath5 , is rather high , largely because of the nearby ( 14.5@xmath24 away ) bright radio galaxy pks 1610 - 60 .
future more sensitive radio hi and infrared observations are required to better quantify the ism content in the galaxy and in the tail .
our observations reveal at least 29 hii regions , all downstream of .
interestingly , the hii regions closest to the galactic disk form a bow - like front with the axis close to the tail .
their projected distances from the galactic nucleus are up to 39 kpc .
most of them ( if not all of them ) appear away from the galactic disk plane .
similar intracluster hii regions have been found before in the virgo cluster : one hii region 21 kpc from ngc 4388 s nucleus in projection ( gerhard et al . 2002 ) and one hii region 3 kpc off the disk and 6.5 kpc from ngc 4402 s nucleus in projection ( cortese et al .
both galaxies are 2 - 3.2 times more luminous than in the @xmath113 band , while their isolated hii regions have h@xmath0 luminosities of only 1.3@xmath136 ergs s@xmath3 and 2.9 @xmath137 ergs s@xmath3 respectively ( intrinsic extinction corrected ) . intergalactic hii regions in poorer environments
have also been found ( ryan - weber et al .
2004 ; mendes de oliveira et al . 2004 ) .
however , none of these known examples match the high number density and luminosities of hii regions downstream of .
for example , 17 hii regions found in this work have h@xmath0 luminosities of @xmath138 ergs s@xmath3 before correction for intrinsic extinction , while only one in previous work ( the brightest one in mendes de oliveira et al .
2004 ) is this luminous .
moreover , five hii regions found in this work are up to five time more luminous than the brightest one in mendes de oliveira et al .
thus , the isolated star formation activity downstream of is unprecedented .
as traverses a3627 , its dark matter halo is tidally truncated by the cluster tidal field .
as we do nt know any characteristic velocity of the galaxy , only a crude estimate can be made . in the simulations by gnedin ( 2003 ) , a dark matter halo of a large spiral galaxy with a circular velocity of 250
km / s is truncated at 30@xmath156 kpc in a cluster similar to the virgo cluster ( @xmath13 = 660 km / s ) . in smaller galaxies ,
the truncation radius is proportional to the circular velocity . from the tully - fisher relation
( e.g. , gnedin et al .
2007 ) , a disk galaxy with s stellar mass has a typical circular velocity of @xmath8 100 km / s .
thus , we take a conservative estimate of 15 kpc for the tidal truncation radius of s dark matter halo .
this estimate is also consistent with the analytical estimate by merritt ( 1988 ) . as shown in fig . 1
, most hii regions may still be in the halo , although three objects at 29 - 39 kpc from the galaxy in projection are hardly still bound . without velocity measurements ,
it is unknown whether most hii regions are still bound in the galactic halo .
however , they may easily escape s potential via weak tidal interaction that barely affects the galactic disk .
we also notice that both the h@xmath0 and x - ray tails are brightest in the halo .
how did these hii regions form ?
there is no indication in the optical that is merging with another galaxy .
elo1 is an emission - line source with insignificant continuum ( fig . 1 and table 1 ) .
in fact , the merger rate in a massive cluster like a3627 should be very small as the cluster velocity dispersion is high ( @xmath13 = 925 km / s for a3627 ) . within 100 kpc ( or 5.3@xmath24 ) from the end of the x - ray tail , there are only two other galaxies identified from dss2 , 2mass and our data , wkk 6166 and another small galaxy without ned i d ( g1 and g2 in fig .
neither has velocity information in the literature , and both are in the fov of our soar data . in the @xmath26 band ,
wkk 6166 is 1.6 - 1.9 mag fainter than , depending on whether its center is blended with a star .
its radius is only @xmath139 , or 3.2 kpc if it is in a3627 . the other galaxy ( g2 in fig .
1b ) is blended with a star . in the @xmath26 band , it is 2.5 - 2.7 mag fainter than , with a major axis extending to @xmath140 . as both galaxies are redder than with @xmath49 of 1.9 - 2.0 mag , their stellar mass is @xmath8 30% - 75% of s , from the mass - to - light ratios in bell & de jong ( 2001 ) . however , even if these two galaxies are in a3627 and have had recent fly - bys with to tidally strip the gas clouds responsible for these hii regions , stars should also be tidally stripped .
we would expect both stellar trails and the hii regions to be present in both directions from . on the contrary ,
the hii regions are present only in a 120 deg cone downstream of .
there are some continuum features downstream of up to 6 kpc from the nucleus ( fig . 1 and fig . 6 ) .
however , all of them have bright hii regions within or nearby .
the upstream side of the galaxy appears undisturbed and free of any tidal features in both the @xmath28 and @xmath26 bands .
cluster galaxies also undergo tidal interaction with the cluster potential ( e.g. , gnedin 2003 ) . recently , cortese et al .
( 2007 ) reported two peculiar galaxies in the massive clusters a1689 ( @xmath141 ) and a2667 ( @xmath142 ) .
images reveal stellar trails composed of bright blue knots and fainter streams behind both galaxies extending to 20 kpc and 75 kpc respectively . the fainter galaxy 131124 - 012040 in a1689 has a similar nir luminosity to , although the star formation in the galaxy has stopped @xmath8 100 myr ago .
they argue that these features are produced through tidal interaction with the cluster potential .
however for , there are no stellar trails connecting the hii regions , at least in the current data .
the detected stellar features 6 kpc downstream of the galaxy are confined around the emission - line regions and hardly extend much .
thus , it is unclear how tidal interactions can strip the clouds responsible for the current hii regions and star clusters , yet not produce stellar trails .
future observations , like optical spectroscopy along both axes and imaging , will allow us to further quantify the dynamical state of ( e.g. , how dynamically cool is the disk ? ) and the stellar diffuse emission downstream of the galaxy .
the significance of ram pressure stripping in and the bow - like front of the hii region distribution drive us to consider another formation mechanism for these hii regions , star - formation in the ram - pressure displaced ism clouds . is undergoing a strong interaction with the surrounding icm , judging by the compactness of the remnant h@xmath0 disk . assuming an ambient icm density of 6@xmath143 @xmath88 and an icm temperature of 6 kev ( see s06 ) , and for a velocity of the galaxy of 1600 km / s ( @xmath144 ) , the thermal pressure is 0.4@xmath145 k @xmath88 , while the ram pressure is 1.1@xmath145 k @xmath88 .
therefore , the ram pressure is high enough to strip 10 k - 100 k ism clouds with density up to 10@xmath71 @xmath88 .
hi gas will be stripped , as well as some less dense dark clouds and the outskirts of molecular clouds , even though the dense cores of molecular clouds may remain in the disk plane .
schaye ( 2004 ) found a critical surface density of @xmath146 @xmath48 for star formation in the outer parts of the galactic disks .
some stripped dense clouds from the disk of should have higher surface density than this critical value , even if clouds are very clumpy .
thus , the ability of high ram pressure to displace dense clouds from the disk may affect the efficiency of star formation in the stripped ism .
the initial instantaneous stripping can be fast in a high ram pressure environment ( @xmath147 10 myr from rh05 ) , while most materials stripped off the disk still remain bound .
the subsequent evolution of the stripped ism is complicated .
simulations ( vollmer et al .
2001 ; ss01 ; rh05 ) show that a fraction of stripped materials ( mainly from the outer disk , @xmath8 10% from rh05 ) can still be bound in the galactic halo ( the `` hang - up '' effect ) for a few 100 myr even in a high ram pressure environment .
some material may even fall back to the disk , while most of the stripped ism concentrates within several galaxy diameters downstream , in the lee of the galaxy where the ram pressure is reduced ( e.g. , ss01 ) .
the details of the process rely on many factors , for example , the porosity of the ism in the inner disk or the effective ram pressure downstream of the galaxy , the halo potential , and the efficiency of heat conduction and viscosity .
simulations ( rh05 ; roediger , brggen & hoeft 2006 ) also show that the wake can be very turbulent .
the stripped clouds will collide with each other , which may produce shocks and trigger star formation .
moreover , being in the halo separates them from the strong stellar uv heating flux present in the disk , allowing cooling to be dominant .
thus , star formation may proceed in the halo or even in unbound clouds far away from the galaxy ( e.g. , elo8 and 9 ) .
we notice that oosterloo et al .
( 2004 ) reported that two intergalactic hii regions at least 100 kpc from the nearest galaxy ( an elliptical ) are in a massive isolated hi cloud .
hii regions in hi clouds were also reported in stephan s quintet by mendes de oliveira et al .
thus , star formation in isolated hi clouds is plausible , although the efficiency may be low as suggested by oosterloo & van gorkom ( 2005 ) .
this scenario can explain the spatial distribution of the hii regions although the details are certainly complicated .
ss01 even suggested formation of dwarf galaxies by these means .
we indeed notice that the brightest hii region , elo1 , has an estimated total mass of @xmath148 m@xmath5 .
if star formation in elo1 is still active , elo1 may grow into a dwarf galaxy eventually .
future and high - resolution hi observations are required to better understand elo1 , other bright hii regions and the stripped ism .
the estimated ages of the hii regions are less than @xmath8 8 myr .
the length of the x - ray tail implies that stripping has lasted for @xmath8 50 myr ( assuming a velocity of @xmath8 1500 km / s in the plane of the sky ) .
thus , it is possible that many hii regions may have already faded .
nevertheless , the long spatial distribution of the hii regions along the tail implies that after the ism has been stripped , the time lapse to the beginning of star formation spans a range up to @xmath8 20 myr or more ( e.g. , for elo8 and 9 ) .
contrary to elo8 and 9 , some luminous hii regions ( e.g. , elo1 - elo3 ) are only @xmath8 1.6 - 5 kpc from the galactic disk in projection , but their ages are comparable to those of elo8 and 9 ( if not longer ) .
their estimated ages are longer than the required time to produce the observed projected offsets .
thus , this may be the evidence that a significant part of stripped ism from the disk can still be bound in the halo .
we also explore the possibility of detecting faded hii regions from the imaging data .
the six additional emission - line regions selected with looser criteria ( fig .
2 ) are certainly candidates with stronger @xmath26 band continuum .
they are also close to the 29 hii regions .
the embedded star clusters are blue if the intrinsic extinction is not too high . from starburst99 model , we find that star clusters like those in elo1 - 3 should be detected in @xmath28 and @xmath26 bands with our data , even at ages of 50 - 60 myr .
however , it is difficult to distinguish them from the various types of galactic stars in the foreground .
about 10 myr after the initial starburst , red supergiants appear and the galactic colors are highly metallicity dependent over this period . for star clusters with metallicities larger than 0.4 solar , starburst99 also predicts a big increase for the @xmath49 color at @xmath8 10 myr for instantaneous starbursts .
@xmath49 color can be as high as 1.7 mag in that period , which makes it almost impossible to select them from galactic stars with imaging data alone ( fig .
2 ) . in fact , ten non - emission - line objects with @xmath149 0.8 mag ( fig .
2 ) are mostly bright stars and none of them are close to the tail .
thus , faded hii regions may be selected only through a spectroscopic survey .
if formation of massive stars proceeds in the stripped ism , high - mass x - ray binaries may form in those star clusters .
the existing shallow ( 14 ks ) observation reveals only three hard x - ray point sources in the tail ( @xmath150 ergs s@xmath3 if in a3627 ) , but both the and the spectra reveal an unresolved hard component in the tail ( s06 ) .
interestingly , the brightest point source p1 ( s06 ) is only @xmath77 from the putative hii region elo5 ( fig .
1 ) , while the combined positional uncertainty from and the soar data can be up to @xmath151 .
the point source p3 positionally coincides with a red optical source ( @xmath120=3.2 ) with an h@xmath0 - h@xmath1 color of 0.01 , which is likely a star .
the point source p2 is 0.7@xmath152 from two star - like objects , but only 2.5@xmath23 away from an emission - line object candidate ( fig .
1 ) . we have a 150 ksec acis - s observation approved for cycle 9 , which can probe the x - ray point source distribution 20 times deeper than the existed data .
the data strongly imply a connection between stripping of the ism and star formation in the halo and intracluster space .
the stripped ism not only contributes to the icm , but also adds intracluster light by subsequent star formation after stripping .
the total mass of starbursts in the 29 hii regions is about 10@xmath4 m@xmath5 ( with large uncertainty from age and extinction ) , while the total gas mass of the h@xmath0 tail and the x - ray tail is @xmath153 m@xmath5 ( depending on the filling factor and density variation ) .
as many hii regions may have already faded and star formation may still proceed in the stripped ism , the final total stellar mass formed in the halo and intracluster space should be at least several times larger than the 10@xmath4 m@xmath5 estimated above . as most of hii regions may still be in the halo , it is possible that star formation in stripped ism is more efficient in the halo than in intracluster space . for
the star clusters in the halo , some of them may remain bound , while others may escape the galactic potential through tidal interactions .
thus , this single transforming galaxy contributes over 10@xmath4 m@xmath5 of new stars in the halo and intracluster space through star formation in the ram - pressure displaced clouds , although it is unclear how many will end in intracluster space .
if some large star clusters are bound to the galaxy for a significant amount of time , the configuration may look like galaxy aggregates found in the coma cluster ( conselice & gallagher 1998 ) .
there are not many cluster galaxies with known h@xmath0 tails , two irregular galaxies in the northwest of a1367 ( gavazzi et al .
2001 ) , ngc 4388 in the virgo cluster ( yoshida et al . 2002 ; 2004 ) , a post - starburst galaxy near the center of the coma cluster ( yagi et al .
2007 ) , and several galaxies in an infalling compact group towards a1367 ( sakai et al . 2002 ; gavazzi et al . 2003
; cortese et al . 2006 ) .
both galaxies in the northwest of a1367 and the post - starburst galaxy in coma have similar @xmath113-band absolute magnitudes as s ( from 0.49 mag fainter to 0.10 mag brighter ) , implying similar total stellar mass as s .
ngc 4388 is 1.3 mag more luminous than in the @xmath113 band and also larger . both h@xmath0 tails in a1367
are suggested to be produced by ram pressure stripping , while the tidal interaction between the two galaxies may trigger star formation in both galaxies ( gavazzi et al .
the h@xmath0 tail behind the galaxy d100 in coma is remarkably narrow and straight ( 60 kpc @xmath154 kpc ) .
it may be produced through either ram pressure stripping , or comes from the gas of a merging dwarf ( yagi et al . 2007 ) .
the h@xmath0 filaments in the northeast of ngc 4388 are suggested to be ionized by ngc 4388 s agn , while ram pressure stripping of the ism provides the cool gas trail ( yoshida et al .
the remarkable infalling compact group in a1367 is composed of several giant galaxies , many dwarf galaxies and h@xmath0-emitting knots .
we notice that the @xmath28 band absolute magnitude of the elo1 ( before correction on the intrinsic extinction ) is comparable to the faint dwarf galaxies in that infalling group ( m@xmath155 = -15.3 mag compared to -15.8 mag ) .
cortese et al . ( 2006 ) suggested that the rich star formation phenomena in this infalling group is triggered by tidal interactions between group members and the ram pressure by the icm .
the galaxy c153 in a2125 ( @xmath12=0.247 ) is another similar case ( wang et al .
2004 ; owen et al .
2006 ) , with a wide 80 kpc [ oii ] tail that is also likely shown in the x - rays .
the galaxy has a distorted disk with substantial star formation activity , just when it is penetrating the cluster core .
however , it is far more distant than all other galaxies discussed here so the interaction can not be studied in the same detail as others .
there are no detections of x - ray tails behind the other galaxies with h@xmath0 tails in the or data .
there are also no reports of isolated hii regions near the two galaxies in the northwest of a1367 and d100 in coma .
the spectroscopically confirmed hii region by gerhard et al .
( 2002 ) is north of ngc 4388 but is far away from the h@xmath0 filaments revealed by yoshida et al .
other emission - line objects in that part of the virgo cluster are even farther away from the h@xmath0 filaments ( fig . 1 of okamura et al .
2002 ) . how does the h@xmath0 tail of form ?
the net h@xmath0 image of the galaxy shows the truncation of the h@xmath0 emission on the front and the sides ( fig . 1 and fig . 6 ) .
in fact , the leading edge of the x - ray emission in coincides positionally with the h@xmath0 front .
the compactness of the remaining h@xmath0 core demonstrates the strength of the icm wind . compared with simulations by roediger & hensler ( 2005 ) , over 90% of the gas in the original gas disk
should have been removed from the disk , either hanging up in the halo or in the tail .
thus , ram pressure stripping should play an important role in the formation of the observed h@xmath0 tail .
the stellar ionization field in the tail may be too weak to ionize the stripped cold gas to produce the observed h@xmath0 tail , at least from the current data .
the positional coincidence of the h@xmath0 tail and the x - ray tail implies a connection .
can heat conduction from the surrounding icm to the cold stripped ism provide enough energy ?
heat conduction across the embedded cold clouds is most likely saturated .
we use the saturated heat conductivity derived by cowie & mckee ( 1977 ) to estimate the flow - in heat flux . the surface area
is still estimated from a cylinder with a 3.5 kpc diameter and a 40 kpc length ( @xmath335 ) , although the real contact surface is almost certainly higher ( e.g. , for clumpy clouds ) .
the total heat luminosity is 4.7@xmath156 ergs s@xmath3 , for an icm temperature of 6.5 kev and an icm density of 10@xmath157 @xmath158 .
although the energy emitted through h@xmath0 is only a small fraction of the total optical line cooling ( several percent from voit , donahue & slavin 1994 ) , there is enough heat energy for the observed h@xmath0 tail .
however , heat conduction can be largely suppressed by magnetic field at the boundaries of the clouds .
it is likely that both stripping and heat conduction contribute to the observed h@xmath0 tail .
spectroscopic studies of the h@xmath0 tail , though difficult , will better clarify its nature .
radio observations are also required to study the distribution of the cold atomic and molecular gas in and behind .
is a star - forming galaxy with a dramatic x - ray tail ( s06 ) . in this work ,
we present optical imaging and spectroscopic data of from soar and the ctio 1.5 m telescopes .
an h@xmath0 tail is found behind , extending to at least 40 kpc from the galaxy .
the h@xmath0 tail positionally coincides with the x - ray tail .
we conclude that ram pressure stripping is responsible for the ism tail , while heat conduction from the hot icm may also contribute to the energy of the optical lines .
this discovery makes the only known cluster galaxy with an x - ray tail and an h@xmath0 tail .
the h@xmath0 emission in is confined to the central @xmath8 1 kpc ( radius ) region .
the current nuclear sfr ( @xmath8 1.2 m@xmath5/yr ) and the largely enhanced stellar continuum ( after correction of intrinsic extinction ) around the center imply that a galactic bulge may be building up .
at least 29 emission - line objects with high h@xmath0 flux are also revealed by the data .
they all spatially cluster downstream of the galaxy , in or around the stripped ism tail . those closest to the galactic disk form a bow - like front . from analysis of the h@xmath1 frame and the estimate of the background emission - line objects ,
we conclude that the expected number of background objects is much less than one in the whole 3.43@xmath40 field and all 29 emission - line objects are very likely hii regions associated with .
the high number density and luminosities of these hii regions downstream of dwarf the previously known examples of isolated hii regions in clusters , making a dramatic example full of rich phenomena .
interestingly , one point source ( an ulx with @xmath159 ergs s@xmath3 if in a3627 ) is in the same position as an hii region .
all these detections indicate significant star formation activity in the halo of and in intracluster space .
our data imply a connection between these hii regions and the stripping of the ism .
we suggest that star formation may proceed in the stripped ism , in both the galactic halo and the intracluster space .
the star formation in the halo may have a higher efficiency as the stripped ism is still bound .
cooling can become dominant when the surrounding uv radiation field is much weaker than in the disk .
cloud collisions may also induce collapse of the ism clouds .
the total mass of the current starbursts in these 29 hii regions is about 10@xmath4 m@xmath5 .
since many hii regions and star clusters may have already faded , the total stellar mass formed in the stripped ism may be several times higher .
therefore , stripping of the ism not only contributes to the icm , but also adds intracluster stellar light through subsequent star formation .
there is no reason to believe is unique .
however , its current active stage may last only for a short time ( @xmath160 yr ) so a large sample of cluster late - type galaxies has to be studied to find more similar ones like , either in x - rays or h@xmath0
. nevertheless , the proximity of makes it a good target to study galaxy transformation , the evolution of stripped ism , star formation and x - ray binaries in the halo and the intracluster space , with the help of more data .
m. s. is very grateful to allan hornstrup , who made an early h@xmath0 observation using the danish 1.54 m telescope with observers jan - erik ovaldsen , josefine h. selj , and dong xu .
we thank nathan de lee for the help on the soar data analysis .
we thank fred walter for the helps on the ctio 1.5 m data and the observation planning .
we acknowledge the work of the soar operators sergio pizarro and daniel maturana , as well as the ctio 1.5 m service observers alberto miranda and alberto pasten .
we thank the referee , giuseppe gavazzi , for helpful and prompt comments .
the financial support for this work was provided by the nasa grant ar6 - 7004x and nasa ltsa grant nng-05gd82 g .
we are presenting data obtained with the southern observatory for astrophysical research , which is a joint project of the brazilian ministry of science , the national optical observatories , the university of north carolina and michigan state university .
we made use of the nasa / ipac extragalactic database and the hyperleda database .
bell , e. f. , & de jong , r. s. 2001 , apj , 550 , 212 bell , e. f. , mcintosh , d. h. , katz , n. , weinberg , m. d. 2003 , apjs , 149 , 289 butcher , h. , & oemler , a. jr .
1984 , apj , 285 , 426 ciardullo , r. , jacoby , g. h. , feldmeier , j. j. , bartlett , r. e. 1998 , apj , 492 , 62 cardelli , j. a. , clayton , g. c. , mathis , j. s. 1989 , apj , 345 , 245 conselice , c. j. , & gallagher , j. s. , iii 1998 , mnras , 297 , l34 cortese , l. et al . 2004 , a&a , 416 , 119 cortese , l. et al .
2006 , a&a , 453 , 847 cortese , l. et al .
2007 , mnras , 376 , 157 cowie , l. l. , mckee , c. f. 1977 , apj , 211 , 135 gavazzi , g. et al .
2001 , apj , 563 , l23 gavazzi , g. et al .
2003 , apj , 597 , 210 gavazzi , g. et al .
2006 , a&a , 446 , 839 gerhard , o. et al .
2002 , apj , 580 , l121 gnedin , o , y. 2003 , apj , 589 , 752 gnedin , o , y. et al .
2007 , apj , in press , arxiv : astro - ph/0607394 gunn , j. e. , & gott , j. r. , iii 1972 , apj , 176 , 1 hamuy , m. , suntzeff , n. b. , heathcote , s. r. , walker , a. r. , gigoux , p. , phillips , m. m. 1994 , pasp , 106 , 566 leitherer , c. et al .
1999 , apjs , 123 , 3 jarrett , t. h. 2000 , pasp , 112 , 1008 kennicutt , r. c. , jr .
1983 , apj , 272 , 54 kennicutt , r. c. , jr .
1998 , ara&a , 36 , 189 kewley , l. j. , & dopita , m. a. 2002 , apjs , 142 , 35 kewley , l. j. , groves , b. , kauffmann , g. , heckman , t. 2006 , mnras , 372 , 961 kraan - korteweg , r. c. , woudt , p. a. , cayatte , v. , fairall , a. p. , balkowski , c. , henning , p. a. 1996 , nature , 379 , 519 malhotra , s. , rhoads , j. e. 2002 , apj , 565 , l71 mendes de oliveira , c. , cypriano , e. s. , sodr , l. , jr . , &
balkowski , c. 2004 , apj , 605 , l17 merritt , d. 1988 , in asp conf .
5 , the minnesota lectures on clusters of galaxies and large - scale structure , ed .
j. dickey ( san francisco : asp ) , 175 oconnell , r. w. 2004 , in asp conf .
322 , the formation and evolution of massive young star clusters , ed .
g. j. l. m. lamers , l. j. smith , & a. nota ( san francisco : asp ) , 551 okamura , s. et al . 2002 ,
pasj , 54 , 883 oosterloo , t. , morganti , r. , sadler , e. m. , ferguson , a. , van der hulst , t. , & jerjen , h. 2004 , in iau symp .
217 , recycling intergalactic and interstellar matter , ed .
duc , j. braine , & e. brinks ( san francisco : asp ) , 486 oosterloo , t. , & van gorkom , j. 2005 , a&a , 437 , 19l osterbrock , d. e. 1989 , astrophysics of gaseous nebulae and active galactic nuclei ( mill valley : university science books ) owen , f. n. , keel , w. c. , wang , q. d. , ledlow , m. j. , morrison , g. e. 2006 , aj , 131 , 1974 pascual , s. et al .
2001 , a&a , 379 , 798 quilis , v. , moore , b. , & bower , r. g. 2000 , science , 288 , 1617 roediger , e. , & hensler , g. 2005 , a&a , 433 , 875 ( rh05 ) roediger , e. , brggen , m. , hoeft , m. 2006 , mnras , 371 , 609 ryan - weber , e. v. et al .
2004 , aj , 127 , 1431 sakai , s. , kennicutt , r. c. , jr .
, van der hulst , j. m. , moss , c. 2002 , apj , 578 , 842 schaye , j. 2004 , apj , 609 , 667 schulz , s. , & struck , c. 2001 , mnras , 328 , 185 ( ss01 ) sun , m. , & vikhlinin , a. 2005 , apj , 621 , 718 sun , m. et al .
2006 , apj , 637 , l81 ( s06 ) sun , m. et al .
2007 , apj , 2007 , 657 , 197 van zee , l. , et al .
1998 , aj , 116 , 2805 voit , g. m. , donahue , m. , slavin , j. d. 1994 , apjs , 95 , 87 vollmer , b. , cayatte , v. , balkowski , c. , duschl , w. j. 2001 , apj , 561 , 708 vollmer , b. et al .
2001 , a&a , 369 , 43 wang , q. d. , owen , f. , ledlow , m. 2004 , apj , 611 , 821 woudt , p. a. , kraan - korteweg , r. c. , fairall , a. p. 1999 , a&a , 352 , 39 woudt , p. a. et al .
2004 , a&a , 415 , 9 woudt , p. a. , kraan - korteweg , r. c. , lucey , j. , fairall , a. p. , moore , s. a. w. 2007 , mnras , submitted , arxiv:0706.2227 yagi , m. et al .
2007 , apj , 660 , 1209 yoshida , m. et al .
2002 , apj , 567 , 118 yoshida , m. et al .
2004 , aj , 127 , 90 - h@xmath1 vs. h@xmath41 for 1708 sources detected in the h@xmath0 band in the field ( 3.42@xmath40 , shown in fig .
both the h@xmath0 and the h@xmath1 fluxes include continuum . over 97% of them are galactic stars with h@xmath0 - h@xmath1 color close to zero .
we define 29 emission - line objects ( in red ) as sources with h@xmath161 -0.2 mag and h@xmath0 - h@xmath162 -0.7 mag ( dotted lines ) .
the h@xmath0 and h@xmath1 magnitudes are ab magnitudes ( hamuy et al .
we have corrected the galactic extinction for all sources ( including galactic stars , 0.15 mag reddening for the h@xmath41 color ) but no correction on the intrinsic extinction associated with has been applied ( also the case for the next three plots ) .
the blue star marks the position of s central region where the giant h@xmath0 nebula is , measured within a 3@xmath163 ( semi - axes ) ellipse centered on the h@xmath0 peak ( also in the next three plots ) .
if we relax the selection criteria ( see @xmath333.1 ) , six more candidate emission - line objects are selected ( in cyan ) . *
upper right * : h@xmath0 magnitude vs. h@xmath41 color .
* lower left * : h@xmath0 - h@xmath1 color vs. @xmath49 color for sources where the @xmath49 color can be constrained .
the emission - line objects appear blue and most of them are bluer than s core ( no correction on the intrinsic extinction ) .
the galactic reddening for the @xmath49 color ( 0.49 mag ) has been applied to all sources . with a small amount of intrinsic extinction ,
the @xmath49 colors of emission - line objects are close to that for pure o / b stars ( -0.2 to -0.8 mag ) .
* lower right * : h@xmath1 magnitude vs. h@xmath56 color for sources detected in the h@xmath1 band .
there are no emission - line objects detected in the h@xmath1 band with h@xmath164 mag
. moreover , no objects are detected with h@xmath1 - h@xmath165 -0.5 mag and h@xmath166 mag . ] 4 ) .
we assumed 1 mag of intrinsic extinction for all of them .
for comparison , the two isolated hii regions in the virgo cluster ( gerhard et al .
2002 ; cortese et al .
2004 ) have h@xmath0 luminosities of 1.3@xmath136 ergs s@xmath3 and 2.9@xmath137 ergs s@xmath3 respectively ( intrinsic extinction corrected ) . ]
image ( the lower right panel ) .
the dispersion axis is always horizontal and the spatial axis is always perpendicular on the detector plane ( the west is upward ) .
note that each slit has a length of @xmath38 and a width of 3@xmath23 centered on the galaxy , while the slits plotted are truncated only for better viewing .
the three 47/ib 2d spectra are spatially aligned .
the # 1 position completely misses the h@xmath0 emission of the galaxy , while the # 3 position covers the h@xmath0 peak of the galaxy .
emission - line objects ( elo1 , and a - f as marked ) were caught in one or more slit positions as shown in the spectra .
the small scale - bar in the # 3 plot ( near `` g '' ) is 6.56 , or 300 km / s .
the slit for the 26/ia observation runs across the peak of the galaxy and a large part of elo1 .
the 26/ia spectrum shows extension to the direction of elo1 ( upward on the detector plane ) for the three strong lines , which implies that elo1 has nearly the same velocity as s .
note that the best instrumental spatial seeing for the spectrograph is 3.5@xmath23 ( or 2.7 pixels ) . ]
spectrum was taken at the slit position # 3 ( fig .
4 ) . emission lines are marked .
the line widths are all small ( e.g. , 3.7 fwhm for the h@xmath0 line ) . as the spectroscopic standards
were observed with 2@xmath23 slits , the shown flux of each spectrum is over - estimated by an unknown constant , but the spectral shape and line ratios are not affected . ]
width ) in four bands ( color coding in the right plot ) .
we chose the origin as the peak of the h@xmath0 profile ( same in the next two plots ) .
stars are masked but emission - line objects are included ( also for the plot shown in the middle panel ) .
the h@xmath1 light traces the h@xmath0 light well upstream ( or east ) until @xmath167 ( or @xmath103 kpc ) from the h@xmath0 peak .
the three h@xmath0 peaks downstream are emission - line objects .
the enhanced @xmath28 band light downstream ( relative to @xmath26 band ) is related to the emission - line objects .
the @xmath26 band profile at -15@xmath23 -5@xmath23 can be fitted with an exponential profile with a scale height of @xmath168 . *
middle * : surface brightness profiles measured parallel to the major axis of the galaxy ( shown as a green box in fig .
1b , 4@xmath23 width ) in four bands . the h@xmath1 light and the scaled continuum
light trace the h@xmath0 light well beyond the central 5@xmath23 ( or @xmath169 kpc ) .
the continuum light within the central 5@xmath23 is flattened with substructures ( also shown in fig .
1e and 1f ) .
surface brightness of some area in the north is absent because of bright stars there .
if the profiles at -26@xmath23 -6@xmath23 and 9@xmath23 37@xmath23 are fitted with exponential profiles respectively , the derived scale heights are @xmath170 in the south and @xmath171 in the north . *
right * : surface brightness profiles measured in elliptical annuli are plotted against the semi - major axis ( @xmath108 ) .
the elliptical annuli are centered on the h@xmath0 peak of the galaxy and have a fixed axis ratio of 2 ( @xmath172 , @xmath37 : semi - minor axis ) .
the positional angle is the same as that used to measure the surface brightness profiles along the major axis in the middle panel ( fig .
all unresolved sources ( including all emission - line objects but elo1 ) are masked .
the @xmath28 and @xmath26 profiles at 7@xmath23 32@xmath23 can be well fitted with an exponential profile with a scale height of @xmath109 ( or 3 kpc ) . ] | we present the discovery of a 40 kpc h@xmath0 tail and at least 29 emission - line objects downstream of a star - forming galaxy in the rich , nearby cluster a3627 .
the galaxy is known to possess a dramatic 70 kpc x - ray tail .
the detected h@xmath0 tail coincides positionally with the x - ray tail . the h@xmath0 emission in the galaxy is sharply truncated on the front and the sides near the nucleus , indicating significant ram pressure stripping .
is thus the first cluster late - type galaxy known unambiguously with both an x - ray tail and an h@xmath0 tail .
the emission - line objects are all distributed downstream of the galaxy , with projected distance up to 39 kpc from the galaxy . from the analysis on the h@xmath1 frame and the estimate of the background emission - line objects ,
we conclude that it is very likely all 29 emission - line objects are hii regions in a3627 .
the high surface number density and luminosities of these hii regions ( up to 10@xmath2 ergs s@xmath3 ) dwarf the previously known examples of isolated hii regions in clusters .
we suggest that star formation may proceed in the stripped ism , in both the galactic halo and intracluster space .
the total mass of formed stars in the stripped ism of may approach several times 10@xmath4 m@xmath5 .
therefore , stripping of the ism not only contributes to the icm , but also adds to the intracluster stellar light through subsequent star formation .
the data also imply that is in an active stage of transformation , accompanied by the build - up of a central bulge and depletion of the ism . |
the discovery@xcite of superconductivity in @xmath4 at @xmath12 opens some very interesting opportunities for practical applications of this material as a conductor in power applications.@xcite typically , the operating temperature of these conductors would be at about half the superconducting transition temperature , @xmath13 , or about @xmath14 where closed cycle refrigerators are quite efficient .
early studies of the temperature and magnetic field dependence of the superconducting critical current density , @xmath15,@xcite indicated that grain boundaries in this material would carry very substantial supercurrents , @xmath16 , in conductors prepared in a variety of ways.@xcite in addition , a very early attempt to explore the use of powder - in - tube methods by jin and co - workers@xcite showed that practical conductors are feasible . on the fundamental side , the measurement of an isotope@xcite effect with @xmath17 where @xmath18 and @xmath19 is the isotopic mass of the boron atom indicated that phonons play a major role in the superconducting interaction in @xmath4 .
for clean samples , resistivity measurements at @xmath20 indicated that @xmath21 which , combined with the calculated fermi velocity , @xcite @xmath22 , gives a mean - free - path , @xmath23 . from the measured upper critical field ,
@xmath24.@xcite , the coherence distance of a clean sample can be determined from @xmath25 $ ] to be @xmath26 . here , @xmath27 is the flux quantum . with @xmath28
substantially greater than @xmath29 , these materials are in the clean limit .
several groups@xcite found @xmath15 values in the range of @xmath30 in self field at @xmath6 , but these values fell below @xmath31 at @xmath32 at @xmath6 . broadly speaking , the irreversibility field , @xmath33 was roughly half of @xmath34 .
the primary drawback for this material appeared to be the relatively low values of the upper critical field , @xmath35,@xcite and the resulting rapid drop in @xmath15 with field .
the most straight forward way to raise @xmath34 is to shorten the mean - free - path , @xmath28 , which in turn shortens the coherence distance , @xmath36 .
thus @xmath37 $ ] would rise with decreasing @xmath28 .
the introduction of point defects is an effective way to raise @xmath34 . in pure @xmath4
, the volume pinning force seems to decrease as @xmath38 ^ 2 $ ] indicating that grain boundary pinning dominates.@xcite hence , reducing the grain size should help .
alternatively , the strength of the pinning and @xmath15 , can be increased by introducing precipitates .
typical dimensions might be an array of pinning sites with a characteristic dimension of the coherence distance , @xmath39 , and a spacing between pinning centers about @xmath40 times larger .
doping of @xmath4 has proven to be difficult because relatively few impurities go into the lattice in solid solution .
early reports by slusky and co - workers@xcite show that @xmath41 added by reacting powders in a solid state reaction substitute for @xmath3 in the @xmath4 lattice .
there is a smooth change in lattice parameter and smooth decrease in @xmath13 for @xmath42 for @xmath43 out to @xmath44 and then a two phase region is entered . at @xmath45 , @xmath13 is still about @xmath46 and thus still relatively high . for @xmath10 doping by solid state reaction of powders
, it appears that relatively little @xmath10 is incorporated in the the @xmath4 grains .
work by zhao and coworkers@xciteshowed that @xmath10 appears as very thin layers of @xmath2 precipitates on the @xmath4 grain boundaries .
samples were made via solid state reaction by grinding the @xmath10 , @xmath3 , and @xmath11 powders , pressing pellets , and heat treating .
microstructural analysis seemed to indicate that @xmath2 may act as a sintering agent leading to higher density , very fine @xmath4 grain - size of about @xmath47 , and excellent grain - to - grain coupling . some high resolution transmission electron microscopy evidence points to the thickness of some of these grain boundary layers being as low as one @xmath2 unit cell.@xcite
because the @xmath10 does not enter the unit cell , there is relatively little diminution in @xmath13 and very substantial increases in the critical current density , @xmath15 .
optimum doping was in the range of a @xmath48 ratio of about @xmath49 .
one of the more important outcomes of this work was the observation that the @xmath10 seems to refine the grain structure to a grain size of about @xmath47 . for carbon doping
, rather large amounts of carbon can replace boron in the structure .
takenobu and co - workers@xcite found that carbon substitutes for boron in @xmath50 out to @xmath51 as indicated by a smooth decrease in a - axis lattice parameter from @xmath52 at @xmath53 to @xmath54 at @xmath55 while @xmath13 drops from @xmath56 to about @xmath57 .
there was relatively little change in the c - axis dimension . here
, samples were made by mixing powders of the elements and firing at @xmath58 for @xmath59 . at higher @xmath60-values
, they found that the ( 110 ) x - ray peak of @xmath4 broadens indicating two phases.@xcite .
bharathi and co - workers,@xcite by contrast , found that reacting under high pressure @xmath61 , permitted the extention of carbon content to @xmath62 .
they find a relatively smooth depression of @xmath13 from @xmath63 at @xmath53 to about @xmath64 at @xmath62 .
using a somewhat different procedure , we have found that reacting @xmath65 with @xmath3 vapor at @xmath66 for @xmath59 gives a material with @xmath67 .
in other work , the change in lattice constant and @xmath13 with carbon content also has been reported to be somewhat less than these values.@xcite for si doping , a solid state reaction of @xmath3 and amorphous @xmath11 powder with nano - size @xmath68 particles show substantial @xmath69 lines in the x - ray data and relatively little diminution in the transition temperature to about @xmath70.@xcite this seems to imply that very little si seems to enter the @xmath4 structure . in an alternate preparation method , cooley and co - workers@xcite have shown @xmath3 can be reacted with @xmath71 at temperatures as low as @xmath72 to give @xmath69 precipitates in a @xmath4 matrix with @xmath73 .
these preliminary experiments seem to indicate that the enhancements in flux pinning arise from @xmath69 precipitates rather than @xmath68 substitution into the @xmath4 lattice .
the purpose of this review is to summarize progress in attempts to make @xmath4 conductors using a cvd method for depositing the starting boron material .
there is a well established commercial scale method to make boron fiber in kilometer lengths for use in high strength composites .
the primary goal of this work is to find ways to make doped boron fiber that will permit the preparation of long lengths of high performance @xmath4 conductor .
tasks include 1 ) the development of processing methods to convert pure boron fiber to @xmath4 , 2 ) the development of methods to dope the starting boron fiber with point defects that will raise the upper critical field , @xmath34 , 3 ) the development of boron fibers that will produce nano - scale precipitates in the @xmath4 , and 4 ) the development of methods to handle the brittle character of the superconductor .
chemical vapor deposition of the starting boron is a particularly attractive route to materials preparation because the impurity atoms can possibly be introduced with atomic scale dispersion .
initial experiments have focussed on titanium - doping to produce small @xmath2 precipitate pinning centers with the next step to be the introduction of carbon point defects to shorten the electronic mean free path . in a typical deposition process for boron fibers , a conducting filament such as tungsten or a carbon coated commercial silicon carbide fiber
is introduced into a long glass chamber through a @xmath74 seal at a rate of a few @xmath75 . a flowing stream of @xmath76 and hydrogen gas move through the full length of the chamber .
the fiber is heated electrically to temperatures in the @xmath77 range , and the boron is deposited in the hot zone .
if a suitable carrier gas can be mixed with this gas stream , then impurity atoms such as @xmath10 or @xmath78 can be co - deposited with the boron to form filaments that can subsequently converted to impurity doped-@xmath4 .
critical current densities were measured via magnetization as reported earlier.@xcite
soon after the discovery of superconductivity@xcite in @xmath4 , small wire segments of commercial boron fiber were converted to @xmath4.@xcite these wire segments were only a few @xmath79 long and rather fragile , but they showed exceptionally low resistivity of @xmath80 at @xmath20 and a sharp superconducting transition at @xmath12 . in a subsequent series of upper critical field measurements,@xcite @xmath81 curves indicated an extrapolation to @xmath82 of about @xmath83 .
critical current densities , @xmath15 for a @xmath84 diameter pure boron fiber reacted in @xmath3 vapor at @xmath58 for @xmath59 are shown for both @xmath6 and @xmath7 in fig .
1 . at @xmath6 , @xmath85 in self field , and @xmath15 is about @xmath86 at @xmath87 . at @xmath7 , @xmath88 in self field , and @xmath15
is about @xmath86 at @xmath89 . also shown on fig .
1 are data for a sample made by depositing @xmath90 of pure boron on a carbon substrate .
the @xmath6 data are the open circles and the @xmath7 data are the solid triangles . within the reproducibility of the experiments ,
the @xmath84 pure boron fibers give the same @xmath15 values as samples made from @xmath90 of pure b on a c substrate within the accuracy of the measurement .
for all samples , the reaction is carried out at a single temperature , and the time can then be adjusted to optimize @xmath15 if that is desired . a series of measurements were undertaken to determine the lowest temperatures where these boron fibers would transform to @xmath4 at a high rate .
a sample reacted at @xmath91 for @xmath59 , as shown in fig .
2 , shows a @xmath3 region only part way across the boron fiber in the relatively light area near the edge of the filament .
a more detailed analysis shows that this is a uniform wall of a high boron compound , @xmath92 , progressing across the fiber .
energy dispersive spectra in the scanning electron microscope indicate that this phase has @xmath60 near @xmath93 or @xmath94 .
because this phase is not the study of this work , further analysis was not undertaken . for this time and temperature , essentially no @xmath95 or @xmath4 forms .
the white core of the filament is @xmath96 .
by contrast , if a similar pure boron fiber is heated in @xmath3 vapor at @xmath97 for @xmath98 , the fiber transforms to @xmath4 nearly everywhere as shown in fig .
the darker gray areas are small regions of @xmath95 as shown by the eds line scan along the line shown , where the circle with a cross is the starting point of the scan .
the large dip in @xmath3 count rate is the @xmath96 core and the @xmath99 percent dip in count rate to the left of the core illustrates the decrease in @xmath3 density going from @xmath4 to @xmath95 .
also visible in the micrograph are several faults along the edge of the fiber and a large void near the right end of the portion of the @xmath96 core .
the filament is not completely flat , so that the @xmath96 core does not show along the entire length of the filament for this particular depth of polish .
these flaws are typical in @xmath4 filaments prepared from commercial pure @xmath11 fibers . in fig .
4 , an electron micrograph is shown in end view for the long time limit where a @xmath84 diameter fiber that has been reacted in @xmath3 vapor at @xmath66 for @xmath100 .
the white core of @xmath96 in the center is measured here to be @xmath101 in diameter and the starting @xmath84 boron fiber has grown to about @xmath102 diameter as @xmath3 is added .
there is a small patch of @xmath95 in the upper left quadrant , and there the void pattern is very typical of all the samples grown .
in the early experiments with @xmath0 doping of @xmath11 with @xmath10 , the material was deposited on either a @xmath103 substrate or a @xmath78 coated substrate , and a preliminary report of this work has been presented.@xcite most of the experiments reported here , however , were done by simultaneously depositing @xmath11 and a few percent @xmath10 on a continuous @xmath104 diameter silicon carbide filament coated with several micrometers of glassy carbon ( commercially available as scs-9a ) .
deposition was accomplished in a reactor similar to that used for commercial boron filament production.@xcite the substrate filament was drawn through the reactor at @xmath105 ; the filament in the reactor was resistively heated to peak temperatures of approximately @xmath106 .
the reactant gas was predominantly a stoichiometric mixture of hydrogen and boron trichloride at atmospheric pressure .
titanium dopant was added by bubbling the hydrogen component through titanium tetrachloride held at @xmath107 .
the resulting partial pressure of titanium tetrachloride was approximately @xmath108 .
the doped boron - coated substrate was spooled continuously upon emerging from the reactor .
an x - ray powder diffraction study of the resulting fibers before @xmath3 reaction indicates comparable amounts of @xmath1 and @xmath2 as well as a small amount of @xmath109 .
subsequent reaction with @xmath3 vapor produced @xmath4 filaments with precipited @xmath2 .
the intensity of the x - ray lines indicates less @xmath1 in the reacted fiber than in the unreacted fiber .
a series of samples were prepared in which @xmath11 and @xmath10 were co - deposited on the carbon surface with thicknesses of @xmath110 , @xmath111 and @xmath112 .
after reaction in @xmath3 vapor at both @xmath66 for @xmath59 and at @xmath106 for @xmath113 , the conversion of @xmath11 to @xmath4 was complete.@xcite an extensive study of the @xmath10 distribution was undertaken with @xmath114 in a @xmath115 where the beam size was approximately @xmath8 micrometer . generally , the @xmath116 ratio was in the neighborhood of @xmath117 in the bulk of the material .
this is illustrated in fig .
5 where a sample with @xmath111 of @xmath118 is deposited on a carbon coated substrate .
there are two regions of the sample , however , that are consistently high @xmath10 .
this is near the @xmath78 substrate and near the outer surface .
reaction kinetics indicate that the deposition of @xmath10 compared to @xmath11 is higher at lower temperatures . hence ,
as the fiber moves through the deposition zone , the @xmath10 deposition will be faster than that of boron during the initial warm - up on entrance to the hottest regions and during the cool - down as the fiber exits the hottest regions .
another specimen illustrated in fig .
6 shows a @xmath110 thick layer of @xmath119 . here , there are two rather distinct regions of @xmath10 content . in most of the boron layer ,
the @xmath10 content is about @xmath120 , but sometimes there are bright lacy regions where the @xmath10 content is about @xmath121 . by studying polished regions of the @xmath4 with polarized light ,
the grain structure is readily apparent . if a particular region of the sample is studied both with polarized light and with @xmath115 with @xmath114 area maps , the @xmath10 is found to be spread throughout the grains .
the grain boundaries are not seen to be high in @xmath10 on these area maps .
this result differs from the beautiful decoration of @xmath4 grain boundaries with @xmath2 seen for samples made from powders in solid state reactions.@xcite titanium doping using @xmath0 methods gives a dispersion of @xmath2 in @xmath4 that is different from that seen for solid state reaction of powders .
the addition of @xmath10 greatly enhances the @xmath15 values if the level of @xmath10 doping is properly adjusted . as reported previously , anything approaching a @xmath116 ratio of 1 is far too much @xmath10.@xcite the volume fraction of superconductor is reduced by half , and the value of @xmath13 drops to about @xmath14 . in accord with earlier findings,@xcite
a @xmath116 ratio in the neighborhood of @xmath122 greatly enhances both @xmath15 and @xmath33 while depressing @xmath13 only a few degrees . at @xmath6 and self field ,
@xmath15 increases from about @xmath123 for pure @xmath4 to about @xmath124 for the @xmath125 sample . with increasing magnetic field , @xmath15 crosses @xmath9 at @xmath126 with @xmath125 as compared to @xmath89 for the pure @xmath4 sample as shown by the solid triangles in fig .
7 . at @xmath7 , the @xmath127 sample crosses @xmath9 at about @xmath128 compared with the pure @xmath4 sample which crosses @xmath9 at @xmath129 .
in addition to simply adding @xmath10 to the boron fibers , there is a great deal of optimization that can yet be done .
when the starting @xmath130 fibers are prepared , there is a large amount of both @xmath1 and @xmath2 in the boron . during the reaction to form @xmath4 ,
some of the @xmath1 is converted to @xmath2 , and there may be some migration of @xmath78 from the substrate into the sample . much more needs to be learned about both of these processes .
an optimization of @xmath15 at @xmath7 with @xmath130@xcite shows that a reaction time of @xmath113 is best for a reaction temperature of @xmath106 and a reaction time of @xmath59 is best for a reaction temperature of @xmath66 .
an optimization for @xmath15 values at @xmath6 , shown in fig .
8 , indicates that @xmath113 at @xmath106 gives the highest @xmath15 values .
for pure @xmath11 fibers deposited on a @xmath131 @xmath103 core , a series of @xmath11 fibers having diameters ranging from @xmath84 to @xmath132 were fully reacted in @xmath3 vapor at @xmath66 for times ranging from @xmath59 to @xmath133 . in all cases ,
the reaction to form @xmath4 increases the fiber diameter by about @xmath134 . at @xmath6 and self field ,
the @xmath15 values range from @xmath135 to @xmath136 and have a magnetic field dependence similar to those shown in fig .
1 . at @xmath7 , the data of fig . 9
indicate that @xmath137 nearly is independent of diameter . to the accuracy and reproducibility of the measurement ,
all of these fiber diameters have the same @xmath138 plots . for the @xmath139 deposited on a carbon coated substrate ,
samples having a thickness of @xmath110 , @xmath111 , and @xmath112 have been studied .
we expected a difference in @xmath15 among these samples caused by different amounts of @xmath78 impurity or caused by a different distribution of @xmath1 or @xmath2 during the @xmath10 and @xmath11 deposition on the @xmath78 substrate .
these are relatively small effects as shown on fig . 10 .
these data were taken for samples reacted at @xmath106 for @xmath140 .
a reaction time less than the @xmath113 for optimum @xmath15 was intended to limit possible diffusion of @xmath78 from the substrate . in broad terms , the @xmath6 @xmath15 data are about @xmath141 at low field and the @xmath15 data cross the @xmath9 line at about @xmath142 .
there is not a large thickness dependence of @xmath15 , but at high fields , the thicker samples seem to show a higher @xmath15 values . at low fields ,
the thinner samples seem to show the higher @xmath15 values .
similar data have been reported elsewhere.@xcite there is lots of room for improvement by optimizing the @xmath10 content , the reaction temperatures , and the reaction times .
co - deposition of boron along with the desired impurities offers a distinctive method of doping .
when these doped fibers are converted to @xmath4 by heating in @xmath3 vapor , the resulting @xmath2 is distributed relatively uniformly within the grains .
this distribution is quite different from @xmath4 samples prepared by solid state diffusion of mixed powders .
critical current densities of @xmath143 to @xmath40 million @xmath144 are obtained at @xmath6 and self field .
critical current densities of @xmath9 are obtained at @xmath7 and @xmath89 . a number of important variables are still not well understood .
improvement is needed in the control of the distribution of @xmath2 , and a more thorough study of carbon doping needs to be undertaken so that a quantitative contrast can be developed between the effects of @xmath78 substitution in the lattice with the effects of @xmath2 precipitates .
although there is a fairly wide sweet spot for the reaction of doped @xmath11 fibers to form doped @xmath4 , there is lots of room for improvement .
the performance of the conductor can yet be optimized with respect to @xmath10 content and carbon content as well as the time and temperature of @xmath3 reaction .
work is supported by the u.s .
department of energy , basic energy sciences , office of science , through the ames laboratory under contract no .
w-7405-eng-82 .
j. nagamatsu , n. nakagawa , t. muranaka , y zenitani , and j. akimitsu , nature * 410 * ( 2001 ) 63 .
paul m. grant mrs symposium proceedings , * 689 * ( 2002 ) 3 .
d. k. finnemore , j. e. ostenson , s. l. budko , g. lapertot , p. c. canfield , phys .
b * 86 * ( 2001 ) 2420 .
d. c. larbalestier , m. rikel , l. d. cooley , a. a. polyanskii , j. y. jinang , s. putnaik , x. y. cai , d. m. feldmann , a. gurevich , a. a. squitier , m. t. naus , c. b. eom , e. e. hellstrom , r. j. cava , k. a. regan , n. rogado , m. a. hayward , t. he , j. s. slusky , k. iaumaru , m. haas , nature * 410 * ( 2001 ) 186 .
y. bugoslavsky , g. k. perkins , x. qi , l. f. cohen , and a. d. caplin , nature * 410 * ( 2001 ) 563 .
s. jin , h. mavoori , and r. b. vandover , nature , * 411 * ( 2001 ) 563 .
s. l. budko , g. lapertot , c. petrovic , c. e. cunningham , n. e. anderson , and p. c. canfield , phys .
lett . * 86 * , ( 2001 ) 1877 .
j. kortus , i. i. mazin , k. d. belashchenko , v. p. antropov , and l. l. boyer , phys .
* 86 * ( 2001 ) 4656 .
s. l. budko , c. petrovic , g. lapertot , c. e. cunningham , p. c. canfield , m - h , jung , and a. h. lacerda , phys .
b , * 63 * ( 2001 ) 220503(r ) .
j. s. slusky , n. rogado , m. a. hayward , p. khalifah , t. he , k. inumaru , s. loureire , m. k. haas , h. w. zandbergen , and r. j. cava , nature , * 410 * 2001 343 .
y. zhao , y. feng , c. h. cheng , l. zhou , y. wu , t. machi , y fudamoto , n. koshizuka , and m. murakami , appl .
. lett . * 79 * , 1154 ( 2001 ) ; y. zhao , d. x. huang , y. feng , c. h. cheng , t. machi , n. koshizuka , and m. murakami , appl .
80 * ( 2002 ) 1640 .
t. takenobu , t. ito , d. h. chi , k. prassides , and y. iwasa , phys .
* 64 * ( 2001 ) 134513 .
a. bharathi , s. k. balaselvi , s. kalavathi , g. l. n. reddy , v. s. sastry , y. haribaran , and t. s. radfakrishnan , physica c * 370 * ( 2002 ) 211 .
m. paranthaman , j. r. thompson , and d. k. christen , physica c * 355 * ( 2001 ) 1 .
x. l. wang , s. h. zhou , m. j. qin , p. r. monroe , s. soltanian , h. k. liu , and s. x. dou , lanl cond .
mat./0208349 l. d. cooley , v. braccini , j. waters , p. hellenbrand , b. senkowicz , j. y. jiang , p. j. lee , e. e. hellstrom , and d. c. larbalestier , ( private communication ) p. c. canfield , d. k. finnemore , s. l. budko , j. e. ostenson , g. lapertot , c. e. cunningham , and c. petrovic , phys .
