article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
by now the standard procedure employed for calculating the scattering length of hadrons from euclidean lattice qcd is to use the finite - size method proposed by lscher , which relates the finite volume shift of energy eigenvalues @xmath5 to the scattering lengths .
the formula reads @xcite @xmath6 where @xmath7 , and @xmath8 and @xmath9 are geometrical constants .
the energy shift @xmath5 is extracted from the time dependence of two - pion four - point functions . in this way
a number of calculations has been made for the @xmath0 @xmath1-wave two - pion scattering length using the staggered @xcite , the standard @xcite and the improved wilson fermion actions @xcite .
it should be noted that the condition @xmath10 is assumed for the two - pion interaction range @xmath11 and the lattice volume @xmath12 in derivation of the formula ( [ luscher.eq ] ) @xcite .
so far there have been studies of the lattice volume dependence of the scattering length , but no direct investigation of the interaction range @xmath11 .
it is extremely important that we examine the validity of the necessary condition for the lscher s formula in our current lattice simulations .
this can be done by investigating the two - pion wave functions , which is one of the purpose of this article .
similar work has been carried for 2 dimensional xy model @xcite . once one has access to the wave functions
, one can try to extract the scattering length from them .
this is the second , and perhaps more interesting , purpose of this article .
our idea is based on the derivation of the lscher s formula ( [ luscher.eq ] ) .
lscher found that the two - pion wave function @xmath13 on a finite periodic box @xmath12 satisfies the following effective schrdinger equation @xcite .
@xmath14 where @xmath15 is the relative coordinate of the two pion and @xmath16 is related to the two - pion energy eigenvalue through @xmath17 .
the function @xmath18 is the fourier transform of the modified bethe - salpeter kernel introduced in @xcite .
it is non - local and generally depends on energy . in the derivation of the formula it is assumed that @xmath19 for @xmath20 . in this region
the wave function satisfies the helmholtz equation @xmath21 .
the general solution of helmholtz equation on a finite periodic box @xmath12 can be written as @xmath22 up to over all constant , where @xmath23 . assuming the interaction range is smaller than one half of box size , _ ie .
_ @xmath10 , and expanding the general solution ( [ luscher.div_two ] ) in terms of spherical bessel @xmath24 and noeman @xmath25 functions for @xmath26 , we obtain @xmath27 where neglected terms are contributions from states with angular momentum @xmath28 .
the expansion coefficients @xmath29 and @xmath30 yield the scattering phase shift in infinite volume by @xmath31 .
in particular for the lowest energy state of the two - pion , it gives the scattering length by @xmath32 .
the constant @xmath30 are also related to @xmath29 geometrically and the relation leads to the lscher s formula ( [ luscher.eq ] ) .
= 7.2 cm in this work we define the two - pion wave function by @xmath33 where @xmath34 is an element of cubic group , and summation over @xmath34 and @xmath35 projects out the @xmath36 sector of the cubic group and that of the zero center of mass momentum . in order to enhance signals we use a source constructed with two wall sources given by @xmath37 , with @xmath38 the wall source at @xmath39 , in coulomb gauge fixed configurations .
we work in quenched lattice qcd employing an rg - improved action for gluons at @xmath40 and an improved wilson action for quarks at @xmath41 .
the corresponding lattice cutoff is estimated as @xmath42 gev from @xmath43 .
the volumes of lattices ( number of configuration ) are @xmath44 @xmath45 and @xmath46 @xmath47 which correspond to @xmath48 and @xmath49 in physical units .
quark masses are chosen to be @xmath50 , @xmath51 , and @xmath52 .
quark propagators are solved with the dirichlet boundary condition imposed in the time direction and the periodic boundary condition in the space directions .
the sources are set at @xmath53 and @xmath54 for both lattices . in figure [ fig.1.fig ]
we show the ratio of the wave function @xmath55 on a @xmath56 lattice at @xmath57 and @xmath58 . here
we choose @xmath59 for normalization point .
we find that the statistical error is very small .
figure [ fig.2.fig ] shows the ratio @xmath60 at the same parameters for several @xmath39 . this ratio should take a negative constant @xmath61 for the @xmath0 two - pion system for @xmath62 .
clear plateau appears for @xmath63 at @xmath64 as shown in the figure .
this indicates that the necessary size for application of the lscher s formula is @xmath65 ) .
we try to extract the value of @xmath16 from the wave function by fitting the data with the function defined by ( [ luscher.div_two ] ) .
the results are plotted by large cross symbols in figure [ fig.1.fig ] , where we choose @xmath66 as the fitting range .
the fit works very well . the result for @xmath16
is @xmath67 @xmath68 , which is consistent with that obtained from the time dependence of the two - pion four - point function , @xmath69 @xmath68 .
we check the consistency between the two determinations at all simulation points on @xmath56 lattice in figure [ fig.3.fig ] , where the scattering length is calculated by substituting @xmath16 to the lscher s formula ( [ luscher.eq ] ) .
we observe that the statistical error of the scattering length obtained from the wave function is smaller than that from the time correlation function .
this feature was also found for the 2 dimensional xy model @xcite . the wave function on a larger lattice of @xmath70 at @xmath71 and the results of the fitting are shown in figure [ fig.4.fig ] , where we choose @xmath57 .
we find a systematic discrepancy between the data and the fit for large @xmath72 region ( @xmath73 ) .
further there is a large discrepancy in the scattering length obtained from two methods as shown in figure [ fig.3.fig ] . a possible cause of the discrepancy is smallness of the time extent .
the strength of the two - pion interaction on a finite @xmath12 periodic box is proportional to @xmath74 .
if we consider a time - dependent effective shrdinger equation , the wave function satisfies a heat diffusion equation .
we expect that a larger time extent is necessary to obtain the correct wave function for a larger volume . in order to confirm this anticipation
, we are going to explore simulations on a temporal extent @xmath75 in our next study .
this work is supported in part by grants - in - aid of the ministry of education ( no.15740134 ) .
simulations were performed on the parallel computer pilot3 at the center for computational physics , university of tsukuba . 9 s.r .
et.al._ , nucl .
* b383 * ( 1992 ) 309 .
y. kuramashi _ et.al._ , phys .
* 71 * ( 1993 ) 2387 ; m. fukugita _ et.al._ , phys . rev .
* d52 * ( 1995 ) 3003 .
r. gupta _ et.al._ , phys . rev .
* d48 * ( 1993 ) 388 .
cp - pacs collaboration , s. aoki _ et.al._ , phys . rev . *
d67 * ( 2003 ) 014502 .
jlqcd collaboration , s. aoki _ et.al._ , phys .
* d66 * ( 2002 ) 077501 . c. liu _
et.al._ , hep - lat/0109010 ; nucl . phys .
* b624 * ( 2002 ) 360 .
m. lscher , commun .
* 105 * ( 1986 ) 153 ; nucl .
phys . * b354 * ( 1991 ) 531 .
j. balog _ et.al._ , nucl .
* b618 * ( 2001)315 . | we present a preliminary report on a calculation of scattering length for @xmath0 @xmath1-wave two - pion system directly from the two - pion wave function .
results are compared with those calculated from the time dependence of two - pion four - point functions .
calculations are made with an rg - improved action for gluons and improved wilson action for quarks at @xmath2 gev on @xmath3 and @xmath4 lattices . |
while the theoretical analysis discussed in this talk does involve the observed mass values of the top - quark , @xmath12 boson , and the @xmath13-quark , it is not a matter of any presently available empirical data disagreeing with a standard model ( sm ) prediction .
instead , the interest is because of some theoretical patterns of the helicity amplitudes for @xmath0 decay .
the theoretical patterns involve both the standard model s decay helicity amplitudes , @xmath14 , and the amplitudes @xmath15 in the case of an additional @xmath16 tensorial coupling of relative strength @xmath17 . to focus the discussion , in this talk
the @xmath5 amplitudes are interpreted as corresponding to the observed top - quark decays @xmath0 [ 1 ] .
this identification hypothesis might be excluded by future theoretical analysis and/or empirical data ; in ( i ) , alternatives to this identification were considered [ 2 ] .
experimental tests and measurements in ongoing and forthcoming [ 1,3,4 ] top - quark decay experiments at hadron and @xmath18 colliders should be able to significantly clarify matters .
the explicit expressions for these amplitudes , and other details , are given in ( i ) and in a hep - ph " preprint ( ii ) [ 2 ] .
measurement of the sign of the @xmath19 , due to the large interference between the @xmath20 and @xmath21 amplitudes , could exclude such a large chiral weak - transition - moment in @xmath22 decay in favor of the sm prediction . on the other hand , measurement of
the sm predicted fraction of final @xmath20 versus final @xmath21 bosons for this decay mode would not distinguish between the two cases .
the definitive empirical test must establish the sign of @xmath23 where @xmath24 is the relative phase of the two @xmath25 amplitudes , @xmath26 and @xmath27 , c.f .
table 1 below . in the @xmath28-quark rest frame ,
the matrix element for @xmath29 is @xmath30 where @xmath31 in terms of the @xmath32 and @xmath13-quark helicities . due to rotational invariance
, there are four independent @xmath33 amplitudes for the most general lorentz coupling .
we use the jacob - wick phase - convention for the amplitudes and use the subscript @xmath34 " to identify the amplitude s associated coupling ; in this paper @xmath35 sm , @xmath36 for only the additional @xmath37 tensorial coupling , and @xmath38 for @xmath39 when @xmath40 . with respect to the latter case ,
the lorentz coupling involving both the sm s @xmath41 coupling and an additional @xmath3 tensorial coupling of arbitrary relative strength @xmath8 is @xmath42 where @xmath43 , and @xmath44 thus , for @xmath45 in @xmath46 units , which corresponds to the ( + ) amplitudes , the lorentz structure of the effective coupling is very simple @xmath47 where @xmath48 and @xmath49 is the w - boson s relativistic four - velocity .
the interest in these particular couplings arose as a by - product of a consideration [ 6 ] of future measurements of competing observables in @xmath50 decay .
in particular , we considered the sm s the @xmath51 coupling values of helicity decay parameters versus those for
@xmath41 @xmath52 single additional lorentz structures .
" it was found that versus the sm s dominant l - handed @xmath13-quark amplitudes , there are two dynamical phase - type ambiguities " produced respectively by an additional @xmath53 coupling and by an additional @xmath3 tensorial coupling , see the @xmath54 and @xmath55 columns of table 1 .
such a dynamical - ambiguity produced physically by the additional lorentz structure is to be contrasted to the mathematical forcing of a phase - ambiguity " by simply changing by - hand the sign of one , or more , of the four helicity amplitudes @xmath56 . by tuning the effective - mass - scale associated with the additional coupling constant ,
the additional @xmath53 coupling , @xmath36 coupling , has respectively changed the sign of the @xmath57 , @xmath58 amplitude . in @xmath59 units ,
the corresponding effective - mass scales are @xmath60 , @xmath61 .
the numerical patterns shown in the table in the case of the additional @xmath62 coupling are not surprising for the @xmath53 coupling because it only contributes to the @xmath20 amplitudes .
however , associated with the additional @xmath3 tensorial coupling , labeled @xmath36 in this table , three interesting numerical puzzles arise at the @xmath63 level in the ( + ) amplitudes versus the sm s pure @xmath41 amplitudes .
the 1st puzzle is that the @xmath64 amplitude has the same value as the @xmath65 amplitude in the sm ; see the corresponding two 220 " entries in the top of table 1 . from the empirical t - quark and w - boson mass values , the mass ratio @xmath66 .
this can be compared with the puzzle s associated mass relation @xmath67 @xmath68 which follows by setting @xmath69 and then expanding in @xmath70 the @xmath64 amplitude , with @xmath71 $ ] in @xmath72 units . since empirically @xmath73 ,
there is only a 4th significant - figure correction from the finite b - quark mass to the only real - valued solution @xmath74 ( @xmath75 ) of this mass relation .
the @xmath63 level of agreement of the two 220 " entries of table 1 is due to the present central value of @xmath76 , and to the central value and @xmath77 precision of @xmath78 .
the error in the empirical value of the mass ratio @xmath79 is dominated by the current @xmath80 precision of @xmath81 .
the 2nd and 3rd numerical puzzles are the occurrence of the same magnitudes of the two r - handed b - quark amplitudes @xmath82 for the sm and for the additional @xmath37 tensorial coupling .
this is shown in the @xmath83 and @xmath84 columns in the bottom half of table 1 .
as explained below , for @xmath45 the magnitudes of these two r - handed moduli are actually exactly equal and not merely numerically equal to the @xmath85 level .
we will next discuss different types of helicity amplitude relations involving both the standard model s decay helicity amplitudes , @xmath1 , and the amplitudes @xmath86 in the case of an additional @xmath3 tensorial coupling of relative strength @xmath8 .
these relations in some cases explain " and in other cases analytically realize as theoretical patterns , these and other numerical puzzles of table 1 .
helicity amplitude relations of types ( i ) and ( ii ) are exact ratio - relations holding for all @xmath87 and @xmath88 values , . by the type ( iii )
ratio - relations holding for all @xmath89 and @xmath90 values , the twb - transformation @xmath91 @xmath92 where @xmath93 @xmath94 characterizes the mass scale @xmath95 .
the parameter @xmath96 is the velocity of the @xmath12-boson in the t - quark rest frame .
somewhat similarly , the amplitude condition ( iv ) , @xmath97 with @xmath98 , and the amplitude condition ( v ) , @xmath99 with @xmath100 , determine respectively the scale of two additional 4x4 twb - transformation matrices @xmath101 and @xmath102 .
thereby , ( iv ) and ( v ) characterize the values of the mass ratios @xmath103 and @xmath104 .
@xmath105 denotes small corrections .
it is not understood why the values are simple for the parameters @xmath106 and @xmath13 .
the first type of ratio - relations holds separately for @xmath107 , @xmath38 ; ( i ) : @xmath108 the second type of ratio - relations relates the amplitudes in the two cases ( ii ) : two sign - flip relations , note sign changes of amplitudes in table 1 , @xmath109 @xmath110 and two non - sign - flip relations @xmath111 @xmath112 the third type of ratio - relations , follows by determining the effective mass scale , @xmath8 , so that there is an exact equality for the ratio of left - handed amplitudes ( iii ) : @xmath113 this was the tuning condition used to produce the dynamical phase - ambiguities of table 1 [ 6 ] .
equivalently , @xmath114 follows from each of : @xmath115 @xmath116 @xmath117 alternatively , the value of @xmath8 can be characterized by postulating the existence of a twb - transformation @xmath91 @xmath92 where @xmath93 @xmath94 , with @xmath118 $ ] and analogously for @xmath5 . assuming ( iii ) , the fourth type of relation is the equality ( iv ) : @xmath119 where @xmath98 .
this is equivalent to the velocity formula @xmath120 for @xmath121 . for @xmath122 , ( iv )
leads to the mass relation discussed above , eq.(5 ) .
however , for @xmath122 , ( iv ) also leads to @xmath123 so @xmath124 without input of a specific value for @xmath125 .
but by lorentz invariance @xmath96 must depend on @xmath125 . accepting ( iii ) as exact
, we interpret this to mean that @xmath126 . as shown in ( ii ) , the @xmath127 corrections in @xmath106 , required by lorentz invariance , arise from @xmath128 and @xmath129 .
equivalently , for @xmath106 arbitrary , ( 15 ) can be expressed postulating the existence of a second twb - transformation @xmath130 @xmath92 where @xmath131\ ] ] the above two twb - transformations do not relate the @xmath132 amplitudes with the @xmath133 amplitudes . from ( i ) thru ( iv ) , in terms of a parameter @xmath13 , the equality ( v ) : @xmath134 is equivalent to @xmath135 @xmath92 @xmath136\ ] ] the choice of @xmath137 , gives @xmath138\ ] ] and corresponds to the mass relation @xmath139 = 4.407 ... gev $ ] for @xmath140 .
if one does not distinguish the @xmath38 versus sm indices , respectively of the rows and columns , these three twb - transformation matrices have some simple properties , for details see ( ii ) : the anticommuting 4x4 matrices @xmath141 with @xmath106 arbitrary , and @xmath142 satisfy the closed algebra @xmath143=2\overline{q},[\overline{m},\overline{q } ] = 2\overline{p},[\overline{p},\overline{q}]=2\overline{m}$ ] .
the bar denotes removal of the overall @xmath96 " factor , @xmath144 .
note that @xmath142 is not a twb - transformation ; @xmath142 is obtained from the first listed commutator . including the b matrix with both @xmath106 and @xmath13 arbitrary , the
commutator + anticommutator " algebra closes with 3 additional matrices @xmath145 obtained by @xmath146=2\overline{h}$ ] ; and @xmath147 .
this has generated an additional twb transformation @xmath148 ; but @xmath149 and @xmath150 are not twb transformations .
the elements of the three logically - successive twb transformations are constrained by the exact helicity amplitude ratio - relations ( i ) and ( ii ) . thereby , the type ( iii ) ratio - relation fixes @xmath45 and the overall scale of the twb - transformation matrix @xmath151 .
somewhat similarly , the amplitude condition ( iv ) with @xmath98 and the amplitude condition ( v ) with @xmath100 determine respectively the scale of the twb - transformation matrices @xmath101 and @xmath102 and characterize the values of @xmath9 and @xmath10 .
the overall scale can be set here by @xmath76 or @xmath152 .
from the perspective of further unification " , @xmath78 is more appropriate since its value is fixed in the sm .
the additional @xmath3 tensorial coupling violates the conventional gauge invariance transformations of the sm and traditionally in electroweak studies such anomalous couplings have been best considered as induced " or effective " .
the @xmath153 component corresponds to a
second class current " [ 7 ] .
@xmath153 has a distinctively different reality structure , and time - reversal invariance property versus the first class @xmath154 [ 8 ] . in the present context
, supersymmetry could provide a more general and useful off - shell theoretical framework in which to consider these theoretical patterns of the helicity amplitudes for @xmath155 decay .
form factor effects would naturally occur . in the extant mssm literature ,
see more complete references in ( ii ) , sizable one - loop - level " reductions in the @xmath155 partial decay width have been reported : from sm higgs and additional mssm s higg s there is a small @xmath156 correction .
however , from susy electroweak corrections , there is in [ 9 ] an up to @xmath157 reduction , depending on @xmath158 . from qcd
including some two - loop - level corrections and susy qcd corrections in the summary of [ 10 ] a @xmath159 reduction is reported .
it is to be emphasized that , firstly , the ( + ) partial width considered in the present paper constitutes a very large , net @xmath160 reduction versus the born - level sm value and that , secondly , these cited susy calculations have been for the partial width , so other couplings instead of an additional effective @xmath37 tensorial coupling , might be predominantly responsible for these reported reductions .
empirically , important tests of the physical relevance of the theoretical patterns to the observed top - quark decays are : ( a ) measurement of the sign of the @xmath161 helicity parameter via determination of stage - two spin - correlation observables [ 5 ] for the @xmath162 channel . these values for @xmath163
are essentially the maximal possible deviations since @xmath164 is the kinematic limit .
the differences from @xmath164 are due to @xmath165 .
( b ) measurement of the closely associated @xmath166-violation , @xmath167 [ 6 ] .
@xmath168-violation can occur due to intrinsic time - reversal violation and/or large @xmath169 final - state interactions .
it is very important to exclude sizable @xmath168-violation and/or @xmath170-violation in top - quark decays .
( c ) measurement of the partial width for @xmath171 , e.g. by single top - quark production at a hadron collider [ 11 ] .
the @xmath172 factor which differs their associated partial widths corresponds to the sm s @xmath173 , versus @xmath174 and a longer - lived ( + ) top - quark if this mode is dominant .
we thank experimental and theoretical physicists for discussions , and the fermilab theory group for a visit during the summer of 2002 .
this work was partially supported by u.s .
dept . of energy contract
no . de - fg 02 - 86er40291 .
cdf collaboration , t. affolder , et.al .
, phys.rev.lett . * 84 * , 216(2000 ) ; d collaboration , b. abbott , et.al .
, phys.rev.lett . * 85 * , 256(2000 ) .
nelson , phys .
* d65 * , 074033 ( 2002 ) ; hep - ph/0304198 . atlas technical proposal , cern / lhcc/94 - 43 , lhcc / p2 ( 1994 ) ; cms technical design report , cern - lhcc- 97 - 32 ; cms - tdr-3 ( 1997 ) . pro . of aps
/ dpf / dpb summer study , snowmass-2001-e3020 , econf c010630 ; tesla tech .
design report , desy 2001 - 011 , http://tesla.desy.de/ ; japanese linear collider group , jlc - i , kek - report 92 - 16(1992 ) ; linear collider physics resource book , american linear collider working group , slac - r-570 , http://www.slac.stanford.edu / grp / th / lcbook/. c.a .
nelson , b.t .
kress , m. lopes , and t.p .
mccauley , phys . rev . * d56 * , 5928(1997 ) .
nelson , p. 369 in physics at extreme energies , " eds
. nguyen van hieu and jean tran thanh van , the gioi publishers , vietnam 2001 ; c.a .
nelson and l.j .
adler , jr .
j. * c17 * , 399(2000 ) ; and c.a .
nelson and a.m. cohen , eur .
j. * c8 * , 393(1999 ) .
s. weinberg , phys .
rev . * 112*,1375(1958 ) .
n. cabbibo , phys .
* 12 * , 137(1964 ) ; c.f .
nelson , phys . lett . *
b355 * , 561(1995 ) .
d. garcia , w. hollik , r.a .
jimenez , j. sola , nucl .
phys . * 427 * , 53(1994 ) .
a. dabelstein , w. hollik , c. junger , r.a .
jimenez , j. sola , nucl .
, * b454 * , 75(1995 ) .
s. willenbrock and d. a. dicus , phys . rev .
* d34 * , 155(1986 ) ; c .- p .
ibid._*d41 * , 42 ( 1990 ) ; r.k .
ellis and s. parke , _
ibid._*d46 * , 3785(1992 ) ; g. bordes and b. van eijk , z. phys . * c57 * , 81(1993 ) ; and t. stelzer , z. sullivan , and s. willenbrock , phys . rev . * d56 * , 5919(1997 ) .
table 1 : numerical values of the helicity amplitudes @xmath175 for the standard model and for the two dynamical phase - type ambiguities ( with respect to the sm s
dominant @xmath176 amplitudes ) .
the values are listed first in @xmath177 units , and second as @xmath178 where @xmath179 is the partial width for @xmath0 .
[ @xmath180 ] . | this talk concerns some theoretical patterns of the helicity amplitudes for @xmath0 decay .
the patterns involve both the standard model s decay helicity amplitudes , @xmath1 , and the amplitudes @xmath2 in the case of an additional @xmath3 tensorial coupling of relative strength @xmath4 .
such an additional electroweak coupling would arise if the observed top - quark has a large chiral weak - transition - moment .
the @xmath5 amplitudes are interpreted as corresponding to the observed top - quark decays .
three twb - transformations @xmath6 @xmath7 , are used in simple characterization of the values of @xmath8 , @xmath9 , and @xmath10 .
measurement of the sign of the @xmath11 helicity parameter , due to the large interference between the @xmath12 longitudinal and transverse amplitudes , could exclude such a chiral weak - transition - moment in favor of the sm prediction .
= -5 mm * consequences of a large top - quark chiral weak - moment * + charles a. nelson + _ department of physics , state university of new york at binghamton + binghamton , n.y . 13902 _ + |
cataclysmic variable stars ( cvs ) consist of a roche - lobe filling main sequence star , called the secondary star , orbiting a white dwarf - the primary star .
mass flows from the secondary star through the inner lagrangian point towards the primary star , ultimately leading to the formation of an accretion disc ( ad ) . after the ad s formation ,
a bright spot ( bs ) is formed at the contact point where the gas stream hits the ad .
dwarf novae ( dne ) are a subclass of cataclysmic variables with two different brightness states , quiescence and outburst . in the latter ,
the system usually brightens by 25mag with a time - scale that can vary from tens to hundreds of days .
dne can also show a third brightness state , the superoutburst , where the system brightens by typically an additional magnitude or two compared to normal outburst .
ip peg is the brightest eclipsing ( @xmath5 ) dn above the period gap , and was discovered by @xcite .
it has a 3.8-h orbital period and a @xmath6 magnitude in the range of 1412mag , the extreme values corresponding to quiescence and outburst , respectively . from the first photometric light curves during quiescence and
return to quiescence @xcite , it was evident that the bs produces a prominent orbital hump , when it rotates into view from behind the disc with a single - stepped ingress and double - stepped egress superimposed on a more gradual disc - eclipse .
infrared light curves showed 0.2mag ellipsoidal variations from the secondary star superposed on a deep eclipse of the wd and the bs @xcite .
@xcite , based on the analysis of 39 eclipses , found out that ip peg has a rapidly increasing orbital period .
however , a new study performed by @xcite and combining 49 eclipses along with previously reported ones , reached the conclusion that the orbital period varies sinusoidally with a period of 4.7 years .
this variation they attributed to a third body , a late m dwarf star with a most probable mass of 0.10@xmath7 .
they also showed that the rate of shrinkage in disc radius during decline from outburst is constant during different outburst cycles , following an exponential decay .
a study of the flickering in ip peg @xcite showed that the bs is the dominant source of flickering , with the ad and wd contributing practically nothing .
spectroscopy of the red star @xcite showed tio bands and the nai doublet ( red dwarf features ) , several broad double - peaked emission lines ( disc features ) and a narrow third peak in the emission lines moving in anti - phase with the double peaks ( a chromospheric emission feature ) .
@xcite performed a tio study of ip peg and concluded that tio instead of the nai doublet should be used for radial velocity studies of the secondary star . in this way they found that the contamination from disc emission , which in other studies returned an elliptical orbit , could be avoided and higher signal - to - noise ratio could be achieved .
they gave a final projected orbital velocity of the companion star @xmath8 and refined values of the primary and secondary star masses @xmath9 and @xmath10 , respectively .
covering a complete cycle from quiescence to outburst and quiescence again , the spectral characteristics of ip peg will now be described .
_ quiescence spectra _
@xcite show broad double - peaked emission lines on a blue continuum , a strong orbital hump in the continuum ( but much weaker in the lines ) and a flat balmer decrement indicative of optically thick emission . the trailed spectra show the signature of a prograde rotation of the ad , where the blue - shifted emission is eclipsed prior to
the red - shifted one , a weak s - wave component in the red - shifted wing and a 10% blue / red asymmetry before eclipse , most probably arising from line emission from the bs .
_ early rise - to - outburst spectra _
@xcite show balmer emission , tio absorption bands , ki and nai absorption doublets , a weak s - wave and two bright spots in both doppler and eclipse maps .
the study of @xcite , as well as a re - analysis of their results using an extension of the classical eclipse mapping method @xcite , concluded that the first spot should be identified as the bs while two possibilities existed for the second spot .
it either was the beginning of formation of a spiral arm or a second inner bs due to gas - stream overflow as predicted by , e.g. , simulations performed by @xcite .
_ rise - to - outburst spectra _
@xcite show a decrease in the equivalent width of h@xmath4 emission line , down to a level similar to that observed immediately after outburst and , simultaneously , weak h@xmath4 emission from the secondary due to irradiation of the secondary star by the boundary layer .
_ near - outburst spectra _
@xcite show unusually strong high - excitation lines , balmer lines in emission , indications of a very disturbed ad with a non - uniform he emission distribution , an outflowing wind from the wd vicinity and chromospheric emission from the secondary .
_ outburst spectra _ as presented by @xcite show : i ) strong balmer and heii @xmath14686 double - peaked emission lines rising from the ad , ii ) non - keplerian emission filling in the centre of heii @xmath14686 , probably due to a compact outflowing wind or inflowing magnetic accretion columns from close to the wd , iii ) balmer but not heii emission from the secondary , located in the polar region on the side facing the ad , iv ) no evidence for an enhanced gas - stream compared to quiescence , v ) large azimuthal asymmetry in the outer disc , with the blue - shifted side of the disc being brighter and possibly more extended at phase 0 and vi ) a sharp transient component in the heii line possibly generated from the interaction of the gas stream and disc .
@xcite suggested that he emission can be explained as photoionisation of the disc by the boundary layer and found that balmer emission in outburst can be attributed by 3/4 to the disc , 1/4 to irradiation of the red star and also to boundary layer irradiation .
another study performed by @xcite showed , from trailed spectra and doppler maps , asymmetric disc emission , strong secondary star emission and a low velocity emission in the h@xmath4 line only .
this stationary emission , appearing as an emission blob in the centre of the doppler maps , was interpreted as slingshot prominences , material trapped in magnetic loops and co - rotating with the secondary star . during another outburst maximum , @xcite found metal line emission such as mgii .
doppler tomography of the he and metal lines locates those lines emission on the inner roche lobe of the secondary star .
_ early decline - from - outburst spectra _
@xcite reveal for the first time in such strength a blend of high - excitation lines of heii , ciii and niii along with a very strong chromospheric emission - line component from the secondary .
_ decline - from - outburst spectra _
@xcite show irradiation - induced emission from the companion star in balmer , he , mgii and cii lines .
this emission , located only near the poles of the secondary , suggests shielding of the equatorial regions of the secondary star by the vertically extended ad .
_ early - quiescence spectra _ in the infrared @xcite reveal the existence of mirror eclipses in ip peg .
a mirror eclipse is an eclipse of the secondary star by an optically thin ad and appears in the trailed spectra as a reduction in the equivalent widths of the lines , at phase 0.5 .
a mirror eclipse also has a mirror - symmetry with the classical rotational disturbance , i.e. the red - shifted part of the line is eclipsed prior to the blue - shifted one . even though ip peg had shown many peculiarities and novelties in the studies cited previously
, it had to reveal another breakthrough .
@xcite were the first to find convincing evidence for the existence of spiral structure in the ad of a close binary .
spiral waves had long been predicted in studies of tidal interactions and numerical simulations ( see @xcite and references therein ) and may play a crucial role in our understanding of the angular momentum transport .
@xcite obtained spectrophotometric observations of ip peg during the rise to an outburst , covering h@xmath4 and hei @xmath16678 .
disc emission was centred on the wd and had a strong azimuthal asymmetry in the forms of two spiral arms .
the repeatability of the two - armed spiral structure was confirmed during another outburst maximum .
@xcite applied doppler tomography of the heii @xmath14686 line and clarified the shock nature of the spiral arms which showed a jump of a factor of more than 2 in intensity .
additionally they calculated an azimuthal extent of 90@xmath11 , a 30@xmath11 upper limit in opening angle and a 15% contribution of the shocks to the total disc emission .
@xcite showed that the spiral pattern persists even 5 and 6 days after the outburst maximum , without showing any diminution in strength .
time - resolved optical spectra obtained during late outburst by @xcite reveal , through the eclipse mapping technique , that spirals are present even 8 days after the onset of an outburst .
their study showed a similar contribution to the continuum emission , as that measured during the peak of the outburst but also a retrograde azimuthal rotation of 58@xmath11 compared to outburst maximum .
@xcite performed a re - analysis of previous spectroscopic data of ip peg during outburst @xcite , based on the eclipse mapping technique .
the difference in their new study was based on the fact that they separated the inner and outer regions of the emission line light curves , obtaining in this way the `` core '' and `` spiral arms '' lights curves , respectively . in this way they improved the previously noisy view of the spiral arms and
were able to trace them in detail .
the two - armed spiral structures were clearly visible in all their eclipse maps , albeit the `` blue '' arm appearing farther out from the disc s centre than the `` red '' one . however , their crucial findings were the sub - keplerian velocities along the spiral structures , as well as the clear correlation between the opening angle of the spirals and the outburst stage ( the later in the outburst , the smaller the opening angle of the spirals ) .
both these findings greatly favour the interpretation of the spiral asymmetric structures as tidally - induced spiral shocks instead of mere radiation patterns on an otherwise unperturbed ad .
a study performed by @xcite presents a review of the spiral structure that far , and plots in detail the azimuthal angles and extent of the spirals for the heii @xmath14686 emission line during an outburst presented by @xcite . in more detail
, it is shown that the two spirals can be traced for almost 180@xmath11 , while their velocity varies from 495 to 780kms@xmath12 .
under the assumption of keplerian velocities , they show that the arms cover a substantial part of the disc , 0.30.9 times the distance to the inner langragian point @xmath13 .
it is pointed out that tracing the spiral arms and plotting their position and intensity as a function of azimuth throughout an outburst , or even better applying this procedure to a range of emission lines simultaneously , can yield important results on the evolution of the spirals . a future paper , as stated in the discussion section ,
is intended to address this issue .
a first approach on tracing the spiral arms of two different emission lines ( h@xmath4 , hei @xmath16678 ) , as well as plotting their emissivity as a function of azimuth , is presented in @xcite .
it is shown that , even though there is a similar position for the spirals in terms of velocity as a function of azimuth , the intensity modulation is significantly different .
our work is part of a project of studying cvs through high resolution time - resolved echelle spectra and applying the indirect imaging technique of doppler tomography .
the simultaneous study of several emission lines can yield important results on the ad structure , as presented for another cv system in @xcite . following the introduction on ip peg one can conclude on the benefits of applying such an analysis to this object , especially since it was observed during outburst . in this way we can study the behaviour of the spiral arms and map the structure of the ad in detail .
the application of modulation doppler tomography , an extension of the classic doppler tomography method @xcite , gives us the opportunity to map emission sources which can vary harmonically with the orbital period .
such an analysis is a first for ip peg .
this paper is organised as follows : in sect . 2 we present our observations , in sec .
3 the data reduction , in sect .
4 we perform the spectral analysis and sect . 5 deals with the application and results of modulation doppler tomography .
last , sect .
6 is devoted to the discussion and summary of our results .
the observing run was conducted on the @xmath14 of august 1999 , at eso s ntt .
echelle spectroscopy was performed using the remd mode of the emmi spectrograph @xcite .
the echelle grating # 9 and cross - disperser # 3 resulted in 21 orders covering the wavelength range 40007500 .
the dispersion increased between 0.09 and 0.17pix@xmath12 , from blue to red , respectively .
ip peg was observed for 5.7h , covering 1.5 binary orbits . in total , 41 spectra with an exposure time of 300sec each , were obtained .
[ f_aavso ] shows a long - term light curve of our system as generated from the aavso light curve generator tool , choosing only visual - validated data to appear , and a window of a few days around outburst .
an inspection of the outburst maximum reveals that ip peg was at the peak of its outburst just 1 day before our observing run , when it was also 1 magnitude brighter .
the resulting spectra were reduced using the standard echelle reduction procedures in _ iraf _ and the slit spectra reduction task doecslit .
whenever , both an interactive and non - interactive mode existed for a task , the interactive one was always chosen .
initially , the bias and lampflat frames were combined into one bias and flat frame with the zerocombine and flatcombine tasks , respectively .
the spectra were then de - biased and a normalised flat was produced with the task apflatten in order to apply the flat - fielding process .
apflatten , after finding and tracing all apertures ( defined as the sum of dispersion lines containing the signal of an order ) , also models both the profile and overall spectrum shape and removes it from the flat - field leaving only the sensitivity variations .
the resulting normalised flat , if plotted perpendicular to the dispersion gives alternating smooth and noisy sections , smooth sections corresponding to the regions between the orders and having a value of 1 , and noisy sections corresponding to the regions inside the orders and having values fluctuating around 1 .
the spectra were then divided by the normalised flat and thus flat - fielded . at this stage
, the spectra were also corrected for cosmic rays with the use of the cosmicrays task , part of the crutil package .
the task doecslit connects and combines several individual tasks to provide a single , complete data reduction path .
it provides a degree of guidance , automation and record keeping . additionally , the user with the choice of the interactive mode ( used in this set of spectra ) and the use of a graphical interface , is able to be in control of the whole procedure , and if necessary , fine - tune the reduction process while it runs .
the several steps involved in doecslit and therefore applied in the dataset are summarised as follows : * in an aperture reference image , preferably a bright object , the orders were located and given a default fixed aperture width , which could be changed interactively if needed . * the background sample regions were inspected .
the function fitting parameters could be set interactively in order to best subtract the background later on .