* 86 * ( 2001 ) 2423 .
n. e. anderson , w. e. straszheim , s. l. budko , p. c. canfield , and d. k. finnemore , physica c , submitted .
r. j. suplinskas and j. v. marzik , `` boron and silicon carbide filaments '' , in , j. v. milewski and h. s. katz , ed . , van nostrand reinhold , new york , ( 1987 ) . | processing methods are described for the development of magnesium diboride wire using the chemical vapor deposition ( @xmath0 ) to produce long lengths of suitably doped starting boron fiber .
it is found that titanium can be co - deposited with the boron to make long lengths of doped fiber that contain both @xmath1 and @xmath2 . when this fiber is reacted in @xmath3 vapor to transform boron into @xmath4 ,
the resulting conductor has a superconducting critical current density of about @xmath5 at @xmath6 and self field .
the critical current density at @xmath7 and @xmath8 tesla is @xmath9 . using optical methods to define grain boundaries and energy dispersive x - rays to determine @xmath10 and @xmath3 concentration ,
these samples show a fine dispersion of @xmath10 through out the grains and no conspicuous precipitation of @xmath2 on the @xmath4 grain boundaries .
this is to be contrasted with the precipitation of @xmath2 on @xmath4 grain boundaries observed for samples prepared by solid state reaction of @xmath10 , @xmath3 , and @xmath11 powders .
introducing @xmath10 impurities into the @xmath11 during the @xmath0 deposition of the @xmath11 gives a distribution of @xmath2 in the resulting @xmath4 different from solid state reaction of powders . |
numerical simulations of freely decaying turbulence show that vortex interactions play a crucial role in the dynamics of two - dimensional viscous flows @xcite . in particular , vortex mergers are responsible for the appearance of larger and larger structures in the flow , a process which is directly related to the celebrated `` inverse energy cascade '' @xcite .
although nonperturbative interactions such as vortex mergers are extremely complex and desperately hard to analyze from a mathematical point of view @xcite , rigorous results can be obtained in the perturbative regime where the distances between the vortex centers are large compared to the typical size of the vortex cores . as an example of this situation ,
consider the case where the initial flow is a superposition of @xmath3 point vortices .
this means that the initial vorticity @xmath4 satisfies @xmath5 where @xmath6 are the initial positions and @xmath7 the circulations of the vortices .
let @xmath8 be the solution of the two - dimensional vorticity equation @xmath9 with initial data @xmath4 , where @xmath10 is the velocity field defined by the biot - savart law @xmath11 solutions of , with singular initial data of the form were first constructed by benfatto , esposito , and pulvirenti @xcite .
more generally , if @xmath12 is any finite measure , giga , miyakawa , and osada @xcite have shown that the vorticity equation has a global solution with initial data @xmath4 , which moreover is unique if the total variation norm of atomic part of @xmath4 is small compared to the kinematic viscosity @xmath13 .
this last restriction has been removed recently by i. gallagher and the author @xcite , so we know that has a unique global solution @xmath14 with initial data , no matter how small the viscosity coefficient is .
this solution is uniformly bounded in @xmath15 , and the total circulation @xmath16 is conserved . in the vanishing viscosity limit
, the motion of point vortices in the plane is described by a system of ordinary differential equations introduced by helmholtz @xcite and kirchhoff @xcite .
if @xmath17 denote the positions of the vortices , the system reads @xmath18 and the initial conditions @xmath19 for @xmath20 are determined by .
a lot is known about the dynamics of the _ point vortex system _ , see e.g. @xcite for a recent monograph devoted to this problem .
most remarkably , is a hamiltonian system with @xmath3 degrees of freedhom , which always possesses three independent involutive first integrals . in particular , system is _ integrable _ if @xmath21 , whatever the vortex circulations @xmath22 may be . in the simple situation where @xmath23 , both vortices rotate with constant angular speeed around the common vorticity center , see fig.1(left ) . in the exceptional case where @xmath24
, there is no center of vorticity and the vortices move with constant speed along parallel straight lines , see fig.1(right ) . [ cols="^,^ " , ] the discussion above shows that the dynamics of weakly interacting viscous vortices is essentially driven by two different mechanisms : _ diffusion _ , which is responsible for the continuous growth of the vortex cores , and _ advection _ , which creates the motion of the vortex centers and the deformation of the vortex profiles .
the latter effect persists in the vanishing viscosity limit , and it is therefore reasonable to expect that , if we can find solutions of euler s equation describing widely separated gaussian vortices , these inviscid solutions will provide an accurate approximation of the viscous @xmath3-vortex solution considered in theorem [ thm1 ] , if @xmath13 is sufficiently small .
the aim of this contribution is to explore this idea in the particular case of a single _
vortex pair_. in this simple situation , we only need to consider solutions of euler s equation that are stationary in a uniformly rotating or translating frame .
assume thus that @xmath23 and , for definiteness , that both circulations @xmath25 are positive . given @xmath26 such that @xmath27 , let @xmath28 and @xmath29 .
we consider the inviscid vorticity equation in a rotating frame with angular speed @xmath30 : @xmath31 formally , the vorticity distribution @xmath32 is a stationary solution of . in the laboratory frame , this corresponds to a time periodic solution of where both vortices rotate around the origin with angular velocity @xmath30 . now , let @xmath33 be a nonnegative vorticity profile which is radially symmetric , decreasing along rays , and normalized in the sense that @xmath34 . given @xmath35
, we look for a stationary solution of of the form @xmath36 where @xmath37 and @xmath38 as @xmath39 .
existence of such `` desingularized '' solutions of euler s equation has been investigated in a recent work by d. smets and j. van schaftingen @xcite .
in fact , the authors of @xcite do not consider rotating vortex pairs of the form , but they treat a variety of other interesting cases , including a single stationary vortex in a bounded or unbounded domain , a rotating vortex in a disk , and a translating vortex pair in the plane .
they use the `` stream function method '' , which consists in constructing ( by variational methods ) a nontrivial solution to an elliptic equation of the form @xmath40 , where @xmath41 is a power - like nonlinearity which depends on @xmath0 in an appropriate way .
the vorticity @xmath42 is then a stationary solution of . for technical reasons ,
the desingularized vorticity profiles obtained in @xcite are always compactly supported , but it is reasonable to expect that similar results can be obtained with gaussian profiles too .
we hope to clarify this issue and to extend the results of @xcite to pairs of vortices of the same sign in a future work . in section [ sec2 ]
below , we study in detail the case of two identical vortices ( @xmath43 ) . for a large class of
radially symmetric profiles @xmath44 , we prove the existence of approximate stationary solutions of of the form .
in other words , we construct an asymptotic expansion in powers of @xmath0 of the vorticity distribution as a steady state of . under natural symmetry assumptions ,
we show that this expansion can be performed to arbitrarily high order .
then , in section [ sec3 ] , we prove that the solution @xmath45 of the rotating viscous vorticity equation @xmath46 with initial data is very close to the inviscid stationary solution @xmath47 with asymptotic profile @xmath48 , if @xmath2 is sufficiently small .
this means that , when the viscosity is small , the solution of with initial data slowly travels through a family of uniformly rotating solutions of euler s equation , whose vorticity profiles are approximately gaussian and evolve diffusively .
we expect a similar picture to be relevant in the general situation considered in theorem [ thm1 ] , although the corresponding inviscid solutions may be more difficult to identify in that case
. * acknowledgements . *
i am grateful to felix otto for suggesting the point of view adopted in this paper , and to didier smets for many stimulating discussions .
the aim of this section is to construct an asymptotic expansion for a family of stationary solutions of the inviscid vorticity equation , which correspond to weakly interacting vortex pairs . for simplicity
, we only consider the particular case where both vortices have the same circulation @xmath49 .
as is explained in the introduction , weak interaction means that the distance @xmath50 between the vortex centers is large compared to the size of the vortex cores .
if the vorticity distribution is given by , this condition is satisfied if @xmath35 is sufficiently small . here , we find it more convenient to fix the size of the vortex cores , and to assume that the distance @xmath50 between the centers is large .
this alternative point of view is of course equivalent to the previous one , up to a rescaling .
note , however , that the rotation speed @xmath30 will now depend on @xmath50 and behave like @xmath51 as @xmath52 . from now on ,
we fix @xmath49 , @xmath53 , and we look for a stationary solution @xmath54 of describing a pair of identical vortices with circulation @xmath55 .
we make the ansatz @xmath56 where @xmath57 , @xmath58 is a localized vorticity profile to be determined , and @xmath59 $ ] is the velocity field obtained from @xmath58 via the biot - savart law .
we assume that @xmath58 belongs to the schwartz class @xmath60 , is nonnegative , and satisfies the normalization condition @xmath61 we also impose the following symmetry @xmath62 which implies that @xmath63 for all @xmath64 .
finally , without loss of generality , we assume that @xmath65 this means that the vorticity distribution @xmath54 defined in is indeed a superposition of two localized vortices centered at the points @xmath66 .
the distribution @xmath54 will be a stationary solution of the rotating vorticity equation if the profile @xmath58 satisfies @xmath67 replacing @xmath68 by @xmath69 and denoting @xmath70 , we obtain the equivalent equation @xmath71 note that the rotating term @xmath72 behaves like the velocity field @xmath73 when @xmath74 is changed into @xmath75 ; this implies that the symmetry is indeed compatible with eq . .
to determine the rotation speed @xmath30 , we multiply both members of by @xmath74 and we integrate by parts over @xmath76 .
we easily obtain @xmath77 this identity can be simplified if we use , , and the fact that @xmath78 .
we thus arrive at the following relation , @xmath79 \,:=\ , -\frac{2}{d}\int_{{{\mathbb r}}^2}v_2(-x-2x_d)w(x){\,{\rm d}}x~,\ ] ] which determines @xmath80 as a function of @xmath58 .
as we shall see , since the vorticity profile @xmath58 is normalized by , the corresponding velocity field @xmath59 $ ] satisfies @xmath81 for any fixed @xmath82 . in view of
, this means that @xmath83 as @xmath52 .
in particular , if we take formally the limit @xmath52 in , we obtain the limiting equation @xmath84 , which simply means that @xmath58 ( or @xmath85 ) is a stationary solution of euler s equation . in
what follows , we assume that the limiting profile @xmath44 is _ radially symmetric _ and _ stable _ in the sense of arnold @xcite . roughly speaking
, this means that @xmath86 is a strictly decreasing function of @xmath87 .
given any such profile , we shall construct perturbatively a family of approximate solutions of , indexed by the parameter @xmath53 , which converge to @xmath44 as @xmath52 . under natural assumptions ,
these approximate solutions are uniquely determined by the asymptotic profile @xmath44 .
an important question that we leave open here is whether we can actually construct _ exact _ solutions of which converge to @xmath44 as @xmath52 , in which case our approximate solutions could be recovered by truncating the asymptotic expansion of the exact solutions . as was mentioned in the introduction
, it should be possible to prove the existence of such solutions , at least for a particular class of compactly supported profiles @xmath44 , by adapting the variational techniques of d. smets and j. van schaftingen @xcite .
it might also be possible to construct exact solutions of for more general profiles using a fixed point argument of nash - moser type .
we hope to clarify these issues in a future work .
let @xmath33 be a radially symmetric , nonnegative function satisfying the normalization condition , and let @xmath88 $ ] be the velocity field obtained from @xmath44 via the biot - savart law .
we can thus write @xmath89 where @xmath90 is a smooth , rapidly decreasing function and @xmath91 . note that @xmath92 as @xmath93 .
we assume that @xmath44 satisfies the following _ strong stability conditions : _
@xmath94 the first condition in implies of course that @xmath95 for all @xmath96 , so that @xmath97 for all @xmath98 . in particular
, compactly supported asymptotic profiles @xmath44 are excluded .
this condition also implies that @xmath44 is a stable solution of the two - dimensional inviscid vorticity equation , with respect to perturbations in @xmath99 @xcite .
the second assumption in is more technical in nature , and can probably be relaxed .
it is satisfied , for instance , if @xmath100 for some @xmath101 .
note that we always have @xmath102 indeed , if we assume on the contrary that @xmath103 for all @xmath98 , then the function @xmath104 is nonnegative and satisfies @xmath105 as @xmath106 , and @xmath107 as @xmath93 . thus @xmath108 , but if we integrate by parts we obtain @xmath109 which yields a contradiction .
finally , in addition to , we also assume that @xmath110 decays rapidly at infinity : @xmath111 as was already observed , the asymptotic profile @xmath44 is already an approximate solution of , in the sense that , if we substitute @xmath112 for @xmath113 in , the left - hand side converges to zero as @xmath114 .
our goal is to construct here more accurate approximations , which take into account the interaction of the vortices .
we look for solutions of the form @xmath115 where @xmath116 $ ] is the velocity field obtained from @xmath54 via the biot - savart law .
the symmetry implies that @xmath117 and in agreement with , we impose @xmath118 finally , we assume without loss of generality that @xmath54 has no radially symmetric component , namely @xmath119 we can always realize by including , if necessary , the radially symmetric part of @xmath54 into the asymptotic profile @xmath44 . in this respect , it is important to note that both conditions in are open . inserting into
, we obtain for @xmath54 the following equation @xmath120 \,=\ , 0~,\ ] ] where @xmath121 is the linearized operator defined by @xmath122 and @xmath123 $ ] is a remainder term which depends on the distance @xmath50 between the vortex centers @xmath124(x ) \,=\ , \bigl(v_*(x+2x_d ) - u(-x-2x_d ) -\tilde
\omega[w_*+\omega](x+x_d)^\perp\bigr)\cdot\nabla ( w_*(x)+\omega(x))~.\ ] ] in , it is understood that @xmath116 $ ] is the velocity field associated to @xmath54 .
we look for solutions @xmath54 of in the hilbert space @xmath125 equipped with the scalar product @xmath126 and with the associated norm @xmath127 . here
the weight @xmath128 is defined by @xmath129 the reason for this particular choice is that the linear operator @xmath121 has nice properties in the space @xmath130 , see section [ lamss ] below . in view of
, the asymptotic profile @xmath44 and all its moments belongs to @xmath130 : for any @xmath131 , we have @xmath132 it is easy to verify that the operator @xmath121 commutes with the rotations about the origin in @xmath76 , see lemma [ lampol ] below .
it is thus natural to use polar coordinates @xmath133 in the plane , and to decompose our space @xmath130 as a direct sum @xmath134 where @xmath135 is the orthogonal projection in @xmath130 defined by the formula @xmath136 in particular , @xmath137 is the subspace of all radially symmetric functions , and for @xmath138 the subspace @xmath139 contains functions of the form @xmath140 . with this notation
, condition means that @xmath141 .
we now discuss the main properties of the linearized operator @xmath121 defined in . in the particular case where @xmath44 is the profile @xmath142 of oseen s vortex ,
the operator @xmath121 was studied in detail in @xcite , and we shall obtain here analogous results in a more general situation . from we know that @xmath143 , where @xmath144 and @xmath145\cdot\nabla w_*$ ] . as is easily verified , @xmath146 is compact in @xmath130 , while @xmath147 is unbounded .
the maximal domain of @xmath121 is therefore @xmath148 the most remarkable property of this operator is that it is _ skew - symmetric _ in @xmath130 .
[ skew ] for all @xmath149 , we have @xmath150 .
* proof .
* we shall prove in fact that both operators @xmath151 are skew - symmetric . first , since the weight @xmath152 is radially symmetric , we have @xmath153 because the velocity field @xmath154 is divergence - free .
next , since @xmath155 , we have @xmath156 see e.g. ( * ? ? ?
* lemma 4.8 ) . combining both equalities
, we obtain the desired result . as in the gaussian case @xcite , the operator @xmath121 is invariant under rotations about the origin in the plane @xmath76 .
it is thus natural to work in polar coordinates @xmath133 , and to develop the vorticity @xmath157 in fourier series with respect to the angular variable @xmath158 . in these variables ,
the action of @xmath121 can be described fairly explicitly .
let @xmath159 then we have the following result : [ lampol ] fix @xmath160 .
if @xmath161 , then @xmath162\cos(n\theta)$ ] , where @xmath163 is the regular solution of the differential equation @xmath164 similarly , if @xmath165 , then @xmath162\sin(n\theta)$ ] .
* proof . * if @xmath166 , namely if @xmath54 is radially symmetric , it is straightforward to verify that @xmath167 .
thus we assume that @xmath138 , and that @xmath161 .
if @xmath168 denotes the stream function defined by @xmath169 , we have @xmath170 , where @xmath163 is the regular solution of , namely @xmath171 the velocity field @xmath116 = -\nabla^\perp\psi$ ] is thus @xmath172 since @xmath173 , we conclude that @xmath174 which , in view of , is the desired result .
the case where @xmath165 is similar . as an application of lemma [ lampol
] , we can characterize the kernel of the operator @xmath121 .
we already know that @xmath175 if @xmath54 is radially symmetric .
moreover , differentiating the identity @xmath176 with respect to @xmath177 and @xmath74 , we obtain @xmath178 . as in the gaussian case @xcite ,
we conclude : [ lamker ] @xmath179 .
* proof .
* since the decomposition is invariant under the action of @xmath121 , it is sufficient to characterize the kernel in each subspace @xmath180 .
the case @xmath166 is trivial , because @xmath181 , hence we assume from now on that @xmath138 . if @xmath161 satisfies @xmath167 , we know from lemma [ lampol ] that @xmath182 . in view of ,
this can be written in the equivalent form @xmath183 now , the second assumption in means that @xmath184 thus , if @xmath185 , the `` potential '' term @xmath186 in is positive , and since @xmath187 as @xmath106 and @xmath93 , the maximum principle implies that @xmath188 , hence also @xmath189 .
thus @xmath190 if @xmath185 . in the particular case @xmath191 , it is easy to verify that @xmath192 is the regular solution of .
using we find @xmath193 , so that @xmath194 .
similarly , @xmath195 , hence the kernel of @xmath121 in @xmath196 is spanned by the functions @xmath197 . using the same arguments as in @xcite
, one can show that the operator @xmath121 is not only skew - symmetric , but also _ skew - adjoint _ in @xmath130 .
this implies that @xmath198 , hence @xmath199 let @xmath200 we now show that @xmath201 , and we establish a semi - explicit formula for the inverse of @xmath121 on that subspace .
[ laminv ] if @xmath202 for some @xmath185 , there exists a unique @xmath203 such that @xmath204 . specifically , if @xmath205 , then @xmath206 , where @xmath207 and @xmath163 is the regular solution of the differential equation @xmath208 similarly , if @xmath209 , then @xmath210 .
* proof . * if @xmath161 , then @xmath211\cos(n\theta)$ ] by lemma [ lampol ] , where @xmath163 satisfies .
the equation we have to solve is therefore @xmath212 , which gives .
moreover , combining and , we obtain . proceeding as in ( * ? ? ?
* lemma 3.4 ) , we now show that has a unique regular solution , and we establish a representation formula . as we observed in the proof of lemma [ lamker ] ,
the `` potential '' term @xmath186 in is positive if @xmath185 .
let @xmath213 , @xmath214 be the ( unique ) solutions of the homogeneous equation such that @xmath215 by the maximum principle , the functions @xmath213 , @xmath214 are strictly monotone and linearly independent . the wronskian determinant @xmath216 satisfies @xmath217 , hence @xmath218 for some @xmath219 , and we also have @xmath220 with these notations , the unique regular solution of has the following expression : @xmath221 if @xmath222 , it is straightforward to verify that the function @xmath163 defined by is continuous and vanishes at the origin and at infinity .
moreover , we know from that @xmath223 as @xmath93 . thus , if we assume that @xmath202 , we see that the function @xmath224 defined by satisfies @xmath225 . as a consequence ,
if @xmath161 , we conclude that @xmath203 , and @xmath204 by construction .
[ none ] if @xmath191 , the conclusion of proposition [ laminv ] fails because @xmath226 for @xmath227 .
however , if @xmath228 satisfies @xmath229 for @xmath227 , one can show that there exists a unique @xmath230 such that @xmath231 .
equipped with the technical results of the previous section , we now go back to equation , which we want to solve perturbatively for large @xmath50 .
this equation can be written as @xmath232 = 0 $ ] , where @xmath233 = u\cdot\nabla\omega + r_d[\omega]$ ] . before starting the calculations
, we briefly explain why we expect to find a unique solution , under our symmetry assumptions .
first , if @xmath234 satisfies , then @xmath235 , hence @xmath54 is uniquely determined by @xmath236 .
indeed , as was already observed , means that @xmath141 .
moreover , it follows from , that @xmath237 hence @xmath235 by lemma [ lamker ] .
next , if @xmath54 and @xmath116 $ ] have the symmetries , it is straightforward to verify that the nonlinearity in satisfies @xmath233(x_1,-x_2 ) = -n_d[\omega](x_1,x_2)$ ] for all @xmath238 .
this implies that @xmath239 = 0 $ ] and @xmath240\rangle = 0 $ ] .
moreover , we have @xmath241\rangle = 0 $ ] by construction , because this is the relation we imposed to determine the angular speed @xmath80 in .
thus , we see that @xmath233 \in { \mathrm{ker}}(\lambda)^\perp$ ] , and if we can prove in addition that @xmath242 \in x$ ] , then proposition [ laminv ] ( and remark [ none ] ) will imply that @xmath233 \in { \mathrm{ran}}(\lambda)$ ] .
we can therefore hope to find a unique @xmath243 such that @xmath232 = 0 $ ] . to begin our perturbative approach , we compute the remainder term for @xmath244 , namely @xmath245(x ) \,=\ , \bigl(v_*(x+2x_d ) - \tilde \omega[w _ * ] ( x+x_d)^\perp \bigr)\cdot\nabla w_*(x)~ , \qquad x \in { { \mathbb r}}^2~.\ ] ] from , we know that @xmath246 by assumption , the term @xmath247 decays faster than any inverse power of @xmath87 as @xmath248 , hence we can neglect its contribution in our calculations . for any fixed @xmath82 we thus have @xmath249 where @xmath250 setting @xmath251 , @xmath252 , and proceeding as in ( * ? ? ?
* lemma 3.2 ) , we find @xmath253 in particular , returning to and using definition , we obtain @xmath254 \,=\ , \frac{2}{d}\int_{{{\mathbb r}}^2}(v_*)_2(x+2x_d)w_*(x ) { \,{\rm d}}x \,=\ , \frac{1}{\pi d^2 } + { { \cal o}}\bigl(\frac{1}{d^\infty}\bigr)~ , \qquad \hbox{as}\quad d \to \infty~.\ ] ] note that the term @xmath255 in gives no contribution to the angular velocity @xmath256 $ ] . on the other hand , inserting , into and using the expansion together with the relation @xmath257 , where @xmath258 is defined in , we find for @xmath251 : @xmath259 motivated by this result , we now construct inductively an approximate solution of of the form @xmath260 where each velocity profile @xmath261 is obtained from @xmath262 via the biot - savart law .
the order @xmath263 of the approximation is in principle arbitrary , but the complexity of the calculations increases rapidly with @xmath263 , and we shall restrict ourselves to @xmath264 for simplicity .
of course , we assume that the symmetry and normalization conditions hold at each order of the approximation . in particular , we have @xmath265 for all @xmath266 . in view of ( * ? ? ? * appendix b ) , this implies that the velocity field @xmath267 decays at least as fast as @xmath268 when @xmath248 .
it follows that the term @xmath269 in is @xmath270 as @xmath52 , and will therefore not contribute to @xmath262 for @xmath271 . for the same reason , @xmath272 \,=\ , \frac{2}{d}\int_{{{\mathbb r}}^2 } \bigl((v_*)_2(x+2x_d ) - u_2(-x-2x_d)\bigr)\bigl(w_*(x ) + \omega(x)\bigr){\,{\rm d}}x \,=\ , \frac{1}{\pi d^2 } + { { \cal o}}\bigl(\frac{1 } { d^6}\bigr)~,\ ] ] as @xmath52 .
indeed , the leading term @xmath273 $ ] was computed in , and we know that the contribution of @xmath274 is negligible . moreover , using , , and , it is easy to verify that @xmath275 , as @xmath114 . summarizing , we have shown that @xmath276(x ) \,&=\ , \bigl(v_*(x+2x_d ) -\tilde \omega[w _ * ] ( x+x_d)^\perp\bigr)\cdot\nabla ( w_*(x)+\omega(x ) ) + { { \cal o}}\bigl(\frac{1 } { d^5}\bigr ) \\
\label{rdtrunc } \,&=\ , r_d(x ) + \bigl(\frac{1}{2\pi}v(x,2x_d ) - \frac{x^\perp}{\pi d^2}\bigr)\cdot\nabla \omega(x ) + { { \cal o}}\bigl(\frac{1}{d^5}\bigr)~ , \end{aligned}\ ] ] as @xmath52 .
similarly , the quadratic term @xmath277 in satisfies @xmath278 it is now a straightforward task to determine the first vorticity profiles in the expansion . from , , we know that the nonlinearity @xmath233 = u\cdot\nabla\omega + r_d[\omega]$ ] in satisfies , for @xmath279 , @xmath280(x ) \,=\ , \frac{g(r)}{2\pi}\left(\frac{r^2}{d^2 } \sin(2\theta ) - \frac{r^3}{d^3}\sin(3\theta)\right ) + { { \cal o}}\bigl(\frac{1}{d^4}\bigr)~ , \qquad \hbox{as } \quad d \to \infty~.\ ] ] thus , to ensure that @xmath232 = { { \cal o}}(d^{-4})$ ] , we must impose @xmath281 by proposition [ laminv ] , eq .
has a unique solution @xmath282 of the form @xmath283 where @xmath284 are given by , with @xmath285 .
as is easily verified , the symmetry conditions , are satisfied by the profiles @xmath286 for @xmath287 , and the velocity @xmath288 decays like @xmath289 as @xmath248 .
computing the profiles @xmath290 is more cumbersome , but also more representative of what happens in the general case .
first of all , the quadratic term is no longer negligible , and using for @xmath291 we find @xmath292 note that , to ensure that @xmath293 , we need an assumption on the second derivative of the function @xmath294 appearing in .
for instance , in analogy with , one can impose @xmath295 next , we must compute the contribution of @xmath296 to the right - hand side of . using , ,
we obtain @xmath297 where @xmath298 combining these results , we find instead of : @xmath299(x ) \,=\ , \frac{g(r)}{2\pi}\sum_{n=2}^3 ( -1)^n \frac{r^n}{d^n}\sin(n\theta ) + \frac{b(r)}{d^4 } \sin(4\theta ) + \frac{c(r)}{d^4}\sin(2\theta ) + { { \cal o}}\bigl(\frac{1}{d^5}\bigr)~,\ ] ] as @xmath52 , where @xmath300 .
therefore , in addition to , we must impose @xmath301 using again proposition [ laminv ] , we see that has a unique solution @xmath302 of the form @xmath303 , where @xmath304 is given by , with @xmath305 and @xmath306 , while @xmath307 is given by the same relations with @xmath291 and @xmath308
. an explicit expression of the velocity profile @xmath309 can also be obtained , as in . summarizing ,
we have shown : [ pexp ] let @xmath44 be a radially symmetric vorticity profile of the form , where the function @xmath294 satisfies , , , and let @xmath310 where the vorticity profiles @xmath311 satisfy , , and the velocity profiles @xmath261 are obtained by the biot - savart law .
then @xmath58 is an asymptotic solution of eqs .
, in the sense that @xmath312(x+x_d)^\perp\bigr ) \cdot \nabla w(x ) \,=\ , { { \cal o}}\bigl(\frac{1}{d^5}\bigr)~,\ ] ] in the topology of @xmath130 and uniformly on @xmath76 , as @xmath52 .
the asymptotic expansion is very natural , and it is clear that it can be performed to any finite order @xmath313 if we make appropriate assumptions on the derivatives of the profile @xmath294 , as in .
as was already mentioned , we also conjecture that there exists an _
exact _ solution of for @xmath53 which coincides with up to corrections of order @xmath270 .
it is important to notice that , under the symmetry and normalization conditions , the exact solution ( if it exists ) and the asymptotic expansion are uniquely determined by the limiting profile @xmath44 . [ moregen ] rather strong assumptions on the limiting profile @xmath44 were made in this section to ensure that the asymptotic expansion holds in the function space @xmath130 defined in , which we believe is naturally associated to the problem .
this does not restrict the scope of our results here , because these conditions are automatically fulfilled by the gaussian profiles created by the navier - stokes evolution .
however , within the framework of euler s equation , it is certainly interesting to construct interacting vortex pairs with more general profiles , including compactly supported ones . if we do not insist on controlling our expansion in the space @xmath130 , the calculations presented in this section show that the assumptions on the function @xmath294 can be considerably relaxed .
the most important point is that eq .
should have a unique solution for @xmath314 , and for @xmath191 if the right - hand side satisfies some orthogonality conditions .
this is definitely the case if the second inequality in holds , but that condition does not imply that @xmath294 is strictly decreasing and can well be satisfied if @xmath294 is compactly supported . to conclude this section ,
we indicate how approximate solutions of of the form can be obtained from proposition [ pexp ] by a simple rescaling .
given @xmath315 , we consider the situation described in with @xmath316 and @xmath317 . if @xmath44 is a radially symmetric vorticity profile satisfying the assumptions of proposition [ pexp ] , we define , for all sufficiently small @xmath35 , @xmath318 where @xmath286 are as in
. then , by construction , the vorticity distribution @xmath319 where @xmath57 , is an approximate solution of eq . with @xmath320 .
more precisely , it follows from that @xmath321 where @xmath322 is obtained from @xmath47 via the biot - savart law .
in this final section , we describe in some detail the result of @xcite in the particular case of a vortex pair , and we interprete it using the approximate solutions of euler s equation constructed in section [ sec2 ] .
given @xmath323 , we set @xmath324 and we denote by @xmath4 the vorticity distribution where @xmath316 and @xmath317 . for any @xmath325 , we consider the ( unique ) solution @xmath45 of the rotating viscous vorticity equation @xmath326 with initial data @xmath4 .
up to a rotation of angle @xmath327 , the vorticity distribution @xmath45 coincides with the solution of the nonrotating equation with the same initial data , which is studied in @xcite .
the advantage of using a rotating frame is that the vortex centers remain fixed , instead of evolving according to the point vortex dynamics . as a matter of fact
, theorem 2.1 in @xcite establishes that @xmath328 as @xmath329 , for any @xmath330 . to obtain a more precise convergence result
, we decompose the solution of into a sum of viscous vortices : @xmath331 where @xmath332 . as is shown in @xcite , both vorticity profiles @xmath333 can be approximated by the same function @xmath334 where @xmath142 is the gaussian profile and the first order correction @xmath335 is constructed as follows .
let @xmath336 be the fokker - planck operator @xmath337 and @xmath121 be the linearized operator with @xmath48 , namely @xmath338\cdot\nabla g~.\ ] ] here we use the functional setting of section [ sec2 ] in the particular case where the asymptotic profile @xmath44 is the oseen vortex @xmath142 .
this means that @xmath339 in , and the assumptions , , are clearly satisfied . with these notations ,
the profile @xmath335 is the unique solution of the linear equation @xmath340 where @xmath341 , see ( * ? ? ?
* section 3.3 ) . in polar coordinates @xmath342 , we thus have @xmath343 where @xmath258 is defined in .
in particular , if we use the angular decomposition of the function space , we see that @xmath344 , and it follows that @xmath345 too .
now , setting @xmath346 in , we obtain the simple equation @xmath347 , which coincides with for @xmath291 . since @xmath348
, we conclude that @xmath349 .
it is clear from that the actual profile @xmath335 is close to @xmath350 if the viscosity @xmath13 is small compared to the circulation @xmath55 of the vortices . as a matter of fact
, it is shown in ( * ? ? ?
* lemma 3.5 ) that @xmath351 to formulate our main approximation result , we introduce a function space with a weaker norm than @xmath130 .
given any @xmath352 , we denote by @xmath353 the space @xmath354 applying theorem 2.5 of @xcite to the particular situation considered here , we obtain [ omapprox ] fix @xmath355 , and let @xmath45 be the solution of the rotating viscous vorticity equation with initial data @xmath4 .
there exist positive constants @xmath356 , depending only on the product @xmath357 , such that , if @xmath45 is decomposed as in , then the vorticity profiles @xmath358 satisfy @xmath359 for all @xmath360 $ ] , where @xmath361 is given by .
this result can be reformulated in a slightly different way , using the approximate solutions of euler s equation constructed in section [ sec2 ] . indeed , if we set @xmath2 , and if we remember that @xmath48 in the present case , we see that the inviscid profile @xmath362 defined in satisfies @xmath363 thus , combining and , we obtain : finally , using the continuous inclusion @xmath365 , we can formulate an approximation result for the original function @xmath45 .
if we compare , and use the fact that @xmath366 and @xmath367 are even functions of @xmath368 , we arrive at : comparing this last result with theorem [ thm1 ] , we see that replacing a linear superposition of oseen vortices by a more accurate solution of euler s equation , which takes into account the deformation of the vortex cores due to mutual interaction , results in a better approximation .
if we believe that there exists an _
exact _ solution of euler s equation which is close to @xmath47 for @xmath0 sufficiently small , then shows that the solution @xmath45 of the rotating viscous vorticity equation slowly travels through a family of steady states of the inviscid equation indexed by the length parameter @xmath0 , which evolves diffusively according to @xmath370 .
[ final ] the right - hand side of suggests that both dimensionless quantities @xmath371 and @xmath372 play an important role in the evolution of viscous vortices .
it is tempting to eliminate one of these quantities by considering , for instance , the limit @xmath329 while @xmath2 is kept fixed . under the assumptions of proposition [ omapprox ]
, one may conjecture that the solution @xmath45 of satisfies , if @xmath35 is sufficiently small , @xmath373{}~ \omega_\epsilon(x)~ , \qquad \hbox{for all~ } x \in { { \mathbb r}}^2~,\ ] ] where @xmath374 is an exact stationary solution of of the form .
unfortunately , the results of @xcite do not provide any control on @xmath45 in the limit where @xmath375 . | the aim of this contribution is to make a connection between two recent results concerning the dynamics of vortices in incompressible planar flows .
the first one is an asymptotic expansion , in the vanishing viscosity limit , of the solution of the two - dimensional navier - stokes equation with point vortices as initial data .
in such a situation , it is known @xcite that the solution behaves to leading order like a linear superposition of oseen vortices whose centers evolve according to the point vortex system , but higher order corrections can also be computed which describe the deformation of the vortex cores due to mutual interactions . the second result is the construction by d. smets and j. van schaftingen of `` desingularized '' solutions of the two - dimensional euler equation @xcite .
these solutions are stationary in a uniformly rotating or translating frame , and converge either to a single vortex or to a vortex pair as the size parameter @xmath0 goes to zero .
we consider here the particular case of a pair of identical vortices , and we show that the solution of the weakly viscous navier - stokes equation is accurately described at time @xmath1 by an approximate steady state of the rotating euler equation which is a desingularized solution in the sense of @xcite with gaussian profile and size @xmath2 . |
this paper concerns a category - theoretic question which arose in a model - theoretic context . in model theory , specifically in the construction of frass limits ( see @xcite ) , one considers a class @xmath2 of structures ( in some first - order language ) with the following properties : * the : ( @xmath2 is nonempty and ) for every two structures @xmath3 , there are embeddings @xmath4 and @xmath5 into some @xmath6 : @xmath7 & x \\ a
\urar[dashed]{f } & & b \ular[dashed][swap]{g } \end{tikzcd};\ ] ] * the : every diagram of embeddings @xmath7 b & & c \\ & a \ular{f } \urar[swap]{g } \end{tikzcd}\ ] ] between structures @xmath8 can be completed into a commutative diagram @xmath7 & x \\ b \urar[dashed]{h } & & c \ular[dashed][swap]{k } \\ & a \ular{f } \urar[swap]{g } \end{tikzcd}\ ] ] for some structure @xmath6 and embeddings @xmath9 and @xmath10 .
common examples of classes @xmath2 with these properties include : finitely generated groups ; posets ; nontrivial boolean algebras ; finite fields of fixed characteristic @xmath11 . from the ap ( and optionally
the jep ) , one has various `` generalized amalgamation properties '' , whereby more complicated diagrams ( of embeddings ) can be completed into commutative diagrams ( of embeddings ) , e.g. , the following diagram by two uses of ap : @xmath7 & & y \\ & x \urar[dashed ] \\ a \urar[dashed ] & & c \ular[dashed ] & & e \ar[dashed]{uull } \\ & b \ular \urar & & d \ular \urar \end{tikzcd}.\ ] ] however , the following diagram can not be completed using just the ap ( and the jep ) : @xmath12 a & ~ & b \\
c \uar{f } \ar{urr}[pos=.8]{h } & & d \ar[crossing over]{ull}[pos=.8,swap]{g } \uar[swap]{k } \end{tikzcd}.\ ] ] for example , take @xmath13 the class of finite sets , @xmath14 , @xmath15 , and @xmath16 .
this leads to the following question can we characterize the shapes of the diagrams which can always be completed using the ap , i.e. , the `` generalized amalgamation properties '' which are implied by the ap ?
if so , is such a characterization decidable ?
this question concerns only abstract properties of diagrams and arrows , hence is naturally phrased in the language of category theory .
let @xmath17 be a category . recall that
a over a diagram in @xmath17 consists of an object @xmath18 , together with morphisms @xmath19 in @xmath17 for each object @xmath20 in the diagram , such that the morphisms @xmath21 commute with the morphisms in the diagram ; this is formally what it means to `` complete '' a diagram . recall also that a of a diagram is a universal cocone , i.e. , one which admits a unique morphism to any other cocone .
( see [ sec : prelims ] for the precise definitions . )
we say that @xmath17 has the ap if every pushout diagram ( i.e. , diagram of shape @xmath22 ) in @xmath17 has a cocone , and that @xmath17 has the jep is monic ( although see [ sec : general ] ) . ] if every diagram in @xmath17 consisting of finitely many objects ( without any arrows ) has a cocone . when @xmath17 is the category of structures in a class @xmath2 and embeddings between them , this recovers the model - theoretic notions defined above .
category - theoretic questions arising in frass theory have been considered previously in the literature ; see e.g. , @xcite , @xcite .
the possibility of answering the above question in the generality of an arbitrary category is suggested by an analogous result of par @xcite ( see [ thm : pushout ] below ) , which characterizes the diagram shapes over which a colimit may be built by pushouts ( i.e. , colimits of pushout diagrams ) .
there , the crucial condition is that the diagram shape must be ( see [ defn : pi1 ] ) ; failure to be simply - connected is witnessed by the of the diagram shape , whose morphisms are `` paths up to homotopy '' . for example , the fundamental groupoid of the shape of is equivalent to @xmath23 , with generator given by the `` loop '' @xmath24 .
however , simply - connectedness of a diagram s shape does not guarantee that a cocone over it may be built using only the ap ( see [ ex : boat ] below ) . intuitively , the discrepancy with par s result is because the universal property of a pushout allows it to be used in more ways to build further cocones .
simply - connectedness nonetheless plays a role in the following characterization , which is the main result of this paper : [ thm : amalgam ] let @xmath0 be a finitely generated category .
the following are equivalent : 1 .
every @xmath0-shaped diagram in a category with the ap and the jep has a cocone .
every @xmath0-shaped diagram in the category of sets and injections has a cocone .
( when @xmath0 is finite , it suffices to consider finite sets . ) 3 .
@xmath0 is upward - simply - connected ( see [ defn : l ] ) .
when @xmath0 is a finite poset , these are further equivalent to : 1 .
every upward - closed subset of @xmath0 is simply - connected .
2 . @xmath0 is forest - like ( see [ defn : tree ] ; this means @xmath0 is built via some simple inductive rules ) .
similarly , every @xmath0-shaped diagram in a category with the ap has a cocone , iff @xmath0 is connected and any / all of ( ii ) , ( iii ) ( also ( iv ) , ( v ) if @xmath0 is a poset ) hold .
a corollary of our proof yields a simple decision procedure for these conditions ( for finite @xmath0 ) .
this is somewhat surprising , because par s result ( [ thm : pushout ] ) implies that the analogous question of whether every @xmath0-shaped diagram has a colimit in a category with pushouts is _ _
un__decidable .
this paper is organized as follows . in [ sec : prelims ] , we fix notations and review some categorical concepts . in [ sec : setinj ] , we introduce an invariant @xmath1 , similar to the fundamental groupoid , and use it to prove the equivalence of ( ii ) and ( iii ) in [ thm : amalgam ] for arbitrary small ( not necessarily finitely generated ) @xmath0 . in [ sec : posetal ] , we analyze upward - simply - connected posets in more detail , deriving the conditions ( iv ) and ( v ) equivalent to ( iii ) and proving that they imply ( i ) when @xmath0 is a finite poset . in [ sec : general ] , we remove this restriction on @xmath0 and complete the proof . finally , in [ sec : decidability ]
, we discuss decidability of the equivalent conditions in [ thm : amalgam ] and of the analogous conditions in par s result . _ acknowledgments .
_ we would like to thank alexander kechris for providing some feedback on an earlier draft of this paper .
we begin by fixing notations and terminology for some basic categorical notions ; see @xcite . for a category @xmath17 and objects @xmath25
, we denote a morphism between them by @xmath26 { } y$ ] .
we use the terms and interchangeably .
we use @xmath27 to denote the category of sets and functions , @xmath28 to denote the category of sets and injections , and @xmath29 to denote the category of sets and partial injections .
we use @xmath30 to denote the categories of small categories , resp . , small groupoids .
we regard a preordered set @xmath31 as a category where there is a unique arrow @xmath32 j$ ] iff @xmath33 .
we say that a category @xmath17 is if every morphism @xmath26 { } y$ ] in it is monic ( i.e. , if @xmath34 then @xmath35 , for all @xmath36 { } x$ ] ) .
similarly , @xmath17 is if every endomorphism @xmath26 { } x$ ] is idempotent ( i.e. , @xmath37 ) .
a category @xmath0 is if there are finitely many arrows in @xmath0 whose closure under composition is all arrows in @xmath0 .
note that such @xmath0 necessarily has finitely many objects , and that a preorder is finitely generated iff it is finite .
for a category @xmath17 and a small category @xmath0 , a of shape @xmath0 in @xmath17 is simply a functor @xmath38 .
a @xmath39 over a diagram @xmath40 consists of an object @xmath18 together with a family of morphisms @xmath41 { } x)_{i \in \*i}$ ] , such that for each @xmath42 { } j$ ] , we have @xmath43 .
a @xmath39 and @xmath44 over the diagram @xmath40 is a morphism @xmath45 { } y$ ] such that for each object @xmath46 , we have @xmath47 . a cocone @xmath39 over @xmath40 is a of @xmath40 if it is initial in the category of cocones over @xmath40 , i.e. , for any other cocone @xmath44 there is a unique cocone morphism @xmath48 ; in this case we write @xmath49 , and usually use a letter like @xmath50 for the cocone maps @xmath51 .
@xmath52{f_i } \ar{drr}[swap , pos=.4]{g_i } & f(j ) \dar[crossing over][pos=.3]{f_j } \drar{g_j } \\[2em ] & x = \injlim f \rar[dashed][swap]{h } & y \end{tikzcd}\ ] ] as mentioned above , a category @xmath17 has the if every ( i.e. , diagram of shape @xmath22 ) in @xmath17 has a cocone ( colimits of such diagrams are called ) , and @xmath17 has the ( regardless of whether @xmath17 is monic ) if every finite coproduct diagram ( i.e. , diagram of finite discrete shape ) in @xmath17 has a cocone .
( so the empty category does not have the jep . )
a category @xmath17 is if it has exactly one , where @xmath25 are in the same connected component if they are joined by a zigzag of morphisms @xmath53 ( so the empty category is not connected . )
we use @xmath54 to denote the set ( or class , if @xmath17 is large ) of connected components of @xmath17 .
note that in the presence of the ap , connectedness is equivalent to the jep , since the ap may be used to turn `` troughs '' into `` peaks '' in a zigzag .
[ defn : pi1 ] the of a category @xmath0 , denoted @xmath55 , is the groupoid freely generated by @xmath0 ( as a category ) .
thus @xmath55 has the same objects as @xmath0 , while its morphisms are words made up of the morphisms in @xmath0 together with their formal inverses , modulo the relations which hold in @xmath0 ( and the relations which say that the formal inverses are inverses ) .
we say that @xmath0 is if @xmath55 is an equivalence relation , i.e. , has at most one morphism between any two objects .
[ rmk : nerve ] there is also a topological definition : @xmath55 is the same as the fundamental groupoid of the ( simplicial ) nerve of @xmath0 ; see @xcite for the general case .
when @xmath0 is a poset , the of @xmath0 can be defined as the ( abstract ) simplicial complex whose @xmath56-simplices are the chains of cardinality @xmath57 in @xmath0 ; see ( * ? ? ?
* , 2.4 ) ( in which the nerve is called the _ order complex _ ) .
we now state par s result @xcite , mentioned in the introduction , characterizing colimits which can be built using pushouts : [ thm : pushout ] let @xmath0 be a finitely generated category .
the following are equivalent : 1 .
every @xmath0-shaped diagram in a category with pushouts has a colimit .
@xmath0 is simply - connected .
[ ex : boat ] let @xmath0 be the shape of the diagram in the introduction . as mentioned there , @xmath55 is equivalent to @xmath23 ( i.e. , it is connected and its automorphism group at each object is @xmath23 ) .
now let @xmath58 be the shape of the ( commuting ) diagram @xmath59 a & & b \\ c \uar{f } \ar{urr}[pos=.8]{h } & & d \ar[crossing over]{ull}[pos=.8,swap]{g } \uar[swap]{k } \\ & e \ular{u } \urar[swap]{v } \end{tikzcd}.\ ] ] unlike @xmath0 , @xmath60 is simply - connected .
thus by [ thm : pushout ] , a colimit of a @xmath60-shaped diagram can be constructed out of pushouts .
however , a cocone over a @xmath60-shaped diagram can not necessarily be constructed from the ap : take @xmath14 , @xmath15 , and @xmath16 in @xmath28 as in the introduction , and @xmath61 .
note that there is no contradiction , since @xmath28 does _ not _ have pushouts .
the `` reason '' that @xmath60 is simply - connected even though @xmath0 is not is that the generating `` loop '' @xmath62 in @xmath55 becomes trivial in @xmath63 : @xmath64 this suggests that to characterize when a @xmath60-shaped diagram has a cocone in any category with the ap , we need a finer invariant than @xmath65 , which does not allow the use of @xmath66 to simplify the loop @xmath62 above .
an is a category @xmath17 such that every morphism @xmath26 { } y$ ] has a unique @xmath67 { } x$ ] obeying @xmath68 and @xmath69 .
we write @xmath70 for the category of small inverse categories . for basic properties of inverse categories ,
see e.g. , @xcite ; the one - object case of _ inverse monoids _ is well - known in semigroup theory @xcite .
here are some elementary facts about inverse categories we will use without mention : let @xmath17 be an inverse category .
* @xmath71 is an involutive functor @xmath72 .
* idempotents in @xmath17 commute . *
@xmath26 { } y$ ] is monic iff it is split monic iff @xmath73 .
the archetypical example of an inverse category is @xmath29 , the category of sets and partial injections ( where @xmath74 is given by the partial inverse of @xmath75 ) .
in fact , the axioms of an inverse category capture precisely the algebraic properties of @xmath29 , in the sense that we have the following representation theorem , generalizing the wagner - preston representation theorem for inverse semigroups and the yoneda lemma ( see @xcite , @xcite ) : [ thm : invcat - yoneda ] let @xmath17 be a small inverse category .
we have an embedding functor @xmath76{f } y ) & |-- > \left(\begin{aligned } \sum_z f^{-1 } \circ f \circ \operatorname{hom}_\*c(z , x ) & \overset{\sim}{-- > } \sum_z f \circ
f^{-1 } \circ \operatorname{hom}_\*c(z , y ) \\ g & |-- > f \circ g \end{aligned}\right).\end{aligned}\ ] ] here @xmath77 denotes disjoint union , and @xmath78 denotes the set of all composites @xmath79 for @xmath80 { } x$ ] , equivalently the set of all @xmath80 { } x$ ] such that @xmath81 ( and similarly for @xmath82 ) .
in this section , we characterize the small categories @xmath0 such that every @xmath0-shaped diagram in @xmath28 has a cocone .
we begin with the following easy observation : [ thm : setinj - colim ] a diagram @xmath83 has a cocone iff the colimit of @xmath40 in @xmath27 is such that the canonical maps @xmath84 are injective , for all @xmath46 .
if the @xmath50 are injective , then @xmath85 is a cocone in @xmath28 .
conversely , if @xmath40 has a cocone @xmath86 in @xmath28 , then the unique cocone morphism @xmath87 is such that @xmath88 is injective for each @xmath89 , hence @xmath50 is injective for each @xmath89 .
now recall that for a diagram @xmath83 , the standard construction of @xmath90 in @xmath27 is as the quotient of the disjoint sum : @xmath91 { } j,\ , x \in f(i)\}.\end{aligned}\ ] ] two elements @xmath92 are thus identified iff they are connected by a zigzag @xmath93 & x_1 \mathrlap{{}\in f(i_1 ) } & & x_3 \mathrlap{{}\in f(i_3 ) } & ~ & x_{2n-1 } \mathrlap{{}\in f(i_{2n-1 } ) } \\ x \mathrlap{{}\in f(i ) } \urar[mapsto]{f(i_1 ) } & & x_2 \mathrlap{{}\in f(i_2 ) } \ular[mapsto][swap]{f(i_2 ) } \urar[mapsto]{f(i_3 ) } & & \hspace{3em } \uar[draw = none][pos=.5,anchor = center]{\textstyle\dotsm } & & y \in f(i ) \ular[mapsto][swap]{f(i_{2n})}. \end{tikzcd}\ ] ] since the @xmath94 are injective , the endpoint @xmath95 of this zigzag is determined by @xmath96 together with the `` path '' @xmath32{i_1 } i_1 < ---[\*i]{i_2 } i_2 --->[\*i]{i_2 } \dotsb < ---[\*i]{i_{2n } } i$ ] in @xmath89 ; in other words , @xmath40 induces an action of such `` paths '' via partial injections between sets .
this motivates defining the category of such `` paths '' , while keeping in mind that they will be acting via partial injections : [ defn : l ] the of a category @xmath0 , denoted @xmath1 , is the inverse category freely generated by @xmath0 such that every morphism in @xmath0 becomes monic in @xmath1 .
thus , we have a functor @xmath97 , such that @xmath1 is generated by the morphisms @xmath98 for @xmath42 { } j$ ] , such that @xmath99 for each such @xmath100 , and such that any other functor @xmath38 into an inverse category with @xmath101 for each @xmath100 factors uniquely through @xmath102 .
( we write @xmath100 for @xmath103 when there is no risk of confusion . )
we say that @xmath0 is if @xmath1 is idempotent .
( see [ thm : usc - etc ] below for equivalent conditions when @xmath0 is a poset . )
thus @xmath104 extends in an obvious manner to a functor @xmath105 , which is left adjoint to the functor @xmath106 taking an inverse category to its subcategory of monomorphisms . to see that @xmath1 is a finer invariant than @xmath55 , note that the forgetful functor @xmath107 ( to which @xmath65 is left adjoint ) factors through @xmath108 ; indeed , a groupoid is precisely an inverse category in which every morphism is monic .
thus @xmath55 is a quotient of @xmath1 ( by the least congruence which makes every arrow monic ) .
since idempotents in a groupoid are identities , we get [ thm : usc - sc ] if a category @xmath0 is upward - simply - connected , then it is simply - connected .
[ thm : setinj - amalgam ] let @xmath0 be a small category .
the following are equivalent : 1 . every diagram @xmath83 has a cocone .
the diagram @xmath109 has a cocone , where @xmath110 is the embedding from [ thm : invcat - yoneda ] ( whose restriction along @xmath111 lands in the subcategory @xmath28 ) .
@xmath0 is upward - simply - connected .
first we remark on why @xmath112 in ( ii ) lands in @xmath28 .
this is because every morphism @xmath100 in @xmath0 becomes monic in @xmath1 , hence in @xmath29 ; and the monomorphisms in @xmath29 are precisely the total injections .
( i)@xmath113(ii ) is obvious .
( ii)@xmath113(iii ) : let @xmath114 { } i$ ] be an endomorphism . to show that @xmath75 is idempotent , it suffices to show that @xmath115 , since then @xmath116 .
since @xmath1 is generated by the morphisms in @xmath0 and their pseudoinverses , we have @xmath117 for some zigzag `` path '' in @xmath0 @xmath118 & i_1 & & i_3 & ~ & i_{2n-1 } \\
i = i_0 \urar{i_1 } & & i_2 \ular[swap]{i_2 } \urar{i_3 } & & \uar[draw = none][pos=.5,anchor = center]{\textstyle\dotsm } & & i_{2n } = i. \ular[swap]{i_{2n } } \end{tikzcd}\ ] ] this yields a zigzag ( with some obvious abbreviations for clarity ) @xmath59 & i_1 f^{-1 } f \mathrlap{{}\in \psi(i_1 ) } & & i_3 i_2^{-1 } i_1 f^{-1 } f \mathrlap{{}\in \psi(i_3 ) } & ~ & i_{2n-1 } \dotsm i_2^{-1 } i_1 f^{-1 } f \mathrlap{{}\in \psi(i_{2n-1 } ) } \\ f^{-1 } f \mathrlap{{}\in \psi(i ) } \urar[mapsto]{i_1 } & & i_2^{-1 } i_1 f^{-1 } f \mathrlap{{}\in \psi(i_2 ) } \ular[mapsto][swap]{i_2 } \urar[mapsto]{i_3 } & & \hspace{2em } \uar[draw = none][pos=.5,anchor = center]{\textstyle\dotsm } & & f f^{-1 } f \in \psi(i ) \ular[mapsto][swap]{i_{2n } } \end{tikzcd}\ ] ] where the even - numbered mappings are by the following calculation : @xmath119 by the discussion following [ thm : setinj - colim ] , this last zigzag implies that @xmath120 in @xmath121 , whence by ( ii ) and [ thm : setinj - colim ] , the same equality holds in @xmath122 and hence in @xmath1 , as desired .
( iii)@xmath113(i ) : let @xmath83 be a diagram . by the universal property of @xmath1 , @xmath40 extends along @xmath102 to a functor @xmath123 such that @xmath124 .
this functor @xmath125 takes a pseudoinverse @xmath126 { } i$ ] , for @xmath42 { } j$ ] , to the ( partial ) inverse of @xmath127 , hence takes a `` path '' @xmath128 { } i$ ] to the partial injection @xmath129 { } f(i)$ ] mapping @xmath130 to the endpoint @xmath95 of the zigzag in the remarks following [ thm : setinj - colim ] . since every @xmath114 { } i$ ] is idempotent ,
so is every @xmath131 . since the idempotents in @xmath29 are precisely the partial identity functions , it follows ( by the remarks following [ thm : setinj - colim ] ) that the canonical maps @xmath84 are injective .
this proves ( ii)@xmath132(iii ) in [ thm : amalgam ] , for arbitrary small @xmath0 ; the parenthetical in ( ii ) follows from let @xmath0 be a finite category . if a diagram @xmath83 does not have a cocone , then some diagram @xmath133 which is a pointwise restriction of @xmath40 to finite subsets also does not have a cocone . if @xmath40 does not have a cocone , then there is some @xmath46 and @xmath134 which are identified in @xmath90 .
take a zigzag @xmath135 witnessing that @xmath136 are identified in @xmath90 , as in the remarks following [ thm : setinj - colim ] , where @xmath137 .
put @xmath138 { } j\}$ ] .
in this section , we prove that ( iii)@xmath113(i ) in [ thm : amalgam ] for finite posets @xmath0 .
to do so , we first examine the structure of upward - simply - connected posets ; this will lead to the conditions ( iv ) and ( v ) in [ thm : amalgam ] .
[ thm : l - cosieve - faithful ] let @xmath0 be any category and @xmath139 be an upward - closed subcategory ( or ) , i.e. , a full subcategory such that if @xmath140 and @xmath42 { } j$ ] then @xmath141
. then the canonical induced functor @xmath142 is faithful .
let @xmath143 be the inverse category obtained by taking @xmath144 , adding the objects in @xmath145 , and freely adjoining zero morphisms between every pair of objects ( taking the zero morphism from an object in @xmath145 to itself as the identity ) .
let @xmath146 send morphisms @xmath100 in @xmath60 to @xmath147 and all other morphisms to @xmath148 .
then @xmath40 sends all morphisms to monomorphisms , hence extends along @xmath111 to @xmath149 .
@xmath150{\eta_\*j } \rar[hook ] & \*i \dar[swap]{\eta_\*i } \ar{ddr}{f } \\
\@l(\*j ) \ar[hook , bend right=15]{drr } \rar & \@l(\*i ) \drar[dashed][swap , pos=.2]{\~f } \\ & & \*k \end{tikzcd}\ ] ] the composite @xmath151 is equal to the inclusion @xmath152 , because it takes morphisms @xmath147 for @xmath100 in @xmath60 to @xmath153 .
it follows that @xmath142 must be faithful , as desired .
[ thm : usc - cosieve - sc ] if @xmath0 is upward - simply - connected , then every upward - closed subcategory @xmath139 is simply - connected . by [ thm : l - cosieve - faithful ] and [ thm : usc - sc ] .
we will show that for finite posets @xmath0 , the two conditions in [ thm : usc - cosieve - sc ] are equivalent to each other and to the following combinatorial notion : [ defn : tree ] the class of finite posets is defined inductively by the following rule : 1 .
if @xmath154 are finite tree - like posets , and @xmath155 is a connected upward - closed subset for each @xmath156 , then a new finite tree - like poset is formed by taking the disjoint union of the @xmath157 s and adjoining a single point which lies below each @xmath158 . a finite poset is a finite disjoint union of tree - like posets .
[ thm : posetal - cosieve - sc - decomp ] let @xmath0 be a finite connected simply - connected poset and @xmath46 be minimal . then for each connected component @xmath159 , the subposet @xmath160 ( where @xmath161 ) is connected
this is straightforward to show using topological arguments , by considering the nerve of @xmath0 as in [ rmk : nerve ] ; however , in the interest of keeping this paper self - contained , we will give a more elementary proof .
let @xmath162 and @xmath163 .
note that we must have @xmath164 , since @xmath165 and @xmath0 is connected .
consider the partition of connected subposets @xmath166 of @xmath0 , ordered by @xmath167 for @xmath168 .
then @xmath169 is connected ( because @xmath0 is ) , @xmath170 is connected ( because @xmath143 is ) , and we have two antichains @xmath171 and @xmath172 , with only elements of the former below elements of the latter ( and @xmath173 below every element of the latter ) .
so the hasse diagram of @xmath169 looks like @xmath174 & \*u_1 \dar & \*u_2 \dlar \dar \drar & \*u_3 \dar \ar{drr } & \dotsb & \*u_m \ar{dll } \\
\*i \setminus \*k \urar \ar[crossing over]{urr } \ar[crossing over]{urrr } \ar[crossing over]{urrrrr } & \*v_1 & \*v_2 & \*v_3 & \dotsb & \*v_n \end{tikzcd}\ ] ] using that each element of @xmath169 is connected , it is easy to check that the functor @xmath175 induced by the quotient map @xmath176 is full .
so since @xmath0 is simply - connected , so must be @xmath169 .
but since @xmath169 has no chains of cardinality @xmath177 , @xmath178 is just the graph - theoretic fundamental groupoid of its hasse diagram ( depicted above ) .
so for @xmath169 to be simply - connected , its hasse diagram must be acyclic , which clearly implies @xmath179 , as desired .
[ thm : cosieve - sc - tree ] let @xmath0 be a finite ( connected ) poset such that every upward - closed subset is simply - connected .
then @xmath0 is forest - like ( tree - like ) .
[ thm : amalgam - tree ] let @xmath0 be a finite tree - like poset .
then for every category @xmath17 with the ap , every diagram @xmath38 has a cocone . by induction on the construction of @xmath0 , there is @xmath46 minimal such that @xmath180 is the disjoint union of tree - like posets @xmath154 and @xmath181 is connected for each @xmath156 . for each @xmath156
, there is some @xmath182 since @xmath158 is connected , and there is a cocone @xmath183 over @xmath184 by the induction hypothesis . by the ap in @xmath17 , there is a cocone @xmath185 over the pushout diagram consisting of the composite maps @xmath186{f(i , u_k ) } f(u_k ) --->[\*c]{f^k_{u_k } } x_k\end{aligned}\ ] ] ( where @xmath187 denotes @xmath40 applied to the unique morphism @xmath32 { } u_k$ ] ) for all @xmath156 , where @xmath188 { } y$ ] and @xmath189 { } y$ ] with @xmath190 for @xmath191 , let @xmath192 be the composite @xmath193 we claim that @xmath194 is a cocone over @xmath40 . for @xmath195
where @xmath191 for some @xmath156 , also @xmath196 , whence @xmath197 because @xmath183 is a cocone over @xmath184 .
so we only need to check that for @xmath198 , i.e. , @xmath199 , we have @xmath200 . by connectedness of @xmath158 ,
there is a path @xmath201 in @xmath158 , whence @xmath202 [ thm : amalgam - forest ] let @xmath0 be a finite forest - like poset . then for every category @xmath17 with the ap and the jep , every diagram @xmath38 has a cocone .
find cocones over @xmath40 restricted to each connected component of @xmath0 , then apply jep .
[ thm : usc - etc ] a finite poset @xmath0 is upward - simply - connected , iff every upward - closed subset @xmath139 is simply - connected , iff @xmath0 is forest - like . by [ thm : usc - cosieve - sc ] , [ thm : cosieve - sc - tree ] , [ thm : amalgam - forest ] , and [ thm : setinj - amalgam ] .
this proves ( iii)@xmath132(iv)@xmath132(v)@xmath113(i ) in [ thm : amalgam ] for finite posets @xmath0 .
we begin this section by explaining why in a context such as frass theory , where we are looking at diagrams in a monic category @xmath17 ( of embeddings in the case of frass theory ) , the only diagram shapes @xmath0 worth considering are posets .
the of a category @xmath0 is the monic category @xmath203 freely generated by @xmath0 .
explicitly , @xmath204 for the least congruence @xmath205 on @xmath0 such that @xmath206 is monic .
thus for a diagram @xmath38 in a monic category @xmath17 , @xmath40 factors through @xmath203 , say as @xmath207 ; and pullback of cocones over @xmath133 along the projection @xmath208 is an isomorphism of categories ( between the category of cocones over @xmath133 and the category of cocones over @xmath40 ) .
so in a monic category @xmath17 , we may as well only consider diagrams whose shape @xmath0 is monic .
and by the following , an upward - simply - connected monic @xmath0 is necessarily a preorder ; clearly @xmath0 may be replaced with an equivalent poset in that case .
[ thm : incat - mono ] let @xmath17 be an idempotent inverse category
. then there is at most one monomorphism between any two objects in @xmath17 .
let @xmath209 { } y$ ] be two monomorphisms .
then @xmath210 is idempotent , whence @xmath211 .
this implies that @xmath212 is a pseudoinverse of @xmath75 , whence @xmath213 , whence @xmath214 .
[ thm : usc - mon - posetal ] an upward - simply - connected small monic category @xmath0 is necessarily a preorder .
since @xmath0 is monic , we have an embedding @xmath215 { } i ) \\
( i --->[\*i]{i } j ) & |-- > ( j |- > i \circ j);\end{aligned}\ ] ] this embedding factors through the canonical functor @xmath216 , so @xmath102 must be faithful . by [ thm : incat - mono ] , for @xmath217 { } j$ ] , we have @xmath218 , whence @xmath219 .
now we check that @xmath1 is invariant when passing from @xmath0 to @xmath203 : [ thm : l - m ] for any ( small ) category @xmath0 , @xmath220 .
@xmath221 is left adjoint to the composite @xmath222 ( where @xmath223 is the category of ( small ) monic categories ) where the first functor takes an inverse category to its subcategory of monomorphisms , and the second functor is the full inclusion , which has left adjoint @xmath224 . [
thm : amalgam - monic ] let @xmath0 be a connected upward - simply - connected category with finitely many objects .
then every diagram @xmath38 in a monic category @xmath17 with the ap has a cocone .
since @xmath17 is monic , @xmath40 factors through @xmath203 , as @xmath207 , say .
since @xmath0 is upward - simply - connected , so is @xmath203 , whence @xmath203 is a finite preorder . clearly replacing @xmath203 with an equivalent poset
does not change whether @xmath133 has a cocone , so by [ thm : amalgam - tree ] , @xmath133 has a cocone , which induces a cocone over @xmath40 .
we now deduce the general case of ( iii)@xmath113(i ) in [ thm : amalgam ] : [ thm : ap - moniq ] let @xmath17 be a category with the ap . then @xmath225 defines a congruence on @xmath17 , thus @xmath226 .
clearly @xmath205 is reflexive , symmetric , and right - compatible . to check left - compatibility , suppose @xmath227 { } y$ ] and @xmath228 { } z$ ] ; we show @xmath229 .
let @xmath230 { } w$ ] witness @xmath231 , so that @xmath232
. then @xmath229 is witnessed by @xmath233 such that the following diagram commutes : @xmath234{l } & & w \ular[dashed ] \\ & y \ular{h } \urar[swap]{k } \end{tikzcd}\ ] ] to check transitivity , suppose @xmath235 { } y$ ] ; we show @xmath236 .
let @xmath230 { } z$ ] witness @xmath231 and @xmath237 { } w$ ] witness @xmath238 .
then @xmath236 is witnessed by an amalgam of @xmath239 .
let @xmath0 be a finitely generated connected upward - simply - connected category .
then every diagram @xmath38 in a category @xmath17 with the ap has a cocone .
let @xmath240 be the monic reflection of @xmath17 . by [ thm : amalgam - monic ] , @xmath241 has a cocone @xmath44 , where @xmath242 { } y$ ] for each @xmath46 .
pick for each @xmath46 a lift @xmath243 { } y$ ] of @xmath244 . for each @xmath42 { } j$
] , since @xmath44 is a cocone over @xmath245 , we have @xmath246 ; by [ thm : ap - moniq ] , this means there is some @xmath247 { } z_i$ ] such that @xmath248 . now letting @xmath228 { } z$ ]
be an amalgam of @xmath249 for all arrows @xmath100 in some finite generating set of arrows in @xmath0 , we have @xmath250 for all @xmath42 { } j$ ] in the generating set , so @xmath251 is a cocone over @xmath40 .
let @xmath0 be a finitely generated upward - simply - connected category .
then every diagram @xmath38 in a category @xmath17 with the ap and the jep has a cocone .
this proves ( iii)@xmath113(i ) in [ thm : amalgam ] . since ( i)@xmath113(ii ) is obvious , to complete the proof of the theorem it only remains to check let @xmath0 be a category such that every @xmath0-shaped diagram in a category with the ap has a cocone .
then @xmath0 is connected .
consider the diagram @xmath252 where @xmath253 is regarded as a discrete category .
suppose we are given a finite category @xmath0 in some explicit form ( say , a list of its morphisms and a composition table )
. then our proof of [ thm : amalgam ] yields a simple procedure for testing whether a `` generalized amalgamation property '' holds for @xmath0-shaped diagrams : first , compute the monic reflection @xmath203 ; this can be done in finite time , since @xmath204 for the least congruence @xmath205 such that @xmath206 is monic , and @xmath205 can be computed by taking the equality congruence and closing it under finitely many conditions , which takes finitely many steps since @xmath0 is finite .
if @xmath203 is not a preorder , then @xmath0 is not upward - simply - connected by [ thm : usc - mon - posetal ] .
if @xmath203 is a preorder , then replace it with an equivalent poset @xmath60 and recursively test whether @xmath60 is forest - like using [ thm : posetal - cosieve - sc - decomp ] , i.e. , for each connected component @xmath254 , test whether @xmath143 is tree - like by picking some minimal @xmath255 and then for each @xmath256 testing whether @xmath257 is connected and whether @xmath258 is tree - like .
this can not be extended to finitely presented @xmath0 , since if @xmath0 is a group regarded as a one - object category then @xmath259 is idempotent iff @xmath0 is trivial , and it is undecidable whether a finite group presentation presents the trivial group ( see e.g. , @xcite ) .
we end by pointing out the following simple , but somewhat surprising , consequence of par s result , [ thm : pushout ] , which shows that the analogy between the ap and pushouts breaks down when it comes to decidability : there is a standard procedure to turn a finite presentation of a group @xmath260 into a finite @xmath261-dimensional ( abstract ) simplicial complex @xmath262 with fundamental group @xmath260 ; then the nerve ( see [ rmk : nerve ] ) of the face poset of @xmath262 is the barycentric subdivision of @xmath262 , hence the face poset of @xmath262 has fundamental group(oid equivalent to ) @xmath260 ( see @xcite for details ) . c. f. miller , decision problems for groups survey and reflections , in _ algorithms and classification in combinatorial group theory _ , mathematical sciences research institute publications , vol .