* apertures were then defined for all object and standard star images , using as initial apertures , those defined from the aperture reference image .
all images sequentially passed through this step .
possible deviations between the aperture reference image and the rest of the images , were fixed interactively by re - tracing and re - centering the apertures of the images . * the actual extraction of the spectra
was done using the _ variance _ option in the _ weights _ parameter of the sparams task of the echelle package .
variance weighting , often called optimal extraction , is performed by summing the pixel values across each aperture at each point along the dispersion .
each pixel in the sum is weighted by its estimated variance based on a spectrum model and the detector readout noise and gain parameters . for detailed information on the _ optimal _ extraction method
the reader should refer to @xcite and @xcite , whose work was followed by _
iraf_. * dispersion correction was then applied through the use of thorium - argon arc lamp spectra .
initially , a few arc lines were identified in each order ( beginning , end and middle of the dispersion axis ) , making use of the th - ar line atlas provided by eso for the specific _ emmi _
setup used in the observing run .
additional lines were then automatically identified with the use of an identification th - ar line list available in _
iraf_. a dispersion function was then fit to the total of identified lines in each order , and then applied to the corresponding orders of the rest of the images .
the typical root mean square of the fits was @xmath150.012 . * for the flux calibration we used spectra of the spectrophotometric standard ltt 7379 .
first , the bandpass regions were interactively defined and then the standard calibration data of ltt 7379 , available in _ iraf _ , were interpolated to the user - defined bandpasses .
the flux values for each aperture were then fit with a smooth curve , in order to relate system sensitivity to wavelength for each aperture , thus resulting in a sensitivity curve for every aperture . *
finally the standard star spectra were extinction corrected ( using the extinction file provided by eso for the location of the la silla observatory ) , and flux calibrated using the derived sensitivity curves . as a last step in doecslit ,
all previous steps were applied to all object spectra , resulting in the final flux - calibrated spectra of ip peg .
the average out - of - eclipse ( phases 0.150.85 , where phase 0 is that of mid - eclipse ) spectrum is shown in fig .
[ f_avgsp_ippeg ] . to check that spectra of incomplete orbits do not affect the line profiles of the resulting average out - of - eclipse spectrum we first computed and compared the average out - of - eclipse spectrum of the whole dataset ( 1.5 orbits ) and that of one full orbit .
the spectrum resulting from the whole dataset differed only in the balmer lines which showed an enhancement of the blue - peak intensity .
the rest of the lines showed minimal or no differences at all .
therefore , the out - of - eclipse spectra of one full orbit were used in fig .
[ f_avgsp_ippeg ] .
the orbital phases of the individual spectra were calculated according to the linear orbital ephemeris given by @xcite : @xmath16=2445615.4224 + 0.^{\rm d}15820616 e\ ] ] where @xmath17 is the time of mid - eclipse .
.characteristics of emission lines .
[ cols= " < , < , < , < " , ] , while top x - axis gives it as a factor of the volume radius of the primary roche lobe @xmath18,width=510 ] the outburst spectrum ( fig .
[ f_avgsp_ippeg ] ) is characterised by several double - peaked emission lines , along with absorption features .
there is strong balmer and high - excitation emission from heii @xmath19 and the ciii - niii blend . also present
is weaker emission from hei ( @xmath14143 , 4388 , 4471 , 4713 , 4922 , 5016 , 5048 , 5876 , 6678 , 7065 and 7281 ) , heii ( @xmath14199 , 4541 and 5411 ) and cii @xmath14267 lines .
ip peg also shows emission from the heavier elements of feii ( feii 42 triplet and feii @xmath15169 , @xmath15303 ) , tiii @xmath14418 and mgii @xmath14481 . finally , there are the nad doublet and tio absorption bands .
all emission lines appear with symmetric line wings , except the lines of the blue part of the spectrum where one can not tell with certainty due to their severe blending with neighbouring lines .
concerning the blue - to - red peak strength , the balmer lines seem to differentiate from the high - excitation lines .
they have an equal blue - to - red - peak strength ( except h@xmath20 which has a stronger red peak ) while the high excitation lines ( blended or stand - alone ) appear to have a stronger blue peak . for the hei lines it is harder to distinguish between the strength of the peaks , as they are either located in the wings of stronger lines ( @xmath14143 , @xmath14388 , @xmath14713 ) or blended with equally strong lines ( @xmath14471 , @xmath14922 , @xmath15016 , @xmath14143 ) .
however , the stand - alone hei lines ( @xmath15876 , @xmath16678 , @xmath17065 ) point to a similar behaviour with the balmer lines . in this study
, we will focus on the 9 most prominent emission lines i.e. the balmer lines , heii @xmath19 and hei ( @xmath14922 , 5876 , 6678 , 7065 ) .
table [ t_lin_ippeg ] shows the characteristics of the emission lines ( full width at zero intensity -fwzi , equivalent width -ew , and integrated flux ) .
however , caution should be drawn in the interpretation of these values for the severely blended lines , namely h@xmath3 , h@xmath20 , heii @xmath19 and hei @xmath21 .
the balmer decrement from h@xmath4 to h@xmath3 1:0.77:0.68:0.55 shows a mix of optically thick and thin conditions . assuming a keplerian velocity field and ad location of line emission
, we can then estimate the radial extension of the emission lines given the fact that the velocities measured from the double - peak separation and fwzi are located in the outer and inner disc radius , respectively . using equations ( 2 ) and ( 3 ) from @xcite and taking the tidal effect into account the resulting radial extension of the emission lines
is shown in fig .
[ f_radii_ippeg ] .
the bottom x - axis is measured in terms of the wd radius @xmath22 , while the top x - axis in terms of the volume radius of the primary roche lobe @xmath18 .
the system parameters used to derive the inner and outer radius , @xmath23 and @xmath24 , are the ones published by @xcite .
the parameters @xmath22 and @xmath18 were calculated according to equations ( 2.82 ) and ( 2.4c ) , respectively , in @xcite . a look at fig .
[ f_radii_ippeg ] reveals a variety of behaviours .
the balmer line emission commences from close to the wd , @xmath25 and extends up to @xmath26 for h@xmath4 . both @xmath23 and @xmath24
gradually increase from h@xmath27 to h@xmath4 , albeit much more abruptly in @xmath28 .
the high - excitation line of heii @xmath14686 also appears from close to the wd but is constrained to a smaller outer radius @xmath29 and therefore is the most centrally located line .
the hei @xmath17065 , 5876 lines have a similar @xmath30 and @xmath31 but hei @xmath14922 appears more centrally located , most probably due to its blending with the feii line which also causes an asymmetry of the blue - to - red peak not present in the rest of the he lines as seen in fig .
[ f_avgsp_ippeg ] .
however , hei @xmath16678 displays a different behaviour , being more centrally located and having a greater radial extension of @xmath32 , compared to the rest of the hei lines that are more confined with an extension of @xmath33 .
non - uniform he distribution was also reported by @xcite , when ip peg was observed near outburst maximum .
after normalising the spectra , and choosing the regions of interest , the trailed spectra of the 9 emission lines were computed .
they are ordered according to orbital phase and shown in fig .
[ f_trsp_ippeg ] . in order to enhance the emitting regions ,
the corresponding spectrograms are presented in fig .
[ f_spec_ippeg ] , with black corresponding to maximum emission .
they have been folded on phases 01 and repeated over two cycles .
both figures show characteristics typical of ip peg : ( i ) prograde rotation of a disc - like emitting region , with the blue peak being eclipsed prior to the red one , ( ii ) double - peaked emission lines signature of an ad ( ii ) highly phase - dependent double - peaked separation , ( iii ) an s - wave moving in anti - phase with the peaks , signature of chromospheric emission from the secondary star .
additional features , seen around @xmath34 in the spectrograms of some emission lines , are caused by blendings with weaker emission lines in the red - wings of the principal lines .
each spectrogram has been plotted in a scale that best enhances its emitting regions , and therefore the intensities of the different emission lines should not be directly compared . at this point , the application of the modulation doppler tomography , presented in the next section , will permit us to gain further insight of the ad structures .
modulation doppler tomography , developed by @xcite , is an extension of the classical doppler tomography which relaxes one of its fundamental axioms , that the flux from any point is constant in time .
this allows the mapping of time - dependent emission sources , and this variability information is not considered in standard doppler tomography studies .
observations show that emission sources can modulate in time and this can happen either because there is anisotropic line - emission or because all sources are not equally visible at all times due to the system s geometry or both .
modulation doppler tomography gives three output images ; one describes the average non - varying flux distribution and the other two describe the sine and cosine term of the harmonically varying emission sources as a function of orbital phase . for a detailed description of modulated doppler tomography , the implementation of the modmap code and test reconstructions
please refer to @xcite and references therein .
the resulting images after running modmap , for the 9 most prominent lines , are shown in fig .
[ f_maps1 ] . from top to bottom and from left to right each panel shows : the observed and predicted data ( the latter calculated from the constant and variable components represented by the maps ) , the non - variable emission distribution and the total variable emission component and finally the individual sine and cosine terms of the variable component .
the input and predicted line profiles are plotted in the same scale , reflected by the colour wedge on their right .
the average non - varying and total modulated emission maps are plotted in their own scale , with the colour wedge reflecting the scale of the modulated map .
the scale is given as a percentage of the average emission to easily depict how strong the modulation is , relative to the constant component .
the cosine and sine amplitude maps are also plotted in their own ( common ) scale , reflected by the colour wedge on their right , again as % of the average emission , in order to ascertain their relative amplitudes .
the roche lobe shape and ballistic stream trajectory have also been computed by the modmap code , for the mass ratio and binary separation of ip peg , and are over - plotted for reference .
the average non - varying - emission maps of all lines show the characteristic spirals plus donor star emission that is well known for ip peg .
the balmer lines show strong modulation of the donor star , mostly attributed to its cosine term , while h@xmath4 also shows modulated emission near the stream trajectory .
the modulation of the donor star is at a 40% level of the average emission in h@xmath3 and decreases when moving to higher balmer series reaching about 1% for h@xmath4 .
there appears to be some modulated disc emission in h@xmath4 though only at a 12% level .
no modulated flux , at the signal - to - noise ratio ( @xmath35 ) of our data , is seen from the spiral arms .
the hei lines reveal a strong component at the position of the secondary star which is of course modulated .
two disc regions appear to show some modulation , though only at the 1 - 2% level .
they also show a modulated component at the inner lagrangian point , albeit of low level compared to average emission .
the only exception is hei @xmath14922 ( see the cosine map ) but one should note that this line is blended with feii relatively strong in our set of spectra .
again , no significant amount of modulated flux from the spiral arms was needed to reproduce the observed data .
finally , the high - ionisation line of heii @xmath14686 also shows the characteristic spiral arms and secondary star emission like the rest of the lines and similarly to the maps doppler maps presented by @xcite .
the nearby bowen blend causes some contamination , most likely the cause of the modulated structure .
it is the only line in which there is a hint of modulated emission from the lower spiral arm at a 6% level .
this study presents the first time - resolved echelle spectra of ip peg , one day after the peak of an outburst .
our spectra show the ensemble of the characteristics previously seen in spectroscopic studies of ip peg , albeit not covering all lines together . the spectroscopic results , deduced from the average out - of - eclipse spectra are analytically given in sec .
5 and can be summarised as follows : * strong balmer double - peaked emission arising from the ad .
the lines show an equal blue - to - red - peak strength , as opposed to the higher excitation lines that have a stronger blue peak , therefore pointing to a less uniform emission .
the balmer decrement shows a mix of optically thin and thick conditions .
the radial extension of the emission originates close to the white dwarf and extends up to @xmath36 for h@xmath4 .
the outer radius of the radial extension gradually increases when moving from h@xmath3 to h@xmath4 . * strong double - peaked emission from the high excitation line of heii @xmath14686 with a stronger blue peak .
it is the most centrally located emission with its outer radius reaching @xmath37 .
* weaker hei and heavier elements ( feii , tiii and mgii ) double - peaked emission . out of the four hei lines ( @xmath17065 , 6678 , 5876 , 4922 ) , hei @xmath16678 is the most centrally located and shows the greater radial extension .
the trailed spectra and spectrograms , reveal the prograde rotation , the double peaked emission lines arising from the disc , an s - wave component indicative of chromospheric emission from the secondary star and highly phase - dependent double - peaked separation , indicative of a distorted ad resulting in two asymmetric emission components , the spiral waves .
all of these features , are present and obvious in the constructed doppler maps .
we have applied to our data set , the modulation doppler tomography method , which in addition to the classic doppler maps reveals if any harmonically varying components are present in the system configuration .
this is of great importance in sight of the spiral waves , which nevertheless have not shown any significant harmonically varying emission and therefore we can not claim anisotropic emission .
the only line to show a hint of anisotropic emission , is the heii line , and only in the lower spiral arm and at a 6% level of the average emission , but this could be due to the blending of the line .
however , the fact that for the majority of lines no modulated emission is seen from the spiral arms does not mean that the spirals do not modulate at all , but that no significant modulated contribution was needed to reproduce the observed data . at higher
@xmath35 we may have seen some modulation .
but , since we see features of amplitude a few % , this means that our @xmath35 rules out strong anisotropy in disc emission . we could therefore conclude that , if present , the modulated emission of the spiral arms is relatively weak at that our @xmath35 was good enough to put a lower detection limit of any modulated emission at 56% .
the balmer and hei lines show strong secondary star emission , part of which is modulated .
its level of emission decreases when moving from h@xmath3 to h@xmath4 . as stated in the previous section secondary star emission
is naturally modulated with the cosine term .
this is normal since irradiation induced emission from the secondary is phase dependent , so it is natural to see a flux variation when the phase is changing .
given equation 3 of @xcite and that at phase 0.50.7 , the face on secondary should emit more , then the cosine term is the dominant component .
all hei lines show modulated emission in 2 regions of the ad , but are at a very low level . only in @xmath14922
do they appear more pronounced .
these ad modulated regions , seem to appear in the `` gaps '' between the spiral arms and modulate with the sine - term
. this could be in accordance with the spiral arms shielding the disc areas just behind them and hiding a possible cosine - term of disc modulated emission as well .
moreover , the sine - term is expected considering their phasing is in between the spiral arms . geometric shielding caused by the spirals
was also proposed from @xcite and shielding of the secondary star by the ad was proposed by @xcite . finally , most of the hei lines show modulated emission near the inner lagrangian point also covering a small region at the beginning of the stream trajectory .
though , shows pronounced modulated emission for the bigger part of the stream trajectory . in a future paper
it is our aim to study the same dataset , but concerning the spiral arms of the different lines only .
tracing , providing the amplitude , and the contribution to the total flux of the spiral arms of all lines , as well as finding the ratio of the brightness between the two spiral arms can yield valuable results for this unique set of time - resolved spectra .
* fig3.dat : an ascii file that gives the radial extension of the emission lines as a factor of both the white dwarf radius @xmath22 and the volume radius of the primary roche lobe @xmath18 , as seen in fig . 3 .
* hjd_phase.dat : an ascii file giving the heliocentric julian date and the orbital phase of the 41 spectra .
* spectra : a folder that contains , * * the average out - of - eclipse spectrum of fig . 2 as a gzipped fits file , named + ippeg_avr.fits.gz and * * the normalised trailed spectra of the 9 most prominent emission lines as shown in fig .
these trailed spectra were the ones used for the reproduction of fig .
the naming convention is \{emission line}_\{sequence number}.fits.gz , resulting in 9 ( number of emission lines)@xmath3841 ( number of spectra ) different files .
c.p . gratefully acknowledges a doctoral research grant by the belgian federal science policy office ( belspo ) .
, h. m. j. 2001 , in lecture notes in physics , berlin springer verlag , vol.573 , astrotomography , indirect imaging methods in observational astronomy , eds .
boffin , h. m. j. and steeghs , d. and cuypers , j. , p. 69 , h. , delabre , b. & dodorico , s. 1986 , presented at the society of photo - optical instrumentation engineers ( spie ) conference , vol.627 , instrumentation in astronomy vi ; proceedings of the meeting , tucson , az , mar .
4 - 8 , 1986 .
part 1 ( a87 - 36376 15 - 35 ) .
bellingham , wa , society of photo - optical instrumentation engineers , 1986 , p. 339
- 348 , d. 2001 , in lecture notes in physics , berlin springer verlag , vol.573 , astrotomography , indirect imaging methods in observational astronomy , eds .
boffin , h. m. j. and steeghs , d. and cuypers , j. , p. 45 | we analyse a unique set of time - resolved echelle spectra of the dwarf nova ip peg , obtained at eso s ntt with emmi .
the dataset covers the wavelength range of 4000 - 7500@xmath0 and shows balmer , hei , heii and heavier elements in emission .
ip peg was observed one day after the peak of an outburst .
the trailed spectra , spectrograms and doppler maps show characteristics typical of ip pegasi during the early stages of its outburst , such as prograde rotation from the accretion disc flow , chromospheric emission from the secondary and spiral arms .
the high - ionisation line of heii @xmath14686 is the most centrally located line and has the greatest radial extension compared to the hei lines .
the balmer lines extend from close to the white dwarf up to @xmath2 , with the outer radius gradually increasing when moving from h@xmath3 to h@xmath4 .
the application , for the first time , of the modulation doppler tomography technique , maps any harmonically varying components present in the system configuration , and this variability information is not considered in standard doppler tomography studies .
we find , as expected , that part of the strong secondary star emission in balmer and hei lines is modulated predominantly with the cosine term , consistent with the emission originating from the irradiated front side of the mass - donor star , facing the accreting white dwarf . for the balmer lines
the level of the modulation , compared to the average emission , decreases when moving to higher series .
emission from the extended accretion disk appears to be only weakly modulated , with amplitudes of at most a few % of the non - varying disk emission .
we find no evidence of modulated emission in the spiral arms , which if present , is relatively weak at that our signal - to - noise ratio was good enough to put a lower detection limit of any modulated emission at 56%
. only in one arm of the heii @xmath14686 line , is there a possibility of modulated emission , but again , we can not be sure this is not caused by blending with the nearby bowen complex of lines . |
the preceding letter of this volume is devoted to the measurement of the @xmath6meson parameters in the kedr experiment performed during energy scans from 3.67 to 3.92 gev at the vepp-4 m @xmath7collider . in this letter
we describe the application of the developed tools to the measurement of @xmath0parameters omitting details common for @xmath6and @xmath0 .
since the discovery of the @xmath0 , seven experiments contributed to the determination of its parameters , nevertheless the situation with the mass , total width and electron partial width is still not clear .
the incomplete compilation of results reported on @xmath0mass is presented in table [ tab : mcomp ] .
it does not include the results of refs .
@xcite with the analysis of the @xmath8 data of bes @xcite and the @xmath9 data of belle @xcite in which the @xmath0electron width has been fixed in the fits causing a mass bias .
in addition , the bin size in belle data around @xmath0 seems too large for a simple center - of - bin fitting .
these works encouraged us to employ the vector dominance model in the analysis @xcite .
the values presented form three partially overlapping clusters .
the first one with @xmath10 comes from the analyses in which interference between resonant and nonresonant @xmath1production has been ignored @xcite .
the statistical uncertainty in this case is small ( in @xcite the influence of @xmath11 and higher @xmath12 s included in the analysis increases the @xmath0mass uncertainty ) .
the second cluster of @xmath13 analyses @xcite has @xmath14 ( the result of @xcite is not included because of its uncertain status ) .
the third , highest mass , cluster is formed by the analyses accounting for interference @xcite and gives @xmath15 . as was mentioned in section 5.2 of the previous letter ,
taking into account the resonance
continuum interference is essential for a determination of the @xmath0parameters .
a close @xmath1production threshold significantly increases the importance of that .
a consideration of the interference effects is one of the primary goals of this experiment .
if interference is ignored in a fit of the measured @xmath1 or multihadron cross section , a bias appears in the growing continuum contribution that causes a bias in the resonance amplitude and a shift of the mass value .
the signs of these effects depend on the relative position of the interference peak and dip .
the @xmath1cross section at the threshold is fixed at zero , therefore the weights of the more distant data points in a fit are larger than those of the less distant ones .
evidence for a dip after the @xmath1cross section maximum is visible in all published data with large enough statistics ( see , for example , fig . 1 of ref .
@xcite ) , therefore , the artificial mass shift should be negative ( undercounted events move the resonance peak to the left ) .
that is exactly what we observe analyzing the published mass results . if the result on mass of @xcite is ignored , the @xmath0mass value obtained in @xmath16 decays does not contradict neither to @xmath17 nor @xmath18 .
the interference of the resonant and nonresonant @xmath19 yields also takes place in this case but the relation between them can differ from that in @xmath7collisions , besides , the interference effect can be partially compensated by subtraction of the combinatorial background .
thus , the intermediate mass value does not seem surprising .
below we briefly describe the theoretical basis of the analysis performed , enter some details concerning the analysis procedure and not covered in , present the results on the @xmath0parameters and discuss their systematic uncertainties and model dependence .
@xmath20[mev ] & comments + mark - i @xcite & @xmath21 & @xmath22 + delco @xcite & @xmath23 & @xmath22 + mark - ii @xcite & @xmath24 & @xmath22 + belle @xcite & @xmath25 & @xmath26 + kedr @xcite & @xmath27 & @xmath28 + bes - ii @xcite & @xmath29 & @xmath22 + bes - ii @xcite & @xmath30 & @xmath31 + belle@xcite & @xmath32 & @xmath33 + babar @xcite & @xmath34 & @xmath35 + babar @xcite & @xmath36 & @xmath37 + kedr @xcite & @xmath38 & @xmath39 + + + + + + +
a few approaches can be employed to determine the resonance parameters using a multihadron cross section data . in the @xcite the fit of the @xmath40 ratio
was performed , in the @xcite the efficiency corrected cross section was analyzed .
there are many different sources of multihadron events such as the @xmath6and @xmath0production , the light quark production etc .
, thus the variation of the net detection efficiency in the whole experiment range can exceed 20% @xcite .
the calculation of the net efficiency implies knowledge of the resonance parameters and accounting for the interference effects , therefore an iterative analysis is required . in this work
we fit the observed multihadron cross section not corrected for the detection efficiency which allows iterations to be avoided . in the energy range from slightly below the @xmath6 peak to slightly above the @xmath42 threshold the variation of the light quark contribution to @xmath40 ( @xmath43 )
is small , so that the multihadron cross section observed in the experiment can be written as @xmath44 where @xmath45 s are theoretical cross sections , @xmath46 s are corresponding detection efficiencies , and @xmath47 s are terms treated empirically as described below .
the @xmath48 superscript means that the cross section has been corrected for initial state radiation ( isr ) effects , @xmath49 stands for the direct @xmath0decay to light hadrons , the other ( super / sub)scripts seem self - explanatory , @xmath50 is a branching fraction .
all detection efficiencies explicitly entering eq . can be kept energy independent with sufficient accuracy for the event selection criteria employed ( see sec . [
sec : deteff ] ) .
the first four terms have no peculiarities in the whole energy range of the experiment , while the last four are responsible for the excess of the cross section in the @xmath0region . the fourth term of eq .
corresponding to the light quark contribution can be scaled as @xmath51 where a relatively small parameter @xmath52 is due to the energy dependence of the detection efficiency and radiative corrections . possible variation of @xmath43 can also contribute to @xmath52 .
this term can be easily removed from the consideration in the fit of the cross section provided that the @xmath52 value is known .
the @xmath42 cross section can be treated as a small correction .
we took it into account using the approximately known shape and an additional fit parameter .
calculations for @xmath53 and @xmath54 are described in , a small contribution of the @xmath55 tail was calculated similarly to the @xmath6one , for the @xmath1production cross section ( here and below @xmath41 stands for @xmath56 or @xmath57 ) one has @xmath58 where @xmath59 is the probability to lose a fraction of @xmath60 in the initial state radiation @xcite , @xmath61 describes a distribution of the total collision energy , which can be assumed to be gaussian with an energy spread @xmath62 .
for the charged mode ( @xmath63 ) the factor @xmath64 describing the coulomb interaction between the mesons produced @xcite is taken according to sommerfeld - sakharov @xcite : @xmath65 for the neutral mode ( @xmath66 ) there is no such interaction , thus @xmath67 the step functions @xmath68 are shown explicitly to simplify some expressions below .
the cross section @xmath69 can be expressed via the form factor @xmath2 and @xmath41-meson velocity in the c.m.system @xmath70 : @xmath71 to determine the parameters of resonances above the @xmath1threshold , their amplitudes should be separated in @xmath2 : @xmath72 where @xmath73 is the phase of the @xmath74-th resonance @xmath75 relative to @xmath76 . for the resonance with the partial widths @xmath77 and @xmath78 and the total width @xmath79 , one has a breit - wigner amplitude @xmath80 ( the vacuum polarization factor is included in @xmath77 ) . considering @xmath81 as a nominal resonance width and introducing the sum of the branching fractions to all non-@xmath1modes @xmath50 , one obtains the energy - dependent @xmath1partial width @xmath82 in line with the pdg prescriptions ( p. 808 of ref .
@xcite ) . here
@xmath83 and @xmath84 are the blatt - weisskopf damping factors for a vector resonance @xcite : @xmath85 where @xmath86 represents the meson radius and @xmath87 is the c.m .
momentum of the meson @xmath88 .
the partial width dependence according to eq . corresponds to the approach of ref .
its simplified form was used in the experiments @xcite .
the approach is somewhat different from that employed in which does not lead to noticeable changes of the @xmath0parameters .
the @xmath42 cross section entering ( [ eq : sigmhobs ] ) as a small correction can be calculated with sufficient accuracy using @xmath89 & \beta_{{\ensuremath{d\overline{d}{}}\xspace}\pi } = \sqrt{(1\!-\!(m_{d^{*}}\!+\!m_d)^2/w^2 ) ( 1\!-\!(m_{d^{*}}\!-\!m_d)^2/w^2)}\ , .
\end{split } \label{eq : cmsbeta}\ ] ] the quantity @xmath90 is treated as a fit parameter .
the nonresonant part of the form factor can be written as @xmath91 with @xmath92 for point - like particles . here
@xmath93 is the vacuum polarization operator except the contributions of all resonances which are written separately in ( [ eq : ff ] ) .
we remind that the full polarization operator is calculated using the total cross section of @xmath94 that already includes all resonances , therefore use of the full operator @xmath95 instead of @xmath93 in the nonresonant amplitude leads to double counting of the resonances and thus incorrect values of the leptonic widths ( see also the discussion in sec .
5.3 of ) .
there are no precise theoretical predictions for @xmath96 .
the model - independent result can be obtained using the expansions of @xmath97 and @xmath98 at the point @xmath99 with the coefficients free in the fit .
our statistics are not sufficient for that , thus we have to rely on some model or use a pure empirical approach as in also taking into account the resonance continuum interference . the most certain prediction of the form factor can be obtained with an application of the vector dominance model ( vdm ) to charm production .
standard vdm assumes that the inclusive cross section @xmath100 at low energy is saturated by the interfering contributions of the limited number of vector mesons .
a similar assumption can be accepted for the inclusive @xmath101 cross section and its exclusive modes such as @xmath102 .
the vdm like analysis of the @xmath40 ratio in the energy range of @xmath103 gev has been performed by bes in ref .
@xcite , where the light quark contribution @xmath43 was calculated using pqcd .
the work cited accounts for @xmath0 , @xmath11 , @xmath104 and @xmath105 resonances but does not account for a possible contribution of @xmath6decays to @xmath1above the threshold .
studies of this contribution refs .
@xcite include a theoretical consideration and some analysis of the @xmath1cross section measured by bes as well as by belle . in this work
we employ vdm in a simplified form @xmath106 where @xmath107 is a real constant representing the contributions of the @xmath11 and higher @xmath12 s .
the @xmath6contribution to the @xmath1form factor @xmath108 was calculated using eq . with the @xmath109 and @xmath110partial widths defined similarly to eq . with a specific value of the effective radius @xmath86 .
the value of @xmath111 at some reference point @xmath112 , as well as the constant @xmath107 , should be obtained from the data fit ( @xmath113 mev was used ) .
the partial width ratio @xmath114 is presumably close to that of @xmath0 . to evaluate the model dependence of the @xmath0parameters we tried a few nonresonant form factor parameterizations , which do not assume vector dominance
the most popular empirical parameterization is probably exponential : @xmath115 where @xmath87 is the c.m .
momentum @xcite .
it is well motivated far above the threshold but has few parameters to describe the low energy region . instead of it
we used @xmath116 the minus sign is chosen to match the @xmath6dominance expectations . in the case
@xmath117 , @xmath118 , the nonresonant cross section acquires the blatt - weisskopf factor with @xmath119 .
the case @xmath120 corresponds to a more rapid form factor fall .
use of two parameters @xmath121 and @xmath122 allows us to take into account in the limited energy range the increase of the @xmath1cross section described by the @xmath123 structure in the alternatively , the dependence on @xmath124 @xmath125 and combined dependences @xmath126 were considered . to check validity of the @xmath6domination hypothesis in eq .
the following parameterizations were used : @xmath127 where @xmath70 is the @xmath41meson velocity .
they are expansions of the breit - wigner amplitude with the mass @xmath128 treated as a free parameter , the values of @xmath129 correspond to different assumptions on @xmath79 dependence . in case of @xmath6dominance
the fitted value of @xmath128 would be close to @xmath130 .
to perform a fit of the observed multihadron cross section with eq . , it is necessary to know six detection efficiencies explicitly entering the equation and the detection efficiency @xmath131 implicitly contained in the term @xmath132 related to the continuum light quark production .
they were determined from monte carlo simulation .
the efficiency @xmath133 enters eq . in the product with the non-@xmath1branching fraction @xmath50 , which is rather uncertain
that allows one to assume @xmath134 . the event selection criteria , which are different for 2004 and 2006 scans , and the procedure of the detection efficiency determination for the @xmath6 decay simulation are described in detail in .
the tuned version of the bes generator @xcite was employed to the obtained @xmath6detection efficiency in the vicinity of the peak .
the same version of the generator with parameters optimal for @xmath6simulation was used to simulate the @xmath6and @xmath55tails and the continuum @xmath135 production . to simulate @xmath136 events ,
@xmath1pairs were first generated with the proper angular distribution .
decays of @xmath41 mesons were simulated using the routine _
lu2ent _ of the _ jetset 7.4 package @xcite .
the decay tables of _ jetset were updated according to those of the pdg review @xcite . _ _
.[tab : def]detection efficiency for the processes of interest and its variation in the experiment energy range @xmath137 mev . [ cols="<,^,^,^",options="header " , ] the main sources of systematic uncertainty in @xmath0parameters are listed in table [ tab::dperr ] . lcccc & @xmath138[mev ] & @xmath139[mev ] & @xmath77[@xmath140 & @xmath141[% ] + + @xmath50 & @xmath142 & @xmath143 & @xmath144 & @xmath145 + @xmath146 value in @xmath79 & 0.3 & 0.3 & 2 . & 1.5 + @xmath147 & 0.1 & 0.1 & 0.4 & 0.8 + @xmath148 masses & 0.06 & 0.04 & 0.3 & 0.5 + @xmath42 cross section & 0.15 & 0.05 & 1 . & 2 .
+ + det .
efficiency variation & 0.03 & 0.04 & 2.4 & 5 .
+ hadronic event selection & 0.3 & 0.3 & 3 . & 5 .
+ residual background & 0.06 & 0.3 & 2.9 & 3 .
+ luminosity measurement & 0.1 & 0.1 & 2 . & 2 .
+ beam energy & 0.03 & & & + _ sum in quadrature _ &
@xmath149 & @xmath150 & @xmath151 & @xmath152 + when the resonance - continuum interference is taken into account , the multihadron cross section becomes rather sensitive to the non-@xmath1fraction of @xmath0decays .
it was varied from zero to 0.16 as was mentioned in section [ sec : fitting ] .
the variations of the @xmath0mass and total width were 0.3 and 0.1 mev , respectively .
the shift of the electron width was + 8.8% for the first solution and -2.3% for the second one . the uncertainty on the @xmath146 value used to specify the energy - dependent width , of about @xmath153 ( ) leads to these of @xmath154 mev both in the mass and total width .
when the interference is ignored , the sensitivity to @xmath86 variations reduces by a factor of 3 .
the due to of the branching fraction ratio for @xmath109 and @xmath110 are about 0.1 mev for the mass and total width .
approximately the same uncertainties are obtained because of the @xmath41 meson masses .
the estimates were obtained by variation of the values within their errors quoted by pdg . to estimate due to the @xmath42 cross section at the edge of the energy range of the experiment , we used two methods : shrinking of the fit range and assumption of the linear dependence on the @xmath41-meson c.m .
velocity instead of the cubical one in eq . .
the latter corresponds to variation of the effective interaction radius @xmath86 for @xmath42 states from zero to infinity .
the variations of the mass , total width and electron width do not exceed @xmath155 mev , @xmath156 mev and @xmath157 , respectively .
the systematic uncertainties due to the energy dependence of the detection efficiencies shown in table [ tab : def ] can be neglected in all cases except @xmath131 .
the latter together with the energy dependence of the radiative correction factor and possible @xmath43 variation determine the power in the expression used to parameterize the light quark contribution to the multihadron cross section .
the radiative correction factor @xmath158 was calculated according to ref .
@xcite using the vacuum polarization data compilation by the cmd-2 group reviewed in ref .
the error quoted includes the uncertainty of the detection efficiency dependence on the mass of the hadronic system produced via isr and that of the vacuum polarization data .
we explicitly considered the @xmath55 tail in the cross section , thus the correction factor is @xmath159% less than that used in and its variation in the experiment energy range does not reach 0.1% .
the precise @xmath40 measurements at @xmath160 and @xmath161 mev @xcite do not indicate essential @xmath43 variation , thus we concluded that the @xmath135 efficiency variation dominates in the uncertainty of the power @xmath162 .
performing the fits with different values of @xmath52 we evaluated the uncertainty of the @xmath0parameters as 0.03 mev , 0.04 mev and 2.4% for the mass , total width and electron width , respectively .
compared to that , the energy dependence of @xmath163 gives only a 0.5% bias of the electron width and a few kev shifts of the mass and total width .
the sensitivity of the mass and width to the criteria of the multihadron event selection was checked by changing cuts on the energy deposited in the calorimeter and conditions on the number of tracks .
the results were stable within 0.3 mev .
the detection efficiency uncertainty due to inaccuracy of the @xmath41-meson decay ratios @xcite used for the simulation contributes @xmath164 to the electron width .
the dependence on the choice of the selection criteria increases it up to 3% .
the sensitivity to the event selection criteria is partially due to the influence of the residual background .
we ignore that and treat the background correction as an independent source which makes the estimates more conservative .
uncertainties due to the luminosity measurement instability are less than @xmath165 mev for the mass and width .
the accuracy of the absolute luminosity measurements discussed in contributes less than @xmath164 to the electron width . the uncertainty on @xmath0mass due to the beam energy determination does not exceed @xmath166 kev .
the parameters of the @xmath0meson have been measured using the data collected with the kedr detector at the vepp-4 m @xmath167 collider .
interference of resonant and nonresonant production essential in the near - threshold region has been taken into account . our final results on the mass and width of @xmath0are : + [ -3pt ] @xmath168 the corrections applied to the fit results are listed in table [ tab : fitcor ] .
the third error arises from the model dependence .
it was estimated comparing the results obtained under the assumption of vector dominance in the @xmath41-meson form factor ( quoted values ) and under a few alternative assumptions which do not imply vector dominance .
the quoted model errors do not include possible deviations of the resonance shape from the breit - wigner one with usual assumptions about the total width energy dependence , which are predicted , e.g. , in the coupled - channel model @xcite .
the result on the @xmath0mass agrees with that by babar also taking into account interference ( ref .
@xcite ) and is significantly higher than all results obtained ignoring this effect .
the mass values obtained studying @xmath16-meson decays by babar @xcite and belle @xcite are lower but do not contradict to our measurement .
we got two possible solutions for the @xmath0electron partial width and the radiatively corrected nonresonant @xmath1cross section at the mass of @xmath0 : @xmath169 the phase shifts of the @xmath0amplitude relative to the negative nonresonant amplitude are @xmath170 and @xmath171 degrees for solutions ( 1 ) and ( 2 ) , respectively .
most of potential models support the first solution and can barely tolerate the second one .
the increase of the @xmath0mass according to the babar and kedr measurements implies the decrease of the @xmath172-@xmath173 mixing used in potential models to rise the electron width value above 100 ev ( refs .
@xcite and the reviews @xcite ) .
the correct choice of the true solution is extremely important for a determination of the non-@xmath1fraction of @xmath0decays .
because of the large uncertainty the solution ( 1 ) does not contradict formally to the previously published results , which do not take the interference effect into account , the solution ( 2 ) is only two standard deviations higher than the current world average .
however , the qualitative consideration and numerical estimates confirm that the impact of the resonance continuum interference on the resulting electron width value is large , therefore the resonance parameters obtained taking into account interference can not be directly compared with the corresponding values obtained ignoring this effect .
we greatly appreciate permanent support of the staff of the experimental , accelerator and electronics laboratories while preparing and performing this experiment . stimulating discussions with v. p. druzhinin
are acknowledged .
this work was partially supported by the ministry of education and science of the russian federation and grants sci.school 6943.2010.2 , rfbr 10 - 02 - 00695 , 11 - 02 - 00112 , 11 - 02 - 00558 .
99 p. a. rapidis _
et al . _ ,
* 39 * ( 1977 ) 526 .
w. bacino _ et al .
* 40 * ( 1978 ) 671 .
r. h.shindler _ et al .
d * 21 * ( 1980 ) 2716 .
k. abe _ et al . _ [ belle collaboration ] , phys .
* 93 * ( 2004 ) 051803 .
k. y. todyshev _ et al . _
[ kedr collaboration ] , pos * hep2005 * ( 2006 ) 115 .
m. ablikim _ et al . _
[ bes collaboration ] , phys .
b * 652 * ( 2007 ) 238 .
m. ablikim _ et al . _
[ bes collaboration ] , econf * c070805 * ( 2007 ) 02 [ phys . lett .
b * 660 * ( 2008 ) 315 ] .
j. brodzicka _ et al . _ [ belle collaboration ] , phys .
* 100 * ( 2008 ) 092001 .
b. aubert _ et al . _ [ babar collaboration ] , phys . rev .
d * 77 * ( 2008 ) 011102 .
b. aubert _
et al . _ ,
d * 76 * ( 2007 ) 111105(r ) .
k. y. todyshev [ kedr collaboration ] , pos * ichep2010 * ( 2010 ) 218 .
h. b. li , x. s. qin and m. z. yang , phys .
d * 81 * ( 2010 )
011501 . y. j. zhang and q. zhao , phys .
d * 81 * ( 2010 ) 034011 .
m. ablikim _ et al . _
[ bes collaboration ] , phys .
b * 668 * ( 2008 ) 263 .
g. pakhlova _ et al . _ [ belle collaboration ] , phys . rev .
d * 83 * ( 2011 ) 011101 m. ablikim _ et al .
_ , phys .
* 101 * ( 2008 ) 102004 .
d. zhang _ et al . _ ,
d * 74 * ( 2006 ) 054012 e. a. kuraev and v. s. fadin , sov .
* 41 * ( 1985 ) 466 .
m. b. voloshin , prog.part.nucl .
* 61 * ( 2008 ) 455 .
a. sommerfeld , atombau und spektrallinien , vol.*ii * ( vieweg , braunschweig,1939 ) . a. d. sakharov , sov . phys .
jetp * 18 * ( 1948 ) 631 .
k. a. milton , i. l. solovtsov , mod .
lett . a * 16 * ( 2001 ) 2213 .
k. nakamura _ et al .
_ , j. phys .
g * 37 * ( 2010 ) 075021 .
blatt and v.f .
weisskopf , theoretical nuclear physics , wiley , new york , ( 1952 ) .
et al . _ ,
. letters * 37 * ( 1976 ) 398 .
et al . _ ,
d * 62 * ( 2000 ) 034003 .
t. sjostrand and m. bengtsson , comp .
* 43 * ( 1987 ) 367 .
et al . _ ,
* 95 * ( 2005 ) 121801 e. eichten _
et al . _ , phys .
d * 21 * ( 1980 ) 203 .
w. buchmuller and s. h. h. tye , phys .
d * 24 * ( 1981 ) 132 .
s. godfrey and n. isgur , phys.rev .
d * 32 * , ( 1985 ) 189 .
m. ablikim _ et al . _ [ bes collaboration ] , phys .
b * 641 * ( 2006 ) 145 .
m. ablikim _ et al . _
[ bes collaboration ] , phys .
b * 677 * ( 2009 ) 239 a. d. bukin ,
arxiv:0710.5627 [ physics.data-an ] . j. d. jackson and d. l. scharre , nucl .
instrum .
* 128 * ( 1975 ) 13 .
m. ablikim _ et al . _
[ bes collaboration ] , phys .
b * 659 * ( 2008 ) 74 .
d. besson _ et al .
_ [ cleo collaboration ] , phys . rev . lett .
* 96 * ( 2006 ) 092002 [ erratum - ibid .
* 104 * ( 2010 ) 159901 ] [ arxiv : hep - ex/0512038 ] .
s. actis _ et al .
_ , eur .
j. c * 66 * ( 2010 ) 585 .
a. m. badalian and i. v. danilkin , phys . atom .
* 72 * ( 2009 ) 1206 .
s. f. radford and w. w. repko , phys .
d * 75 * ( 2007 ) 074031 .
p. gonzlez _ et al .
_ , phys .
d * 68 * ( 2003 ) 034007 .
j. m. richard , z. phys .
c * 4 * ( 1980 ) 211 .
n. brambilla _ et al .
j. c * 71 * ( 2011 ) 1534 .
n. brambilla _ et al . _
[ quarkonium working group ] , arxiv : hep - ph/0412158 . | we report the final results of a study of the @xmath0 meson using a data sample collected with the kedr detector at the vepp-4 m electron - positron collider .
the data analysis takes into account interference between the resonant and nonresonant @xmath1production , where the latter is related to the nonresonant part of the energy - dependent form factor @xmath2 . the vector dominance approach and several empirical parameterizations
have been tried for the nonresonant @xmath3 . our results for the mass and total width of @xmath0are @xmath4 where the first , second and third uncertainties are statistical , systematic and model , respectively . for the electron partial width
two possible solutions have been found : @xmath5 our statistics are insufficient to prefer one solution to another .
the solution ( 2 ) mitigates the problem of non-@xmath1decays but is disfavored by potential models .
it is shown that taking into account the resonance
continuum interference in the near - threshold region affects resonance parameters , thus the results presented can not be directly compared with the corresponding pdg values obtained ignoring this effect . |
the electron density is a fundamental parameter in plasma physics .
knowledge of the three - dimensional ( 3d ) electron density structure is very important for our understanding of physical processes in the solar corona , such as the coronal heating and the acceleration of the solar wind ( _ e.g. _ , ; ) .
the density structure of the corona strongly affects the propagation of cmes @xcite .
the density is also important for estimates of the alfvn mach number and compression rate of cme - driven shocks @xcite , and for the interpretation of solar radio emission such as type ii and type iv radio bursts produced by coronal eruptions @xcite .
the k corona arises from thomson scattering of photospheric white light from free electrons ( _ e.g. _ , ) .
because the emission is optically thin , the measured signal is a contribution from electrons all along the line of sight ( los ) .
the derivation of the electron density in the k corona from the total brightness ( @xmath3 ) or polarized radiance ( @xmath0 ) is a classical problem of coronal physics , first addressed by and . because of difficulties in generally separating the k - coronal component from the f - coronal component arising from interplanetary dust scattering ( _ e.g. _ , ; ) , most of the inversion techniques are practical for @xmath0 measurements . the f - coronal polarization is not very well understood ( see reviews by ; ) .
it is generally accepted that the polarized contribution of the f corona can be ignored within 5 @xmath4 @xcite , but some observations show that the f corona is almost unpolarized even at elongations ranging from 10 to 16 @xmath4 ( , @xcite ; ) .
in addition , the f corona dominates the total brightness of the corona beyond about 4 or 5 @xmath4 and make it difficult to recover the much fainter k - corona emission @xcite . to retrieve the electron density of the corona from a single 2d @xmath0-image , one needs to assume some special geometries for the distribution of electrons along the los .
the previous studies have modeled the electron density distribution in several ways , including the simple spherically symmetric model @xcite , the axisymmetric model @xcite , or the models that take into account large - scale structures , such as polar plumes in the coronal holes @xcite , or active streamers in the equatorial regions @xcite . among the above inversion methods , the spherically symmetric inversion ( ssi ) developed by ( called the _ van de hulst _ inversion thereafter ) , was the most representative and commonly used .
he found that the density integral for @xmath0 signals becomes invertible if the latitudinal and azimuthal gradients in electron density are weaker than the radial gradient ( _ i.e. _ , a local spherical symmetry approximation ) .
this classic inversion technique has been applied to establish the standard density models of the coronal background at equator and pole in the solar minimum and maximum ( see ) and the density models of near - symmetric coronal structures such as streamers and coronal holes @xcite .
the ssi method was also used to analyze detailed density distribution of fine coronal structures observed in eclipses ( _ e.g. _ , ; ) , and to derive the 2d density distribution of the entire corona @xcite , when the spherical symmetry is assumed holding locally .
the importance of the ssi method for coronal density determination has been demonstrated by wide applications of the derived densities such as in testing models of the acceleration mechanism of the fast solar wind @xcite , interpreting sources of type ii and type iv radio bursts @xcite , and determining the coronal magnetic field strengths from fast magnetosonic waves by global coronal seismology @xcite
. however , in contrast to extensive applications , studies on evaluation of the ssi method are few in the literature .
compared the white light densities to those determined from the density - sensitive euv line ratios of si 350/342 observed by soho / cds , and found that densities determined from these two different analysis techniques match extremely well in the low corona for a very symmetric solar minimum streamer structure .
similarly , compared densities of various coronal structures determined by inverting mlso mk4 @xmath0 maps and from the line ratios of o 1032/1037.6 observed by soho / uvcs , and found that the mean densities in a streamer by the two methods are consistent , while the coronal densities for a coronal hole and an active region are within a factor of two .
these results are encouraging , and suggest that the 2d white - light density distribution in coronal structures can be very useful for other studies , but a detailed assessment is required for its better application , such as information about the limitation of the ssi method and the uncertainty of derived densities . to achieve this goal
, one may use synthetic @xmath0 images from 3d densities of the corona reconstructed by tomographic techniques (;
, @xcite ; ; ; ; ) , or simulated by global 3d mhd models @xcite .
the tomographic technique is a sophisticated method which reconstructs optically thin 3d coronal density structures using observations from multiple viewing directions .
the use of this method in solar physics was previously proposed by , and later this method has been applied to the soho / lasco @xcite and stereo / cor1 data @xcite . for a solar coronal tomography based on observations made by a single spacecraft or only from the earth - based coronagraph , data typically need to be gathered over a period of half solar rotation , so , generally , only structures that are stationary over about two weeks can be reliably reconstructed .
therefore , this technique is not applicable to eclipses or , perhaps , periods of high level of solar activity , although the 3d coronal electron density can be routinely computed @xcite .
similarly , as global mhd models of the corona need measurements of photospheric magnetic field data over a solar rotation , it is also difficult using the mhd method to reconstruct dynamic or rapidly - evolving coronal structures matching to observations
. thus , the ssi analysis could be very useful in some cases when the tomography is not suitable .
in addition , the ssi method is also useful in order to investigate the coronal density variability over a long term period ( several solar cycles ) when modern quality synoptic observations were not available and relate it to the modern state of the art reconstructions . in this study
, we choose the 3d coronal density obtained by tomography as a model in order to estimate uncertainties of the ssi method . compared the tomographic reconstruction and a 3d mhd model of the corona , and found that at lower heights the mhd models have better agreement with the tomographic densities in the region below 3.5 @xmath4 , but become more problematic at larger heights .
they also showed that the tomographic reconstruction has more smaller - scale structures within the streamer belt than the model can reproduce .
moreover , the tomographic reconstruction is entirely based on _ coronal _ observations , while the mhd models are primary based on photospheric boundary conditions .
this suggests that the tomographic reconstructions are more realistic and thus may be more suitable to be used as a model in order to evaluate uncertainties of the ssi method .
this article is organized as follows .
section [ sctssi ] describes two ssi methods and their relationship .
section [ sctrlt ] presents the evaluation of the ssi method .
we demonstrate the 3d density reconstruction by the ssi method based on real data in section [ sctd3d ] .
the discussion and conclusions as well as the potential extension of our work are given in section [ sctdc ] .
following derivations in , @xmath0 at a point , p , on the plane of sky ( pos ) can be expressed as , @xmath5\frac{\rho^2}{r^2 } n(s)ds , \label{eqpb}\ ] ] where the pos is defined as a plane crossing the sun s center and perpendicular to the los , @xmath6 is the electron density , @xmath7 is the perpendicular distance between the los and sun center , @xmath8 is the radial distance from the sun center , and @xmath9 is the distance measured from p along the los .
these distances are related by @xmath10 .
@xmath11 is the mean solar brightness .
@xmath13 , is the thompson scattering cross section for a single electron , where @xmath14 is the classical electron radius .
note that @xmath15 @xmath13 is the
thompson scattering cross section " as referred to in @xcite .
@xmath16 is the limb darkening coefficient .
in addition , @xmath17 and @xmath18 are geometrical factors given by @xcite @xmath19 , \label{eqbr}\end{aligned}\ ] ] where the angle @xmath20 is defined by @xmath21 .
if electron density is a function of @xmath8 only , equation ( [ eqpb ] ) can be written in the form @xmath22n(r)\frac{\rho^2 dr}{r\sqrt{r^2-\rho^2 } } , \label{eqpbr}\ ] ] where @xmath8 and @xmath7 are in units of @xmath4 , and @xmath23 .
developed a technique for the inversion of @xmath0 measurements ( called the _ van de hulst _ inversion ) . to implement this technique
, one needs first to fit the @xmath0 data using a curve in the polynomial form , specifically , @xmath24 once the coefficients @xmath25 are determined , the solution of equation ( [ eqpbr ] ) is then given by @xmath26 where @xmath27 where @xmath28 is the gamma function . note that the constant @xmath29 here is different from that in , because we have used the expression of @xmath0 in .
more generally , the index @xmath30 of radial power law in equation ( [ eqpbsum ] ) is real but not necessarily integer valued .
this leads to the modified van de hulst technique assuming @xmath31 , where @xmath25 and @xmath32 are the fit coefficients , then @xmath33 and @xmath34 in equation ( [ eqnr ] ) need to be replaced with @xmath35 and @xmath36 ( _ e.g. _ , ; ; )
. developed another ssi technique by assuming the radial electron density distribution in the polynomial form , @xmath37 , and have used it to the inversion of total brightness observations .
we here adopt this technique for the @xmath0 inversion , and call it the spherically symmetric polynomial approximation ( sspa ) method for short . by substituting the polynominal for @xmath38 in equation ( [ eqpbr ] ) ,
we obtain @xmath39 where @xmath40(r^{-k})\frac{\rho^2 dr}{r\sqrt{r^2-\rho^2 } } , \label{eqgkr}\ ] ] for easier calculation , using the substitution @xmath41 the integral can be transformed into @xmath42{\rm cos}^k\theta\,d\theta , \label{eqgktht}\ ] ] where @xmath43 is the angle from the pos to a direction from the sun center to a point of distance @xmath9 along the los . since the integral @xmath44 can be numerically calculated for all desired impact distances @xmath7 and exponents @xmath30 , substituting the observed @xmath0 curve along a radial trace for the left - hand side of equation ( [ eqpbg ] ) the only unknowns are the coefficients @xmath45 .
we determine the coefficients by a multivariate least - squares fit to the curve of @xmath46 ( using the function of _ svdfit _ provided by idl , interactive data language ) .
the radial distribution of electron density is then obtained by substituting the resulting coefficients directly into the polynomial form . to select appropriate degrees of polynominal we did some experiments using coronal density models for the region between 1.5 and 6 @xmath4 in , and found that choosing the first five terms ( @xmath47 ) can determine @xmath38 with the relative errors within 5% and reproduce @xmath0 measurements with the relative errors within 1% .
we thus use the 5-degree polynomial fits for all the sspa inversions in our study . to look into the relationship between the sspa and van de hulst inversions theoretically ,
we use taylor series to approximate the functions @xmath17 and @xmath18 in the case when @xmath48 with @xmath8 in unit of @xmath1 , @xmath49 for the taylor polynomial approximations of degree six above , the relative errors are less than 1% when @xmath50 . since @xmath17 and @xmath18 are on the order of @xmath51 for very large @xmath8 , by keeping only the terms of @xmath52 we have @xmath53 , which corresponds to the point source approximation @xcite .
applying this approximation to equation ( [ eqnr ] ) for the van de hulst inversion , we obtain @xmath54 this verifies the polynomial form of @xmath38 assumed in the sspa method .
conversely , if @xmath38 is given in the same form as equation ( [ eqnrapx ] ) , likewise we can recover @xmath55 as assumed in the van de hulst inversion from the sspa inversion using equations ( [ eqpbg ] ) and ( [ eqgktht ] ) with the point source approximation for @xmath56 .
thus , we have proved that the sspa and van de hulst inversions are identical on the order of @xmath51 , and their difference in higher orders can be reduced by increasing the degrees of @xmath30 .
note that the above analysis is also valid for the case when @xmath30 is real . in section [ sctstrm ]
, we will show that the coronal densities determined from cor1 images by the sspa and van de hulst inversions with @xmath30=5 are almost identical .
therefore , the assessment of the sspa inversion in the following is equivalent to that of the van de hulst inversion .
as a model for evaluation of ssi method we used a 3d coronal electron density obtained by tomography method applied to white - light coronagraph data @xcite .
the data were acquired by the inner coronagraph ( cor1 ) telescopes aboard the twin _ solar terrestrial relation observatory _ ( stereo ) spacecraft .
cor1 observes the white - light k corona from about 1.4 to 4 @xmath1 in a waveband of 22.5 nm wide centered on h@xmath57 line at 656 nm with a time cadence of 5 minutes .
regularized tomographic inversion method by with the limb darkening coefficient @xmath16=0.6 provided the reconstruction of a 3d coronal electron density for the period of 2008 february 114 ( consisting of 28 @xmath0 images from cor1-b ) that corresponds to carrington rotation ( cr ) 2066 .
the scattered light in the @xmath0 data was removed by subtracting a combination of the monthly minimum and the roll minimum backgrounds @xcite .
the reconstruction domain is a @xmath58 rectangular grid covering a spherical region between 1.5 and 4 @xmath1 .
we evaluate the sspa inversion method using this 3d density reconstruction as a model . in order to synthesize @xmath0 images observed by cor1-a and -b at a given time , we first determine the orbital positions of the spacecraft in carrington coordinates using the routine @xmath59 provided by solarsoftware ( ssw ) .
the 3d density grid is then transformed from the carrington heliographic system to the projected coordinate system viewed from stereo - a or -b @xcite .
therefore , the los integral of @xmath38 in equation ( [ eqpb ] ) becomes a simple summation in the @xmath60-direction ( defined along los ) .
figure [ fig : necr ] shows the tomographic reconstructed 3d coronal electron density at 2 @xmath1 .
figure [ fig : nemap ] shows the density distributions of the corona in the pos for cor1-a and -b at 12:00 ut on 8 february 2008 .
figure [ fig : pbmap ] shows the corresponding synthetic @xmath0 images , which represent the ideal measurements without contamination by the scattered light and instrumental noises .
coronal streamers are the most conspicuous , large - scale structures in the extended corona .
the streamer belt is a usually continuous sheet of enhanced density associated with the magnetic neutral sheet or the current sheet (; ; ; , @xcite ) .
its shape gets progressively deformed from a rather flat plane at minimum solar activity to a highly warped surface at maximum solar activity ( _ e.g. _ , ) .
some previous studies have suggested that the ssi assumptions are suitable to symmetric streamers at low latitude , in particular , the solar minimum streamer belt ( _ e.g. _ , ; ; ) .
here we first test the sspa inversion of the solar minimum streamer belt .
a streamer belt during cr 2066 corresponding to solar minimum of solar cycles 23@xmath6124 is shown in figure [ fig : necr ] .
for instance , we use the synthetic @xmath0 images at 12:00 ut on 8 february 2008 when the separation angle between stereo - a and -b was about 45@xmath2 .
figure [ fig : pbmap ] shows that the streamer belt on the east limb is almost edge - on , as inferred from the similar appearance in cor1-a and -b . in the edge - on condition , the coronagraph is looking along the streamer belt ; _ i.e. _ , where all the streamers are at the same latitude behind each other along the los .
in contrast , the streamer belt on west limb in cor1-b is face - on , showing a distinctly different shape from that in cor1-a .
we choose a radial trace near the middle of the east - limb streamer ( see marked positions in figure [ fig : pbmap ] ) , where the @xmath0 profiles for cor1-a and -b are nearly identical ( figure [ fig : pblc ] ) , and suppose that this location best meets the ssi condition .
we then derive the radial distribution of electron density by fitting the @xmath0 data along this radial trace between 1.6 and 3.9 @xmath1 using both the sspa method and the van de hulst technique ( @xmath62 with @xmath30=5 in ssw ) for an illustrated comparison , and show the inversion results in figure [ fig : pbinv ] .
note that since cor1 does not perform well at @xmath631.55 @xmath1 at some position angles due to interference from the occulter ( see ) , this led to defective @xmath0 signals and thus unreliable reconstructions within this region , so we constrain all sspa inversions in the following to regions with @xmath64 .
we find that the electron densities obtained from both @xmath0 inversion methods agree well in radial distribution with the model 3d densities in the pos for selected longitudes corresponding to the edge - on streamer positions ( see figure [ fig : pbinv ] ( c ) and ( d ) ) .
this confirms our theoretical prediction that the ppsa and the van de hulst inversions are equivalent .
the average ratio of the sspa to the model densities along the selected radial cut between 1.6 and 3.0 @xmath4 is [email protected] and [email protected] for cor1-a and -b , respectively .
note that weak increase in density with height between 3.5 and 4.0 @xmath4 may be due to fov effects ( see discussion in section [ sctdc ] ) .
in addition , the lower panels of figure [ fig : pbinv ] show that both the tomographic and sspa density profiles determined for streamers are comparable ( with differences by @xmath6620% ) to the previous obtained by ssi method in the solar minimum @xcite . to examine to
what extent the sspa solution is consistent with the 3d density model , we compare the sspa density profile with the model density profiles in 13 angular sections to the pos within 30@xmath2 in figures [ fig : pbinm](a ) and ( b ) , which show that the sspa solution is closest to the distribution of 3d densities in the pos .
this is as expected because the @xmath0 integrals along the los are most heavily weighted toward the regions near the pos ( see ; ) .
we evaluate the sspa inversion of the streamer as a function of longitude in the following .
this corresponds to make the sspa inversion of synthetic @xmath0 images at different virtual observing times " .
we synthesize the @xmath0 images at nine times for cor1-a and -b , in a period from 20:00 ut , 6 february to 04:00 ut , 10 february with intervals of 10 hours , centered at 12:00 ut on 8 february 2008 ( the time for the instance analyzed in the last section ) .
the sun rotated by 44@xmath2 over this period , approximately equal to the separation angle between stereo - a and b. so the ending - time location analyzed in cor1-a is approximately superposed with the starting - time location in cor1-b as shown in figure [ fig : necr ] .
we fit the @xmath0 data along the same radial trace between 1.6 and 3.9 @xmath4 at these nine times using the sspa method . figures [ fig : neva ] and [ fig : nevb ] show comparisons of the obtained density profiles with the model densities in the pos for cor1-a and cor1-b , respectively . for the analyzed locations with longitude between @xmath67 and @xmath68 ( approximately between the limbs @xmath69 and @xmath70 as shown in figure [ fig : necr ] ) , we find that the density profiles derived by sspa method have the similar shape as the model profiles over almost the whole fov range ( 1.63.8 @xmath4 ) , but the magnitudes are smaller than the model by about 20%50% .
the reason for good agreements during this period could be that the streamers in the los at the analyzed location are near the pos .
the model density increases near the edge of fov seen in figures [ fig : neva](c)-(e ) and figures [ fig : nevb](g)-(i ) likely result from the fov effects in the tomographic inversion @xcite .
figure [ fig : netim ] shows the sspa density as a function of longitude at five heights ( 1.6 , 2.0 , 2.5 , 3.0 and 3.5 @xmath4 ) .
the comparison with the model confirms the result above that the better agreement lies at the locations with the longitude in 60@xmath2106@xmath2 .
we also find that the agreement tends to be better at lower heights .
the average difference between the sspa and model densities are @xmath6620% for the region between 1.6 and 2.0 @xmath4 , while @xmath6640%50% for 2.53.5 @xmath4 .
the reason could be that the ssi condition is better meet at lower heights where the streamer has a larger extent in the los , and thus the los integration in equation ( [ eqgktht ] ) over a very long distance ( compared to the width of the streamer ) becomes more reasonable .
now we evaluate the latitudinal dependence of sspa inversions of the corona . by assuming that the ssi condition holds locally for all angular positions around the sun
, the 2d coronal density can be derived from a @xmath0 image by fitting the radial profile at each angular position using the sspa method .
for the case analyzed in section [ sctstrm ] , we reconstruct 2d coronal densities in the pos by fitting radial profiles between 1.6 and 3.7 @xmath4 for 360 position angles with intervals of 1@xmath2 from both synthetic and observed @xmath0 images , and show the results in figure [ fig : ne2d ] .
compared to the tomographic densities in the pos ( see figure [ fig : nemap ] ) , the sspa coronal densities determined from the observed images do not show very low - density regions around streamers , while those from the synthetic images have a small region of zero ( or negative ) densities at the southern side of the west - limb streamer , where the synthetic @xmath0 signals are very low ( below the background level in non - streamer regions ) .
the zero density regions in tomographic reconstructions could be caused either ( or both ) due to coronal dynamics or due to real very small density in these coronal regions ( see discussions in section [ sctdc ] ) .
for a quantitative comparison , we plot the sspa and model density profiles as a function of position angles at four heights ( 1.6 , 2.0 , 2.5 , and 3.0 @xmath4 ) in figure [ fig : neap ] . for the three edge - on streamers marked with s1s3
, we measure their peak densities and angular widths ( in fwhm ) .
the ratio of peak densities from the sspa to the model is on average [email protected] , [email protected] , and [email protected] , and the ratio of angular widths is on average [email protected] , [email protected] , and [email protected] at 2.0 , 2.5 , and 3.0 @xmath4 , respectively .
the results indicate that the sspa inversion underestimates the peak density of streamers by about 20%40% , while overestimates the angular width by about 20%90% .
the increase of deviations with height in the peak density of streamers by the sspa method from the model may be caused by narrowing of the streamer ( belt ) with height ( see figure 2 of ) . as the streamer width along
the los decreases with the height , the condition for the ssi assumption becomes worse .
the reason for the latitudinal spreading of the sspa density relative to the model may lie in that the analyzed streamers are not exactly edge - on , _
i.e. _ , the streamer belt at the analyzed locations is actually tilted somewhat away from the equatorial direction ( or the los near the limb , see figure [ fig : necr ] ) .
in such a case at the locations near the edge of streamer in the @xmath0 image the contribution is larger from points along the los behind or in front of the pos , leading to an overestimation of the model density in the pos by the sspa inversion at these places .
in addition , we notice for the face - on streamer ( marked s4 ) in cor1-b , its sspa density profiles are consistent with the model densities as well , this is because this streamer is by coincidence located near the pos at this instance ( figure [ fig : necr ] ) . in figure
[ fig : neap ] , we also compare the sspa densities inverted from the synthetic @xmath0 image with those from the observed @xmath0 image , and find that they are generally consistent except in the region near the occulter of cor1 , where the sspa inversions from the observed @xmath0 are much larger ( @xmath6623 times ) than the model densities ( see panels ( a ) and ( e ) in figure [ fig : neap ] ) .
the version of tomographic reconstruction used here most likely underestimated the density there as the solution at grid points close to the occulter is less constrained by the observational data .
in the above sections , we evaluated the sspa method by comparing the derived density distribution as a function of radial distance , longitude and latitude with the 3d tomographic model .
the assessments indicate that the sspa inversion can determine the 2d coronal density from a @xmath0 image , which approximately agrees with the 3d tomographic densities for streamers near the pos .
this suggests that we may reconstruct a 3d density of the corona by applying the sspa inversion to a 14 day data set from cor1-a or -b . in this section
, we demonstrate the sspa 3d density reconstructions using the same data set ( consisting of 28 @xmath0 images from cor1-b ) as used for the reconstruction of 3d tomographic model , and the simultaneous cor1-a data set .
the data cadence of about 2 images per day corresponds to a longitudinal step of about @xmath71 .
first we determine the 2d density distribution by fitting the radial @xmath0 data between 1.6 and 3.7 @xmath4 using the sspa inversion at 142 angular positions ( with intervals of 2.5@xmath2 ) surrounding the sun for each image .
then we map the east - limb and west - limb density profiles of all images to make a synoptic map at a certain height , based on their carrington coordinates ( neglecting the inclination of the solar rotational axis to the ecliptic ) .
a 3d density reconstruction is made of 25 synoptic maps for the radial heights from 1.5 to 3.9 @xmath4 with an interval of 0.1 @xmath4 .
for each synoptic map , we convert the irregular grid into the regular grid using the idl function , _
trigrid_. although the 3d density reconstruction can be made using the sspa method from the cor1 data with a cadence as high as 5 minutes , the higher temporal resolution may not help improve its actual angular resolution in longitude due to intrinsic limitations of the sspa method ( see appendix ) .
figure [ fig : synab ] shows the sspa - reconstructed 3d coronal density at 1.6 , 2.0 and 2.5 @xmath4 for cor1-a and -b .
some discontinuities of the density are seen at two longitudes which separate the two regions made of the east - limb and west - limb inversions
. these flawed structures may be due to temportal changes of streamers and/or the effect due to neglecting the solar axial tilt in the reconstruction .
we smooth these discontinuities by averaging the two reconstructions from cor1-a and cor1-b and then make a @xmath72 smoothing . in figure
[ fig : syncp ] , we show the ratio of tomographic density to the sspa average density for the cor1-a and -b reconstructions with smoothing for two cases , one for the sspa densities obtained using the synthetic data ( panels ( a)-(c ) ) , and the other for the sspa densities obtained using the real data ( panels ( d)-(f ) ) .
both cases indicate that the density ratios in the streamer belt are very close to 1 , within a factor of two or so ( @xmath660.51 at 1.6 @xmath1 , @xmath6612 at 2.0 @xmath1 , and @xmath6613 at 2.5 @xmath1 ) .
the good consistency validates that the ssi assumptions are very appropriate for the streamer belt in the solar minimum . for a quantitative comparison between the tomography and the sspa densities ( obtained from real data )
, we also show their density profiles along an equatorial cut in figure [ fig : synpf ] .
we find that they best match at 2 @xmath1 .
the same result is also indicated by the pixel - to - pixel scatter plots in figure [ fig : pxsct ] .
we obtain the ratio of the sspa to the tomographic average density is 1.02 , and their linear pearson correlation coefficient is 0.93 for the reconstructions at 2 @xmath1 .
in addition , the scatter plots show that the dispersion is larger for those pixels with smaller values , where the sspa densities are much larger than those obtained by the tomography .
in this study , we have , for the first time , evaluated the ssi method using the 3d model of the coronal electron densities reconstructed by tomography from the stereo / cor1 @xmath0 data .
our study is instructive for more efficient use of the ssi technique to invert the @xmath0 observations from ground- and space - based coronagraphs , in particular , the cor1 data .
we demonstrate both theoretically and observationally that the sspa method and the van de hulst inversion are equivalent ssi techniques when the radial densities or @xmath0 signals are assumed in the polynomial form of high degrees ( more than two ) .
the polynomial degree of five is suitable for cor1 data inversions .
thus , assessment results of the sspa method can also be applied to the van de hulst inversion technique .
we determine radial profiles of the streamer density from the cor1-a and -b synthetic @xmath0 images as well as their longitudinal and latitudinal dependencies .
we find that the sspa density values are close to the model for the core of streamers near the pos , with differences ( or uncertainties if we regard the model input as a true solution ) typically within a factor of @xmath662 .
this result is consistent with those evaluated using uv spectroscopy @xcite .
we find that the sspa density profiles tend to better match the model at lower heights ( @xmath73 2.5 @xmath1 ) .
our results confirm the suggestion in some previous studies that the ssi assumption is appropriate for the edge - on streamers or the streamer belt during the solar minimum .
we suggest that the edge - on condition for streamers may be determined by tomography method or by examining the consistency between simultaneous @xmath0 measurements from cor1-a and -b in radial distribution , when the two spacecraft have a small angular separation ( _ e.g. _ , less than 45@xmath2 ) .
we also find the sspa streamer densities are more spread in both longitudinal and latitudinal directions than in the model .
we demonstrate the application of the sspa inversion for reconstructions of the 3d coronal density near the solar minimum , and show that the sspa 3d density for the streamer belt is roughly consistent in both position and magnitude with the tomographic reconstruction .
the synoptic density maps derived by the sspa method show some discontinuities at the longitudes that separate the regions made of the east- and west - limb inversions
. these discontinuities may be due to temporal changes of coronal structures and/or the effect of tilt of the solar rotation axis on the poles visibility that is not considered .
they can be smoothed out during post - processing by smoothing and averaging the density distributions from cor1-a and -b , but such treatments will reduce the spatial resolution . in comparison , the tomographic inversion can fully take into account the tilt effect of the solar pole and produce the density distributions smoother in these discontinuity regions @xcite .
we estimate that the sspa method may resolve the coronal density structure near the pos with an angular resolution of @xmath6650@xmath2 in longitudinal direction . given this limitation , the sspa reconstruction using @xmath0 data with higher cadence ( _ e.g. _ , more than three images per day ) would not help improve its actual angular resolution .
although the current state of the tomographic method has allowed to routinely obtain the 3d coronal densities , we speculate that the ssi method could be complementary to the tomography when used for the interpretation of observations in such cases as during maximum of solar activity , in some regions where tomography gives zero density values , or in the regions near the edges of fov . the zero density values in the tomographic reconstruction ( so called zero - density artifacts " , zdas ) could be caused either ( or both ) due to coronal dynamics @xcite , or due to real very small density in these coronal regions which are below the error limit in the tomography method @xcite .
the latter reason is also supported by results of mhd modeling @xcite . in the former case , the ssi method could be complementary to tomography , while in the latter case the sspa method gives much larger values than in the tomographic model .
the fov artifacts in the tomography are due to the finite coronagraph fov that causes the reconstructed density to increase at the regions close to the outer reconstruction domain @xcite .
however , this fov artifacts can be reduced by extending the outer reconstruction domain beyond the coronagraph fov limit @xcite .
another way to obtain more correct density values in this region could be by using the sspa method which does not imply strict outer boundaries for los integration .
thus the estimate of uncertainties of the sspa method should be limited to the regions with distances less than about 3.5 @xmath4 for the tomographic model used .
also , the use of mhd model in order to produce artificial data can be useful for this test .
however this will be a subject of future research .
the tomography generally assumes that the structure of the corona is stable over the observational interval , _
e.g. _ , two weeks of observations made by a single spacecraft , although for some coronal regions that are exposed to the spacecraft for only about a week during the observation the stationary assumption can be reduced to about a week @xcite .
however , such an assumption is hard to meet during solar maximum or times of enhanced coronal activity .
the cme catalog in the nasa cdaw data center shows that the cme occurrence rate increases from @xmath660.5 per day near solar minimum to @xmath666 near solar maximum during solar cycle 23 @xcite .
although our assessment results for the sspa method are based on a static coronal model , their validity may not be limited to the static assumption . because the key factor for the sspa inversion for obtaining a good estimate of the 2d coronal density is an instantaneous ( local ) symmetric condition for coronal densities along the los
. the minimal size of this local symmetry is limited by the angular resolution in longitude which is about @xmath74 .
therefore , the ssi method can be used to estimate the density of a dynamic coronal structure in terms of weighted average over the region with this angular size in longitude . for this reason
it may be a better choice to use a combination of the tomography and the ssi inversions for interpretations of radio bursts and shocks produced by cmes in the case when coronal structures of interest evolve quickly with time @xcite .
in addition , the ssi method is also often applied to the cases when observational data are not suitable for tomography , _
e.g. _ , solar eclipses .
the 3d mhd models of the corona using the synoptic photospheric magnetic field data have been successfully used to interpret solar observations , including total eclipses and ground - based ( _ e.g. _ mlso / mk4 ) images of the corona ( _ e.g. _ , ; ) .
we suggest that the evaluation of the ssi method based on such a global mhd model may be necessary in the future .
the profits using such mhd models to estimate the uncertainties of the ssi method could be in avoidance of the zda and fov effects .
the modeled corona also allows evaluations of the ssi method down to very low heights ( _ e.g. _ , the region between 1.1 and 1.5 @xmath4 ) where the corona is much more structured .
moreover , a simulated time dependent corona from a time - evolving mhd model would allow us to estimate the uncertainties in tomography and the ssi inversion when they are applied to the dynamic corona , especially during the solar maximum .
this needs detailed investigations in the future .