23 , springer , 1992 . | a category has the amalgamation property ( ap ) if every pushout diagram has a cocone , and the joint embedding property ( jep ) if every finite coproduct diagram has a cocone .
we show that for a finitely generated category @xmath0 , the following are equivalent : ( i ) every @xmath0-shaped diagram in a category with the ap and the jep has a cocone ; ( ii ) every @xmath0-shaped diagram in the category of sets and injections has a cocone ; ( iii ) a certain canonically defined category @xmath1 of `` paths '' in @xmath0 has only idempotent endomorphisms .
when @xmath0 is a finite poset , these are further equivalent to : ( iv ) every upward - closed subset of @xmath0 is simply - connected ; ( v ) @xmath0 can be built inductively via some simple rules .
our proof also shows that these conditions are decidable for finite @xmath0 .
= 1 |
the study of interstellar hydrides has received a great push in the recent years , much of it due to observations made from the _ herschel space observatory_. having relatively large rotational constants , many rotational transitions of hydrides lie in the millimiter- and sub - millimitr wave regions , which are difficult , albeit not impossible , to access from ground - based observatories , even for sites as good as those of alma or apex . since hydrides are some of the first molecules to form in space from atomic gas and molecular hydrogen , they provide invaluable information about the environment in which they are found @xcite .
among some of the recent molecules identified in _ herschel _ s spectra are the chlorine - bearing compounds h@xmath1cl@xmath0 @xcite and hcl@xmath0 @xcite .
the related hcl neutral had already been identified in the interstellar medium by @xcite .
h@xmath1cl@xmath0 has later been detected in a variety of lines of sight @xcite , both in absorption and emission , also from _ herschel _ s spectra , and in an extragalactic source in alma spectra @xcite .
the only published detection of hcl@xmath0 that we are aware of is that reported by @xcite in the line of sight of the luminous continuum sources w49n and w31c near the galactic center .
the key chemical pathways leading to the formation and destruction of chlorine - containing molecules in the insterstellar medium ( ism ) are described in @xcite and references therein .
succinctly , since the ionization potential of cl is 12.8 ev , slightly lower that that of h , chlorine is mostly in ionized form in the diffuse medium , and it is known to react exothermically with h@xmath1 to form hcl@xmath0 .
further reaction with h@xmath1 forms h@xmath1cl@xmath0 , which can produce neutral hcl or cl upon electron recombination .
chemical models @xcite predict the highest concentrations of hcl@xmath0 in regions with low extinction @xmath2 ) , with molecular h@xmath1 accounting for less than 1 % of the available h. the identification of hcl@xmath0 by @xcite , relied on the calculated transition frequencies and hyperfine patterns derived from a dunham - type isotope - independent fit of two sets of laboratory data @xcite : the first one was an extensive set of vis - uv emission measurements @xcite comprising more that 8000 transitions in the @xmath3 band of all isotopologues . the precision of those lines was @xmath4 @xmath5 , and the spectral resolution was too low to provide any information on the hyperfine structure ( hfs ) . the second set of data @xcite provided 34 accurate frequencies in the millimiter - wave region , resolving the chlorine hyperfine patterns of three rotational transitions for h@xmath6cl@xmath0 and h@xmath7cl@xmath0 .
this allowed @xcite to perform a joint analysis of both sets of data yielding highly accurate predictions , enough to identify the hcl@xmath0 transitions in _ herschel _ s spectra .
however , the fit did not contain any rotationally resolved data from vibrational transitions in the ground state . adding that information to the global fit
will improve the accuracy of the predictions of the model for other transitions and may facilitate possible observations in the future in the infrared ( ir ) region of the spectrum .
indeed , ir observations from ground platforms at high spectral resolution are a way to build up in the study of interstellar hydrides , providing an alternative and complementary tool to the millimiter and sub - millimeter observations , especially now that the _ herschel _ mission is over .
the instrument great onboard sofia covers the frequency region of the lowest - frequency rotational transitions of h@xmath6cl@xmath0 and h@xmath7cl@xmath0 , although telluric lines at 1441.1 and 1444.0 ghz overlap some of the hyperfine components of the transitions and might hinder the observations . nevertheless , there are current observation programs of hcl@xmath0 with great toward sgr b2 ( m ) , w31c ( g10.6 - 0.4 ) , g29.96 - 0.02 , w49n , w51 , w3(oh ) , and ngc6334i @xcite .
in this paper we provide a set of accurate wavenumbers for vibration - rotation transitions of the @xmath8 band of h@xmath6cl@xmath0 and h@xmath7cl@xmath0 in the mid - ir , as well as an extended and improved fit with millimeter - wave , optical and ir data .
there is one previous laboratory ir study by diode laser spectroscopy , described in a preliminary note @xcite , where 20 fine transitions of h@xmath6cl@xmath0 and 13 fine transitions of h@xmath7cl@xmath0 , were measured with an uncertainty of @xmath9 @xmath5 .
the limited number of observations was most likely due to the patchy coverage of lead salt diode lasers . on account of the much more extensive and accurate set of wavenumbers obtained in this investigation ,
those results have not been included in the present work .
the apparatus used in this experiment is basically the same that has been used recently for the investigation of the ir spectra of nh@xmath10d@xmath0 and arh@xmath0 ions @xcite .
it is based on an ir difference - frequency laser spectrometer , a hollow - cathode discharge reactor , and a double modulation technique with phase - sensitive detection .
the only major change with respect to those previous works is the substitution of the linbo@xmath10 crystal for a mgo - doped periodically poled linbo@xmath10 chip with different poling periods , covering without gaps the 1900 - 4300 @xmath5region when pumped by the 514 nm line of an ar@xmath0 laser , and providing significantly more ir power than our previous setup ( @xmath11 mw vs. @xmath12w . )
the accuracy and repeatability of the frequency scale rely on the active frequency locking of the ar@xmath0 laser to an hyperfine transition of @xmath13i@xmath1 , and a high - accuracy wavemeter , calibrated with the same laser , resulting in a 3@xmath14 accuracy of 10 mhz ( 3.3@xmath15 @xmath5 ) in the ir frequency scale .
the instrumental resolution is @xmath16 mhz ( 10@xmath17 @xmath5 ) , so the observed linewidths are limited by the doppler effect . in most of the experiments , the cathode was water - cooled , although , in a second step , some lines were recorded again using a flow of liquid nitrogen . a thermocouple resting at the mid - point of the cathode read @xmath18330 and @xmath18180 k , respectively , under discharge conditions .
the plasma was operated in a flow of hcl diluted in he , passed through a dry - ice / ethanol trap to remove possible water traces .
the strongest signals were obtained with a partial pressure of he of 1.2 mbar , and a partial pressure of hcl below the resolution of the available capacitance pressure meter ( 0.001 mbar ) .
the discharge current was 325 ma , with 400 v rms between electrodes .
the rest of operating conditions are similar to those reorted in @xcite .
line positions were calculated with the polynomial expressions given by @xcite , that do not account for any hyperfine splitting , and later they were taken from the prediction available at the jpl database @xcite , entry c036005 , with frequencies for all hyperfine components .
the accuracy was high enough to find the lines within less than 0.02 @xmath5from their expected positions . between 6 and 200 scans
were averaged for each line , depending on its intensity , attaining signal - to - noise ratios between 10 and 500 .
the investigated region spanned the 2337 - 2774 @xmath5interval .
atmospheric co@xmath1 absorption hampered the detection at lower frequencies ( higher-@xmath19 lines in the @xmath20-branch ) , and we kept a similar @xmath21 range for the @xmath22-branch .
no attempt was made to cancel the earth s magnetic field in the experiments ( 44.6 @xmath23 t in madrid ) , since the calculation of zeeman energies by @xcite showed a worst - case splitting ( that of the @xmath24 lines ) of only 3 mhz , much smaller than their doppler width .
. in all panels ,
the bottom trace is a stick spectrum as derived from the fit showing the hfs , and the top trace is the experimental observation . ( a )
the middle trace is the convolution of the stick spectrum with a gaussian function of 0.0065 @xmath5fwhm .
( b ) the numbers on the top of the sticks are the total angular momentum quantum numbers @xmath25 .
the group of lines marked with an asterisk correspond to @xmath26 .
the middle trace is the stick spectrum convolved with a gaussian function of 0.0055 @xmath5fwhm .
( c ) the middle trace is the convolution of the stick spectrum with a gaussian function of 0.0055 @xmath5fwhm . in this case , the @xmath27 and @xmath28 components in the @xmath29 ladder are not resolved , unlike those in the @xmath30 ladder .
, title="fig : " ] . in all panels ,
the bottom trace is a stick spectrum as derived from the fit showing the hfs , and the top trace is the experimental observation . ( a )
the middle trace is the convolution of the stick spectrum with a gaussian function of 0.0065 @xmath5fwhm .
( b ) the numbers on the top of the sticks are the total angular momentum quantum numbers @xmath25 .
the group of lines marked with an asterisk correspond to @xmath26 .
the middle trace is the stick spectrum convolved with a gaussian function of 0.0055 @xmath5fwhm .
( c ) the middle trace is the convolution of the stick spectrum with a gaussian function of 0.0055 @xmath5fwhm . in this case , the @xmath27 and @xmath28 components in the @xmath29 ladder are not resolved , unlike those in the @xmath30 ladder .
, title="fig : " ] . in all panels ,
the bottom trace is a stick spectrum as derived from the fit showing the hfs , and the top trace is the experimental observation . ( a )
the middle trace is the convolution of the stick spectrum with a gaussian function of 0.0065 @xmath5fwhm .
( b ) the numbers on the top of the sticks are the total angular momentum quantum numbers @xmath25 .
the group of lines marked with an asterisk correspond to @xmath26 .
the middle trace is the stick spectrum convolved with a gaussian function of 0.0055 @xmath5fwhm .
( c ) the middle trace is the convolution of the stick spectrum with a gaussian function of 0.0055 @xmath5fwhm . in this case , the @xmath27 and @xmath28 components in the @xmath29 ladder are not resolved , unlike those in the @xmath30 ladder .
, title="fig : " ] the energy levels structure of hcl@xmath0 has been described in previous papers , so we just summarize it here .
the ground electronic state of hcl@xmath0 is @xmath31 , so there is electron spin and orbital angular momenta interaction , originating two ladders of levels with @xmath32 .
since the ratio of the constants @xmath33 is -66 , it is an _ inverted _ doublet ( the @xmath29 state lies lower in energy by
@xmath34 @xmath5 ) that can be well described by hund s coupling case
_ a_. levels in each ladder are split by @xmath35-doubling into @xmath27 and @xmath28 parity doublets that are further split into hyperfine components because of the interaction with the nuclear spins ( @xmath36 , @xmath37 ) .
the selection rules @xmath38 , @xmath39 for @xmath40 and @xmath41 , @xmath42 for @xmath43 thus give rise to @xmath20 , @xmath44 and @xmath22 branches with four lines for each rotational transition .
these lines split into multiplets due to the hfs of the levels ( @xmath45 ) arising from the interaction with the cl nucleus spin . in low-@xmath19 transitions ,
the hfs is clearly evident , although not completely resolvable , and its amplitude rapidly diminishes as @xmath19 increases .
figure [ fig : figf]a shows an example of lines where the hyperfine components can not be resolved due to their proximity and the doppler broadening of the lines . in turn , figure [ fig : figf]b shows the @xmath46 line of h@xmath6cl@xmath0 line recorded at @xmath47 k , where the hfs is almost completely resolved .
there are intermediate cases where the hfs manifests as broadening or asymmetries of the lines , an example is shown in figure [ fig : figf]c . in this case , both @xmath27 and @xmath28 components of the @xmath48 line of h@xmath7cl@xmath0 overlap , and , together with the hfs , yield a complex , unresolved asymmetric profile . the same line in the @xmath30 ladder is split into @xmath27 and @xmath28 components , also with an asymmetric profile ( the @xmath28 component is outside the frequency range of the figure ) .
it must be noted that the proton hfs could not be resolved in the thz study by @xcite . in the present work , we access lines more sensitive to this interaction , but , as could be expected from the larger doppler widths and the magnitude of the splittings predicted by theory @xcite , we have not been able to resolve it either , and only the cl hyperfine interaction has been considered . from the line width of a gaussian profile fitted to unresolved transitions with negligible hfs broadening ,
we have obtained kinetic temperatures of @xmath49 k for the spectra recorded with water cooling , and @xmath50 k for those with liquid nitrogen cooling .
the ratio of intensities of different @xmath19 lines under rotational equilibrium is also in agreement with this estimation .
the ratio of intensities of lines with the same @xmath19 in @xmath51 leads to a much higher spin temperature , near @xmath52 k , as can be inferred from the higher intensity of the @xmath30 lines as compared with the calculations assuming boltzmann equilibrium in figures [ fig : figf]a and [ fig : figf]c .
this is not surprising since there is no reason to expect an equilibrium between rotational , vibrational , or spin temperatures in a plasma .
moreover , @xcite also observed a high spin temperature , and the emission data from @xcite showed vibrational excitation up to @xmath53 .
no hot - band lines were recorded , thus we can not give an estimate of the vibrational temperature of hcl@xmath0 in the discharge .
the list of assignments and observed ir frequencies , together with those calculated and their residuals after the fit ( see the appendix ) , is given in table [ tab : linelist ] .
cccccccclclcc 37 & 1 & 1 & 2 & 2 & 2 & 2 & @xmath54 & 2536.09722 & 30 & & 2536.09701 & 21.3 + 37 & 1 & 1 & 2 & 2 & 2 & 1 & @xmath54 & 2536.10792 & 36 & & 2536.10792 & -0.3 + 37 & 1 & 1 & 2 & 2 & 2 & 3 & @xmath55 & 2536.67608 & 40 & & 2536.67600 & 7.9 + 35 & 1 & 1 & 2 & 2 & 2 & 3 & @xmath55 & 2538.49530 & 30 & & 2538.49519 & 10.6 + 35 & 8 & 9 & & 8 & 9 & & @xmath56 & 2543.90798 & 30 & & 2543.90823 & -25.3 + 35 & 8 & 9 & & 8 & 9 & & @xmath57 & 2544.52117 & 30 & & 2544.52074 & 43.3 + 35 & 1 & 1 & 1 & 1 & 1 & 1 & @xmath57 & 2568.54858@xmath58 & 30 & * & 2568.54624 & 33.9 + 35 & 1 & 1 & 1 & 1 & 1 & 2 & @xmath57 & 2568.54858@xmath58 & 30 & * & 2568.54893 & -35.3 + a new combined isotope - independent fit with thz , optical , and ir data was carried out and is described in the appendix . using the dunham equations and the correlated error propagation formulae , the band parameters can be determined directly from the fit @xcite , and are given in table [ tab : band_constants ] . in this format ,
comparisons to other data analyses in the literature are possible .
a few things are remarkable : ( 1 ) the value of @xmath59 is in much better agreement with @xcite than with any other study , including @xcite ; ( 2 ) the values are generally of slightly lower precision than gupta et al . ,
opposite to the trend observed in the dunham fit ( see the appendix . ) the former result is presumably due to brown s careful approximations of higher - order parameters . the latter may be due to truncation error in the prior analysis in which parameter uncertainties are underestimated when higher - order parameters that are not well approximated as zero are not included in the analysis .
when these terms are introduced they have uncertainty and this correlates with the prior uncertainties .
rrrrrr @xmath60 & 77003804.2(21 ) & 77003551(75 ) & & 77005284(99 ) & + @xmath61 & @xmath628882(40 ) & @xmath629163(34 ) & @xmath628993(165 ) & & @xmath629114 + @xmath63 & @xmath628882(40 ) & @xmath629163(34 ) & & & + @xmath64 & @xmath6219279455.6(47 ) & @xmath6219279138(38 ) & @xmath6219278775(1000 ) & @xmath6219430419 ( 92 ) & @xmath6218137444 + @xmath65 & @xmath6219264103.2(47 ) & @xmath6219263430(69 ) & & @xmath6219417878(138 ) & + @xmath66 & 45.0(12 ) & 54.33(99 ) & 45.3(42 ) & 62.60(48 ) & + @xmath67 & 68.9(12 ) & 80.98(113 ) & & 84.03(75 ) & + @xmath68 & 0.01853(113 ) & & & & + @xmath69 & 0.01853(113 ) & & & & + @xmath70 & 293443.752(64 ) & 293444.013(48 ) & 293420.5(60 ) & 293607.7(23 ) & + @xmath71 & 283806.268(103 ) & 283812.603(324 ) & & 283959.2(30 ) & + @xmath72 & 16.39011(163 ) & 16.39759(157 ) & & 16.419(13 ) & + @xmath73 & 16.16813(186 ) & 16.20489(518 ) & & 16.161(18 ) & + @xmath74 & 4.823(103 ) & 4.468(126 ) & & 4.57(23 ) & + @xmath75 & 4.823(103 ) & 4.468(126 ) & & 4.22(32 ) & + @xmath76 & 18281.2(10 ) & 18280.17(104 ) & 18279.60(94 ) & 18259(12 ) & + @xmath77 & 17983.9(11 ) & 17972.6(471 ) & & 17953(17 ) & 17598 + @xmath78 & -1.266(24 ) & @xmath621.290(25)@xmath79 & & @xmath620.975(51 ) & + @xmath80 & -1.266(24 ) & @xmath621.290(25 ) & & @xmath620.932(98 ) & + @xmath81 & 5.80(150 ) & 6.16(172 ) & & & + @xmath82 & 5.80(150 ) & 6.16(172 ) & & & + @xmath83 & @xmath62332.76(12 ) & @xmath62332.198(87 ) & @xmath62343.0(11 ) & @xmath62335.1(18 ) & @xmath62360 + @xmath84 & @xmath62320.03(18 ) & @xmath62320.72(210 ) & & @xmath62318.7(33 ) & + @xmath85 & 0.04052(326 ) & 0.04106(478 ) & & 0.0470(60 ) & + @xmath86 & 0.04052(326 ) & 0.04106(478 ) & & 0.017(12 ) & + @xmath87 & 357.2(18 ) & 357.04(60 ) & & & 369.5 + @xmath88 & 409.10(113 ) & 409.16(88 ) & & & 408 + @xmath89 & 167.4(38 ) & 169.6(34 ) & & & 166 + @xmath90 & 77.0 ( 36 ) & 78.34(254 ) & & & 85 + @xmath91 & -271.2 ( 51 ) & -274.64(313 ) & & & -243 + @xmath92 & 521.56(256 ) & 523.12(282 ) & & & 489 + @xmath93 & 88.8(540 ) & 60.6(414 ) & & & + @xmath94 & -9.67(301 ) & -8.03(210 ) & & & -12.5 + @xmath95 & -351.6(272 ) & -370.4(184 ) & & & -169 +
using the einstein @xmath96 coefficient for the @xmath97 transition of 205.6 s@xmath98 calculated by @xcite , we can obtain the vibrational transition dipole moment ( 0.1959 d ) and the individual line intensities as a function of temperature . with this information , it is possible to estimate the detectable column densities in hypothetical ir observations from the ground . for 5
k ( the estimated excitation temperature of hcl@xmath0 in the clouds toward w31c and w94n @xcite ) , the strongest ir absorption of @xmath99 would be that of the @xmath24 transition at 2568.21258 @xmath5(76,993,076 mhz ) . neglecting hfs and @xmath35 doubling ( that would not be observed for this line even with resolving powers r@xmath100100000 )
, we obtain the following integrated intensities ( in units of @xmath101kms@xmath98molec@xmath98 ) @xmath102 for temperatures @xmath103 k , respectively . @xcite
estimated the column densities to be @xmath104 @xmath105molec@xmath98 for w31c and @xmath106 @xmath105molec@xmath98 for w49n .
the observed absorption peak intensity will depend inversely on the observed linewidth , that , in turn , will depend on the velocity spread of the cloud and the spectrograph resolution .
@xcite found @xmath107 kms@xmath98 from the analysis of the hfs of the lines of hcl@xmath0 , in good agreement with the @xmath108 kms@xmath98 derived by @xcite from sh@xmath0 and ch@xmath0 in the same and other similar diffuse clouds .
for an instrument line width of 3 kms@xmath98 ( that of spectrographs with a resolving power of @xmath109 , as available from e.g. crires+ at the vlt in paranal or texes at mauna kea ) the combined line width would be @xmath18 6 kms@xmath98 , and one can expect peak absorptions of 3.85 , 3.60 , and 2.64 % , respectively , for the above temperatures and a column density of @xmath110 @xmath101molec@xmath98 . infrared emission from the dust envelope around the proto - stars in those sources
can be bright enough in the ir , and , indeed , the absorption of neutral hcl toward crl2136 at @xmath111 with peak intensities of @xmath112 has been reported by @xcite , using the crires instrument at the vlt .
therefore , the absorptions estimated above should be clearly detectable against bright ir sources in a reasonable amount of observing time . +
we have provided accurate wavenumber measurements of 183 vibration - rotation lines of h@xmath6cl@xmath0 and h@xmath7cl@xmath0 , measured with a difference - frequency laser spectrometer in a hollow - cathode discharge cell .
the new wavenumbers have improved the previous dunham - type fit to optical and millimiter - wave data , allowing more accurate predictions of other transitions .
notably , the wavenumbers of the @xmath24 ) transitions of the @xmath113 band are 2568.21258@xmath114 and 2566.34959@xmath115 @xmath5(@xmath116 ) for h@xmath6cl@xmath0 and h@xmath7cl@xmath0 , respectively
. these wavenumbers should help in future searches for these molecules .
in the conditions of the diffuse clouds where this ion has been observed with _
herschel _ , absorptions of a few percent should be observable from the ground with high - resolution ir telescopes .
jld , jc , vjh and it acknowledge the financial support from the spanish mineco through the consolider astromol project , grant csd2009 - 00038 , and from the erc through the synergy grant erc-2013-syg nanocosmos .
additional support through grants fis2012 - 38175 ( jld ) , fis2013 - 408087-c2 - 1p ( vjh , it ) and aya2012 - 32032 ( jc ) is also acknowledged .
we thank j.r .
goicoechea for his reading and comments on the manuscript .
portions of the research described in this paper were performed at the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
government sponsorship acknowledged .
rrr @xmath117 & 40117164(205 ) & 40117378(217 ) + @xmath118 & -10997 ( 54 ) & -11111 ( 82 ) + @xmath119 & 250.2(131 ) & + @xmath120 & -405371 ( 86 ) & -405382 ( 90 ) + @xmath121 & 1305.3 ( 99 ) & 1310.5(107 ) + @xmath122 & 298527.4 ( 47 ) & 298529.7 ( 47 ) + @xmath123 & -4902.66(100 ) & -4898.54(122 ) + @xmath124 & -210.3 ( 47 ) & -216.6 ( 48 ) + @xmath125 & 18.202(197 ) & 17.797(240 ) + @xmath126 & 11.11 ( 63 ) & 11.65 ( 91 ) + @xmath127 & -16.5654(100 ) & -16.5593(100 ) + @xmath128 & 0.11099 ( 37 ) & 0.09637(248 ) + @xmath129 & 0.0643 ( 97 ) & 0.0653(101 ) + @xmath130 & 0.4823(103 ) & 0.4468(126 ) + @xmath131 & -8882 ( 40 ) & -9163(34 ) + @xmath132(@xmath133 ) & -19289074(211 ) & -19288744(222 ) + @xmath134 & 11468(341 ) & 10596(377 ) + @xmath135 & -691 ( 58 ) & + @xmath136 & -1329(140)&-1124(160 ) + @xmath137 & 170.4(164 ) & 134.9(195 ) + @xmath138 & 33.33(120 ) & 41.02(114 ) + @xmath139 & 11.8413(298 ) & 13.32(39 ) + @xmath140 & 0.01852(113 ) & + @xmath141 & -9214.94 ( 53 ) & -9217.0(33 ) + @xmath142 & 74.331(163 ) & 76.9 ( 32 ) + @xmath143 & 0.6328(118 ) & 0.6448(127 ) + @xmath144 & -0.290 ( 74 ) & -0.308 ( 86 ) + @xmath145 & -169.566 ( 74 ) & -168.97 ( 54 ) + @xmath146 & 3.184 ( 36 ) & 2.87 ( 53 ) + @xmath147 & 0.02026(163 ) & 0.02053(239 ) + a full isotope - independent analysis including the fine - structure and magnetic hfs hamiltonian was carried out utilizing the spfit program @xcite .
the full hamiltonian and its mass and magnetic relationships among parameters are described in the appendix of @xcite , in @xcite and in @xcite .
note that in @xcite the signs should be reversed in the definition of @xmath148 parameters involving @xmath149 and @xmath150 related terms , and that the two rightmost factors in equations a4 and a5 should be removed .
the new fit increases the number of fitted lines from 8137 to 8324 ( 8103 lines from the optical data from @xcite ; 34 lines from @xcite and 187 lines from this work ) .
the present data were weighted according to their estimated uncertainty .
lines that showed no significant broadening or asymmetries were fitted to gaussian profiles to obtain the line center , while for clearly asymmetric lines the centroid frequency was used .
some blends were appropriately intensity weighted and severely blended lines , where @xmath151 components could not be resolved , were not used . the new assignments , observed and calculated frequencies , and their residuals are shown in table [ tab : linelist ] . the present high - precision data , together with the high - precision thz data of @xcite , provide anchor points for an improved analysis of the molecular parameters for the @xmath31 state , which are shown in table [ tab : dunham_parameters ] as determined by @xcite and presently
the precision of many of the dunham parameters is improved , especially the higher - order @xmath152 terms that were previously completely dependent on the optical data . unlike some of the more heavily correlated parameters associated with the spin - orbit distortion and electron - spin rotation ,
the @xmath149 and @xmath150 terms from the prior analysis do not deviate much outside the prior uncertainty ( all less than 1.1@xmath14 ) .
the precision afforded by the new data refines these values by factors of 6.1 and 7.3 for @xmath141 and @xmath153 , respectively , and by factors of 2.0 and 1.5 for @xmath154 and @xmath155 , respectively .
the most dramatically modified parameters in the present fit are , not surprisingly , the most heavily correlated ones , @xmath131 and @xmath138 , which change by about 7 - 8@xmath14 outside the previously determined error bars .
the changes stem from the new sensitivity to the @xmath139 parameter , which changes by about 4@xmath14 outside its prior uncertainty and is determined 13@xmath156 more precisely .
the combined changes in @xmath138 and @xmath139 produce the anti - correlated change in @xmath157 , which is then further removed from its near equivalence with @xmath141 and thus implies that hcl@xmath0 is less consistent with the pure - precession hypothesis , but only by 3% .
the basic dunham parameters and the associated bob terms are fairly stable upon addition of the higher fidelity vibrational data .
this is seen in @xmath158 and @xmath159 , which can be very sensitive to changes in the vibrational data set , but here only change by @xmath1601.5@xmath14 .
this indicates that the previous data set achieved a broad enough scope in vibration and mass to accurately determine these subtle effects
. however , higher - order terms in the dunham model remain susceptible to the breadth and scope of the data and statistically significant changes occur in many of these terms . the only other parameters to have changed by more that 3@xmath14 of its prior uncertainty are @xmath123 and @xmath128 , which appear to have correlated ( and anti - correlated ) commensurate changes with @xmath159 , which is not outside of 3@xmath14 of the prior uncertainty due to the larger uncertainty in that parameter .
due to correlation , any improvement in the precision of these two parameters is not obvious .
new frequency predictions for all isotopologues are available at http://spec.jpl.nasa.gov/ftp/pub/catalog/catdir.html[http://spec.jpl.nasa.gov/ftp/pub/catalog/catdir.htm ] . | the chloroniumyl cation , hcl@xmath0 , has been recently identified in space from _ herschel _ s spectra . a joint analysis of extensive vis - uv spectroscopy emission data together with a few high - resolution and high - accuracy millimiter - wave data provided the necessary rest frequencies to support the astronomical identification .
nevertheless , the analysis did not include any infrared ( ir ) vibration - rotation data .
furthermore , with the end of the _ herschel _ mission , infrared observations from the ground may be one of the few available means to further study this ion in space . in this work
, we provide a set of accurate rovibrational transition wavenumbers as well as a new and improved global fit of vis - uv , ir and millimiter - wave spectroscopy laboratory data , that will aid in future studies of this molecule . |
from the current experimental evidence of neutrino oscillations @xcite , we know that neutrinos are massive particles , and each neutrino flavor eigenstate @xmath5 is a coherent superposition of mass eingenstates @xmath6 , connected by the pmns mixing matrix @xmath7 , through the relation @xmath8 .
however , oscillation experiments are sensitive only to the mass squared differences ( @xmath9 ) and not on the absolute value of each mass eigenstate which is one of the most challenging measurement in particle physics
. furthermore , one must be careful about the interpretation of experiments on neutrino masses since different techniques result in different observables .
for instance , neutrino - less double beta decay experiments search for majorana neutrinos with mass @xmath10 and single-@xmath11 decay experiments observe features driven by massive neutrinos with @xmath12 not dependent on majorana phases . since @xmath13 measurements which are not exclusively sensitive to majorana neutrinos can be safely taken as conservative ones @xcite .
if we can measure the mass of any neutrino flavor @xmath14 the masses of @xmath15 can be obtained from the boundaries of the mixing matrix and other @xmath16 obtained from different oscillation experiments . the strongest experimental limit in the neutrino mass , as quoted in the particle data group ( pdg ) @xcite , was obtained by the troitsk experiment on tritium single-@xmath11 decay : @xmath17 2.05 ev at 95% confidence level ( c.l ) @xcite .
this kind of experiment measures the modification of the @xmath11 spectrum near its endpoint caused by the neutrino mass .
the katrin experiment @xcite will improve the technique and plan to set a stronger limit by enhancing the sensitivity at least in one order of magnitude .
neutrinos emitted from stellar collapse , for example type ii supernovae , can give a valuable contribution to the determination of the mass of neutrinos , as idealized by zatsepin @xcite .
using neutrino data @xcite from the sn1987a , which exploded in the large magellanic cloud ( @xmath1850 kpc ) , and the likelihood of detection analysis proposed in @xcite , an upper bound of 5.8 ev at 95% c.l @xcite was obtained . with a different likelihood approach and astrophysical parametrization of the neutrino emission , lamb & loredo @xcite
reached a 5.7 ev at 95% c.l for the neutrino mass limit .
there are good perspectives to explore the sub - ev region in the future experimental scenario for detection of galactic supernovae @xcite .
other mass limits obtained from the supernovae can be seen in the table presented in ref .
@xcite . the perspective to detect @xmath0 from a supernova , but in a liquid scintillator detector , was recently considered in @xcite .
water cherenkov detectors with gadolinium could also improve the number of @xmath0 events according to @xcite . in the present work
, we aim to discuss bounds on neutrino masses from supernovae motivated by the zatsepin s idea .
we are going to consider electronic neutrinos ( @xmath0 ) emitted from the neutronization burst of a galactic supernova at a distance of 10 kpc and 15 m@xmath1 and detected by a liquid argon time projection chamber ( lartpc ) .
the paper is organized as follow : sec .
[ sn ] presents the main assumptions about electronic neutrino production and its propagation ; sec .
[ detec ] presents the basic features about the neutrino detection in the lartpc ; sec .
[ method ] presents the main assumptions about our event generated monte carlo simulation and sec .
[ results ] presents the boundaries on the neutrino mass .
[ discuss ] and sec .
[ conclu ] present the discussion about our results and our main conclusions , respectively .
as the shock evolves against the infalling dense matter of the outer core of the star , some of its energy is lost by the photodissociation of nuclei into protons and neutrons . at this moment , there is an abundant production of electronic neutrinos ( @xmath0 ) via @xmath19 .
these electronic neutrinos accumulate in a very dense and opaque environment behind the shock wave .
when the shock reaches the zone with densities around @xmath20 g @xmath21 in a few miliseconds after the bounce the @xmath0 s are released .
this process is called neutronization burst .
it has a peak luminosity of @xmath22 erg s@xmath23 and lasts @xmath24 miliseconds .
for a recent review of the current knowledge and status of the explosion mechanism , see @xcite . at a distance @xmath25 of the progenitor star , for each flavor @xmath26 ( @xmath27 ) ,
the unoscillated sprectral number fluxes ( @xmath28 ) are : @xmath29 where @xmath30 is the luminosity for the respective flavor @xmath14 , @xmath31 is the @xmath14 mean energy and @xmath32 is the quasi - thermal spectrum written as follows @xcite : @xmath33 the @xmath34 is a parameter defined by @xmath35 and @xmath36 is the normalization constant factor : @xmath37 .
several of these parameters , such as @xmath30 in eq .
( [ eq4 ] ) , and @xmath38 and @xmath39 in eq .
( [ eq5 ] ) , change their respective values with time after bounce and depend on the simulation of the star explosion @xcite . after taking into account the effect of neutrino oscillations , in the three neutrino families framework in a supernova environment , the fluxes ( @xmath40 ) at earth for each flavor
@xmath14 are @xcite : @xmath41 @xmath42 and @xmath43 where @xmath44 are the primary unoscillated neutrino fluxes at the production region , @xmath45 or @xmath46 and @xmath47(@xmath48 ) is the survival probability of an electron ( anti)neutrino after propagation through the sn mantle and the interestellar medium . for the actual oscillated parameters , @xmath49 for normal hierarchy ( nh ) ( @xmath50 ) and @xmath51 for inverted hierarchy ( ih ) ( @xmath52 ) . for massive neutrinos , the arrival time ( @xmath53 ) of the neutrino in the detector at earth will be delayed @xcite : @xmath54 where @xmath55 is the emission time at the source and @xmath56 is the speed of light .
the delay introduced by the neutrino mass @xmath57 in eq .
( [ delay1 ] ) is represented by @xmath58 and can be written as : @xmath59
the neutrino flux of the neutronization burst is mainly populated by electronic neutrinos .
therefore , only experimental technniques with high sensitivity to @xmath0 can have a possibility to investigate this particular phase of supernova explosion .
the liquid argon time projection chamber ( lartpc ) , such as the icarus - like detectors @xcite , have all the features required to identify the neutronization burst .
the `` long baseline neutrino experiment '' ( lbne ) @xcite foresees a 34 kton ( fidutial volume ) lartpc as its far detector .
the experiment is dedicated for determination of cp violation phase and neutrino mass hierarchy .
though , its dimensions and underground site are suitable for exploring rare events , such as neutrinos from supernovae and proton decay .
this experimental design is taken into account for our calculations .
the most favourable electronic neutrinos ( @xmath0 ) detection channel in lartpcs is the charged current ( cc ) interaction : @xmath60 .
there are other detection channels , but all of them with smaller cross section .
these detection channels can be tagged by spectral analysis of the photon emission from de - excitation of @xmath61 , @xmath62 or @xmath63 , which exhibit specific spectral lines .
the @xmath64 cross sections can be found in fig . 3 of ref .
the number of expected events in the detector can be evaluated in the following way : @xmath65 where @xmath66 is the @xmath67 cc cross section , @xmath68 is the number of targets in the 34 kton argon detector and @xmath69 is the oscillated neutrino flux , which is evaluated by eqs .
( [ eq4]-[eq8 ] ) .
we performed a monte carlo simulation ( mc ) with very optimistic assumptions .
nevertheless , all of our assumptions have solid grounds and can be taken as a `` realistic case '' .
the number of expected events depends on the considerations we make about the neutrino production and their propagation till the detector .
we considered a comprehensive knowledge of the astrophysical process related to the sn explosion , which means that the astrophysical parameters related to neutrino production , ie the neutrino fluxes , are well known .
however , to introduce uncertainties related to the explosion mechanism , we considered a 5% error in the overall normalization of the evaluated counting rates . concerning the sn distance , we assumed a galactic sn of 10 kpc .
the detector configuration , as pointed out before , is 34 kton lartpc with 80% of efficiency for @xmath0 charged current ( cc ) detection . for now , background effects have been disconsidered since the detector is assumed to be placed underground . in this situation
we expect a suppression of seven orders of magnitude @xcite .
the @xmath0 energy threshold for detection was set to 5 mev . for the energy resolution we used a parametrization given by : @xmath70 @xcite . for the time resolution
, we adopted a @xmath7120 @xmath72 flat distribution .
also , we stress the fact that we performed our mc ignoring aditional complications given by the lack of knowledge if the first detected event is syncronized or not with the beginning of the burst . in the mc we included the two possible scenarions for mass hierarchy , as mentioned in sec .
[ sn ] , with the best - fit values of the oscillation parameters with no cp violation effects @xcite .
our mc simulation results in thinner lines are shown in fig .
nh and ih hierarchy are represented by solid and dashed lines , respectively .
the error bars in each bin are assumed to follow a poissonian distribution and each bin has @xmath73 s width . in fig .
[ mc ] , we also show the expected number of events in the thicker lines considering a 0.5 ev neutrino mass , just for comparison with our mc .
non vanishing neutrino mass causes a delay in the experimental signal , changing the expected number of events in each time bin , as we can notice analysing eq .
( [ delay2 ] ) and eq .
( [ eq3 ] ) .
these theoretical modifications can be compared with the number of events in each bin generated by our mc simulation using the @xmath74 test : @xmath75 where @xmath76 bins , @xmath77 is the number of events generated by our mc in the @xmath78-bin , @xmath79 is the number of events evaluated by eq .
( [ delay2 ] ) and eq .
( [ eq3 ] ) for different neutrino masses ( @xmath57 ) and @xmath80 is the associated error .
the `` realistic case '' provided @xmath81 for a neutrino mass of @xmath82 ev and considering nh . from the @xmath83 , the upper bound on neutrino mass from the @xmath0 neutronization burst stage is , at 95% c.l , 2.71 ev . on the other hand , for the ih case ,
we have obtained @xmath84 for a neutrino mass of @xmath85 ev . at 95% c.l , @xmath86 ev .
we present our @xmath87 in terms of the neutrinos mass , @xmath57(ev ) , in fig .
[ fig1 ] , for both hierarchies . in these plots ,
dotted lines are bounds with 68.27% c.l , dashed lines are for 90% c.l ; and dotted - dashed lines are the 95% c.l .
limit ; dashed line is the 2@xmath88 limit ; and dotted - dashed line is the 3@xmath88 limit . the strong red dashed curve is for ih and the strong solid black curve is for nh . ]
we also performed other mc simulations and did the @xmath74 analysis for the mass as a parameter for other situations that are described in the following : 1 .
`` perfect case '' : we consider that the normalization is completely known , 100% of detector efficiency and a perfect time and energy resolution .
we mantain an energy cutoff in 5 mev ; 2 .
`` perfect norm '' : returning to the conditions estabilished in `` realistic case '' , but now for a completely known supernova normalization ; 3 . ``
perfect eff '' : considering a detector with 100% of efficiency and other conditions of the `` realistic case '' maintained ; 4 .
`` cut ener '' : keeping the `` realistic case '' , but instead cutting events with energy less than 5 mev , we perform a energy cut at @xmath89 mev ; 5 .
`` time resol '' : considering a temporal resolution of 1 ms instead of 20 @xmath72 of the `` realistic case '' ; 6 .
`` bin1 '' : if we increase the bin size from 5 ms to 10 ms using the `` realistic case '' and 7 .
`` bin2 '' : if we decrease the bin size from 5 ms to 2 ms using the other conditions of the `` realistic case '' we present in table [ tabresumonh ] our mass boundaries at 68.27@xmath90 , 90@xmath90 and 95@xmath90 c.l for nh .
the numbers in parentheses are the boundaries for ih . in this table
we also show the minimum value of the mass ( @xmath91 ) with its respective @xmath74 , the @xmath92 .
.upper bounds on @xmath57(ev ) considering a supernova at 10 kpc and a liquid argon detector with 34 kton .
we show the results for normal hierarchy ( nh ) and the numbers inside the parentheses are the mass evaluation for inverted hierarchy ( ih ) .
also we show the the minimum value of the mass ( @xmath91 ) with its respective @xmath74 , the @xmath92 . for details of the meaning of each test , see the text . [
cols="<,<,<,<,<,<",options="header " , ]
from table [ tabresumonh ] we notice that , despite we change the configuration of the detector , statistical fluctuation of the monte carlo simulation plays a fundamental role in the mass bound determination , since we have few events in the neutronization burst and this process lasts only @xmath93 25 ms .
a very good example of this kind of behaviour is the one shown in the ih curve of fig .
[ mc ] ( red dashed ) . in the interval of 35 - 40
ms , there is a peak that was not theoreticaly predicted .
so , this generates a worse @xmath74 and an interval in the neutrino mass bound ( 0.18 - 1.70 ev ) with a @xmath92 at 1.00 ev .
another example of the impact of this statistical fluctuation is that , for the nh case , we have obtained a more stringent bound in the `` realistic '' case compared to the `` perfect '' case .
our bound , at 95% c.l , for the nh is very similar to the one obtained by the troitsk experiment of tritium @xmath11 decay .
so , if we really have nh and comparing with terrestrial experiments , which are independent of mass hierarchy , like katrin ( @xmath94 ev ) , it will be very difficult that lbne will put a more stringent limit on neutrino mass . from the planck cosmic microwave background measurements and
galaxy clustering information from the baryon oscillation spectroscopic survey ( boss ) , part of the sloan digital sky survey - iii , and assuming a @xmath95cdm cosmological model , ref .
@xcite obtained , at 95% c.l , @xmath96 ev . ref .
@xcite obtained that @xmath97 ev , when one takes into account observations of the large - scale matter power spectrum from the wigglez dark energy survey , planck sattelite and baryon acoustic oscillation . if we consider that mass eigenstates are degenerated , @xmath98 ev @xcite and @xmath99 ev @xcite . for the nh and ih case , our bounds , respectively , at 95% c.l , are @xmath2 ev and @xmath3 ev .
these bounds are almost similar with other prospects of neutrino mass bounds obtained for sn neutrinos . just for comparison ,
authors in @xcite considering only normal hierarchy , using a 20 kton liquid scintillator at jiangmen underground neutrino observatory ( juno ) and a galatic supernova @xmath100 signal obtained a limit on the absolute neutrino mass scale : @xmath101 ev at 95% c.l .
this limit is also similar with the ones obtained in the other cases discussed in this work - see table [ tabresumonh ] .
so , our conclusion here is that , for a sn at 10 kpc of distance and 15 m@xmath1 , lbne 34 kton argon detector can probe neutrino masses of order of 1 ev , at 95% c.l .
we stress the fact that knowing neutrino mass hierarchy is fundamental to put limits on neutrino masses from sn and lbne has as one of its objectives determine this property .
even though our supernova neutrino bound seems not to be competitive with the prospective katrin bound , the neutrino mass bound from sn has a lot merit and importance , and , of course , it should be a source of strong investigation . with improvements in the detector efficiency ,
the number of events per bin would increase and a more competitive bound could be obtained .
we remember that efficiency includes several effects , such as , loss of events due to light attenuation , fluctuations of the number of photoelectrons , the geometry of the detector etc . these are still unknown sources of errors , and their knowledge is crutial for a more detailed work than presented here .
we are aware that our assumptions are very crude , however , as stated by the lbne collaboration in @xcite , work in understanding the physics and choices of detector is still underway .
it is necessary to determine , for example , the low energy response of the detector , its geometry , its real threshold , the reconstruction of the events , event vertex resolution , the energy resolution , absolute event @xmath102 time precision and other details .
that is why we are being very conservative , but at the same time , very optimistic in this work .
however , we should stress that considering this time binned analysis , we have demonstrated that statistical fluctuation may play a major role in the determination of the neutrino mass bound .
we point out that a more dedicated mc simulation , with real and proper detector information , with an effective study of a possible event - by - event likelihood of detection for this flux of @xmath0 of neutronization burst and , finally , the importance of the time of the first event detected should be studied deeply in the foreseeable future .
f. rossi - torres would like to thank cnpq for the full financial support .
e. kemp would like to thank fapesp and cnpq for partial financial suppport .
m. m. guzzo also would like to thank cnpq for partial financial support to this work .
also , we would like to thank prof .
janka for providing the data of supernova simulation used to calculate neutrino fluxes .
m. steidl , arxiv:0906.0454 [ nucl - ex ] .
j. beringer _
[ particle data group collaboration ] , phys .
d * 86 * , 010001 ( 2012 ) .
v. n. aseev _ et al . _
[ troitsk collaboration ] , phys .
d * 84 * , 112003 ( 2011 ) [ arxiv:1108.5034 [ hep - ex ] ] .
r. g. h. robertson [ katrin collaboration ] , arxiv:1307.5486 [ physics.ins-det ] .
zatsepin , pisma zh .
* 8 * ( 1968 ) 333 .
hirata _ et al .
_ , phys .
d * 38 * ( 1988 ) 448 ; k. hirata _ et al .
_ [ kamiokande - ii collaboration ] , phys .
* 58 * ( 1987 ) 1490 .
bionta _ et al .
_ , phys .
lett . * 58 * ( 1987 ) 1494 ; c.b .
bratton _ et al . _
[ imb collaboration ] , phys .
d * 37 * ( 1988 ) 3361 .
alekseev , l.n .
alekseeva , v.i .
volchenko and i.v .
krivosheina , jetp lett .
* 45 * ( 1987 ) 589 ; e.n .
alekseev , l.n .
alekseeva , i.v .
krivosheina and v.i .
volchenko , phys .
b * 205 * ( 1988 ) 209 .
a. ianni , g. pagliaroli , a. strumia , f. r. torres , f. l. villante and f. vissani , phys .
d * 80 * , 043007 ( 2009 ) [ arxiv:0907.1891 [ hep - ph ] ] .
g. pagliaroli , f. rossi - torres and f. vissani , astropart .
phys . * 33 * , 287 ( 2010 ) [ arxiv:1002.3349 [ hep - ph ] ] .
loredo and d.q .
lamb , phys .
d * 65 * ( 2002 ) 063002 .
e. nardi and j.i .
zuluaga , phys .
d * 69 * ( 2004 ) 103002 .
e. nardi and j.i .
zuluaga , nucl .
b * 731 * ( 2005 ) 140 .
zuluaga , phd thesis defended on 2005 at the univ .
de antioquia , medellin , astro - ph/0511771v2 .
j. ellis , h. -t .
janka , n. e. mavromatos , a. s. sakharov and e. k. g. sarkisyan , phys .
d * 85 * , 105028 ( 2012 ) [ arxiv:1202.0248 [ hep - ph ] ] .
f. vissani , g. pagliaroli and f. rossi - torres , int .
j. mod .
d * 20 * , 1873 ( 2011 ) .
r. laha , j. f. beacom and s. k. agarwalla , arxiv:1412.8425 [ hep - ph ] . c. lujan - peschard , g. pagliaroli and f. vissani , jcap * 1407 * , 051 ( 2014 ) [ arxiv:1402.6953 [ astro-ph.sr ] ] .
r. laha and j. f. beacom , phys .
d * 89 * , no .
6 , 063007 ( 2014 ) [ arxiv:1311.6407 [ astro-ph.he ] ] . h. -t .
janka , f. hanke , l. huedepohl , a. marek , b. mueller and m. obergaulinger , ptep * 2012 * , 01a309 ( 2012 ) [ arxiv:1211.1378 [ astro-ph.sr ] ] .
m. t. .keil , g. g. raffelt and h. -t .
janka , astrophys . j. * 590 * , 971 ( 2003 ) [ astro - ph/0208035 ] .
m. kachelriess , r. tomas , r. buras , h. -t . .janka
, a. marek and m. rampp , phys .
d * 71 * , 063003 ( 2005 ) [ astro - ph/0412082 ] .
a. s. dighe and a. y. .smirnov , phys .
d * 62 * , 033007 ( 2000 ) [ hep - ph/9907423 ] . c. rubbia , 1977 , cern - ep - int-77 - 08 .
s. amoruso _ et al . _
[ icarus collaboration ] , eur .
j. c * 33 * , 233 ( 2004 ) [ hep - ex/0311040 ] .
d. v. forero , m. tortola and j. w. f. valle , phys .
d * 86 * , 073012 ( 2012 ) [ arxiv:1205.4018 [ hep - ph ] ] .
elena giusarma , roland de putter , shirley ho , olga mena , phys.rev .
* d88 * , 063515 ( 2013 ) , arxiv:1306.5544 [ astro - ph ] . | this work presents an upper bound on the neutrino mass using the emission of @xmath0 from the neutronization burst of a core collapsing supernova at 10 kpc of distance and a progenitor star of 15 m@xmath1 .
the calculations were done considering a 34 kton liquid argon time projection chamber similar to the far detector proposal of the long baseline neutrino experiment ( lbne ) .
we have performed a monte carlo simulation for the number of events integrated in 5 ms bins .
our results are @xmath2 ev and @xmath3 ev , at 95% c.l , assuming normal hierarchy and inverted hierarchy , respectively .
we have analysed different configurations for the detector performance resulting in neutrino mass bound of @xmath4 ev . |
one of the primary motivations for studying the production of mesons off nucleons is to investigate the structure and properties of nucleon resonances and , in the case of heavy - meson productions , to learn about hadron dynamics at short range . in particular , a clear understanding of the production mechanisms of mesons heavier than the pion still requires further theoretical and experimental investigation .
apart from pion production , the majority of theoretical investigations of meson - production processes are performed within phenomenological meson - exchange approaches .
such approaches force us to correlate as many independent processes as possible within a single model if one wishes to extract meaningful physics information .
indeed , this is the basic motivation behind the coupled - channel approaches . here , we present some selected results of our investigation of the @xmath0- and @xmath1-meson productions in both the photon- and hadron - induced reactions , @xmath2 based on a relativistic meson - exchange approach , where @xmath3 .
we consider the three reactions mentioned above in the following manner : the photoproduction reaction is calculated by considering the @xmath4- , @xmath5- and @xmath6-channel feynman diagrams plus a generalized contact term @xcite , which ensures gauge invariance of the total amplitude , in addition to accounting for the final - state interaction ( fsi ) effects .
( see ref .
@xcite for details . )
the @xmath7 reaction is described in the tree - level approximation including the @xmath4- , @xmath5- , and @xmath6-channels . to the extent that this reaction is dominated by the excitation of the @xmath8 resonance at least for energies close to the threshold
, this should be a reasonable approximation if we confine ourselves to energies not too far from the threshold .
the @xmath9 process is calculated in the distorted - wave born approximation ( dwba ) , where both the @xmath10 fsi and the initial - state interaction ( isi ) are taken into account explicitly @xcite .
the @xmath10 fsi is known to be responsible for the dominant energy dependence observed in the total cross section ( apart from the dependence due to the phase space ) arising from the very strong interaction in the @xmath11-wave states at very low energies . as for the basic meson - production amplitudes for the reactions in eq .
( [ eq:1 ] ) , our model @xcite includes the nucleonic , mesonic , and nucleon resonance currents which are derived from the relevant effective lagrangians .
the free parameters of our model the resonance parameters , the @xmath12 coupling constant , and the cutoff parameter @xmath13 at the electromagnetic vector - meson exchange vertex are fixed such as to reproduce the available data in a global fitting procedure of the reaction processes listed in eq .
( [ eq:1 ] ) .
the details of the present approach are fully described in refs .
@xcite . ultimately , our goal is to perform a more complete model calculation where the relevant fsis are taken into account explicitly .
however , before undertaking a complex calculation that couples many channels , we need to learn some of the basic features of meson productions within a simplified model where these basic features may be revealed and analyzed in an easier manner . in this
regard , one of the purposes of the present investigation is to show that consideration of the hadronic reactions @xmath14 , in conjunction with more basic two - body reactions , would greatly help in the study of nucleon resonances , especially , in imposing much stricter constraints on the extracted resonance - nucleon - meson ( @xmath15 ) coupling strength involving a meson @xmath16 other than the pion . in fact , our current knowledge of the branching ratios of the majority of the known resonances is still very limited @xcite .
in this section we discuss some selected results for @xmath0-meson production obtained according to the procedure described in the previous section .
we refer the details and the numerical results to ref .
@xcite . for photoproduction of @xmath0 off nucleons ,
the available database is considerable , in particular , for the proton target .
this covers not only the total and differential cross section data over a wide range of energy starting from threshold but also the beam and target asymmetry data .
when combined , these data sets offer better opportunities for investigating the properties of nucleon resonances .
in particular , much more detailed studies than in the past are possible now for resonances that may perhaps couple strongly to @xmath18 but only weakly to @xmath19 . in this respect ,
the recent data on the quasi - free @xmath20 process @xcite have attracted much interest in @xmath0-production processes in connection to the possible existence of a narrow ( crypto - exotic ) baryon resonance with a mass near @xmath21 gev , which is still under debate . to date , the narrow bump structure in the @xmath20 reaction cross section was reported by several groups @xcite . in this work ,
we restrict our discussion to the @xmath17 reaction .
as far as the resonance currents are concerned , we follow the strategy adopted in ref .
@xcite to include the resonances one by one until a reasonable fit is achieved .
we found that the basic features of the total and differential cross section data as well as the beam asymmetry data can be described by considering the @xmath8 , @xmath22 , @xmath23 , @xmath24 , and @xmath25 resonances .
inclusion of more resonances , especially , the higher spin resonances ( @xmath26 and @xmath27 ) did not change the overall quality of the fit .
the spin-3/2 resonances are important in describing the measured angular distributions in the range of @xmath28 gev and the beam asymmetry .
but we found that the magnitude of the measured beam asymmetry near threshold requires a more dedicated analysis .
also , it is difficult to explain the measured target asymmetry ; the same difficulty has been encountered also in other investigations @xcite . we emphasize that the resonance parameters in our model are highly correlated to each other and that the existing data are insufficient to establish a unique set of these parameters . in ref .
@xcite , cross section data at energies close to threshold have been used to extract the @xmath29 coupling constant assuming that all the other relevant reaction mechanisms are known at these energies .
close to threshold this reaction is dominated by the @xmath8 resonance contribution . here , as an alternative
, we show that the beam asymmetry at higher energies can also constrain this coupling constant .
the relatively small cross sections measured at higher energies and backward angles favor a very small @xmath5-channel nucleonic current contribution , compatible with a vanishing @xmath29 coupling constant .
however , we found that the cross section data at higher energies can not constrain the @xmath29 coupling constant uniquely because of a possible interference effect with the nucleon resonance current .
most data for the @xmath7 reaction have been obtained in the 1960s through 1970s ; they are rather scarce and less accurate than for the @xmath17 and @xmath31 reactions .
recently , the crystal ball collaboration has measured the differential and total cross sections of this reaction near threshold @xcite .
theoretically , this reaction has been studied mostly in conjunction with other reactions in a combined analysis or in coupled - channel approaches ( see , e.g. , the references quoted in ref .
@xcite ) in order to constrain some model parameters . recently , zhong _ et al . _
@xcite have extended the chiral constituent quark model for meson photoproduction to this reaction .
also , @xmath7 has been studied within a heavy - baryon chiral perturbation theory @xcite .
et al . _
@xcite have investigated the role of @xmath7 on the @xmath18 scattering length within a coupled - channel analysis of this reaction and the elastic @xmath32 scattering . in the present work , the total cross section is nicely reproduced up to @xmath33 gev , where it is dominated by the @xmath34 resonances , especially , the @xmath8 . in our study , the measured total cross section at higher energies are underestimated due to the absence of the @xmath35 contribution via the coupled channel @xcite .
we also found that the @xmath25 resonance is important in reproducing the structure exhibited by the measured angular distributions at higher energies .
this process is particularly relevant in connection to the role of the @xmath37 fsi .
most of the existing calculations take into account the effects of the @xmath10 fsi in one way or another , which is well - known to influence the energy dependence of the cross section near threshold .
calculations which include the @xmath37 fsi to the lowest order reproduce the bulk of the energy dependence exhibited by the data . however , they fail to explain the @xmath38 invariant mass distribution measured by the cosy - tof @xcite and cosy-11 collaborations @xcite . in ref .
@xcite , the importance of the three - body nature of the final state in the @xmath11-wave has been emphasized in order to account for the observed @xmath38 invariant mass distribution .
other authors have suggested an extra energy dependence in the basic production amplitude @xcite to reproduce the existing data . yet
another possibility has been offered , which is based on a higher partial wave ( @xmath39-wave ) contribution @xcite .
we observe that all that is required to reproduce the measured @xmath38 invariant mass distribution is an extra @xmath40 dependence , where @xmath41 denotes the relative momentum of the two protons in the final state .
obviously , this can be achieved either by an @xmath11-wave or by a @xmath39-wave contribution .
note that the @xmath10 @xmath39-wave @xmath42 ) can also yield a flat proton angular distribution as observed in the corresponding data . as pointed out in ref .
@xcite , the measurement of the spin correlation function should settle down the question on the @xmath11-wave versus @xmath39-wave contributions in a model independent way .
although the model calculation of ref .
@xcite , based on a stronger @xmath39-wave contribution , reproduces nicely the shape of the measured @xmath38 invariant mass distributions , it underestimates the total cross section data near threshold ( @xmath43 mev ) .
here , we discuss the new results based on a combined analysis of the @xmath17 , @xmath30 , and @xmath44 reactions , which reproduce the currently existing data on @xmath44 .