* estimates of angular resolution of the sspa method in longitude * + to determine the angular resolution of the sspa method in the longitudinal direction , a numerical experiment is performed using 2d coronal models .
we construct a 2d density model by first using the equatorial background density model to build a background corona of rotational symmetry in the equatorial plane , and then inserting two structures into it .
the structures have the angular width of @xmath75 , the density contrast ratio of @xmath76 to the background , and the longitudinal profile following a step function or gaussian function . for the gaussian - type structure ,
its fwhm is set as @xmath75 .
figures [ fig : reso](a ) and ( b ) show the two types of 2d coronal models with about the same fov as cor1 , where two structures are separated by an angle of 2@xmath75 , thus the width of the gap between them is also equal to @xmath75 in the step - profile case .
we assume that the two structures ( marked a and b ) are located at the longitudes of @xmath75 and @xmath77 , respectively , i.e. defining their middle position as the origin of longitude .
so for the cases shown in figures [ fig : reso](a ) and ( b ) , the origin of longitude is just located in the pos at the west limb .
we synthesize @xmath0 data for the west - limb region from 1.5 to 4 @xmath4 at longitudes in the range from @xmath78 to 90@xmath2 using equation ( [ eqpb ] ) , and then derive the electron density using the sspa method by fitting the synthetic @xmath0 data .
the panels ( c)-(e ) show comparisons of the sspa density with the model density in the pos at 2 @xmath4 as a function of longitudes for different angular distances between two artificial structures with the density contrast ratio @xmath76=10 .
we find that for both types of the structure ( in the step or gaussian profile ) , the minimum resolvable distance ( or angular resolution ) is about 50@xmath2 .
to examine the effects of radial distance , structure width , and density contrast on the obtained resolution , we make a parametric study , and show the results in panels ( f)-(h ) .
we find that the resolution is only slightly dependent on the radial distance and structure width , which is better for the lower heights and narrower widths , but almost independent on the density contrast .
in addition , we also find that the obtained sspa peak density is about a half of the model density for the analyzed structures in most of the cases , which is consistent with our results for edge - on streamers .
these numerical experiments suggest that for coronal structures with more smoothing profile and larger extension in the longitudinal direction and with lower density contrast to the background , which can be regarded as better conditions meeting the spherically symmetric assumption , the sspa solutions are more accurate .
the work of tw was supported by the nasa cooperative agreement nng11pl10a to the catholic university of america and nasa grant nnx12ab34 g .
we very much appreciate to dr .
maxim kramar for his suggestions that led to an improved estimation of angular resolution of the sspa method in appendix .
we also thank the anonymous referee for his / her valuable comments in improving the manuscript .
airapetian , v. , ofman , l. , sittler , e. c. , kramar , m. : 2011 , _ astrophys .
j. _ * 728 * , 67 .
doi : 10.1088/0004 - 637x/728/1/67 allen , c. w. : 2000 , allen s astrophysical quantities ( 4th edition ) , arthur n. cox ( ed . ) , springer - verlag , isbn 0 - 387 - 98746 - 0 .
barbey , n. , auchre , f. , rodet , t. , vial , j .- c . : 2008 , _ solar phys . _ * 248 * , 409 .
doi : 10.1007/s11207 - 008 - 9151 - 6 barbey , n. , guennou , c. , auchre , f. : 2013 , _ solar phys .
_ * 283 * , 227 .
doi : 10.1007/s11207 - 011 - 9792 - 8 billings , d. e. : 1966 , a guide to the solar corona , academic press , new york .
blackwell , d. e. , petford , a. d. : 1966a , _ mon . not .
soc . _ * 131 * , 383 .
ads : http://adsabs.harvard.edu/abs/1966mnras.131..383b blackwell , d. e. , petford , a. d. : 1966b , _ mon . not .
_ * 131 * , 399 .
ads : http://adsabs.harvard.edu/abs/1966mnras.131..399b blackwell , d. e. , dewhirst , d. w. , ingham , m. f. : 1967 , _ advances in astronomy and astrophysics _ , z. kopal ( ed . ) , new york , * 5 * 1 .
butala , m. d. , frazin , r. a. , kamalabadi , f. : 2005 , _ j. geophys .
_ * 110 * , a09s09 .
doi : 10.1029/2004ja010938 butala , m. d. , hewett , r. j. , frazin , r. a. , kamalabadi , f. : 2010 , _ solar phys .
_ * 262 * , 495 .
doi : 10.1007/s11207 - 010 - 9536 - 1 caroubalos , c. , hillaris , a. , bouratzis , c. , alissandrakis , c. e. , preka - papadema , p. , polygiannakis , j. _ et al . _ : 2004 , _ astron .
astrophys . _ * 413 * , 1125 .
doi : 10.1051/0004 - 6361:20031497 cho , k .- s .
, lee , j. , moon , y .-
j . dryer , m. , bong , s .- c . , kim , y .- h . , park , y. d. : 2007 , _ astron .
astrophys . _ * 461 * , 1121 .
doi : 10.1051/0004 - 6361:20064920 cranmer , s. r. , kohl , j. l. , noci , g. , antonucci , e. , tondello , g. , huber , m. c. e. _ et al . _ : 1999 , _ astrophys .
j. _ * 511 * , 481 .
doi : 10.1086/306675 davila , j. m. : 1994 , _ astrophys .
j. _ * 423 * , 871 .
doi : 10.1086/173864 feng , x. s. , zhou , y. f. , wu , s. t. : 2007 , _ astrophys . j. _ * 655 * , 1110 .
doi : 10.1086/510121 frazin , r. a. , janzen , p. : 2002
, _ astrophys .
j. _ * 570 * , 408 .
doi : 10.1086/339572 frazin , r. a. , vsquez , a. m. , kamalabadi , f. , park , h. : 2007 , _ astrophys .
j. _ * 671 * , l201 .
doi : 10.1086/525017 frazin , r. a. , lamy , p. , llebaria , a. , vsquez , a. m. : 2010 , _ solar phys .
_ * 265 * , 19 .
doi:10.1007/s11207 - 010 - 9557 - 9 frazin , r. a. , vsquez , a. m. , thompson , w. t. , hewett , r. j. , lamy , p. , llebaria , a. , vourlidas , a. , burkepile , j. : 2012 , _ solar phys . _ * 280 * , 273 .
doi : 10.1007/s11207 - 012 - 0028 - 3 gibson , s. e. , bagenal , f. : 1995 , _ j. geophys
_ , * 100 * , 19865 .
doi : 10.1029/95ja01905 gibson , s. e. , fludra , a. , bagenal , f. , biesecker , d. , del zanna , g. , bromage , b. : 1999 , _ j. geophys .
_ * 104 * , 9691 .
doi : 10.1029/98ja02681 gopalswamy , n. , lara , a. , yashiro , s. , nunes , s. howard , r. a. : 2003 , _ solar variability as an input to the earth s environment , international solar cycle studies ( iscs ) symposium _ , wilson , a. ( ed . ) , esa publ .
division , noordwijk , * sp-535 * , 403 .
groth , c. p. t. , de zeeuw , d. l. , gombosi , t. i. , powell , k. g. : 2000 , _ j. geophys .
* 105 * , 25053 .
doi : 10.1029/2000ja900093 guhathakurta , m. , fisher , r. r. : 1995 , _ geophys .
* 22 * , 1841 .
doi : 10.1029/95gl01603 guhathakurta , m. , holzer , t. e. , macqueen , r. m. : 1996 , _ astrophys .
j. _ * 458 * , 817 .
doi : 10.1086/176860 hayashi , k. : 2005 , _ astrophys .
_ * 161 * , 480 .
doi : 10.1086/491791 hayes , a. p. , vourlidas , a. , howard , r. a. : 2001 , _ astrophys .
j. _ * 548 * , 1081 .
doi : 10.1086/319029 hu , y. q. , feng , x. s. , wu , s. t. , song , w. b. : 2008 , _ j. geophys .
res . _ * 113 * , 3106 .
doi : kimura , h. , mann , i. : 1998 , _ earth planets space _ * 50 * , 493 .
ads : http://adsabs.harvard.edu/abs/1998ep koutchmy , s. : 1994 , _ adv .
space res . _
* 14 * , 29 .
doi : 10.1016/0273 - 1177(94)90156 - 2 koutchmy , s. , lamy , p. l. : 1985 , _ properties and interactions of interplanetary dust _
, r. h. giese & p. l. lamy ( eds . ) , _ assl _ * 119 * , 63 .
kramar , m. , jones , s. , davila , j. m. , inhester , b. , mierla , m. : 2009 , _ solar phys . _ * 259 * , 109 .
doi : 10.1007/s11207 - 009 - 9401 - 2 kramar , m. , davila , j. , xie , h. , antiochos , s. : 2011 , _ ann .
geophys . _
* 29 * , 1019 .
doi : 10.5194/angeo-29 - 1019 - 2011 kramar , m. , inhester , b. , lin , h. , davila , j. : 2013 , _ astrophys . j. _ * 775 * , 25 .
doi : 10.1088/0004 - 637x/775/1/25 kramar , m. , airapetian , v. , miki , z. , davila , j. : 2014,_solar phys .
_ , in press .
doi : 10.1007/s11207 - 014 - 0525 - 7 kwon , r .- y .
, kramar , m. , wang , t. j. , ofman , l. , davila , j. m. , chae j. : 2013 , _ astrophys . j. _ * 776 * , 55 . doi : 10.1088/0004 - 637x/776/1/55 lallement , r. , qumerais , e. , lamy , p. , bertaux , j .- l . ,
ferron , s. , schmidt , w. : 2010 , _
soho-23 : understanding a peculiar solar minimum _ , s. r. cranmer , j. t. hoeksema , and j. l. kohl .
( eds . ) , _ asp conf . _
* 428 * , 253 .
lee , k .- s . , moon , y .-
j . , kim , k .- s . ,
lee , j .- y . , cho , k .- s . , choe , g. s. : 2008 , _ astron .
astrophys . _ * 486 * , 1009 .
doi : 10.1051/0004 - 6361:20078976 linker , j. a. , miki , z. , biesecker , d. a. , forsyth , r. j. , gibson , s. e. , lazarus , a. j. _ et al . _ : 1999 , _
j. geophys .
* 104 * , 9809 .
doi : 10.1029/1998ja900159 lionello , r. , linker , j. a. , miki , z. : 2009 , _ astrophys .
j. _ * 690 * , 902 .
doi : 10.1088/0004 - 637x/690/1/902 manchester , w. b. , gombosi , t. i. , roussev , i. , de zeeuw , d. l. , sokolov , i. v. , powell , k. g. , tth , gbor : 2004 , _
j. geophys .
res . _ * 109 * , a02107 .
doi : 10.1029/2003ja010150 manchester , w. b. , gombosi , t. i. , de zeeuw , d. l. , sokolov , i. v. , roussev , i. i. , powell , k. g. _ et al . _ : 2005 , _ astrophys .
j. _ * 622 * , 1225 .
doi : 10.1086/427768 mann , i. : 1992 , _ astron .
astrophys . _ * 261 * , 329 .
ads : http://adsabs.harvard.edu/abs/1992a miki , z. , linker , j. a. , schnack , d. d. , lionello , r. , tarditi , a. : 1999 , _ phys . plasma _
* 6 * , 2217 .
doi : 10.1063/1.873474 minnaert , m. : 1930 , _ zs .
* 1 * , 209 .
ads : http://adsabs.harvard.edu/abs/1930za......1..209m munro , r. h. jackson , b. v. : 1977 , _ astrophys .
j. _ * 213 * , 874 .
doi : 10.1086/155220 november , l. j. , koutchmy , s. : 1996 , _ astrophys .
j. _ * 466 * , 512 .
doi : 10.1086/177528 odstril , d. , pizzo , v. j. : 1999 , _ j. geophys .
* 104 * , 493 .
doi : 10.1029/1998ja900038 odstril , d. , linker , j. , lionello , r. , miki , z. , riley , p. , pizzo , v. j. , luhmann , j. g. : 2002 , _
j. geophys .
res . _ * 107 * , 1493 .
doi : 10.1029/2002ja009334 qumerais , e. lamy , p. : 2002 , _ astron .
astrophys . _ * 393 * , 295 .
doi : 10.1051/0004 - 6361:20021019 qumerais , e. , lallement , r. , koutroumpa , d. , lamy , p. : 2007
, _ astrophys .
j. _ * 667 * , 1229 .
doi : 10.1086/520918 ramesh , r. , kishore , p. , mulay , s. m. , barve , i. v. , kathiravan , c. , wang , t. j. : 2013 , _ astrophys .
j. _ * 778 * , 30 .
doi : 10.1088/0004 - 637x/778/1/30 reames , d. v. : 1999 , _ space sci .
rev . _ * 90 * , 41 .
doi : 10.1023/a:1005105831781 riley , p. , linker , j. a. , miki , z. : 2001 , _
j. geophys .
res . _ * 106 * , 15889 .
doi : 10.1029/2000ja000121 riley , p. , linker , j. a. , miki , z. , lionello , r. , ledvina , s. a. , luhmann , j. g. : 2006 , _ astrophys
. j. _ * 653 * , 1510 .
doi : 10.1086/508565 saez , f. , llebaria , a. , lamy , p. , vibert , d. : 2007 , _ astron .
astrophys . _ * 473 * , 265 .
doi : 10.1051/0004 - 6361:20066777 saito , k. : 1970 , _ ann .
tokyo astron .
, ser . 2 _ , * 12 * , 53 .
saito , k. , poland , a. i. , munro , r. h. : 1977 , _ solar phys .
_ * 55 * , 121 .
doi : 10.1007/bf00150879 schulz , m. 1973 , _ astrophys .
space sci . _
* 24 * , 371 .
doi : 10.1007/bf02637162 shen , c. , liao , c. , wang , y. , ye , p. , wang , s. : 2013 , _ solar phys .
_ * 282 * , 543 .
doi : 10.1007/s11207 - 012 - 0161-z sokolov , i. v. , roussev , i. i. , gombosi , t. i. , lee , m. a. , kta , j. , forbes , t. g. , manchester , w. b. , sakai , j. i. : 2004 , _ astrophys . j. lett .
_ * 616 * , l171 .
doi : 10.1086/426812 thompson , w. t. , wei , k. , burkepile , j. t. , davila , j. m. , st .
cyr , o. c. : 2010 , _ solar phys .
_ * 262 * , 213 .
doi:10.1007/s11207 - 010 - 9513 - 8 thompson , w. t. : 2006 , _ astron .
astrophys . _ * 449 * , 791 .
doi : 10.1051/0004 - 6361:20054262 usmanov , a. v. , goldstein , m. l. , besser , b. p. , fritzer , j. m. : 2000 , _
j. geophys .
res . _ * 105 * , 12675 .
doi : 10.1029/1999ja000233 van de hulst , h. c. : 1950 , _ bull .
* 11 * , 135 .
ads : http://adsabs.harvard.edu/abs/1950ban....11..135v van der holst , b. , sokolov , i. v. , meng , x. , jin , m. , manchester , w. b. , iv , tth , g. , gombosi , t. i. : 2014 , _ astrophys . j. _ * 782 * , 81 .
doi : 10.1088/0004 - 637x/782/2/81 vsquez , a. m. , frazin , r. a. , hayashi , k. , sokolov , i. v. , cohen , o. , manchester , w. b. , iv , kamalabadi , f. : 2008 , _ astrophys .
j. _ * 682 * , 1328 .
doi : 10.1086/589682 yashiro , s. , gopalswamy , n. , michalek , g. , st .
cyr , o. c. , plunkett , s. p. , rich , n. b. , howard , r. a. : 2004 , _ j. geophys .
res . _ * 109 * , 7105 .
doi : 10.1029/2003ja010282 zucca , p. , carley , e. p. , bloomfield , d. s. , gallagher , p. t. : 2014 , _ astron .
astrophys . _ * 564 * , 47 .
doi : 10.1051/0004 - 6361/201322650 | determination of the coronal electron density by the inversion of white - light polarized brightness ( @xmath0 ) measurements by coronagraphs is a classic problem in solar physics .
an inversion technique based on the spherically symmetric geometry ( spherically symmetric inversion , ssi ) was developed in the 1950s , and has been widely applied to interpret various observations .
however , to date there is no study about uncertainty estimation of this method . in this study
we present the detailed assessment of this method using a three - dimensional ( 3d ) electron density in the corona from 1.5 to 4 @xmath1 as a model , which is reconstructed by tomography method from stereo / cor1 observations during solar minimum in february 2008 ( carrington rotation , cr 2066 ) .
we first show in theory and observation that the spherically symmetric polynomial approximation ( sspa ) method and the van de hulst inversion technique are equivalent .
then we assess the sspa method using synthesized @xmath0 images from the 3d density model , and find that the sspa density values are close to the model inputs for the streamer core near the plane of the sky ( pos ) with differences generally less than a factor of two or so ; the former has the lower peak but more spread in both longitudinal and latitudinal directions than the latter .
we estimate that the sspa method may resolve the coronal density structure near the pos with angular resolution in longitude of about 50@xmath2 .
our results confirm the suggestion that the ssi method is applicable to the solar minimum streamer ( belt ) as stated in some previous studies .
in addition , we demonstrate that the sspa method can be used to reconstruct the 3d coronal density , roughly in agreement with that by tomography for a period of low solar activity ( cr 2066 ) .
we suggest that the ssi method is complementary to the 3d tomographic technique in some cases , given that the development of the latter is still an ongoing research effort . |
* atomic and electronic structure of ipcm .
* figure 1 shows the crystal structure of the ( gete)@xmath0(sb@xmath0te@xmath1)@xmath8 superlattice in the highly resistive reset state with sb@xmath0te@xmath1-te - ge - ge - te- stacking sequence ( inverted - petrov structure ) and the highly conductive set state with sb@xmath0te@xmath1-ge - te - ge - te- stacking sequence ( ferroelectic - gete structure ) along the rhombohedral [ 111 ] direction .
the sb@xmath0te@xmath1 block consists of one te - sb - te - sb - te quintuplet layer . despite the large contrast in electrical conductivity ,
a sole difference in atomic structure is the stacking sequence of ge - te layers sandwiched by sb@xmath0te@xmath1 layers .
the mbjlda band structures along the high - symmetry directions for the ( a ) inverted - petrov and ( b ) ferro - gete ( gete)@xmath0(sb@xmath0te@xmath1)@xmath8 superlattice are shown in figs .
1(e ) and ( f ) , respectively .
as expected , the reset state shows an inverted band structure at @xmath5 where the conduction band minimum is derived primarily from te-@xmath9 states while the valence band maximum from cation ( ge or sb ) @xmath9 states , indicating that the reset state is a topological insulator in agreement with previous studies@xcite .
the direct energy gap of 38 mev at the @xmath5 point smaller by a factor of 5 than that of sb@xmath1te@xmath2@xcite indicates a reduction in the band inversion strength and thus proximity of the system to the boundary between topological and normal insulator phases . despite the conducting nature of the surface states protected by the nontrivial band topology , the electric conduction is mostly dominated by the insulating bulk states rendering the reset state highly resistive . on the other hand
, the emergence of ferroelectric polarization in the set state induces a giant rashba effect that lifts the degeneracy around @xmath5 and @xmath10 . as a results , the valence band maxima and conduction band minima shift away from these symmetry points . despite these differences , both reset and set
states show insulating behavior along the high symmetry lines may not explain the striking conductivity change in the phase transition . obviously , it is natural to examine if the set state is a wsm by searching the weyl nodes in the entire bz .
* effect of weyl semimetal phase on electrical properties . *
interestingly , we find that the set state is indeed a wsm with twelve gapless weyl nodes at the fermi level in the vicinity of the @xmath10-point .
figure 2(a ) and ( b ) show three pairs of weyl nodes located at the @xmath11 plane and the other three pairs at the @xmath12 plane , respectively , where @xmath13 @xmath14 is the deviation from the bz cross section at @xmath15 .
the distances between the nearest and next nearest weyl nodes are 0.036 @xmath14 and 0.079 @xmath14 , respectively .
the two - dimensional band structure of the set state in the @xmath16 planes at @xmath11 , @xmath17 , and @xmath18 are shown in figs .
2(d ) , ( e ) and ( f ) , respectively
. one can clearly see that the set state is gapped at the bz boundary @xmath17 , while it is gapless , having six weyl nodes each at two planes away from the bz boundary .
calculated berry curvature shows more detailed characteristics of the weyl phase in the set state .
each weyl node behaves as an effective magnetic monopole in the momentum space with either positive or negative chirality , corresponding to a source or a sink of the berry curvature , respectively . in fig .
2(g ) , we show the berry curvature with a dense @xmath7-point grid on the @xmath19 plane that passes through the two weyl nodes , clearly indicating the outgoing and incoming flux of berry curvature from one monopole ( red ) to the other monopole ( blue ) . unlike dirac nodes , which may easily become massive by weak perturbations , weyl nodes are persistently massless until they are mutually annihilated .
furthermore , the chirality of weyl nodes prohibits the back - scattering of conduction electrons from each chiral channel .
the emergence of gapless nodes characterized by an exceptional high mobility , therefore , may explain the giant conductivity enhancement as ipcm undergoes the reset to set phase transition . by employing the maximally localized wannier functions and the semiclassical boltzmann transport equation in the constant relaxation time approximation
, we calculate the room - temperature electric conductivity of the set and reset states , with the assumption that both state have the same relaxation time .
2(h ) shows the ratio of the set to the reset state of both the in- and out - of - plane conductivity as a function of the chemical potential .
our calculations show that for both transport directions , @xmath20 at @xmath21 . on the other hand ,
the conductivity ratio may increase by up to two orders of magnitude for p - doped cases , which commonly occur in pcm due to ge vacancies .
therefore , we can attribute the giant change of conductivity observed in experiment@xcite to the presence of weyl nodes . to our knowledge , this is the first connection between weyl state to the transport property of a real material .
* fermi arcs and surface states . * for the experimental verification of the weyl state , we now address another key characteristic of a wsm ; the presence of the fermi arc surface states that connect the weyl nodes in pairs in the surface bz . using the green s function method based on the tight - binding hamiltonian generated by the wannier projection scheme , we obtain the surface band structures of the ge - te - terminated ( gtt ) and the sb - te - terminated ( stt ) surfaces . the band structure of the gtt surface [ fig .
3(a ) ] illustrates typical features of the weyl phase , namely it is gapped along the @xmath22 direction while it is gapless along the @xmath23 direction .
this is consistent with the fermi arcs in fig .
3(c ) connecting weyl pairs across the @xmath23 axis . on the other hand , the band structure on the stt surface [ fig .
3(b ) ] is not trivial as the gtt surface .
the stt surface bands cross the fermi level along both the high symmetry lines indicating a closed loop on the surface [ fig .
interestingly , the stt surface indeed displays one closed loop together with six fermi arcs which rather connect counter weyl node pairs compared to the gtt surface .
the relation between the fermi - arc - connected pairs and the emergent closed loop is discussed below . due to shorter out - of - plane lattice constant ( 17.385 ) of the set state compared to that ( 18.309 ) of the reset state
, the domain in the set state may experience tensile uniaxial strain from surrounding domains in the reset state .
the evolution of the surface states at the fermi level on the gtt ( stt ) surface as a function of the te - te interlayer distance , @xmath24 , between the sb@xmath0te@xmath1 and gete building blocks ( fig .
1 ) is shown in the top ( bottom ) panels fig .
3(c ) [ 3(d ) ] .
the red circles denote the surface - projected weyl nodes and the white - blue color gradient denotes spectral weight of surface states in the same scale with the band structures .
figures 3 ( c ) and ( d ) demonstrate that the weyl nodes of the set state are robust under tensile strain where they mutually merge and disappear under compressive strain . by decreasing the vacancy layer thickness @xmath24 by 0.1 ,
the six pairs of weyl nodes merge along the @xmath25 high symmetry directions in the bulk bz , reopening the band gap of 7 mev .
it is worthwhile to note that , in comparison to bitei , the type of strain ( tensile or compressive ) required to shift the weyl nodes from their creation points is opposite , namely it is compressive for bitei while it is tensile for ferro .- gete .
the fermi arcs appear regardless of the surface termination .
* two types of fermi arc connectivity .
* the distinct behavior of the fermi arcs on the gtt and stt surfaces of the set state can be understood as an adiabatic transition of surface states from the normal to the topological insulator via the weyl semimetal phase . for insulating nonmagnetic materials ,
an even ( odd ) number of surface bands crossing the fermi level indicates trivial ( nontrivial ) topological phase .
figure 4 shows schematically two possible transition pathways of surface states from the normal to topological insulators passing through wsm by changing an external parameter _
m_. in the first type of transition [ fig .
4(a ) ] , the initial trivial insulating phase has no surface states along the high symmetry line connecting the two time - reversal - invariant momenta , @xmath26 and @xmath27 , of the surface bz .
fermi arcs subsequently emerge connecting weyl nodes corresponding to monopole - antimonopole creation pairs ; @xmath28-@xmath29 and @xmath30-@xmath31 . in the weyl phase , where the fermi arcs evolve as the external parameter changes , there are no surface states along the @xmath26-@xmath27 line until the system undergoes a transition to the topological insulator phase via the mutual monopole - antimonopole pair annihilation . on the other hand , in the second type of transition [ fig .
4(b ) ] , the initial trivial insulating phase exhibits two surface states crossing the fermi level along the @xmath26-@xmath27 line . as the system enters in the weyl phase ,
the emerging fermi arcs connect weyl nodes , @xmath28-@xmath30 and @xmath29-@xmath31 , corresponding to monopole - antimonopole annihilation pairs so that two fermi surface states cross the @xmath26-@xmath27 symmetry direction _ before _ the system undergoes a phase transition to the topological insulator . in this case the surface states in the weyl phase consist of both closed fermi surface and open fermi arcs .
note that in both transition paths , there are even ( odd ) number of surface state along the @xmath26-@xmath27 line in the normal ( topological ) insulator phase .
consequently , the surface of a weyl semimetal with fermi arcs connecting creation ( annihilation ) monopole - antimonopole pairs should exhibit an even ( odd ) number of closed loops .
the gtt ( stt ) surface of the set state of ipcm corresponds to the former ( later ) case .
a recent study of the band - bending effect @xcite also showed that an emergent closed loop on the surface is accompanied by a partner exchange of fermi arcs .
* impilcation of photoemssion spectroscopy . *
the appearance of the closed loop at the stt surface may hinder the observation of fermi arcs in angle - resolved photoemission spectroscopy measurements due to its close proximity in momentum space .
however , even with limited resolution , the spectral intensity on the @xmath32-@xmath33 lines should be twice of that on the @xmath32-@xmath34 lines .
it is also worthwhile to note that the distance of about 0.1 @xmath35 of the fermi arcs from @xmath32 is larger than that of other predicted hexagonal weyl semimetals@xcite .
* effect of breaking time reversal on the weyl phase .
* weyl semimetals require breaking of either the time - reversal or inversion symmetry .
we have also explored the consequences of breaking the time reversal symmetry in the reset phase under an external magnetic field by including a zeeman splitting term , @xmath36 in the tight - binding hamiltonian the effective magnetic field is along [ 001 ] direction and @xmath37 is the pauli spin operator .
figure 5 shows that the direct band gap of the reset inverted - petrov structure decreases with increasing zeeman splitting and eventually closes above the critical field strength ( @xmath38 mev ) .
the number of weyl nodes depend on the symmetry of the system .
for instance , the number of weyl nodes in a system with @xmath39 symmetry , _ i.e. _ bitei or ge@xmath0sb@xmath0te@xmath2 , is always a multiple of twelve .
the applied magnetic field , however , breaks the mirror symmetries and hence reduces the symmetry to @xmath40 , giving rise to only six weyl nodes .
the spin degeneracy of the initial band structure [ fig .
5(b ) ] is lifted due to the zeeman term . with increasing splitting [ fig .
5(c ) ] , a pair of valence and conduction bands eventually touch at @xmath5 point and six gapless nodes persist along the @xmath41 direction at higher magnetic field . because of the minuscule separation of weyl nodes in momentum space and the complicated dependence on magnetic field direction , the detailed weyl properties with broken time - reversal symmetry are not discussed here .
perhaps a more practical way in introducing magnetic exchange splitting in ipcm without external magnetic field is via magnetic doping .
since there have been previous studies@xcite on magnetic doping in ge@xmath0sb@xmath0te@xmath2 , it seems highly plausible to tune the exchange field by varying the magnetic impurity concentration .
the ipcm undergoes a reversible structural transition between the reset and the set state .
we predict that the high conductive set state exhibits the _ weyl semimetal phase _ due to the ferroelectric polarization of the gete block in contrast to the _ topological insulating phase _ of the low conductive reset state .
the emergent twelve weyl nodes in the set state cause a giant enhancement in the conductivity by a factor of 5 @xmath42 100 in comparison to the reset state . to the best of our knowledge ,
this is the first memory device in which the conductivity contrast between the set and the reset state is attributed to the emergent weyl nodes .
our results , implying a robust weyl phase in the set state , open a way to functionalize the ipcm and design novel devices .
we also predict two types of fermi arc connectivity in the set state depending on the surface termination .
* calculation details . *
the density functional theory calculations were carried out using the vienna _ ab initio _ simulation package ( vasp ) @xcite . the pseudopotential and wave functions
are treated within the projector - augmented wave ( paw ) method @xcite .
structural relaxations were carried using the generalized gradient approximation as parameterized by perdew _
@xcite with van der waals corrections@xcite .
the plane wave cutoff energy is 400 ev and a 15 @xmath43 15 @xmath43 5 @xmath7 point mesh is used in the bz sampling .
the electronic properties such as the band structures , surface states , and conductivities have been calculated using the tight - binding parameters obtained from vasp - wannier90 calculations @xcite . in the wannier function projection ,
the modified becke - johnson local density approximation ( mbjlda ) @xcite potential was employed , which has been shown to yield accurate band gaps , effective masses , and frontier - band ordering .
spin - orbit coupling was included in the projection with a 10 @xmath43 10 @xmath43 4 @xmath7-point mesh in the bz sampling .
the weyl nodes were traced using a steepest decent method while the longitudinal electric conductivity was calculated within the semiclassical boltzmann transport theory@xcite .
37 s. r. ovshinsky , reversible electrical switching phenomena in disordered structures , phys .
* 21 * , 1450 ( 1968 ) .
w. welnic , a. pamungkas , r. detemple , c. steimer , s. blgel , and m. wuttig , unravelling the interplay of local structure and physical properties in phase - change materials , nat .
mater . * 5 * , 56 ( 2006 ) .
m. wuttig , and n. yamada , phase - change materials for rewriteable data storage , nat . mater .
* 6 * , 824 ( 2007 ) .
j. kim , j. kim , and s .- h .
jhi , prediction of topological insulating behavior in crystalline ge - sb - te , phys .
b * 82 * , 201312 ( 2010 ) .
j. kim , j. kim , k .- s .
kim , and s .- h .
jhi , topological phase transition in the interaction of surface dirac fermions in heterostructures , phys .
lett . * 109 * , 146601 ( 2012 ) .
r. e. simpson , p. fons , a. v. kolobov , t. fukaya , m. krbal , t. yagi , and j. tominaga , interfacial phase - change memory , nat .
* 6 * , 501 ( 2011 ) .
h. zhang , c .- x .
qi , x. dai , z. fang , s .- c .
zhang , topological insulators in bi@xmath0se@xmath1 , bi@xmath0te@xmath1 and sb@xmath0te@xmath1 with a single dirac cone on the surface nat .
phys . * 5 * , 438 ( 2009 ) .
d. d. sante , p. barone , r. bertacco , and s. picozzi , electric control of the giant rashba effect in bulk gete , adv .
mater . * 25 * , 509 ( 2013 ) .
m. j. polking , m .-
han , a. yourdkhani , v. petkov , c. f. kisielowski , v. v. volkov , y. zhu , g. caruntu , a. paul alivisatos , and r. ramesh , ferroelectric order in individual nanometre - scale crystals , nat .
11 * , 700 ( 2012 ) .
m. j. polking , j. j. urban , d. j. milliron , h. zheng , e. chan , m. a. caldwell , s. raoux , c. f. kisielowski , j. w. ager , r. ramesh , and a. p. alivisatos , size - dependent polar ordering in colloidal gete nanocrystals , nano lett . * 11 * , 1147 ( 2011 ) .
j. tominaga , a. v. kolobov , p. j. fons , x. wang , y. saito , t. nakano , m. hase , s. murakami , j. herfort and y. takagaki , giant multiferroic effects in topological gete - sb2te3 superlattices , sci .
technol . adv .
* 16 * , 014402 ( 2015 ) .
d. bang , h. awano , j. tominaga , a. v. kolobov , p. fons , y. saito , k. makino , t. nakano , m. hase , y. takagaki , a. giussani , r. calarco , and s. murakami , mirror - symmetric magneto - optical kerr rotation using visible light in [ ( gete)2(sb2te3)1]n topological superlattices , sci .
rep . * 4 * , 5727 ( 2014 ) .
b. sa , j. zhou , z. sun , j. tominaga , and r. ahuja , topological insulating in gete / sb2te3 phase - change superlattice , phys .
* 109 * , 096802 ( 2012 ) .
s. murakami , phase transition between the quantum spin hall and insulator phases in 3d : emergence of a topological gapless phase , new j. phys .
* 9 * , 356 ( 2007 ) .
s. murakami and s .-
kuga , universal phase diagrams for the quantum spin hall systems , phys .
b * 78 * , 165313 ( 2008 ) .
h. weyl , elektron und gravitation .
i , z. phys .
* 56 * , 330 ( 1929 ) .
h. nielsen and m. ninomiya , the adler - bell - jackiw anomaly and weyl fermions in a crystal , phys .
b * 130b * , 389 ( 1983 ) .
a. a. burkov , m. d. hook , and leon balents , topological nodal semimetals , phys .
b * 84 * , 235126 ( 2011 ) .
yang , y .- m .
lu , and y. ran , quantum hall effects in a weyl semimetal : possible application in pyrochlore iridates , phys .
b * 84 * , 075129 ( 2011 ) .
x. wan , a. m. turner , a. vishwanath , and s. y. savrasov , topological semimetal and fermi - arc surface states in the electronic structure of pyrochlore iridates , phys . rev .
b * 83 * , 205101 ( 2011 ) .
xu , i. belopolski , n. alidoust , m. neupane , g. bian , c. zhang , r. sankar , g. chang , z. yuan , c .- c .
huang , h. zheng , j. ma , d. s. sanchez , b.wang , a. bansil , f. chou , p. p. shibayev , h. lin , s. jia , and m. z. hasan , discovery of a weyl fermion semimetal and topological fermi arcs , science * 349 * , 613 ( 2015 ) .
l. x. yang , z. k. liu , y. sun , h. peng , h. f. yang , t.zhang , b. zhou , y. zhang , y. f. guo , m. rahn , d. prabhakaran , z. hussain , s. k. mo , c. felser , b. yan , and y. l. chen , weyl semimetal phase in the non - centrosymmetric compound taas , nat . phys . * 11 * , 728 ( 2015 ) .
j. liu and d. vanderbilt .
, weyl semimetals from noncentrosym - metric topological insulators , phys . rev .
b * 90 * , 155316 ( 2014 ) .
i. a. nechaev , i. aguilera , v. de renzi , a. di bona , a. lodi rizzini , a. m. mio , g. nicotra , a. politano , s. scalese , z. s. aliev , m. b. babanly , c. friedrich , s. blgel , and e. v. chulkov , quasiparticle spectrum and plasmonic excitations in the topological insulator @xmath44 , phys .
b * 91 * , 245123 ( 2015 ) . i. p. rusinov , t. v. menshchikova , i. yu .
sklyadneva , r. heid , k .-
bohnen , and e. v. chulkov , pressure effects on crystal and electronic structure of bismuth tellurohalides , new j. phys .
* 18 * , 113003 ( 2016 ) .
song , l .- p .
shi , x .- s .
miao , and t .- c .
chong , synthesis and characteristics of a phase - change magnetic material , adv .
* 20 * , 2394 ( 2008 ) .
f. tong , j. h. hao , z. p. chen , g. y. gao , h. tong , and x. s. miao , anomalous second ferromagnetic phase transition as a signature of spinodal decomposition in fe - doped gete diluted magnetic semiconductor , appl .
. lett . * 99 * , 202508 ( 2011 ) .
song , l .- p .
shi , and t .- c .
chong , magnetic properties and phase change features in fe - doped ge?sb?te , j. nanosci .
nanotechnol . * 11 * , 2648 - 2651 ( 2011 ) .
g. kresse and j. furthmller , efficient iterative schemes for ab initio total - energy calculations using a plane - wave basis set , phys .
b * 54 * , 11169 ( 1996 ) .
g. kresse and j. furthmller , efficiency of ab - initio total energy calculations for metals and semiconductors using a plane - wave basis set , comput .
* 6 * , 15 ( 1996 ) .
p. e. blchl , projector augmented - wave method , phys .
b * 50 * , 17953 ( 1994 ) .
g. kresse and d. joubert , from ultrasoft pseudopotentials to the projector augmented - wave method , phys .
b * 59 * , 1758 ( 1999 ) .
j. p. perdew , k. burke , and m. ernzerhof , generalized gradient approximation made simple , phys .
lett . * 77 * , 3865 ( 1996 ) .
s. grimme , semiempirical gga - type density functional constructed with a long - range dispersion correction , j. comp . chem . * 27 * , 1787 ( 2006 ) .
f. tran and p. blaha , accurate band gaps of semiconductors and insulators with a semilocal exchange - correlation potential , phys .
102 * , 226401 ( 2009 ) .
d. koller , f. tran , and p. blaha , improving the modified becke - johnson exchange potential , phys .
b * 85 * , 155109 ( 2012 ) .
a. a. mostofi , j. r. yates , g. pizzi , y .- s .
lee , i. souza , d. vanderbilt , and n. marzari , an updated version of wannier90 : a tool for obtaining maximally - localised wannier functions , comput .
. commun . * 185 * , 2309 ( 2014 ) .
the work at csun is supported by nsf - partnership in research and education in materials ( prem ) grant dmr-1205734 and by the us army of defense ( dod ) , grant contract number w911nf-16 - 1 - 0487 .
s. and s .- h .
j. were supported by the national research foundation of korea through src program ( contract no .
2011 - 0030046 ) .
j. kim and r. wu were supported by doe - bes ( grants nos . de - fg02 - 05er46237 and sc0012670 ) .
calculations were performed on parallel computers at nersc supercomputer centers .
j.k . and n.k .
designed this project . j.k . ,
j.k . , and y .- s.s . performed the calculation .
r.w . , s .- h.j . , and n.k . wrote the manuscript .
all authors commented on the manuscript and the results .
[ t ] , where @xmath46 is the pauli spin operator .
calculated band structures with ( * b * ) @xmath47 and ( * c * ) @xmath48 ev .
color of the bands denotes the orbital character as in fig . | the interfacial phase - change memory ( ipcm ) gete / sb@xmath0te@xmath1 , promising candidates for the next generation non - volatile random - access memories , exhibits fascinating topological properties . depending on the atomic - layer - stacking sequence of the gete block , the ipcm can be either in the set ( ge - te - ge - te ) or reset ( te - ge - ge - te ) states , where the former exhibits ferroelectric polarization and electrical conductivity two orders of magnitude larger than that of the reset state .
but , its origin remains elusive .
here , we predict the emergence of a weyl semimetal phase in the set state induced by the ferroelectric polarization which breaks the crystal inversion symmetry .
we show that the giant conductivity enhancement of the set phase is due to the appearance of gapless weyl nodes .
the ge - te- or sb - te - terminated surfaces of weyl semimetal ipcm produce surface states with completely distinctive topology , where the former consists solely of fermi arcs while the latter consists of both closed fermi surface and open fermi arcs .
the ipcm with van der waals interfaces offers an ideal platform for exploiting the exotic weyl properties for future memory device applications .
* phase change memory . * among the chalcogenide phase - change materials , ge@xmath0sb@xmath0te@xmath2 ( gst ) , continues to be one of the most attractive subjects for active experimental and theoretical investigations because of its extensive use in optical memory devices ( dvd - ram ) , as well as its high potential for non - volatile phase change memory ( pcm ) devices@xcite .
both applications rely on the large contrast in optical and electric properties between the crystalline and amorphous phases of the phase - change materials which can be switched reversibly and rapidly by sweeping temperature either via laser irradiation ( dvd ) or the joule effect ( pcm )
. * interfacial phase change memory . * recently , a new type pcm structure consisting of a two - component ( gete)@xmath3/(sb@xmath0te@xmath1)@xmath4 superlattice was proposed , referred to as interfacial phase change memory ( ipcm)@xcite .
it offers reduction in switching energies , improved write - erase cycle lifetimes , faster switching speeds and surprisingly large room - temperature magnetoresistance even without any magnetic element .
it is noticeable that the two end members of the ipcm exhibit distinct properties ; sb@xmath0te@xmath1 is a three - dimensional topological insulator ( ti ) with a band inversion@xcite at the @xmath5-point , while gete is a narrowgap ( 0.55 ev ) ferroelectric material with a giant rashba effect@xcite .
the ferroelectricity of gete persists down to the nanometer scale@xcite . compared to their pcm counterparts , ipcm remains in the crystalline phase in both the high- ( set ) and low - conductivity ( reset ) states , and the switch process involves a one - dimensional atomic motion normal to the interface .
more explicitly , the stacking sequences within the gete film in the set and reset phases are @xmath6-te - ge - te - ge-@xmath6 and @xmath6-te - ge - ge - te-@xmath6 , respectively .
while the former exhibits ferroelectric polarization due to the breaking of the crystal inversion symmetry@xcite , the later has no polarization .
the underlying mechanisms of the phase change dynamics and the giant change of resistance between the set and reset states still remain elusive@xcite .
one important property of the reset state of the ipcm is the topologically nontrivial order inherited from the bulk sb@xmath0te@xmath1 , although the band inversion strength is somewhat reduced by the intercalating gete block@xcite . in the set state
, the gete block not only reduces the band inversion strength but also breaks the spatial inversion symmetry .
consequently , this raises a question whether the set phase is a weyl semimetal .
it was proposed @xcite that the absence of the inversion symmetry guarantees the weyl phase to emerge at the phase boundary between topological and trivial insulator phases .
* weyl semimetal .
* weyl semimetals ( wsms ) exhibit a plethora of unusual and intriguing properties .
wsms are three - dimensional crystals , which break either time - reversal or inversion symmetry , where pairs of energy bands touch at certain @xmath7 points , referred to as weyl nodes@xcite , and the bands disperse linearly in the vicinities around the weyl nodes .
the weyl nodes have chiralities ( topologically protected chiral charges ) and act as sources or sinks of the berry curvature flux in the brillouin zone ( bz)@xcite .
wsms host unusual surface states which form disjoint open fermi arcs rather than closed fermi curves , connecting the projections of the weyl nodes onto the surface bz where the fermi arcs on opposite surfaces of the sample are linked through the bulk states@xcite . despite extensive search for wsms ,
the most convincing signature of wsm has been observed only in the covalently - bonded taas family@xcite . to prevent the intervention of surface dangling bonds for the observation of the fermi arcs , it is typical to find materials whose layers are weakly bonded by the van der waals forces .
the ferroelectric insulator bitei with van der waals layers was predicted to undergo a phase transition from normal insulator to topological insulator with an intervening weyl semimetal phase under a hydrostatic pressure@xcite .
however ,
the pressure window is too narrow to realize steadily the weyl semimetal phase . * objective . * here , employing _ ab initio _ electronic structure calculations we predict that the set phase of the van der waals type ipcm structure is a weyl semimetal .
this is due to the ferroelectric polarization in the gete building block that breaks the inversion symmetry .
we also elucidate that the underlying mechanism of the giant enhancement ( by up to two orders of magnitude ) of the electrical conductivity as the ipcm undergoes the phase transition from the reset to the set state is the emergence of the gapless chiral monopole - antimonopole weyl pairs .
we further demonstrate that the shape of the fermi surface depends sensitively on the surface termination exhibiting different topology .
the ge - te - terminated surface has solely fermi arcs and the sb - te - terminated surface has both closed fermi surface and open fermi arcs .
our findings show a possibility of realizing and exploiting the weyl state in materials without the need of particular experimental conditions such as high pressure or low temperature . |
blue compact dwarf ( bcd ) galaxies pose an interesting puzzle for standard galaxy evolution models . their extremely low gas
phase metallicities and blue colors suggest that these galaxies are relatively unevolved systems , perhaps undergoing their first bursts of star formation at the present epoch ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
recently , however , deep optical and infrared imaging of bcds have shown that the majority have extensive low surface brightness stellar halos , indicating that most have an underlying older stellar population in addition to bright young stars from the present starburst ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
thus , despite their low luminosity , low metallicity nature , most bcds are not young systems ; rather , the burst phase in bcds is a transitory event , and perhaps cyclical in nature .
this immediately raises the question : where are the quiescent phase bcds today ?
are they included as part of the large , diverse , class of dwarf irregulars ?
do bcds use up their gas and fade into dwarf ellipticals ? or , do they fade beyond detectability , and thus lie hidden and ignored in the vastness of space ?
the possibility of morphological evolution among the dwarf galaxy classifications ( de , di , bcd , in order of increasing star formation activity ) has been suggested many times over the last several decades ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ) , with a long list of positive and negative attributes for each of the various possible transformations .
one of the popular evolutionary scenarios is that the starburst phase uses up ( or blows out , * ? ? ?
* ) the interstellar medium . as the young stars fade , the system will then evolve into a gas poor , red , dwarf elliptical .
alternatively , if the gas is not entirely consumed , the system could continue to form stars at a much lower rate , and thus be classified as a gas rich dwarf irregular .
the possibility of evolutionary connections between bcds , dis , and des has largely been based on the remarkable similarity in their stellar distributions . while exceptions exist , the majority of these low mass objects are well fit by exponential disks ; furthermore , in a plot of magnitude vs. surface brightness , the de class is distinct from the giant elliptical class , but overlaps with the dis ( e.g. , * ? ? ?
the similarity in stellar distributions ( but not current stellar populations or gas content ) suggested that des could be faded dis , or in some cases , such as the nucleated dwarf ellipticals , faded bcds .
recently , detailed kinematic studies of bcds , des , and dis has shed new light on the interrelationship between these low mass galaxies .
high resolution hi kinematic studies of a number of bcds indicate that these are rotation dominated systems ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in contrast , stellar kinematic observations of des find little to no evidence of rotational support @xcite ; if the gas and stars are kinematically coupled in bcds , these results suggest that bcds can not evolve into des because of angular momentum conservation considerations .
however , the bcd classification covers a large range of morphological types of `` host '' galaxies ( e.g. , * ? ? ?
* ) , and the previous samples did not include systems that were necessarily likely to evolve into des .
thus , finding rotation in those bcds does not preclude this scenario for all types of bcds . in order to test directly the feasibility of the popular bcd to de evolutionary scenario ,
we have obtained hi synthesis observations of a small sample of bcds which appeared to be `` hosted '' by dwarf elliptical like systems ; that is , this sample contains the systems that are _ most likely _ to be morphologically similar to des when their starburst fades . in this paper
, we present the results of high spatial resolution hi synthesis observations of six bcds .
we discuss the sample selection and observational procedures in section 2 .
the hi distributions and kinematics of the six galaxies are presented in section 3 .
a discussion of the neutral gas dynamics and its implications for bcd evolutionary scenarios is presented in section 4 ; section 5 contains a brief summary of the conclusions .
multiconfiguration hi observations of six bcds were obtained with the very large array . in this section , the sample selection and procedures for observation and data reduction are discussed .
the bcd classification covers a large range of morphological types of `` host '' galaxies . within the bcd class ,
there are a handful of galaxies which appear to be hosted by dwarf ellipticals ; that is , aside from the central starburst , the underlying stellar distribution has smooth , elliptical isophotes .
these bcd / des are the _ most likely _ subclass of bcds to look like des after the starburst fades .
we thus selected a small sample of bcd / des in order to investigate their gas kinematics and distributions .
the galaxies in the present sample were selected from a large hi survey of blue compact dwarf galaxies @xcite .
physical properties of the selected galaxies are summarized in table [ tab : global ] .
the distance to each system was calculated from the systemic velocity using a virgocentric infall model based on the outline of @xcite and an h@xmath0 of 75 km s@xmath1 mpc@xmath1 .
optical diameters and luminosities were taken from the tabulation of @xcite .
the absolute blue magnitudes have been corrected for galactic extinction , but not for internal extinction or for nebular contributions to the broadband luminosity .
all six galaxies are at comparable distances and have similar luminosities .
their oxygen abundances , however , span a wide range @xcite .
the l band continuum flux listed in table [ tab : global ] was measured from the line free channels of the hi observations ( see section [ sec : hidata ] ) ; while the hi observations were not optimized for continuum detection , weak 20-cm continuum emission was found associated with all six bcds .
observations of um 38 , um 323 , mk 1418 , mk 750 , mk 900 , and mk 324 were obtained with the vla in its b and cs configurations during 19982000 .
as currently implemented , the cs configuration is a modified version of the c configuration where a telescope from the middle of the north arm is re positioned at an inner d configuration station to provide improved coverage of the short spacings .
a summary of the observing sessions and the total on source integration times are listed in table [ tab : vlaobs ] . during all
observing sessions , the correlator was used in 2ad mode with the right and left circular polarizations tuned to the same frequency ; the total bandwidth was 1.56 mhz .
the on line hanning smoothing option was selected , producing final spectral data cubes of 127 channels , each 2.6 km s@xmath1 wide .
standard tasks in aips @xcite were employed for calibration and preliminary data reduction .
each set of observations was calibrated separately , using 3c 48 or 3c 286 as the flux and bandpass calibrator and nearby continuum sources as phase calibrators . for all except mk 324 ,
continuum emission was removed with the aips task uvlin after the combination of _ u v _ data sets and prior to transformation to the _ x y _ plane .
the line data were transformed to the image plane with several weighting schemes and combinations of configurations . to check the data quality ,
maps were made for each observing session .
the data cubes presented in this paper were made from the combined data sets of the b and cs configuration observations .
a robust weighting technique was employed by the aips task imagr to optimize the beam shape and noise levels @xcite .
the `` robustness parameter '' in imagr controls the weighting of the _ u v _ data , permitting a fine tuning between sensitivity and resolution in the final data cubes .
as currently implemented , a robustness of 5 corresponds to natural weighting of the _ u v _ data ( maximizing sensitivity at the cost of spatial resolution ) while a robustness of 5 corresponds to uniform weighting ( maximal spatial resolution with lower sensitivity ) .
additionally , some maps were made with _
_ tapers to increase their sensitivity to low column density material .
the relevant imagr parameters for a selected sample of the maps are listed in table [ tab : maps ] . throughout this paper
we will refer to the lowest resolution , tapered data cubes as the `` tapered cubes '' , to the robustness of 5 cubes as the `` natural weight cubes '' , and to the robustness of 0.5 cubes as the `` intermediate weight cubes . ''
all subsequent analysis was performed within the gipsy package @xcite .
the reduction stream for the mk 324 observations differed slightly from the above synopsis .
first , line emission from mk 324 and exg 2323 + 1816 filled much of the bandpass , making it impossible to remove the continuum emission in the _ u v _ plane .
a continuum image was created for each data cube by averaging the few line free channels ; this image was then subtracted from the original data cube .
second , due to the low signal to noise of the line emission in mk 324 , the data were smoothed during the transformation to the _ x y _ plane by averaging 2 consecutive channels , resulting in a cube of 63 channels , each 5.2 km s@xmath1 wide . to determine if the total flux density was recovered in the hi synthesis observations , total integrated profiles were constructed from the natural weight data cubes after correcting for the primary beam shape with the gipsy task pbcorr ( figure [ fig : flux ] ) .
for all systems , the total flux density recovered in the vla maps is in good agreement ( within 10% ) with previous single dish flux measurements ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
the total hi masses listed in table [ tab : global ] were calculated from the observed vla flux density and the assumed distance to the system .
moment maps of each data cube were computed in the following manner . for each galaxy , both
the tapered and natural weight cubes were smoothed to a resolution of twice the beam ; the smoothed cubes were clipped at the 2@xmath2 level ; the resultant clipped cubes were then interactively blanked to remove spurious noise spikes .
signal was identified based on spatial continuity between channels .
a conditional transfer was used to blank the corresponding locations in both of the original data cubes , and in the intermediate weight data cube , after correcting for primary beam attenuation .
moment maps of the tapered , natural , and intermediate weight data cubes were created with the gipsy task moments and are presented in figures [ fig : um38][fig : mk324 ] . in the process of making moment maps ,
each data cube was searched for possible companion galaxies .
nearby systems were identified in the lowest resolution data cubes of both um 323 and mk 324 ( figures [ fig : um323comp ] and [ fig : mk324comp ] ) ; we designate these objects as exg 0123 - 0040 and exg 2323 + 1816 , respectively .
both exg 0123 - 0040 and exg 2323 + 1816 have good positional coincidences with faint optical galaxies .
other hi synthesis observations of um 323 also identified an hi signal near the location of exg 0123 - 0040 and at a similar velocity , designated as um 323a @xcite ; while this original detection was subsequently retracted ( see * ? ? ?
* ) , this is probably the same object .
note that the present data set is completely independent from the previous d configuration observations of @xcite .
the integrated hi profiles of these systems are shown in figure [ fig : flux ] ; the observed hi flux densities , systemic velocities , and velocity widths are tabulated in table [ tab : others ] .
the primary beam correction may underestimate the total hi flux density of these systems since both of these galaxies are located slightly beyond the half power point of the main beam ( see d@xmath3 in table [ tab : others ] ) .
both systems are at a projected distance of @xmath4130 kpc from the bcd , and have similar hi masses ( 30% and 70% of the target galaxy , respectively )
. these systems may not be either sufficiently massive , or close enough , to have `` triggered '' the current burst of star formation in um 323 and mk 324 , but the continuing detection of faint , but gas rich , systems which have been missed by optical surveys ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) suggests that hi observations are a crucial step in the process of determining the local environment of a galaxy .
complementary optical images were obtained during several observing runs at kpno .
h@xmath5 and b band images of um 323 , mk 750 , mk 900 , and mk 324 were obtained in 19891990 with the kpno no . 1 0.9 m telescope
; the observation and data reduction procedures for these images were similar to those described in @xcite .
h@xmath5 and b band images of um 38 , mk 900 , and mk 324 were obtained with the kpno 0.9 m telescope in june 1998 ; the observation and reduction procedures for these images are described in @xcite .
since no optical images were available for mk 1418 , or for the extended fields around um 323 and mk 324 , images of these fields were extracted from the digitized sky survey .
plate solutions for the optical images were derived from coordinates of at least five stars listed in the apm catalogapmcat/ ] and are accurate to 0.5 .
star formation rates were calculated from the h@xmath5 images of um 38 , um 323 , mk 750 , mk 900 , and mk 324 .
first , the observed h@xmath5 flux was corrected for internal and external absorption based on the observed spectroscopic balmer line ratios @xcite .
the derived h@xmath5 luminosity was then converted to the current star
formation rate ( sfr ) , using the conversion factor from @xcite : @xmath6 the current sfr and the gas depletion time scale ( m@xmath7/sfr ) are tabulated in table [ tab : global ] . in general , the current sfrs of the bcds are quite high , with typical gas depletion times of approximately 1 - 5 gyr . however , it is important to note that the star formation rate conversion factor used here may lead to an overestimate of the star formation rate in low metallicity galaxies , since the uv opacity will be lower in metal poor stars .
thus , it is possible that the gas depletion time scales quoted here are underestimates .
the neutral gas distributions of the six bcds are shown in figures [ fig : um38][fig : mk324 ] . as has been seen in other observations of bcds ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the hi distribution extends well beyond the optical system . the hi to optical diameters ( measured at the 10@xmath8 atoms @xmath9 and 25 mag arcsec@xmath10 isophotes , respectively )
are tabulated in table [ tab : global ] . for all except mk 900
, the neutral gas extends approximately 2 times the optical diameter , which is quite typical for gas rich galaxies ( see , e.g. , * ? ? ?
* ; * ? ? ?
both mk 1418 and mk 324 have extended gaseous distributions which show kinematic peculiarities , reminiscent of tidal tails .
optically , these galaxies have smooth , regular isophotes , so it is not clear if they are merger remnants , recently had an interaction , or if some other mechanism produced their extended gaseous disks .
overall , the neutral gas distribution in these bcd / des is quite similar to that found in other bcds .
azimuthal averages of the gas distributions are shown in figure [ fig : gas ] for the present sample , a random sample of bcds @xcite , and a sample of gas rich dis @xcite .
the gas is much more centrally concentrated in the bcds [ see also @xcite and @xcite ] , which presumably facilitates their high star formation rates .
more specifically , the neutral gas is concentrated in the regions of star formation activity .
slices through the neutral and ionized gas images of um 38 , um 323 , mk 750 , mk 900 , and mk 324 are shown in figure [ fig : slice ] .
the position angle of each slice was chosen so that the slice would intersect the brightest hii regions in each galaxy . as was seen with other bcds @xcite , the neutral gas density peaks coincide with regions of star formation .
however , in the highest spatial resolution observations there is some evidence of feedback between the star forming regions and the neutral medium .
for instance , the brightest hii region in um 38 is actually located in a slight depression in the neutral gas surface density , as would be expected if the neutral medium has been partially used , or ionized , by the star formation activity .
the observed peak column densities are tabulated in table [ tab : colden ] . since the observed column density depends on the beam dilution ,
the value determined in each data cube is presented . in general , the observed peak column densities are between 24 @xmath11 10@xmath12 atoms @xmath9 ; when corrected for inclination and neutral helium content , the peak neutral gas surface densities are on the order of 3 @xmath11 10@xmath12 atoms @xmath9 for all six galaxies .
interestingly , these peak values are very similar to the predicted critical density as calculated from the @xcite instability criterion ( e.g. , * ? ? ?
* ) .
the velocity fields are shown in figures [ fig : um38]d[fig : mk324]d . in all cases ,
a clear velocity gradient is present ; interestingly , the velocity gradient is aligned along the optical major axis of only mk 900 and mk 324 . both um 38 and mk 900 have relatively smooth , ordered velocity fields while those for um 323 , mk 1418 , mk 750 and mk 324 are significantly more complicated .
less than perfect ordered rotation has also been found in other high spatial resolution observations of bcds ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . nonetheless , despite the kinematic complexities
, it is clear that these systems all have large rotation velocities .
model rotation curves for the bcds were derived from tilted ring analysis of the observed velocity fields using the gipsy task rotcur .
the process of fitting a rotation curve is iterative .
first , the kinematic center and systemic velocity were determined using both sides of the velocity field to constrain these global parameters .
next , the center and systemic velocity were held fixed while the inclination angle and position angle were derived .
unfortunately , the complex nature of the velocity fields imply a large degree of uncertainty in the kinematic parameters ; for all except um 38 , the derived kinematic parameters ( table [ tab : hiparms ] ) should only be considered rough estimates . in particular ,
it was not possible to constrain the inclination angle based on the velocity field alone ; an estimate of the inclination angle was derived from the observed axial ratios ( both hi and optical ) .
finally , despite the apparent solid body nature of the velocity fields , the final rotation curves produced by rotcur had a tendency to flatten in the outer parts ; thus , in the end , only um 38 was satisfactorily fit by this method .
nonetheless , the rotcur fits were sufficient to indicate the major axis of each galaxy so that position velocity diagrams could be created ( figure [ fig : pv ] ) . as expected from the velocity fields , the p
v diagrams indicate solid body rotation throughout the gas disk for um 323 , mk 1418 , mk 750 , mk 900 , and mk 324 , while um 38 appears to be differentially rotating in the outer disk . also shown in figure
[ fig : pv ] is the expected observed rotation curve from the rotcur model for um 38 , and estimates of solid body rotation curves for the other galaxies .
in fact , the p
v diagrams provide a good estimate of the rotation curves for um 323 , mk 1418 , mk 750 , mk 900 , and mk 324 , assuming that no warps or other kinematic abnormalities are present .
this is probably a reasonable assumption for mk 900 , but the derived kinematic parameters for the other four galaxies are significantly more uncertain . the radii and velocities ( corrected for inclination ) of the last measured points of the rotation curves are tabulated in table [ tab : hiparms ] . since the rotation curves were poorly constrained for all except um 38 and mk 900 , these values are only rough estimates .
if the bcds are dark matter dominated , as has been found for virtually all dwarf galaxies studied ( e.g. , * ? ? ?
* ) , the dynamical mass can be calculated from the derived rotation curves . for a spherical mass distribution ( appropriate for most dark matter halos ) ,
the dynamical mass within radius @xmath13 is : @xmath14 where @xmath15 is in km s@xmath1 and @xmath13 is in kpc .
the derived dynamical masses for the last measured point of the rotation curves are tabulated in table [ tab : hiparms ] . the hi to dynamical mass ratios for um 38 ,
mk 1418 , mk 750 and mk 900 are quite typical for dwarf galaxies in general ( e.g. , * ? ? ?
* ; * ? ? ?
* ) while those of um 323 and mk 324 are significantly higher than average .
however , both um 323 and mk 324 had extremely complex velocity fields , so this may simply reflect on the poorly derived rotation curves , rather than an intrinsic difference in galaxy properties . an estimate of the neutral gas velocity dispersion was obtained from gaussian fits to the spectral profiles in the natural weight data cubes .
the median @xmath2 is tabulated in table [ tab : hiparms ] .
velocity dispersions are quite typical of dwarf galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
in all cases , the observed velocity dispersions are significantly less than the rotational component in these galaxies , indicating that these systems are rotation dominated .
based on these and other observations of bcds and dis ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , it appears that all types of moderate luminosity gas rich dwarf galaxies are rotation dominated systems . in particular , as illustrated in figure [ fig : pv ] , the bcd / des show significant velocity gradients across the optical and hi disks , which implies that the gaseous disks have significant angular momenta .
the specific angular momenta for a given radius in the disk can be calculated from the derived rotation curves : @xmath16 where @xmath17 is the specific angular momenta , @xmath18 is the rotation velocity , and @xmath19 is the radius .
it is difficult to specify a `` characteristic '' value for the specific angular momentum of the disk in low mass galaxies because the specific angular momentum increases with radius for galaxies undergoing solid body rotation .
however , it is useful to compare the specific angular momenta of the outermost neutral hydrogen gas in bcds and low surface brightness ( lsb ) dis .
the specific angular momenta of the extended hi disks were computed from the maximum rotation velocities and radii tabulated in table [ tab : hiparms ] ( bcds ) and in table 4 of @xcite ( lsbdgs ) . while the definition of `` the outermost '' gas is sensitivity dependent , the two studies shown in figure [ fig : angmom ] have similar sensitivities , and the galaxies have similar hi to optical ratios . as illustrated in figure [ fig : angmom]a , the bcds have lower specific angular momenta than the lsb dis , as one would expect for rotation curves which reach similar maximum velocities at smaller radii ( figure [ fig : angmom]b ) . since the rotation curves of dwarf galaxies are primarily solid body , the fact that bcds reach a similar maximum velocity within a smaller radius implies that bcds have steeply rising rotation curves .
the fact that bcds appear to have steeply rising rotation curves ( low angular momenta ) has several interesting consequences for evolutionary studies of bcds .
first , the shape of the rotation curve suggests that the matter distribution is centrally concentrated ( see also * ? ? ?
* ) . as expected for galaxies which are classified as `` compact , '' studies of their stellar distributions
indicate that the stars are centrally concentrated in bcds @xcite ; similar results are found for the gaseous distribution as well ( figure [ fig : gas ] ) .
however , the rotation curve results also suggest that the dark matter distributions may be centrally concentrated in bcds .
given the poor constraints on the derived rotation curves in this sample , we have elected to forgo extensive analysis at this time ( i.e. , dark matter decompositions ) .
however , more extensive analysis could place strong constraints on dark matter models , since dwarf galaxies are usually dark matter dominated .
bcds clearly inhabit an important , and as yet unexplored , region of parameter space .
a second major implication from the observed rotation curve shapes ( angular momenta ) is that bcds have higher threshold densities for the onset of star formation . using the classic @xcite instability analysis for a rotating disk ,
the threshold density for star formation can be derived from the balance between gravitational collapse and the effects of both thermal pressure and rotational shear . for a solid body rotation curve , this type of analysis results in a star formation threshold density which is proportional to the slope of the rotation curve . in other words ,
a galaxy with a steep rotation curve ( low specific angular momenta ) will have a higher threshold density for star formation . while the instability threshold is a global dynamical property
, it is particularly important to note that the current observed gas column densities appear to hover around the instability threshold densities in both the bcds ( this paper and * ? ? ? * ) and dis @xcite , and that the local column density peaks are correlated with sites of current star formation activity .
thus , there appears to be a strong connection between crossing the global instability threshold and the onset of star formation , a local process .
angular momentum appears to play a crucial role in determining the ultimate star formation history of a low mass galaxy .
the gas associated with a galaxy with high angular momentum will have difficulty collapsing , and thus at the present day it will appear as a diffuse , low surface brightness galaxy . on the other hand , the gas associated with galaxies with low angular momenta will be able to collapse further , leading to smaller , compact , high surface brightness galaxies . concurrently , in these compact systems , the onset of star formation may be delayed because the threshold density is extremely high .
for example , if the infalling gas accretes at a steady rate , it will take longer for the gas density to exceed the critical density in a galaxy with lower angular momenta .
further , once the gas density does exceed the critical density , there will be a large fuel supply readily available for star formation .
it is thus quite likely that these low angular momenta , high density systems will be more susceptible to a `` burst '' mode of star formation than similar mass , but high angular momenta , low density systems .
it is also logical to hypothesize that starbursts will occur episodically in these compact systems , with the starburst duty cycle regulated by the frequency at which the central gas density reaches the critical density for star formation .
as mentioned in the introduction , one of the key questions in dwarf galaxy evolution is the ultimate fate of the star bursting dwarf galaxies .
since the starburst phase in bcds is expected to be a transitory event which lasts only a few 100 myr , one logical hypothesis is that bcds must evolve from ( and into ) other types of low mass galaxies .
several of the major unresolved issues in dwarf galaxy evolutionary scenarios are : ( 1 ) do all dwarf galaxies undergo a starburst phase ?
( 2 ) what mechanisms initiate or trigger a starburst event ? ( 3 ) what mechanisms terminate a starburst event ? and ( 4 ) how does the galaxy evolve after the burst phase ? for the last 30 years , most discussions of dwarf galaxy evolutionary scenarios have centered around the morphological similarities between starbursting dwarf galaxies , dwarf irregular galaxies , and dwarf elliptical galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these studies have shown that the stellar distributions of all three types of low mass galaxies are well fit by exponential profiles , with a few notable exceptions ( e.g. , * ? ? ?
* ; * ? ? ?
thus , based on the stellar distributions alone , it has been argued that bcds can evolve into either the dwarf elliptical class or dwarf irregular class with only minor changes in the stellar distribution and optical morphology .
arguments against such passive morphological evolution usually focus on the fact that the stellar distributions of bcds are much more compact than the typical dwarf irregular or dwarf elliptical galaxy ; however , since there is some overlap in the distribution functions for all three type of galaxies , such morphological evolution can not be excluded based on the stellar distributions alone . in the next two sections we outline the arguments for and against the canonical bcd evolutionary scenarios
first , in section 4.2.1 , we discuss whether gas rich bcds can evolve into gas poor dwarf ellipticals .
second , in section 4.2.2 , we discuss whether post burst bcds can be found within the general population of gas rich dwarf irregular galaxies .
a similar discussion with a focus on the evolution of dwarf irregular galaxies may be found in @xcite . secular evolution from bcd
to de requires more than just fading of the central starburst .
all of the dwarf ellipticals in the local group are gas poor ( with m@xmath20 , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , while the bcds are gas rich systems , with m@xmath21 .
thus , evolution from bcd to de requires either efficient gas consumption or removal of the remaining interstellar medium . the latter process can be accomplished in at least two distinct ways : ( 1 ) @xcite proposed that a starburst might have sufficient kinetic energy to disrupt the interstellar medium in a low mass galaxy , and ( 2 ) @xcite proposed that the ism might be removed via ram pressure stripping .
hydrodynamic models initially indicated that a starburst could have a disastrous effect on the interstellar medium within a low mass galaxy ( e.g. , * ? ? ?