( see ref .
the major difference from the previous calculation @xcite is a much stronger resonance contributions at low energies .
we emphasize that it is not possible to determine uniquely the resonance parameters involved in the present work . in this process ,
in addition to the couplings of the nucleon resonances to pseudoscalar mesons , we need their couplings to the vector mesons @xmath45 and @xmath46 . as a consequence ,
we obtain either the dominance of the @xmath34 resonance or @xmath47 resonance in the cross sections at low energies depending on different interference patterns among the various meson exchanges in the excitation of these resonances .
the analyzing power is found to be very sensitive to these interference patterns .
unfortunately , however , the existing data for this observable are not accurate enough to disentangle them and more precise measurements of the analyzing power is strongly required .
in contrast to the @xmath0 meson production , not many data exist for the @xmath1 meson production reactions . in photoproduction ,
only the earlier total cross section data from abbhhm collaboration @xcite were available until late 1990s .
more recently , the differential cross section data from the saphir @xcite and the clas @xcite collaborations became available .
currently , the beam asymmetry in this reaction is being measured by the clas collaboration . also , the cb - elsa collaboration has reported measurements of cross sections in quasi - free @xmath1 photoproduction process on neutron @xcite . for @xmath48
, there are data for total cross sections for excess energies up to @xmath49 mev and for the differential cross sections at two excess energies .
also , the @xmath38 and @xmath50 invariant mass distribution data will become available soon @xcite .
but there are still no data for @xmath51 , apart from the very preliminary total cross section data by the cosy-11 collaboration @xcite . with the only data set available so far for @xmath52 , it is not possible to have stringent constraints on the resonance parameters .
in fact , there are many sets of parameter values which can reproduce the data equally well @xcite . in order to impose more stringent constraints on these parameters ,
one requires more exclusive and precise data , in particular , for the spin observables , such as the beam asymmetry .
a common feature of the model results corresponding to different sets of parameters is the bump structure , which is likely to be due to the @xmath53 and/or @xmath54 resonance .
therefore , experimental confirmation of this structure will be very interesting .
the angular distributions at higher energies and at large scattering angles restrict the value of the @xmath55 coupling constant to be not much greater than @xmath56 , which is of particular interest in connection to the spin content of the nucleon . in this reaction ,
the dominant production mechanism is found to be the excitation of the @xmath34 resonance .
the energy dependence exhibited by the @xmath58 reaction is markedly different from that of @xmath59 .
both the ( measured ) total and differential cross sections are rather well reproduced within our model .
the preliminary data for @xmath38 and @xmath60 invariant mass distributions @xcite are also rather well explained .
it is interesting to note that , within the statistical uncertainties , the @xmath38 invariant mass distributions for @xmath0 and @xmath1 productions exhibit similar features .
the present model also reproduces the very preliminary total cross section data ( upper limit ) for @xmath61 .
from the results for @xmath62 ( @xmath3 ) , it is clear that the present investigation should be extended to include the vector meson productions as well in order to impose more stringent constraints on the resonance parameters . in particular , the @xmath46 and @xmath45 meson production reactions will be useful to constrain the couplings of the nucleon resonances to these mesons which are required to understand the @xmath0- and @xmath1-meson production mechanisms in @xmath10 collisions .
it is a part of our ongoing effort to consider pseudoscalar and vector meson productions within a single model , and , here , we briefly report on the progress made in this direction so far .
the existing total and differential cross section data for the @xmath63 reaction can be described very well , consistently with the @xmath0-meson production reactions considered in this work . however , the recent differential cross section data @xcite for @xmath64 at higher energies and backward angles require a very small nucleonic current contribution within the current approach , leading to a small @xmath65 coupling strength ( or a very soft form factor @xcite ) which is much smaller than those found in the description of other reactions such as @xmath10 scatterings @xcite .
the present model does not consider the fsi ; therefore , this could be a natural explanation for the discrepancy . however , the calculations of refs .
@xcite based on coupled - channel approaches , which take into account the fsi in one way or another , have also shown that a much smaller @xmath65 coupling is needed in order to describe the photoproduction data . here
we also mention that the same problem is found in the description of @xmath45-meson photoproduction within the present model .
these findings indicate that this is not a trivial issue and that it should be addressed in order to understand the mechanism of vector meson photoproduction .
a lot of data for meson production in photon - induced reactions have been accumulated and many new data are expected to be reported soon . the existing world database for meson production ( other than pion ) from two - body _ hadronic _ reactions , however , is limited and most data are old and subject to rather large uncertainties .
currently there is a little effort to improve and expand this database , which limits severely our ability to extract certain information on nucleon resonances .
in particular , information on the various decay rates ( branching ratios ) of these resonances is strongly required .
meson productions in @xmath10 collisions , therefore , can be used to improve this situation : the existing data are rather accurate and the corresponding database can be expanded with the existing facilities .
so , this reaction should be included as a part of hadronic model calculations together with photoproduction reactions aiming at investigating nucleon resonances in conjunction with more basic two - body reactions .
k. nakayama and h. haberzettl , phys .
c * 69 * , 065212 ( 2004 ) ; * 73 * , 045211 ( 2006 ) . h. haberzettl , k. nakayama , and s. krewald , phys .
c * 74 * , 045202 ( 2006 ) .
k. nakayama , j. speth , and t-.s .
h. lee , phys .
c * 65 * , 045210 ( 2002 ) .
k. nakayama , yongseok oh , and h. haberzettl , arxiv:0803.3169 .
w. m. yao _
( particle data group ) , j. phys .
g * 33 * , 1 ( 2006 ) .
v. kuznetsov _ et al .
_ , phys .
* b647 * , 23 ( 2007 ) .
h. shimizu , talk at nstar2007 , http://nstar2007.uni-bonn.de .
l. tiator _ et al .
_ , phys .
c * 60 * , 035210 ( 1999 ) .
l. tiator , c. bennhold , and s. s. kamalov , nucl .
* a580 * , 455 ( 1994 ) . s. prakhov _ et al .
_ , phys .
c * 72 * , 015203 ( 2005 ) .
h . zhong , q. zhao , j. he , and b. saghai , phys .
c * 76 * , 065205 ( 2007 ) .
b. krippa , phys .
c * 64 * , 047602 ( 2001 ) .
r. a. arndt _
_ , phys .
c * 72 * , 045202 ( 2005 ) .
a. gasparyan _ et al .
_ , phys .
c * 68 * , 045207 ( 2003 ) .
m. abdel - bary _ et al .
j. a * 16 * , 127 ( 2003 ) .
p. moskal _
c * 69 * , 025203 ( 2004 ) .
a. fix and h. arenhvel , phys .
c * 69 * , 014001 ( 2004 ) .
a. deloff , phys .
c * 69 * , 035206 ( 2004 ) .
k. nakayama _
et al . _ ,
c * 68 * , 045201 ( 2003 ) .
abbhhm collaboration , phys . rev . * 175 * , 1669 ( 1968 ) ; nucl .
b108 * , 45 ( 1976 ) .
r. pltzke _ et al .
_ , phys . lett .
* b444 * , 555 ( 1998 ) ; j. barth _ et al .
_ , nucl . phys .
* a691 * , 374c ( 2001 )
. m. dugger _
* 96 * , 062001 ( 2006 ) , * 96 * 169905(e ) ( 2006 ) . b. krusche ( cbelsa / taps collaboration ) , in these proceedings .
j. barth _ et al .
j. a * 18 * , 117 ( 2003 ) .
y. oh , a .- i . titov , and t .- s .
h. lee , phys .
c * 63 * , 025201 ( 2001 ) ; y. oh and t .- s .
lee _ ibid . _ * 69 * , 025201 ( 2004 ) .
g. janssen , k. holinde , and j. speth , phys .
c * 54 * , 2218 ( 1996 ) . v. shklyar _
et al . _ ,
c * 71 * , 055206 ( 2005 ) , * 72 * , 019903(e ) ( 2005 ) . m. paris , arxiv:0802.3383 . | the production of @xmath0 and @xmath1 mesons in photon- and hadron - induced reactions are studied in a combined analysis based on an effective lagrangian approach in order to learn about the relevant production mechanisms and the possible role of nucleon resonances in these reactions . |
in the most venerable scenario for forming planetesimals , dust particles in circumstellar gas disks are imagined to settle vertically into thin sublayers ( `` subdisks '' ) sufficiently dense to undergo gravitational instability ( @xcite ; @xcite ; for a review of this and other ways in which planetesimals may form , see chiang & youdin 2010 , hereafter cy10 ) .
along with this longstanding hope comes a longstanding fear that dust remains lofted up by turbulence . even if we suppose that certain regions of the disk are devoid of magnetized turbulence because they are too poorly ionized to sustain magnetic activity @xcite , the dusty sublayer is susceptible to a kelvin - helmholtz shearing instability ( khi ; * ? ? ?
the khi arises because dust - rich gas at the midplane rotates at a different speed from dust - poor gas at altitude .
the background radial pressure gradient @xmath9 causes dust - free gas at disk radius @xmath10 to rotate at the slightly non - keplerian rate @xmath11 where @xmath12 is the kepler angular frequency , @xmath13 is a dimensionless measure of centrifugal support by pressure , and @xmath14 is the density of gas ( e.g. , * ? ? ?
* ; * ? ? ?
the numerical evaluation is based on the minimum - mass solar nebula derived by cy10 . unlike dust - free gas , dust - rich gas is loaded by the extra inertia of solids and must rotate at more nearly the full keplerian rate to remain in centrifugal balance .
variations in the dust - to - gas ratio @xmath15 with height @xmath16 result in a vertical shear @xmath17 from which free energy is available to overturn the dust layer .
the shearing rate across a layer of thickness @xmath18 is given to order of magnitude by @xmath19 where @xmath20 at the midplane and @xmath21 at altitude ( for more details see cy10 , or
[ ssec : wcfe][ssec : initialconditions ] of this paper ) .
for @xmath22 the velocity difference @xmath23 saturates at a speed @xmath24 , well below the gas sound speed @xmath25 . that the flow is highly subsonic motivates what simulation methods we employ in our study .
we might expect the flow to be stabilized if the brunt - visl frequency @xmath26 of buoyant vertical oscillations is much larger than the vertical shearing rate . for the order - of - magnitude evaluation in ( [ eqn : brunt ] ) we approximate the vertical gravitational acceleration @xmath27 as the vertical component of stellar gravity @xmath28 ( no self - gravity ) , and the density gradient @xmath29 .
the last approximation relies in part on the dust density @xmath30 changing over a lengthscale @xmath18 much shorter than the gas scale height .
both @xmath31 and @xmath32 shrink as @xmath1 decreases .
, the vertical shearing and brunt frequencies reach minima set by pressure and temperature gradients in gas ( see , e.g. , * ? ? ?
the limit @xmath33 is not relevant for our study and not captured by either ( [ eqn : verticalshear ] ) or ( [ eqn : brunt ] ) . ] for two - dimensional , heterogeneous , unmagnetized flow , a necessary but not sufficient condition for instability is given by the richardson number : @xmath34 ( @xcite ; see the textbook by @xcite ) .
the richardson number is simply the square of the ratio of the stabilizing brunt frequency ( [ eqn : brunt ] ) to the destabilizing vertical shearing frequency ( [ eqn : verticalshear ] ) .
the critical value of 1/4 arises formally but can also be derived heuristically by energy arguments ( e.g. , * ? ? ?
* ) . the richardson criterion does not formally apply to our dusty subdisk , which represents a three - dimensional flow : the khi couples vertical motions to azimuthal motions , while the coriolis force couples azimuthal motions to radial motions .
( for how the richardson criterion may not apply to magnetized flows , see @xcite . )
nevertheless we may hope the richardson number is useful as a guide , as previous works have assumed @xcite . in this spirit
let us use the richardson criterion to estimate the thickness of a marginally kh - unstable dust layer .
substitution of ( [ eqn : verticalshear ] ) and ( [ eqn : brunt ] ) into ( [ eqn : richardson ] ) reveals that ) which approximates the vertical shear , are each smaller than their counterparts given by ( * ? ? ?
* , first full paragraph ) by a factor of @xmath35 .
this is because @xcite evaluate quantities deep inside the layer , within a density cusp at the midplane , whereas we are interested in quantities averaged across the entire layer .
the difference does not change either our conclusions or theirs . ]
@xmath36 since the gas scale height @xmath37 and @xmath38 , equation ( [ eqn : deltaz ] ) indicates that for @xmath39 the marginally unstable dust sublayer is @xmath6@xmath40 times as thin as the gas disk in which it is immersed .
those kh - unstable modes that disrupt the layer should have azimuthal wavelengths and by extension radial wavelengths , because the kepler shear turns azimuthal modes into radial ones that are comparable to @xmath18 .
shorter wavelength modes can not overturn the layer , while longer wavelength modes grow too slowly @xcite .
how does disk rotation affect the development of the khi ? in a linear analysis , @xcite highlight the role played by the keplerian shear , characterized by the strain rate @xmath41 in limiting the growth of kh - unstable modes .
the radial shear is implicated because azimuthal motions excited by the khi are converted to radial motions by the coriolis force ; moreover , the non - axisymmetric pattern excited by the khi is wound up , i.e. , stretched azimuthally by the radial shear .
the kepler rate @xmath42 is at least as large as @xmath32 , and can dominate the latter when @xmath1 is small .
this suggests that @xmath0 does not capture all the relevant dynamics
a concern already clear on formal grounds . in this paper
we address this concern head - on , using fully 3d numerical simulations to assess the role of the richardson number in governing the stability of the dust layer .
three - dimensional shearing box simulations of the khi in dusty subdisks , performed in the limits that dust is perfectly coupled to gas and disk self - gravity is negligible , demonstrate the importance of the kepler shear .
compared to rigidly rotating disks @xcite , radially shearing disks are far more stable @xcite .
the relevance of @xmath0 , or lack thereof , may be assessed by simulating flows with initially spatially constant @xmath0 @xcite , and varying @xmath0 from run to run to see whether dust layers turn over .
* hereafter c08 ) found that when @xmath39 , dust layers for which @xmath43 overturn , while those for which @xmath44 do not .
in retrospect , we might have anticipated this result , that the critical value @xmath2 dividing stable from unstable runs lies near the canonical value of 1/4 , at least for @xmath39 , because in this regime of parameter space all the frequencies of the problem are comparable to each other : latexmath:[$|\partial v_\phi/\partial z| \sim \omega_{\rm brunt } \sim @xmath39 and @xmath461 .
but other simulations of c08 also make clear that @xmath0 does not alone determine stability under all circumstances . for @xmath470.4
, @xmath2 was discovered to drop substantially to @xmath60.02 ( see his runs s9s12 ) .
@xcite speculated that the baroclinic nature of the flow may be responsible @xcite , but no details were given .
in addition to being left unexplained , the findings of c08 require verification .
parameter space was too sparsely sampled to discern trends with confidence .
concerns about numerics e.g . , biases introduced by box sizes that were too small , resolutions too coarse , and runs terminated too early also linger .
at least one numerical artifact marred the simulations of c08 : the khi manifested first at the `` co - rotation '' radius where the mean azimuthal flow speed was zero ( see his figure 8) .
but in a shearing box , by galilean invariance , there should be no special radius .
it was suspected , but not confirmed , that errors of interpolation associated with the grid - based advection scheme used by c08 artificially suppressed the khi away from co - rotation . for the problem at hand , the spectral code developed by @xcite and modified by ( * ? ? ?
* hereafter b09 ) to treat mixtures of dust and gas is a superior tool to the grid - based zeus code utilized by c08 . working in fourier space rather than configuration space , the simulations of the khi by b09
did not betray the co - rotation artifact mentioned above .
spectral methods , often used to model local ( wkb ) dynamics , are appropriate here because the structures of interest in the subdisk have dimensions tiny compared to the disk radius ( by at least a factor @xmath48 according to equation [ eqn : deltaz ] ) and even the gas scale height . at the same computational expense ,
spectral algorithms typically achieve greater effective spatial resolution than their grid - based counterparts @xcite .
another advantage enjoyed by the b09 code is that it employs the anelastic approximation , which is designed to treat subsonic flows such as ours .
having filtered away sound waves , anelastic codes are free to take timesteps set by how long it takes fluid to advect across a grid cell ( which themselves move at the local orbital velocity in a shearing coordinate system ) . by contrast , codes such as zeus take mincing steps limited by the time for sound waves to cross a grid cell .
the latter constraint is the usual courant condition for numerically solving problems in compressible fluid dynamics .
it was unnecessarily applied by c08 to a practically incompressible flow .
in this paper we bring all the advantages of the spectral , anelastic , shearing box code of b09 to bear on the problems originally addressed by c08 .
we assess numerically the stability of flows characterized by constant richardson number @xmath0 , systematically mapping out the stability boundary in the parameter space of @xmath0 , midplane dust - to - gas ratio @xmath1 , and bulk metallicity @xmath49 ( the height - integrated surface density ratio of dust to gas ) .
though our simulations may still be underresolved , we rule out box size as a major influence on our results .
we offer some new insight into why @xmath0 is not a sufficient predictor of stability . and in the restricted context of our constant @xmath0 flows
, we assess the conditions necessary for the midplane to become dense enough to trigger gravitational instability on a dynamical time .
following c08 and b09 , we continue to work in the limit that dust is perfectly coupled to gas , i.e. , in the limit that particles are small enough that their frictional stopping times @xmath50 in gas can be neglected in comparison to the dynamical time @xmath51 .
the perfect coupling approximation allows us to screen out the streaming instability which relies on a finite stopping time and which is most powerful when particles are marginally coupled , i.e. , when @xmath521 @xcite .
numerical simulations have shown that when an order - unity fraction of the disk s solids is in particles having @xmath531 , the streaming instability clumps them strongly and paves the way for gravitational instability ( e.g. , * ? ? ?
* ; * ? ? ?
the particle sizes corresponding to @xmath54 depend on the properties of the background gas disk , as well as on the particle s shape and internal density ; under typical assumptions , marginally coupled particles are decimeter to meter - sized .
it remains debatable whether a substantial fraction of a disk s solid mass is in marginally coupled particles at the time of planetesimal formation , as current proposals relying on the streaming instability assume .
particle size and shape distributions are not well constrained by observations ( though see , e.g. , @xcite , who showed that centimeter - wavelength fluxes from a few t tauri stars are consistent with having been emitted by predominantly centimeter - sized particles )
. measuring @xmath55 in disks also requires knowing the gas density , but direct measurements of the gas density at disk midplanes do not exist . marginally coupled particles
sometimes referred to as `` meter - sized boulders''also face the longstanding problem that they drift onto the central star too quickly , within hundreds of years from distances of a few au in a minimum - mass disk .
@xcite claimed to solve this problem by agglomerating all the boulders into ceres - mass planetesimals via the streaming instability before they drifted inward .
their simulation presumed , however , that all of the disk s solids began boulder - sized .
the concern we have is that even if particle - particle sticking could grow boulders ( and sticking is expected to stall at centimeter sizes ; @xcite ; cy10 ) , the disk s solids may not be transformed into boulders all at once .
rather , marginally coupled bodies may initially comprise a minority population on the extreme tail of the particle size distribution .
unless they can transform themselves from a minority to a majority within the radial drift timescale , they would be lost from the nebula by aerodynamic drag . by focussing on the dynamics of the smallest , most well entrained particles having @xmath56
, our work complements that which relies on the streaming instability .
we would argue further that the well coupled limit is potentially more relevant for planet formation .
if even the smallest particles having sizes @xmath57 cm can undergo gravitational collapse to form kilometer - sized or larger planetesimals , nature will have leapfrogged over the marginally coupled regime , bypassing the complications and uncertainties described above .
particle clumping is not restricted to marginally coupled particles via the streaming instability .
small @xmath55 particles also clump within the interstices of turbulent , high vorticity eddies ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ; for a review , see cy10 ) .
this particle concentration mechanism presumes some gas turbulence , which may be present in the marginally kh - unstable state to which dust settles .
our simulations can not capture this phenomenon .
however , on the scales of interest to us , turbulent clumping might only be of minor significance . particles of given @xmath50 are concentrated preferentially by eddies that turn over on the same timescale .
thus the degree of concentration depends sensitively on particle size and the turbulent spectrum . at least in kolmogorov turbulence ,
the smallest eddies concentrate particles most strongly because they have the greatest vorticity .
the smallest eddies at the inner scale of kolmogorov turbulence have sizes @xmath58 , where @xmath59 is the molecular kinematic viscosity , and @xmath60 and @xmath61 are the turnover time and velocity of the largest , outer scale eddy . given @xmath62 m / s , @xmath63 , and values of @xmath59 based on the nebular model of cy10 , we estimate that @xmath64 cm .
this is far smaller than the sublayer thicknesses @xmath65 cm considered in this paper .
moreover , the lifetimes of the particle clumps on a given eddy length scale should roughly equal the eddy turnover times , which for the smallest eddies are of order @xmath66 s. we do not expect such rapid fluctuations in particle density , occurring on such small length scales , to affect significantly the evolution of the slower , larger scale khi .
turbulent clumping may only serve as a source of noise on tiny scales .
the possibility that turbulent clumping could still be significant on larger scales is still being investigated @xcite .
the perfect coupling approximation prevents us from studying how particles sediment out of gas into dusty sublayers , but it does not stop us from identifying what kinds of sublayers are dynamically stable to the khi . a subdisk with a given density profile is either dynamically stable or it is not , and we can run the b09 code for many dynamical times ( typically 60 or more ) to decide the answer . in a forthcoming paper
we will combine the b09 code with a settling algorithm that will permit us to study how dust settles from arbitrary initial conditions , freeing us from the assumption that the density profile derives from a constant richardson number .
our numerical methods , including our rationale for choosing box sizes and resolutions , are described in [ sec : method ] .
results are presented in [ sec : results ] and discussed in [ sec : discussion ] .
the equations solved by the b09 code are rederived in [ ssec : wcfe ] .
initial conditions for our simulations are given in
[ ssec : initialconditions ] .
the code itself is briefly described in [ ssec : code ] .
our choices for box size and resolution are explained in [ ssec : bsnr ] .
the equations we solve are identical to equations ( 12a e ) of b09 .
we outline their derivation here , filling in steps skipped by b09 , adjusting the notation , and providing some clarifications .
this section may be skimmed on a first reading .
we begin with the equations for an ideal gas perfectly coupled to pressureless dust in an inertial frame : @xmath67 where @xmath68 is the convective derivative , @xmath69 is the density of gas ( dust ) , @xmath70 is the gas pressure , and @xmath71 is the gas temperature . under the assumption that they are perfectly coupled , gas and dust share the same velocity @xmath72 , and the dust - to - gas ratio is conserved in a lagrangian sense .
the background potential is provided by the central star of mass @xmath73 : @xmath74 , where @xmath10 is the cylindrical radius and @xmath16 is the vertical distance above the disk midplane .
there are five equations for the five flow variables @xmath72 , @xmath14 , @xmath30 , @xmath70 , and @xmath71 .
the thermodynamic constants include the specific heat @xmath75 at constant volume , the ideal gas constant @xmath76 , the specific heat @xmath77 at constant pressure , and @xmath78 .
equation ( [ eqn : adiabat ] ) is equivalent to the condition that the flow be isentropic [ @xmath79 .
the code which solves the fluid equations actually employs an artificial hyperviscosity to damp away the smallest scale perturbations ( [ ssec : code ] ) ; in writing down equations ( [ eqn : momentum])([eqn : idealeos ] ) , we have omitted the hyperviscosity terms for simplicity .
we move to a frame co - rotating with dust - free gas at some fiducial radius @xmath80 .
this frame has angular frequency @xmath81 given by ( [ eqn : omegaf ] ) with @xmath82 .
we define a velocity @xmath83 using the pressure support parameter @xmath84 , as given by ( [ eqn : eta ] ) : @xmath85 the velocity @xmath86 is the difference in azimuthal velocity between a strictly keplerian flow and dust - free gas ; it is the maximum possible difference in velocity , attained at large @xmath1 , between gas at the midplane and gas at altitude .
the quantities @xmath86 , @xmath84 , and the background radial pressure gradient are equivalent ; specifying one specifies the other two .
our numerical models are labeled by @xmath86 .
in addition to moving into a rotating frame , we also replace the usual cylindrical coordinates @xmath87 with local cartesian coordinates @xmath88 , @xmath89 , and @xmath16 . and @xmath87 . ] keeping terms to first order in @xmath90 ( see the discussion surrounding equation [ eqn : deltaz ] ) and dropping curvature terms , the momentum equation ( [ eqn : momentum ] ) reads @xmath91 where @xmath92 ( @xmath93 ) . on the right - hand side ,
the first term is the coriolis acceleration , the second combines centrifugal and radial gravitational accelerations , the third represents the vertical gravitational acceleration from the star , and the last term arises from the centrifugal acceleration in a frame rotating at @xmath94 .
the remaining fluid equations appear the same as ( [ eqn : gascontinuity])([eqn : idealeos ] ) , except that @xmath72 is now measured in a ( rigidly ) rotating frame .
we measure all flow variables relative to a time - independent reference state ( subscripted `` ref '' ) : @xmath95 @xmath96 @xmath97 @xmath98 @xmath99 the reference state is defined as follows .
it is dust - free ( @xmath100 ) and has constant gas temperature @xmath101 .
the gas in the reference state does not shear , either in the radial or vertical directions , but rotates with a fixed angular frequency @xmath81 in the inertial frame ( hence @xmath102 in the rotating frame ) .
in the reference state there exists a radial pressure gradient directed outward @xmath103 and a vertical pressure gradient balanced by vertical tidal gravity @xmath104 equation ( [ eqn : dpdzref ] ) together with equation ( [ eqn : idealeos ] ) and the assumption of constant @xmath101 implies that the reference gas density @xmath105 and pressure @xmath106 have gaussian vertical distributions in @xmath16 with scale height @xmath107 . for simplicity
we neglect the radial density gradient ( @xmath108 ) , as did b09 .
this reference state should not be confused with our equilibrium states of interest ( [ ssec : initialconditions ] ) , which do shear and which do contain dust .
the reference state merely serves as a fiducial .
the flows of interest are subsonic .
mach numbers @xmath109 peak at @xmath110 for gas sound speeds @xmath111 km / s at @xmath112 au .
such flow is nearly incompressible : @xmath113 . invoking the anelastic approximation
, we keep only terms leading in @xmath114 in any given equation .
equations ( [ eqn : gascontinuity ] ) , ( [ eqn : noslip ] ) , and ( [ eqn : idealeos ] ) reduce to : @xmath115 where we define @xmath116 and the pressure - like enthalpy @xmath117 , and henceforth for convenience drop all tildes on @xmath30 , @xmath118 , and @xmath72 ( but not the other variables related to gas ) . the rightmost equalities of ( [ eqn : gascontinuity_an ] ) , (
[ eqn : noslip_an ] ) , and ( [ eqn : idealeos_an ] ) match equations ( 12b ) , ( 12c ) , and ( 12e ) of b09 .
the anelastic approximation has been employed in the study of atmospheric convection @xcite , stars @xcite , and vortices in protoplanetary disks @xcite . by eliminating the time derivative in the continuity equation ( [ eqn : gascontinuity_an ] ) , we effectively `` sound - proof '' the fluid .
the simulation timestep is not limited by the sound - crossing time but rather by the longer advection time .
we rewrite our energy equation ( [ eqn : adiabat ] ) as follows : replace @xmath119 with @xmath120 to find that @xmath121 where for the second line we dropped @xmath122 in comparison to @xmath123 , and for the third line we replaced @xmath124 using ( [ eqn : dpdrref ] ) and ( [ eqn : dpdzref ] ) .
equation ( [ eqn : adiabat_an ] ) matches ( 12d ) of b09 except that for the right - hand side he has a coefficient equal to @xmath125 , which we have set to unity .
finally , to recover the form of the momentum equation ( 12a ) of b09 , first consider the pressure acceleration and isolate the contribution from dust - free gas ( @xmath126 ) : @xmath127 now expand @xmath128 where for the last line we used ( [ eqn : dpdrref ] ) and ( [ eqn : dpdzref ] ) .
insertion of ( [ eqn : isolate ] ) and ( [ eqn : napbfinal ] ) into ( [ eqn : momentum_ss ] ) yields the anelastic momentum equation ( 12a ) of b09 : @xmath129 .
\nonumber \\ \end{aligned}\ ] ] which isolates the driving term due to dust .
equilibrium initial conditions ( superscripted `` @xmath130 '' ) are specified by five functions : @xmath131 , @xmath132 , @xmath133 , @xmath134 , and @xmath135 . for @xmath136 , we use flows characterized by a globally constant richardson number @xcite .
the conditions @xmath137 constant , @xmath138 , and @xmath139 ( no self - gravity ) yield @xmath140^{1/2 } - 1,\ ] ] where @xmath1 is the initial midplane dust - to - gas ratio and @xmath141 is a characteristic dust height .
the dust density peaks at the midplane and decreases to zero at @xmath142 which is consistent with our order - of - magnitude expression ( [ eqn : deltaz ] ) .
neither equation ( [ eqn : mu ] ) nor the code accounts for self - gravity and therefore we are restricted to modeling flows whose densities are less than that required for the toomre parameter of the subdisk to equal unity ( cy10 ; see also [ sec : discussion ] ) . for the minimum - mass disk of cy10 ,
this restriction is equivalent to @xmath143 .
input model parameters include @xmath1 , @xmath0 , and @xmath83 . for the gas
, we assume @xmath144 ( initially isothermal ) and solve vertical hydrostatic equilibrium for @xmath145 ( the @xmath16-component of equation [ eqn : momentumanelastic ] ) : @xmath146 the functional form for @xmath147 is not especially revealing and so we do not write it out here .
for simplicity we assume that @xmath145 does not depend on @xmath148 . from @xmath145 and @xmath149
it follows from ( [ eqn : idealeos_an ] ) that @xmath150 the fractional deviations @xmath151 and @xmath152 from the reference state are very small , of order @xmath153 .
it remains to specify @xmath154 . using the conditions on @xmath145 stated above , we solve for the equilibrium ( steady - state ) solution to equation ( [ eqn : momentumanelastic ] ) : @xmath155 \vmax \ , .
\label{eqn : vy}\end{aligned}\ ] ] in our reference frame rotating with the velocity of dust - free gas at @xmath156 , the first term on the right side of ( [ eqn : vy ] ) accounts for the standard kepler shear , while the second term describes how dust , which adds to inertia but not pressure , speeds up the gas .
to @xmath136 we add random perturbations @xmath157 \ , .\ ] ] the amplitude @xmath158 ) is constructed in fourier space so that each fourier mode has a random phase and an amplitude inversely proportional to the horizontal wavenumber : @xmath159 . because our box sizes are scaled to @xmath160 , our fourier noise amplitudes are largest on scales comparable to the dust layer thickness .
thus those modes which are most likely to overturn the layer are given the greatest initial power .
the perturbations are also chosen to be antisymmetric about the @xmath148-axis so that no extra energy is injected into the system .
we take the root - mean - squared amplitude @xmath161 of the perturbations to be @xmath162 or @xmath163 . in summary ,
three input parameters @xmath1 , @xmath0 , and @xmath86 determine our isothermal equilibrium initial conditions ( equations [ eqn : mu ] , [ eqn : enthalpy ] , and [ eqn : vy ] ) .
the equilibrium solution for @xmath164 is then perturbed ( equation [ eqn : perturbations ] ) by a root - mean - squared fractional amount @xmath165 .
the parameters of primary interest are @xmath1 and @xmath0 . for the remaining parameters @xmath86 and @xmath165 we consider three possible combinations : @xmath166 for our standard runs ; @xmath167 to probe larger initial perturbations ; and @xmath168 to assess the effect of a stronger radial pressure gradient .
note that specifying @xmath1 and @xmath0 ( and @xmath86 , though this last variable is fixed for all of our standard runs ) specifies the entire dust and gas vertical profiles , @xmath169 and @xmath170 , and by extension the bulk height - integrated metallicity , @xmath171 .
we do not give an explicit expression for @xmath49 because it is cumbersome and not particularly revealing .
the bulk metallicity is in some sense the most natural independent variable because its value is given by the background disk ( for ways in which the bulk metallicity may change , e.g. , by radial particle drifts , see cy10 ) .
we will plot our results in the space of @xmath1 , @xmath0 , and @xmath49 , keeping in mind that only two of these three variables are independent .
we use the spectral , anelastic , shearing box code developed by @xcite and modified by b09 to simulate well - coupled gas and dust . the code employs shearing periodic boundary conditions in @xmath10 , periodic boundary conditions in @xmath172 , and closed lid boundaries in @xmath16 ; the vertical velocity @xmath173 is required to vanish at the top and bottom of the box ( @xmath174 ) .
spectral methods approximate the solution to the fluid equations as a linear combination of basis functions .
the basis functions describe how the flow varies in space , and the coefficients of the functions are determined at every timestep . for each of the periodic dimensions ,
a standard fourier basis is used , while for the vertical direction , chebyshev polynomials are employed .
whereas in @xmath10 and @xmath172 grid points are evenly spaced , the use of chebyshev polynomials in @xmath16 has the effect that vertical grid points are unevenly spaced ; points are concentrated towards the top and bottom boundaries of the box , away from the midplane where the dust layer resides . thus to resolve the dust layer vertically
, we need to increase the number of vertical grid points @xmath175 by an amount disproportionately large compared to the numbers of radial and azimuthal grid points @xmath176 and @xmath177
. see [ ssec : bsnr ] for further discussion .
spectral codes have no inherent grid dissipation ; energy is allowed to cascade down to the smallest resolved length scales through nonlinear interactions . to avoid an energy `` pile - up '' at the highest wavenumbers , we dissipate energy using an artificial hyperviscosity , given in 3.3.3 of @xcite .
simulations satisfy the courant - friedrichs - lewy ( cfl ) condition which states that the cfl number , defined as the code timestep divided by the shortest advection time across a grid cell , be small . in the shearing coordinates in which the code works ,
that advection time is the cell dimension divided by the local velocity over and above the keplerian shear , i.e. , orbital velocities are subtracted off before evaluating the cfl number .
all simulations reported in this paper are characterized by cfl numbers less than about 0.1 .
our standard box dimensions are @xmath178 and the corresponding numbers of grid points are @xmath179 . by scaling our box lengths @xmath180 to @xmath160 and fixing the numbers of grid points @xmath181 , we ensure that each standard simulation enjoys the same resolution ( measured in grid points per physical length ) regardless of @xmath0 , @xmath1 , and @xmath182 .
the vertical extent of the dust layer between @xmath183 is resolved by 22 grid points ( this is less than @xmath184 \times 2z_{\rm max}=32 $ ] because the chebyshev - based vertical grid only sparsely samples the midplane ) .
the radial and azimuthal directions are resolved by 10 grid points per @xmath185 length .
we choose our resolution in the vertical direction to be greater than that of the horizontal directions because the dust layer has finer scale structure in @xmath16 : the dust layer becomes increasingly cuspy at the midplane as @xmath1 increases .
we prescribe the same resolution in the radial and azimuthal directions ( @xmath186 ) ; experiments with different resolutions in @xmath10 and @xmath172 generated spurious results .
too small a box size can artificially affect the stability of the dust layer , because a given box can only support modes having integer numbers of wavelengths inside it .
small boxes may be missing modes that in reality overturn the layer .
we verify that for all runs in which the dust layer overturns , the kh mode that most visibly disrupts the layer spans more than one azimuthal wavelength .
typically 35 wavelengths are discerned across the box .
to more thoroughly test our standard choices for @xmath180 , we study how systematic variations in box length affect how the instability develops . for this test
, we adopt a fixed set of physical input parameters , @xmath187 , which should lead to instability @xcite .
our diagnostic is the time evolution of the vertical kinetic energy at the midplane : @xmath188 , where the average is over all @xmath10 and @xmath172 at fixed @xmath189 and time @xmath190 .
we vary @xmath180 and @xmath181 in tandem to maintain the same resolution from run to run , thereby isolating the effect of box size .
figure [ fig : lopt ] shows how doubling one of the box dimensions while fixing the other two alters the time history of @xmath191 .
panel ( a ) demonstrates that our standard choice for @xmath192 is sufficiently large because the curves for @xmath192 and @xmath193 practically overlap .
panels ( b ) and ( c ) show that our standard choices for @xmath194 and @xmath195 are somewhat less adequate .
the peak of the curve for @xmath196 is delayed by two orbits compared to that for @xmath197 , and the curve for @xmath198 peaks an orbit earlier than that for @xmath199 .
nevertheless these time differences are small compared to the total time to instability , about 10 orbits .
moreover , the errors point in opposite directions .
thus we expect our choices for @xmath200 and @xmath201 to partially compensate for each other so that any error due to our box size in calculating the time to instability will be less than @xmath61 orbit .
we test how robust our results are to numerical resolution by re - running a few simulations at twice the normal resolution ( doubling @xmath181 while fixing @xmath180 ) .
results at high resolution are given in
[ ssec : tests ] . every simulation is run for at least ten orbits .
a typical run performed at our standard resolution takes approximately 2.5 wall - clock hours using 56 processors on the purdue steele cluster .
a high - resolution run takes about 32 wall - clock hours .
in our standard simulations , we fix @xmath86 and @xmath165 while systematically varying @xmath0 and @xmath1 from run to run .
our systematic variations of @xmath0 and @xmath1 correspond to systematic variations in @xmath49 ; recall that only two of the three parameters @xmath0 , @xmath1 , and @xmath49 are independent . for each @xmath202
we adjust @xmath0 until the threshold value @xmath2 dividing stable from unstable runs is determined to within 0.1 dex . deciding by numerical simulation whether a given dust layer is stable or not is unavoidably subject to the finite duration of the simulation .
we define our criteria for deciding stability in [ ssec : criteria ] .
results are given in
[ ssec : resultsnow ] and tested for robustness in [ ssec : tests ] .
stability is assessed by two quantities : the midplane vertical kinetic energy @xmath203 where the average is performed over @xmath10 and @xmath172 at fixed @xmath204 and @xmath190 , and the dust density profile @xmath205 where the average is performed over @xmath10 and @xmath172 at fixed @xmath16 and @xmath190 . by definition , in an `` unstable '' run , @xmath206 grows exponentially over several orbital periods , and @xmath207 deviates from its initial value @xmath208 by more than 15% .
`` stable '' simulations satisfy neither criterion .
some runs are `` marginally unstable '' in that they satisfy the first but not the second criterion . at the end of the standard ten - orbit duration of a marginally unstable run , we find the kinetic energy continues to rise , suggesting that were the run to be extended for longer than ten orbits , the dust layer would eventually overturn . in every instance where we extend the duration of a marginally unstable run
, we verify that this is the case .
thus `` marginally unstable '' is practically synonymous with `` unstable .
'' examples of unstable and stable runs are shown in figure [ fig : profile ] . in the unstable simulation , after @xmath209 orbits , the kinetic energy rises exponentially . at @xmath210 orbits , the dust layer overturns and
the midplane dust - to - gas ratio falls by more than 60% .
by contrast , in the stable simulation , after an initial adjustment period lasting @xmath63 orbits during which the midplane value of @xmath211 decreases by 10% , the kinetic energy drops by orders of magnitude to a nearly constant value and shows no evidence of further growth .
figure [ fig : vmuevolve ] shows the evolution of @xmath212 and @xmath213 for the same unstable run of figure [ fig : profile ] .
the velocity data are sampled at a single @xmath214 position at the center of our simulation box .
the radial and vertical velocities @xmath215 and @xmath216 , initially zero , grow to become comparable with the shearing velocity @xmath217 . figure [ fig : mucolor ] displays corresponding snapshots of @xmath218 , taken at a single radius @xmath148 near the center of our box .
though the data in figures [ fig : vmuevolve ] and [ fig : mucolor ] are sampled at particular radial locations in our box , we verify that the instability develops similarly at all locations as it should unlike the zeus - based simulations of @xcite .
figure [ fig : m0001 ] maps the stable and unstable regions in @xmath219 space , for fixed @xmath220 and @xmath221 .
figures [ fig : m0001ris ] and [ fig : m0001sm ] portray the same data using alternate but equivalent projections of parameter space : @xmath222 and @xmath223 , respectively .
these plots demonstrate that there is no unique value of @xmath224 .
rather @xmath2 is a function of @xmath1 , or equivalently a function of @xmath49 . for bulk metallicities @xmath49 near the solar value , @xmath2 is found to be close to the classical value of @xmath225 .
but as @xmath49 decreases below the solar value , @xmath2 shrinks to @xmath60.01 or even lower .
a least - squares fit to the four midpoints ( evaluated in log space ) in figure [ fig : m0001 ] dividing neighboring stable points ( in black ) and unstable points ( in red or red outlined with black ) yields @xmath226 .
this same fit , projected into metallicity space , is shown in figures [ fig : m0001ris ] and [ fig : m0001sm ] ; in metallicity space the stability boundary is not a power law . as figure [ fig : m0001sm ] attests , dust - to - gas ratios @xmath1 as high as @xmath68 can be attained in disks of solar metallicity without triggering a shear instability : see the intersection between the dashed curve fitted to our standard resolution data , and the dotted line representing solar metallicity .
this intersection occurs at @xmath227 .
were we to re - fit the dashed curve using the higher resolution data represented by triangles , the intersection with solar metallicity would occur at @xmath1 closer to 8 . a dust - to - gas ratio of @xmath228 is within a factor of @xmath64 of the toomre threshold for gravitational fragmentation in a minimum - mass disk ( cy10 ; [ sec : discussion ] ) .
we can achieve the toomre threshold by simply allowing for a gas disk that is @xmath6@xmath7 more massive than the minimum - mass nebula .
alternatively we can enrich the disk in metals to increase @xmath49 . extrapolating the boundary of stability ( dashed curve ) in figure [ fig : m0001sm ] to higher
@xmath49 suggests that the toomre threshold @xmath229 could be achieved for minimum - mass disks having @xmath6@xmath8 the solar metallicity . the sensitivity to metallicity
is also exemplified by figure [ fig : profile ] .
for the same @xmath230 , the dust layer based on a near - solar metallicity of @xmath231 overturns , whereas one derived from a supersolar metallicity of @xmath232 remains stable .
we test how robust our determination of @xmath2 is to numerical resolution by redoing our simulations for @xmath234 and @xmath235 with double the number of grid points in each dimension .
the results are overlaid as blue triangles in figures [ fig : m0001 ] , [ fig : m0001ris ] , and [ fig : m0001sm ] . at @xmath236 ,
increasing the resolution does not change @xmath2 from its value of 0.009 . at @xmath230 , @xmath2 shifts downward from 0.3 to 0.2 .
although we have not strictly demonstrated convergence of our results with resolution , and although high resolution data at other values of @xmath1 are missing , it seems safe to conclude that the slope of the stability boundary in @xmath0-@xmath1 space is close to , but decidedly shallower than , linear .
we also test the sensitivity of our results to @xmath165 .
increasing @xmath165 by an order of magnitude to @xmath163 shifts @xmath2 upward by @xmath237 0.2 dex at @xmath238 , but leaves @xmath2 unchanged at larger @xmath1 ( figure [ fig : m001 ] ) .
b09 also reported some sensitivity to @xmath165 .
tests where @xmath86 was doubled to @xmath239 reveal no change in @xmath2 ( data not shown ) .
where a protoplanetary disk is devoid of turbulence intrinsic to gas , dust particles settle toward the midplane , accumulating in a sublayer so thin and so dense that the dust - gas mixture becomes unstable .
if the first instability to manifest is self - gravitational , dust particles are drawn further together , possibly spawning planetesimals .
if instead the layer is first rendered unstable by a kelvin - helmholtz - type shearing instability ( khi ) , the resultant turbulence prevents dust from settling further , pre - empting gravitational collapse . in this paper
we investigated the conditions which trigger the khi , hoping to find a region of parameter space where the khi might be held at bay so that planetesimals can form by self - gravity .
a fundamental assumption underlying our work is that turbulence intrinsic to gas can , in some regions of the disk , be neglected .
there is some consensus that near disk midplanes , in a zone extending from @xmath61 to at least @xmath610 aus from the parent star , gas may be too poorly ionized to sustain magnetohydrodynamic turbulence @xcite . presumably if the magnetorotational instability ( e.g. , * ? ? ? * ) can not operate at the midplane , disk gas there is laminar pending the uncertain ability of magnetically active surface layers to stir the disk interior ( e.g. , * ? ? ?
* ) , or the discovery of a purely hydrodynamic form of turbulence @xcite . to get a sense of how laminar disk gas must be to permit dust sublayers to form , chiang & youdin ( 2010 ) compared the height to which dust particles are stirred in an `` alpha''-turbulent disk to the thickness of the sublayer ( [ eqn : deltaz ] ) .
they estimated that the former is smaller than the latter when the dimensionless turbulent diffusivity @xmath240 for @xmath241 .
to place this requirement in context , @xmath242 values for magnetically active zones are typically quoted to be greater than @xmath6@xmath163 .
whether magnetically dead zones are sufficiently passive for dust to settle into sublayers remains an outstanding question .
that models gas and dust as two perfectly coupled fluids @xcite .
we found that stability is not characterized by a single critical richardson number .
rather the value of @xmath2 distinguishing layers that overturn from those that do not is a nearly linear function of the midplane dust - to - gas ratio @xmath1 ( figure [ fig : m0001 ] ) .
dust - rich sublayers having @xmath243 have @xmath5near the canonical value of 1/4while dust - poor sublayers having @xmath244 ( still orders of magnitude dustier than well - mixed gas and dust at solar abundances ) have @xmath2 as low as 0.009 .
previous studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) assumed a universal critical richardson number of 1/4 .
this popular assumption seems correct only for dust - rich layers having @xmath1 so large they are on the verge of gravitational instability . for less dusty midplanes , the assumption appears to be incorrect .
our numerical results are roughly consistent with those of @xcite , who also found evidence that @xmath2 decreases with decreasing @xmath1 .
comparing his table 2 with our figure [ fig : m0001 ] shows that his constraints on @xmath2 are , for the most part , compatible with those presented here , for the range @xmath24510 where our respective data overlap .
our findings supersede those of @xcite insofar as we have explored parameter space more finely and systematically , at greater and more uniform resolution , with numerical methods better suited for subsonic flows .
our results turn out to be consistent with the classical richardson criterion which states only that @xmath246 is necessary , not sufficient , for instability even though the criterion as derived by @xcite applies only to two - dimensional flows , which our dust layers are not .
our simulations demonstrate that the criterion can still serve as a useful guide for assessing stability in disks having bulk metallicities ranging from subsolar to slightly super - solar values with the proviso that the actual richardson number dividing kh - stable from kh - unstable flows , while @xmath247 , is generally not equal to 1/4 . why is nt the richardson criterion for instability sufficient in rotating dust disks ? the criterion considers the competition between the destabilizing vertical shear and the stabilizing influence of buoyancy , which causes fluid parcels to oscillate about their equilibrium positions at the brunt - visl frequency .
however , there exists another stabilizing influence , ignored by the richardson number , provided by the radial kepler shear @xcite . in the limit @xmath248 ,
the brunt frequency ( [ eqn : brunt ] ) becomes negligible relative to the kepler shearing frequency ( [ eqn : keplershear ] ) , suggesting stability now depends on the competition between the destabilizing vertical shear and stabilizing radial kepler shear .
we expect the flow to be stable as long as the kepler shear can wind up unstable eigenmodes to higher radial wavenumbers before their amplitudes grow large enough to trigger nonlinear effects .
this suggests that we replace the richardson number with a `` shearing number , '' defined by analogy as the square of the ratio of the kepler shearing frequency to the vertical shearing frequency : @xmath249 where we have used ( [ eqn : verticalshear ] ) and ( [ eqn : deltaz ] ) . by assuming @xmath250 is constant for marginally stable dust profiles , we arrive at the relation @xmath251 what is surprising is that this trend , although expected to hold only for @xmath3 , appears to hold approximately for all @xmath1 , according to our simulation results in figure [ fig : m0001 ] . for @xmath252
, we would have expected from ( [ eqn : sh ] ) that @xmath2 asymptote to a constant ; but it does not .
our higher resolution runs do suggest the stability curve slightly flattens at @xmath253 , but such deviations seem too small to be fully explained using arguments relying purely on the shearing number . to explain the observed trend , we might co - opt the methods of @xcite , who linearized and numerically integrated the 3d equations of motion for the dust layer . for their particular choice of background vertical density profile , they solved for the maximum growth factors for the most unstable kh modes ( see also @xcite who considered the axisymmetric problem ) .
we would need to replace their assumed profile with our profiles having spatially constant @xmath0 .
perhaps our numerically determined stability curve @xmath254 corresponds to a locus of fixed maximum growth factor .
gravitational instability occurs on a dynamical time when the dust layer s toomre @xmath255 $ ] reaches unity @xcite . for @xmath256 given by the minimum - mass solar nebula
, this occurs when @xmath229 , fairly independently of @xmath10 @xcite .
of course in more massive gas disks ( greater @xmath14 ) , the requirement on @xmath1 is proportionately lower .
figure [ fig : m0001sm ] shows that for disks having bulk metallicities @xmath49 equal to the solar value of 0.015 , the dusty sublayer can achieve @xmath257 before it becomes kh unstable .
taken at face value , such a marginally kh - stable subdisk , embedded in a gas disk having @xmath258 times the mass of the minimum - mass solar nebula , would undergo gravitational instability on the fastest timescale imaginable , the dynamical time . the case that planets form from disks several times more massive than the minimum - mass solar nebula is plausible ( e.g. , * ? ? ?
* ; * ? ? ?
. an alternate way of crossing the toomre threshold is to allow the bulk metallicity @xmath49 to increase above the solar value of 0.015 . extrapolating the boundary of stability in figure [ fig : m0001sm ]
to @xmath229 suggests that metallicities roughly @xmath63 times solar would be required for dynamical gravitational instability in a minimum - mass disk .
there are several proposed ways to achieve supersolar metallicities in some portions of the disk , among them radial pileups @xcite or dissipative gravitational instability ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; see also the introduction of @xcite ) .
none of the ways we have outlined for achieving gravitational instability rely on the streaming instability or turbulent concentration of particles , mechanisms that we have criticized in [ ssec : versus ] .
nevertheless our scenarios may be too optimistic because all our dust profiles are predicated on the assumption of a spatially constant @xmath0 .
this assumption tends to generate strong density cusps at the midplane that might not be present in reality . in a forthcoming paper
we will relax the assumption of spatially constant @xmath0 and measure the maximum @xmath1 attainable , as a function of metalllicity @xmath49 , by simulating explicitly the settling of dust towards the midplane .
we thank daniel lecoanet , eve ostriker , prateek sharma , jim stone , and yanqin wu for discussions .
an anonymous referee provided a thoughtful and encouraging report that helped to place our work in a broader context .
this research was supported by the national science foundation , in part through teragrid resources provided by purdue university under grant number tg - ast090079 . | gravitational instability ( gi ) of a dust - rich layer at the midplane of a gaseous circumstellar disk is one proposed mechanism to form planetesimals , the building blocks of rocky planets and gas giant cores .
self - gravity competes against the kelvin - helmholtz instability ( khi ) : gradients in dust content drive a vertical shear which risks overturning the dusty subdisk and forestalling gi . to understand the conditions under which the disk can resist the khi , we perform three - dimensional simulations of stratified subdisks in the limit that dust particles are small and aerodynamically well coupled to gas , thereby screening out the streaming instability and isolating the khi .
each subdisk is assumed to have a vertical density profile given by a spatially constant richardson number @xmath0 .
we vary @xmath0 and the midplane dust - to - gas ratio @xmath1 and find that the critical richardson number dividing kh - unstable from kh - stable flows is not unique ; rather @xmath2 grows nearly linearly with @xmath1 for @xmath1 = 0.310 .
plausibly a linear dependence arises for @xmath3 because in this regime the radial kepler shear replaces vertical buoyancy as the dominant stabilizing influence .
why this dependence should persist at @xmath4 is a new puzzle .
the bulk ( height - integrated ) metallicity is uniquely determined by @xmath0 and @xmath1 . only for disks of bulk solar metallicity
is @xmath5 , close to the classical value .
our empirical stability boundary is such that a dusty sublayer can gravitationally fragment and presumably spawn planetesimals if embedded within a solar metallicity gas disk @xmath6@xmath7 more massive than the minimum - mass solar nebula ; or a minimum - mass disk having @xmath6@xmath8 solar metallicity ; or some intermediate combination of these two possibilities .
gravitational instability seems possible without resorting to the streaming instability or to turbulent concentration of particles . |
it has been known since the earliest observations of rich clusters ( see e.g. , * ? ? ?
* ) that galaxies which inhabit these high - density regions are quite distinct from the general field galaxy population .
the cluster population is dominated by galaxies with early - type morphologies , primarily ellipticals and s0s , which have colours and spectral types consistent with their undergoing passive evolution .
in contrast , the field population is dominated by actively star - forming galaxies generally of late morphological types , such as spirals and irregulars . explaining the segregation of galaxies into these two broad classes is one of the outstanding problems of extragalactic astronomy and cosmology .
early work suggested that elliptical galaxies may have an intrinsically different formation process to disc galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . however , since the general acceptance of the hierarchical process as the preferred model of structure formation ( see e.g. , * ?
* ; * ? ? ?
* ; * ? ? ?
* ) , much attention has been focused on mechanisms that could transform star - forming , late - type galaxies into passively evolving , early - type systems in dense environments .
some examples of suggested evolutionary processes include : galaxy mergers the results of numerical simulations suggest that the collision and merging of two equal mass disc galaxies can produce a product with properties typical of elliptical galaxies ( e.g. , * ? ? ?
* ) ; ram - pressure stripping of gas from galaxies falling into clusters @xcite ; and galaxy harassment , where tidal forces strip disc galaxies to make dwarf spheroidal systems @xcite .
modern large surveys such as the 2df galaxy redshift survey ( 2dfgrs , * ? ? ?
* ) and the sloan digital sky survey ( sdss , * ? ? ?
* ) , provide excellent quality data and have allowed the dependence of galaxy properties upon environment in the local universe to be examined in some depth in recent years . @xcite studied the environmental dependence of the galaxy star formation rate near clusters in the 2dfgrs and found that it converged to the field value beyond @xmath9 times the virial radius , indicating that relatively small increments in local density at the outskirts of clusters lead to a decrease in star formation rates .
they found no dependence of the average star formation rate upon cluster velocity dispersion , a tracer of mass and hence a measure of the global environment .
@xcite used the early data release of the sdss to study the galaxy star formation rate as functions of local galaxy density and clustercentric radius .
they found that the star formation rate is strongly correlated with the local galaxy density .
similarly to @xcite , @xcite also found that the star formation rate converges to the field value at clustercentric distances of 3 - 4 virial radii .
@xcite extended these studies of ` galaxy ecology ' to groups , using both 2dfgrs and sdss data .
they found little evidence that the distribution of the h@xmath10 line strength , an indicator of star formation rate , depends strongly upon environment amongst the actively star - forming galaxy population .
similarly to both @xcite and @xcite , they also found that the fraction of galaxies with significant ongoing star formation decreased steadily with increasing density , and interpreted the persistence of this correlation at low densities as indicating that ram pressure stripping is not the only mechanism responsible for the truncation of star formation in galaxies .
in addition , @xcite found little dependence of the star formation rate upon group velocity dispersion .
the luminosity function ( lf ) of galaxies is a probability distribution over absolute magnitude , and by measuring the shape of the distribution in a number of environments , one is able to obtain clues as to which processes are important in shaping galaxy evolution .
* hereafter dp03 ) , constructed @xmath6 composite luminosity functions ( lfs ) of rich 2dfgrs clusters and found that the lfs of early spectral type galaxies have brighter characteristic magnitudes @xmath11 and steeper faint end slopes @xmath10 than the 2dfgrs field lfs of @xcite .
in contrast , the lfs of star - forming galaxies were found to be essentially identical .
dp03 also constructed lfs for a variety of cluster subsamples , divided by several environmental variables including richness , velocity dispersion , substructure , and b
m @xcite type , but found no evidence of differences in the derived lf parameters .
however , one measure of the large scale cluster environment that dp03 were unable to test was cluster x - ray emission , which is produced from the thermal bremsstrahlung radiation of the hot intracluster gas ( see e.g. , * ? ? ?
* ) , and is perhaps a better tracer of cluster mass than the velocity dispersion .
there have of course been many other studies of the lf in different environments in recent years ; we refer the reader to the references of dp03 for details of these . in this paper
we seek to build upon the previous work described above and investigate the dependence of cluster galaxy populations upon the x - ray luminosity of the host cluster , by utilising the rosat - eso flux limited x - ray galaxy cluster survey ( reflex ) catalogue @xcite in conjunction with galaxy data taken from the 2dfgrs .
this subject is also being explored by the rass - sdss galaxy cluster survey @xcite , who have examined cluster scaling relations and galaxy luminosity functions for a sample of clusters selected from rosat x - ray observations , with galaxy photometry and spectroscopy drawn from the sdss .
the outline of this paper is as follows . in the next section ,
we briefly describe the two datasets used in this survey , 2dfgrs and reflex . in section [ s_lxsbi ] , we describe our initial method of correlating reflex and 2dfgrs data , and measure the relation between x - ray luminosity and cluster velocity dispersion . in section [ s_reflexnonreflex ] , we supplement the reflex cluster sample with rich clusters taken from the catalogue compiled by ( * ? ? ?
* hereafter dp02 ) , and divide the combined cluster sample in x - ray luminosity .
we describe the construction of composite cluster luminosity functions in section [ s_lfs ] for low- and high-@xmath4 cluster subsamples , to complement the work of dp03 .
finally , we end with a discussion of the results in section [ s_discuss ] and our conclusions in section [ s_conclude ] . throughout
we assume a concordance cosmology of @xmath12 , @xmath13 , and @xmath14 km s@xmath2 mpc@xmath2 .
since the details of the 2dfgrs and reflex surveys have been covered in considerable depth elsewhere ( see in particular @xcite for the 2dfgrs , @xcite for reflex ) , here we only briefly review the gross characteristics of the datasets used in this paper . the 2dfgrs is the largest completed spectroscopic survey of galaxies to date , covering approximately 1500 deg@xmath15 of the apm galaxy survey region @xcite , and contains spectroscopic information on approximately 250,000 galaxies down to a magnitude limit of @xmath16 .
the survey area consists primarily of two strips : the sgp strip centred close to the south galactic pole , covering an area of @xmath17 , and the ngp strip in the northern galactic hemisphere that covers an area of @xmath18 .
in addition there are 99 widely scattered random @xmath19 fields .
the median redshift of the galaxies in the survey is @xmath20 , and up to this redshift the survey is @xmath21 per cent complete at the stated @xmath6 magnitude limit . in this paper we restrict ourselves to spectra with a 2dfgrs quality flag @xmath22 .
these galaxies have measured redshifts with an rms uncertainty of 85 km s@xmath2 . in our examination of galaxy populations ,
we make use of the 2dfgrs @xmath23 parameter defined by @xcite .
the value of @xmath23 is correlated with the present- to past - averaged star formation rate , and so passively - evolving galaxies have low values of @xmath23 , and the opposite is true for galaxies that have formed a significant fraction of their stars more recently . in this paper
we take a simple approach to classifying galaxies using this parameter , adopting @xmath24 for early - type , passively evolving galaxies ( spectral type 1 of * ? ? ?
* ) , and @xmath25 for late - type , actively star - forming galaxies ( encompassing spectral types 2 - 4 of * ? ? ?
* ) . throughout this paper
we adopt the following @xmath26-corrections for each population : @xmath27 @xmath28 where @xmath29 is the correction found for spectral type 1 galaxies by @xcite and @xmath30 is that found for spectral type 3 , an appropriate average to use for the late - types . the rms error on the 2dfgrs @xmath6-band photometry
is 0.15 mag .
reflex is the largest homogeneous catalogue of x - ray selected galaxy clusters assembled to date .
reflex cluster candidates were selected from the region @xmath31 of the second processing of the rosat all sky survey ( rass ii , * ? ? ?
* ) , avoiding the crowded stellar fields of the magellanic clouds and the region @xmath32 either side of the galactic plane .
reflex therefore covers the entire 2dfgrs region down to a nominal flux limit of @xmath33 erg s@xmath2 @xmath34 in the ( 0.12.4 kev ) energy band .
the final catalogue @xcite contains 447 clusters with measured redshifts , and is @xmath35 per cent complete . the mean fractional error on the measured reflex fluxes in the ( 0.12.4 kev ) band is 16.7 per cent .
throughout this paper we quote reflex luminosities corrected for flux lost outside the detection aperture as described by @xcite .
these can be found in table 6 of @xcite . [
cols="^,^,^,^,^",options="header " , ] since both @xmath4 and @xmath36 should trace the cluster mass , and @xmath23 is correlated with the equivalent width of the h@xmath10 line , we can also make a comparison with the results of @xcite , who examined the correlation between the fraction of galaxies with h@xmath10 equivalent width @xmath37 and local density in the 2dfgrs and sdss .
@xcite divided a sample of @xmath38 group and cluster galaxies into two subsamples of low-@xmath36 ( @xmath39 km s@xmath2 ) and high-@xmath36 ( @xmath40 km s@xmath2 ) .
they found marginal evidence that the fraction of star - forming galaxies was lower in the high-@xmath36 clusters relative to the low-@xmath36 groups at fixed local density .
this is similar to the result we see for the low- and high-@xmath4 cluster samples studied in this paper , although we make no attempt to distinguish between the effects of variation in local galaxy density and the large scale cluster environment ( as traced by @xmath4 ) .
@xcite also studied star - formation rates in 2dfgrs cluster galaxies as traced by the equivalent width of h@xmath10 .
@xcite considered 17 clusters drawn from dp02 and divided their sample into clusters with @xmath41 km s@xmath2 and @xmath42 km s@xmath2 , with 10 and 7 clusters in each respective subsample .
they found no dependence of galaxy star - formation rates upon the velocity dispersion of the host cluster , in contrast to the results that we obtain for our low- and high-@xmath4 cluster samples .
this is perhaps due to the comparatively small size of the cluster sample used by @xcite .
we present a comparison of our derived lf parameters with the values obtained from 2dfgrs data by dp03 for clusters and @xcite for the field in table [ t_lfcomp ] , divided by spectral type and transformed to @xmath14 km s@xmath2 mpc@xmath2 . here , spectral type 1 corresponds to our definition of early - type galaxies and spectral types 2 - 4 collectively form our sample of late - type galaxies . in this table
we also quote average values of @xmath11 and @xmath10 for the late - type lfs of @xcite , weighted by the respective normalisation constants , to ease comparison with our late - type lf parameter values .
the most surprising result of this work is the disagreement of the high-@xmath4 sample early - type lf parameter values with both the low-@xmath4 sample and the results of dp03 . @xmath11 and @xmath10 are both lower in the high-@xmath4 , early - type lf in comparison to the corresponding dp03 early - type lf at @xmath43 significance .
the characteristic magnitude of the high-@xmath4 sample lf is also significantly lower than found in the field by @xcite , although the faint - end slope is steeper . in contrast ,
both schechter function parameter values for the early - type lf of the low-@xmath4 sample are within @xmath44 of agreement with the results of dp03 , with @xmath11 brighter than in the field , and steeper @xmath10 .
the late - type lf parameter values for the high-@xmath4 sample agree within @xmath45 of the type 2 and types 3 + 4 lfs of dp03 , and the weighted average of the field results of @xcite .
there is therefore marginal disagreement
although note that there is no significant disagreement between the low- and high-@xmath4 late - type lfs .
the low-@xmath4 sample late - type lf parameters are well within @xmath46 of agreement with the dp03 results and the @xcite field results .
these results are therefore consistent with a picture in which the late - type lf varies little with environment .
we are also able to compare our results to similar studies undertaken using sdss @xmath47-band data , following a suitable transformation to the @xmath48 band as used by the 2dfgrs , i.e. , @xmath49 as used by @xcite .
we present results from @xcite for rass - sdss galaxy clusters , and @xcite for the field in table [ t_sdsslfcomp ] .
note that we assume a @xmath50 colour of 0.8 when converting the @xmath11 values using equation ( [ e_g2bj ] ) , a value we estimated from inspection of the colour distribution of early spectral type 2dfgrs galaxies , after transforming the 2dfgrs @xmath48 and @xmath51 bands to sloan @xmath47 and @xmath52 using the prescriptions given in @xcite .
@xcite computed lfs for a sample of 69 x - ray selected clusters drawn from the rass - sdss survey .
we quote the lf parameter values for the composite lf produced from galaxies within 1.5 mpc clustercentric distance in table [ t_sdsslfcomp ] .
we can see from comparison of the results presented in section [ s_results ] that there is a significant disagreement between @xcite and the results that we obtain for both lf parameters for the high-@xmath4 sample lf of all spectral types in fact , we obtain a value of @xmath11 that is @xmath53 magnitude fainter .
in contrast , there is mild disagreement ( at the @xmath54 level ) between the characteristic magnitude of this lf and the value that we obtain for the low-@xmath4 sample lf for galaxies of all spectral types .
however , our low-@xmath4 sample lf has a significantly flatter faint - end slope than this lf .
@xcite constructed lfs for the same 69 rass - sdss clusters using galaxies selected within @xmath5 , rather than using a fixed metric aperture , and fitted a composite of two schechter functions to each lf . using this method
, they obtained results consistent with those previously obtained in @xcite .
it is appropriate to compare the lf values that we obtained in section [ s_results ] with the bright component of the @xcite double - schechter function fits .
we find in this case a much better agreement of our low-@xmath4 sample lf for galaxies of all spectral types , with the values of @xmath11 and @xmath10 in agreement within @xmath55 of each other .
the agreement with the lf parameters we derived for the equivalent high-@xmath4 sample lf is also much closer we find @xmath56 and @xmath57 in this case .
@xcite also constructed lfs for galaxies divided into early and late types on the basis of their @xmath58 colour .
because of uncertainties in the @xmath26-correction , they used a reduced sample of sixteen rass - sdss clusters located within @xmath59 .
as table [ t_sdsslfcomp ] shows , @xcite found a value of @xmath11 for the early - type galaxy lf intermediate between the values we obtained for our low- and high-@xmath4 cluster samples .
in fact , both @xmath11 and @xmath10 for this lf are in good agreement with the values we obtain for our corresponding high-@xmath4 sample lf .
the agreement with the low-@xmath4 sample lf is worse , but the @xmath11 values are still in agreement within @xmath46 , and the @xmath10 values differ at the @xmath60 level .
however , for the late - type lf @xcite find a much brighter characteristic magnitude and steeper faint - end slope than we find for both our low- and high-@xmath4 cluster samples .
we note that unlike the late - type lfs produced from 2dfgrs data by @xcite for the field , and both dp03 and the present work for clusters , @xcite obtain a brighter characteristic magnitude for the late - type lf in comparison to the early - type lf .
the reasons for this discrepancy are not clear .
it is prudent to look for selection effects that could explain the discrepancy between the early - type luminosity functions in our low- and high-@xmath4 cluster samples . in the case of the low-@xmath4 sample , which are almost all
exclusively optically selected , we may worry that they are not in fact bound systems , in which case the effect of interloper galaxies on the derived lf parameters could be significant . to investigate the effect of changing the selection of cluster members in redshift space on the derived lf parameters , we constructed lfs using galaxies selected within @xmath61 , @xmath62 , @xmath63 , @xmath64 , and @xmath65 km s@xmath2 slices ( @xmath66 ) around each cluster redshift , within the previously calculated selection radius of @xmath5 ( i.e. instead of using @xmath67 the velocity dispersion to select cluster members ) .
[ f_sbislices ] shows the resulting schechter function parameters , with 1@xmath68 statistical errors derived from bootstrap resampling , obtained using the same method as the results presented in section [ s_lfs ] .
we can see from this figure that the values of @xmath11 and @xmath10 are not significantly affected by this selection effect , remaining consistent with the results presented in section [ s_results ] .
we also considered the possibility that a selection effect could arise due to galaxy crowding in the centres of the high-@xmath4 clusters , because these objects are on average more distant than the low-@xmath4 clusters ( panel ( b ) of fig .
[ f_sbidist ] ) .
there are two possible sources of unaccounted - for incompleteness in this case : incompleteness in the parent apm galaxy catalogue , due to object blending ; and/or incompleteness in the 2dfgrs spectroscopic catalogue , due to fibre crowding . to determine the importance of this effect , we produced luminosity functions for the high-@xmath4 sample after excising galaxies within the central @xmath69 of each cluster .
we present the lf for early spectral types in fig .
[ f_coreless ] .
the derived lf parameters ( @xmath70 , @xmath71 ) are clearly consistent with those obtained in section [ s_results ] , and so we conclude that incompleteness effects of this type have no significant impact upon our results . the physical aperture sampled by a 2df fibre naturally becomes smaller as redshift increases , and so the spectral type determined for a particular galaxy may have a mild redshift dependence .
however , due to seeing effects and fibre positioning errors , the effect is likely to be small ( see * ? ? ?
. nevertheless , we attempted to gauge the impact of the aperture bias by constructing lfs for both the low- and high-@xmath4 samples using only clusters located within a common redshift range of @xmath72 .
of the high-@xmath4 cluster sample , only 8 clusters of the 19 are found in this redshift range , and we found the best fitting early - type lf parameters were @xmath73 , @xmath74 .
the low-@xmath4 sample in this redshift range is made up of 39 clusters , for which the best fitting schechter function was specified by @xmath75 , @xmath76 .
we therefore conclude from a comparison of these results with those for the full sample that the effect of fibre aperture bias is negligible .
the faint - end of the high-@xmath4 sample is defined by relatively few clusters .
in fact , at magnitudes fainter than @xmath77 , only 11 clusters contribute to the lf .
although the early - type high-@xmath4 sample lf is made up from 17 per cent more galaxies than the equivalent low-@xmath4 lf , we may simply be unfortunate in that the faint - end behaviour of these 11 clusters is not typical of the rest of the sample . to address this concern
, we performed a one - parameter fit to the early - type , high-@xmath4 sample lf , with the value of the faint - end slope fixed to that found for the low-@xmath4 sample , i.e. @xmath78 . the resulting lf is shown in fig .
[ f_fixedalpha ] .
although the value of @xmath11 for the best fit is now brighter than found from the two - parameter fit , we find that @xmath11 remains lower than the value obtained from the low-@xmath4 sample lf .
this result is marginally significant , at the @xmath79 level ( @xmath80 ) .
nevertheless , this exercise confirms that the low - value of @xmath11 found for the high-@xmath4 cluster sample arises primarily from the deficit of early - type galaxies with @xmath8 ( panel ( c ) of fig .
[ f_earlylfs ] ) , rather than from the behaviour of the faint - end of the lf .
since simple selection effects are unable to explain our results , we now consider how the difference in the early - type lf parameters could be driven by environmental differences between the low- and high-@xmath4 clusters .
other studies of the luminosity function using 2dfgrs data have found evidence to suggest that @xmath11 brightens and @xmath10 steepens with increasing galaxy density @xcite or mass @xcite .
the results presented here seem to suggest that the opposite is true when considering the effect of increasing x - ray luminosity on the lf .
it could therefore be suggested that the high-@xmath4 cluster sample is actually composed of low - mass , merging systems
@xmath4 is known to be a good tracer of mass , but the results of numerical simulations suggest that @xmath4 can be boosted during mergers . for example , ( * ? ? ?
* see also @xcite ) demonstrated that an equal mass merger with impact parameter @xmath81 ( i.e. a head - on collision ) can boost the bolometric x - ray luminosity of the resultant object by a factor of 5 - 6 .
the boost in @xmath4 in this case lasts for approximately half a sound crossing time , @xmath53 gyr for the high-@xmath4 cluster sample . without knowledge of the merger histories of the individual high-@xmath4 clusters ,
the impact of this effect is difficult to gauge .
however , in the limiting case of all the high-@xmath4 clusters being boosted in @xmath4 by a factor of six , the median unboosted ( 0.12.4 kev ) x - ray luminosity of the sample is @xmath82 erg s@xmath2 . by comparison , for the majority optically - selected low-@xmath4 cluster sample
the median maximum @xmath4 set by the reflex flux limit is @xmath83 erg s@xmath2 .
therefore even in this extreme case both samples would be of similar mass rather than the high-@xmath4 clusters being significantly less massive .
in addition , as shown in section [ s_sampledefinitions ] , the median velocity dispersion of the high-@xmath4 sample is higher than that for the low-@xmath4 sample .
these facts imply that the high-@xmath4 clusters are more likely to be higher mass systems than the low-@xmath4 clusters .
we suggest that a consistent explanation of the results of sections [ s_properties ] and [ s_results ] can be achieved by a scenario in which the high-@xmath4 clusters are in general older , more dynamically evolved systems than the low-@xmath4 clusters . in this case , the higher fraction of early - type , passively evolving galaxies present in the high-@xmath4 sample comes about naturally from there being more crossing times available for infalling late - type galaxies to be converted into early - type galaxies within the cluster environment .
this holds irrespective of the mechanism responsible for driving the transformation , whether it is by interactions with the intracluster medium ( as in ram - pressure stripping ) , or by interactions between cluster member galaxies ( as in the harassment scenario ) .
the excess galaxies that we see at the very bright end of the high-@xmath4 cluster lf could be explained by these systems having a more extensive merger history than the low-@xmath4 clusters , as suggested by the results of @xcite . in particular ,
a ` galactic cannibalism ' scenario ( e.g. * ? ? ?
* ) where the brightest cluster galaxies grow by the accretion of the successively fainter members could help to bring about the deficit of galaxies seen at @xmath8 in the high-@xmath4 sample , early - type lf , which seems to be the cause of the low - value of @xmath11 found in this case .
@xcite studied the @xmath26-band lfs of 93 galaxy clusters divided into two samples of low- and high - mass ( estimated from the x - ray temperature ) .
although they found the high - mass lf had a brighter characteristic magnitude than the low - mass lf in contrast to the results of this work for low- and high-@xmath4 clusters they found a deficit of galaxies below @xmath11 in the high - mass sample in comparison to the low - mass lf .
@xcite ( see also * ? ? ?
* ) interpreted their result as an indication that a major contribution to the masses of the very brightest galaxies in clusters comes from mergers with other bright ( @xmath84 ) cluster galaxies . perhaps because our lfs are constructed purely from spectroscopically - selected cluster members rather than employing a more uncertain statistical subtraction of non - cluster members the effect that we see on the derived schechter function parameters is more dramatic .
we have correlated the reflex catalogue of x - ray selected galaxy clusters with the 2df galaxy redshift survey and performed a study of the dependence of cluster galaxy populations upon x - ray luminosity . using this large , homogeneous combined dataset , supplemented with additional rich clusters from the catalogue compiled by @xcite
, we found : \(1 ) an @xmath0 relation consistent with that derived by @xcite for the reflex survey team using a combination of literature and reflex optical follow - up data .
\(2 ) the fraction of early - type , passively evolving galaxies is 12 per cent higher in high-@xmath4 clusters compared to low-@xmath4 clusters out to @xmath5 , as determined from a volume - limited sample of 2dfgrs galaxies with @xmath85 , @xmath86 . in both cases the early - type galaxy fraction falls off smoothly with increasing distance from the cluster centre .
\(3 ) both the characteristic magnitude @xmath87 and faint - end slope @xmath10 of the galaxy luminosity function are lower in high-@xmath4 clusters compared to low-@xmath4 clusters .
this is caused primarily by the low value of @xmath11 found for the lf of early spectral type galaxies in the high-@xmath4 sample lf , which is driven by an underabundance of galaxies with @xmath8 relative to the corresponding low-@xmath4 sample lf .
we believe these results are consistent with a scenario in which the high-@xmath4 clusters are more dynamically evolved systems than the low-@xmath4 clusters . the higher fraction of passively evolving galaxies found in the high-@xmath4 sample could arise through the accretion and conversion of a greater number of star - forming field galaxies . the lower value of @xmath11 in the lf of early spectral types galaxies in high-@xmath4 clusters could be explained by ` cannibalism ' of bright ( @xmath8 ) members to produce the very brightest cluster galaxies .
we thank everyone involved in the success of both the 2df galaxy redshift survey and the reflex cluster survey , without which of course this work would not have been possible .
thanks to phil james for some useful suggestions that improved the clarity of this paper . thanks also to the referee , hans bhringher , who provided useful feedback which improved this manuscript .
m. hilton acknowledges the support of a studentship from liverpool john moores university .
h. , schuecker p. , guzzo l. , collins c. a. , voges w. , cruddace r. g. , ortiz - gil a. , chincarini g. , de grandi s. , edge a. c. , macgillivray h. t. , neumann d. m. , schindler s. , shaver p. , 2004
, a&a , 425 , 367 p. l. , nichol r. c. , miller c. j. , balogh m. l. , goto t. , zabludoff a. i. , romer a. k. , bernardi m. , sheth r. , hopkins a. m. , castander f. j. , connolly a. j. , schneider d. p. , brinkmann j. , lamb d. q. , subbarao m. , york d. g. , 2003 , apj , 584 , 210 | the rosat - eso flux limited x - ray ( reflex ) galaxy cluster survey and the 2df galaxy redshift survey ( 2dfgrs ) respectively comprise the largest , homogeneous x - ray selected cluster catalogue and completed galaxy redshift survey . in this work
we combine these two outstanding datasets in order to study the effect of the large - scale cluster environment , as traced by x - ray luminosity , on the properties of the cluster member galaxies .
we measure the @xmath0 relation from the correlated dataset and find it to be consistent with recent results found in the literature . using a sample of 19 clusters with @xmath1 erg s@xmath2 in the ( 0.12.4 kev ) band , and 49 clusters with lower x - ray luminosity , we find that the fraction of early spectral type ( @xmath3 ) , passively - evolving galaxies is significantly higher in the high-@xmath4 sample within @xmath5 .
we extend the investigation to include composite @xmath6 cluster luminosity functions , and find that the characteristic magnitude of the schechter - function fit to the early - type luminosity function is fainter for the high-@xmath4 sample compared to the low-@xmath4 sample ( @xmath7 ) .
this seems to be driven by a deficit of such galaxies with @xmath8 .
in contrast , we find no significant differences between the luminosity functions of star - forming , late - type galaxies .
we believe these results are consistent with a scenario in which the high-@xmath4 clusters are more dynamically evolved systems than the low-@xmath4 clusters .
[ firstpage ] x - rays : galaxies : clusters galaxies : evolution galaxies : luminosity function , mass function
cosmology : observations |
spinal cord injury , brain computer interface , virtual reality environment , electroencephalogram , kinesthetic motor imagery , gait , ambulation , locomotion .
spinal cord injury ( sci ) can leave the affected individuals with paraparesis or paraplegia , thus rendering them unable to ambulate .
since there are currently no restorative treatments for this population , technological approaches have been sought to substitute for the lost motor functions .
examples include robotic exoskeletons @xcite , functional electrical stimulation ( fes ) systems @xcite , and spinal cord stimulators @xcite .
however , these systems lack the able - body - like supraspinal control , and so the ambulation function of these devices is controlled manually .
in addition to being unintuitive , these systems may be costly and cumbersome to use , and therefore have not yet garnered popular appeal and adoption among potential sci users . due to these limitations ,
wheelchairs remain the primary means of mobility after sci .
unfortunately , the extended reliance on wheelchairs typically lead to a wide variety of comorbidities that constitute the bulk of chronic sci - related medical care costs @xcite .
consequently , to address the above issues associated with the treatment of paraparesis and paraplegia after sci , novel brain - controlled prostheses are currently being pursued @xcite .
recent results by our group @xcite suggest that an electroencephalogram ( eeg ) based brain - computer interface ( bci ) controlled lower extremity prosthesis may be feasible .
more specifically , these studies demonstrated the successful implementation of a bci system that controls the ambulation of an avatar ( a stand - in for a lower extremity prosthesis ) within a virtual reality environment ( vre ) . by using a data - driven machine learning approach to decode the users kinesthetic motor imageries ( kmis ) , this bci - controlled walking simulator enabled a small group of subjects ( one with paraplegia due to sci @xcite ) to achieve intuitive and purposeful bci control after a short training session .
while the single sci subject outperformed most able - bodied subjects in this study , the operability of this system has not yet been tested in a sci population .
the successful implementation of the bci - controlled walking simulator in a population of subjects with sci will establish the feasibility of future bci - controlled lower extremity prostheses and will represent an important step toward developing novel gait rehabilitation strategies for sci . extending the application of the bci - controlled walking simulator to a sci population
is faced with several problems .
first , cortical reorganization , which is common after sci @xcite , may cause the cortical representation of walking kmi to vary vastly from one sci subject to another .
second , this representation may dramatically evolve over time when sci subjects are engaged in kmi training @xcite .
finally , subjects with sci may interpret walking kmi either as motor imagery or as attempted walking , which in turn may result in multiple patterns of cortical activation across these individuals .
therefore , intuitive bci operation under these conditions requires a system that can accommodate for the variations of brain physiology across sci individuals , time , and strategies . to address these problems
, we used a data - driven machine learning method to decode walking kmis in a small population of sci individuals .
this approach enabled 5 subjects to achieve intuitive and self - paced operation of the bci - controlled walking simulator after only minimal training .
furthermore , they were able to maintain this level of control over the course of several weeks .
the goal of this study was to determine if individuals with complete motor sci can use intuitive control strategies to purposefully operate a bci - controlled walking simulator . to achieve this goal ,
5 subjects with sci underwent a short training procedure where they performed alternating epochs of idling and walking kmi while their eeg were recorded .
these training eeg data were then analyzed to build decoding models for subsequent online bci operation . to ascertain purposeful bci control ,
subjects then performed 5 sessions of an online bci goal - oriented virtual walking task @xcite .
this entire procedure was performed 5 times over the course of several weeks to determine if subjects performances improved with additional practice .
this study was approved by the university of california , irvine institutional review board .
four subjects with paraplegia and one with tetraplegia due to sci were recruited via physician referral from the long beach veterans affairs spinal cord injury center and other sci outreach programs . the subjects ( see table 1 ) gave their informed consent to participate in the study .
note that all subjects were bci nave and most of them performed the experimental procedures at a rate of once per week .
.list of participants with demographic data and level of sci .
asia = american spinal injury association impairment scale . [ cols="^,^,^,<",options="header " , ]
the results of this study show that subjects with paraplegia or tetraplegia due to sci can operate a non - invasive bci - controlled avatar within a vre to accomplish a goal - oriented ambulation task .
all subjects gained purposeful online bci control on the first day after undergoing a 10-min training session , with the exception of subject 2 , who did not attain control until day 2 .
in addition , bci control was maintained and continued to improve over the course of the study .
these findings suggest that a bci - controlled lower extremity prosthesis for either gait rehabilitation or restoration may be feasible .
the offline classification accuracies varied across subjects and experimental days , but were significantly above the chance level performance ( 50% ) . similar to able - bodied subjects engaged in the same task @xcite , a short 10-min training session was sufficient for the data - driven machine learning algorithm to generate accurate subject - specific eeg decoding models for this population .
the topographic maps of these models ( e.g. fig . 3 and fig .
4 ) showed that the spatio - spectral features underlying the differences between walking and idling kmis varied across subjects and evolved over experimental days .
the differences in the brain areas and eeg frequencies across subjects may be due to variations in cortical reorganization following sci @xcite , or due to differences in imageries employed by each subject ( e.g. the kmi of walking instructions may have been interpreted differently by each subject ) . nevertheless , all subjects showed activation of mid - frontal areas , which likely overlay the pre - motor and supplementary motor areas , as well as the pre - frontal cortex .
their activation during walking kmi is consistent with functional imaging findings , such as those in @xcite .
another common pattern across subjects was the presence of activity near bilateral , lateral central - parietal electrodes , which likely represents the arm sensorimotor areas .
a similar pattern was observed in able - bodied individuals @xcite , and is hypothesized to originate from arm swing imagery .
finally , the evolution of the feature extraction maps over the 5 experimental days may be indicative of a neuro - plasticity process associated with practice and learning @xcite .
the spatio - temporal variations of walking kmi activation patterns demonstrate the necessity of a data - driven machine learning approach for rapid acquisition of intuitive bci control .
first , our approach accommodates for the variations of these activity patterns across subjects , as well as their evolution over time .
second , it facilitates rapid acquisition of online bci control , presumably by enabling subjects to utilize intuitive mental strategies .
the user training time necessary to acquire purposeful bci control in this study is significantly shorter than those of other bci studies where users must learn a completely new cognitive skill to modulate pre - selected eeg features , such as the @xmath2-rhythm over lateral central areas @xcite .
finally , this approach carries a significant potential value in the future practical implementation of bci - prostheses , as it may drastically reduce the training time needed to attain purposeful and useful bci control from a timescale of weeks to months to one of minutes to days .
this in turn may significantly reduce the cost of training users to operate future bci - prostheses .
the results presented in table 4 show that once purposeful control was achieved , it was maintained in 96% of all online sessions . in addition , 3 out of 5 subjects achieved successful stop scores similar to those obtained using a manually controlled joystick .
even though no subjects were able to complete the course as fast as manual control , it is encouraging that the average composite scores increased significantly over the course of the study .
furthermore , the average composite scores ( table 5 ) improved over time , with the best scores approaching 100% for subjects 3 and 5 by the end of the study . therefore , not only was online control significantly different from random walk , but it was also meaningful . given this trend ,
additional training and practice may help further improve performance , possibly to the point of approaching that of the manually controlled joystick . in conclusion
, the high level of online control achieved by sci subjects over the course of 5 experimental days suggests that it may be feasible to apply this bci system to control a lower extremity prosthesis for ambulation after sci .
furthermore , the proposed bci - vre system may serve as a training platform for operation of bci lower extremity prostheses once they become widely available .
this study shows that sci subjects can purposefully operate a self - paced bci - vre system in real time , and that the current bci design approach is able to overcome the potential problems associated with variations in neurophysiology due to cortical reorganization after sci , learning and plasticity processes , and differences in kmi strategies .
furthermore , the system satisfies the requirements of an ideal bci - lower extremity prosthesis set forth in @xcite , namely : intuitiveness , robustness , and short training time .
the operation of the system is intuitive as it enabled subjects to use walking kmi to control the ambulation of the avatar .
the system is robust in that the data - driven decoding methodology was able to successfully accommodate for subject - to - subject and day - to - day variations in the neurophysiological underpinnings of idling and walking kmi behaviors .
in addition , subjects were able to maintain purposeful online control over the course of several weeks , further underscoring the system s robustness over time .
finally , the system required only a short training time , as bci control was generally attained after only a 10-min training data collection procedure followed by a 2-min calibration session on the 1@xmath6 experimental day ( for 4 out of the 5 subjects ) .
the successful outcome of this study indicates that bci - controlled lower extremity prostheses for gait rehabilitation or restoration after sci may be feasible in the future .
sci , spinal cord injury ; bci , brain - computer interface ; vre , virtual reality environment ; kmi , kinesthetic motor imagery ; eeg , electroencephalogram ; fes , functional electrical stimulation ; emg , electromyogram ; fft , fast fourier transform ; cpca , classwise principal component analysis ; lda , fisher s linear discriminant analysis ; aida , approximate information discriminant analysis ; 1d , one - dimensional ; cv , cross - validation ; npc , non - player character ; mc , monte carlo ; pdf , probability density function ; asia , american spinal injury association .
cek received salary from hrl laboratories , llc ( malibu , ca ) .
the remaining authors declare that they have no competing interests .
cek carried out the experiments , collected and analyzed the data , and wrote the article .
ptw programmed the brain - computer interface software , assisted with carrying out the experiments , collecting the data , and analyzing the data .
lac contributed to conception of the study .
ahd conceived and designed the study , implemented the vre , recruited and consented subjects , supervised the experiments , and co - wrote and proofread the article .
zn conceived and designed the study , designed the signal processing , pattern recognition , and classification algorithms , and co - wrote and proofead the article .
all authors read and approved the final manuscript .
this study was funded by the roman reed spinal cord injury research fund of california ( rr 08 - 258 and rr 10 - 281 ) , and was partially funded by the long beach va advanced fellowship research award . | [ [ background ] ] background : + + + + + + + + + + + spinal cord injury ( sci ) can leave the affected individuals with paraparesis or paraplegia , thus rendering them unable to ambulate . since there are currently no restorative treatments for this population , novel approaches such as brain - controlled prostheses have been sought .
our recent studies show that a brain - computer interface ( bci ) can be used to control ambulation within a virtual reality environment ( vre ) , suggesting that a bci - controlled lower extremity prosthesis for ambulation may be feasible .
however , the operability of our bci has not yet been tested in a sci population .
[ [ methods ] ] methods : + + + + + + + + five subjects with paraplegia or tetraplegia due to sci underwent a 10-min training session in which they alternated between kinesthetic motor imagery ( kmi ) of idling and walking while their electroencephalogram ( eeg ) were recorded .
subjects then performed a goal - oriented online task , where they utilized kmi to control the linear ambulation of an avatar while making 10 sequential stops at designated points within the vre .
multiple online trials were performed in a single day , and this procedure was repeated across 5 experimental days .
[ [ results ] ] results : + + + + + + + + classification accuracy of idling and walking was estimated offline and ranged from 60.5% ( p=0.0176 ) to 92.3% ( p=1.36@xmath010@xmath1 ) across subjects and days .
offline analysis revealed that the activation of mid - frontal areas mostly in the @xmath2 and low @xmath3 bands was the most consistent feature for differentiating between idling and walking kmi . in the online task , subjects achieved an average performance of [email protected] successful stops in 273@xmath451 sec .
these performances were purposeful , i.e. significantly different from the random walk monte carlo simulations ( [email protected] ) , and all but one subject achieved purposeful control within the first day of the experiments .
finally , all subjects were able to maintain purposeful control throughout the study , and their online performances improved over time .
[ [ conclusions ] ] conclusions : + + + + + + + + + + + + the results of this study demonstrate that sci subjects can purposefully operate a self - paced bci walking simulator to complete a goal - oriented ambulation task .
the operation of the proposed bci system requires short training , is intuitive , and robust against subject - to - subject and day - to - day neurophysiological variations .
these findings indicate that bci - controlled lower extremity prostheses for gait rehabilitation or restoration after sci may be feasible in the future .
[ 1995/12/01 ] |
most studies of structure formation in the universe assume that the primordial density perturbations are gaussian .
standard inflationary theories predict gaussian perturbations , and the central limit theorem tells us that any theory involving the superposition of many random processes will give rise to approximately gaussian fluctuations .
however , many well motivated theories predict non - gaussian initial conditions , including topological defect theories ( kibble 1976 ) and certain forms of inflation ( peebles 1983,1997 ) .
so far , no convincing observational evidence has been found to confirm or refute the gaussian hypothesis .
clusters of galaxies , however , provide us with a unique probe of possible non - gaussianity .
being the most massive collapsed structures , they correspond to rare peaks in the primordial density field , so their statistics respond very sensitively to non - gaussianity in the initial matter distribution . in addition , an analysis of the formation of clusters from given initial conditions requires primarily gravitational physics , and is largely free of complications from star formation and feedback . finally ,
although the gravitational evolution which must be modeled is non - linear , well - tested analytical approximation schemes exist for studying the statistics of the resulting cluster distribution . in this work we adapt these analytic schemes to the case of non - gaussian fluctuations , and
use them to make predictions for two simple observations , the number density of clusters and the amplitude of cluster correlations .
we show that the possibility of non - gaussianity introduces a degeneracy in the value of @xmath0 ( where @xmath4 is the _ rms _ overdensity in a sphere of radius @xmath5mpc ) preferred by each of these observations .
however , combining the two constraints breaks the degeneracy , allowing a precise determination of @xmath0 and the degree of non - gaussianity in the universe .
we describe the primordial density field by specifying both the power spectrum @xmath6 and the probability density function ( pdf ) @xmath7 of the fractional over - density @xmath8 after a top - hat smoothing on scale @xmath9 .
we make the simplifying assumption that only the _ rms _ level of fluctuations changes as a function of scale , that is @xmath10 where @xmath11 is a rescaled pdf , whose distribution has zero mean and an _ rms _ value of 1 .
this assumption should hold reasonably well for topological defect theories , whose perturbations are scale invariant over a range of scales , and more generally , will not affect our results provided the form of the pdf does not change significantly over the narrow range of scales relevant to cluster formation .
the press - schechter ( 1974 ) formalism ( hereafter ps ) allows us to compute the number density of clusters within a given background cosmology for gaussian fluctuations .
the derivation suggests a simple way to adapt the ps formula for theories with a general pdf ( as done by chiu , ostriker & strauss 1997 ) : the probability that a given region of space lies inside a collapsed region with pre - collapse radius larger than @xmath9 is taken to be @xmath12 , where @xmath13 is a critical threshold for collapse .
differentiating with respect to @xmath9 and dividing by the pre - collapse volume we obtain the number density @xmath14 of collapsed objects with pre - collapse radius r , where @xmath15 \right| dr.\ ] ] as in the gaussian case ( bond et al .
1991 ) , we have multiplied by a correction factor @xmath16 , to ensure that the entire mass of the universe is accounted for .
we integrate the above expression to obtain cumulative number counts @xmath17 for all objects with a pre - collapse radius greater than @xmath9 : @xmath18 \frac{dy_c(r')}{dr'}\end{aligned}\ ] ] where @xmath19 .
we convert to cumulative number counts as a function of mass and temperature using the method of eke , cole & frenk ( 1996 ) .
the peak - background split , first suggested by kaiser ( 1984 ) , provides a simple scheme for estimating the bias , and hence the correlation amplitude , of collapsed objects .
consider the manner in which the addition of a long wavelength background fluctuation @xmath20 modulates the local number density of collapsed objects .
the threshold for critical collapse due to short wavelength fluctuations is modified locally from @xmath21 to @xmath22 , so the number density of collapsed objects is modified from its background value @xmath17 to @xmath23 .
the factor @xmath24 comes from the linear growth of perturbations .
taylor expanding to first order in @xmath25 we obtain @xmath26 the large scale cluster - cluster correlation function is given by @xmath27 where @xmath28 and @xmath29 is the fractional overdensity in the cluster distribution .
defining bias @xmath30 via @xmath31 where @xmath32 is the correlation function of the dark matter on large scales , we find @xmath33 .
substituting using equation [ eqn - n ] yields the bias : @xmath34 / ( r^{3 } \sigma_r ) -3 \int_r^\infty dr'\ ; p[y_c(r ' ) ] / ( r'^4 \sigma_{r ' } ) } { \int_r^\infty dr'p[y_c(r')]y_c(r')\left(-d\sigma_{r'}/{dr'}\right)/ ( r'^{3 } { \sigma_{r'}})}.\ ] ] for each of our models we specify the background cosmology in terms of @xmath35 and @xmath36 .
we then parameterize the power spectrum @xmath6 of matter fluctuations as a cdm spectrum ( bardeen et al .
1986 , eq . g3 ) with shape parameter @xmath37 ( where @xmath38 ) and normalization specified by @xmath0 . over the range of scales relevant to cluster formation
this form gives a useful fit to power spectra in a variety of structure formation scenarios .
we choose the pdf @xmath11 to be either a @xmath39 or a log - normal distribution , translated and renormalized to have mean zero and standard deviation 1 . in each case
there is one free parameter which allows us to dial the amount of deviation from gaussianity , which we quantify in terms of a new parameter @xmath40 , where @xmath41 is the probability of obtaining a three or higher standard deviation peak for the pdf in question relative to that for a gaussian pdf . to obtain log - normal and @xmath39 distributions with @xmath41 less than 1 ,
we reflect the pdf by making the transformation @xmath42 .
some sample pdfs , along with the value of @xmath41 , are illustrated in fig .
our models are thus specified by five parameters : @xmath35 , @xmath36 , @xmath37 , @xmath0 and @xmath41 .
the ps formula is based on analytic results for the collapse of spherically symmetric perturbations . in the case of gaussian fluctuations
, it provides remarkably good fits to the multiplicity function of cluster scale objects ( gross et al .
1997 ) . for non - gaussian fluctuations
, one may worry that the spherical collapse picture is less applicable , particularly in the case @xmath43 , where the highest density objects will tend to form from matter swept out of low density regions .
as for cluster correlations , our result for the cluster bias in the gaussian case is identical to that of mo & white ( 1996 ) , which has been shown by mo , jing & white ( 1996 ) to give a good fit to the amplitude of cluster correlations in n - body simulations . for highly non - gaussian fluctuations , problems could arise in the separation of the long and short wavelength fluctuations , which are not necessarily uncorrelated .
however , if most clusters form from rare peaks in the initial density field , and if the long and short wavelength modes are nearly uncorrelated , then there is no reason to expect the modified ps or peak - background split formalisms to lead to errors .
our predictions seem to be born out qualitatively by the results of published simulations of non - linear gravitational clustering from non - gaussian initial conditions ( park , spergel & turok 1991 ; weinberg & cole 1992 ; borgani et al .
however , a detailed quantitative comparison with numerical simulations remains to be carried out .
we compare our predictions with two sets of observational data .
firstly , we compare predictions for the cluster temperature function at @xmath44 with observations computed from the catalogue of nearby clusters compiled by henry & arnaud ( 1991 ) . we compute the cumulative number counts at two values of the temperature , and estimate the covariance matrix by bootstrap resampling . for each resampling
, we add gaussian random errors to the temperature consistent with the quoted observational uncertainty , and an additional 10% systematic error to allow for uncertainties in the @xmath45 vs. @xmath41 normalization ( eke , navarro & frenk 1998 ) . the observational data , along with @xmath46errorbars for our two chosen temperatures , are shown in the top panel of fig . 2
secondly , we compare with the amplitude of the cluster - cluster correlation function , quantified by @xmath47 , the value of @xmath48 for which @xmath49 .
croft et al .
( 1997 ) ( hereafter c97 ) have computed @xmath47 as a function of mean cluster separation @xmath50 for subsamples of the apm cluster survey with different lower bounds on the cluster richness .
their results are shown in the bottom panel of fig . 2 .
we compare model predictions with just the @xmath51mpc data point , since this point is derived from the largest number of clusters , and the points are highly correlated .
we use @xmath52mpc , where we have corrected the correlation amplitude for redshift space distortions by subtracting @xmath53mpc , as suggested by the simulations in c97 .
the first errors are their @xmath46 statistical errors and the second are @xmath46 errors we have added to account for systematic uncertainties in the redshift space correction . for each model
we use the techniques described above to compute the cluster correlation length @xmath47 at @xmath54 for clusters with pre - collapse radius determined by their mean separation @xmath55mpc . in fig .
2 we compare observations and predictions for a set of four models , each with @xmath1 , @xmath56 and @xmath2 .
models 1 , 2 and 3 have a normalization given by @xmath57 .
model 1 has gaussian fluctuations , model 2 has @xmath58 ( a reflected @xmath39 distribution with 50 degrees of freedom ) , and model 3 has @xmath59 ( a @xmath39 with 1 degree of freedom ) .
model 4 has the same pdf as model 3 , but a lower normalization , @xmath60 . from the top panel of fig . 2 we see that increasing @xmath41 increases the number density of clusters of a given temperature
the reason for this is clear from fig . 1 : for the value of @xmath0 considered here , only peaks of height roughly @xmath61 and higher ( shown by the heavily shaded region under each curve ) will collapse , since the ratio of the critical overdensity to the standard deviation @xmath0 in a sphere which will collapse to form a typical mass cluster is @xmath62 .
so in this `` standard '' case , the parameter @xmath41 directly counts the expected number of clusters .
now we consider the effect of changes in the pdf on the amplitude of the cluster correlation function . from the bottom panel of fig . 2 we see that increasing @xmath41 decreases the correlation amplitude , for reasons which are also illustrated in fig . 1 . adding a small background fluctuation modifies the local value of the critical overdensity for collapse , leading to a spatial modulation in the number density of collapsed objects . the lightly shaded regions under the curves in fig
1 show the effect for each pdf of modifying @xmath63 from 3.0 to 2.5 . for the @xmath64 pdf which has a long flat tail , this modification produces only a small enhancement in the local number of objects capable of collapse , and
hence there is only a small enhancement in the clustering amplitude of those collapsed objects . for the @xmath65 pdf where the positive tail falls off very sharply
, this modification greatly enhances the local number of collapsed objects , and consequently there is large enhancement in the clustering amplitude of those objects .
finally , we consider the effect on each of the above results of changing @xmath0 for a fixed pdf .
model 4 illustrates the effect of decreasing @xmath0 for the case @xmath64 . as @xmath0 is decreased , the number density of clusters decreases , the reason being that the number of peaks with high enough over - density to collapse is decreased .
the top panel of fig .
2 shows that for @xmath59 , @xmath60 provides a good fit to the temperature function data . in the bottom panel of fig . 2 , we see that as @xmath0 is decreased , the amplitude of the cluster - cluster correlation function also decreases .
this is due to the fact that decreasing @xmath0 decreases the amplitude of the underlying dark matter correlation function , but leaves the bias roughly unchanged .
model 4 shows that for @xmath59 , @xmath60 provides a worse fit to the correlation function data than @xmath57 .
thus we see that the two datasets we have chosen are complementary : variations in the parameter @xmath41 introduce a degeneracy in the preferred value of @xmath0 for each dataset , but the degeneracies are `` orthogonal '' . for the temperature function ,
increasing @xmath41 requires lower @xmath0 , while for the correlation amplitude , increasing @xmath41 requires higher @xmath0 .
so taken in combination , these data allow a precise determination of both @xmath0 and @xmath41 .
we quantify this by computing values of the @xmath39 statistic for each model relative to the two datapoints in the top panel of fig . 2 and the @xmath51mpc data point in the bottom panel . in fig
3 we show goodness of fit contours in the @xmath0 vs. @xmath41 plane for log - normal pdf models with a range of background cosmologies and values of the parameter @xmath37 .
( we use three values of @xmath37 which span the range allowed by measurements of the apm cluster power spectrum by tadros , efstathiou & dalton 1997 . )
the dotted blue line ( @xmath66 ) shows the location in the @xmath41 vs. @xmath0 plane of a gaussian pdf .
the constraints on @xmath0 from present day cluster number abundances have been extensively investigated in this restricted case ( white , efstathiou & frenk 1993 ; eke et al .
1996 ; viana & liddle 1996 ) and the results obtained here are in close agreement .
in addition , mo et al .
( 1996 ) have investigated the combined constraints of cluster correlation amplitude and number abundance measurements for the case of gaussian fluctuations .
our results along this `` line of gaussianity '' can be considered an update of their analysis , as we use more recent cluster correlation data ( c97 ) with smaller observational errors .
the results depend both on the background cosmology and the power spectrum shape parameter @xmath37 . for low values of @xmath35 ,
the value of @xmath0 preferred by the temperature function data increases .
this is just as observed in the gaussian case , and is due to the fact that for low values of @xmath35 , clusters of typical mass form from spheres with a larger pre - collapse radius @xmath9 , for which @xmath4 is less than @xmath0 .
the value of @xmath0 preferred by the correlation amplitude data , however , is almost independent of @xmath35 and @xmath36 .
this is because the correlation amplitude is measured for clusters with a fixed value of @xmath50 , and consequently for clusters with the same pre - collapse radius ( not the same mass ) . finally , as @xmath37 increases , the value of @xmath0 preferred by the correlation amplitude also increases .
this is because the corresponding change in the shape of the power spectrum decreases the amount of power on scales of order the cluster correlation length for a given value of @xmath0 .
results for the best fit values of @xmath0 and @xmath41 from the combined datasets for the family of log - normal pdfs in the various background cosmologies are summarized in table 1 .
we have performed the same calculation for @xmath39-parameterized pdfs and find very similar results .
thus we expect that we can infer results for other models provided we work out the appropriate value of the parameter @xmath41 ( for instance @xmath67 for textures
park et al . 1991 ) . in future work
we will expand on the analysis presented here by including observations of the evolution of cluster abundance and making use of updated datasets at low redshift .
we have modified the press - schechter and the peak - background split formalisms to compute the cluster number density and cluster correlation amplitude in non - gaussian models .
the best - fit value of @xmath0 and level of non - gaussianity depend on the choice of background cosmology , but we have demonstrated how these two quantities can be constrained simultaneously . for a flat , @xmath56 , @xmath2 universe we find that the fluctuations are consistent with gaussian , with @xmath3 ( 95% limits ) . for low-@xmath35 models with @xmath2 , gaussian fluctuations are ruled out with a high degree of confidence , while non - gaussian fluctuations ( @xmath64 ) are strongly preferred . taken in conjunction with accurate knowledge of the background parameters ( which we are rapidly gaining from supernovae and cmb observations ) , this method can provide a powerful constraint on the nature of density fluctuations in the universe .
some concerns remain regarding the validity of our modification of the ps and peak - background split formalisms to the case of non - gaussian fluctuations , and some systematic errors may not yet be accounted for .
this issue should be settled by an analysis of n - body simulations with non - gaussian initial conditions .
however , our formulae have been well tested for the gaussian case , so the result that gaussianity is ruled out for certain choices of background model is not subject to these uncertainties .
at the very least , we have demonstrated the power of two simple and well measured datasets , the cluster temperature function and the amplitude of cluster correlations , to simultaneously constrain the amplitude of mass fluctuations and the level of non - gaussianity in the universe .
we would like to thank p. ferreira , m. davis and m. markevitch for helpful discussions .
this work has been supported in part by a grant from the nsf , and e.g. acknowledges the support of an nsf graduate fellowship .
ccccc @xmath35 & @xmath36 & @xmath37 & @xmath0 & @xmath41 + 1.0 & 0.0 & 0.1 & @xmath68 & @xmath69 + 1.0 & 0.0 & 0.2 & @xmath70 & @xmath71 + 1.0 & 0.0 & 0.3 & @xmath72 & @xmath73 + 0.3 & 0.0 & 0.1 & @xmath74 & @xmath75 + 0.3 & 0.0 & 0.2 & @xmath76 & @xmath77 + 0.3 & 0.0 & 0.3 & @xmath78 & @xmath79 + 0.3 & 0.7 & 0.1 & @xmath80 & @xmath81 + 0.3 & 0.7 & 0.2 & @xmath82 & @xmath83 + 0.3 & 0.7 & 0.3 & @xmath84 & @xmath85 + | we consider constraints on the amplitude of mass fluctuations in the universe , @xmath0 , derived from two simple observations : the present number density of clusters and the amplitude of their correlation function .
allowing for the possibility that the primordial fluctuations are non - gaussian introduces a degeneracy in the value of @xmath0 preferred by each of these constraints .
however , when the constraints are taken together this degeneracy is broken , yielding a precise determination of @xmath0 and the degree of non - gaussianity for a given background cosmology . for a flat , @xmath1 universe with a power spectrum parameterized by a cdm shape parameter @xmath2 , we find that the perturbations are consistent with a gaussian distribution with @xmath3 ( 95% limits ) . for some popular choices of background model , including the favored low matter density models , the hypothesis that the primordial fluctuations are gaussian
is ruled out with a high degree of confidence . |
galactic evolution is driven at the end - points of stellar evolution , where copious mass loss returns chemically - enriched and sometimes dusty matter back to the interstellar medium ( ism ) ; the stellar winds of evolved stars and the violent deaths of the most massive stars also inject energy and momentum into the ism , generating turbulence and galactic fountains when superbubbles pop as they reach the `` surface '' of the galactic disc .
evolved stars are also excellent tracers , not just of the feedback processes , but also of the underlying populations , that were formed from millions to billions of years prior to their appearance .
the evolved phases of evolution generally represent the most luminous , and often the coolest , making evolved stars brilliant beacons at ir wavelengths , where it is also easier to see them deep inside galaxies as dust is more transparent at those longer wavelengths than in the optical and ultraviolet where their progenitors shine .
the final stages of evolution of stars with birth masses up to @xmath18 m@xmath4 asymptotic giant branch ( agb ) stars and red supergiants ( rsgs ) are characterised by strong radial pulsations of the cool atmospheric layers , rendering them identifiable as long - period variables ( lpvs ) in photometric monitoring campaigns spanning months to years ( e.g. , whitelock , feast & catchpole 1991 ; wood et al . 1992 ; wood 2000 ; ita et al . 2004a , b ) .
the prevalent theory of disc galaxy formation hinges on the formation of isolated discs through the dissipational collapse of a gaseous proto - galaxy embedded in cold dark matter halos ( white & rees 1978 ; fall & efstathiou 1980 ; peebles 1984 ) .
thin discs are built through gas accretion , while spheroids and thick discs are built via mergers ( steinmetz & navarro 2002 ; brook et al .
2004 ; brooks et al . 2009 ) .
an alternative scenario for disc formation is based on gas - rich major mergers ( robertson et al .
2006 ; stewart et al .
in fact , each of these scenarios may operate depending on galaxy environments ( blanton et al .
spatially - resolved observational estimates of the star formation histories ( sfhs ) and stellar chemical abundances in spiral galaxies are a robust way to constrain models for their formation and evolution ( eggen et al . 1962 ; freeman & bland - hawthorn 2002 ) .
this has become possible within galaxies of the local group ( e.g. , sarajedini et al .
2000 ; harbeck et al . 2001 ; ferguson et al .
2002 ; brown et al . 2003
; cole et al .
all those methods rely on theoretical models of stellar evolution to provide stellar ages and evolutionary lifetimes , as well as stellar yields , and an assumption about the initial mass function ( imf ) at which stars are born .
m33 is a low luminosity , late type disc spiral galaxy in the local group of galaxies . with a stellar mass of @xmath19@xmath20 m@xmath4 and neutral gas mass of @xmath21 m@xmath4 ( corbelli 2003 ) , m33 is less massive but more gas - rich than the milky way disc , which has @xmath22 m@xmath4 ( licquia & newman ; cf .
mcmillan 2011 ) and @xmath23 m@xmath4 ( nakanishi & sofue 2015 ) , with no sign of recent mergers .
in addition to its proximity ( distance modulus @xmath24 mag ; bonanos et al .
2006 ; u et al .
2009 ) , its large but not too large angular size ( @xmath25 ) and low inclination ( @xmath26 ) offer us a unique opportunity to study its stellar populations , their history and their feedback across an entire spiral galaxy . using colour magnitude diagrams ( cmds ) of resolved stars in m33 , barker et al .
( 2007 ) found a positive age gradient in three fields just outside the break radius on the southern minor axis .
more recently , williams et al .
( 2009 ) found a negative age gradient within the break radius in the radial mass profile , @xmath27 kpc , in four fields on m33 s southern major axis .
this hints at multiple mechanisms at play , that shaped the m33 disc as we see it today , and calls for independent , homogeneous analyses across the entire span of the m33 galaxy . here
we provide such study .
the main objectives of our project are described in javadi , van loon & mirtorabi ( 2011c ) : to construct the mass function of lpvs and derive from this the sfh in m33 ; to correlate spatial distributions of the lpvs of different mass with galactic structures ( spheroid , disc and spiral arm components ) ; to measure the rate at which dust is produced and fed into the ism ; to establish correlations between the dust production rate , luminosity , and amplitude of an lpv ; and to compare the _ in situ _ dust replenishment with the amount of pre - existing dust .
paper i in the series presented the photometric catalogue of stars in the inner square kpc ( javadi et al .
2011a ) , with paper ii presenting the galactic structure and sfh ( javadi , van loon and mirtorabi 2011b ) , and paper iii presenting the mass - loss mechanism and dust production rate ( javadi et al .
paper iv presented the photometric catalogue of stars in a nearly square degree area covering much of the m33 optical disc .
this fifth paper in the series covers the sfh across the enlarged area ; the next paper will present an analysis of the mass return to the m33 disc .
in paper iv we described the method employed to identify large - amplitude lpvs across the galactic disc of m33 with the wfcam imager on the united kingdom ir telescope ( ukirt ) on mauna kea , hawaii .
the area that was monitored covers almost a square degree ( @xmath28 ) ; given the inclination and distance of m33 , we thus probe all of its disc out to a galactocentric radius of 7.5 kpc , i.e. 0.88 @xmath29 ( optical radius ) , and out to 15 kpc ( 1.76 @xmath29 ) in certain directions .
the observations were done between 20052007 in the k@xmath30-band ( @xmath31 @xmath32 m ) with occasionally observations in j- and h - bands ( @xmath33 and 1.68 @xmath32 m , respectively ) for the purpose of obtaining colour information .
the photometric catalogue comprises 403734 stars , among which 4643 stars were identified as lpvs agb stars , super - agb stars and rsgs . their distribution over magnitude and colour
is shown in fig . 1 .
some brighter variable rsgs are found ( around @xmath34 mag ) , but the clump of variable stars with @xmath35 mag are foreground stars and all are saturated .
these stars are removed from further analysis .
the amplitude is generally a@xmath36 mag , but a small fraction of variable stars ( 8% ) reach a@xmath37 mag . among these extreme variables ,
311 have six or fewer measurements and their amplitudes are therefore unreliable . disregarding those , only 20 stars remain with a@xmath38 mag .
we separate the stars in our catalogue into three categories : massive stars , agb stars and rgb stars , on the basis of k@xmath30-band magnitude and @xmath39 colour criteria ( see fig . 1 ) .
we define a demarcation line between hot massive stars and cooler , less - massive giant stars to run from @xmath40 to @xmath41 mag , with massive stars those that have @xmath42 colours bluer than this ( down to @xmath43 mag ) or that have @xmath44 mag , and agb stars and rgb stars those that have @xmath39 colours redder than this and that have @xmath45 mag ( agb ) or @xmath46 mag ( rgb , which does extend to much fainter magnitudes than our survey completeness ) .
we left small gaps to avoid contamination .
the distributions of different types of near - ir populations across the galactic disc of m33 is shown in fig . 2 , deprojected onto the m33 galactic plane , with dashed lines showing the distributions for the variables that were identified in our near - ir monitoring programme .
there is a clear difference between the stellar populations within the inner @xmath47 kpc , with rgb stars showing a core but massive stars in particular those identified by their variability showing a cusp .
further out , all populations show a distribution consistent with an exponential disc .
an exponential profile fit of the form @xmath48 yields @xmath49 for the agb stars ( @xmath50 kpc for the variables ) and @xmath50 kpc for the massive stars ( @xmath51 kpc for the variables , with a sersi core ) .
the rgb population , dominated by old stars , is fit with a sersi profile , with @xmath52 kpc .
this inflated distribution could be a sign of dynamical relaxation .
the relatively large scalelength of the massive stars , which still reflect their birth sites , compared to the agb stars may be an indication of the inside out formation of the disc of m33 , whilst the central hike in density of the massive stars may indicate accretion - induced star formation ( see paper ii ) .
the method used here to derive the sfh strongly depends on the agb phase of stellar evolution . to date , the padova models ( marigo et al . 2008 ) are the most realistic . the full agb evolution is calculated except for the super - agb phase of stellar evolution ( @xmath539 m@xmath4 : siess 2007 ; doherty et al .
2015 ) , which is too fast to be calculated without a separate treatment .
it accurately accounts for the third dredge - up , and hot bottom burning ( hbb ; iben & renzini 1983 ) in the most massive agb stars ( @xmath546 m@xmath4 ) .
they also include the effects of molecular opacities in the cool stellar mantles and atmospheres , which becomes especially apparent in the transformation from oxygen - dominated ( m - type ) agb stars to carbon stars in the birth mass range @xmath554 m@xmath4 ( girardi & marigo 2007 ) .
in addition , and crucial for our method , the padova models predict the pulsation properties and the effects of circumstellar dust production on the spectral energy distribution ( sed ) of the star plus its dust envelope .
separate models cover the massive stars , of which the lower mass range ( @xmath5630 m@xmath4 become rsgs .
the full suite of models thus enables us to evaluate the sfh from as long ago as 10 gyr , to as recent as @xmath57 myr ago .
that said , those models do not consider the effects of magnetic fields or rotation on the mixing process inside the stars . including more efficient mixing would make stars live longer and massive stars evolve to cooler rsgs ( meynet & maeder 2000 ; heger , woosley & spruit 2005 )
likewise , the life times of intermediate - mass stars are prolonged in the presence of convective overshoot ( marigo & giardi 2007 ) .
also , the thermal pulsing agb is computed by adopting parameterisations for the efficiency of third dredge - up , and the threshold temperature at which it sets in , not by solving the equations governing the stellar internal structure .
in paper ii , we presented a novel way to derive the sfh by using cool , large - amplitude lpvs which are identified in ir monitoring programmes .
we use the fact that these variables have reached the very final stages of their evolution , and their brightness can thus be transformed into their mass at birth by employing theoretical evolutionary tracks , or by isochrones .
the lpvs are located at the cool end of each of the isochrones . hence , we convert the observed k - band magnitude ( which is close to the peak in the sed , thus minimising uncertainties in the bolometric corrections ) into the birth mass .
this was done in paper ii for four different metallicities , from super - solar
suitable for massive elliptical galaxies and stellar populations in the bulges of massive spiral galaxies such as the milky way to sub - solar values appropriate for galaxies such as the magellanic clouds ( rezaeikh et al .
as discussed in paper ii , the central region of m33 has an approximately solar metallicity ( rosolowsky & simon 2008 ; magrini et al.2009 ) , for which we adopt @xmath58 .
the disc of m33 outside that central region is characterised by sub - solar metallicity ; in paper iv we showed that @xmath59 agrees well with our cmds as well as the literature .
magrini et al .
( 2007 ) derived an oxygen abundance gradient based on hii regions of @xmath60 dex kpc@xmath5 for @xmath61 kpc and @xmath62 dex kpc@xmath5 for @xmath63 kpc , suggesting a possibly slightly larger range of metallicities .
however , in paper ii we showed that the sfhs derived using @xmath64 and @xmath58 are very similar and hence we are not worried about small deviations from these typical values .
since the method was described in detail in paper ii , here we only provide a brief summary .
the total mass of the stars created between times @xmath65 and @xmath66 is defined as : @xmath67 where @xmath68 is the star formation rate ( sfr ) , in m@xmath4 yr@xmath5 .
the mass @xmath69 corresponds to the number of stars formed , @xmath70 , as : @xmath71 where @xmath72 is the imf , which is described by : @xmath73 where @xmath74 is the normalisation factor and @xmath75 is defined for different ranges in stellar mass , @xmath76 , by kroupa ( 2001 ) : @xmath77 for the minimum and maximum of the stellar mass ranges we adopt @xmath78 m@xmath4 and @xmath79 m@xmath4 , respectively .
the question now is how many of these stars , @xmath80 , are lpvs around time @xmath65 , which is when we observe them .
the number of stars with mass between @xmath81 and @xmath82 that were created between times @xmath65 and @xmath66 is : @xmath83 combining the above equations we have : @xmath84 when we determine @xmath68 over an age bin @xmath85 , the number of lpvs observed in that age bin , @xmath86 , depends on the duration of the evolutionary phase of large - amplitude , long - period variability , the `` pulsation duration '' @xmath87 : @xmath88 inverting the above equations , we obtain a relation between the observed number of lpvs in a certain age bin , and the sfr that long ago : @xmath89 note that the value of the normalisation constant @xmath74 in eq .
( 3 ) does not matter .
the relation between k - band magnitude and mass was constructed in paper ii from the points on each of the padova isochrones when stars reach their peak brightness in the k - band .
the relation was monotonic for agb stars and rsgs ; in the mass regime @xmath901.3 the padova isochrones suggest an excursion towards fainter k - band magnitudes .
the stellar evolution models for these super - agb stars are truncated .
we compared two different approaches to deal with this : [ 1 ] interpolation of the @xmath91@xmath92 relation through the mass range of super - agb stars , or [ 2 ] accepting the jump to fainter magnitudes from agb to super - agb stars .
the second approach yielded unrealistic results for the sfh while the first approach was credible , so we adopt the interpolation approach here also .
dusty lpvs suffer from attenuation . while the isochrones of marigo et al.(2008 ) include an estimate of the photometric effects of circumstellar dust their values are very uncertain .
instead , we correct for the effects of reddening by tracing the reddened stars , along the reddening vector , back to the point on the isochrone where they would lie in the absence of reddening . as fig
. 1 illustrates , there exist two types of reddening ; carbon stars ( e.g. , at @xmath93 gyr , @xmath94 ) redden very rapidly , resulting in a relatively shallow trajectory in the cmd , whereas oxygen - rich stars ( such as low - mass stars at @xmath95 gyr , @xmath96 ) experience a `` greyer '' type of reddening due to the greater transparency of silicates .
we thus obtain reddening - free k - band magnitudes , @xmath97 , from the relation : @xmath98 with @xmath99 mag mag@xmath5 for carbon stars and @xmath100 mag mag@xmath5 for oxygen - rich stars at @xmath58 , and @xmath101 mag mag@xmath5 for carbon stars and @xmath102 mag mag@xmath5 for oxygen - rich stars at @xmath59 .
the reddening correction will be applied to stars with @xmath103 mag . in the case where no j - band data are available we use a similar reddening correction prescribed as a function of @xmath104 colour . in the case
neither j- nor h - band data are available we assume a detection limit of @xmath105 mag .
it is therefore important to distinguish between carbon stars and oxygen - rich stars .
unfortunately , such information is not available for the majority of our lpvs in m33 .
instead , we fix the mass range for carbon stars ; based on theory and observation in the lmc ( groenewegen & de jong 1993 ; van loon , marshall & zijlstra 2005 ; girardi & marigo 2007 ) we expect at solar and slightly sub - solar metallicity carbon stars to descend from stars in the range @xmath554 m@xmath4 at low mass the third dredge - up is not strong enough to change the composition of the surface to carbon dominated , while hbb prevents carbon to enrich the surface in higher mass stars . in practice , we first apply the carbon dust correction to all reddened stars ; if the derived mass is in the 1.54 m@xmath4 range then we accept that star as a carbon star , otherwise we re - apply an oxygenous dust correction .
we refer to paper ii for the relations converting the de - reddened k - band magnitude into birth mass , and the birth mass into lpv age , and for the parameterisation of the pulsation duration . to have a meaningful errorbar on the sfr in each age bin , instead of using equal age bins we adjust the sizes of the age bins such that each contain the same number of stars . despite the rarity of massive lpvs born relatively recently , we are still more able to detect hikes in the star formation rate within the past gyr than 10 gyr ago .
so a recent `` bursty '' sfh does not preclude this from having been a feature throughout the assembly of the galaxy .
finally , in paper iii we realised that the duration of the lpv phase that we detect in our survey is overestimated in the padova models , resulting in too low values for the sfr .
we thus applied a correction to the pulsation duration , and hence to the sfrs that had been derived in paper ii .
we will revisit this important issue in detail for the disc and central region in the discussion section .
furthermore , we stress that , while our method is model dependent , the padova models are the only models that predict the pulsation behaviour , which is essential information for our method to work .
we would like to encourage other groups to also predict the pulsation behaviour , to offer another avenue to empirically explore the differences between models .
other models that computate the thermal - pulsing agb evolution ( but , like the padova models , not for super - agb stars ) include basti ( pietrinferni et al .