* ; * ? ? ?
* ) ; however , recent models , which include a large dark matter component , have shown that it is much more difficult to remove the interstellar medium with a single starburst than previously thought ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
rather , while some of the hot , ionized gas may escape ( as possibly seen in ngc 1569 , * ? ? ?
* ) , the neutral medium is largely unaffected by the starburst . further , detailed star formation histories of the local group des indicate that these dwarf galaxies appear to have had several discrete star
formation episodes @xcite , rather than the single burst envisioned by @xcite .
thus , it now appears that it is unlikely that a single starburst will completely use up , or blow out , the interstellar medium of a dwarf galaxy .
ram pressure stripping , on the other hand , is still a viable mechanism to remove the remnants of the ism in a post burst low mass galaxy .
ram pressure stripping will occur predominantly in high density regions , the exact type of environment favored by dwarf elliptical galaxies .
however , this process will operate regardless of the star formation activity in the low mass galaxy ; thus , while this is a viable mechanism to transform a gas rich star forming galaxy into a gas poor evolved galaxy , it is tangential at best to the main discussion of the evolutionary fate of star bursting dwarf galaxies .
the neutral gas kinematics presented here and in @xcite provide the first definitive evidence that the majority of bcds will not evolve passively into dwarf elliptical galaxies .
all of the bcds studied so far are rotation dominated and have significant angular momenta .
in contrast , dwarf elliptical galaxies are not rotationally supported @xcite . of the 5 dwarf ellipticals
observed , only one has a maximum stellar rotation velocity @xmath22 15 km s@xmath1 ; the remainder are consistent with zero stellar rotation .
as mentioned previously , the bcds in the present study were selected specifically to have optical properties similar to dwarf elliptical galaxies , since these would be the _ most likely _ to be classified as dwarf ellipticals after the starburst ; however , even these galaxies have significant rotation velocities .
thus , these observations indicate that bcds can not evolve passively into dwarf elliptical galaxies since such a morphological evolution would require the loss of angular momentum .
further , the possibility that ram pressure stripping may be a necessary process to remove the interstellar medium during the formation of dwarf elliptical galaxies leads directly to another strong argument against generic morphological evolution from a starbursting dwarf galaxy to a dwarf elliptical ( such as the one envisioned by * ? ? ?
* ) : bcds and des have very different clustering properties .
dwarf ellipticals are generally found in clusters or as companions to massive galaxies while bcds are typically field objects .
in fact , des are among the most strongly clustered galaxies known @xcite , while bcds are among the least clustered @xcite .
thus , it may be necessary to consider the evolution of field and cluster dwarf galaxies separately : bcds formed in the field will probably retain a substantial fraction of their ism , even after the starburst phase , and thus will be classified as dwarf irregular galaxies if identified either pre or post burst ; meanwhile , some bcds formed in cluster environments may evolve into dwarf elliptical galaxies , but such morphological evolution will not occur via passive secular evolution .
one of the difficulties for discussions of bcd to di evolutionary scenarios is a nomenclature issue : in many regards , bcds _ are _ dis . since the dwarf irregular classification is usually based on irregular morphology and gas richness , star
bursting dwarf galaxies are included in this more general classification .
thus , rather than true morphological evolution , the bcd to di evolutionary scenario is a matter of degree : can post burst bcds be found within the general dwarf irregular class , and also , is a starburst phase common to all gas rich low mass galaxies ?
these questions have been addressed extensively by @xcite , who argues that the majority of dis do not go through a starburst phase .
rather , the star formation histories of most dwarf irregular galaxies are best described by a low level of star formation activity which percolates across the gaseous disk . however , large imaging studies of dwarf irregular galaxies have shown that there are a few dis which have structural parameters similar to the bcd class ( e.g. , * ? ? ?
* ; * ? ? ?
these `` compact dis '' have similarly small scale lengths for their luminosity , but do not have extremely high star formation rates at the present time .
thus , these galaxies are likely to be examples of post burst bcds .
indeed , unlike the other dwarf irregular galaxies in the samples , the `` compact dis '' have detectable color gradients from the inner to outer regions of the stellar disk , similar to those found in samples of starbursting dwarf galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the kinematic data presented here provides additional evidence that bcds do not represent a common phase in the star formation history of gas rich low mass galaxies .
bcds have steeper rotation curves and lower specific angular momenta than `` typical '' dwarf irregular galaxies ( figures 12 and 13 , see also * ? ? ?
* ) . however , this is not to say that there is no overlap between bcds and dis in this regard as well . in order to accentuate the differences between bcds and dis
, we have focussed here on examples taken from the extremes of both populations ( lsb and hsb ) .
in fact , dis embody a continua of properties , from the low surface brightness regime to the high surface brightness regime , and with a wide range of star formation rates and gas content .
thus , the marked differences seen in figures [ fig : gas ] and [ fig : angmom ] should not be used to infer that post
burst bcds are morphologically and dynamically distinct from general di classification .
instead , it appears that post
burst bcds can be found within the general di classification , but rather than being a necessary phase in dwarf galaxy evolution , starbursting episodes are rare events . as the first structures to collapse in hierarchical galaxy formation models , dwarf galaxies play a crucial role in galaxy formation .
however , the interplay between structure formation , the onset of star formation activity , and the subsequent evolution of baryonic matter is complex and not well understood at the present time .
most cosmological models treat the physics of the dark matter particles in great detail , while the complex processes of star formation and gas dynamics are applied as simplified `` recipes '' ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
unfortunately , this broad brush approach to star formation means that all dwarf galaxies are treated in a similar manner , regardless of their physical attributes ; further , none of the current models have the spatial resolution to track the gas distribution in individual galaxies , and thus they can not determine if a galaxy is likely to form stars in a large burst of star formation ( centrally concentrated gas distributions ) , or in a more quiescent and constant manner ( low density gas distributions ) .
ultimately , the lack of spatial resolution and the poor treatment star formation processes means that current cosmological models can not differentiate between low surface brightness dwarf galaxies and high surface brightness dwarfs .
as discussed in section 4.2.2 , low surface brightness dwarf galaxies and high surface brightness ( starbursting ) dwarf galaxies probably represent the extreme ends of a continua of dwarf galaxy properties .
one of the critical parameters which still needs to be determined is the relative number density of starbursting dwarf galaxies to quiescent dwarfs . in particular , bcds have gained attention recently due to the identification of bursting systems at intermediate and high redshifts .
@xcite proposed that very intense and short bursts of star formation in low mass galaxies ( similar to the present day bcds ) could be responsible for the faint blue galaxy excess found in studies of high redshift luminosity functions ( e.g. , * ? ? ? * ) .
however , if bcds represent the extreme members of the dwarf irregular class , and occur only rarely , then only a small fraction of dwarf galaxies will go through a starburst phase . as a consequence ,
the number density of starbursting dwarf galaxies as a function of redshift may be lower than previously thought .
the possibility that there may be fewer starbursting dwarf galaxies at high redshift also has severe implications for galaxy evolution models .
most cosmological models assume that the net effect of any star formation activity in a dwarf galaxy is the immediate injection of heavy elements into the intergalactic medium via galactic winds and other `` feedback '' processes ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these models predict rapid enrichment of the intergalactic medium , as low mass halos ( dwarf galaxies ) collapse in the early universe .
if the majority of dwarf galaxies do not go through a starburst phase , however , this enrichment process may be less efficient than currently assumed .
in summary , we have presented the results of high spatial resolution hi synthesis observations of six bcd / des .
our results and conclusions are summarized below .
\(1 ) all six of the bcds in this sample are rotation dominated systems .
the velocity fields and position velocity diagrams clearly indicate that these galaxies have significant velocity gradients across the gaseous disks ( figure [ fig : pv ] ) .
\(2 ) as in previous studies of bcds ( e.g. , * ? ? ?
* ) , the neutral gas density is centrally concentrated ( figure [ fig : gas ] ) , which presumably facilitates the high star formation rate in these compact galaxies .
\(3 ) starbursting dwarf galaxies appear to have lower specific angular momenta than similar luminosity low surface brightness dwarf irregular galaxies ( figure [ fig : angmom ] ) .
the intrinsic angular momenta differences may explain both the compact nature of bcds and their tendency to have starburst episodes .
\(4 ) the observed gas dynamics places strong constraints on the evolutionary fate of starbursting dwarf galaxies . based on their gas kinematics , bcds can not evolve passively into dwarf elliptical galaxies .
in contrast , both bcds and dis are rotationally supported systems , and thus it is likely that post
burst bcds can be found within the general dwarf irregular classification .
these observations place the first strong kinematic constraints on bcd evolutionary scenarios .
the fact that the gaseous disks are supported by rotation , even for these most `` de like '' of the bcd class , implies that starbursting dwarf galaxies will not evolve passively into dwarf elliptical galaxies .
however , since it is possible to transfer angular momentum via merging or interactions between galaxies , bcds in high density regions could evolve into dwarf ellipticals via a more catastrophic evolutionary scenario .
in fact , the formation of dwarf elliptical galaxies via ram pressure stripping of either gas rich dwarf irregular galaxies or star bursting dwarf galaxies is further substantiated by the strong morphology density relationship in low mass galaxies .
one potential concern is that the kinematic constraints presented here are for the gaseous disks of the starbursting dwarf galaxies , while the kinematic data for dwarf elliptical galaxies is almost entirely derived from observations of the stellar population . if the stars and gas are kinematically de coupled in bcds , the angular momentum arguments presented here are less compelling . in this regard
, it is important to note that of the few dwarf ellipticals which have been mapped in hi , at least one , ngc 205 , has a detectable velocity gradient in the gas disk @xcite which is not detected in the stellar motions @xcite .
thus , further observations of the stellar kinematics of bcds will be necessary to definitely exclude the possibility that bcds evolve passively into dwarf elliptical galaxies when the starburst fades .
nonetheless , the observed dynamics of bcds appear to favor the hypothesis that post burst bcds should be found within the large , diverse , dwarf irregular classification .
the definitive test of the bcd to di evolutionary scenario will be to find examples of post burst bcds which have been classified as dwarf irregular galaxies .
studies of the structural parameters of dwarf irregular galaxies suggest that the majority of dis are not related to the bcd class , but that a small percentage have similar morphology and color gradients @xcite . if these compact dis are post
burst bcds , they should also have similar gas distributions and gas dynamics as the bcds .
thus , future hi synthesis observations of compact dis should provide a direct test of the bcd to di evolutionary scenario , and definitively determine if bcds fade gracefully into more quiescent dis , or if an altogether different , heretofore unknown , fate lies ahead .
we acknowledge the financial support by nsf grant ast9553020 to jjs .
partial support from a nasa ltsarp grant no .
nagw3189 ( eds ) is gratefully acknowledged .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
bender , r. , & nieto , j.l . 1990 , , 239 , 97 bender , r. , paquet , a. , & nieto , j.l . 1991 , , 246 , 349 briggs , d. 1995 , phd thesis , new mexico tech broeils , a. h. , & rhee , m.h .
1997 , , 324 , 877 broeils , a. h. , & van woerden , h. 1994 , , 107 , 129 carignan , c. , & beaulieu , s. 1989 , , 347 , 760 cole , s. , aragon salamanca , a. , frenk , c. s. , navarro , j. f. , & zepf , s. e. 1994 , , 271 , 781 cole , s. , lacey , c. g. , baugh , c. m. , & frenk , c. s. 2000 , , 319 , 168 cowie , l. l. , songalia , a. , & hu , e. m. 1991 , , 354 , 460 davies , j. i. , & phillipps , s. 1988 , , 233 , 553 de young , d. s. , & heckman , t. m. 1994 , , 431 , 598 de young , d. s. , & gallagher , j. s. 1990 , , 356 dekel , a. , & silk , j. 1986 , , 303 , 39 doublier , v. , caulet , a. , comte , g. 1999 , , 138 , 213 doublier , v. , comte , g. , petrosian , a. , surace , c. , & turatto , m. 1997 , , 124 , 405 drinkwater , m. , & hardy , e. 1991 , , 101 , 94 efstathiou , g. 2000 , , 317 , 697 ferguson , h. c. , & babul , a. 1998 , , 296 , 585 ferguson , h. c. , & binggeli , b. 1994 , , 6 , 67 ferguson , h. c. , & sandage , a. 1989 , , 346 , 53 heckman , t. m. , dahlem , m. , lehnert , m. d. , fabbiano , g. , gilmore , d. , & waller , w. h. 1995 , , 448 , 98 heyl , j. s. , cole , s. , frenk , c. s. , & navarro , j. f. 1995 , , 274 , 755 hoffman , g. l. , helou , g. , salpeter , e. e. , & lewis , b. m. 1989 , , 339 , 812 huchtmeier , w. k. , & richter , o. g. 1986 , , 63 , 323 iovino , a. , melnick , j. , & shaver , p. 1988 , , 330 , l17 izotov , y. i , & thuan , t. x. 1999 , , 511 , 639 james , p. a. 1994 , , 269 , 176 kauffmann , g. , white , s. d. w. , & guiderdoni , b. 1993 , , 264 , 201 kennicutt , r. c. 1989 , , 344 , 685 kennicutt , r. c. 1998 , , 36 , 189 kormendy , j. 1985 , , 295 , 73 lee , j. c. , salzer , j. j. , law , d. a. , & rosenberg , j. l. 2000 , , 536 , 606 lin , d. n. c. & faber , s. m. 1983 , , 266 , l21 loose , h.h . , & thuan , t. x. 1985 , in `` star forming dwarf galaxies and related objects '' kunth , d. , thuan , t. x. , & vn , t. t. eds .
paris : editions frontires , p. 73
loose , h.h .
, & thuan , t. x. 1986 , , 309 , 59 mac low , m.m . , & ferrara , a. 1999 , , 513 , 142 marlowe , a. t. , heckman , t. m. , wyse , r. f. g. , & schommer , r. 1995 , , 438 , 563 mateo , m. l. 1998 , , 36 , 435 meurer , g. r. , carignan , c. , beaulieu , s. f. , & freeman , k. c. 1996 , , 111 , 1551 meurer , g. r. , staveley smith , l. , killeen , n. e. b. 1998 , , 300 , 705 moore , b. 1994 , , 370 , 629 napier , p. j. , thompson , r. t. , & ekers , r. d. 1983 , proc .
ieee , 71 , 1295 norton , s. a. , & salzer , j. j. 2001 , in prep papaderos , p. , loose , h.h . , fricke , k. j. , & thuan , t. x. 1996a , , 314 , 59 papaderos , p. , loose , h.h .
, thuan , t. x. , & fricke , k. j. 1996b , , 120 , 207 patterson , r. j. , & thuan , t. x. 1996 , , 107 , 103 pisano , d. j. , & wilcots , e. m. 1999 , , 117 , 2168 pustilnik , s. a. , ugryumov , a. v. , lipovetsky , v. a. , thuan , t. x. , & guseva , n. g. 1995 , , 443 , 499 putman , m. e. , bureau , m. , mould , j. r. , staveley
smith , l. & freeman , k. c. 1998 , , 115 , 2345 rosenberg , j. l. , salzer , j. j. , & moody , j. w. 1994 , , 108 , 1557 salzer , j. j. , 2000 , private communication salzer , j. j. , di serego aligheri , s. , matteucci , f. , giovanelli , r. & haynes , m. p. 1991
, , 101 , 1258 salzer , j. j. , & norton , s. a. 1999 , in iau colloquium 171 : the low surface brightness universe , ed .
j. i. davies , c. impey , and s. phillipps ( asp conf . series ) , 170 , 253 salzer , j. j. , rosenberg , j. l. , weisstein , e. , mazzarella , j. m. , & bothun , g. d. 2001 , submitted sargent , w. l. w. , & searle , l. 1970 , , 162 , l155 schechter , p. 1980 , aj , 85 , 801 silich , s. a. , & tenorio tagle , g. 1998 , , 299 , 249 silk , j. , wyse , r. f. g. , shields , g. a. 1987 , , 322 , l59 simpson , c. e. , & gottesman , s. t. 2000 , , 120 , 2975 skillman , e. d. , & bender , r. 1995 , rev .
aasc , 3 , 25 skillman , e. d. , bothun , g. d. , murray , m. a. , & warmels , r. h. 1987 , , 185 , 61 somerville , r. s. , & primack , j. r. 1999 , , 310 , 1087 sommer
larsen , j. , gelato , s. , & vedel , h. 1999 , , 519 , 501 sung , e.c .
, han , c. , ryden , b. s. , chun , m.s .
, & kim , h.i . 1998 , , 499 , 140 swaters , r. a. , 1999 , phd thesis , university of groningen taylor , c. l. , brinks , e. , grashius , r. m. , & skillman , e. d. 1995 , , 99 , 427 taylor , c. l. , brinks , e. , grashius , r. m. , & skillman , e. d. 1996 , , 102 , 189 ( erratum ) taylor , c. l. , brinks , e. , pogge , r. w. , & skillman , e. d. 1994 , , 107 , 971 terlevich , r. , melnick , j. , masegosa , j. , moles , m. , & copetti , m. v. f. 1991 , , 91 , 285 thuan , t. x. , izotov , y. i. , & foltz , c. b. 1999 , , 525 , 105 toomre , a. 1964 , , 139 , 1217 van der hulst , j. m. , terlouw , j. p. , begemen , k. , zwitser , w. , & roelfsema , p. r. 1992 , in astronomical data analysis software and systems , ed .
d. worall , c. biemesderfer , and j. barnes ( asp conference series ) , 25 , 131 van zee , l. 2000 , 119 , 2757 van zee , l. 2001 , in press van zee , l. , haynes , m. p. , & giovanelli , r. 1995 , , 109 , 990 van zee , l. , haynes , m. p. , salzer , j. j. , & broeils , a. h. 1997b , , 113 , 1618 van zee , l. , skillman , e. d. , & salzer , j. j. 1998a , , 116 , 1186 van zee , l. , westpfahl , d. , haynes , m. p. , & salzer , j. j. 1998b , , 115 , 1000 walter , f. , brinks , e. , duric , n. , & klein , u. 1997 , , 113 , 2031 white , s. d. m. , & frenk , c. s. 1991 , , 379 , 52 wilcots .
e. m. , lehman , c. , & miller , b. 1996 , , 111 , 1575 young , l. m. 1999 , , 117 , 1758 young , l. m. 2000 , , 119 , 188 young , l. m. , & lo , k. y. 1996 , , 462 , 203 young , l. m. , & lo , k. y. 1997 , , 476 , 127 | we present the results of high spatial resolution hi synthesis observations of six blue compact dwarf ( bcd ) galaxies .
optically , the selected galaxies have smooth , symmetric isophotes , and thus are the most likely of the bcd class to fade into an object morphologically similar to a dwarf elliptical when the current starburst ends .
the neutral gas in all six galaxies appears to be rotationally supported , however , indicating that true morphological transformation from a bcd to a de will require significant loss of angular momentum .
based on the observed neutral gas dynamics of these and other bcds , it is unlikely that present day bcds will evolve directly into dwarf ellipticals after a starburst phase .
we discuss alternative evolutionary scenarios for bcds and place them within the larger context of galaxy formation and evolution models . in general , bcds appear to have steeper rotation curves than similar luminosity low surface brightness dwarf galaxies .
bcds have centrally concentrated mass distributions ( stars , gas , and dark matter ) and have lower specific angular momenta . based on disk instability analyses , steeply rising rotation curves result in higher threshold densities for the onset of star formation .
these results suggest that angular momenta may play a crucial role in the morphological evolution of low mass galaxies : galaxies with low angular momenta will be able to collapse into small , compact galaxies while galaxies with high angular momenta will be more diffuse systems .
further , because the star formation threshold density is higher in low angular momenta systems , star formation will be delayed until an extremely high surface density is reached .
thus , angular momentum may be the fundamental parameter which determines if a low mass galaxy will have centrally concentrated stellar and gaseous distributions , and be more susceptible to a burst mode of star formation . |
dwarf galaxies in our local universe are ideal laboratories for studying the interplay between the ism and star formation in low - metallicity environments ( hunter & gallagher 1989 ) .
they are at relatively early epochs of their chemical evolution , possibly resembling distant protogalaxies in their early stages of star formation .
although only a few metal - poor galaxies have been observed in the mir using iso , the characteristics of the mir dust components appear to differ remarkably from those of normal - metallicity starbursts ( madden 2000 ) . whether this is an abundance or composition effect is not yet clear . for this reason
, we are studying their detailed luminosity budget by modeling their spectral energy distributions ( seds ) from optical to millimeter wavelengths , thereby constructing templates to study conditions in primordial galaxies to help to constrain galaxy evolution models .
the galaxy we present here is ngc 1569 .
it is a hi - rich , metal - poor ( @xmath0 ) dwarf irregular galaxy which lies near the galactic plane at a distance of @xmath1 ( israel 1988 ) .
ngc 1569 is presently in the aftermath of a massive burst of star formation ( israel 1988 , israel & de bruyn 1988 , waller 1991 ) and exhibits very compact hii regions and 2 super - star - clusters ( hunter et al 2000 ) .
figure [ fig : obs ] shows some of our broadband images of ngc 1569 with isocam and scuba .
the isocam 7.75 @xmath2 band traces the aromatic bands ( pah ) and the scuba band traces the cold grain continuum . [
cols="^,^ " , ] figure [ fig : cvf ] shows the isocam cvf spectrum for ngc 1569 . compared to a normal disk galaxy , there are very strong ionic lines due to o - b stars .
the ratio @xmath3/[neii ] } \sim 10 $ ] limits the age of the stellar clusters to @xmath4 5 myr .
there are weak aromatic bands , probably due to the destruction of the carriers as a consequence of the hard penetrating radiation field in the low - metallicity environement .
moreover , there s a very steeply rising continuum due to very small grains emission ( see the the dsert et al ( 1990 ) model ) which are stochastically heated .
the dsert et al ( 1990 ) model empirically computes the dust emissivity from near - infrared to submillimetre wavelengths .
it is a coherent interpretation of both the interstellar extinction and the infrared emission .
the emission originates from three different components : the pah ( polycyclic aromatic hydrocarbon ) which are aromatic 2-dimensional molecules producing the fuv non - linear part of the extinction curve , the vsg ( very small grains ) which are 3-dimensional carbon grains apparently responsible for the absorption bump at 217.5 nm , the bg ( big grains ) which are 3-dimensional silicate grains explaining the nir and visible rise of the extinction curve .
each component is described by four parameters : a minimum and maximum grain size ( @xmath5 and @xmath6 ) , a power - law index , @xmath7 , for the size distribution and a scaling factor which is the mass abundance relative to hydrogen ( @xmath8 ) .
the number density of grains of radius between @xmath9 and @xmath10 is @xmath11 .
we model the sed of this galaxy , fitting the dsert et al ( 1990 ) dust model to our observations ( broadbands and spectrum ) .
we make the link between the stellar radiation field and the full dust spectrum by modeling the stellar populations with pegase ( fioc , rocca - volmerange , 1997 ) using uv to nir data from the literature and mir ionic lines as a further constraint .
we then use our modeled stellar radiation field as input to the dust model . to obtain the best @xmath12 fit ,
we vary different parameters of the model : the mass abundances ( @xmath13 ) , the minimum and maximum sizes ( @xmath5 and @xmath6 , except @xmath14 ) , assuming the continuity of the size distribution , and the extinction of the synthesized radiation field .
however , if we try to fit our data with only the three components , there remains a submillimetre excess we ca nt explain .
we decide to add a fourth component corresponding probably to very cold grains ( vcgs ) , modeled by a black body with an emissivity index @xmath15 to reproduce this excess and to obtain a correct fit varying the mass abundance of this component .
finally we get @xmath16 for 15 data points and 8 free parameters .
all the parameters used are summarized in the following table and compared to the values for the milky - way .
the best fit we obtained is shown on figure [ fig : sed ] . & milky & ngc & milky & ngc & milky & ngc & & ngc + & way & 1569 & way & 1569 & way & 1569 & & 1569 + @xmath13 & @xmath17 & @xmath18 & @xmath19 & @xmath20 & @xmath21 & @xmath22 & @xmath13 & @xmath23 + @xmath5 & 4 @xmath24 & 4 @xmath24 & 12 @xmath24 & 37 @xmath24 & 150 @xmath24 & 47 @xmath24 & t & 7 k + @xmath6 & 12 @xmath24 & 37 @xmath24 & 150 @xmath24 & 47 @xmath24 & 1100 @xmath24 & 8500 @xmath24 & @xmath15 & 1.0 + @xmath7 & 3 & 3 & 2.6 & 2.6 & 2.9 & 2.9 & +
the dust masses deduced from our modeling are @xmath25 , @xmath26 , @xmath27 , @xmath28 and the gas - to - dust mass ratio is @xmath29 .
the dust mass deduce from the extinction is @xmath30 which is the same order of magnitude as the mass deduced from the model .
our results seem to show that the dust mass in ngc 1569 is mainly concentrated in cold dust : big grains and very cold grains which both have roughly the same mass abundance .
the pahs as well as the vsgs are very sparse , while the sed is dominated by bigger grains . 99 dsert , f.x . , boulanger f. , puget , j.l . , 1990 , a&a , * 237 * , 215 fioc , m. , rocca - volmerange , b. , 1997 , a&a , * 326 * , 950 - 962 galliano , f. , madden s. , jones , a. , wilson , c. , 2001 , in preparation hunter , d.a . ,
gallagher iii , j.s .
, 1989 , science , * 243 * , 1557 hunter , d.a . ,
gallagher iii , j.s . , rice , w.l . ,
gillett , f.c .
, 1989 , apj , * 336 * , 152 hunter , d.a . ,
oconnel , r.w . ,
gallagher , j.s .
, smecker - hane , t.a .
, 2000 , apj , * 120 * , 2383 - 2401 israel , f.p . , 1988 , * 194 * , 24 - 32 israel , f.p .
, de bruyn , a.g . , 1988 , *
198 * , 109 - 115 madden , s.c . , 2000 , new astronomy reviews , * 44 * , 249 stil , j. , 1999 , phd thesis waller , w.h . , 1991 , *
370 * , 144 - 157 | we are modeling the spectra of dwarf galaxies from infrared to submillimeter wavelengths to understand the nature of the various dust components in low - metallicity environments , which may be comparable to the ism of galaxies in their early evolutionary state .
the overall nature of the dust in these environments appears to differ from those of higher metallicity starbursting systems . here
, we present a study of one of our sample of dwarf galaxies , ngc 1569 , which is a nearby , well - studied starbursting dwarf . using isocam , iras , isophot and scuba data with the dsert et al ( 1990 ) model , we find consistency with little contribution from pahs and very small grains and a relative abundance of bigger colder grains , which dominate the fir and submillimeter wavelengths .
we are compelled to use 4 dust components , adding a very cold dust component , to reproduce the submillimetre excess of our observations . |
this paper introduces a technique which we call _ reverse engineering _ that can be used to construct infinite families of distinct smooth structures on many @xmath0-manifolds . as one example of the utility of this technique we will construct infinitely many distinct smooth structures on @xmath1 .
exotic smooth structures on these manifolds were first constructed in @xcite .
reverse engineering is a three step process for constructing infinite families of distinct smooth structures on a given simply connected @xmath0-manifold .
one starts with a model manifold which has nontrivial seiberg - witten invariant and the same euler number and signature as the simply connected manifold @xmath2 that one is trying to construct , but with @xmath3 .
the second step is to find @xmath4 essential tori that carry generators of @xmath5 and to surger each of these tori in order to kill @xmath5 and , in favorable circumstances , to kill @xmath6 .
the third step is to compute seiberg - witten invariants .
after each of the first @xmath7 surgeries one needs to preserve the fact that the seiberg - witten invariant is nonzero .
the fact that the next to last manifold in the string of surgeries has nontrivial seiberg - witten invariant allows the use of the morgan , mrowka , szab formula @xcite to produce an infinite family as was done in @xcite . in many instances this procedure
can be successfully applied without any computation , or even mention , of seiberg - witten invariants .
if the model manifold for @xmath2 is symplectic and @xmath7 of the tori are lagrangian so that a luttinger surgery will reduce @xmath4 , then there are infinitely many distinct smooth manifolds with the same cohomology ring as @xmath2 . if the resulting manifold is simply connected , then one can often show that there are infinitely many distinct smooth structures on @xmath2 . aside from finding interesting model manifolds
, it seems that the most difficult aspect to the reverse engineering procedure is the computation of fundamental groups .
we will prove the main theorem that shows that this procedure provides infinitely many distinct manifolds in 2 .
we then provide two examples . in
3 we apply the reverse engineering procedure to a model for @xmath1 , the @xmath8-fold symmetric product of a genus 3 surface .
we will identify the lagrangian tori , show that the luttinger surgeries result in a simply connected manifold , and produce infinitely many distinct smooth structures on @xmath1 . in
4 we apply the reverse engineering procedure to the product of two genus @xmath8 surfaces , a model for @xmath9 .
we will identify lagrangian tori that kill @xmath10 and construct infinitely many distinct smooth manifolds with the cohomology ring of @xmath9 .
we have been unsuccessful in showing that these manifolds are simply connected .
one of the key questions in smooth @xmath0-manifold topology is whether a fixed homeomorphism type containing a smooth @xmath0-manifold must actually contain infinitely many diffeomorphism types .
the idea of this section is to state and prove a general theorem pointing in this direction which may be useful to those who are constructing exotic @xmath0-manifolds . to state our theorem
, we need to discuss some notation related to surgery on a torus with trivial normal bundle .
suppose that @xmath11 is such a torus with tubular neighborhood @xmath12 .
let @xmath13 and @xmath14 be generators of @xmath15 and let @xmath16 and @xmath17 be loops in @xmath18 homologous in @xmath12 to @xmath13 and @xmath14 respectively .
let @xmath19 denote a meridional circle to @xmath11 in @xmath2 . by _
@xmath20-surgery _ on @xmath11 with respect to @xmath14 we mean @xmath21 where the gluing map satisfies @xmath22 ) = q [ s^1_{\b } ] + p [ \mu_t]$ ] in @xmath23 .
we denote the ` core torus ' @xmath24 by @xmath25 .
we have framed @xmath26 using @xmath16 and @xmath17 ; so the pushoffs of @xmath13 and @xmath14 in this framing are @xmath16 and @xmath27 . when the curve @xmath27 is nullhomologous in @xmath28 , then @xmath29 .
in addition , when @xmath11 itself is nullhomologous , then @xmath30 .
if @xmath2 is a symplectic manifold and @xmath11 is any lagrangian torus , then there is a canonical framing , called the lagrangian framing , of @xmath26 .
this framing is uniquely determined by the property that pushoffs of @xmath11 in this framing remain lagrangian .
if one performs @xmath31 surgeries with respect to the pushoff in this framing of any curve on @xmath11 , then the result is also a symplectic manifold .
we refer the reader to @xcite for a full discussion of this phenomenon , which is referred to there as _ luttinger surgery_. one must be careful to note that if the pushoff of a curve using the lagrangian framing is not nullhomologous in @xmath28 , then @xmath31 surgery may in fact change @xmath10 .
our theorem is : [ sw ] let @xmath2 be a smooth closed oriented @xmath0-manifold which contains a nullhomologous torus @xmath32 , and let @xmath33 be a simple loop on @xmath32 so that @xmath34 is nullhomologous in @xmath35 . if the seiberg - witten invariant of @xmath36 is nontrivial in the sense that for some basic class @xmath37 , @xmath38)\ne0 $
] , then among the manifolds @xmath39 , infinitely many are pairwise nondiffeomorphic . the meaning of ` @xmath40 ' is explained below .
the following is a very simple but effective corollary to the proof .
[ csw ] suppose that @xmath41 has , up to sign , just one seiberg - witten basic class .
then the manifolds @xmath42 are pairwise nondiffeomorphic . in case @xmath2
is simply connected and the @xmath39 are also simply connected then all of the manifolds @xmath39 are homeomorphic . as outlined in the introduction ,
one very useful application of the theorem is to start with a model manifold with @xmath3 and nontrivial seiberg - witten invariant and with the same euler number and signature as a ( say ) @xmath43 manifold that we are trying to construct .
then , provided that we can find them , we surger essential tori which carry generators of @xmath5 .
if we can do this @xmath44 times , we kill @xmath44 .
we will see that if we can arrange the seiberg - witten invariant to be nonzero after each surgery , then we will satisfy the hypothesis that @xmath36 have nontrivial seiberg - witten invariant .
for example , if we start with a symplectic @xmath0-manifold with @xmath45 and each time perform a luttinger surgery on an embedded lagrangian torus , this will be true .
the fact that the next to last manifold in our string of surgeries has nontrivial seiberg - witten invariant will allow the application of the theorem and/or its corollary .
we will discuss two examples in 3 and 4 .
the proof of theorem [ sw ] involves calculation of seiberg - witten invariants .
we give a short discussion for the purpose of setting notation .
the seiberg - witten invariant of a smooth closed oriented @xmath0-manifold @xmath2 with a homology orientation , _ i.e. _ an orientation of @xmath46 , and with @xmath47 is an integer - valued function @xmath48 which is defined on the set of @xmath49 structures over @xmath2 .
corresponding to each @xmath49 structure @xmath50 over @xmath2 is the bundle of positive spinors @xmath51 over @xmath2 .
set @xmath52 to be the poincar dual of @xmath53 .
each @xmath54 is a characteristic element of @xmath55 ( i.e. its poincar dual @xmath56 reduces to @xmath57 mod 2 ) .
we shall work with the modified seiberg - witten invariant @xmath58 defined by @xmath59 .
this is a diffeomorphism invariant of @xmath2 together with its homology orientation .
( as usual , the choice of homology orientation is suppressed from notation . ) if @xmath60 has no @xmath8-torsion , then @xmath61 . in case @xmath62 ,
the invariant requires the choice of a class @xmath63 with @xmath64 .
we now need to be a bit more explicit .
suppose we have a given orientation of @xmath65 and a given metric for @xmath2 .
the seiberg - witten invariant depends on the metric @xmath66 and a self - dual 2-form as follows .
there is a unique @xmath66-self - dual harmonic 2-form @xmath67 with @xmath68 and corresponding to the positive orientation .
fix a characteristic homology class @xmath69 . given a pair @xmath70 , where @xmath71 is a connection in the complex line bundle whose first chern class is the poincar dual @xmath72 $ ] of @xmath73 and
@xmath74 a section of the bundle of self - dual spinors for the associated @xmath75 structure , the perturbed seiberg - witten equations are : @xmath76 where @xmath77 is the self - dual part of the curvature @xmath78 , @xmath79 is the twisted dirac operator , @xmath80 is a self - dual 2-form on @xmath2 , and @xmath81 is a quadratic function .
write @xmath82 for the corresponding invariant .
as the pair @xmath83 varies , @xmath82 can change only at those pairs @xmath83 for which there are solutions with @xmath84 .
these solutions occur for pairs @xmath83 satisfying @xmath85 .
this last equation defines a wall in @xmath86 .
the point @xmath87 determines a component of the double cone consisting of elements of @xmath86 of positive square .
we prefer to work with @xmath88 .
the dual component is determined by the poincar dual @xmath89 of @xmath90 .
an element @xmath91 of positive square lies in the same component as @xmath89 if @xmath92 .
if @xmath93 for a generic @xmath80 , @xmath94 is well - defined , and its value depends only on the sign of @xmath95 .
write @xmath96 for @xmath82 if @xmath97 and @xmath98 in the other case . the invariant @xmath99 is defined by @xmath100 if @xmath101 , or dually , if @xmath102 , and @xmath103 if @xmath104 . as in the case
above , we work with the modified invariant @xmath105
. we now proceed to the proof of the theorem .
set @xmath41 .
recall that @xmath106 is the torus in @xmath107 which is the core torus of the surgery .
there is a surface in @xmath107 which intersects @xmath106 once , and it follows that @xmath106 is essential ( in fact primitive in @xmath108 ) .
the surface in question is the union of a normal disk to @xmath106 and the surface in @xmath109 bounded by @xmath34 .
let @xmath110 .
then @xmath111 is the core torus of the surgery in @xmath112 .
its meridian @xmath113 represents @xmath114+[\mu_{\l}]$ ] , which in @xmath115 is homologous to @xmath116 $ ] , a nontorsion class .
this means that @xmath111 is nullhomologous in @xmath112 .
let @xmath117 be characteristic .
the classes in @xmath118 which are characteristic and which agree with the restriction of @xmath119 in @xmath120 in the diagram : @xmath121 differ by an even multiple of @xmath122 $ ] .
the adjunction inequality implies that each such class with a nontrivial seiberg - witten invariant is orthogonal to @xmath122 $ ] .
there is just one class in @xmath112 which restricts to a fixed class in @xmath123 because @xmath111 is nullhomologous , and the same is true for @xmath2 ( and @xmath32 ) .
it follows from @xcite that @xmath124)\ ] ] where @xmath73 and @xmath37 are classes which restrict as in the paragraph above . recalling the hypothesis that there is a @xmath125 , satisfying @xmath126)\ne0 $ ]
, it follows that the integer invariants @xmath127 will distinguish an infinite family of pairwise nondiffeomorphic manifolds among the @xmath112 . in case @xmath62
, we need to check issues with chambers .
the inclusions of @xmath128 in @xmath2 and @xmath112 induce isomorphisms on @xmath108 and thus an isomorphism of @xmath129 with @xmath55 .
the gluing formula of @xcite relates chambers using this isomorphism .
so , for example , if @xmath130 this means that @xmath102 .
the isomorphism above gives a @xmath131 and an @xmath132 ( and @xmath133 ) , and it also gives @xmath134 ; so @xmath135 .
thus the gluing formula applies to the invariant @xmath136 .
this works for any choice of period point @xmath89 .
hence the argument in the @xmath45 case applies directly to @xmath137 as well .
corollary [ csw ] follows similarly . when @xmath107 has just one basic class up to sign , it is straightforward to see that if @xmath138 the collection of values of the seiberg - witten invariant @xmath139 is different from @xmath140 .
[ swcor ] let @xmath2 be a smooth closed oriented @xmath0-manifold which contains a nullhomologous torus @xmath32 and let @xmath33 be a simple loop on @xmath32 so that @xmath34 is nullhomologous in @xmath35 .
suppose also that there is a square @xmath141 torus @xmath142 that satisfies @xmath143 .
if @xmath107 has a basic class , _ i.e. _ a class @xmath37 with @xmath144 , then among the manifolds @xmath39 , infinitely many are pairwise nondiffeomorphic .
if @xmath142 is a torus of square @xmath141 that satisfies @xmath143 , then the adjunction inequality implies that in each collection @xmath145 , there is at most one basic class .
thus the hypothesis of theorem [ sw ] is satisfied .
we now provide two examples to illustrate the reverse engineering procedure .
the 2-fold symmetric product @xmath146 is the quotient of @xmath147 by the action of the involution @xmath148 given by @xmath149 .
let @xmath150 , @xmath151 denote standard generators for @xmath152 .
it follows from @xcite that the natural singular khler form @xmath153 on @xmath154 derived from the khler curve @xmath155 admits a cohomologous smoothing to a khler form which equals @xmath153 away from a chosen neighborhood of the diagonal .
( we thank paul kirk for pointing out the necessity of this reference . )
we obtain a basis for @xmath156 as follows .
the tori @xmath157 , @xmath158 , @xmath159 , and @xmath160 , @xmath161 , in @xmath147 descend to twelve tori of self - intersection @xmath141 in @xmath154 , and we also denote these by @xmath157 , @xmath158 , @xmath159 , and @xmath162 .
the three tori @xmath163 in @xmath147 descend to tori @xmath164 of square @xmath165 , and together with the image of @xmath166 , a genus 3 surface which represents a homology class @xmath167 with self - intersection @xmath168 , we get a basis for @xmath169 . the euler number @xmath170 and its signature @xmath171 , in agreement with the characteristic numbers for @xmath1 . to establish some notation , consider figure 1 . for example , we see loops @xmath172 , @xmath173 , and @xmath174
. we also have based loops ( with basepoint @xmath175 , the vertex ) which we shall denote by @xmath176 , @xmath177 , @xmath178 , etc .
the based loop @xmath179 , for example , is the one which starts at the lower left vertex @xmath175 , proceeds backwards along @xmath180 to the initial point @xmath181 of @xmath182 , then traverses @xmath182 until it gets to its endpoint @xmath183 , and then heads vertically downward back to @xmath175 .
the based loop @xmath184 starts at the vertex at the initial point of @xmath180 , travels along @xmath180 to @xmath185 , traverses @xmath186 , then at @xmath187 it heads upward back to @xmath175 .
the based loop @xmath176 is equal to @xmath172 . using @xmath188 as basepoint , the abelian group @xmath189 is generated by the @xmath190 and @xmath191 .
we will perform six surgeries on disjoint lagrangian tori to kill these generators .
we need some notation to describe our surgeries .
first notice that the lagrangian tori @xmath192 , @xmath193 , @xmath194 , @xmath195 , @xmath196 , and @xmath197 in @xmath147 are disjoint from the diagonal , and so they descend to lagrangian tori in @xmath146 .
( we still denote these tori in @xmath154 by @xmath192 , etc . )
they are mutually disjoint .
if , for example , we consider the torus @xmath192 and do @xmath198-framed surgery in @xmath154 along the loop @xmath199 with respect to the lagrangian framing , we denote this as the surgery @xmath200 .
we now perform surgeries along disjoint lagrangian tori @xmath201 denote these six lagrangian tori by @xmath202 , @xmath203 .
we consider tubular neighborhoods @xmath204 with their lagrangian framings .
for example , if @xmath205 then @xmath206 can be visualized in figure 1 as the product of the two annuli given by the trapezoids between @xmath173 and @xmath207 , @xmath208 . then we get lagrangian pushoffs @xmath209 and @xmath210 on @xmath211 .
let @xmath212 be the boundary circle to the normal disk passing through the intersection point @xmath213 of @xmath214 and @xmath215 .
in like manner , for each @xmath73 we have a ` distinguished triple of loops ' @xmath216 which pass through @xmath217 on @xmath218 . for each @xmath202
there is a path , @xmath219 , from the basepoint @xmath188 of @xmath154 to @xmath217 which except for its endpoint lies in @xmath220 .
if we base the distingushed triple of loops for @xmath202 at @xmath175 using @xmath219 and @xmath221 , then the corresponding based loops are respectively homotopic in @xmath222 to @xmath223\ } , \quad \ { \b_1\a_1\b_1^{-1 } , \b_2 ; [ \b_1,\a_2^{-1}]\},\\ & & \ { \a_1,\a_3 ; [ \b_1^{-1},\b_3^{-1}]\ } , \quad \{\b_1 , \b_3\a_3\b_3^{-1 } ; [ \a_1^{-1},\b_3]\},\\ & & \{\a_2 , \a_3 ; [ \b_2^{-1},\b_3^{-1}]\},\quad \{\b_2\a_2\b_2^{-1 } , \b_3 ; [ \b_2,\a_3^{-1}]\}.\end{aligned}\ ] ] ( cf .
@xcite ) we give the proof for the first two tori @xmath224 and @xmath225 .
the other cases are similar .
first consider @xmath226 .
let @xmath227 be the path from the basepoint @xmath175 to @xmath228 traveling backwards along @xmath44 as discussed above , and let @xmath229 be a similar path starting at @xmath175 and ending at @xmath230 , traveling backwards along @xmath180 .
set @xmath231 .
the trapezoid in figure 1 between @xmath232 and @xmath233 gives rise to a homotopy between the based loop @xmath234 and @xmath235 .
for each @xmath236 $ ] , let @xmath237 and @xmath238 .
thought of as a homotopy , the trapezoid between @xmath232 and @xmath233 is composed of parallel closed paths @xmath239 , where @xmath240 , @xmath241 , and the initial point of @xmath239 is @xmath242 .
this defines the based homotopy of @xmath234 to @xmath243 whose path at level @xmath244 is the image in @xmath154 of the product of paths @xmath245 it is easily seen that the trace of this homotopy is disjoint from all the tori @xmath202 .
similarly , the trapezoid between @xmath246 and @xmath247 gives the homotopy between the based loop @xmath248 and @xmath249 .
the meridian @xmath212 lies in the orthogonal torus @xmath250 which is lagrangian away from the basepoint @xmath188 of @xmath154 .
this torus also contains the base path @xmath251 .
since @xmath251 runs backwards along both @xmath44 and @xmath180 to @xmath252 , it follows that when @xmath212 is based using @xmath251 , it is given by the commutator of the appropriately oriented @xmath253-generators of @xmath250 ; _ viz . _
@xmath254 $ ] .
( there is a choice of orientation for this meridian .
the opposite choice would cause us to change the signs of our surgeries . for definiteness , we choose the orientation implied by the statement of the lemma . )
next consider @xmath255 .
its tubular neighborhood @xmath256 with lagrangian framing is given by the product of the two annuli described by trapezoids in figure 1 between @xmath257 and @xmath258 and between @xmath259 and @xmath260 .
( the curve @xmath260 is not shown in figure 1 , but it is analogous to @xmath232 . ) let @xmath261 be the path traveling positively along @xmath44 from @xmath262 to @xmath263 , and let @xmath264 be the path traveling backwards along @xmath247 from @xmath175 to @xmath265 . set @xmath266 .
then , analogously to the argument above , the trapezoid between @xmath258 and the nearby parallel dashed line describes a homotopy from @xmath267 to the based loop given by this dashed line which is @xmath268 .
the homotopy from @xmath269 to @xmath270 is given by the trapezoid between @xmath260 and @xmath180 .
to see the meridian @xmath271 we use the dual torus @xmath272 which contains the path @xmath273 which runs positively along @xmath44 and negatively along @xmath247 , and the same reasoning as above shows that when @xmath271 is based using @xmath273 , it is given by @xmath274 $ ] . denote the manifold obtained from these six surgeries by @xmath2 .
the result of each surgery is to reduce the first betti number by one , reduce the second betti number by two , and introduce a relation in @xmath253 .
for example , because of lemma 1 , the surgery @xmath275 introduces the relation @xmath276 $ ] , and the surgery @xmath277 introduces the relation @xmath278 $ ] .
note that @xmath279 , and since the surgeries change neither the euler number nor signature , @xmath280 , and @xmath2 is a rational homology @xmath1 .
the following relations hold in @xmath281 : @xmath282,\ \
\b_2= [ \b_1,\a_2^{-1}],\ \ \a_1= [ \b_1^{-1},\b_3^{-1}],\\ \b_1= [ \a_1^{-1},\b_3],\ \ \a_3= [ \b_2^{-1},\b_3^{-1}],\ \ \b_3=[\b_2,\a_3^{-1}],\\ [ \a_1,\b_1]=1 , \ \ [ \a_1,\a_2]=1,\ \ [ \a_1,\b_2]=1,\ \ [ \a_1,\a_3]=1,\ \ [ \b_1,\a_3]=1,\\ [ \a_2,\b_2]=1,\ \ [ \a_2,\a_3]=1,\ \ [ \a_2,\b_3]=1,\ \ [ \a_3,\b_3]=1.\end{gathered}\ ] ] thus we have @xmath278=[[\a_1^{-1},\b_3],\a_2^{-1}]=1 $ ] , using the commutativity relations @xmath283=1 $ ] and @xmath284=1 $ ] .
now it follows from the other relations that @xmath285 . since the surgeries that we perform on the lagrangian tori all have surgery coefficient @xmath286 with respect to the lagrangian framing , the resultant manifolds all have induced symplectic structures .
one simple way to see that @xmath2 is not diffeomorphic to @xmath1 is to use the fact from @xcite that @xmath1 has a unique symplectic form up to diffeomorphism and symplectic deformation .
this means that for any symplectic form on @xmath1 , the canonical class must pair negatively with the symplectic form . on @xmath146 , which is a surface of general type , the canonical class pairs positively with the symplectic form , and since we have constructed @xmath2 by surgeries on lagrangian tori of @xmath154 , the same is still true in @xmath2 .
( the point here is that if @xmath287 is the result of a luttinger surgery on @xmath154 , then the complements of tubular neighborhoods of the respective lagrangian tori in each can be identified , and the restrictions of the symplectic forms can as well .
the canonical classes are supported in the complements of these tori and agree over the complements of the tubular neighborhoods .
it follows that as elements of @xmath288 the poincar duals satisfy @xmath289 $ ] .
so @xmath290 since @xmath291 is lagrangian . )
hence @xmath2 can not be diffeomorphic to @xmath1 .
the symplectic manifold @xmath2 is irreducible and homeomorphic but not diffeomorphic to @xmath1 .
the irreducibility of @xmath2 follows from @xcite once we show that @xmath2 is minimal .
this follows from the seiberg - witten calculations below .
it is interesting to ask whether @xmath2 is actually diffeomorphic to the symplectic manifolds constructed in @xcite . in order to produce an infinite family of exotic @xmath1 s ,
let @xmath292 denote the result of the first five luttinger surgeries on @xmath154 .
thus @xmath293 and @xmath294 .
we construct @xmath2 by performing a surgery @xmath295 in @xmath292 . in @xmath2 ,
the surgery gives us a nullhomologous torus @xmath32 , the core " of the surgery .
there is a loop @xmath33 on @xmath32 so that surgery on @xmath296 gives @xmath292 back . the framing for this surgery must be the nullhomologous framing .
we apply theorem [ sw ] to @xmath297 . in fact
, corollary [ csw ] will tell us that the manifolds @xmath298 are pairwise nondiffeomorphic once we see that @xmath292 has exactly two basic classes .
( the manifold denoted @xmath107 in corollary [ csw ] is our manifold @xmath292 . ) note that @xmath299 is the result of performing the surgery @xmath300 in @xmath292 .
( these are _ not _ luttinger surgeries . )
[ inf ] the manifolds @xmath1 , @xmath2 , and @xmath112 , @xmath301 , are pairwise homeomorphic and ( except for @xmath1 ) are minimal , but no two are diffeomorphic .
the homeomorphism statement will follow once we see that each @xmath112 is simply connected . a presentation for @xmath302
is obtained from the one above for @xmath303 by replacing the relation @xmath304 $ ] by @xmath304^{-(n+1)}$ ] , and @xmath305 follows as above .
next , we need to show that the manifold @xmath292 has just two basic classes , @xmath306 its canonical class , and then call on corollary [ csw ] . since @xmath154 is a surface of general type , its only basic classes are @xmath306 its canonical class , i.e. @xmath307 , where @xmath167 and @xmath164 are the classes described above . according to @xcite , each time we do a surgery ,
the seiberg - witten invariant of the result is calculated in terms of the seiberg - witten invariants of the original manifold and those of the result of the surgery that kills the curve on the torus .
for example , if @xmath308 is the result of the surgery @xmath275 on @xmath154 , then let @xmath309 be the result of the surgery that kills @xmath182 directly ( @xmath141-surgery ) . in @xmath309 , the surface @xmath310 , which represents @xmath167 , has its genus reduced by one .
applying the adjunction inequality to this situation , we see that any basic class of @xmath309 has the form @xmath311 .
since the square of a basic class must be @xmath312 , in fact none of these classes can be basic ; so the seiberg - witten invariant of @xmath309 vanishes .
the result of this argument is that the manifold @xmath308 also has just two basic classes , @xmath306 its canonical class .
the very same argument works for each surgery and finally shows that @xmath292 has just two basic classes .
thus @xmath2 and @xmath112 have just two basic classes , @xmath313 , and the difference is a class of square @xmath314 . if one of these manifolds failed to be minimal , it would have to have a pair of basic classes , @xmath315 , whose difference has square @xmath316 .
thus @xmath2 and @xmath112 are minimal . in order to obtain infinitely many smooth structures
, we did not need to perform this last step which shows that @xmath2 and @xmath112 have just two basic classes .
we did this to explicitly show that all the @xmath112 are distinct .
the hypothesis of corollary [ swcor ] is satisfied because each of the lagrangian tori on which surgery is performed has a dual torus of square @xmath141 .
each @xmath317 contains disjoint embeddings of a minimal genus @xmath318 surface representing @xmath167 and the three tori @xmath319 , @xmath320 , @xmath321 with self - intersection @xmath165 .
more interestingly , each @xmath317 contains a sphere of self - intersection @xmath322 representing @xmath323 -[t_{2 } ] -[t_{3}]$ ] that is the image in @xmath324 of a pushoff of the diagonal in @xmath147 .
these surfaces can be useful for other constructions .
the symmetric product @xmath325 of a genus @xmath326 surface with itself has @xmath327 and that @xmath328 , @xmath329 , and @xmath330 .
thus @xmath331 is a model for @xmath332 .
a straightforward generalization of the above application of reverse engineering provides infinitely many distinct smooth structures on these manifolds , one of which is symplectic .
we now give an example to point out that the computation of fundamental groups in the reverse engineering procedure can be difficult .
let @xmath333 , the product of two genus @xmath8 surfaces , and denote the standard generators of @xmath253 by @xmath334 and @xmath335 for @xmath208 .
( since the calculations in this section parallel those of 3 , for the sake of simplicity we no longer denote the distinction between based and unbased loops . )
so @xmath336 has these eight generators with relations @xmath337[a_2,b_2]=1 $ ] , @xmath338[c_2,d_2]=1 $ ] and all @xmath172 and @xmath339 commute with all @xmath340 and @xmath341 .
the area forms on the two copies of @xmath342 induce a symplectic form on the product @xmath154 , all tori of the form @xmath343 , @xmath344 , @xmath345 , and @xmath346 are lagrangian , and the lagrangian framing is the obvious one coming from the product structure . the euler number @xmath347 and its signature @xmath348 , in agreement with the characteristic numbers for @xmath9 .
let @xmath351 be the complement of the above eight lagrangian tori in @xmath352 .
inside @xmath351 , there are basepaths from the basepoint of @xmath154 to the basepoint of the boundaries of the tubular neighborhoods of these tori , so that if we base the distinguished triple of loops using these basepaths then the corresponding based loops are homotopic in @xmath351 to @xmath353\ } , \quad \ { b_1 , d_1c_1d_1^{-1 } ; [ a_1^{-1},d_1]\},\\ & & \ { a_2 , c_2 ; [ b_2^{-1},d_2^{-1}]\ } , \quad \ { b_2 , d_2c_2d_2^{-1};[a_2^{-1},d_2]\ } , \\ & & \ { a_2 , c_1 ; [ b_2^{-1},d_1^{-1}]\},\quad \ { b_2a_2b_2^{-1 } , d_1 ; [ b_2,c_1^{-1}]\},\\ & & \ { a_1 , c_2 ; [ b_1^{-1},d_2^{-1}]\},\quad \ { b_1a_1b_1^{-1 } , d_2 ; [ b_1,c_2^{-1}]\}.\end{aligned}\ ] ] the following relations hold in @xmath281 : @xmath354=a_1,\ \ [ a_1^{-1},d_1]=b_1,\ \ [ b_2^{-1},d_2^{-1}]=a_2,\ \ [ a_2^{-1},d_2]=b_2,\\ [ d_1^{-1},b_2^{-1}]=c_1,\ \ [ c_1^{-1},b_2]=d_1,\ \ [ d^{-1}_2,b^{-1}_1]=c_2,\ \ [ c_2^{-1},b_1]=d_2,\\ [ a_1,c_1]=1 , \ \ [ a_1,c_2]=1,\ \ [ a_1,d_2]=1,\ \ [ b_1,c_1]=1,\\ [ a_2,c_1]=1 , \ \ [ a_2,c_2]=1,\ \ [ a_2,d_1]=1,\ \ [ b_2,c_2]=1,\\ [ a_1,b_1][a_2,b_2]=1,\ \ [ c_1,d_1][c_2,d_2]=1.\end{gathered}\ ] ] from these relations it is clear that @xmath355 , and since the surgeries change neither the euler number nor signature , @xmath356 . in fact , the only homology classes that survive are those represented by @xmath357 and @xmath358 ; so @xmath2 is a homology @xmath9 .
we have been unable to determine if the perfect group @xmath281 is trivial or not . also , there are other surgeries that can be performed , and also other sets of eight lagrangian tori that can be surgered , to obtain many presentations of perfect groups that we have not succeeded in showing are trivial
. we can produce an infinite family of distinct homology @xmath9 s , exactly as in 3 .
the presentation for @xmath253 is exactly the one given above for @xmath281 except that the relation @xmath359=d_2 $ ] is replaced by @xmath359^{n+1}=d_2 $ ] compounding the difficulty of determining whether the group is trivial . | we introduce a general procedure called ` reverse engineering ' that can be used to construct infinite families of smooth @xmath0-manifolds in a given homeomorphism type . as one of the applications of this technique , we produce an infinite family of pairwise nondiffeomorphic @xmath0-manifolds homeomorphic to @xmath1 .
.5 cm |
* the inequality . * the quantum mechanical system considered here is a particle present in one of two distinguishable paths , denoted respectively as @xmath3 and @xmath4 .
additionally , the particle is equipped with a @xmath5-dimensional internal degree of freedom spin which we will treat as a subsystem @xmath6 .
the complete state of the particle prepared in the interferometer is characterized by a certain density operator @xmath7 . both paths
are equiprobable , i.e. @xmath8 . inside the interferometer ,
the particle interacts with an environment @xmath9 producing a joint state @xmath10 .
the interaction is constrained in that it can not transfer the particle between the paths . in this
setting , which - way information contained in the environment corresponds to the ability to discriminate between two normalized density matrices @xmath11 , @xmath12 . in the general case of mixed states
, distinguishability is given explicitly by the expression@xcite @xmath13 where @xmath14 denotes the trace norm .
let us stress here that this quantity is not conditioned upon any measurement performed eventually on subsystems @xmath15 .
our aim is to infer an upper bound on @xmath0 from the properties of the particle , i.e. subsystems @xmath15 , after the interaction with the environment .
the state of the particle at this stage can be written as a result of an action of a certain quantum channel@xcite @xmath16 on the input state @xmath7 , with @xmath17 .
because the particle is not transferred between interferometer arms , for any spin operator @xmath18 we have @xmath19 our central result is that @xmath0 satisfies the inequality : @xmath20 where @xmath21 , named generalized visibility , reads : @xmath22 here @xmath23 is a normalized maximally entangled state of two replicas of the spin subsystem , @xmath24 is the identity channel , @xmath25 is the identity operator on the spin subsystem , and @xmath26 are normalized spin states for the particle in either arm of the interferometer before the interaction with the environment .
the complete proof of the inequality is presented in methods .
* examples . * the inequality ( [ eq : dandv ] ) resembles that for a spinless particle given in eq .
( [ eq : spinless ] ) , but the concept of generalized visibility covers a much wider range of physical scenarios . in order to illustrate its breadth ,
let us consider here three cases .
first , when no interaction with the environment occurs , @xmath27 . in this case
@xmath28 , i.e. no matter what internal states for individual paths are initially chosen , the environment does not gain any which - way information .
second , let us take @xmath29 , where @xmath30 is a fixed unit - trace hermitian operator .
physically , this occurs when the internal degree of freedom is entirely transferred to the environment and replaced with a state @xmath30 .
then @xmath31 which is the standard expression for quantum fidelity@xcite between density operators @xmath32 and @xmath33 . consequently , the only way to ensure that no which - way information leaks out is to associate both the interferometer paths with identical internal states .
finally , suppose that @xmath34 where @xmath35 denotes transposition .
this channel gives @xmath36 .
the resulting expression has non - intuitive properties : for a particle prepared in any pure state @xmath37 the generalized visibility equals @xmath38 , which is strictly less than one for a non - trivial spin subsystem when @xmath39 .
generalized visibility reaches one only when the two paths are associated with maximally mixed spin states , @xmath40 , which follows immediately from the inequality between arithmetic and quadratic means .
this indicates that which - way information deposited in the environment may be erased by introducing classical noise in the input state of the spin subsystem . to illuminate this point ,
let us take a single photon traveling through a mach - zehnder interferometer and consider its polarization , spanned by two orthogonal states @xmath41 and @xmath42 , as the two - dimensional internal degree of freedom .
let the interaction of the photon with the environment be given by a unitary map @xmath43 \label{eq : channelexplicit}\end{gathered}\ ] ] where @xmath44 , @xmath45 , are four orthogonal states of the environment and @xmath46 , @xmath47 are the probability amplitudes in the rectilinear basis .
it is easy to verify that this map indeed induces @xmath48 , while the polarization state in an individual arm is scrambled to a completely mixed state , @xmath49 . if we prepare @xmath50 , the two states of the environment correlated with the paths : @xmath51 are partly distinguishable .
similarly , for the vertical input polarization , @xmath52 , the environment states are @xmath53 however , if the input polarization is prepared in a completely mixed state @xmath54 , one sees immediately that both the paths are associated with the same mixed environment state @xmath55 and which - way information is not available anymore .
the inequality ( [ eq : dandv ] ) demonstrates that this effect is an intrinsic feature of the channel under consideration rather than its specific realization given in eq .
( [ eq : channelexplicit ] ) . * estimating generalized visibility . * generalized visibility @xmath21 is given by a rather intricate expression involving the spin preparation and the effective quantum channel experienced by the particle . in principle
, full information about the channel can be obtained from quantum process tomography,@xcite but this approach may be resource consuming , especially for a high dimension of the internal subsystem @xmath6
. we will now give a recipe to construct estimates for @xmath21 from direct visibility measurements for specific spin preparations and selections of individual spin components before interfering the particle paths .
consider the following procedure .
the particle is prepared in one of states @xmath56 where the normalized kets @xmath57 and @xmath58 describe the spin state in the upper and lower interferometer path .
the index @xmath59 labels different preparations . after the interaction with the environment ,
the spin subsystem is filtered individually in each interferometer arm to extract probability amplitudes corresponding to normalized spin states @xmath60 and @xmath61 .
these components are subsequently made indistinguishable by a suitable spin transformation and interfered on a balanced beam splitter with a relative phase shift @xmath62 .
the probabilities of detecting the particle at the two output ports @xmath63 of the beam splitter read @xmath64.\ ] ] here @xmath65\ ] ] is the overall probability of detecting the particle in the filtered components .
the modulation depth of interference fringes is characterized by the fractional visibility , given explicitly by @xmath66 for the quantities introduced here , we always have @xmath67 .
full - depth modulation of fringes is observed for the equality sign .
let us now take any set of complex coefficients @xmath68 such that a linear combination of rank - one operators @xmath69 can be written as @xmath70 for a certain operator @xmath71 acting on the duplicated spin subsystem that satisfies @xmath72 .
we show in methods that under these assumptions the generalized visibility is bounded by a linear combination of fractional visibilities with the same coeffcients @xmath68 : @xmath73 combining this result with eq .
( [ eq : dandv ] ) yields a family of bounds on distinguishability in the form @xmath74 in the upper arm and @xmath75 in the lower arm . after interaction with the environment described by the channel @xmath16 the phase between interferometer arms
is modulated by @xmath62 .
next , the spin states are filtered in the orthonormal bases @xmath76 for the upper arm and @xmath77 for the lower arm , and individual components are interfered pairwise to determine fractional visibilities @xmath78.,scaledwidth=48.0% ] special cases of eq .
( [ eq : dlesum ] ) are intuitive .
consider a scenario shown in fig .
[ fig : singlepreparation ] .
we use only a single preparation @xmath79 ( henceforth in this example we drop the superfluous label @xmath59 ) , while the sets of filter states in the upper @xmath80 and lower @xmath81 arm form two orthonormal bases for the spin subsystem with @xmath82 indexing the basis elements . in eq .
( [ eq : sumklalphas ] ) we can take arbitrary phase factors as coefficients @xmath83 , with the operator @xmath84 satisfying the condition @xmath72 . according to eq .
( [ eq : genvisbound ] ) we have @xmath85 , which maximized over phases @xmath86 estimates the generalized visibility by a sum of absolute values of fractional visibilities : @xmath87 if for each filter we measure the maximum possible fractional visibility @xmath88 , the right hand side of ( [ eq : visorthogonal ] ) becomes one , because completeness of the bases @xmath80 and @xmath81 implies that @xmath89 .
in other words , when the interaction with the environment introduces different phase shifts between distinguishable components of the spin subsystem but maintains overall coherence between the paths , we can observe full visibility by suitably sorting spin components at the output and interfering them pairwise .
this suffices to verify that no which - way information is deposited in the environment .
however , as noted earlier the interplay between the path and the spin subsystems can exhibit more intricate behaviour .
we illustrate this point with a single photon experiment . * experiment . * in order to expose more complex aspects of the trade - off between which - way information and interference visibility for a particle with an internal structure , we investigated experimentally interference of a single photon in a noisy mach - zehnder interferometer shown in fig .
[ fig : setup ] , using photon polarization as the internal subsystem @xmath6 .
single photons with @xmath90 nm central wavelength were generated by type - ii spontaneous parametric down - conversion in a @xmath91 mm long ppktp crystal pumped with 14 mw of @xmath92 nm wavelength light from continuous wave diode laser , and heralded by detection of orthogonally polarized conjugate photons from the same pairs .
the photons , after sending through a @xmath93 nm bandwidth interference filter and transmitting via a single mode fiber , were split between two paths using a calcite displacer .
equal splitting was ensured by using a fiber polarization controller .
removable half - wave plates @xmath94 and @xmath95 were used to prepare any combination of horizontal @xmath96 and vertical @xmath97 polarization states for the two paths .
* a ) * and @xmath95 .
noise @xmath16 is introduced using four sets of orientations of hwps @xmath98 and @xmath99 and quarter - wave plates @xmath100 and @xmath101 mounted on computer - controlled motorized rotation stages .
polarization components of the output state are filtered using a hwp @xmath102 and a second calcite crystal .
interference is realized in the common - path configuration using hwp @xmath103 oriented at @xmath104 degrees with respect to the setup plane and a polarizing beam splitter @xmath105 with output ports monitored by single photon detectors @xmath106 and @xmath107 .
two other detectors @xmath108 and @xmath109 are used to determine the overall count rate .
the phase shift @xmath62 is introduced by rotating the second calcite crystal about an axis perpendicular to the setup plane using a closed loop piezo actuator . *
b * , orientations of principal axes of wave plates @xmath110 that realize path - dependent unitaries @xmath111.,scaledwidth=48.0% ] * b ) * [ cols="^,>,>,>,>",options="header " , ] because individual paths inside the interferometer are completely depolarized by the interaction with the environment , the theoretical probability of filtering out the photon for any selection of the preparation @xmath59 and the filter @xmath82 equals @xmath112 .
theoretical predictions for fractional visibilities are collected in tab .
[ tab : fracvis](a ) .
it is seen that only few combinations of preparations and filters are expected to produce interference fringes .
let us select for a moment a single preparation @xmath59 .
after filtering a pair of polarization components one could use the remaining components to measure also the second fractional visibility for complementary polarizations .
for example , @xmath113 is complementary to @xmath114 , and @xmath115 is complementary to @xmath116 .
this is an example of a scenario covered by the inequality ( [ eq : visorthogonal ] ) .
although the photon is equally likely to choose either of the two complementary filters , interference fringes are visible only for one of them and the right hand side of ( [ eq : visorthogonal ] ) is at most @xmath117 .
suppose now that before interaction with the environment the photon polarization is averaged to the completely mixed state and @xmath118 , for example by switching randomly between four preparations . to estimate which - way information available in this case ,
let us take on the left hand side of eq .
( [ eq : sumklalphas ] ) four non - zero coefficients : @xmath119 where @xmath120 are arbitrary phases .
the corresponding operator @xmath121 is unitary for any choice of phases @xmath120 .
therefore eq . ( [ eq : genvisbound ] ) provides a bound on generalized visibility in the form @xmath122 .
maximization over phases yields @xmath123 this bound on generalized visibility can be applied directly to experimental data in order to estimate available which - way information for mixed polarization preparation .
( @xmath124 , black ) and @xmath107 ( @xmath125 , red ) conditioned upon detection of conjugate heralding photons from the down - conversion source as a function of the phase shift in the interferometer .
panels are labelled as @xmath126 , where @xmath59 specifies polarization preparation and @xmath82 defines polarizations filtered from the two paths .
solid lines are sinusoidal fits used to determine fractional visibilities .
experimental uncertainties of count rates are smaller than the size of the symbols.,width=332 ] in order to measure fractional visibilities , the phase in the interferometer was adjusted by rotating one of the calcite crystals about an axis perpendicular to the plane of the setup using a closed loop piezo - electric actuator .
for each selected phase value we averaged over four unitaries simulating the interaction with the environment .
experimentally measured interference fringes for combinations of preparations @xmath59 and filters @xmath82 entering the inequality ( [ eq : swapestimate ] ) are shown in fig .
[ fig : fringes ] along with results obtained for complementary filters .
probabilities @xmath127 and fractional visibilities @xmath128 determined from these measurements are collected in tab .
[ tab : fracvis](b ) . inserting these data in eq .
( [ eq : swapestimate ] ) gives @xmath129 .
this translates into an upper bound on distinguishability in the form @xmath130 according to eq .
( [ eq : dlesum ] ) .
remarkably , this stringent bound is determined from just few measurements of fractional visibilities rather than complete quantum process tomography .
in contrast , consider an individual pure preparation and complementary projective filters , corresponding to a pair of graphs shown in a single row of fig .
[ fig : fringes ] . in this case ,
experimental results yield at most @xmath131 using eq .