2004 , 2006 ) . in the appendix
we compare the mass luminosity and mass age relations between the basti and padova models .
the main difference is that the faintest stars , around @xmath106 mag ) are less massive and older when using the basti models ( @xmath107 gyr compared to 10 gyr when using the padova models ) , but the difference is negligible for intermediate - age stars ( few gyr old ) .
since we already derived the sfh for the central region of m33 using uist data in paper ii , we first show how this compares to the sfh derived from the wfcam data .
the distribution over brightness and the present day ( birth ) mass function of lpvs are shown in fig .
two peaks can be seen , one at @xmath108 and another at @xmath109 .
as described in paper iv , the branch of stars brighter than @xmath110 mag was removed from further analysis because these stars are affected by saturation .
however , rsgs in m33 are fainter than this , so the derived mass function is not affected .
the sfh of the central square kpc of m33 is shown in fig . 4 , separately as derived from the wfcam data and from the uist data .
the oldest bin is only shown in part as it extends to ( unrealistically ) large ages .
the only reason it is plotted is to show that , as expected , the star formation rate we estimate for ages exceeding the hubble time is negligible .
two main epochs of star formation are obvious ; a major epoch of formation @xmath1115 gyr ago ( @xmath1129.7 ) peaking around 4 gyr ago at a level @xmath113 times as high as during the subsequent couple of gyr .
this is slightly shifted to more recent times compared with the uist result for which this epoch is @xmath1118 gyr ago ( @xmath1129.9 ) with a peak around 6 gyr ago ( @xmath114 ) .
this could arise from the inferior angular resolution of wfcam , resulting in incompleteness and blends rendering stars brighter ( hence appearing younger ) than they really are .
the uist data may be affected too , if less so ( the ancient sfr is higher than in the wfcam data ) , and hence the peak at 46 gyr may not be real and star formation could well have started and peaked earlier .
however , this does not change our final conclusion that the sfr in the central regions of m33 has dropped considerably some 4 gyr ago . a second epoch of star formation
is seen to occur from 300 myr20 myr ago ( @xmath1157.2 ) , in both data sets , with a rate @xmath116 times higher than the aforementioned peak .
there is a hint for the sfr to have been decreasing slightly over the course of this more recent epoch , suggesting that something may have happened @xmath117 myr ago which triggered an enhanced sfr .
the wfcam images reach out to a galactocentric radius of nearly 15 kpc , which includes the pseudo - bulge and most of the disc and spiral arms of m33 . the distribution over brightness and the present day ( birth ) mass function of lpvs in this larger area are shown in fig . 5
again , the mass distribution is dominated by two populations of lpvs ; one around @xmath118 and another around @xmath119 .
there are relatively many intermediate - age agb stars in the disc compared with the central region , and this suggests that the average age of stars in the disc is lower than in the central region .
the sfh across the disc of m33 is shown in fig .
the double - peaked mass function is again translated into two epochs of enhanced star formation .
the old epoch took place @xmath1206 gyr ago ( @xmath1219.8 ) , peaking @xmath122 gyr ago ( @xmath123 ) at a level almost twice that in the subsequent few gyr .
barker et al .
( 2011 ) showed that the outer disc of m33 had a major epoch of star formation @xmath1244 gyr ago ( @xmath1250.4 for a standard wmap @xmath126cdm cosmology ; dunkley et al . 2009 ) .
the offset by one gyr compared to the timing of this epoch of star formation in the central region ( derived from the superior uist data ) , suggests that the outer regions of m33 are younger than the central region .
this would confirm the results by li et al .
this conclusion is , however , controversial since cioni et al .
( 2008 ) suggested the opposite .
because their results were based on photometry with wfcam , they may have been affected by the crowding in the central regions . the second
, recent epoch of enhanced star formation occurred @xmath127300 myr ago ( @xmath1288.5 ) , reaching a level of almost 4 times that during the earlier epoch of star formation .
we caution , though , that this is almost certainly due to the difference in time resolution : the peak of the earlier epoch is an average over @xmath129 gyr and peaks lasting only @xmath17 myr may have been much more intense than that .
what is beyond doubt , however , is that the more recent peak in sfr is relatively strong compared to the same epoch in the central region of m33 .
while our approach may favour the assignment of the status of carbon star over oxygen - rich star , the sfh results do not change substantially when assuming that all stars are oxygen - rich ( see appendix ) .
the variation of sfh as a function of radial distance from the centre of m33 ( @xmath130 ) is shown in fig .
each radius bin contains the same number of variable stars ; as a result , the bins grow in size with increasing distance .
two important results become clear immediately ; firstly , the old epoch of star formation weakens and then practically disappears as one moves outward through the disc . in the outermost parts of the disc , the star formation rate gradually increases until @xmath131 myr ( @xmath128 ) ago , and then decreases again in more recent times .
secondly , the very recent epoch of star formation is seen across the entire disc at a similar rate , but lower in the very central region .
more subtly , star formation at intermediate ages , @xmath132 gyr ( @xmath94 ) is most pronounced in the most prominent part of the disc , @xmath133 kpc , but weaker both in the central and the outermost regions .
gratier et al . (
2010 ) showed that the radial profile of total baryonic mass in m33 displays a double exponential component , with a break around @xmath134 kpc . within that radius , while the star formation efficiency is relatively constant the sfh shifts to more recent times as one moves outward ; this is referred to as the `` inside out '' formation scenario . our analysis ( fig .
7 ) clearly confirms this scenario for the disc of m33 .
fig . 7 also suggests a change in sfh around @xmath134 kpc ; beyond that radius , the star formation efficiency decreases ( see also robles - valdez et al .
2013 ) . to examine the difference between the disc within @xmath135 kpc and the outer regions at @xmath136 kpc ,
the mass functions for these two areas are shown separately in fig .
apart from an overall drop in stellar content , the outer regions display a conspicuous dearth of intermediate - mass stars ( between @xmath1370.4 ) . on the other hand ,
the outer regions contain more even in an absolute sense low - mass , presumably old lpvs ( @xmath138 ) . in these sparsely populated regions , stars at @xmath139
could be foreground stars ; we provide an estimate of this in section 4.4 .
the corresponding sfhs are shown in fig .
both at @xmath135 kpc and at @xmath136 kpc there is strong evidence for recent star formation .
the old epoch of star formation is present in both areas as well .
it is possible that the old stars in the outermost regions were formed further in , but migrated outwards through dynamical relaxation or interaction with spiral arms ( grand , kawata & cropper 2012 ) .
we have divided the stellar populations into four different age bins , and plot their spatial distributions in fig .
all populations are centrally concentrated , but this is especially true for the youngest and to a lesser extent the oldest .
the intermediate - age populations are distributed more uniformly across the disc .
it is interesting to note also the asymmetry in the distributions , with a higher density towards the south .
the youngest population exhibits a local overdensity in the north as well , around @xmath140 ( @xmath141 ) in right ascension and @xmath142 in declination .
the spiral structure is not obvious in these graphs . to assess the level of contamination by foreground stars we used the trilegal simulation tool ( girardi et al .
we used the default parameters for the structure of the galaxy , simulating a 0.68 square degree field in the direction of m33 ( @xmath143 , @xmath144 ) .
the region with which we compare this simulation is the area in the galaxy plane that avoided the central @xmath27 kpc .
the result of this comparison is shown in fig .
the variable stars with @xmath145 mag experience only a modest level of contamination ( @xmath146% ) ; those with @xmath147 mag and @xmath148 mag are not affected at all .
the level of contamination for all sources with @xmath149 mag and those with @xmath150 mag and @xmath151 mag is @xmath152% .
these are all upper limits , as most foreground stars would not appear as lpvs .
m33 has two main , inner spiral arms in ( the northern branch ) and is ( the southern branch ) , indicated by the spiral curves in fig .
these logarithmic arms have a pitch angle of @xmath153 and can be traced out to @xmath154 kpc ; they include some of the most prominent hii regions in m33 : ngc604 ( @xmath155 , @xmath156 ) , ic143 ( @xmath157 , @xmath158 ) , ic142 ( @xmath159 , @xmath160 ) , ngc 595 ( @xmath161 , @xmath162 ) in in and ic137 ( @xmath163 , @xmath164 ) in is . however , the spiral arm pattern of m33 is much more complex , with at least four additional sets of arms that can be traced out to radii of @xmath165 , i.e. @xmath166 kpc ( e.g. , newton 1980 ; hagen et al .
2015 ; sandage & humphreys 1980 ) .
a sketch of these spirals based on the location of stellar associations is shown in fig .
12 ( humphreys & sandage 1978 ) . most of the arms are covered completely in our survey except for a small part of arms ivn and vn .
we note previous attempts were made by block et al .
( 2007 ) to trace the spiral arm structure using near - ir maps of evolved stars ; they identified major arcs or carbon stars .
logarithmic spirals have a form @xmath167 when seen in a face - on view .
the pitch angle is defined by @xmath168 $ ] where @xmath169 is assumed to be constant .
parameters for the ten optical arms in m33 have been given in sandage & humphreys ( 1980 ) , including @xmath169 and @xmath170 and the angles ( north through east ) over which each of the arms exists .
to compare the distributions of stars with these logarithmic arm structures , the positions of the stars in the image plane must be deprojected onto the galaxy plane .
then , we can determine the angular distance of stars with respect to each of the arms .
if the spiral arms are the result of a rigidly rotating density wave , then we can express the arc distance , @xmath171 , in terms of the time since the passage of the density wave : @xmath172}\ ] ] we adopt @xmath173 km s@xmath5 kpc@xmath5 from newton ( 1980 ) .
for the stellar velocity function we adopt ( corbelli & schneider 1997 ; in km s@xmath5 ) : @xmath174,\ ] ] where @xmath130 is the radial distance expressed in arcminutes .
this places the co - rotation radius at @xmath175 kpc , beyond which the number of hii regions drops significantly .
obviously , near the co - rotation radius we would not measure any asymmetric lag between stars just born in the density enhancement and older stars dying in the disc behind the wave ( diffusion processes would lead to symmetric dispersal ) .
the calculated time lag @xmath65 of lpvs with respect to arm iin becomes meaningless within a few hundred pc from the co - rotation radius .
we should therefore limit calculations to an area outside the co - rotation radius or set a limit on @xmath65 ( @xmath176 gyr ) .
the resulting distributions over the time lag , by avoiding an area in the disc within @xmath177 kpc from the co - rotation radius , is shown in fig . 13 .
we also inspected the results we would obtain by placing a limit on @xmath65 , as well as by adopting a slightly more extended exclusion zone see appendix .
we divided our populations of lpvs into five age groups : ( 1 ) those formed @xmath178 gyr ago ; ( 2 ) those formed between 5 gyr @xmath179 gyr ago during the first epoch of intense star formation ; ( 3 ) those formed between 1 gyr @xmath180 gyr ago , during the period of moderate star formation activity ; ( 4 ) those formed between 100 myr @xmath181 gyr during the recent epoch of enhanced star formation ; and ( 5 ) massive ( super-)agb stars and rsgs formed @xmath182 myr ago .
the southern arms show a concentration near @xmath183 in all age bins .
arms is and iis show the strongest concentration of massive stars .
the northern arms also show a concentration of stars , in all age bins , near @xmath183 but not generally at @xmath183 itself .
more surprisingly perhaps , there is no evidence whatsoever for a lag associated with the density wave having passed through the position where we see the stars dying now , or any asymmetry at all .
if spiral arms do not leave a lasting imprint on the distribution of young and intermediate - age stellar populations , one could ask the question whether there is any other structure which evolves with time ? to this aim we construct the distribution over nearest - neighbour separations ( fig . 14 ) .
the youngest stars ( @xmath182 myr ) are indeed more agglomerated than the older groups of stars .
the 100 myr @xmath181 gyr population no longer shows any greater agglomeration than still older populations , suggesting that the relaxation time for these structures to dissolve is a few 100 myr at most . to investigate this in more detail , we divided the group of 100 myr @xmath181 gyr stars into two subgroups : those formed between 100 myr @xmath184 myr ( dotted line in fig .
14 ) , and those formed between 300 myr @xmath181 gyr ( dashed line in fig .
the younger of these two groups still shows marginally greater agglomeration , implying a relaxation time of @xmath15 myr ( definitely slower than 100 myr and faster than 300 myr ) . because stars form from dense gas , we would expect a correlation between the ism density and the density of young stars , but not old stars . using the cold dust density map constructed in tabatabaei et al .
( 2014 ) from _ herschel _ space observatory and _ spitzer _ space telescope data of m33 , we show in fig . 15 histograms over ism dust density , per each age group defined previously .
the different age populations exhibit very similar distributions over density , probably because both stars and gas accumulate in the deepest parts of the potential well ( towards the centre , in the disc , and in the spiral arms ) .
the peak ism dust density for all populations is around @xmath185 , which for a gas : dust mass ratio of @xmath15 ( braine et al
. 2010 ) corresponds to ism densities of @xmath186 cm@xmath6 typical of galaxy discs . the youngest stars ( @xmath182 myr ) ,
however , are slightly biased towards higher ism dust density , as was to be expected .
in paper iii we realised that the duration of long period variability is over - estimated by the theoretical padova models , causing mass loss in excess of the birth mass and an under - estimation of the sfr in our method . while we determined an approximate correction factor @xmath57 ( cf .
rezaeikh et al.2014 ) , this issue needs to be revisited in the light of the new data and larger area of our survey .
first , we examine how the sfrs we determined compare with those determined using other methods .
the current or recent sfr in m33 has been estimated from h@xmath75 emission , or the far - uv or far - ir luminosity by numerous groups ( e.g. , hippelein et al.2003 ; engargiola et al . 2003 ; heyer et al .
2004 ; gardan et al .
2007 ; boissier et al . 2007 ; verley et al .
2009 ; kang et al . 2012 ) .
those sfrs range mostly between 0.020.04 m@xmath4 yr@xmath5 kpc@xmath6 for the central regions of m33 .
our new estimate of the current sfr in the central square kpc , which is based on only a small number of rsgs , @xmath188 m@xmath4 yr@xmath5 kpc@xmath6 ( fig .
4 ) is 36 times lower .
verley et al .
( 2009 ) estimated the sfr over the past 100 myr across the disc of m33 to be @xmath189 m@xmath4 yr@xmath5 .
our estimate for the entire disc of m33 ( cf .
6 , multiplied by the area covered ) is @xmath190 m@xmath4 yr@xmath5 , i.e. 78 times lower .
we thus conclude that our srfs require multiplication by a factor @xmath191 to be in line with other , independent estimates .
is there any evidence for this correction factor to depend on mass and/or metallicity ?
the sfrs for different epochs across the disc of m33 are shown in fig .
16 , assuming the same correction factor of 7 applies to all .
if this is correct , then we see that for @xmath192 gyr ago , the sfr reaches a maximum value for @xmath193 kpc and further out it diminishes . at intermediate times , @xmath132 gyr , the radial profile of the sfr becomes much flatter , and in more recent times the sfr peaks somewhere in the middle of the disc , around @xmath1945 kpc ( i.e. not far from the co - rotation radius if there is such a thing as a density wave ) .
how does this compare to independent estimates ?
williams et al .
( 2009 ) determined the sfh based on _ hubble _ space telescope images of four fields located at @xmath195 , 2.5 , 4.3 and 6.1 kpc .
the innermost field was characterised by a first epoch of intense star formation at @xmath1961014 gyr ago , declining to @xmath197 m@xmath4 yr@xmath5 kpc@xmath6 before increasing again to @xmath198 m@xmath4 yr@xmath5 kpc@xmath6 at @xmath19613 gyr ago .
the ( corrected ) sfr at that location derived from our wfcam data is @xmath1990.02 m@xmath4 yr@xmath5 kpc@xmath6 over the past 0.56 gyr . at @xmath200 gyr our sfr is lower , possibly due to incompleteness at low masses . around @xmath201
kpc the hst analysis suggested a much steadier sfr without the initial `` burst '' of star formation , around @xmath202 m@xmath4 yr@xmath5 kpc@xmath6 .
this compares favourably with our estimates at @xmath93 and 5 gyr , of @xmath2030.18 m@xmath4 yr@xmath5 kpc@xmath6 ( fig.16 ) .
further out in the disc , williams et al .
( 2009 ) suggest star formation started at a low rate and gradually increased over time .
this is very similar to what we see ( figs . 7 , 16 and a2 ) . around @xmath204 kpc , at @xmath2058
gyr ago williams et al .
( 2009 ) found @xmath203 m@xmath4 yr@xmath5 kpc@xmath6 , indistinguishable from our result ( fig .
16 ) ; the same is true around @xmath206 kpc .
apart from the oldest stars in the central region , where a larger correction factor might be needed , the consistency with the hst analysis strongly affirms the use of a constant correction factor of 7 for all populations across the disc of m33 . in another study , barker et al .
( 2007 ) observed three fields located at projected radii of @xmath20713 kpc , southeast of m33 s nucleus .
they reported a peak in star formation at @xmath2085 gyr ago ( @xmath2099.7 ) for their fields a1 and a2 with a sfr of @xmath210 m@xmath4 yr@xmath5 kpc@xmath6 and @xmath211 m@xmath4 yr@xmath5 kpc@xmath6 , respectively .
apart from that peak , the sfrs are typically 3.5 , 1.8 and @xmath212 m@xmath4 yr@xmath5 kpc@xmath6 , for their fields a13 . at @xmath213 kpc ,
@xmath2145 gyr ago , we derive @xmath215 m@xmath4 yr@xmath5 kpc@xmath6 ( corrected ; see fig . a1 ) ; further out the sfr drops to @xmath216 m@xmath4 yr@xmath5 kpc@xmath6 .
these rates compare very favourably with the results obtained by barker et al .
( 2007 ) , vindicating the correction by a factor 7 .
the question arises : to what extent is the correction factor due to survey incompleteness ?
we can address this by looking at the central square kpc of m33 .
the uist survey ( paper i ) found @xmath217 variable stars ; the wfcam survey ( paper iv ) found @xmath218 .
they had @xmath219 variable stars in common .
this already suggests that the uist survey is no more complete than @xmath220% , and likewise the wfcam survey is no more complete than 52% .
this is most likely due to the limited sampling of the lightcurves , as most of the variable stars that are sought have amplitudes that exceed the photometric uncertainties ; blending is unlikely to be a major factor as wfcam recovers 95% of the stars ( variable and not variable ) detected with the sharper uist camera .
thus , the survey incompleteness of the uist and wfcam for the large - amplitude lpvs is likely to be close to 60% and 50% , respectively , leaving another factor 3.5 that we attribute to the over - estimation by the models of the pulsation duration .
the latter correction can not be much smaller than that and , consequently , the survey incompleteness can not be much worse than stated above as the mass - loss catastrophy that first triggered the inquest into the correction factor ( paper iii ) required a total correction of a factor 10 .
( we expect that we need to seek to reduce the mass loss estimates by a factor 0.7 , for the mass - loss catastrophy to be avoided when adopting a correction to the number of variable stars of a factor 7 . ) as time passes , the metallicity of the ism and hence that of new generations of stars changes as a result of nucleosynthesis and feedback from dying stars .
older stars are thus expected to have formed in more metal poor environments than younger stars , though this age
metallicity relation may vary spatially within a galaxy .
most of the chemical enrichment occurred @xmath221 gyr ago , when a significant fraction of the total stellar mass was put in place ( note that chemical enrichment naturally slows down dramatically once the metallicity approaches solar values ) . while this means that the assumed metallicity in section 4 is quite reliable for young and intermediate - age populations , the oldest populations ( @xmath222 gyr ) might be deficient in metals . to examine the metallicity effect on the early sfh
, we consider the following metallicities : @xmath59 , 0.004 , 0.0024 , 0.0019 , 0.0015 and 0.0012 .
the parameterisations of the mass luminosity relation , mass pulsation relation and mass - age relation for these different metallicities are provided in the appendix ( see golshan et al .
2016 for details of the fitting ) . while the incorporation in our method of an age
metallicity relation is non - trivial and beyond the scope of the present work , the resulting sfhs for different but uniform values of the metallicity are shown in fig .
interpretation of this diagram must take notice of what is known about the metallicities of populations of different ages within m33 .
barker et al .
( 2007 ) used optical cmds to determine the sfh and age
metallicity relation for three fields at deprojected radii of 913 kpc we consider this the most extreme scenario for any deviation from our assumed metallicity .
sfrs peaked @xmath223 gyr ago in the innermost two fields but @xmath222 gyr in the outermost field .
their metallicities are @xmath2240.006 for most of the past @xmath11 gyr ( @xmath225 ) , but they were @xmath2260.003 around @xmath200 gyr . for @xmath227 gyr ago ( @xmath228 ) we see ( in fig .
17 ) that changing the adopted metallicity from @xmath59 to @xmath229 has a negligible effect on the derived sfr .
such change also does not affect the sfrs at the oldest ages ( @xmath178 gyr ) , but it does increase the sfrs at @xmath230 gyr ago by up to a factor three .
the change is not very much larger for a metallicity as low as @xmath231 or even as low as @xmath232 . certainly , while the derived sfrs in the period @xmath233 gyr ago ( @xmath234 ) would increase by an order of magnitude if the metallicity is as low as @xmath2350.0015 , such low metallicities are not appropriate even for the relatively metal - poor outskirts of the m33 disc . in conclusion ,
the results we have derived by adopting @xmath59 across the disc of m33 are robust against reasonable deviations from this value for the metallicity , with a possible modest underestimate of the star formation around @xmath2308 gyr ago . in the following ,
all sfrs are corrected by a factor 7 .
the star formation rate has varied between @xmath0 ( @xmath1 ) and 0.060@xmath2 ( 0.052@xmath3 ) m@xmath4 yr@xmath5 kpc@xmath6 statistically ( systematically ) in the central square kiloparsec of m33 .
the total star formation rate in m33 within a galactocentric radius of 14 kpc has varied between @xmath7 ( @xmath8 ) and @xmath9 ( @xmath10 ) m@xmath4 yr@xmath5 statistically ( systematically ) .
the statistical estimation of sfrs with associated star number statistics ( represented by the vertical errorbars ) are shown for the central region and the disc in fig . 4 and fig . 6 , respectively .
the systematic estimates of sfrs reflect the variation in sfrs arising from the initial assumptions such as the model we choose , the distance modulus we adopt and the metallicity we adopt .
each of these factors has been investigated carefully in this paper and/or in paper ii ( see fig . 10 in paper ii for the effect of distance modulus , fig . 11 in paper ii and fig .
17 in this paper for the metallicity effect and fig .
a4 for the model effect ) .
we find that the total mass of stars formed in the central square kpc of m33 is @xmath236 m@xmath4 , of which @xmath237 m@xmath4 ( 73% ) was formed @xmath192 gyr ago ( @xmath238 ) , but only @xmath239 m@xmath4 ( 6.5% ) was formed in the most recent few hundred myr .
these values agree with our previous results based on the uist data ( paper ii ) . in the disc of m33 , we find a total mass of stars @xmath240 m@xmath4 , of which @xmath241 m@xmath4 ( 71% ) was formed @xmath242 gyr ago ( @xmath243 ) , but only @xmath244 ( 13% ) was formed in the most recent few hundred myr .
the total mass in stars is at the low end of the range given by corbelli et al .
( 2003 ) ; we probably missed a significant number of the oldest , lowest - mass stars that do not develop large amplitude pulsation .
our results suggest that , currently , star formation is more active in the disc than in the central square kpc , and vice versa in the early assembly of m33 an inside out formation scenario . of course
, this does not take into account the effect of stellar migration , which may be important for the older populations and most easily discerned in the outskirts where no new stars form ( cf .
el - badry et al .
2015 ; miranda et al .
2015 ) .
the recent enhancement in the sfr confirms the result of minniti et al.(1993 ) , who found that the central regions in the disc of m33 experienced enhanced star formation activity over the past 1 gyr .
this may have been caused by interaction with the much more massive andromeda spiral galaxy , m31 ( richardson et al .
evidence for interaction comes from the warped shape of the disc ( sandage & humphreys 1980 ) , and could also have caused the accretion of gas into the central regions necessary to sustain the star formation at the rates observed ( paper iii ) .
hagen et al . ( 2015 ) determined the recent sfh on the basis of uv photometry .
the sfr has varied between 0.01 and 1.4 m@xmath4 yr@xmath5 over the past 400 myr , with the higher values associated with peaks in the sfr around 4 and 10 myr ago .
this is partially due to the better time resolution in recent times , with past such peaks having been smeared out in time to much lower averages .
our estimate of 0.42 m@xmath4 yr@xmath5 over the past 100 myr is in good agreement with their results .
kang et al . (
2012 ) constructed a parametrized model to describe the main characteristics of the m33 sfh , assuming a moderate outflow and inside out formation scenario .
the latter agrees with the negative age gradient in the inner disc measured by williams et al .
( 2009 ) , as well as our @xmath245 . beyond a certain `` break '' radius ,
the older populations regain prominence ( barker et al .
we find that in the central regions the peak in sfr is similar between the old and recent star formation , but this ratio gradually declines as one moves out through the disc of m33 , reaching a minimum around @xmath246 kpc where the ancient sfr is only @xmath57% of that in recent times . at @xmath247
kpc the stellar population becomes older again .
this was also seen for 17 out of 44 face - on galaxies by ruiz - lara et al .
( 2015 ) ; they propose that the disc of these systems were formed in their entirety at early times , but that more recent star formation has been quenched from inside out . while younger stars ( @xmath182 myr ) are more prevalent in the spiral arms of m33 , all stellar populations show a similar concentration towards these arms , with a conspicuous lack of any asymmetry .
this means that while stars migrate out of the arms they do not systematically lag behind them .
this , in turn , means that the spiral arms do not reflect a density wave potential which has a pattern speed that is notably different from that of the stellar orbital velocities .
we conclude that the spiral arms are transient features , in accordance with numerical simulations ( e.g. , grand et al .
junichi et al . ( 2015 ) also performed three - dimensional @xmath70-body / hydrodynamic simulations and found that the spiral pattern in barred galaxies could change on time scales of 100 myr , washing out any relation between older stars and spiral arms .
this means that grand - design spiral arms in barred galaxies are not stationary , but transient .
our results suggest this is also the case in lower - mass , non - barred spiral galaxies ; stars are accumulated temporarily within spiral arms , but over time spans of 100 myr or more the star formation takes place throughout the disc so stars quickly lose memory of where they were formed .
the photometric catalogue of near - ir variable stars of paper iv was used to reconstruct the sfh following the method introduced in paper ii across the galactic disc of m33 and with respect to its spiral arm pattern .
we summarise the main results as follows : * most stars in m33 , @xmath248% were formed @xmath122 gyr ago , i.e.redshift @xmath129 . *
the star formation history of m33 has peaked again more recently , and we detect one such peak @xmath14 myr ago , lasting @xmath15 myr and forming @xmath249% of all stars in m33 . * from the centre outwards through the disc of m33 , the early epoch of star formation becomes progressively less pronounced , indicating a youtfull disc consistent with an inside out formation scenario .
* old stars become more prominent again in the outskirts of m33 , at @xmath166 kpc from the centre , possibly indicating a reduced star formation efficiency in the low density gas in combination with a dynamically relaxed , more extended old stellar population . *
we find that stars of all ages trace some of the spiral arms , but only stars younger than @xmath17 myr show a greater concentration in the arms ; likewise only stars younger than @xmath17 myr show a greater affinity with high density gas .
this indicates that dynamical mixing operates on timescales @xmath250 myr .
* we determine a survey incompleteness of @xmath251% , and suggest the stellar evolution models we use may have overestimated the lifetimes of the large - amplitude lpv phase by a factor 3 to 4 .
here , we show additional diagrams referred to in the text , providing alternative or additional information . in fig .
a1 , we show the sfh assuming all stars are oxygen - rich , i.e. we neglect the fact that stars in a certain mass range will be carbon stars with somewhat different photometric properties .
the sfh does not deviate from that in fig . 6 by a great deal , and our conclusions are therefore not sensitive to the exact manner in which we assign the chemical class to the variable stars . in fig .
a2 , we show how the sfh varies between regions at increasing distance to the centre of m33 .
this contains similar information to that in fig . 7 , but in a way which facilitates reading off the precise values for the sfr . in fig .
a3 , we show the results of the investigation of the association of stars of different ages with the spiral arms , employing a different ( from that in the main body of the paper ) exclusion criterion to avoid the singularities arising at the co - rotation radius .
the conclusions are unaltered . in fig .
a4 , we compare the mass luminosity relation ( in the k - band ) for @xmath59 derived from padova isochrones with basti isochrones . in fig .
a5 , we compare the mass age relation for @xmath59 derived from padova isochrones with basti isochrones . in tables
a1 , a2 and a3 , we present parameterisations of the mass luminosity relation , mass age relation and mass pulsation relation derived from the theoretical models of marigo et al .
( 2008 ) and used in our derivation of the sfh in sections 4 and 5.2 .
ccr @xmath252 & @xmath169 & validity range + + @xmath253 & @xmath254 & @xmath255 + @xmath256 & @xmath257 & @xmath258 + @xmath259 & @xmath260 & @xmath261 + @xmath262 & @xmath263 & @xmath264 + @xmath265 & @xmath266 & @xmath267 + @xmath268 & @xmath269 & @xmath270 + @xmath271 & @xmath272 & @xmath273 + + @xmath274 & @xmath275 & @xmath276 + @xmath277 & @xmath278 & @xmath279 + @xmath280 & @xmath281 & @xmath282 + @xmath283 & @xmath284 & @xmath285 + @xmath286 & @xmath287 & @xmath288 + @xmath289 & @xmath290 & @xmath291 + @xmath292 & @xmath293 & @xmath294 + + @xmath295 & @xmath296 & @xmath297 + @xmath298 & @xmath299 & @xmath300 + @xmath301 & @xmath302 & @xmath303 + @xmath304 & @xmath305 & @xmath306 + @xmath307 & @xmath308 & @xmath309 + @xmath310 & @xmath311 & @xmath312 + @xmath313 & @xmath314 & @xmath315 + + @xmath316 & @xmath317 & @xmath318 + @xmath319 & @xmath320 & @xmath321 + @xmath322 & @xmath323 & @xmath324 + @xmath325 & @xmath326 & @xmath327 + @xmath328 & @xmath329 & @xmath330 + @xmath331 & @xmath332 & @xmath333 + @xmath334 & @xmath335 & @xmath336 + + @xmath337 & @xmath338 & @xmath339 + @xmath340 & @xmath341 & @xmath342 + @xmath343 & @xmath344 & @xmath345 + @xmath346 & @xmath347 & @xmath348 + @xmath349 & @xmath350 & @xmath351 + @xmath352 & @xmath353 & @xmath354 + @xmath355 & @xmath356 & @xmath357 + + @xmath358 & .076 & @xmath359 + @xmath360 & .275 & @xmath361 + @xmath362 & @xmath363 & @xmath364 + @xmath365 & @xmath366 & @xmath367 + @xmath368 & @xmath369 & @xmath370 + @xmath371 & @xmath372 & @xmath373 + @xmath374 & @xmath375 & @xmath376 + @xmath377 & @xmath378 & @xmath379 + ccr + @xmath380 & .425 & @xmath381 + @xmath382 & .696 & @xmath383 + @xmath384 & @xmath385 & @xmath386 + @xmath387 & @xmath388 & @xmath389 + @xmath390 & @xmath391 & @xmath392 + @xmath393 & @xmath394 & @xmath395 + @xmath396 & @xmath397 & @xmath398 + @xmath399 & @xmath400 & @xmath401 + ccr @xmath252 & @xmath169 & validity range + + @xmath402 & @xmath403 & @xmath404 + @xmath405 & @xmath406 & @xmath407 + @xmath408 & @xmath409 & @xmath410 + @xmath411 & @xmath412 & @xmath413 + @xmath414 & @xmath415 & @xmath416 + @xmath417 & @xmath418 & @xmath419 + + @xmath420 & @xmath421 & @xmath422 + @xmath423 & @xmath424 & @xmath425 + @xmath426 & @xmath427 & @xmath428 + @xmath429 & @xmath430 & @xmath431 + @xmath432 & @xmath433 & @xmath434 + @xmath435 & @xmath436 & @xmath437 + @xmath438 & @xmath439 & @xmath440 + + @xmath441 & @xmath442 & @xmath443 + @xmath444 & @xmath445 & @xmath446 + @xmath447 & @xmath448 & @xmath449 + @xmath450 & @xmath451 & @xmath452 + @xmath453 & @xmath454 & @xmath455 + @xmath456 & @xmath457 & @xmath458 + @xmath459 & @xmath460 & @xmath461 + + @xmath462 & @xmath463 & @xmath464 + @xmath465 & @xmath466 & @xmath467 + @xmath468 & @xmath469 & @xmath470 + @xmath471 & @xmath472 & @xmath473 + @xmath474 & @xmath475 & @xmath476 + @xmath477 & @xmath478 & @xmath479 + + @xmath480 & @xmath481 & @xmath482 + @xmath483 & @xmath484 & @xmath485 + @xmath486 & @xmath487 & @xmath488 + @xmath489 & @xmath490 & @xmath491 + @xmath492 & @xmath493 & @xmath494 + @xmath495 & @xmath496 & @xmath497 + @xmath498 & @xmath499 & @xmath500 + + @xmath501 & @xmath502 & @xmath503 + @xmath504 & @xmath505 & @xmath506 + @xmath507 & @xmath508 & @xmath509 + @xmath510 & @xmath511 & @xmath512 + @xmath513 & @xmath514 & @xmath515 + @xmath516 & @xmath517 & @xmath518 + @xmath519 & @xmath520 & @xmath521 + @xmath522 & @xmath523 & @xmath524 + ccr + @xmath525 & @xmath526 & @xmath527 + @xmath528 & @xmath529 & @xmath530 + @xmath531 & @xmath532 & @xmath533 + @xmath534 & @xmath535 & @xmath536 + @xmath537 & @xmath538 & @xmath539 + @xmath540 & @xmath541 & @xmath542 + @xmath543 & @xmath544 & @xmath545 + @xmath546 & @xmath547 & @xmath548 + ccccc @xmath549 & @xmath550 & @xmath252 & @xmath169 & @xmath551 + + @xmath552 & 1 & 1.955 & 0.173 & 0.074 + & 2 & 2.545 & 0.545 & 0.196 + & 3 & 2.936 & 1.539 & 0.429 + + @xmath553 & 1 & 0.850 & 0.202 & 0.080 + & 2 & 2.265 & 0.547 & 0.188 + & 3 & 2.748 & 1.483 & 0.410 + + @xmath553 & 1 & 1.02 & 0.23 & 0.056 + & 2 & 1.43 & 0.53 & 0.129 + & 3 & 2.52 & 1.29 & 0.576 + + @xmath554 & 1 & 1.267 & 0.222 & 0.064 + & 2 & 1.831 & 0.529 & 0.147
+ & 3 & 2.567 & 1.302 & 0.427 + + @xmath552 & 1 & 1.510 & 0.229 & 0.065 + & 2 & 1.702 & 0.508 & 0.149 + & 3 & 3.258 & 1.320 & 0.519 + + @xmath555 & 1 & 2.34 & 1.281 & 0.378 + & 2 & 1.32 & 0.460 & 0.165 + & 3 & 0.38 & 0.145 & 0.067 + + @xmath556 & 1 & 2.18 & 1.238 & 0.238 + & 2 & 1.36 & 0.514 & 0.127 + & 3 & 0.30 & 0.228 & 0.084 +
we thank the staff at ukirt for their excellent support of this programme .
jvl thanks the school of astronomy at ipm , tehran , for their hospitality during his visits .
we are grateful for financial support by the leverhulme trust under grant no .
rf/4/rfg/2007/0297 , by the royal astronomical society , and by the royal society under grant no .
fst acknowledges financial support from the spanish ministry of economy and competitiveness ( mineco ) under grant number aya2013 - 41243-p .
finally , we thank the referee for her / his constructive report which prompted us to improve the manuscript .
this work has made use of basti web tools .
barker m. k. , sarajedini a. , geisler d. , harding p. , schommer r. , 2007 , aj , 133 , 1138 barker m. k. , ferguson a. m. n. , cole a. a. , ibata r. , irwin m. , lewis g.f .
, smecker - hane t. a. , tanvir n. r. , 2011 , mnras , 410 , 504 bertelli g. , mateo m. , chiosi c. , bressan a. , 1992 , apj , 388 , 400 bigiel f. , leroy a. , walter f. , brinks e. , de block w. j. g. , madore b. , thornley m. d. , 2008 , aj , 136 , 2846 blanton m. r. , et al .
, 2003 , apj , 594 , 186 block d. l. , et al .
, 2007 , a&a , 471 , 467 boissier s. , et al . , 2007 , apjs , 173 , 524 bonanos a. z. , et al . , 2006 , apj , 652 , 313 braine j. , et al . , 2010 , a&a , 518 , l69 brook c. b. , kawata d. , gibson b. k. , freeman k. c. , 2004 , apj , 612 , 894 brooks a. m. , governato f. , quinn t. , brook c. b. , wadsley j. , 2009 , apj , 694 , 396 brown t. m. , ferguson h. c. , smith e. , kimble r. a. , sweigart a. v. , renzini a. , rich m. r. , vandenberg d. a. , 2003 , apj , 592 , l17 chang r. x. , hou j. l. , shen s. y. , shu c. g. , 2010 , apj , 722 , 380 cioni m .-
l. , irwin m. , ferguson a. m. n. , mcconnachie a. , conn b. c. , huxor a. , ibata r. , lewis g. , tanvir n. , 2008 , a&a , 487 , 13 cole a. a. , tolstoy e. , gallagher j. s. iii .
, smecker - hane t. a. , 2005 , aj , 129 , 1465 corbelli e. , 2003 , mnras , 342 , 199 corbelli e. , schneider s. , 1997 , apj , 479 , 244 doherty c. l. , gil - pons p. , siess l. , lattanzio j. c. , lau h. h. b. , 2015 , mnras , 446 , 2599 dohm - palmer r. c. , et al .
, 1997 , aj , 114 , 2527 dunkley j. , et al . , 2009 , apjs , 180 , 306 eggen o. j. , lynden - bell d. , sandage a. r. , 1962 , apj , 136 , 748 el - badry k. , wetzel a. , geha m. , hopkins p. f. , kere d. , chan t. k. , faucher - gigure c .- a .
, 2015 , arxiv:1512.01235 engargiola g. , plambeck r. l. , rosolowsky e. , blitz l. , 2003 , apjs , 149 , 343 fall s. m. , efstathiou g. , 1980 , mnras , 193 , 189 ferguson a. m. n. , irwin m. j. , ibata r. a. , lewis g. f. , tanvir n. r. , 2002 , aj , 124 , 1452 freeman k. c. , bland - hawthorn j. , 2002 , ara&a , 40 , 487 gallart c. , freedman w. l. , aparicio a. , bertelli g. , chiosi c. , 1999 , aj , 118 , 2245 gardan e. , braine j. , schuster k. f. , brouillet n. , sievers a. , 2007 , a&a , 473 , 91 girardi l. , marigo p. , 2007 ,
a&a , 462 , 237 girardi l. , groenewegen m. a. t. , hatziminaoglou e. , da costa l. , 2005 , a&a , 436 , 895 golshan r. h. , javadi a. , van loon j. th . , khosroshahi h. , saremi e. , 2016 , mnras , submitted grand r. j. j. , kawata d. , cropper m. , 2012 , mnras , 421 , 1529 gratier p. , et al .
, 2010 , a&a , 522 , a3 groenewegen m. a. t. , de jong t. , 1993 , a&a , 267 , 410 hagen l. m. z. , siegel m. h. , gronwall c. a. , hoversten e. a. , vargas a. , immler s. , 2015 , arxiv:1504.6635 harbeck d. , et al . , 2001 , aj , 122 , 3092 harris j. , zaritsky d. , 2004 , aj , 127 , 153 heger a. , woosley s. e. , spruit h. c. , 2005 , apj , 626 , 350 heyer m. h. , corbelli e. , schneider s. e. , young j. s. , 2004 , apj , 602 , 723 hippelein h. , haas m. , tuffs r. j. , lemke d. , stickel m. , klaas u. , volk h.j . , 2003 , a&a , 407 , 137 holtzman j. a. , et al . , 1999 ,
aj , 118 , 2262 humphreys r. , sandage a. , 1978 , ann.rep.carnegie inst .
, 77 , 180 iben i. jr . ,
renzini a. , 1983 , ara&a , 21 , 271 ita y. , et al .
, 2004a , mnras , 347 , 720 ita y. , et al . , 2004b , mnras , 353 , 705 javadi a. , van loon j. th . , mirtorabi m. t. , 2011a , mnras , 411 , 263 ( paper i ) javadi a. , van loon j. th . , mirtorabi m. t. , 2011b , mnras , 414 , 3394 ( paper ii ) javadi a. , van loon j. th . , mirtorabi m. t. , 2011c , in : why galaxies care about agb stars ii , eds . f. kerschbaum , t. lebzelter & r. f. wing , aspc , 445 , 497 javadi a. , van loon j. th .
, khosroshahi h. , mirtorabi m. t. , 2013 , mnras , 432 , 2824 ( paper iii ) javadi a. , van loon j. th .
, khosroshahi h. , mirtorabi m. t. , 2014 , mnras , 432 , 2824 ( paper iv ) kang x. , chang r. , yin j. , hou j. , zhang f. , zhang y. , han z. , 2012 , mnras , 426 , 1455 kennicutt r. c. , 1998 , apj , 498 , 541 kennicutt r. c. , et al . , 2007 , apj , 671 , 333 leroy a. k. , walter f. , brinks e. , bigiel f. , de blok w. j. g. , madore b. , thornley m. d. , 2008 , aj , 136 , 2782 li j. , et al . , 2004 , a&a , 420 , 89 licquia t. c. , newman j. a. , 2015 , apj , 806 , 96 magrini l. , stanghellini l. , villaver e. , 2009 , apj , 696 , 729 magrini l. , corbelli e. , galli d. , 2007a , a&a , 470 , 843 magrini l. , vlchez j. m. , mampaso a. , corradi r. l. m. , leisy p. , 2007a , a&a , 470 , 865 marcon - uchida m. m. , matteucci f. , costa r. d. d. , 2010 , a&a , 520 , a35 marigo p. , girardi l. , 2007 , a&a , 469 , 239 marigo p. , girardi l. , bressan a. , groenewegen m. a. t. , silva l. , granato g. l. , 2008 , a&a , 482 , 883 martnez - delgado d. , aparicio a. , gallart c. , 1999 , aj , 118 , 2229 mcmillan p. j. , 2011 , mnras , 414 , 2446 meynet g. , maeder a. , 2000 , a&a , 361 , 101 miller b. w. , dolphin a. e. , lee m. g. , kim s. c. , hodge p. , 2001 ,
apj , 562 , 713 miranda m. s. , et al .
, 2015 , arxiv:1512.04559 mould j. , kristian j. , 1986 , apj , 305 , 591 nakanishi h. , sofue y. , 2015 , arxiv:1511.08877 newton l. , 1980 , mnras , 190 , 689 olsen k. a. g. , 1999 , aj , 117 , 2244 peebles p. j. e. , 1984 , apj , 277 , 470 pietrinferni a. , cassisi s. , salaris m. , castelli f. , 2004 , apj , 612 , 168 pietrinferni a. , cassisi s. , salaris m. , castelli f. , 2006 , apj , 642 , 797 rezaeikh s. , javadi a. , khosroshahi h. , van loon j. th , 2014 , mnras , 445 , 2214 richardson j. c. , et al . , 2011 , apj , 732 , 76 robertson b. , bullock j. s. , cox t. j. , di matteo t. , hernquist l. , springel v. , yoshida n. , 2006 , apj , 645 , 986 robles - valdez f. , carigi l. , peimbert m. , 2013 , mnras , 429 , 2351 rosolowsky e. , simon j. d. , 2008 , apj , 675 , 1213 ruiz - lara t. , et al . , 2015 , arxiv:1511.03499 sandage a. , humphreys r. m. , 1980 , apj , 263 , 1 sarajedini a. , geisler d. , schommer r. , harding p. , 2000
, aj , 120 , 2437 siess l. , 2007 , a&a , 476 , 893 skillman e. d. , tolstoy e. , cole a. a. , dolphin a. e. , saha a. , gallagher j. s. , dohm - palmer r. c. , mateo m. , 2003 , apj , 596 , 253 steinmetz m. , navarro j. f. , 2002 , new astronomy , 7 , 155 stewart k. r. , bullock j. s. , wechsler r. h. , maller a. h. , 2009 , apj , 702 , 307 tabatabaei f. , et al . , 2014 , a&a , 561 , 95 tolstoy e. , et al . , 1998 , aj , 116 , 1244 tosi m. , greggio l. , marconi g. , focardi p. , 1991
, aj , 102 , 951 u v. , urbaneja m. a. , kudritzki r .-
, jacobs b. a. , bresolin f. , przybilla n. , 2009 , apj , 704 , 1120 van loon j. th . , marshall j. r. , zijlstra a. a. , 2005 , a&a , 442 , 597 verley s. , corbelli e. , giovanardi c. , hunt l. k. , 2009 , a&a , 493 , 453 williams b. f. , dalcanton j. j. , dolphin a. e. , holtzman j. , sarajedini a. , 2009 , apj , 695 , l15 white s. d. m. , rees m. j. , 1978 , mnras , 183 , 341 whitelock p. , feast m. , catchpole r. , 1991 , mnras , 248 , 276 wong t. , blitz l. , 2002 , apj , 569 , 157 wood p. r. , 2000 , pasa , 17 , 18 wood p. r. , whiteoak j. b. , hughes s. m. g. , bessell m. s. , gardner f.f . , hyland a. r. , 1992 , apj , 397 , 552 wyder t. , 2003 , aj , 125 , 3097 zaritsky d. , elston r. , hill j.m .
, 1989 , aj , 97 , 97 | we have conducted a near - infrared monitoring campaign at the uk infrared telescope ( ukirt ) , of the local group spiral galaxy m33 ( triangulum ) . on the basis of their variability
, we have identified stars in the very final stage of their evolution , and for which the luminosity is more directly related to the birth mass than the more numerous less - evolved giant stars that continue to increase in luminosity . in this fifth paper of the series , we construct the birth mass function and hence derive the star formation history across the galactic disc of m33 .
the star formation rate has varied between @xmath0 ( @xmath1 ) and 0.060@xmath2 ( 0.052@xmath3 ) m@xmath4 yr@xmath5 kpc@xmath6 statistically ( systematically ) in the central square kiloparsec of m33 , comparable with the values derived previously with another camera .
the total star formation rate in m33 within a galactocentric radius of 14 kpc has varied between @xmath7 ( @xmath8 ) and @xmath9 ( @xmath10 ) m@xmath4 yr@xmath5 statistically ( systematically ) .
we find evidence of two epochs during which the star formation rate was enhanced by a factor of a few one that started @xmath11 gyr ago and lasted @xmath12 gyr and produced @xmath13% of the total mass in stars , and one @xmath14 myr ago that lasted @xmath15 myr and formed @xmath16% of the mass in stars .
radial star formation history profiles suggest that the inner disc of m33 was formed in an inside out formation scenario .
the outskirts of the disc are dominated by the old population , which may be the result of dynamical effects over many gyr .
we find correspondence to spiral structure for all stars , but enhanced only for stars younger than @xmath17 myr ; this suggests that the spiral arms are transient features and not part of a global density wave potential .
[ firstpage ] stars : luminosity function , mass function galaxies : evolution galaxies : individual : m33 galaxies : spiral galaxies : stellar content galaxies : structure |
let us now introduce some notations that will be used throughout this paper .
by @xmath1 we denote the sets of all natural numbers , integers , real numbers and positive real numbers , respectively . for @xmath2 , @xmath3 ( @xmath4 ) , define _ integer intervals _ @xmath5=\{a , a+1,a+2,\cdots\}$ ] , @xmath6=\{a , a+1,\cdots , b\}$ ] . for @xmath7 , @xmath8 , denote by @xmath9 the open @xmath10neighborhood of @xmath11 . for a convergent bi - infinite sequence @xmath12 , denote by @xmath13 the limits of the sequence as @xmath14 tends to @xmath15 , i.e. @xmath16 and @xmath17 .
consider the following second order equation @xmath18 where @xmath19 , @xmath20 for all @xmath21 , @xmath22 is the forward difference operator defined by @xmath23 and @xmath24 .
a solution @xmath25 of ( [ ( 1.1 ) ] ) is called a _ heteroclinic solution _ ( or _ heteroclinic orbit _ )
if there exist @xmath26 , @xmath27 , such that @xmath28 are two equilibria of ( [ ( 1.1 ) ] ) and @xmath29 we are interested in the problem of the existence and multiplicity heteroclinic solutions of ( [ ( 1.1 ) ] ) .
so far as we are aware , it is the first time in the literature for us to study heteroclinic orbits of difference equations . equation ( [ ( 1.1 ) ] ) can be considered as a discrete analogue of the classical pendulum equation : @xmath30 the phase plane portrait of ( [ ( 1.1 ) ] ) with @xmath31 , shown on figure [ fig : de ] , can be compared to the phase plane portrait of ( [ ( 1.2 ) ] ) with @xmath32 , shown on figure [ fig : ode ] .
the phase plane analysis of ( [ ( 1.2 ) ] ) shows the existence of two heteroclinic solutions for ( [ ( 1.2 ) ] ) joining @xmath33 and @xmath34 . on the other hand , the phase plane of ( [ ( 1.1 ) ] ) is similar to that of ( [ ( 1.2 ) ] ) . on figure
[ fig : de ] , we use colors to distinguish between different orbits . nine ellipses represent nine periodic orbits , while two curves around nine ellipses are non - periodic orbits .
close similarities observed on figures [ fig : de ] and [ fig : ode ] suggest the existence of heteroclinic orbits for ( [ ( 1.1 ) ] ) .
our goal in this paper is to show that there indeed exist two heteroclinic solutions of ( [ ( 1.1 ) ] ) joining @xmath33 and @xmath34 also .
let us now recall briefly the existence and multiplicity heteroclinic orbits for the following hamiltonian system @xmath35 which is a generalization form of ( [ ( 1.2 ) ] ) , where @xmath36 , @xmath37 and @xmath38 . in the past twenty years , many authors had studied the existence and multiplicity of heteroclinic solutions and heteroclinic chains for ( [ ( 1.3 ) ] ) .
the first result in this area was proved in @xcite , where the author discussed ( [ ( 1.3 ) ] ) under assumptions : @xmath39 @xmath40 , @xmath41 @xmath42 , and @xmath43 @xmath44 is periodic in @xmath45 with the period @xmath46 , @xmath47
. conditions @xmath41 and @xmath43 imply that @xmath44 has a global maximum on @xmath48 . without loss of generality , it is assumed that the global maximum of @xmath44 is 0 and put @xmath49 . under the non - degeneracy condition : @xmath50 @xmath51 consists only of isolated points , the following result was obtained in @xcite .
* theorem a. * _ under assumptions @xmath52 , for every @xmath53 , there exist at least two heteroclinic orbits of ( [ ( 1.3 ) ] ) joining @xmath54 to @xmath55 .
at least one of these orbits emanates from @xmath54 and at least one terminates at @xmath54_. let @xmath56 , @xmath57 and @xmath40 . then ( [ ( 1.3 ) ] ) becomes ( [ ( 1.2 ) ] ) , and @xmath58 .
theorem a guarantees at least two heteroclinic orbits of ( [ ( 1.2 ) ] ) through every point of @xmath51 .
further development in this direction was done by felmer ( cf .
@xcite ) , who generalized the above result to first order spatially periodic hamiltonian systems . by using the saddle point theorem
, the author obtained the existence of heteroclinic orbits joining two saddle type critical points . without imposing the non - degeneracy condition , caldiroli and jeanjean ( cf .
@xcite ) studied conservative singular hamiltonian system without forcing term , and were able to establish the existence of heteroclinic orbits joining a global maximum point and a non - constant periodic solution .
in the case when the potential is periodic and time reversible , by using minimization arguments , rabinowitz showed the existence of heteroclinic solutions between pairs of periodic solutions , ( cf .
@xcite ) . under the same assumptions , maxwell ( cf .
@xcite ) proved that there exist heteroclinic chains connecting every pairs of periodic solutions . for non - autonomous hamiltonian systems without forcing term , strobel ( cf .
@xcite ) studied the existence of heteroclinic chains between pairs of equilibria . by using constrained minimization and comparison arguments , rabinowitz and zelati ( cf .
@xcite ) studied ( [ ( 1.3 ) ] ) without forcing term , and found multiple heteroclinic chains joining pairs of equilibria .
subsequent progress was done by bertotti and montecchiari ( cf .
@xcite ) , who generalized the results of strobel , by proving the existence of infinitely many heteroclinic solutions for a class of forced slowly oscillating hamiltonian system with potential @xmath59 of form @xmath60 , with @xmath61 being almost periodic in @xmath62 . next ,
alessio , bertotti and montecchiari ( cf .
@xcite ) obtained a generalization of the results of @xcite , in which @xmath63 is replaced by @xmath64 for @xmath8 small enough .
however , without the non - degeneracy , these results are not as strong as those of strobel . in the case of forced slowly oscillating hamiltonian systems , rabinowitz ( cf .
@xcite ) showed the existence of basic and even more complex heteroclinic orbits , without making any non - degeneracy assumption .
then , zelati and rabinowitz ( cf .
@xcite ) showed that there exist heteroclinic solutions joining two stationary points in different energy levels , which was established by using minimization arguments .
we should also mention the work by chen and tzend ( cf .
@xcite ) , in which variational and penalization methods were being used to study the existence of heteroclinic orbits for the following system @xmath65 where @xmath44 is not periodic nor asymptotically periodic in @xmath62 . in those papers ,
the authors obtained multiple heteroclinic orbits and chains joining pairs of equilibria as well as joining an equilibrium to a non - constant periodic solution .
izydorek and janczewska ( cf .
@xcite ) proved , without assuming periodicity or almost periodicity in @xmath62 of the potential , for ( [ ( 1.3 ) ] ) without forcing term , the existence of heteroclinic solutions joining pairs of equilibria .
however , no results on the existence of heteroclinic solutions of difference equations have been proved . in this paper , by using variational arguments , we will study the existence and multiplicity of heteroclinic solutions for ( [ ( 1.1 ) ] ) . to this end , we need to choose a suitable functional space on which a variational functional , associated with ( [ ( 1.1 ) ] ) , can be constructed .
however , lack of continuity assumption leads to some new problems which were not present in the case of differential systems .
for example , for differential systems , if an orbit contains two points such that one of them is outside of @xmath66 , while the other belongs to inside of @xmath67(@xmath68 ) , then the orbit ( because of its continuity ) contains a point belonging to @xmath69 .
however , such a statement is not valid for orbits of discrete systems .
in this section , we study the existence and multiplicity of heteroclinic orbits of ( [ ( 1.1 ) ] ) joining every two adjacent points of @xmath70 . for simplicity
, we make an image translation . by applying the substitution @xmath71 , ( [ ( 1.1 ) ] ) can be rewritten as @xmath72 we look for heteroclinic orbits of ( [ ( 2.1 ) ] ) which join two adjacent points of @xmath73 .
let @xmath74 be the vector space of all convergent sequences @xmath75 , i.e. @xmath76 we define the space @xmath77 by @xmath78 and put @xmath79^\frac{1}{2},\qquad \forall\ y\in h.\label{norm}\end{aligned}\ ] ] then we have [ 1 ] the bilinear product ( [ inner ] ) is an inner product on @xmath77 and the space @xmath77 equipped with the norm given by ( [ norm ] ) is a hilbert space .
recall that the space @xmath80 of all sequences @xmath81 such that @xmath82^{\frac 12}<\infty,\ ] ] is a hilbert space .
let @xmath83 be a cauchy sequence in @xmath77 , i.e. @xmath84^{\frac 12}<\ve.\ ] ] then @xmath85 is a cauchy sequence in @xmath86 , while @xmath87 , @xmath88 , is a cauchy sequence in @xmath80 . by completeness of @xmath80
, there exists a limit @xmath2 in @xmath80 of @xmath89 .
one can easily observe , that there exists a unique @xmath90 in @xmath77 such that @xmath91 by passing to the limit as @xmath92 goes to @xmath93 , we obtain from ( [ cauchy ] ) @xmath94 which proves that @xmath95 converges to @xmath96 . consequently , @xmath77 is a hilbert space .
similar arguments as those presented in @xcite , one can define variational functional @xmath97 associated with ( [ ( 2.1 ) ] ) by @xmath98.\ ] ] put @xmath99 and @xmath100 .
[ r3]for every @xmath101 , if @xmath102 , then @xmath103 , @xmath104 . indeed ,
suppose for example @xmath105 , then there exist @xmath106 and @xmath107 such that @xmath108 for all @xmath109 .
therefore , @xmath110 given @xmath111 , @xmath112 , define the set @xmath113 of all @xmath114 satisfying * @xmath115 , * @xmath116 , * @xmath117 for all @xmath118 . obviously , @xmath113 is not empty for all @xmath119 .
define @xmath120 now we give a simple but useful lemma . [ 2 ] given a sequence of disjoint integer intervals @xmath121 , @xmath122 and @xmath123 .
let @xmath114 be such that @xmath124 then , @xmath125 let @xmath126 . since for @xmath127 @xmath128 and since @xmath129 and @xmath130 for @xmath131 , we have @xmath132 where @xmath133 .
assume @xmath134 .
we will prove the existence of an orbit minimizing function @xmath135 restricted to @xmath113 .
for this purpose , we need the following two lemmas . [ 4 ] consider @xmath136 and assume that @xmath137 is a minimizing sequence for ( [ ( 2.3 ) ] ) restricted to @xmath113 , such that for any @xmath138 uniformly for @xmath139 $ ] . if @xmath114 and @xmath102 , then @xmath140 . by remark [ r3 ] ,
there exist @xmath141 , @xmath142 such that @xmath143 . by assumption @xmath144 uniformly for @xmath145 $ ] and @xmath146 .
@xmath147 for all @xmath148 .
+ indeed , if there exist @xmath149 and @xmath150 such that @xmath151 , then @xmath152 . since @xmath144
uniformly for @xmath153 $ ] , we have for sufficiently large @xmath92 that @xmath154 and @xmath155 , which is a contradiction .
+ if @xmath157 , then for any @xmath158 ,
@xmath159 @xmath160 @xmath161 . since @xmath144
uniformly for @xmath162 $ ] , there exists @xmath163 , @xmath164 for @xmath165 .
consequently , for those @xmath166 , @xmath167 , @xmath168 and @xmath169 , we have @xmath170 .
thus @xmath171 , which contradicts the fact that @xmath172 .
thus @xmath173 .
a similar argument can be applied to show @xmath174 . @xmath175 .
+ since @xmath172 , for every @xmath176 , there exists @xmath177 such that @xmath178 and @xmath179 for all @xmath180 . for @xmath114 , put @xmath181 and @xmath182 . then @xmath183 .
therefore , we can assume that @xmath184 for all @xmath176 .
consequently @xmath185 and @xmath186 , @xmath187 .
thus , @xmath188 , i.e. @xmath189 . @xmath116 .
+ notice that @xmath190 . choose @xmath191 satisfying @xmath192 and @xmath193 .
in order to show that such @xmath194 exists , put @xmath195 .
then , @xmath196 , @xmath197 and there exists @xmath198 such that @xmath199 for @xmath200 , which implies the existence of @xmath194 with the required properties .
suppose , to the contrary , that @xmath201 , then there exists @xmath202 such that @xmath203 @xmath204 and @xmath205 .
since @xmath144 uniformly for @xmath206 $ ] , there exists a sufficiently large @xmath92 , such that @xmath207 .
thus @xmath208 .
we need to consider the following two cases : * case 1 : * @xmath209 .
+ then @xmath210 , and we have @xmath211.\nonumber\end{aligned}\ ] ] define @xmath212 then @xmath213 and @xmath214\nonumber\\ & = & \frac{1}{2}|y_{n_0 + 1}^m|^2+\sum\limits_{n = n_0 + 1}^{\infty}\left[\frac{1}{2}|\triangle y^m_n|^2+a(1-\cos y^m_n)\right]\nonumber\\ & \leq & \frac{1}{2}|y_{n_0 + 1}^m|^2+j(y^m)-\frac{9\delta^2}{2}\nonumber\\ & < & j(y^m)-\frac{5\delta^2}{2},\nonumber\end{aligned}\ ] ] which leads to the following contradiction @xmath215 + * case 2 : * @xmath216 . + * subcase i : * @xmath217 for all @xmath218 .
+ then @xmath219\nonumber\end{aligned}\ ] ] define @xmath220 then @xmath221 and @xmath222\nonumber\\ & = & \frac{1}{2}|y_{n_0 + 1}^m|^2+\sum\limits_{n = n_0 + 1}^{\infty}\left[\frac{1}{2}|\triangle y^m_n|^2+a(1-\cos y^m_n)\right]\nonumber\\ & \leq & \frac{1}{2}|y_{n_0 + 1}^m|^2+j(y^m)-\sqrt{2a\alpha_{\delta}}\varepsilon/6\nonumber\\ & < & j(y^m)-\delta^2,\nonumber\end{aligned}\ ] ] which yields the following contradiction @xmath223 + * subcase ii : * there exists a @xmath224 $ ] such that @xmath225 $ ] .
+ if @xmath226 , by a similar argument as in case 1 , we get a contradiction . on the other hand ,
if @xmath227 , then by the argument used in subcase i of case 2 , we again obtain a contradiction .
consequently , @xmath228 , which completes the proof .
[ 5 ] for any @xmath229 , there exists @xmath230 such that @xmath231 , i.e. @xmath232 minimizes @xmath233 .
let @xmath234 be a minimizing sequence for ( [ ( 2.3 ) ] ) . there exists a positive number @xmath235 such that @xmath236 .
we claim that @xmath237 is a bounded sequence .
suppose to the contrary that for any @xmath238 there exists @xmath239 such that @xmath240 .
thus @xmath241 , and there exists @xmath242 such that @xmath243 when @xmath244 .
consider @xmath245 .
* case i : * if @xmath246 , then @xmath247 .
let @xmath248 , we have @xmath249 , which contradicts the assumptions .
* case ii : * otherwise , @xmath250 .
denote @xmath251\}$ ] .
then we have @xmath252 let @xmath248 in ( [ ( 2.5 ) ] ) , then @xmath253 .
but @xmath253 if and only if @xmath254 which is equivalent to @xmath255 , which contradicts again the assumptions .
consequently , @xmath256 is a bounded sequence and , by the definition of the norm on @xmath77 , @xmath257 is a bounded sequence in @xmath77 .
therefore , passing to a subsequence if necessary , there is @xmath258 such that @xmath259 weakly converges to @xmath96 in @xmath77 .
we claim @xmath260 . indeed ,
consider @xmath261 and define for @xmath114 @xmath262.\ ] ] the weak convergence of the sequence @xmath263 to @xmath96 in the hilbert space @xmath77 implies that @xmath264 for any @xmath265 .
then , @xmath266 converges uniformly to @xmath267 . clearly , @xmath268 is lower continuous , so it is also lower semi - continuous . combining @xmath269 with the lower semi - continuity of @xmath268 , we have @xmath270 since @xmath258 and @xmath271 are arbitrary , ( [ ( 2.6 ) ] ) implies @xmath272 .
lemma [ 4 ] implies @xmath273 , and we have @xmath274 .
put @xmath275 we will show that , fixed @xmath276 , there are finite @xmath277 s such that @xmath278 , @xmath279 is a critical value of @xmath135 restricted on the set @xmath280 .
[ 6 ] the set @xmath281 is finite .
consider @xmath282 and @xmath283 .
then @xmath175 , @xmath116 , @xmath147 .
put @xmath284 and @xmath285 . if @xmath286 , then by lemma [ 2 ] , @xmath287 notice that @xmath288 if and only if @xmath289 which is equivalent to @xmath290 .
thus @xmath291 as @xmath288 . in the case
@xmath292 , by a similar ( but even simpler ) argument , we obtain the same result .
consider @xmath293 .
then we have @xmath294 and there exists @xmath295 such that @xmath296 for all @xmath297 . consequently , there are only finitely many @xmath298 which can be equal to @xmath299 .
fixed @xmath8 , lemma [ 6 ] implies @xmath299 is achieved at some points @xmath300 .
now by choosing a sequence of @xmath301 , we claim that there exists a subsequence @xmath302 such that , for sufficiently large @xmath303 the points @xmath304 are independent of @xmath303 , i.e. we have the following : [ 7 ] suppose that @xmath305 is a decreasing sequence of positive numbers such that @xmath306 as @xmath307 .
then there exists a subsequence @xmath308 such that , for sufficiently large @xmath303 the points @xmath304 are independent of @xmath303 .
consider @xmath309 .
for any @xmath310 , we have @xmath311 and also @xmath312 for all @xmath265 .
thus @xmath313 and consequently @xmath314 .
by definition of @xmath315 , we have @xmath316 thus @xmath317 is monotone non - increasing bounded sequence . by a similar argument to the one used in the proof of lemma [ 6 ] , the sequence @xmath318 is bounded .
consequently , it contains a convergent subsequence @xmath319 .
since the set @xmath320 consists of isolated points , @xmath321 is a constant sequence for @xmath303 sufficiently large . since for sufficiently large @xmath303 the points @xmath321 are independent of @xmath303 , denote by @xmath322 . by lemma
[ 5 ] , there exists @xmath323 such that @xmath324 .
[ th1 ] for @xmath303 sufficiently large , @xmath325 is a heteroclinic solution joining @xmath326 and @xmath141 .
put @xmath327 . by the definition of @xmath328 and @xmath77 , it is sufficient to show that for large @xmath303 , @xmath329 for all @xmath265 . if not , there would exist a sequence @xmath330 and @xmath331 such that @xmath332 by similar argument used in the proof of lemma [ 6 ] , @xmath333 is bounded .
passing to a subsequence , if necessary , @xmath334 must be a constant sequence , i.e. @xmath335 .
we have the following two possibilities:*case 1 : * there is an increasing sequence of integers @xmath336 such that @xmath337 for all @xmath338 , or**case 2 : * * for every @xmath339 there is a @xmath340 such that @xmath341 .
if case 1 occurs , define @xmath342 then @xmath343 and @xmath344\\ & \hskip.5 cm -\frac{1}{2}|\triangle x_{n_{k^{\prime}}}(k^{\prime})|^2+a(1-\cos x_{n_{k^{\prime}}}(k^{\prime}))\\ & = \sum\limits_{n = n_{k^{\prime}}}^{\infty}\left[\frac{1}{2}|\triangle y_n(k^{\prime})|^2+a(1-\cos y_n(k^{\prime}))\right]-\frac{1}{2}(\varepsilon_{k^{\prime}})^2- a(1-\cos\varepsilon_{k^{\prime}}).\end{aligned}\ ] ] if there exists a @xmath345 such that @xmath346 , then @xmath347 .
otherwise , there exist two adjacent points such that the distance of them is larger than @xmath348 .
then we have @xmath349 .
define @xmath350 .
we have @xmath351 this is a contradiction .
if case 2 occurs , define @xmath352 then @xmath353 and @xmath354-\frac{1}{2}|\triangle z_{m_k}(k)|^2-a(1-\cos z_{m_{k}}(k))\nonumber\\ & = & \sum\limits_{n = m_k}^{\infty}\left[\frac{1}{2}|\triangle y_n(k)|^2+a(1-\cos y_n(k)\right]-\frac{1}{2}{\varepsilon_k}^2-a(1-\cos \varepsilon_k)\nonumber\end{aligned}\ ] ] by applying a similar argument as in the case 1 , we get again a contradiction . as we can see on figure [ fig : de ] ,
every heteroclinic solution join two adjacent points of the set @xmath355 , or , after translation , heteroclinic solution join two adjacent points of the set @xmath356 .
denote by @xmath357 the set of @xmath358 such that there exist a heteroclinic solution joining @xmath326 to @xmath141 .
the above observing gives @xmath359 , which will be proved strictly below .
since @xmath360 is @xmath361-periodic , we have @xmath362 .
this implies that , for any integer @xmath363 , if there exists a heteroclinic orbit joining @xmath364 and @xmath326 , there must exists a heteroclinic orbit joining @xmath326 and @xmath365 .
thus we need only to consider heteroclinic orbits joining @xmath326 to @xmath365 .
[ 8 ] @xmath359 . following the above argument
, we just consider heteroclinic solutions joining @xmath326 and @xmath365 , where @xmath366 is a positive integer .
suppose , to the contrary , theorem [ th1 ] implies that there exist @xmath367 where @xmath368 .
lemma [ 5 ] guarantees existence of @xmath369 which minimizes @xmath370 .
denote @xmath371 .
we have the following two cases : * case 1 : * there exists @xmath372 such that @xmath373 .
+ define @xmath374 then @xmath375 and @xmath376\nonumber\end{aligned}\ ] ] if @xmath377 .
otherwise , there exists at least a suffix @xmath378 $ ] . if @xmath379 $ ] such that @xmath380 , then we have @xmath381 .
otherwise , there must be two adjacent points such that the distant larger than @xmath348 .
and then @xmath382 .
define @xmath383 .
all those situations contrary with @xmath384 .
+ * case 2 : * if there is no @xmath372 such that @xmath373 , denote @xmath385 and two situation maybe meet : * subcase i : * if @xmath386 are two adjacent suffix .
+ define @xmath387 then @xmath375 and @xmath388 which implies the following contradiction @xmath389 * subcase ii : * @xmath390 . then
, we have @xmath391 . + define @xmath392 then @xmath375 and @xmath393-\frac{1}{2}|\triangle x_{n_2}|^2-a(1-\cos x_{n_2})\nonumber\\ & = & \sum\limits_{n = n_2 + 1}^{n_1 - 2}[\frac{1}{2}|\triangle y_n|^2+a(1-\cos
y_n)]+\frac{1}{2}|\triangle y_{n_2}|^2 -\frac{1}{2}|y_{n_1 - 1}-2(k-1)\pi - y_{n_1}|^2\nonumber\end{aligned}\ ] ] a similar argument as case 1 of theorem 1 induces a contradiction .
consequently , we finish our proof . for each @xmath119 , there exist at least two heteroclinic orbits joining @xmath394 to @xmath54 and at least two of heteroclinic orbits joining @xmath54 to @xmath395 . without loss generality ,
we only need to check heteroclinic orbits joining @xmath326 and @xmath396 .
lemma [ 8 ] implies that only @xmath397 and @xmath361 belong to @xmath357 . if @xmath398 is a heteroclinic orbit connecting @xmath326 and @xmath361 , then @xmath399 is also a heterclinic solution joining @xmath361 to @xmath326 . and
@xmath400 also two heteroclinic solutions joining @xmath397 to @xmath326 .
the proof is complete .
for simplicity , we paint phase plane of ( [ ( 1.2 ) ] ) with @xmath32 in section 1 .
we should paint phase plane of ( [ ( 1.1 ) ] ) with @xmath32 to compare with that of ( [ ( 1.2 ) ] ) . however , phase plane of ( [ ( 1.1 ) ] ) with @xmath32 ( figure [ fig : ade ] ) is so different from that of ( [ ( 1.2 ) ] ) .
non - periodic solutions move between the upper and lower half plane of ( [ ( 1.1 ) ] ) . at first glance , the phase plane of ( [ ( 1.1 ) ] ) is different from that of ( [ ( 1.2 ) ] ) in essence .
but it is not .
those phenomena appear because of approximation error .
approximation error depends on amplitude .
when amplitude @xmath401 equal to @xmath402 , we paint the phase plane of ( [ ( 1.1 ) ] ) as figure [ fig : lade ] . all periodic solutions and non - periodic solutions become disordered .
that is why we choose the phase plane of ( [ ( 1.1 ) ] ) with @xmath31 .
f. alessio , m.l.bertotti , p.montecchiari , multibump solutions to possibly degenerate equilibria for almost periodic lagrangian systems , z. angew .
phys . , 50 , ( 1999 ) , 860 - 891 .
m.l.bertotti , p.montecchiari , connecting orbits for some classes of almost periodic lagrangian systems , j. diff .
eq . , 145 , ( 1998 ) , 453 - 468 .
p.caldiroli , l.jeanjean , homoclinics and heteroclinics for a class of conservative singular hamiltonian systems , j. diff .
equa . , 136 , ( 1997 ) , 76 - 114 .
c.n.chen , s.y.tzeng , variational and penalization methods for studying connecting orbits of hamiltonian systems , electronic j. diff .
equa . , ( 2000 ) , 1 - 21 .
c.n.chen , s.y.tzeng , existence of multiplicity results for heteroclinic orbits of second order hamiltonian systems , j. diff .
, 158 , ( 1999 ) , 211 - 250 .
c.n.chen , s.y.tzeng , periodic solutions and their connecting orbits of hamiltonian systems , j. diff .
equa . , 177 , ( 2001 ) , 121 - 145 .
p.l.felmer , heteroclinic orbits for spatially periodic hamiltonian systems , analyse nonlineaire , 8 , ( 1991 ) , 477 - 497 . z.m.guo , j.s.yu , existence of periodic and subharmonic solutions for second - order superlinear difference equations , sci.china a , 46 , ( 2003 ) , 506 - 515 .
m.izydorek , j.janczewska , heteroclinic solutions for a class of the second order hamiltonian systems , j. diff .
equa . , 238 , ( 2007 ) , 381 - 393 .
t.o.maxwell , heteroclinic chains for a reversible hamiltonia system , non .
, 28 , ( 1997 ) , 871 - 887 .
r.h.rabinowitz , heteroclinics for a reversible hamiltonian system ii , diff .
, 14 , ( 1994 ) , 817 - 829 .
r.h.rabinowitz , heteroclinics for a reversible hamiltonian system , ergodic theory and dynamical systems , 14 , ( 1994 ) , 817 - 829 .
r.h.rabinowitz , on a theorem of strobel , calc .
, 12 , ( 2001 ) , 399 - 415 .
r.h.rabinowitz , periodic and heteroclinic orbits for a periodic hamiltonian system , analyse nonlineaire , 6 , ( 1989 ) , 331 - 346 .
r.h.rabinowitz , v.c.zelati , multichain - type solutions for hamiltonian systems , electronic j. diff .
equa . , ( 2000 ) , 223 - 235 .
k.strobel , multibump solutions for a class of periodic hamiltonian systems , university of wisconsin thesis , ( 1994 ) .
v.c.zelati , r.h.rabinowitz , heteroclinic solutions between stationary points at different energy levels , topol .
, 17 , ( 2001 ) , 1 - 21 . | about twenty years ago , rabinowitz showed firstly that there exist heteroclinic orbits of autonomous hamiltonian system joining two equilibria .
a special case of autonomous hamiltonian system is the classical pendulum equation .
the phase plane analysis of pendulum equation shows the existence of heteroclinic orbits joining two equilibria , which coincide with the result of rabinowitz .
however , the phase plane of discrete pendulum equation is similar to that of the classical pendulum equation , which suggests the existence of heteroclinic orbits for discrete pendulum equation also . by using variational method and delicate analysis technique ,
we show that there indeed exist heteroclinic orbits of discrete pendulum equation joining every two adjacent points of @xmath0 . * key words and phrases : * heteroclinic solution , critical point , discrete pendulum equation , minimization arguments . this project is supported by national natural science foundation of china ( no .
10625104 ) and the research fund for the doctoral program of higher education of china ( no .
20061078002 ) . ]
corresponding author .
e - mail : [email protected] . ] |
here we prove that for any entanglement monotone @xmath3 the coherence quantifier _
@xmath112\right)\right\ } \label{eq : ce-1}\ ] ] _ does not increase on average under ( selective ) incoherent operations : @xmath113 with probabilities @xmath114 $ ] , quantum states @xmath115 , and incoherent kraus operators @xmath116 acting on the system @xmath1 .
due to the definition of @xmath4 , the amount of entanglement between the system and ancilla can not exceed @xmath4 for any incoherent operation @xmath54 , i.e. , @xmath117 @xmath118\right)\leq c_{e}\left(\rho^{s}\right).\ ] ] note that this statement is also true if we introduce another particle @xmath119 in an incoherent state @xmath120 . then
, for any tripartite incoherent operation @xmath121 it holds : @xmath122\right)\leq c_{e}\left(\rho^{s}\right).\label{eq : proof-3}\ ] ] we will now prove the claim by contradiction , showing that a violation of eq .
( [ eq : monotone ] ) also implies a violation of eq .
( [ eq : proof-3 ] ) . if eq .
( [ eq : monotone ] ) is violated , then by definition of @xmath4 there exists a set of incoherent operations @xmath123 such that the following inequality is true for @xmath88 large enough : @xmath124\right)>c_{e}\left(\rho^{s}\right).\label{eq : proof}\ ] ] in the next step we introduce an additional particle @xmath119 and use the general relation @xmath125 which is valid for any entanglement monotone @xmath3 . with this in mind , the inequality ( [ eq : proof ] ) implies @xmath126\otimes\ket{i}\bra{i}^{b}\right)>c_{e}\left(\rho^{s}\right).\ ] ] recall that the states @xmath127 are obtained from the state @xmath51 by the means of an incoherent operation , and thus we can use the relation @xmath128 with incoherent kraus operators @xmath116 .
this leads us to the following expression : @xmath129\otimes\ket{i}\bra{i}^{b}\right)>c_{e}\left(\rho^{s}\right).\label{eq : proof-1}\ ] ] it is now crucial to note that the state on the left - hand side of the above expression can be regarded as arising from a tripartite incoherent operation @xmath121 acting on the total state @xmath130 : @xmath131\nonumber \\ & = \sum_{i}\lambda_{i}^{sa}\left[k_{i}\rho^{s}k_{i}^{\dagger}\otimes\ket{0}\bra{0}^{a}\right]\otimes\ket{i}\bra{i}^{b}.\label{eq : proof-2}\end{aligned}\ ] ] this can be seen explicitly by introducing the kraus operators @xmath132 corresponding to the operation @xmath121 : @xmath133 here , @xmath134 are incoherent kraus operators corresponding to the incoherent operation @xmath123 : @xmath135=\sum_{j}l_{ij}\rho^{sa}l_{ij}^{\dagger}.\ ] ] the unitaries @xmath136 are incoherent and defined as @xmath137 with these definitions we see that @xmath132 are indeed incoherent kraus operators .
moreover , it can be verified by inspection that the incoherent operation @xmath121 arising from these kraus operators also satisfies eq .
( [ eq : proof-2 ] ) . finally , using eq .
( [ eq : proof-2 ] ) in eq .
( [ eq : proof-1 ] ) we arrive at the following inequality : @xmath122\right)>c_{e}\left(\rho^{s}\right).\ ] ] this is the desired contradiction to eq .
( [ eq : proof-3 ] ) , and completes the proof of property ( c3 ) for @xmath4 , thus establishing that @xmath4 is a coherence monotone for any entanglement monotone @xmath3 .
here we show that the quantifier of coherence @xmath4 given in eq .
( [ eq : ce-1 ] ) is convex for any convex entanglement measure @xmath3 : @xmath138 for any quantum states @xmath139 and probabilities @xmath9 . for this , note that by convexity of the entanglement quantifier @xmath3 it follows : @xmath140\right)\nonumber \\ & \leq\sum_{i}p_{i}e^{s : a}\left(\lambda^{sa}\left[\rho_{i}^{s}\otimes\ket{0}\bra{0}^{a}\right]\right).\end{aligned}\ ] ] taking the supremum over all incoherent operations @xmath54 together with the limit @xmath141 on both sides of this inequality we obtain the following result : @xmath142\right)\right\}\ ] ] finally , note that the right - hand side of this inequality can not decrease if the supremum over incoherent operations @xmath54 and the limit @xmath141 are performed on each term of the sum individually : @xmath143\right)\right\ } \nonumber \\ & \leq\sum_{i}p_{i}\lim_{d_{a}\rightarrow\infty}\sup_{\lambda^{sa}}e^{s : a}\left(\lambda^{sa}\left[\rho_{i}^{s}\otimes\ket{0}\bra{0}^{a}\right]\right)\\ & = \sum_{i}p_{i}c_{e}\left(\rho_{i}^{s}\right).\nonumber\end{aligned}\ ] ] together with eq .
( [ eq : convexity_proof ] ) , this completes the proof of convexity in eq .
( [ eq : convexity ] ) .
we will now show that the bound provided in theorem 1 of the main text can be saturated for the geometric entanglement @xmath91 and geometric coherence @xmath102 .
the former is defined as @xmath144 where the maximum is taken over all separable states @xmath60 , and @xmath145 is the fidelity .
the geometric entanglement defined in eq .
( [ eq : distance ] ) coincides with its expression obtained via a convex roof construction @xcite : @xmath146 where the minimum is taken over all decompositions of the state @xmath96 .
the latter expression ( [ eq : convex ] ) is the original definition of the geometric entanglement for mixed states @xcite , and the equivalence of eqs .
( [ eq : distance ] ) and ( [ eq : convex ] ) was shown in @xcite .
equipped with these tools , we are now in position to prove the saturation of the bound in theorem 1 for these measures of entanglement and coherence
. in particular , we will show the existence of an incoherent operation @xmath54 such that @xmath148\right)=c_{\mathrm{g}}\left(\rho^{s}\right)\label{eq : main}\ ] ] if the dimension of the ancilla is not smaller than the dimension of the system , @xmath70 . as we will further show , the optimal incoherent operation achieving this equality is the generalized cnot operation @xmath149 given in eq .
( [ eq : cnot ] ) of the main text . to prove eq .
( [ eq : main ] ) , we first recall that the final state after the application of the generalized cnot operation @xmath149 is @xmath150 where @xmath77 are matrix elements of @xmath78 . states of the form ( [ eq : mc ] ) are known as maximally correlated states . as shown in section [ sec:2 ] of this supplemental material
, there always exists a separable maximally correlated state @xmath151 which is a closest separable state to @xmath152 : @xmath153 these results imply that the geometric coherence of @xmath51 is bounded above by the geometric entanglement of @xmath152 : @xmath154 this follows by using eq .
( [ eq : eg-2 ] ) together with the equality @xmath155 where @xmath156 is an incoherent state with the same coefficients @xmath157 as in eq .
( [ eq : sigma ] ) . on the other hand , theorem 1 in the main text implies the inequality @xmath158 and thus we arrive at the desired statement in eq .
( [ eq : main ] ) .
we further note that the arguments presented above can also be applied to any distance - based quantifiers of entanglement @xmath57 and coherence @xmath58 if the distance is contractive under quantum operations , and for any maximally correlated state @xmath152 there exists a separable maximally correlated state @xmath159 ( which may depend on @xmath152 ) such that @xmath160 in particular , this is the case for the geometric entanglement and coherence , where the distance is given by @xmath161 . as discussed in the main text , these results can be immediately extended to any distance @xmath162\ ] ] which is a nonincreasing function of fidelity . in this section
we will show that for any maximally correlated state @xmath163 there exists a separable maximally correlated state @xmath164 such that @xmath165 this can be proven by using results from refs .
@xcite . in particular , given a maximally correlated state @xmath166 , consider its arbitrary decomposition into pure states @xmath167 with positive probabilities @xmath168 such that @xmath169 as is proven on page 6 in @xcite , all states @xmath167 in such a decomposition must be linear combinations of product states @xmath170 : @xmath171 with complex coefficients @xmath172 .
consider now an optimal decomposition of the state @xmath166 , i.e. , a decomposition which minimizes the average entanglement such that @xmath173 we further define product states @xmath174 to be the closest product states to @xmath167 : @xmath175 due to eq .
( [ eq : psi ] ) , all states @xmath176 can be chosen as @xmath177 , where the number @xmath178 corresponds to the coefficient @xmath179 of the state @xmath167 with the maximal absolute value : @xmath180 .
finally , consider the separable maximally correlated state @xmath181 with probabilities @xmath182 defined as @xmath183}{1-e_{\mathrm{g}}(\rho_{\mathrm{mc}})}.\label{eq : q}\ ] ] as we now show , @xmath184 is the desired optimal state , satisfying the equality ( [ eq : eg ] ) .
this can be seen by first recalling that the geometric entanglement of the state @xmath166 is bounded above as @xmath185 since the state @xmath184 is separable . on the other hand ,
the square root of the fidelity @xmath186 satisfies the strong concavity relation @xmath187 for any two states @xmath188 and @xmath189 ; see theorem 9.7 on page 414 in @xcite ( note that the fidelity defined there is the square root of the fidelity used in our paper ) . applied to the maximally correlated states @xmath166 and @xmath184 , and using eqs .
( [ eq : eg-1 ] ) and ( [ eq : q ] ) , this inequality becomes @xmath190 implying that the geometric entanglement of @xmath166 is bounded below as @xmath191 combining this result with eq .
( [ eq : bound-1 ] ) completes the proof of eq .
( [ eq : eg ] ) .
earlier we have proven that the optimal incoherent operation which attains the maximization in eq .
( [ eq : ce-1 ] ) is the generalized cnot when @xmath3 is the geometric measure of entanglement @xmath91 .
if the system @xmath1 is a single qubit ( @xmath192 ) , the output state of system and ancilla after the cnot is a two - qubit state .
the geometric entanglement @xmath91 of any bipartite two - qubit state @xmath193 is computable in closed form and given by @xcite @xmath194\,,\end{aligned}\ ] ] where @xmath195 is the concurrence of @xmath193 @xcite .
let @xmath196 be an arbitrary state of the single qubit @xmath1 , written with respect to a reference basis @xmath7 .
after applying the cnot on the above state and the initially incoherent ancilla , we get a maximally correlated two - qubit state @xmath197 = \sum_{i , j=0}^{1}\rho_{ij}\ket{i}\bra{j}^{s}\otimes \ket{i}\bra{j}^{a}.\end{aligned}\ ] ] the concurrence of the above maximally correlated state is easily evaluated as @xcite @xmath198 the geometric entanglement of @xmath152 can be written then as @xmath199\nonumber\\ & = \frac{1}{2}\left[1- \sqrt{1- 4|\rho_{01}|^2}\right].\end{aligned}\ ] ] finally , the geometric coherence for an arbitrary single - qubit state @xmath31 is given by the expression reported in the main text : @xmath200.\end{aligned}\ ] ] | quantum coherence is an essential ingredient in quantum information processing and plays a central role in emergent fields such as nanoscale thermodynamics and quantum biology .
however , our understanding and quantitative characterization of coherence as an operational resource are still very limited . here
we show that any degree of coherence with respect to some reference basis can be converted to entanglement via incoherent operations .
this finding allows us to define a novel general class of measures of coherence for a quantum system of arbitrary dimension , in terms of the maximum bipartite entanglement that can be generated via incoherent operations applied to the system and an incoherent ancilla .
the resulting measures are proven to be valid coherence monotones satisfying all the requirements dictated by the resource theory of quantum coherence .
we demonstrate the usefulness of our approach by proving that the fidelity - based geometric measure of coherence is a full convex coherence monotone , and deriving a closed formula for it on arbitrary single - qubit states .
our work provides a clear quantitative and operational connection between coherence and entanglement , two landmark manifestations of quantum theory and both key enablers for quantum technologies . *
_ introduction_.*coherence is a fundamental aspect of quantum physics that encapsulates the defining features of the theory @xcite , from the superposition principle to quantum correlations .
it is a key component in various quantum information and estimation protocols and is primarily accountable for the advantage offered by quantum tasks versus classical ones @xcite . in general
, coherence is an important physical resource in low - temperature thermodynamics @xcite , for exciton and electron transport in biomolecular networks @xcite , and for applications in nanoscale physics @xcite .
experimental detection of coherence in living complexes @xcite and creation of coherence in hot systems @xcite have also been reported . while the theory of quantum coherence is historically well developed in quantum optics @xcite , a rigorous framework to quantify
coherence for general states in information theoretic terms has only been attempted recently @xcite .
this framework is based on identifying the set of incoherent states and a class of ` free ' operations , named incoherent quantum channels , that map the set onto itself @xcite .
the resulting resource theory of coherence is in direct analogy with the resource theory of entanglement @xcite , in which local operations and classical communication are the ` free ' operations that map the set of separable states onto itself @xcite .
within such a framework for coherence , one can define suitable measures that vanish for any incoherent state , and satisfy specific monotonicity requirements under incoherent channels .
measures that respect these conditions gain the attribute of coherence monotones , in analogy with entanglement monotones @xcite .
examples of coherence monotones include the relative entropy and the @xmath0-norm of coherence @xcite . intuitively , both coherence and entanglement capture quantumness of a physical system , and it is well known that entanglement stems from the superposition principle , which is also the essence of coherence .
it is then legitimate to ask how can one resource emerge _
quantitatively _ from the other @xcite . can be converted to entanglement via incoherent operations on @xmath1 and an incoherent ancilla @xmath2 .
input coherence and output entanglement are quantitatively equivalent : for any entanglement monotone @xmath3 , the maximum entanglement generated between @xmath1 and @xmath2 by incoherent operations defines a faithful coherence monotone @xmath4 on the initial state of @xmath1.,width=321 ] in this letter , we provide a mathematically rigorous approach to resolve the above question , using a common frame to quantify quantumness in terms of coherence and entanglement .
in particular , we show that any nonzero amount of coherence in a system @xmath1 can be converted to ( distillable ) entanglement between @xmath1 and an initially incoherent ancilla @xmath2 , by means of incoherent operations .
this allows us to formulate a novel , general method to quantify coherence in terms of entanglement ( see fig . [ cofig ] ) .
namely we prove that , given an arbitrary set of entanglement monotones @xmath5 , one can define a corresponding class of coherence monotones @xmath6 that satisfy all the requirements of ref .
@xcite . the input coherence @xmath4 of @xmath1 is defined as the maximum output entanglement @xmath3 over all incoherent operations on @xmath1 and @xmath2 .
we explicitly evaluate the maximization in some relevant instances , defining novel coherence monotones such as the fidelity - based geometric measure of coherence .
these results provide powerful advances for the operational quantification of coherence .
* _ characterizing coherence_.*for an arbitrary fixed reference basis @xmath7 , the incoherent states are defined as @xcite @xmath8 where @xmath9 are probabilities . any state which can not be written as above
is defined _ coherent _ @xcite .
note that , unlike other resources in information theory , coherence is basis - dependent .
the reference basis with respect to which coherence is measured depends on the physical problem under investigation ; it is e.g. the energy basis for transport phenomena in engineered and biological domains @xcite , or the eigenbasis of the generator of an unknown phase shift in quantum metrology @xcite .
a completely positive trace preserving map @xmath10 is said to be an incoherent operation if it can be written as @xmath11={\sum}_{l}k_{l}\rho k_{l}^{\dagger},\label{eq : incoherent_operation}\ ] ] where the defining operators @xmath12 , called incoherent kraus operators , map every incoherent state to some other incoherent state , i.e. @xmath13 , where @xmath14 is the set of incoherent states . following established notions from entanglement theory @xcite , baumgratz _ et al_. proposed the following postulates for a measure of coherence @xmath15 in ref .
@xcite : * ( c1 ) @xmath16 , and @xmath17 if and only if @xmath18 .
+ * ( c2 ) @xmath15 is nonincreasing under incoherent operations , i.e. , @xmath19)$ ] with @xmath20\subseteq{\cal { i}}}$ ] . + * ( c3 ) @xmath15 is nonincreasing on average under selective incoherent operations , i.e. , @xmath21 , with probabilities @xmath22 $ ] , states @xmath23 , and incoherent kraus operators @xmath12 obeying @xmath24 . + * ( c4 ) @xmath15 is a convex function of density matrices , i.e. , @xmath25 . note that conditions ( c3 ) and ( c4 ) automatically imply condition ( c2 ) .
the reason we listed all conditions above is that ( similar to entanglement measures ) there might exist meaningful quantifiers of coherence which satisfy conditions ( c1 ) and ( c2 ) , but for which conditions ( c3 ) and ( c4 ) are either violated or can not be proven . following the analogous notion from entanglement theory
, we call a quantity which satisfies conditions ( c1 ) , ( c2 ) , and ( c3 ) a _ coherence monotone_. examples of functionals that satisfy all the four properties mentioned above include the @xmath0-norm of coherence @xcite @xmath26 and the relative entropy of coherence @xcite @xmath27 with the quantum relative entropy @xmath28-\mathrm{tr}[\rho\log_{2}\varsigma]$ ] . as was shown in @xcite , the relative entropy of coherence
can also be written as @xmath29 , where @xmath30 is the diagonal part of the density matrix @xmath31 in the reference basis @xmath7 and @xmath32 is the von neumann entropy . * _ bipartite coherence_.*we first extend the framework of coherence to the bipartite scenario ( see also @xcite ) ; the following definitions extend straightforwardly to multipartite systems .
in particular , for a bipartite system with two subsystems @xmath33 and @xmath34 , and with respect to a fixed reference product basis @xmath35 , we define bipartite incoherent states as follows : @xmath36 here , @xmath37 are probabilities and the states @xmath38 and @xmath39 are incoherent states on the subsystem @xmath33 and @xmath34 respectively , i.e. @xmath40 and @xmath41 for probabilities @xmath42 and @xmath43 .
note that the states in eq .
( [ eq : incoherent_state ] ) are always separable .
we next define bipartite incoherent operations as in eq .
( [ eq : incoherent_operation ] ) , with incoherent kraus operators @xmath12 such that @xmath44 , where @xmath45 is now the set of bipartite incoherent states defined in eq .
( [ eq : incoherent_state ] ) .
an example of bipartite incoherent operation is the two - qubit cnot gate @xmath46 .
it is not possible to create coherence from an incoherent two - qubit state by using the cnot gate , since it takes any pure incoherent state @xmath47 to another pure incoherent state , @xmath48 .
the cnot gate can be used however to create entanglement , e.g. it is well known that the state @xmath49 is entangled for any coherent state @xmath50 @xcite . * _ converting coherence to entanglement_.*referring to fig . [ cofig ] , we say that the coherence in the initial state @xmath51 of a ( finite - dimensional ) system @xmath1 can be converted to entanglement via incoherent operations if , by attaching an ancilla @xmath2 initialized in a reference incoherent state @xmath52 , the final system - ancilla state @xmath53 $ ] is entangled for some incoherent operation @xmath54 . note that incoherent system states @xmath55
can not be used for conversion to entangled states in this way , since for any incoherent state @xmath55 the state @xmath56 $ ] will be of the form given in eq .
( [ eq : incoherent_state ] ) , and thus separable .
entanglement can instead be generated by incoherent operations if the initial @xmath51 is coherent , as in the two - qubit cnot example .
it is then natural to ask : can _ any _ nonzero amount of coherence be converted to entanglement via incoherent operations ? to answer this , we first consider distance - based measures of entanglement @xmath57 and coherence @xmath58 @xcite : @xmath59 here , @xmath60 is the set of separable states and @xmath45 is the set of incoherent states .
moreover , we demand that the distance @xmath61 be contractive under quantum operations , @xmath62,\lambda[\varsigma])\leq d(\rho,\varsigma)\label{eq : contractive}\ ] ] for any completely positive trace preserving map @xmath10 .
this implies that @xmath57 does not increase under local operations and classical communication @xcite , and @xmath58 does not increase under incoherent operations @xcite .
equipped with these tools we are now in position to present the first result of this letter .
[ thm:1 ] for any contractive distance @xmath61 , the amount of ( distance - based ) entanglement @xmath57 generated from a state @xmath51 via an incoherent operation @xmath54 is bounded above by its ( distance - based ) coherence @xmath58 : @xmath63\right)\leq c_{d}\left(\rho^{s}\right)}}.\label{eq : bound}\ ] ] let @xmath55 be the closest incoherent state to @xmath51 , i.e. , @xmath64 .
the contractivity of the distance @xmath61 further implies the equality : @xmath65 . in the final step , note that the application of an incoherent operation @xmath54 to the incoherent state @xmath66 brings it to another incoherent and thus separable state .
applying eq .
( [ eq : contractive ] ) and combining the aforementioned results we arrive at the desired inequality : @xmath67,\lambda^{sa}[\sigma^{s}\otimes\ket{0}\bra{0}^{a}])\geq e_{d}^{s : a}(\lambda^{sa}[\rho^{s}\otimes\ket{0}\bra{0}^{a}])$ ] .
this result provides a strong link between the resource frameworks of entanglement and coherence .
an even stronger link exists when choosing specifically @xmath61 as the relative entropy .
the corresponding quantifiers are the relative entropy of entanglement @xmath68 @xcite , and the relative entropy of coherence @xmath69 @xcite introduced in eq .
( [ eq : cr ] ) .
importantly , the inequality ( [ eq : bound ] ) can be saturated for these measures if the dimension of the ancilla is not smaller than that of the system , @xmath70 . in this case
there always exists an incoherent operation @xmath54 such that @xmath71\right)=c_{\mathrm{r}}\left(\rho^{s}\right)}}.\label{eq : optimal}\ ] ] to prove this statement , we consider the unitary operation @xmath72 note that for two qubits this unitary is equivalent to the cnot gate with @xmath1 as the control qubit and @xmath2 as the target qubit
. it can be seen by inspection that this unitary is incoherent ( i.e. , the state @xmath73=u\rho^{sa}u^{\dagger}$ ] is incoherent for any incoherent state @xmath74 ) , and maps the state @xmath75 to the state @xmath76={\sum}_{i , j}\ \rho_{ij}\ket{i}\bra{j}^{s}\otimes\ket{i}\bra{j}^{a}}},\ ] ] where @xmath77 are the matrix elements of @xmath78 . in the next step
we use the fact that for any quantum state @xmath79 the relative entropy of entanglement is bounded below as follows @xcite : @xmath80 . applied to the state @xmath53 $ ]
, this inequality reduces to @xmath71\right)\geq h\left({\sum}_{i}\ \rho_{ii}\ket{i}\bra{i}^{s}\right)-h(\rho^{s})}}.\ ] ] noting that the right - hand side of this inequality is equal to the relative entropy of coherence @xmath81 @xcite , we obtain @xmath82)\geq c_{\mathrm{r}}(\rho^{s})$ ] .
the proof of eq .
( [ eq : optimal ] ) is complete by combining this result with theorem [ thm:1 ] .
the results presented above also hold for the distillable entanglement @xmath83 .
namely , the relative entropy of coherence @xmath69 also serves as an upper bound for the conversion to distillable entanglement via incoherent operations , and the equality in eq .
( [ eq : optimal ] ) still holds if @xmath68 is replaced by @xmath83 , and the incoherent unitary of eq .
( [ eq : cnot ] ) is applied .
this follows from theorem [ thm:1 ] , together with the fact that distillable entanglement admits the following bounds @xcite : @xmath84 .
this shows that the degree of ( relative entropy of ) coherence in the initial state of @xmath1 can be exactly converted to an equal degree of ( distillable or relative entropy of ) entanglement between @xmath1 and the incoherent ancilla @xmath2 by a suitable incoherent operation , that is a generalized cnot gate .
we can now settle the general question posed above .
[ thm:2 ] a state @xmath51 can be converted to an entangled state via incoherent operations if and only if @xmath51 is coherent . if @xmath51 is incoherent , it can not be converted to an entangled state via incoherent operations by theorem [ thm:1 ] .
conversely , if @xmath51 is coherent , it has nonzero relative entropy of coherence @xmath85 . by eq .
( [ eq : optimal ] ) , there exists an incoherent operation @xmath54 leading to nonzero relative entropy of entanglement @xmath82)>0 $ ] , concluding the proof . *
_ quantifying coherence with entanglement_.*we are ready to present the central result of the letter . reversing the perspective , theorem [ thm:1 ]
can also be seen as providing a lower bound on distance - based measures of coherence through conversion to entanglement : precisely , the coherence degree @xmath58 of a state @xmath51 is always bounded below by the maximal entanglement degree @xmath57 generated from it by incoherent operations .
going now beyond the specific setting of distance - based measures , we will show that such a maximization of the output entanglement , for any fully general entanglement monotone , leads to a quantity which yields a valid quantifier of input coherence in its own right .
we specifically define the family of _ entanglement - based coherence measures _ @xmath6 as follows : @xmath86\right)\bigg\ } .\label{eq : ce}\ ] ] here , @xmath3 is an arbitrary entanglement measure and the supremum is taken over all incoherent operations @xmath54 . in eq .
( [ eq : ce ] ) is well defined , since the supremum @xmath87\right)$ ] can not decrease with increasing dimension @xmath88 of the ancilla . ]
it is crucial to benchmark the validity of @xmath4 for any @xmath3 as a proper measure of coherence .
remarkably , we find that @xmath4 satisfies all the aforementioned conditions ( c1)(c3 ) given an arbitrary entanglement monotone @xmath3 , with the addition of ( c4 ) if @xmath3 is convex as well .
we namely get the following result : [ thm : monotone]@xmath4 is a ( convex ) coherence monotone for any ( convex ) entanglement monotone @xmath3 . using the arguments presented above it is easy to see that @xmath4 is nonnegative , and zero if and only if the state @xmath51 is incoherent .
moreover , @xmath4 does not increase under incoherent operations @xmath89 performed on the system @xmath1 .
this can be seen directly from the definition of @xmath4 in eq .
( [ eq : ce ] ) , noting that an incoherent operation @xmath89 on the system @xmath1 is also incoherent with respect to @xmath90 . the proof that @xmath4 further satisfies condition ( c3 ) is presented in the supplemental material @xcite .
there we also show that @xmath4 is convex for any convex entanglement monotone @xmath3 , i.e. ( c4 ) is fulfilled as well in this case .
these powerful results complete the parallel between coherence and entanglement , _ de facto _ establishing their full quantitative equivalence within the respective resource theories .
thanks to theorem [ thm : monotone ] , one can now use the comprehensive knowledge acquired in entanglement theory in the last two decades @xcite , to address the quantification of coherence in a variety of operational settings , and to define and validate physically motivated coherence monotones .
for instance , @xmath4 as defined by eq .
( [ eq : ce ] ) amounts to the previously defined relative entropy of coherence @xcite , if @xmath3 is the relative entropy of entanglement or the distillable entanglement .
furthermore , we can now focus on the relevant case of @xmath3 being the geometric entanglement @xcite @xmath91 , defined for a bipartite state @xmath31 as @xmath92 , where the maximum is taken over all separable states @xmath93 , and @xmath94 is the uhlmann fidelity .
the geometric entanglement coincides with its expression obtained via convex roof @xcite , @xmath95 , where the minimum is over all decompositions of @xmath96 . in the supplemental material @xcite
, we show that the _ geometric measure of coherence _
@xmath97 , associated to @xmath91 via eq . (
[ eq : ce ] ) , can be evaluated explicitly and amounts to @xmath98 , where the maximum is taken over all incoherent states @xmath99 . the incoherent operation which attains the maximum in eq .
( [ eq : ce ] ) is again the generalized cnot defined by eq .
( [ eq : cnot ] ) . due to theorem [ thm : monotone ] ,
since the geometric measure of entanglement is a full convex entanglement monotone @xcite , we have just proven that the geometric measure of coherence @xmath97 is a full convex coherence monotone obeying ( c1)(c4 ) .
this settles an important question left open in previous literature @xcite .
remarkably , the geometric measure @xmath97 is also analytically computable for an arbitrary state @xmath31 of one qubit @xcite , as follows : @xmath100 where @xmath101 is the off - diagonal element of @xmath31 with respect to the reference basis .
notice that @xmath102 in this case is a simple monoto - nic function of the @xmath0-norm of coherence @xcite , @xmath103 .
some of these results extend to any distance - based entanglement measure @xmath104 defined via eq .
( [ ecd ] ) with @xmath105 , where @xmath106 is a nonincreasing function of the fidelity @xmath107 .
these include the bures measure of entanglement @xcite , with @xmath108 , and the groverian measure of entanglement @xcite , with @xmath109 . for any such entanglement @xmath104 ,
the corresponding quantifier of coherence is @xmath110 @xcite , and theorem [ thm:1 ] holds with equality for any matching pair @xmath104 and @xmath111 .. ] * _ conclusions_.*in this letter we have established a rigorous and general framework for the interconversion between two quantum resources , coherence on one hand , and entanglement on the other hand , via incoherent operations .
our framework can be interpreted in both ways : on one hand , it demonstrates the formal potential of coherence for entanglement generation ( although not necessarily useful in practical applications , as cheaper schemes for entanglement creation might be available ) ; on the other hand , it demonstrates the usefulness of entanglement to obtain and validate measures of coherence . building on this connection , we proposed in fact a family of coherence quantifiers in terms of the maximal entanglement that can be generated by incoherent operations ( see fig .
[ cofig ] ) .
the proposed coherence quantifiers satisfy all the necessary criteria to be _ bona fide _ coherence monotones @xcite . in particular ,
the relative entropy of coherence and the geometric measure of coherence have been ( re)defined and interpreted operationally in terms of the maximum converted distillable and geometric entanglement , respectively .
our framework bears some resemblance with , and may be regarded as the general finite - dimensional counterpart to , the established ( qualitative and quantitative ) equivalence between input nonclassicality , intended as superposition of optical coherent states , and output entanglement created by passive quantum optical elements such as beam splitters @xcite .
the results presented here should also be compared to the scheme for activating distillable entanglement via premeasurement interactions @xcite from quantum discord , a measure of nonclassical correlations going beyond entanglement @xcite . in the latter approach , which has attracted a large amount of attention recently @xcite , measures of discord in a composite system
are defined in terms of the minimum entanglement created with an ancillary system via fixed premeasurement interactions defined as in eq .
( [ eq : cnot ] ) , where the minimization is over local unitaries on the system regulating the control bases before the interaction .
by contrast , in this work the reference basis is fixed , and a maximization of the output entanglement over all incoherent operations returns a measure of coherence for the initial system .
one might combine the two approaches in order to define a unified framework for interconversion among coherence , discord , and entanglement , whereby discord - type measures could be interpreted as measures of bipartite coherence suitably minimized over local product reference bases ( see e.g. @xcite ) .
exploring these connections further will be the subject of another work .
the theory of entanglement has been the cornerstone of major developments in quantum information theory and has triggered the advancement of modern quantum technologies .
the construction of a physically meaningful and mathematically rigorous quantitative theory of coherence can improve our perception of genuine quantumness , and guide our understanding of nascent fields such as quantum biology and nanoscale thermodynamics . by uncovering a powerful operational connection between coherence and entanglement
, we believe the present work delivers a substantial step in this direction . * _ acknowledgments_.*we thank m. ahmadi , j. asboth , r. augusiak , t. bromley , m. cianciaruso , w. vogel , a. winter , and especially m. piani for fruitful discussions .
a.s .
acknowledges financial support by the alexander von humboldt - foundation , the john templeton foundation , the eu ( ip siqs ) , the erc adg osyris , and the eu - spanish ministry ( chistera diqip ) .
u.s .
, h.s.d . , and
m.n.b .
acknowledge the department of atomic energy , government of india for research fellowship . g.a .
acknowledges financial support by the foundational questions institute ( grant no .
fqxi - rfp3 - 1317 ) and the erc stg gqcop ( grant agreement no
. 637352 ) .
60ifxundefined [ 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:\doibase10.1143/ptps.69.80 [ * * , ( ) ] link:\doibase10.1038/nphoton.2011.35 [ * * , ( ) ] http://dx.doi.org/10.1017/cbo9780511976667[__ ] ( , ) link:\doibase10.1103/physrevlett.113.150402 [ * * , ( ) ] @noop @noop link:\doibase10.1038/ncomms7383 [ * * , ( ) ] link:\doibase10.1103/physrevx.5.021001 [ * * , ( ) ] link:\doibase10.1088/1367 - 2630/10/11/113019 [ * * , ( ) ] link:\doibase10.1021/jp901724d [ * * , ( ) ] http://stacks.iop.org/1742-6596/302/i=1/a=012037 [ * * , ( ) ] link:\doibase10.1038/srep00885 [ * * , ( ) ] link:\doibase10.1080/00405000.2013.829687 [ * * , ( ) ] http://stacks.iop.org/1367-2630/16/i=3/a=033007 [ * * , ( ) ] link:\doibase10.1038/nnano.2012.147 [ * * , ( ) ] link:\doibase10.1103/physrevb.84.113415 [ * * , ( ) ] link:\doibase10.1038/nature05678 [ * * , ( ) ] link:\doibase10.1038/nature08811 [ * * , ( ) ] link:\doibase10.1103/physrevlett.107.273001 [ * * , ( ) ] link:\doibase10.1103/physrev.131.2766 [ * * , ( ) ] link:\doibase10.1103/physrevlett.10.277 [ * * , ( ) ] http://dx.doi.org/10.1017/cbo9781139644105[__ ] ( , ) link:\doibase10.1103/physreva.65.032323 [ * * , ( ) ] link:\doibase10.1103/physrevlett.94.173602 [ * * , ( ) ] link:\doibase10.1103/physrevlett.89.283601 [ * * , ( ) ] link:\doibase10.1103/physreva.89.052302 [ * * , ( ) ] link:\doibase10.1103/physreva.90.033812 [ * * , ( ) ] link:\doibase10.1103/physrevlett.113.140401 [ * * , ( ) ] link:\doibase10.1103/physrevlett.113.170401 [ * * , ( ) ] link:\doibase10.1103/physreva.90.032312 [ * * , ( ) ] @noop link:\doibase10.1103/revmodphys.81.865 [ * * , ( ) ] link:\doibase10.1103/physrevlett.111.250404 [ * * , ( ) ] link:\doibase10.1080/09500340008244048 [ * * , ( ) ] link:\doibase10.1103/physrevlett.78.2275 [ * * , ( ) ] link:\doibase10.1103/physreva.57.1619 [ * * , ( ) ] @noop * * , ( ) link:\doibase10.1103/physrevlett.114.210401 [ * * , ( ) ] link:\doibase10.1088/0305 - 4470/33/22/101 [ * * , ( ) ] link:\doibase10.1103/physrevlett.84.2014 [ * * , ( ) ] link:\doibase10.1098/rspa.2004.1372 [ * * , ( ) ] @noop link:\doibase10.1103/physrevlett.80.2245 [ * * , ( ) ] link:\doibase10.1103/physreva.68.042307 [ * * , ( ) ] link:\doibase10.1088/1367 - 2630/12/12/123004 [ * * , ( ) ] link:\doibase10.1103/physreva.67.062314 [ * * , ( ) ] link:\doibase10.1103/physreva.65.062312 [ * * , ( ) ] link:\doibase10.1103/physreva.73.044301 [ * * , ( ) ] link:\doibase10.1103/physrevlett.106.220403 [ * * , ( ) ] link:\doibase10.1103/physrevlett.106.160401 [ * * , ( ) ] link:\doibase10.1142/s0219749911008258 [ * * , ( ) ] link:\doibase10.1103/revmodphys.84.1655 [ * * , ( ) ] link:\doibase10.1007/978 - 3 - 319 - 09656 - 8 [ _ _ ] ( , ) link:\doibase10.1103/physreva.85.040301 [ * * , ( ) ] link:\doibase10.1103/physreva.88.012117 [ * * , ( ) ] link:\doibase10.1103/physrevlett.112.140501 [ * * , ( ) ] @noop link:\doibase10.1103/physreva.91.042120 [ * * , ( ) ] @noop g. adesso , http://quantumcorrelations.weebly.com/uploads/6/6/5/5/6655648/2015_badhonnef_activations_cde.pdf[_coherence , discord , and entanglement : activating one resource into another and beyond _ ] , talk presented at the 586 we - heraeus - seminar on `` quantum correlations beyond entanglement '' ( bad honnef , april 2015 ) . y. yao , x. xiao , l. ge , and c. p. sun , @noop . |
the dirac operator @xmath0 on a compact riemannian manifold or on a finite simple graph @xcite is the square root of the hodge laplacian @xmath7 , where @xmath8 is the exterior derivative . together with @xmath9 satisfying @xmath10 we can look at the lax pair @xcite @xmath11 \ ; .
\label{lax}\ ] ] this is motivated by @xcite and of course @xcite .
while the laplacian has nonnegative spectrum , the dirac operator @xmath12 has the symmetric spectrum @xmath13 . the deformed operator has the form @xmath14 . in the simplest case , the differential equations @xmath15 preserve @xmath16 and @xmath17 .
we also have @xmath18 .
if @xmath19 is a cocycle , and @xmath20 then @xmath21 so that @xmath22 remains a cocycle .
if @xmath23 is a coboundary and @xmath24 then @xmath25 so that @xmath22 remains a coboundary .
the system therefore deforms cohomology : graph or de rham cohomology does not change if we use @xmath26 instead of @xmath8 .
define the unitary operator @xmath27 by @xmath28 . from @xmath29
we get @xmath30 . as for any lax pair , since @xmath31 , we see that @xmath32 is isospectral to @xmath33 . because @xmath34=[b ,- b^2]=0 $ ] , the laplacian @xmath35 is time independent .
the solution @xmath36 of the linear wave equation @xmath2 only involves the laplacian @xmath35 and therefore is not affected by the deformation .
the nonlinear wave @xmath37 with @xmath38 however depends on @xmath12 .
this time - dependent hamiltonian flow will in the complex be replaced by the much more natural integrable nonlinear flow @xmath39 satisfying @xmath40 where @xmath41 is like @xmath42 a square root of @xmath43 . since @xmath26 is a deformed exterior derivative
, we can look at the new dirac operator @xmath44 and the new laplacian @xmath45 and @xmath46 .
we check that the laplacian @xmath47 decomposes into two commuting operators .
we have @xmath48 $ ] .
the operators @xmath49 and @xmath50 commute because their products @xmath51 are zero : initially , @xmath52=0 $ ] , we differentiate to get @xmath53=0 $ ] .
we also have @xmath54 = [ b ,- bd
d^*-d d^ * b]=0 $ ] showing that the four operators @xmath55 commute and can be simultaneously diagonalized . with @xmath56 , with @xmath57 on @xmath58 forms @xmath59 , we have the supersymmetric relations @xmath60 which describe the individual deformed geometries . as for any exterior derivative @xmath61 , the anti - commutation relations @xmath62
hold , where @xmath63 is the laplacian on @xmath58-forms .
supersymmetry fails for the deformed operator @xmath64 in the sense that @xmath65 and @xmath12 does no more produce an isomorphism between the bosonic subspace @xmath66 and fermionic subspace @xmath67 of the exterior bundle @xmath68 .
the symmetry is still present however because @xmath69 could be deformed also but we can not save the invariant splitting into fermions and bosons . a key observation for the scattering analysis is to see that @xmath70 is selfadjoint with no negative eigenvalues and @xmath71 is selfadjoint with no positive eigenvalues .
proof : the eigenvalues of @xmath72 are initially zero because @xmath73 is zero at @xmath5 .
look at an eigenvalue @xmath74 of @xmath75 . from the rayley formula @xmath76 , where @xmath77 is the unit eigenvector and @xmath78 the dual vector @xmath79 and by symmetry
, we have @xmath80 . we know
that @xmath73 and @xmath81 have the same eigenvectors for @xmath82 because they commute . using @xmath83
we get @xmath84 . if @xmath77 is an eigenvector to @xmath75 to the eigenvalue @xmath74 , then @xmath85 .
this means that @xmath86 if @xmath87 . in other words
, eigenvalues can not cross @xmath88 .
the computation for @xmath89 is similar .
we have lyapunov functions because @xmath90 and so @xmath91 : @xmath92 .
initially , we have @xmath93 and asymptotically , we have @xmath94 so that @xmath95 will have a maximum somewhere .
we see an initial inflation both in the positive and negative time direction .
we have @xmath96 because these matrices do not have anything in the diagonal . from @xmath97
follows @xmath98 .
we see that @xmath99 has its spectrum @xmath100 $ ] with @xmath101 for @xmath6 .
it follows that @xmath26 and @xmath102 converge to zero and @xmath103 converges to an operator @xmath104 satisfying @xmath105 . if @xmath106 then @xmath107 which is not possible for @xmath82 by looking at eigenplanes . for every @xmath108
, we have @xmath109 at @xmath5 .
this is clear for odd @xmath110 , because there is nothing in the diagonal . for even @xmath110 , note that @xmath111 at all times , so that @xmath112 .
but we still know that @xmath113 , by the classical mckean - singer formula @xcite .
the nonlinear analogue @xmath114 of mckean - singer holds : we know that @xmath113 initially because of the linear mckean singer result . because @xmath115 is real analytic and @xmath116 by definition , it is enough to verify that @xmath117 for all @xmath110 at @xmath5 . to see this ,
differentiate @xmath118 at @xmath5 to get @xmath119 . using @xmath120 and @xmath121
one can deduce that all these derivatives are zero .
we have @xmath122 because this is true at @xmath5 and because @xmath123 for all @xmath124 .
in particular , the attractors satisfy @xmath125 .
we have sketched the proof of : both in the graph and manifold case , system ( [ lax ] ) has the property that the limit @xmath126 and @xmath127 exists .
we have @xmath128 for some unitary @xmath118 satisfying the mckean - singer equations @xmath129 for all @xmath124
. the laplacian @xmath1 does not change but it is for all @xmath124 a sum @xmath130 of two commuting operators where @xmath131 is a new laplacian for a new dirac operator @xmath132 belonging to a new exterior derivative which has the same cohomology than @xmath8 , and a block diagonal part @xmath133 whose limit is a square root of @xmath35 .
* 1 ) * in the riemannian manifold case , @xmath12 is a differential operator on the exterior bundle of @xmath134 .
the deformed operator is a pseudo differential operator .
the deformation can be described by matrices , once a basis is chosen . for the circle @xmath135 , the dirac operator is @xmath136 $ ] and @xmath137 $ ] leaves 0-forms and 1-forms invariant .
fourier theory gives an eigenbasis @xmath138 $ ] belonging to eigenvalues @xmath139 so that @xmath140 is the set of integers @xmath141 and @xmath35 has the eigenvalues @xmath142 for @xmath143 .
the zeta function of the circle @xmath144 $ ] is analytic in the entire complex plane as for any odd dimensional riemannian manifold .
the factor @xmath145 has naturally regularized the poles of the minakshisundaram - pleijel zeta function @xmath146 . for the circle , the dirac zeta function has the same roots then the classical riemann zeta function . since @xmath147 , the circle has the dirac ray - singer determinant @xmath148 .
the deformation @xmath149 $ ] satisfies @xmath150 $ ] using @xmath151 $ ] can be written as the matrix differential equation @xmath152 since the quantity @xmath153 is time invariant , @xmath35 is block diagonal with entries @xmath154 and @xmath155 . in fourier space ,
@xmath156 are double infinite matrices .
we have @xmath157 and @xmath158 and @xmath159 . system ( [ 2 ] ) shows initial inflation and asymptotic exponential expansion of the individual circles . also for general initial conditions @xmath160 , system ( [ 2 ] )
satisfies @xmath161 .
the deformed dirac operator @xmath64 describes a larger geometry of two separate circles .
+ * 2 ) * for a finite simple graph @xmath162 with simplex set @xmath163 of cardinality @xmath164 , euler characteristic @xmath165 , the dirac operator @xmath12 is a @xmath166 matrix with @xmath167 . for the two point graph @xmath168 , with @xmath169
the dirac operator is the @xmath170 matrix @xmath171 $ ] with eigenvalues @xmath172 .
the laplacian is @xmath173 $ ] .
the zeta function is @xmath174 , the pseudo determinant @xmath175 . starting with @xmath176 $ ] we end up with @xmath177/\sqrt(2)$ ] .
the differential equation for @xmath178 \ ; , b = \left [ \begin{array}{ccc } 0 & 0 & -d \\ 0 & 0 & d \\ d & -d & 0 \end{array } \right]$ ] simplify to @xmath179 because @xmath180 and @xmath181 . explicit solutions of ( [ 3 ] ) are @xmath182 , @xmath183 and the integral @xmath184 whose level curves are ellipses .
looking at @xmath185 confirms initial inflation for the increasing distance of the two points @xmath186 .
\1 ) * hamiltonian formalism . * to see this system as a hamiltonian system to the hamiltonian @xmath187 in the manifold case , we have to consider dirac zeta function @xmath188 .
note however that to define and run the flow , we do not need the hamiltonian .
the traces @xmath189 and the determinant @xmath190 are then defined by analytic continuation . unlike the zeta function of the laplacian @xcite which is only meromorphic
, we can chose the branch @xmath191 , where @xmath192 is the zeta function of the laplacian , so that the dirac zeta function has an analytic extension everywhere both in the graph or compact riemannian manifold case .
the branch is natural because it leads to the trivial roots @xmath193 agreeing with the fact that the operator @xmath194 has only @xmath88 s in the diagonal . for every observable @xmath195 and for even positive @xmath196
, there is a hamiltonian flow @xmath197 = j \nabla h(d)$ ] .
but because @xmath198 and because @xmath35 commutes with @xmath199 , higher degree flows are related by the first flow with hamiltonian @xmath200 by a energy dependent time change on each eigenplane .
+ 2 ) * broken supersymmetry . *
the eigenfunctions @xmath4 are initially perpendicular superpartners and become parallel or anti - parallel for @xmath6 , hiding supersymmetry : if @xmath201 is a boson , then @xmath202 is only a fermion at @xmath5 .
@xmath102 defines a connes pseudo metric @xmath203 @xcite leading to an expanding space which features a fast inflationary start near the supersymmetric origin . @xmath12
determines via non - commutative geometry a pseudo distance in the larger space @xmath204 , the union of the set @xmath205 of @xmath58-cliques or the exterior bundle @xmath206 which is a union of p - branes @xmath207 , images of the gelfand transform of the banach algebra of @xmath58-forms with pointwise multiplication .
while the linear dirac evolution preserves supersymmetry , the nonlinear evolution provides a mechanism to see it broken for measurements .
+ 3 ) * distances*. @xmath208 always has infinite distance to @xmath209 in the connes pseudo metric @xmath8 because @xmath210 constant @xmath211 .
more generally , if the cohomology group @xmath212 is nontrivial then @xmath205 has infinite distance to all other @xmath213 for @xmath214 and furthermore , @xmath8 is not a metric .
an other distance in @xmath215 or @xmath216 is obtained by measuring how long a wave @xmath217 with @xmath218 and @xmath219 localized at @xmath220 needs to get to @xmath221 . if the minimal @xmath124 such that @xmath222 for @xmath218 , then @xmath223 .
also in the nonlinear case , define @xmath224 to be the minimal @xmath124 such that for some @xmath218 we have @xmath225 .
this is possible if @xmath226 has a solution @xmath218 .
linear or nonlinear quantum mechanics makes it possible that with @xmath215 or @xmath216 , different k - branes @xmath163 or @xmath227 have finite distance .
+ 4 ) * deformed curvature . *
gauss - bonnet - chern @xcite and poincar - hopf @xcite and its link @xcite are the key to define curvature on the larger space .
for @xmath5 , the spaces @xmath205 or @xmath228 are disconnected and curvature @xmath229 is the expectation @xmath230 $ ] of the index @xmath231 defined for functions on @xmath205 or @xmath228 . at @xmath5
we take the product measure @xmath232 . the unitary evolution
@xmath27 pushes forward this probability measure @xmath232 on functions .
the new measure @xmath233 produces a new expectation and so deforms curvature @xmath234 . we have not yet explored the question how curvature defined by the new operator @xmath44 of the expanding space evolves if we rescale length so that the diameter stays constant . the answer will be interesting in any case : if the evolution simplifies space , it can be used as a geometric tool which unlike the ricci flow features global existence . if geometry is not simplified , it should lead to limiting geometries , or a dynamical system on an attractor of limiting geometries .
+ 5 ) * complex structure . * a complex generalization of the system is obtained by defining @xmath235 , where @xmath236 is a parameter .
this is similar to @xcite who modified the toda flow by adding @xmath237 to @xmath199 .
the case @xmath238 is now a very special case .
for @xmath239 , we have @xmath240 and @xmath241 satisfies @xmath242 but no more @xmath29 as in the real case .
the essential features of the system like expansion are @xmath236-independent because @xmath103 and so @xmath99 is @xmath236 independent .
the limit @xmath243 for the complex flow is the same than the limiting real flow .
the lax pair @xmath244 $ ] is now more symmetric and the nonlinear unitary flow @xmath27 is asymptotic to the linear dirac flow because @xmath245 , where @xmath73 is a square root of @xmath35 .
remarkably , a complex differential structure has emerged during the evolution because @xmath64 is complex for @xmath82 even so we have start with a real graphs or manifold .
define @xmath246 and @xmath247 , then @xmath248 and @xmath249 . because cocycles and coboundaries deform in an explicit way , cohomology groups defined by @xmath250 and @xmath251 are both the same than for @xmath8 .
since @xmath252 , the laplacian @xmath1 is the sum of two laplacians @xmath253 and @xmath254 , where @xmath255 and @xmath256 . one could call a graph @xmath162 with a complex structure given by @xmath12 a khler graph , if @xmath257 . the complex structure disappears asymptotically for @xmath258 ( see figure ( [ figure1 ] ) ) justifying that @xmath259 is close to @xmath27 satisfying @xmath260 . while the exterior derivative @xmath8 as well as the new dirac operator @xmath261 are complex for @xmath82 , the operator @xmath131 is real if we start with a real @xmath12 . while asymptotically , @xmath262
, the complex structure is especially relevant in the early stage of the evolution . | we introduce an integrable hamiltonian system which lax deforms the dirac operator @xmath0 on a finite simple graph or compact riemannian manifold . we show that the nonlinear isospectral deformation always leads to an expansion of the original space , featuring a fast inflationary start .
the nonlinear evolution leaves the laplacian @xmath1 invariant so that linear schrdinger or wave dynamics is not affected .
the expansion has the following effects : a complex structure can develop and the nonlinear quantum mechanics asymptotically becomes the linear relativistic dirac wave equation @xmath2 . while the later is not aware of the expansion of space and does not see the emerged complex structure , nor the larger non - commutative geometric setup , the nonlinear flow is affected by it .
the natural noether symmetries of quantum mechanics introduced here force to consider space as part of a larger complex geometry .
the nonlinear evolution equation is a symmetry of quantum mechanics which still features supersymmetry , but it becomes clear why it is invisible : while the mckean - singer formulas @xmath3 still hold , the superpartners @xmath4 are orthogonal only at @xmath5 and become parallel or anti - parallel for @xmath6 . |
the start of the new millenium finds the fundamental sciences at an odd cross - roads .
on one hand we have in the standard model ( which i take to include general relativity and neutrino masses ) an exquisitely accurate description ( at least in principle ) of all experiments that have ever been done .
on the other hand we also understand the limitations of this theory , which tells us it must break down at sufficiently high energies . what might be the new theory which replaces the standard model at higher energies , and are we likely to discover it in the near future ?
much in our science depends on the answer to these questions , and the nature of the answers depends crucially on precisely at which energies the standard model fails . we know that the physics of the longitudinal modes of the @xmath0 and @xmath1 bosons _ i.e. _ the physics of electro - weak symmetry breaking must lie just beyond our present experimental reach ( at a few tev or less ) , and this is the ultimate rationale behind the construction of the _ large hadron collider _ at cern . but what happens beyond this depends on more of the details , depending on what is found .
various clues have been proposed over the years , motivated by understanding the relative strength of the weak and gravitational interactions ( the electro - weak hierarchy ) ; the nature of cp violation in the strong interactions ; the patterns of masses and mixings of the known elementary particles ; or the small observed gravitational response to the energy of the vacuum .
all of the above clues share the feature that they demand an explanation for one of the standard model s small dimensionless numbers .
the reason why such small numbers provide a useful clue for the nature of new physics is based in how we understand small numbers in modern quantum field theories .
suppose a parameter @xmath2 is found to be small when measured in an experiment which is performed at an energy scale @xmath3 .
we wish to understand this in terms of a microscopic theory of physics which is defined at energies @xmath4 , in terms of which the prediction for @xmath2 is given by @xmath5 here @xmath6 represents the direct contribution to @xmath2 due to the parameters in the microscopic theory , and @xmath7 represents the contributions to @xmath2 which are obtained as we integrate out all of the physics in the energy range @xmath8 .
science has encountered an enormous number of examples of small parameters like this ranging from the ratio of the size of the nucleus to the size of atoms in atomic physics , to the ratio of the small superconducting gap to the typical electron energy in condensed - matter physics . in _ every single case _
( so far ) we understand the smallness of @xmath9 in two steps : ( @xmath10 ) the quantity @xmath6 is understood to be small in the microscopic theory , and ( @xmath11 ) the process of integrating out from @xmath12 to @xmath3 gives an equally small contribution @xmath13 . in this way the smallness of @xmath9 can be understood for any choice we may choose to make for the microscopic scale @xmath12 .
the ubiquity of this kind of explanation for small quantities has earned it a name : ` _ technical naturalness _ ' .
red flags are raised when both steps @xmath14 and @xmath15 can not be understood , and a naturalness problem is declared for the corresponding small quantity , @xmath9 .
the significance of such naturalness problems comes from the clues they provide about the existence of new physics .
in particular , although we may not be able to understand why @xmath6 is small until we have the ultimate microscopic theory in our hands , we do expect to be able to understand why the ordinary physics _ which we believe we already understand _ at energies @xmath8 should not make @xmath16 unacceptably large .
if we find @xmath7 to be many orders of magnitude larger than the measured value @xmath9 then we suspect that we do not actually understand the energies @xmath8 as well as we thought - and progress can be made because this suggests changes are required in an energy range which is experimentally accessible .
the gravitational response of the vacuum energy is the case which most egregiously resists such a technically - natural explanation @xcite . in this case
cosmological observations indicate that the vacuum s energy density is at present at most @xmath17 ev@xmath18 ( in units with @xmath19 ) @xcite .
but since a particle of mass @xmath20 typically contributes an amount @xmath21 when it is integrated out , such a small value for @xmath22 can only be understood in a technically natural way if @xmath23 ev or less . since essentially all of the elementary particles we know
( including the electron , for which @xmath24 ev ) have @xmath25 ev , it has proven impossible to understand why our understanding of ordinary particles agrees with experiments so well , and yet so badly predicts the vacuum s observed gravitational response .
a technically natural explanation for the vacuum energy density requires two things .
first , it requires a fundamental modification of how gravity responds to physics at scales @xmath26 ev .
second , whatever provides this modification must not ruin the excellent agreement with all of the many non - gravitational experiments which have been performed over the years for the energies @xmath8 , with @xmath27 ev , to which we have experimental access . it is the mutual contradictions of these two conditions which has made it so difficult to make progress .
remarkably , a framework now exists within which both of these conditions may be able to coexist : the framework of large extra dimensions @xcite . according to this picture
which is motivated by the discovery of @xmath28-branes within string theory all of the observed particles apart from the graviton may be trapped on a ( 3 + 1)-dimensional surface within an extra - dimensional space .
in such a world the presence of the extra dimensions would only make a difference for gravitational observables , since only gravitational probes could be used to probe for their existence . remarkably , the present upper limit for the size of such extra dimensions is @xmath29 m , or @xmath30 ev @xcite very close to the scale @xmath3 above which our natural understanding of the vacuum energy breaks down ! furthermore
, extra dimensions can only be this large if there are precisely two of them , and if the fundamental scale , @xmath31 , of the extra - dimensional physics is around 10 tev , due to the relation @xmath32 which relates @xmath31 and @xmath33 to the observed planck mass : @xmath34 ev ( where here @xmath35 denotes newton s constant ) .
other constraints from astrophysics can also be satisfied for these choices of scales @xcite . from this
we see that it is logically possible that the gravitational response of the vacuum could depart dramatically from our 4-dimensional expectations at precisely the scales @xmath36 ev where these expectations can not account for the vacuum energy in a technically natural way .
this observation leads one to ask what the gravitational response of the vacuum might be in such a framework , and to ask whether the gravitational response of the 6d theory can be much smaller than would be expected from the 4d perspective .
this line of thought leads to the proposal of supersymmetric large extra dimensions ( sled ) , which posits there two large ( _ i.e. _ @xmath37 @xmath3 m ) extra dimensions , arising within a supersymmetric theory @xcite .
the motivation for the large extra dimensions is as above having extra dimensions this large provides a loophole for the usual ( four - dimensional ) arguments which say that the vacuum energy must be of order @xmath38 , and so be too large .
supersymmetry is motivated partially because our best high - energy theories are supersymmetric , and partially because of the cancellations between bosons and fermions which appears in the vacuum energy .
this proposal is briefly summarized here , following the discussion of ref .
@xcite . in the sled picture supersymmetry
must be badly broken on our brane , since we know that there are no super - partners for the observed particles having masses which are much smaller than @xmath31 . given this scale for supersymmetry breaking on the brane
there is also a trickle - down of supersymmetry breaking to the ` bulk ' between the branes , whose size is set by the bulk s kaluza - klein scale , @xmath39 , and so which for unwarped geometries can be as low as @xmath40 ev @xcite .
much of the success of the sled proposal relies crucially on the ability to maintain this hierarchy between the scales of supersymmetry breaking on the brane and in the bulk ( for other approaches to separating the supersymmetry breaking scale see ref .
@xcite ) . within the above framework gravitational physics is effectively 6-dimensional for any energies above the scale , @xmath41 ev , and so the cosmological constant problem must be posed within this new context . in order to see how the cosmological constant problem is phrased in 6 dimensions , one must integrate out the degrees of freedom between the scales @xmath42 tev and @xmath43 ev .
we seek the cosmological constant within the effective 4d theory obtained after performing this integration , which describes gravitational physics ( like present - day cosmology ) on scales much larger than @xmath33 .
imagine , therefore , performing the integration over modes having energies @xmath44 in the following three steps @xcite : 1 .
first , integrate out ( exactly ) all of the degrees of freedom on the branes , to obtain the low - energy brane dependence on the massless 4d graviton mode .
in so doing we obtain ( among other things ) a large effective brane tension , @xmath45 for each of the 3-branes which might be present , which includes the vacuum energies of all of the presently - observed elementary particles .
2 . next ,
perform the classical part of the integration over the bulk degrees of freedom .
this amounts to solving the classical supergravity equations to determine how the extra dimensions curve in response to the brane sources which are scattered throughout the extra dimensions .
it will be argued that it is this classical response which cancels the contributions from the branes obtained in step 1 , above .
3 . finally , perform the quantum part of the integration over the bulk degrees of freedom .
given the cancellation of the previous two contributions , it is this contribution which is responsible for the fact that the present - day dark energy density is nonzero .
it is argued below that in some circumstances this quantum contribution is of order @xmath46 , where @xmath47 ev is the supersymmetry - breaking scale in the bulk .
the small size of the 4d vacuum energy is in this way attributed to the very small size with which supersymmetry breaks in the bulk relative to the scale with which it breaks on the branes . in a 4d world
, the only contribution we would have is that of step 1 , above , and the problem is that this is much too large .
but in a 6d world , because all of the observed particles are localized on our brane their vacuum energy should be thought of as a localized energy source in the extra dimensions , to which the bulk geometry must respond .
we next argue that the classical part of this bulk response ( step 2 , above ) is of the same order as the direct contribution of step 1 , and precisely cancels it in a way which does not depend on the details of the supergravity involved or of the precise extra - dimensional geometry which lies between the various branes . in this way it provides a 6 dimensional realization of self - tuning , whereby the effective 4d cosmological constant
is automatically adjusted to zero by the classical response of the 2d bulk to the brane sources .
the final nonzero result finally comes from step 3 , due to the quantum bulk contributions .
but because the supersymmetry breaking scale in the bulk is so small , these bulk loops are arguably the proper size to agree with the recently - observed dark energy density .
( it is this part of the argument which non - supersymmetric proposals crucially miss @xcite . )
step 1 consists of exactly integrating over all brane fields having masses larger than @xmath48 , and this produces a variety of local interactions in the effective theory for energies @xmath49 on the brane . since our interest is in the dependence of the effective theory on the 4d metric , and we assume a large volume for the extra 2 dimensions
@xmath50 we may expand these effective interactions in powers of the curvature : @xmath51\,,\ ] ] where on dimensional grounds we expect @xmath52 , @xmath53 _ etc _ .. step 2 consists of the classical part of the bulk integration , and so is equivalent to substituting into the classical action the bulk field configurations which are found by solving the classical field equations using the above effective brane action as a source .
it happens that for many situations this classical response of the extra dimensions can be computed explicitly within the approximation that the branes are regarded as delta - function tension sources . in the absence of nontrivial brane couplings to bulk fields like the dilaton or bulk gauge fields ( this assumption is relaxed below ) , co - dimension 2 objects
the extra - dimensional curvature tensor typically acquire a delta - function singularity at the position of the branes , corresponding geometrically to the presence of a conical defect at the brane position .
einstein s equations require that the singular contribution to the two - dimensional curvature is given by @xmath54 here @xmath55 denotes the position of the
@xmath10th brane in the transverse 2 dimensions , and the ` smooth contributions ' are all of those which do not involve a delta - function at the brane positions .
the effective 4d cosmological constant obtained after performing steps 1 and 2 above is obtained by plugging the above expression into the classical bulk action .
the effective 4d cosmological constant obtained at this order is then @xmath56 = 0 \ , , \ ] ] where the sum on ` @xmath57 ' is over the various branes in the two extra dimensions and `
@xmath58 ' denotes all of the other terms besides the einstein - hilbert term in the supersymmetric bulk action .
the final equality here has two parts .
first , the sum over brane tensions , @xmath59 , precisely cancels the contribution of the singular part of the curvature , eq .
( [ ricci2 ] ) , to which they give rise @xcite .
second , for supersymmetric theories a similar cancellation also occurs amongst the various ` smooth ' contributions in @xmath60 once these are evaluated for all of the bulk fields using the classical field equations @xcite .
interestingly , this cancellation does not depend on the details of the bulk geometry , or on the number of branes , since it relies only on a classical scale invariance which all 6d supergravity actions enjoy @xcite .
best of all , this cancellation does not depend at all on the _ value _ of the brane tension , @xmath59 , and so applies equally well even if these tensions are large and include all of the quantum effects due to virtual particles localized on the branes .
we are left with the contribution of quantum effects in the bulk ( step 3 ) , to which we return in more detail in subsequent sections .
these must ruin the brane - bulk cancellation because the scale invariance of the classical supergravity equations is not a bona - fide quantum symmetry .
however the bulk sector of the theory is also one which is almost supersymmetric , since the bulk supersymmetry - breaking scale is very small : @xmath40 ev . as a result
we might expect standard supersymmetric cancellations to suppress the quantum part of the result by powers of @xmath61 , and if the leading term should be of order @xmath46 this would be the right size to account for the observed dark energy density . under certain circumstances
this is indeed what happens for some 6d supergravities @xcite .
quantum corrections do lift the flat directions of the classical approximation , and those loops involving bulk fields do so by an amount which is of order @xmath62 , leading to @xmath63 where @xmath64 and @xmath57 are calculable constants and the logarithmic corrections generically arise due to the renormalization of uv divergences in even dimensions @xcite .
the result is this small despite the fact that the bulk loops include an sum over kaluza - klein modes having 4d masses right up to the tev scale , @xmath31 , due to cancellations which the extra - dimensional supersymmetry enforces .
what is interesting about the potential , eq . ( [ eq : vofr ] ) , is that it falls into a category of potentials which can provide a phenomenologically viable description of the dark energy @xcite .
furthermore , this remains true even though there are additional constraints which arise due to the extra - dimensional interpretation of the dark energy .
in particular , although the cosmological evolution of the extra - dimensional volume can imply a potentially dangerous time - dependence of newton s constant over cosmological epochs @xcite , the existing bounds which constrain how much this can happen are fairly easily satisfied due to the effects of hubble friction during the universe s expansion @xcite .
furthermore , this potential predicts that scalar - potential domination occurs when @xmath65 is of order @xmath66 , which can easily be the required value given a modest hierarchy amongst the coefficients , @xmath67 .
the above arguments assume particularly simple couplings between the branes and the various bulk fields like the dilaton , and these assumptions appear to play an important role in the scale - invariance properties which underly the cancellation between brane and bulk contributions @xcite .
it is therefore natural to wonder what happens if these assumptions are relaxed .
a good test of how these arguments generalize is based on a class of solutions to 6d chiral , gauged supergravity obtained in ref .
@xcite by gibbons , guvens and pope ( henceforth ggp ) .
what makes these solutions so useful as a test of self - tuning is that these authors derive the _
most general _ solution to these field equations subject to two assumptions : ( @xmath10 ) maximal symmetry in 4 dimensions ( _ i.e. _ de sitter , minkowski or anti - de sitter space ) , and ( @xmath11 ) axial symmetry in the internal 2 dimensions .
that is , they find the most general solutions whose metric has the form @xmath68 and for which @xmath69 and @xmath70 . here
the intrinsic 4d metric , @xmath71 , satisfies @xmath72 , for some constant @xmath73 .
the assumed axial symmetry corresponds to shifts of the coordinate @xmath74 , and the metric can have singularities at up to two positions , @xmath75 , within the internal 2 dimensions corresponding to the positions of source branes , but these singularities are not required to be only conical in form .
ggp find that there is a five - parameter family of solutions to the supergravity equations subject to these symmetry conditions .
what is most remarkable about these solutions is that _ every single one of them has a flat intrinsic 4d geometry _
( _ i.e. _ @xmath76 ) , even though none of them is supersymmetric ( except for a single solution containing no branes ) , as is consistent with the expectation that the effective 4d cosmological constant vanishes .
this same intrinsic flatness also appears to apply to the known solutions which lie outside of the ggp assumptions @xcite . having the most general solutions , even subject to a symmetry ansatz ,
also allows some exploration of how generic is the classical cancellation of the effective 4d cosmological constant .
in order to do so it is useful to keep track of the physical meaning of the parameters on which the general ggp solutions depend .
there are 5 such parameters , but one of these simply parameterizes the flat direction whose existence is guaranteed by the classical scale invariance of the supergravity equations .
a second parameter corresponds to another classical scaling property , under which a redefinition of the fields may be used to rescale the gauge coupling @xmath77 to any positive fixed value .
the three remaining parameters are broadly related to the three physical quantities which characterize these geometries : the tensions , @xmath78 , of the two branes which source the bulk geometry ; and the overall magnetic flux of the background magnetic field which ( marginally ) stabilize it .
we now ask what may be said about the natural of self - tuning given the properties of these general solutions .
there is a non - trivial constraint amongst the parameters of the model which hold quite generally for all of the ggp solutions , and it is natural to think that these constraints hide the fine - tunings which underlie the flatness of the 4d geometries .
in fact , they do not because they have their origins in topology , as is now explained .
there are two topological conditions which the ggp solutions all share : one which expresses that the internal 2d geometry is topologically a sphere ; and one which expresses the quantization ( and so also conservation ) of magnetic monopole flux @xcite .
since the first of these turns out to hold for all values of the parameters describing the classical solution , it is of less interest as a potential source of fine - tuning .
it is the quantization of monopole number which directly imposes a relation between the brane tensions , the gauge couplings and one of the 5 parameters characterizing the various ggp solutions .
the resulting topological constraint can be written in the following way @xcite : [ tensiontopology ] ( ) = n^2 ( ) , where @xmath79 is the integer which labels the monopole number . here
@xmath78 are the two brane tensions , @xmath77 is the gauge coupling which appears explicitly in the 6d supergravity action , @xmath80 is the gauge coupling for the background magnetic field and @xmath81 is an additive constant in the dilaton configuration ( and so is one of the parameters describing the solution ) .
although this looks like a hidden fine - tuning , first impressions deceive @xcite .
recall in this regard that the crucial issue for fine - tuning is whether or not the constraint is stable against renormalization .
that is , if eq .
( [ tensiontopology ] ) is imposed amongst the renormalized quantities at the tev scale , does it automatically remain imposed as successive scales are integrated out down to the scales below 1 ev ?
if so , then the constraint is technically natural , in the sense described above , and so is not fine - tuned this ( the most serious ) notion of fine tuning .
but topological constraints are _ always _ natural in this sense , because the integrating out of successive scales of physics is a continuous process , and since topological constraints involve quantization of quantities in terms of integers , they remain unchanged by any such continuous process .
topological constraints express global integrability conditions which must be satisfied in order for solutions to exist , rather than relations which select out a special class of ( flat ) solutions amongst a wider class which do not have this property .
however one thing does emerge from an analysis of the properties of the general solutions @xcite is the observation that not all initial brane configurations can give rise to classically stationary solutions . to see this notice that it turns out that only a subset of the general ggp solutions involve purely conical singularities , with the subset defined by the family of ggp solutions whose tensions satisfy the condition @xcite : [ tconstraint ] ( 1 - ) ( 1 - ) = ( ) = n^2 ( ) . here the last equality follows from using the topological constraint , eq .
( [ tensiontopology ] ) .
purely conical singularities are only possible for a one - parameter locus of tensions within the @xmath82 plane . since brane solutions have conical singularities only in the absence of couplings to the dilaton ( in the einstein frame ) @xcite it should be possible to choose a configuration of branes whose tensions do not satisfy eq .
( [ tconstraint ] ) , and so no bulk solution satisfying
the ggp assumptions can exist for such a choice .
one of the assumptions must fail for such a configuration and this is most likely the assumption of a static bulk geometry , most likely leading to a runaway configuration . such a situation would be similar to what obtains if an arbitrary configuration of electric charges were assembled : generically the forces between them do not balance and so they move to zero or infinite separation
. however the existence of a runaway need not in itself be a problem .
after all , we have seen that even if the classical solution is static , a runaway is generated by quantum corrections .
furthermore , this runaway can describe the dark energy density provided @xmath83 , up to logarithmic corrections .
if we suppose that such classical runaways exist , then the central question is : _ is the classical runaway too steep to describe the dark energy ?
_ this need not be a problem even if so , since we need not demand that _ all _ solutions describe our world , and we know that _ some _ solutions do not admit classical runaways . in this case
we must know if it is technically natural to demand that we find ourselves in a classically - static configuration .
a second question of central importance is : _ are there hidden self - tunings , in particular amongst the brane couplings ( for which supersymmetry can not come to the rescue ) ? _
a proper understanding of this requires a detailed understanding of the matching conditions between brane properties and bulk solutions , and although the answer to this question is not yet clear it is at present under active study @xcite .
the sled proposal states that the world becomes six - dimensional at sub - ev energies , in such a way that the bulk gravitational physics is supersymmetric down to the sub - ev kk scale .
the main motivation for this framework is that it dramatically changes the gravitational response of the energy of the vacuum in a way which appears to be technical natural , at least within the limits with which it has so far been checked @xcite .
regardless as to how the naturalness arguments turn out , the sled proposal dramatically changes how physics works at experimentally - accessible energies , and so it is falsifiable through the host of other robust phenomenological implications it makes beyond those it has for cosmology .
these include : * deviations from the inverse square law for gravity , which more precise estimates show should arise for distances of order @xmath84 @xmath3 m @xcite ; * a particular scalar - tensor theory of gravity at large distances , with the scalar(s ) being the moduli ( like the volume ) which describe the two large extra dimensions .
this is the same scalar whose time - dependence now describes the dark energy @xcite .
* distinctive missing - energy signals in collider experiments at the lhc due to the emission of particles into the extra dimensions @xcite . * potential astrophysical signals ( and bounds ) due to the possibility of having too much energy loss into the extra dimensions by stars and supernovae @xcite .
if the sled proposal is correct , it will be spectacularly so since it requires this entire suite of observational implications to be found .
indeed , it is this unprecedented connection between observables in cosmology and particle physics which is driven by its addressing the fundamental naturalness issues described in previous sections that sets the sled proposal apart from other descriptions of dark energy .
the sled proposal represents the combined efforts of a number of collaborators to whom i am most indebted , including yashar aghababaie , andy albrecht , georges azuelos , hugo beauchemin , petter callin , jim cline , dumitru ghilencea , doug hoover , quim matias , susha parameswaran , fernando quevedo , finn ravndal , constantinos skordis , gianmassimo tasinato , andrew tolley and yvonne zavala .
my funding during its development has come from nserc ( canada ) , fqrnt ( qubec ) , the killam foundation as well as mcgill and mcmaster universities .
many thanks to the organizers of theory canada i for their kind opportunity to present these results .
s. perlmutter et al . , ap .
j. * 483 * 565 ( 1997 ) [ astro - ph/9712212 ] ; a.g .
et al _ , ast .
j. * 116 * 1009 ( 1997 ) [ astro - ph/9805201 ] ; n. bahcall , j.p .
ostriker , s. perlmutter , p.j .
steinhardt , _ science _ * 284 * ( 1999 ) 1481 , [ astro - ph/9906463 ] ; d.n .
spergel , _ et al . _
[ astro - ph/0302209 ] . n. arkani - hamed , s. dimopoulos and g. dvali , phys
. lett . *
b429 * ( 1998 ) 263 [ hep - ph/9803315 ] ; phys . rev .
* d59 * ( 1999 ) 086004 [ hep - ph/9807344 ] ; i. antoniadis , n. arkani - hamed , s. dimopoulos and g. r. dvali , phys .
b * 436 * ( 1998 ) 257 [ hep - ph/9804398 ] . c. d. hoyle , u. schmidt , b. r. heckel , e. g. adelberger , j. h. gundlach , d. j. kapner , h. e. swanson , phys .
lett . * 86 * ( 2001 ) 1418 - 1421 , [ hep - ph/0011014 ] ; e.g. adelberger , b.r .
heckel and a.e .
nelson , ann .
nucl.part .
* 53 * ( 2003 ) 77121 , [ hep - ph/0307284 ] ; c. d. hoyle , d. j. kapner , b. r. heckel , e. g. adelberger , j. h. gundlach , u. schmidt and h. e. swanson , phys . rev .
d * 70 * ( 2004 ) 042004 [ hep - ph/0405262 ] .
t. banks , m. dine and a. e. nelson , jhep * 9906 * , 014 ( 1999 ) [ hep - th/9903019 ] ; s. cullen and m. perelstein , _ phys .
* 83 * ( 1999 ) 268 [ hep - ph/9903422 ] ; v. barger , t. han , c. kao , r .- j .
zhang , _ phys .
_ * b461 * ( 1999 ) 34 - 42 , [ hep - ph/9905474 ] ; c. hanhart , d.r .
phillips , s. reddy , m.j .
savage , _ nucl .
_ * b595 * ( 2001 ) 335 [ nucl - th/0007016 ] ; j.a .
pons , d.r .
phillips and s. reddy , _ phys .
lett . _ * b509 * ( 2001 ) 1 - 9 , [ astro - ph/0102063 ] . s. hannestad and g. raffelt ,
_ phys . rev .
* 87 * ( 2001 ) 051301 , [ hep - ph/0103201 ] ; _ phys
. rev . lett . _ * 88 * ( 2002 ) 071301 , [ hep - ph/0110067 ] ; _ phys
. rev . _
* d67 * ( 2003 ) 125008 , erratum - ibid .
* d69 * ( 2004 ) 029901 , [ hep - ph/0304029 ] ; s. hannestad , _ phys .
_ * d64 * ( 2001 ) 023515 , [ hep - ph/0102290 ] .
y. aghababaie , c.p .
burgess , s. parameswaran and f. quevedo , nucl .
phys . * b680 * ( 2004 ) 389414 , [ hep - th/0304256 ] ; y. aghabababie , c.p .
burgess , j.m .
cline , h. firouzjahi , s. parameswaran , f. quevedo , g. tasinato and i. zavala , jhep 0309 ( 2003 ) 037 [ hep - th/0308064 ] ; c.p .
burgess , f. quevedo , g. tasinato and i. zavala , [ hep - th/0408109 ] ; d. ghilencea , c.p .
burgess and f. quevedo ( hep - th/0506164 ) .
y. aghabababie , c.p .
burgess , j.m .
cline , h. firouzjahi , s. parameswaran , f. quevedo , g. tasinato and i. zavala , jhep 0309 ( 2003 ) 037 [ hep - th/0308064 ] .
burgess , _ ann .
* 313 * ( 2004 ) 283 - 401 [ hep - th/0402200 ] .
c. p. burgess , j. matias and f. quevedo , nucl .
b * 706 * , 71 ( 2005 ) [ hep - ph/0404135 ] .
c. p. burgess , f. quevedo , g. tasinato and i. zavala , jhep * 0411 * , 069 ( 2004 ) [ hep - th/0408109 ] .
g. azuelos , p.h .
beauchemin and c.p .
burgess , [ hep - ph/0401125 ] ; p.h .
beauchemin , g. azuelos and c.p .
burgess , _
j. phys . _ * g30 * ( 2004 ) n17 [ hep - ph/0407196 ] .
c.p . burgess and d. hoover , [ hep - th/0504004 ]
; d. hoover and c.p .
burgess , [ hep - th/0507293 ] ; d. ghilencea , d. hoover , c.p .
burgess and f. quevedo , jhep 0509 ( 2005 ) 050 , [ hep - th/0506164 ] .
c. p. burgess , aip conf .
* 743 * ( 2005 ) 417 [ hep - th/0411140 ] .
d. atwood , c.p .
burgess , e. filotas , f. leblond , d. london and i. maksymyk , _ phys .
* d63 * ( 2001 ) 025007 [ hep - ph/0007178 ] .
e. kiritsis and c. kounnas , _ nucl . phys .
_ * b503 * ( 1997 ) 117 [ hep - th/9703059 ] ; t. banks and m. dine , _ nucl .
_ * b479 * ( 1996 ) 173 [ hep - th/9605136 ] ; e. dudas and c. grojean , _ nucl .
* b507 * ( 1997 ) 553 [ hep - th/9704177 ] ; i. antoniadis and m. quirs , _ phys .
* b416 * ( 1998 ) 327 [ hep - th/9707208 ] ; z. lalak and s. thomas , _ nucl .
phys . _ * b515 * ( 1998 ) 55 [ hep - th/9707223 ] ; e. dudas , _ phys . lett . _ * b416 * ( 1998 ) 309 [ hep - th/9709043 ] ; h. p. nilles , m. olechowski and m. yamaguchi , _ phys . lett . _
* b415 * ( 1997 ) 24 [ hep - th/9707143 ] ; t. li , j. lopez and d. nanopoulos , _ mod .
_ * a12 * ( 1997 ) 2647 [ hep - ph/9702237 ] ; k. choi , _ phys .
_ * d56 * ( 1997 ) 6588 [ hep - th/9706171 ] ; k. choi , h. b. kim and c. muoz , _ phys .
rev . _ * d57 * ( 1998 ) 7521 [ hep - th/9711158 ] ; a. lukas , b. ovrut and d. waldram , _ phys .
_ * d57 * ( 1998 ) 7529 [ hep - th/9711197 ] ; _ nucl .
* b532 * ( 1998 ) 43 [ hep - th/9710208 ] ; e. mirabelli and m. peskin , _ phys .
_ * d58 * ( 1998 ) 065002 [ hep - th/97065002 ] ; k. benakli , _ phys .
rev . _ * d60 * , 104002 ( 1999 ) [ hep - ph/9809582 ] ; c. p. burgess , l. e. ibez and f. quevedo , _ phys .
* b447 * , 257 ( 1999 ) [ hep - ph/9810535 ] ; l. randall and r. sundrum , _ nucl .
phys . _ * b557 * ( 1999 ) 79 , [ hep - th/9810155 ] ; c. p. burgess , r. c. myers and f. quevedo , phys .
b * 495 * ( 2000 ) 384 [ hep - th/9911164 ] ; a. anisimov , m. dine , m. graesser and s. thomas , [ hep - th/0201256 ] .
for some early precursors to these ideas see : v.a .
rubakov and m.e .
shaposhnikov , phys .
lett . * b125 * ( 1983 ) 139 ; c. wetterich , nucl .
* b255 * ( 1985 ) 480 ; s. m. carroll and m. m. guica , [ hep - th/0302067 ] .
h. p. nilles , a. papazoglou and g. tasinato , nucl .
b * 677 * , 405 ( 2004 ) [ hep - th/0309042 ] .
a. albrecht , c.p .
burgess , f. ravndal and c. skordis , _ phys .
* d65 * ( 2002 ) 123506 [ astro - ph/0107573 ] ; k. kainulainen and d. sunhede , [ astro - ph/0412609 ] . g. w. gibbons , r. guven and c. n. pope , phys .
b * 595 * , 498 ( 2004 ) [ hep - th/0307238 ] .
s. randjbar - daemi and e. sezgin , nucl
. phys .
b * 692 * ( 2004 ) 346 [ hep - th/0402217 ] .
a. kehagias , phys .
b * 600 * ( 2004 ) 133 [ hep - th/0406025 ] .
h. m. lee and a. papazoglou , nucl .
b * 705 * ( 2005 ) 152 [ hep - th/0407208 ] .
v. p. nair and s. randjbar - daemi , jhep * 0503 * ( 2005 ) 049 [ hep - th/0408063 ] . s. l. parameswaran , g. tasinato and i. zavala , [ hep - th/0509061 ] ; h. m. lee and c. ludeling , [ hep - th/0510026 ] .
g. w. gibbons and c. n. pope , nucl .
b * 697 * , 225 ( 2004 ) [ hep - th/0307052 ] .
p. bostock , r. gregory , i. navarro and j. santiago , phys .
lett . * 92 * , 221601 ( 2004 ) [ hep - th/0311074 ] ; i. navarro and j. santiago , jhep * 0502 * , 007 ( 2005 ) [ hep - th/0411250 ] ; j. vinet and j. m. cline , phys .
d * 71 * , 064011 ( 2005 ) [ hep - th/0501098 ] ; j. vinet and j. m. cline , phys . rev .
d * 70 * , 083514 ( 2004 ) [ hep - th/0406141 ] ; j. m. cline , j. descheneau , m. giovannini and j. vinet , jhep * 0306 * , 048 ( 2003 ) [ hep - th/0304147 ] ; j. m. schwindt and c. wetterich , [ hep - th/0508065 ] ; e. papantonopoulos and a. papazoglou , [ hep - th/0507278 ] ; g. kofinas , [ hep - th/0506035 ] ; j. m. schwindt and c. wetterich , [ hep - th/0508065 ] . | this article briefly summarizes and reviews the motivations for and the present status of the proposal that the small size of the observed dark energy density can be understood in terms of the dynamical relaxation of two large extra dimensions within a supersymmetric higher - dimensional theory .
+ + pacs nos . :
31.15.pf , 31.30.jv , and 32.10.hq french version of abstract ( supplied by cjp if necessary ) # 1to |
granular metals are composite materials where a conducting phase made of metallic particles or clusters with mean sizes ranging from about 10 @xmath3 to 200 @xmath3 or more is randomly dispersed into an insulating glassy material . depending on the nature of the insulating and conducting phases , their relative concentrations and the typical size of the metallic granules ,
these materials show a quite rich phenomenology and peculiar properties such as metal - insulator transitions , giant hall effect,@xcite and variable - range hopping type of transport at low temperatures for insulating samples.@xcite the main ingredients governing most of the granular metal properties are the concentration @xmath4 of metal in the composite , the activation energy @xmath5 and the tunneling parameter @xmath6 , where @xmath7 is the tunneling distance and @xmath8 is the localization length . by varing @xmath4 ,
the composite undergoes a metal - insulator transition at a critical value @xmath9 which can be interpeted as the percolation thresold.@xcite for @xmath10 , connected metallic grains form a macroscopic cluster which extends over the whole sample and current can flow easily from one end to another when a voltage difference is applied . in this regime
, transport has a metallic character identified by a positive temperature coefficient of resistivity ( tcr ) .
when @xmath11 , transport is dominated by tunneling and the resistivity increases as the temperature is lowered ( so tcr is negative ) . in this regime , the activation energy @xmath5 plays a fundamental role since it defines the scale of energy an electron should overcome to successfully tunnel from one grain to another .
it is this parameter that can give rise , together with @xmath12 , to variable - range hopping type of transport at sufficiently low temperatures.@xcite the experimental determination of @xmath9 is however not as clear cut as the above discussion seems to suggest . for granular metals with low values of @xmath5 and @xmath4 close to the percolative thresold , tcr can change sign as the temperature is varied in such a way that @xmath13 for @xmath14 and @xmath15 for @xmath16 , where @xmath17 is some temperature characteristic of the material.@xcite as pointed out in ref .
, the nonmonotonic behavior of tcr makes the determination of @xmath9 ambiguous and alternative methods should be developed . in this paper
we study the piezoresistive response of granular metals and show how its anisotropy is strictly related to the vicinity to the percolation critical point .
piezoresistance , _
i.e. _ , the variation of resistance @xmath18 or conductance @xmath19 upon an applied strain field @xmath20 , is an effect common also in bulk metals . if we consider an isotropic conducting material having bulk resistivity @xmath21 , length @xmath22 , and cross - sectional area @xmath23 , then the resistance @xmath18 is given by : @xmath24 the fractional change of @xmath18 , @xmath25 , due to a strain @xmath26 is given by : @xmath27 if @xmath28 is the poisson s ratio of the conductor , then it is possible to show that @xmath29 so that : @xmath30 where we have defined the total piezoresistive gauge factor @xmath31.@xcite it is clear that the total piezoresistive response is made of an intrinsic , @xmath32 , and a geometric , @xmath33 , component . for bulk metals ,
gf is about 2 and , since @xmath28 is usually between 0.2 and 0.4 , the geometrical effect is in this case more important than the intrinsic one.@xcite a different situation is encountered instead in the so - called thick - film resistors , a particular class of granular metals typically made of ruo@xmath34 , iro@xmath34 , or bi@xmath34ru@xmath34o@xmath35 granules embedded in a glassy matrix , for which gf values as large as @xmath36 has been reported,@xcite so that the intrinsic piezoresistive effect is much larger than the geometrical one .
such high strain - sensitivity resistances have been exploited to manufacture piezoresistive sensor devices successfully used for pressure and force measurements.@xcite due to their commercial applications , thick - film resistors are the granular materials for which the piezoresistive effect has been studied best and it is thought that inter - grain tunneling processes are responsible for their gf values.@xcite to illustrate this point , consider for example the conductance @xmath37 of a simple tunneling process between two metallic grains : @xmath38 . under an applied strain , the tunneling distance @xmath7
is changed to @xmath39 where @xmath40 is the strain .
this modifies the tunneling parameter , @xmath41 and consequently the conductance @xmath37.@xcite the effect on @xmath37 can be substantial because of its exponential dependence and , in principle , it can easily overwhelm the geometrical piezoresistive effect . of course , this tunneling mechanism of strain dependence should hold true also for other granular metals than the thick - film resistors as long as tunneling is an important element in transport properties .
however the magnitude of the intrinsic piezoresistive effect depends on several factors such as elastic heterogeneity , diffusion of the metallic phase into the glass etc .. in this paper , we point out that a factor affecting in an important way the piezoresistive response is the degree of tortuosity the current has in flowing through a sample of granular metal . as an example , consider fig . 1 where we report the results of a monte carlo calculation of a two dimensional bond - percolation resistor - network model .
in the network , a fraction @xmath0 of bonds distributed at random has a finite conductance @xmath37 , while the remaining fraction @xmath42 has zero conductance . when all bonds of the network are conducting ( @xmath43 , fig .
1a ) , the current flows along the direction of the applied voltage difference , that is the @xmath4 direction .
for @xmath44 the missing bonds force the current to aquire nonzero components also along the direction perpendicular to the applied field ( the @xmath45 direction ) as shown in fig .
1b where @xmath46 . for values of @xmath0 close to the percolation thresold @xmath47 ( @xmath48 for a simple square lattice ) ,
the tortuosity of the current flow is so developed that the contributions of components parallel and perpendicular to the applied voltage drop become comparable .
this is clearly shown in fig .
1c where the current flow is calculated for @xmath49 .
how the tortuosity of the current flow affects the piezoresistive response , or better its anisotropy , is qualitatively described as follows .
immagine that each single bond in the direction @xmath4 of the applied field is slightly stretched so that its conductance becomes @xmath50 , where @xmath20 is the strain.@xcite for simplicity , we keep the bonds along the direction perpendicular to the field unchanged : @xmath51 .
the variation of the total conductance in this situation leads to the longitudinal piezoresistive response ( applied strain parallel to the applied field ) .
such longitudinal response is nonzero for all the three cases depicted in fig .
1 since there are always nonzero components of the current along the direction of the applied strain . consider now the transversal piezoresistive response for which the strain is applied along the direction @xmath45 perpendicular to the field , _ i.e. _ , @xmath52 and @xmath53 .
contrary to the longitudinal case , for @xmath43 there are no components of the current along the @xmath45 direction and the total conductance is therefore unaffected by the applied strain . in this case
the transersal piezoresistive response is zero .
however , for @xmath44 current starts to flow also along the @xmath45 direction and becomes affected by the strain induced variation of @xmath54 leading to a nonzero piezoristance effect . since close to the percolation thresold the components of the current along the @xmath4 and @xmath45 directions are comparable ( fig .
1c ) , it is expected that in this case the longitudinal and the transversal piezoresistive responses have similar values . from the above qualitative discussion , we argue that away from the percolation thresold the piezoresistive response is highly anisotropic with the longitudinal component much larger than the transversal one , while close to the percolative critical point the two components becomes comparable , leading to an isotropic piezoresistive response . by using monte carlo calculations and results from the effective medium theory
we show in the following that for resistor network models this conclusion holds true and that actually the longitudinal and the transversal piezoresistive responses are equal at the percolation thresold .
hence , a measurement of the piezoresistive anisotropy could be an alternative way to determine the critical concentration @xmath9 of the conducting phase in granular metals .
the paper is organized as follows . in the next section
we introduce the model and , by using the effective medium theory , provide analytical formulas for the longitudinal and transversal piezoresistive responses . in sec .
[ numerical ] we report our monte carlo calculations and compare them with the analytical formulas of the effective medium theory .
the last section is devoted to a discussion and to the conclusions .
in this section , we evaluate the piezoresistive response of a resistor network model by using a generalized effective medium theory . before entering the details of our model ,
let us first introduce an operative definition of the piezoresistive effect and the different gauge factors . in fig .
2 we show a schematic apparatus for the measurement of the piezoresistive response . the sample is placed on top of a cantilever bar clamped at one end , and a force @xmath55 acts downward on the opposite end .
we assume that in the absence of the force , transport is isotropic with conductance @xmath19 .
when @xmath56 , a tensile stress is built on the upper surface of the cantilever and a strain @xmath57 is transfered to the sample .
this strain is directed along the main axis of the cantilever which we choose to be directed along the @xmath4 direction .
if the thickness of the cantilever is sufficiently small compared to its width , the strain along @xmath45 can in first approximation be neglected .
if we assume that the strain field is completely transfered to the sample ( which is a good approximation as long as the linear dimensions of the sample are much smaller than those of the cantilever ) , then the strain field acting on the granular metal is : @xmath57 , @xmath58 , and @xmath59 , where @xmath28 is the poisson ratio of the material . by referring to fig .
2 , the longitudinal piezoresistive response is obtained by setting a potential difference between points a and b , and the total conductance up to the linear term in the strain is : @xmath60 where we have defined the longitudinal gauge factor : @xmath61 when the potential difference is applied between points c and d of fig .
2 , that is perpendicular to the strain in the @xmath4 direction , the conductance can be expressed as @xmath62 , where @xmath63 is the transversal gauge factor : @xmath64 although @xmath65 and @xmath63 are the most used parameters , there is also a vertical gauge factor obtained applying a voltage difference between the upper and lower surfaces of the sample : @xmath66 .
let us describe now our resistor network model .
we consider a three dimensional simple cubic network where a bond has either a finite conductance @xmath37 with probability @xmath0 or zero conductance with probability @xmath42 .
we limit ourselves to the case in which the activation energy @xmath5 is sufficiently low and the temperature sufficiently high to approximate @xmath37 by a tunneling exponential .
furthermore , we allow a certain distribution in the tunneling distance in such a way that @xmath67 $ ] , where @xmath68 is a prefactor , @xmath69 is a random variable uniformly distributed between @xmath70 and @xmath71 , and @xmath72 can assume values between zero ( single tunneling distance ) and @xmath73 . for this latter case therefore the tunneling exponent is continuously distributed between @xmath70 and @xmath74 .
the probability distribution @xmath75 of the conductance value of a bond is : @xmath76 \theta\!\left[g - g_0 e^{-k-\delta k/2}\right ] + ( 1-p)\delta(g),\ ] ] where @xmath77 is the heaviside step function . for a simple cubic network , the effective medium equation reads:@xcite @xmath78 which , by using eq.([emt4 ] ) , leads to the following expression for the effective macroscopic conductance @xmath79 : @xmath80 as expected by the effective medium approximation
, @xmath79 vanishes for @xmath81 , which could be interpreted as the effective field metal - insulator transition point .
now , let us consider how the effective medium theory should be generalized to describe the effects of applied strain fields . to simplify the problem
, we make the assumption that the strains are oriented along the directions of the bonds in the simple cubic network .
hence , in the presence of the strain field , the conductances become @xmath82 , @xmath54 , and @xmath83 according to whether their corresponding conducting bonds are oriented towards to the @xmath4 , @xmath45 , and @xmath84 directions , respectively . the probability distribution @xmath85 of a bond in the @xmath86 direction ( @xmath87 ) is then equal to eq.([emt4 ] ) where @xmath12 and @xmath72 are replaced by @xmath88 and @xmath89 , respectively .
therefore , if @xmath90 is the strain along the @xmath86 direction , @xmath91 and @xmath92 . within this model we have made the assumption that the strain field does not affect sensibly the prefactor @xmath68 and from now on we set it equal to unity .
the strain - induced anisotropy leads to a system of three coupled effective medium equations for the three macroscopic conductances @xmath93:@xcite @xmath94=0 ; \,\,\,\,\,\,\ , i = x , y , z,\ ] ] where , by following the work of bernasconi,@xcite @xmath95 and @xmath96 is the total resistance between two neighboring nodes in @xmath86 direction for the effective lattice . for the @xmath4 component
, @xmath97 reads : @xmath98+\bar{g}_y[1-\cos(y')]+ \bar{g}_z[1-\cos(z')]},\ ] ] and @xmath99 and @xmath100 are obtained from ( [ emt8 ] ) by replacing in the numerator @xmath101 with @xmath102 and @xmath103 , respectively . for general values of the @xmath93
, the integrals @xmath96 can be evaluated only numerically.@xcite however , we are interested in the piezoresistive effect which is just the linear response to an applied strain field so that , to our purposes , it is sufficient to evaluate the integrals @xmath96 at the first order in the strain . as we show below this procedure permits an analytical solution of the anisotropic effective field equations .
let us replace in eq.([emt8 ] ) the effective conductances @xmath93 with @xmath104 , @xmath105 , and @xmath106 , where , by following the definitions given above , we have identified the gauge factors for the effective conductances by : @xmath107 as shown in the appendix , up to order @xmath20 , @xmath97 reduces to : @xmath108 where @xmath109 .
similarly , we obtain for @xmath99 and @xmath100 : @xmath110 by substituting eqs.([emt10]-[emt11b ] ) into eq.([emt7 ] ) , and equating to zero the term linear in @xmath20 , we obtain three linear equations for the three gauge factors . after some algebra , we obtain for the longitudinal and transversal components : @xmath111 [ ( 1 - 2\nu)\xi ( 9py-\delta k)-12py(1-\nu ) ] } { 18py(1-\nu)[(4 - 9\xi)py+\xi\delta k]},\ ] ] @xmath112 [ 3p\delta kx-6kpy-\delta k](1 - 2\nu ) } { 18py(1-\nu)[(4 - 9\xi)py+\xi\delta k]},\ ] ] where : @xmath113 we show in the next section how @xmath65 and @xmath63 behave for different values of @xmath0 , @xmath12 , and @xmath72 in comparison with monte carlo numerical calculations . before doing so
, we stress here the main result of the effective medium theory , namely , that @xmath65 and @xmath63 are generally different unless @xmath81.@xcite in fact , from eqs.([emt12]-[emt15 ] ) we obtain : @xmath114.\ ] ] the effective medium theory therefore confirms the qualitative arguments discussed in the introduction and , as we shall see in the next section , it captures the essential physics governing piezoresistive anisotropy .
it is also worth stressing that , at @xmath115 , @xmath116 independently of the particular strain field used .
in this section we compare the results of the effective medium theory , eqs.([emt12]-[emt15 ] ) , with numerical calculations of a three dimensional random resistor network model . in our monte carlo calculations ,
we have considered cubic networks with number of nodes up to @xmath117 and bond conductances as described in the last section .
the total conductance has been obtained by solving numerically the kirchoff equations for all nodes in the network , and periodic boundary conditions have been imposed to the sides of the network where no voltage difference is applied .
the different gauge factors have been obtained by calculating , for a fixed configuration of resistors , the difference in conductances when @xmath118 and @xmath119 .
the results have been then averaged over a minimum of 30 to a maximum of 100 different runs . in fig .
3 we show the results of monte carlo calculations ( symbols ) and effective medium theory ( lines ) for the limiting case for which there is no distribution in tunneling probabilities ( @xmath120 ) . in this limit , the effective medium formulas ( [ emt12]-[emt15 ] ) simplify considerably : @xmath121 } , \\ \label{num2 } \lim_{\delta k\rightarrow 0}{\rm gf}_{\rm t}&= & k\frac{3\xi(1-p)(1 - 2\nu ) } { ( 1-\nu)[9\xi(1-p)-4(1 - 3p)]}.\end{aligned}\ ] ] the results of fig .
3 confirm the qualitative analysis made in the introduction ( see also fig .
namely , for @xmath43 the current flows exclusively along paths parallel to the direction @xmath4 of the applied field leading to a non zero longitudinal piezoresistive effect ( @xmath122 ) and a vanishing transversal response ( @xmath123 ) . for @xmath44 ,
the missing bonds force the current to flow also along directions perpendicular to @xmath4 . in this case
, the piezoresistive response acquires a transversal component while the longitudinal one is lowered .
note that the analytical and numerical results agree quite well for bond probabilities larger than @xmath124 while for lower values of @xmath0 the effective medium results start to deviate from the monte carlo data .
this is of course due the increased inaccuracy of the effective medium theory as the percolation thresold is approached .
however , at the critical bond probability ( @xmath125 for simple cubic lattice,@xcite and @xmath126 for the corrsponding effective medium approximation ) both the monte carlo data and the analytical results predict equal longitudinal and transversal responses . finally , note that for @xmath127 , @xmath63 remains equal to zero also for @xmath44 .
this is due to the fact that , for this value of the poisson ratio and our assumption on the applied strain field , one has @xmath128 , therefore the bonds in the @xmath84 direction are contracted of the same amount as the bonds in the @xmath4 direction are stretched .
the results of fig . 3 show clearly how piezoresistive isotropy is related to the vicinity to the percolation thresold .
another limiting situation is the one depicted in fig .
4 where all bonds are present ( @xmath43 ) but with a finite ditribution of tunneling probability . here
, @xmath72 is varied between @xmath120 and @xmath73 , @xmath12 has been fixed equal to @xmath129 and @xmath130 , @xmath131 , and @xmath132 .
for @xmath120 we recover the limit @xmath43 of fig .
3 , while for finite values of @xmath72 the distribution of different tunneling probabilities induces tortuosity in the current flow which is reflected in a reduction of @xmath65 and an enhancement of @xmath63 . as before , the limit @xmath127 has the peculiarity of having @xmath123 independently of the current tortuosity .
note that , since for @xmath133 and @xmath43 the system is far away from the percolation thresold , effective medium theory agrees excellently well with the monte carlo calculations .
this agreement becomes however less satisfactory as the distribution width increases as shown in fig .
5 where results are reported for @xmath120 , @xmath12 , @xmath134 as a function @xmath12 for @xmath43 .
high values of tunneling distribution widths increase the fluctuactions of bond conductances reducing the validity of the effective medium approximation .
having analyzed two representative cases ( @xmath135 with @xmath120 , and @xmath43 with @xmath136 ) we show in fig .
6 the results for both @xmath135 and @xmath137 with @xmath133 and @xmath138 .
the difference between the longitudinal and transversal piezoresistive effects decreases as the bond conductance fluctuations increases whatever is the origin of such fluctuations ( missing bonds or finite distribution of bond conductances ) .
note that , contrary to the other cases , for @xmath139 @xmath65 increases as @xmath0 decreases . from the effective medium formulas it is found that this feature persists also for higher @xmath12 values if @xmath72 is somewhat larger than @xmath12 .
however it is difficult to test this tendency with our monte carlo calculations.since we have employed the relaxation method for which large values of @xmath72 seriously slow the convergency of the recursive algorithm .
the effect of bond fluctuations on the piezoresistive anisotropy is more clearly shown in fig .
7 where we plot , for the monte carlo data of fig .
6 , the quantity @xmath140 defined as : @xmath141 which measures the degree of anisotropy of the piezoresistive response . for a fixed value of @xmath72
, @xmath140 decreases as @xmath142 from above , and it is expected to vanish at @xmath2 .
@xmath140 decreases also at fixed @xmath0 as @xmath72 is made larger .
we believe that @xmath140 , because of its high-@xmath0 sensitivity for @xmath143 , could be a useful practical way to measure the proximity of a granular metal to its percolation critical point .
the results shown in the previous section clearly indicate how the anisotropy of the piezoresistive effect depends on the vicinity to the percolation thresold .
we have interpreted the reduction of anisotropy as due to the increasing of bond conductance fluctuations as @xmath142 from above .
although our results are quite general , their application to real granular metals needs some additional comments . in our bond percolation model , we have assumed that the external strain fields are applied along the directions of the bonds in the network .
however , in a more realistic situation the bond directions should be considered as random . in a practical calculation
this could be achieved by considering an ensemble of ordered networks with different orientations with respect to the applied strain fields .
the average over such an ensemble should be a satisfying description of a realistic case .
what we expect , and bond - average effective medium results ( not reported here ) confirm , is that even for @xmath120 and @xmath43 , the anisotropy parameter @xmath140 , eq.([chi ] ) , is less than one .
this is due to the random bond orientation which contribute , even if the network have strictly equal bond conductances , to a lowering of the piezoresistive anisotropy .
however , also in this case , the reduction of @xmath140 as @xmath0 is lowered ( or @xmath72 is enhanced ) and the limit @xmath144 for @xmath142 remain valid . another effect which has been ignored in the present analysis is the possibility of having elastic heterogeneity within the granular metal .
a large difference between the elastic constants of the metallic and insulating phases can give rise to important fluctuations in the local strain fields .
since the microscopic tunneling processes are affected by the local rather than the macroscopical strain values , the elastic heterogeneity can influence the piezoresistive response in an important way .
this is actually what is expected in thick - film resistors where very stiff metallic granules are embedded in a relatively elastically soft insulating glass .
this large elastic heterogeneity could be at the origin of the large values of piezoresistive responses actually observed for this class of granular materials.@xcite we expect however that the inclusion of this effect should influence the absolute values of @xmath65 and @xmath63 but leave the anisotropy parameter @xmath140 relatively unaffected .
in this appendix we report the evaluation of the integrals @xmath97 , @xmath99 , and @xmath100 up to the linear term in the strain @xmath20 .
let us consider first @xmath97 , eq.([emt8 ] ) , and substitute the effective conductances @xmath93 , @xmath87 , with @xmath145 , @xmath146 , and @xmath147 , where the gauge factors are defined in eq.([emt9 ] ) .
hence , up to the term linear in @xmath20 we obtain : @xmath148 where @xmath149 ^ 2}{[3-\cos(x')-\cos(y')-\cos(z')]^2},\\ \label{a4 } c&=&\frac{1}{\pi^3}\int_0^\pi\int_0^\pi\int_0^\pi dx'\,dy'\,dz'\,\frac{[1-\cos(x')][1-\cos(y ' ) ] ] } { [ 3-\cos(x')-\cos(y')-\cos(z')]^2}.\end{aligned}\ ] ] it is easy to show that @xmath150 and @xmath151 .
moreover @xmath152 can be rewritten as @xmath153 where : @xmath154 form ref . ,
we have : @xmath155 { \bf k}[(2-\sqrt{3})(\sqrt{3}-\sqrt{2})]^2 \simeq 0.5264,\ ] ] where @xmath156 is the complete elliptic integral of the first kind . equation ( [ emt10 ] ) is obtained by substituting @xmath151 , @xmath157 , and @xmath158 into eq.([a1 ] ) . | in this paper , we study the piezoresistive response and its anisotropy for a bond percolation model of granular metals .
both effective medium results and numerical monte carlo calculations of finite simple cubic networks show that the piezoresistive anisotropy is a strongly dependent function of bond probability @xmath0 and of bond conductance distribution width @xmath1 .
we find that piezoresistive anisotropy is strongly suppressed as @xmath0 is reduced and/or @xmath1 is enhanced and that it vanishes at the percolation thresold @xmath2 .
we argue that a measurement of the piezoresistive anisotropy could be a sensitive tool to estimate critical metallic concentrations in real granular metals .
pacs numbers : 72.20.fr , 72.60.+g , 72.80.ng = 10000 + + |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.