( [ eq : visorthogonal ] ) , which implies @xmath132 .
this separation between estimated on distinguishability for mixed and pure polarization preparations strikingly demonstrates the convoluted effects of internal degrees of freedom in a which - way experiment .
the scenario considered in this work assumed a clear distinction between the environment and the spin that are both external to the path subsystem .
tha amount of available which - way information was determined solely from the quantum state of the environment after the interaction with the particle in the interferometer and no classical information about either spin preparation or measurement results could be used for that purpose .
this makes our approach distinct from previous studies of the quantum erasure phenomenon,@xcite where postselection carried out on the environment can restore conditional interference fringes . here
the environment was treated as an adversary whose information about the path taken by the particle we try to control and estimate by manipulating the spin subsystem at the preparation and the detection stages . in this context , the analysis of a which - way experiment with an internal degree of freedom points at non - trivial issues in prepare - and - measure protocols for quantum key distribution over complex noisy channels .
suppose that a cryptographic key is to be established by sending a particle along a randomly selected one of two paths , while the key security is to be verified by preparing occasionally the particle in a superposition of the two paths and measuring its coherence at the output by sampling interference fringes .
if the particle is equipped with spin , the trade - off between distinguishability and visibility derived in this work implies that certain eavesdropping strategies require the sender to use randomness in spin preparation in order to ensure security .
further , to verify the security the sender and the receiver may need to combine data collected for an array of spin preparations and filterings .
an entanglement - based analog of the scenario analyzed above would be to prepare jointly two particles in a maximally entangled path - spin state and to subject one of them to interaction with the environment modelling an eavesdropping attempt . in this case , spin can play the role of a `` shield '' subsystem fully protecting the privacy of a cryptographic key generated by detecting particles in individual paths , even though the noise present in the bipartite state may prevent entanglement distillation at the same rate as key generation.@xcite to complete the parallel , estimates for generalized visibility determined from fractional visibility measurements can be viewed as a dynamical analog of recently introduced privacy witnesses.@xcite
our basic tool will be the choi - jamiokowski isomorphism@xcite between linear maps and operators acting on the duplicated hilbert space .
let us define a maximally entangled state of two replicas of the system @xmath15 as @xmath133 , where @xmath134 and @xmath135 .
the choi - jamiokowski state corresponding to a channel @xmath136 is given by @xmath137 , where the identity map @xmath24 acts on the subsystem @xmath15 , while @xmath136 acts on @xmath138 .
the transformation of a state @xmath139 can be written as @xmath140\ ] ] where the transposition @xmath35 is performed in the same basis in which the state @xmath141 has been defined .
the input state @xmath142 is encoded in the subsystem @xmath15 , while the output state is inscribed in the subsystem @xmath138 . in order to take into account information deposited in the environment in course of the interaction , we will consider a purification of the state @xmath143 : @xmath144 because the particle is not transferred between the two paths , the purified state has the form @xmath145 where the states @xmath146 , are normalized .
the joint state of the particle and the environment after the interaction can be written as : @xmath147 \nonumber \\ & & \end{aligned}\ ] ] in the first expression transposition is performed on subsystems @xmath15 , while in the second expression only on @xmath6 .
thus , the environment states correlated with the particle present in one or another path are given after normalization by @xmath148 $ ] , where @xmath149 are the normalized internal states of the particle in the upper and the lower arms of the interferometer .
distinguishability can only increase by access to additional subsystems that purify the states .
we will take as purifications of @xmath150 states @xmath151 , where @xmath152 are arbitrary unitaries . for any choice of these unitaries we will have @xmath153 the most stringent bound
is obtained by minimizing the right hand side over @xmath154 and @xmath155 .
this can be equivalently written as @xmath156 , where @xmath157 and @xmath158 is an arbitrary unitary , or using the trace operation @xmath159 the maximum can be expressed in terms of the trace norm as @xmath160 the channel characteristics enters the above formula through @xmath161 where @xmath162 has been defined in eq .
( [ eq : lambda01 ] ) . using the fact that @xmath163 yields eq .
( [ eq : genvis ] ) . the expression for fractional visibilities @xmath128 given in eq .
( [ eq : fracvis ] ) can be rewritten in the tensor form @xmath164 ( { \bf i } \otimes \boldsymbol\lambda_{01 } ) \bigl ( { { { |\phi_+\rangle_{}}}_{ss'}{{}_{}\langle \phi_+| } } \bigr ) \bigr\}.\ ] ] maximization in eq .
( [ eq : v = maxuss ] ) can be extended to all operators @xmath158 satisfying @xmath165 .
therefore for any such operator we have @xmath166 where we used eq .
( [ eq : tre = lambda01 ] ) .
consequently , tracing both sides of eq .
( [ eq : sumklalphas ] ) multiplied by @xmath167 and comparing resulting expressions with eqs .
( [ eq : fracvistensor ] ) and ( [ eq : boundtensor ] ) immediately yields a lower bound on the generalized visibility presented in eq .
( [ eq : genvisbound ] ) .
20 bohr , n. das quantenpostulat und die neuere entwicklung der atomistik .
_ naturwissenschaften _ * 16 * , 245 - 257 ( 1928 ) .
bohr , n. the quantum postulate and the recent development of atomic theory . _ nature _ * 121 * , 580 - 590 ( 1928 ) .
wootters w. k. & zurek w. h. complementarity in the double - slit experiment : quantum nonseparability and a quantitative statement of bohr s principle . _
d _ * 19 * , 473 - 484 ( 1979 ) .
scully m. o. , englert b .- g . & walther h. quantum optical tests of complementarity .
_ nature _ * 351 * , 111 - 116 ( 1991 ) .
greenberger d. m. & yasin a. simultaneous wave and particle knowledge in a neutron interferometer .
. lett . a _ * 128 * , 391 - 394 ( 1988 )
. englert b .-
fringe visibility and which - way information : an inequality .
lett . _ * 77 * , 2154 - 2157 ( 1996 ) .
jaeger g. , shimony a. & vaidman l. two interferometric complementarities .
rev . a _ * 51 * , 54 - 67 ( 1995 )
. bjrk g. & karlsson a. complementarity and quantum erasure in _ welcher weg _ experiments .
rev . a _ * 58 * , 3477 - 3483 ( 1998 ) .
jacques v. _ et al .
_ experimental realization of wheeler s delayed - choice gedanken experiment
. _ science _ * 315 * , 966 - 968 ( 2007 ) .
peruzzo a. , shadbolt p. , brunner n. , popescu s. & obrien j. l. a quantum delayed - choice experiment .
_ science _ * 338 * , 634 - 637 ( 2012 ) .
kaiser f. , coudreau t. , milman p. , ostrowsky d. b. & tanzilli s. entanglement - enabled delayed - choice experiment .
_ science _ * 338 * , 637 - 640 ( 2012 ) .
helstrom c. w. _ quantum detection end estimation theory _
( academic press , 1976 ) .
chuang i. l. & nielsen m. a. prescription for experimental determination of the dynamics of a quantum black box .
opt . _ * 44 * , 2455 - 2467 ( 1997 ) .
poyatos j. f. , cirac j. i. & zoller p. complete characterization of a quantum process : the two - bit quantum gate .
lett . _ * 78 * , 390 - 393 ( 1997 ) . dariano g. m. & lo presti p. quantum tomography for measuring experimentally the matrix elements of an arbitrary quantum operation .
* 86 * , 4195 - 4198 ( 2001 ) .
fuchs c. a. & van de graaf , j. cryptographic distinguishability measures for quantum - mechanical states .
_ ieee trans .
inf . theory _
* 45 * , 1216 - 1227 ( 1999 ) .
nielsen m. a. & chuang i. l. _ quantum computation and quantum information _
( cambridge university press , cambridge , 2000 ) .
uhlmann a. the `` transition probability '' in the state space of a @xmath168-algebra .
_ * 9 * , 273 - 279 ( 1976 ) .
jozsa r. fidelity for mixed quantum states .
_ * 41 * , 2315 - 2323 ( 1994 ) .
scully m. o. & drhl , k. quantum eraser : a proposed photon correlation experiment concerning observation and `` delayed choice '' in quantum mechanics . _
rev . a _ * 25 * , 2208 - 2213 ( 1982 ) .
englert b .-
g . & bergou j. a. quantitative quantum erasure . _ opt .
comm . _ * 179 * , 337 - 355 ( 2000 ) .
kim y .- h .
, yu r. , kulik s. p. & shih y. delayed `` choice '' quantum eraser .
_ * 84 * , 1 - 5 ( 2000 ) .
ma x .- s .
_ et al . _ quantum erasure with causally disconnected choice . _
sci . _ * 110 * , 1221 - 1226 ( 2013 ) .
horodecki k. , horodecki m. , horodecki p. & oppenheim j. secure key from bound entanglement .
lett . _ * 94 * , 160502 ( 2005 ) .
horodecki k. , horodecki m. , horodecki p. & oppenheim j. general paradigm for distilling classical key from quantum states .
_ ieee trans .
inf . theory _ * 55 * , 1898 - 1929 ( 2009 ) .
dobek k. , karpiski m. , demkowicz - dobrzaski r. , banaszek k. & horodecki p. experimental extraction of secure correlations from a noisy private state .
lett . _ * 106 * , 030501 ( 2011 ) .
ozols m. , smith g. & smolin j. a. bound entangled states with secret key and their classical counterpart .
preprint at http://arxiv.org/abs/1305.0848 ( 2013 ) .
banaszek k. , horodecki k. & horodecki p. quantum privacy witness .
rev . a _ * 85 * ,
012330 ( 2012 ) . choi m .- d .
positive linear maps on @xmath169-algebras .
_ * 24 * , 520 - 529 ( 1972 ) . jamiokowski a. linear transformations which preserve trace and positive semidefiniteness of operators .
phys . _ * 3 * , 275 - 278 ( 1972 ) .
we wish to acknowledge insightful discussions with b .-
englert , p. raynal , r. demkowicz - dobrzaski and w. wasilewski .
this research was supported by the european union fp7 project q - essence ( grant agreement no .
248095 ) , era - net project quasar , and the foundation for polish science team project cofinanced by the eu european regional development fund .
all authors contributed extensively to the work presented in this paper .
the authors declare no competing financial interests . | for a particle travelling through an interferometer , the trade - off between the available which - way information and the interference visibility provides a lucid manifestation of the quantum mechanical wave - particle duality .
here we analyze this relation for a particle possessing an internal degree of freedom such as spin .
we quantify the trade - off with a general inequality that paints an unexpectedly intricate picture of wave - particle duality when internal states are involved .
strikingly , in some instances which - way information becomes erased by introducing classical uncertainty in the internal degree of freedom .
furthermore , even imperfect interference visibility measured for a suitable set of spin preparations can be sufficient to infer absence of which - way information .
general results are illustrated with a proof - of - principle single photon experiment .
the duality between wave and particle properties of a microscopic physical system is a founding principle of quantum mechanics.@xcite a canonical illustration is provided by a single particle travelling through a double slit or a mach - zehnder interferometer : an attempt to gain information about the path taken by the particle inevitably reduces the visibility of the interference pattern.@xcite the purpose of this work is to analyze the trade - off between interference visibility and which - way information for a particle equipped with an internal degree of freedom , for example spin . in such a general case
, the interaction with an environment that acquires which - way information can transform in a non - trivial manner the joint path - spin state of the particle .
this opens up an interesting question of how to infer the amount of which - way information deposited in the environment from visibility measurements .
another issue is , to what extent manipulating the spin subsystem can control the information about the path taken by the particle .
a common approach to quantify the amount of which - way information is to use distinguishability @xmath0 , defined as the maximum difference between the probabilities of correct and incorrect identification@xcite of the path taken by the particle inside the interferometer , based on the state of the environment .
we assume here that both paths are equiprobable . in the spinless case ,
the ability for the particle distributed between two paths to interfere is characterized by the visibility @xmath1 , which measures the modulation depth of interference fringes after the paths are combined at the interferometer exit .
these two quantities are related by the inequality@xcite @xmath2 when the particle travelling through the interferometer has an internal structure , the amount of available which - way information depends in principle on the preparation of the spin subsystem and the specifics of the interaction with the environment .
here we demonstrate that in this case the strongest bound on the distinguishability @xmath0 is obtained by replacing @xmath1 with a quantity named generalized visibility that depends on the initial spin preparation and the effective quantum channel experienced by the particle resulting from the interaction with the environment .
further , we present a systematic method to construct estimates for the generalized visibility based on directly measurable quantities , i.e. interference visibilities for particular spin preparations at the input and selections of spin states before combining the paths at the interferometer exit .
this provides an efficient strategy to find an upper bound on the available which - way information without performing full quantum process tomography.@xcite we illustrate the general results with a proof - of - principle single photon experiment that also demonstrates how which - way information can be erased by introducing mixedness in the spin preparation . |
the okubo systems are one of the most important classes of fuchsian systems of ordinary differential equations on the riemann sphere @xmath0 ; for the definition of an okubo system , see section 1 .
a characteristic feature of this class is that each okubo system admits a solution matrix consisting of non - holomorphic solutions around its finite singular points , which we call below _
okubo s canonical solution matrix_. okubo @xcite studied basic properties and in particular established an explicit determinant formula for such a solution matrix ( see theorem [ det : okubo1 ] ) .
this determinant formula guarantees that okubo s canonical solution matrix is in fact a fundamental solution matrix . also , it is because of this determinant formula that okubo systems provide with good models for global analysis of fuchsian systems .
yokoyama @xcite classified the types of irreducible rigid okubo systems under the condition that the nontrivial local exponents are mutually distinct at each finite singular point ; this class consists of eight types @xmath1 , @xmath2 , @xmath3 , @xmath4 and @xmath5 , @xmath6 , @xmath7 , @xmath8 , which is usually referred to as _
yokoyama s list_. for each type in yokoyama s list , haraoka constructed a canonical form of the okubo system @xcite , as well as the corresponding monodromy representation , up to the action of diagonal matrices @xcite .
the purpose of this paper is to propose a method for determining the explicit monodromy representation for okubo s canonical solution matrix , including the diagonal matrix factor which has not been fixed in @xcite .
this problem is in fact reduced to determining the connection coefficients among non - holomorphic solutions for the okubo system .
we solve this connection problem by means of the middle convolutions for schlesinger systems . in this paper
we deal with the particular cases of types @xmath1 , @xmath9 , @xmath2 and @xmath3 ; other cases will be discussed in our forthcoming paper .
it is known by a celebrated work of katz @xcite that any irreducible rigid local system can be constructed by a finite iteration of the so - called katz operations ( additions and middle convolutions ) from a local system of rank one .
dettweiler and reiter @xcite reformulated the katz operations so that they can be applied directly to schlesinger systems and their monodromy representations .
therefore , any irreducible rigid schlesinger system , as well as its monodromy representation , can be constructed in principle from the rank one case . in the case of okubo systems ,
yokoyama independently introduced the notions of extending and restricting operations for differential systems and their monodromy matrices @xcite .
he also proved that any generic rigid okubo system can be constructed by a finite iteration of his operations .
it is clarified by oshima @xcite the katz operations and the yokoyama operations are essentially equivalent as far as the okubo systems are concerned .
it is not completely clarified , however , how the katz ( or yokoyama ) operations can be realized on the level of fundamental solution matrices and their monodromy representations . our method for solving the connection problem is based on the inductive construction of okubo systems by middle convolutions . for a given schlesinger system and a fundamental solution matrix , we give an explicit construction of a fundamental solution matrix of its middle convolution in the sense of @xcite .
we then apply this procedure for constructing the okubo systems in yokoyama s list and their explicit monodromy representations .
this paper is organized as follows .
we propose in section 1 the notion of _ okubo s canonical solution matrix _
@xmath10 for an okubo system @xmath11 of ordinary differential equations .
this solution matrix , introduced by okubo @xcite , consists of non - holomorphic solutions around finite singular points , and forms a fundamental solution matrix under a certain genericity condition .
we also give a remark on the relation between the connection coefficients among non - holomorphic local solutions and the monodromy matrices for okubo s canonical solution matrix .
the monodromy matrices for @xmath10 are determined from the connection coefficients through formula .
our main results are collected in section 2 .
we fix a canonical form for each of the okubo systems of types @xmath1 , @xmath9 , @xmath2 and @xmath3 of yokoyama s list as in theorems 2.1 , 2.2 and 2.3 , and give the explicit monodromy matrices for the corresponding canonical solution matrix in theorem 2.4 , 2.5 and 2.6 .
theses results are proved in subsequent sections by iterations of the katz operations .
our method for determining the connection coefficients and the monodromy matrices is based on the middle convolution for schlesinger systems of @xcite .
the procedure of the middle convolution for a schlesinger system @xmath12 is divided into three steps , which we call the _ convolution _ , the _ k - reduction _ and the _ l - reduction_. analyzing these three steps , in section 3 we clarify how one can construct a fundamental solution matrix for the middle convolution from a given fundamental solution matrix of the system .
a special combination of katz operations @xmath13 ( middle convolution with additions at a singular points ) is used effectively for constructing the okubo systems in yokoyama s list .
in fact , each okubo system in yokoyama s list can be constructed by a finite iteration of such operations from the rank one case .
we formulate in section 4 this type of operations for a general okubo system and its monodromy representation .
we show that the resulting schlesinger system is also an okubo system as explicitly given by . for the monodromy representation ,
provides the moodromy representation of the resulting system specified by .
furthermore we obtain the connection coefficients for the canonical solution matrix of the resulting system as in theorem [ connection ] by applying these results . in section 5
we prove our main theorems for okubo systems of type @xmath1 , @xmath14 , @xmath2 and @xmath3 .
our proof is divided into two steps .
we first compute a certain part of the connection coefficients for each of those okubo systems by applying theorem [ connection ] repeatedly .
second , we determine the other connection coefficients by the symmetry of the okubo system . our argument is based on the fact that the permutaion of characteristic exponents at a singular point can be realized by the adjoint action of a contant matrix .
we remark that our approach is similar to that of yokoyama @xcite in which recursive relations for connection coefficients are investigated in the framework of extending operations for okubo systems .
it is not clear , however , whether the connection coefficients for individual okubo systems in yokoyama s list can be directly determined only from the results of @xcite .
we also remark that the connection problem for rigid irreducible fuchsian differential equations of scalar type has been discussed by oshima @xcite as an application of the katz operations .
in this section we formulate the notion of a _ canonical solution matrix _ for a system of ordinary differential equations of okubo type , and investigate fundamental properties of its connection coefficients and monodromy matrices .
let @xmath15 be @xmath16 distinct points in @xmath17 .
fixing an @xmath16-tuple @xmath18 of positive integers with @xmath19 , we denote by @xmath20 the block diagonal matrix whose diagonal blocks are scalar matrices @xmath21 ( @xmath22 ) , where @xmath23 stands for the @xmath24 identity matrix .
a system of ordinary differential equations on @xmath0 of the form @xmath25 is called an _ okubo system _ of type @xmath18 , where @xmath26 is the column vector of unknown functions . for each @xmath27 , we denote by @xmath28 the @xmath29-block of the matrix @xmath30 : @xmath31 then the okubo system can be equivalently rewritten in the schlesinger form @xmath32 & { \large o } & \end{pmatrix}\quad(k=1,\ldots , r).\ ] ] we denoting by @xmath33 the eigenvalues of the diagonal block @xmath34 ( @xmath35 ) and by @xmath36 the eigenvalues of @xmath37 ; these eigenvalues are subject to the _ fuchs relation _ @xmath38 we assume hereafter that @xmath39 for @xmath35 .
a characteristic feature of a system of okubo type is that , if the local exponents are generic , there exists a _ canonical _ fundamental solution matrix consisting of singular solutions around the finite singular points @xmath40 . in what follows ,
we denote by @xmath41 the vector space of all multivalued holomorphic functions on @xmath42 and by @xmath43 the ring of germs of homomorphic functions at @xmath44 . fixing a base point @xmath45 ,
we identify a multivalued holomorphic function on @xmath46 with a germ of holomorphic function at @xmath47 which can be continued analytically along any continuous path in @xmath46 starting from @xmath48 .
we denote by @xmath49 ( @xmath35 ) and @xmath50 the homotopy classes in @xmath51 of continuous paths encircling @xmath52 and @xmath53 in the positive direction , respectively ; we choose these paths so that @xmath54 in @xmath55 ( see figure 1 ) . in this paper , for two continuous paths @xmath56 , we use the notation @xmath57 to refer to the path obtained by connecting @xmath58 and @xmath59 in this order .
choosing the base point @xmath48 appropriately , we assume that @xmath60 for @xmath61 , and assign the arguments @xmath62 ( @xmath63 ) so that @xmath64 when we consider the behavior of @xmath65 around @xmath52 , we use the analytic continuation of @xmath66 along the line segment connecting @xmath48 and @xmath67 .
= 1.2pt ( 80,80)(0,-14 ) ( 30,46 ) ( 80,46 ) ( 178,46 ) ( 120,43)@xmath68 ( 20,48)@xmath69 ( 70,48)@xmath70 ( 180,48)@xmath71 ( 122,-10)@xmath48 ( 35,25)@xmath72 ( 110,30)@xmath73 ( 165,18)@xmath74 ( 70,34)(-2,1)2 ( 103,32)(-2,3)2 ( 170,29)(3,2)2 if condition is satisfied , for each @xmath35 , system has a unique fundamental solution matrix @xmath75 of the form @xmath76 where we specify the branch of @xmath77 near @xmath47 by @xmath78 .
we call @xmath79 the _ local canonical solution matrix _ of at @xmath52 .
we decompose this @xmath80 as @xmath81 then the @xmath82th block @xmath83 represents a basis of all singular solutions around @xmath52 , and its local monodromy is given by @xmath84 where @xmath85 . collecting the singular solutions @xmath86 ( @xmath35 ) together
, we define the @xmath87 solution matrix @xmath10 by @xmath88 we call this @xmath10 the _ canonical solution matrix _ for the okubo system . as for the determinant of @xmath10 ,
the following theorem is known as okubo s determinant formula ( okubo @xcite , kohno @xcite ) .
[ det : okubo1 ] under the assumption , the determinant of the canonical solution matrix @xmath10 is explicitly given by @xmath89 [ det : okubo2 ] under the assumption , the canonical solution matrix @xmath10 of the okubo system is a fundamental solution matrix if and only if @xmath90 for @xmath91 .
we give some remarks on the relationship between the connection coefficients and the monodromy matrices for the canonical solution matrix @xmath10 of the okubo system .
let @xmath10 be the canonical solution matrix of the okubo system .
the analytic continuation of @xmath10 by @xmath49 ( @xmath35 ) defines an @xmath16-tuple of monodromy matrices @xmath92 as @xmath93 through the analytic continuation along @xmath49 to a neighborhood of @xmath52 , the @xmath94-th column block @xmath95 ( @xmath96 ) is expressed uniquely in the form @xmath97 where @xmath98 is an @xmath99 constant matrix and @xmath100 is an @xmath101 solution matrix which is holomorphic around @xmath52 . since @xmath102
, we have @xmath103 and hence @xmath104 eliminating @xmath105 by .
to summarize , the monodromy matrices for the canonical solution matrix @xmath10 are given as @xmath106 & \ddots & & & \\
( e(a_{kk})-1)c^{(k1)}&\dots & e(a_{kk } ) & \ldots & ( e(a_{kk})-1)c^{(kr ) } \\ & & & \ddots&\\[-4pt ] & & & & 1 \end{pmatrix}\quad(k=1,\ldots , r)\ ] ] in terms of the connection coefficients as in .
we remark that , in order to compute the monodromy matrices of @xmath10 , we need not explicitly determine the holomorphic part @xmath105 of analytic continuation .
yokoyama s list is a class of rigid irreducible okubo systems such that each diagonal block @xmath107 @xmath108 of @xmath30 has @xmath109 mutually distinct eigenvalues and that the matrix @xmath110 is diagonalizable . in this class
we can assume that the diagonal blocks @xmath111 ( @xmath22 ) are diagonal matrices with mutually distinct entries .
we assume that @xmath30 is diagonalized as @xmath112 with mutually distinct @xmath113 .
then the okubo systems in yokoyama s list are characterized by the following eight pairs @xmath18 , @xmath114 of partitions of @xmath115 respectively : @xmath116 ( { \rm ii})_{2n } : & \begin{cases } ( n_1,n_2)=(n , n)\\ ( m_1,m_2,m_3)=(n , n-1,1 ) \end{cases}\quad & ( { \rm ii^\ast})_{2n } : & \begin{cases } ( n_1,n_2,n_3)=(n , n-1,1)\\ ( m_1,m_2)=(n , n ) \end{cases}\quad \\[16pt ] ( { \rm iii})_{2n+1 } : & \begin{cases } ( n_1,n_2)=(n+1,n)\\ ( m_1,m_2,m_3)=(n , n,1 ) \end{cases } & ( { \rm iii^\ast})_{2n+1 } : & \begin{cases } ( n_1,n_2,n_3)=(n , n,1)\\ ( m_1,m_2)=(n+1,n ) \end{cases } \\[16pt ] ( { \rm iv})_{6 } : & \begin{cases } ( n_1,n_2)=(4,2)\\ ( m_1,m_2,m_3)=(2,2,2 ) \end{cases}\quad & ( { \rm iv^\ast})_{6 } : & \begin{cases } ( n_1,n_2,n_3)=(2,2,2)\\ ( m_1,m_2)=(4,2 ) \end{cases } \end{array}\ ] ] in this paper we deal with the four types @xmath117 , @xmath118 , @xmath119 , @xmath120 among these eight types , canonical forms of the okubo systems in yokoyama s list are determined by the work of haraoka @xcite
. we first specify explicit canonical forms of the okubo systems of types @xmath121 , @xmath122 , @xmath123 , @xmath124 @xmath125 as in @xcite , which we will use throughout this paper .
* case @xmath1 * : we express the okubo system of type @xmath121 in the form @xmath126 where @xmath127 and @xmath128 are diagonal matrices defined by @xmath129 the fuchs relation in this case is given by @xmath130 we assume that the parameters of @xmath121 satisfy the following conditions : @xmath131 \rho_i-\rho_j\notin\mathbb{z}\quad(1\le i < j\le n ) , \quad&\rho_i\notin\mathbb{z}\quad(1\le i\le n ) .
\end{array}\ ] ] for the type @xmath121 , the canonical form can be taken as follows [ eq : ci ] canonical form of the okubo system of type @xmath132 is given by the matrices @xmath133 and @xmath134 with the following entries respectively @xmath135 @xmath136 * case @xmath9 * : the okubo system of type @xmath122 is expressed in the form @xmath137 then the fuchs relation is given by @xmath138 we assume that the parameters of @xmath122 satisfy the following conditions : @xmath139 the canonical forms of the type @xmath122 is given by following form : canonical form of the okubo system of type @xmath140 is given by the components @xmath141 with the following entries respectively @xmath135 @xmath142 * cases @xmath2 and @xmath3 * : we express the okubo systems of type @xmath123 and @xmath124 in the form @xmath143 where @xmath127 and @xmath128 are diagonal matrices defined by @xmath144 respectively .
note that the fuchs relations in these cases are given by @xmath145 we always assume that the parameters satisfy the following conditions : @xmath146 \beta_i-\beta_j\notin\mathbb{z}\quad(1\le i < j\le
n ) , \quad & \beta_i\notin\mathbb{z}\quad(1\le i\le n ) , \\[4pt ] \rho_i-\rho_j\notin\mathbb{z}\quad(1\le
i < j\le 3 ) , \quad&\rho_i\notin\mathbb{z}\quad(1\le i\le 3 ) .
\end{array}\ ] ] in the cases of type @xmath123 and @xmath124 , the following canonical forms of the okubo systems are proposed by haraoka @xcite .
canonical forms of the okubo systems of type @xmath123 @xmath147 and @xmath124 @xmath148 are given by the matrices @xmath149 and @xmath150 with the following entries respectively @xmath135 @xmath151 \label{eq : ciii } ( { \rm iii})_{2n+1 } : & & \begin{array}{rll } k_{ij}&={\displaystyle}\frac{\prod_{k\neq i,\ , 1\leq k\leq n+1}(\alpha_k+\beta_j-\rho_1-\rho_2 ) } { \prod_{k\neq j,\ , 1\leq k\leq n}(\beta_j-\beta_k)}\ \ & ( 1\le i\le n+1;\ 1\le j\le n ) \\[16pt ] l_{ij}&=(\alpha_j-\rho_1)(\alpha_j-\rho_2){\displaystyle}\frac{\prod_{k\neq i,\ , 1\leq k\leq n}(\alpha_j+\beta_k-\rho_1-\rho_2 ) } { \prod_{k\neq j,\ , 1\leq k\leq n+1}(\alpha_j-\alpha_k)}\ \ & ( 1\le i\le n;\ 1\le j\le n+1 ) .
\end{array}\end{aligned}\ ] ] to be more precise , the canonical form of haraoka @xcite is the conjugation of our @xmath30 by the diagonal matrix @xmath152 for each okubo system , the monodromy matrices for the canonical solution matrix are expressed in the form by using the connection coefficients .
our main results are the explicit formulas for the connection coefficients and the monodromy matrices for the canonical solution matrices of types @xmath1 , @xmath9 , @xmath2 and @xmath3 . in what follows ,
we use the notation @xmath153 @xmath154 with the convention of arguments such that @xmath155 for @xmath156 , and @xmath157 for @xmath158 , using @xmath159 as specified in section 1.1 . *
case @xmath1 * : for the okubo system of type @xmath121 , the canonical solution matrix @xmath160 is a multivalued holomorphic solution matrix on @xmath161 characterized by the following conditions : @xmath162 around @xmath163 , and @xmath164 around @xmath165 , where @xmath166 denotes the @xmath94-th unit column vector . according to the representation of the previous section , the monodromy matrices @xmath167 are expressed in the form @xmath168 in terms of the connection matrices @xmath169 , @xmath170 in .
these connection matrices @xmath171 and @xmath172 are defined as @xmath173 around @xmath163 and @xmath174 around @xmath165 . as to the monodromy around @xmath53
, we remark that @xmath175 where @xmath176 .
the connection matrices @xmath177 and @xmath46 are given by the following theorem : [ maini ] the monodromy matrices @xmath178 , @xmath179 for the canonical solution matrix @xmath10 of type @xmath121 is expressed as in terms of the connection matrices @xmath171 and @xmath172 determined as follows @xmath135 @xmath180 * case @xmath9 * : for the okubo system of type @xmath122 , the canonical solution matrix @xmath181 is a multivalued holomorphic solution matrix on @xmath182 characterized by the following conditions : @xmath183 around @xmath184 .
the monodromy matrices @xmath185 are expressed in the form @xmath186 these connection matrices @xmath187 are defined as @xmath188 [ maini * ] the monodromy matrices @xmath185 for the canonical solution matrix @xmath10 of type @xmath189 is expressed as in terms of the connection matrices @xmath187 determined as follows @xmath135 @xmath190 & { \displaystyle}e(\tfrac{\rho_1}{2 } ) \frac{\prod_{k\neq i\,1\leq k\leq n}(t_i - t_k)^{\alpha_k-\rho_1 } } { \prod_{k\neq j,\,1\leq k\leq n}(t_j - t_k)^{\alpha_k-\rho_1 } } \frac{\gamma(-\alpha_i)\gamma(\alpha_j+1 ) } { \gamma(\alpha_j-\rho_1)\gamma(1+\rho_1-\alpha_i ) } \quad(i > j ) .
\end{cases}\ ] ] * cases @xmath2 and @xmath3 * : for the okubo system type @xmath123 and @xmath124 , the canonical solution matrix @xmath191 is a multivalued holomorphic solution matrix on @xmath161 characterized by the following conditions : @xmath192 around @xmath163 , and @xmath193 around @xmath165 . according to the representation of the previous section ,
the monodromy matrices @xmath167 are expressed in the form @xmath194 in terms of the connection matrices @xmath169 , @xmath170 in .
these connection matrices @xmath195 and @xmath196 are defined as @xmath197 around @xmath163 and @xmath198 around @xmath165 . as to the monodromy around @xmath53
, we remark that @xmath199 where @xmath200 .
[ mainii , iii ] the monodromy matrices @xmath178 , @xmath179 for the canonical solution matrices @xmath10 of types @xmath123 and @xmath124 are expressed as in terms of the connection matrices @xmath195 and @xmath196 determined as follows @xmath135 @xmath201 d_{ij } & = ( -1)^{n-1 } { \displaystyle}{\frac{(t_2-t_1)^{\rho_3-\beta_i}}{(t_1-t_2)^{\rho_3-\alpha_j } } e(\tfrac{1}{2}(\alpha_j+\beta_i-\rho_3 ) ) \frac{\gamma(-\beta_{i})\gamma(\alpha_j+1)}{\gamma(\alpha_j-\rho_1)\gamma(1+\rho_1-\beta_i)}}\\ & { \displaystyle}{\frac{\prod_{k\neq j,\,1\leq k\leq n}\gamma(\alpha_j-\alpha_{k } ) } { \prod_{k\neq i,\,1\leq k\leq n}\gamma(\alpha_j+\beta_{k}-\rho_1-\rho_2 ) } \frac{\prod_{k\neq i,\,1\leq k\leq n}\gamma(1+\beta_{k}-\beta_i ) } { \prod_{k\neq j,\,1\leq k\leq n}\gamma(1+\rho_1+\rho_2-\alpha_k-\beta_i ) } } \quad ( 1\leq i , j\leq n ) , \end{array}\end{aligned}\ ] ] @xmath202 d_{ij } & = ( -1)^{n-1}{\displaystyle}{e(\tfrac{1}{2}(\alpha_j+\beta_i-\rho_3 ) ) \frac{(t_2-t_1)^{\rho_3-\beta_i}}{(t_1-t_2)^{\rho_3-\alpha_j } } \frac{\gamma(-\beta_{i})\gamma(\alpha_j+1)}{\gamma(1+\rho_1-\alpha_j)\gamma(1+\rho_2-\alpha_j)}}\\ & { \displaystyle}{\frac{\prod_{k\neq j,\,1\leq k\leq n+1}\gamma(\alpha_j-\alpha_{k } ) } { \prod_{k\neq i,\,1\leq k\leq n}\gamma(\alpha_j+\beta_{k}-\rho_1-\rho_2 ) } \frac{\prod_{k\neq i,\,1\leq k\leq n}\gamma(1+\beta_{k}-\beta_i ) } { \prod_{k\neq j,\,1\leq
k\leq n+1}\gamma(1+\rho_1+\rho_2-\alpha_k-\beta_i ) } } \,\left(\begin{array}{ll } 1\leq i\leq n\\ 1\leq j\leq n+1 \end{array}\right ) . \end{array}\end{aligned}\ ] ] in the following sections , we prove these theorems by the method of middle convolutions .
in this section , we briefly recall the definitions of the katz operations ( the addition and the middle convolution ) for a schlesinger system and its monodromy representation ( @xcite ) . for each @xmath16-tuple of matrices
@xmath203 , we consider the schlesinger system @xmath204 of ordinary differential equations on @xmath205 .
we denote by @xmath206 the residue matrix at @xmath53 . for an @xmath16-tuple @xmath207 ,
the addition @xmath208 of @xmath209 by @xmath210 is defined simply as @xmath211 the corresponding schlesinger system is given by @xmath212 let @xmath213 be a fundamental solution matrix of the schlesinger system associated with an @xmath16-tuple of matrices @xmath214 .
then the monodromy matrices @xmath215 ( @xmath35 ) defined as @xmath216 give rise to an @xmath16-tuple @xmath92 of invertible matrices .
the addition for an @xmath16-tuple @xmath217 is defined by @xmath218 for each @xmath219 , where @xmath220 .
we remark that the schlesinger system is obtained from by the operation @xmath221 then the monodromy matrices for the fundamental solution matrix @xmath222 of is given by with @xmath223 @xmath224 .
the middle convolution @xmath225 of @xmath209 with parameter @xmath226 is defined through three steps ( @xcite ) .
* step 1 * ( convolution ) : we first define the _ convolution _ @xmath227 of @xmath209 with parameter @xmath228 as the @xmath16-tuple of matrices @xmath229 by setting @xmath230 where @xmath231 is zero outside the @xmath82-th block row .
the schlesinger system associated with @xmath232 can be written in the form of an okubo system @xmath233 where @xmath234 and @xmath235 .
* step 2 * ( @xmath236-reduction ) : setting @xmath237 , we decompose @xmath238 @xmath224 as @xmath239 with two matrices @xmath240 , @xmath241 , and take the block matrix @xmath242 since the @xmath29-block of @xmath243 is given by @xmath244 , we have @xmath245 , namely @xmath246 hence , the system gives rise to the okubo system @xmath247 for @xmath248 .
equivalently , we obtain the schlesinger system @xmath249 corresponding to @xmath250 , where @xmath251 * step 3 * ( @xmath252-reduction ) : setting @xmath253 , we further decompose @xmath254 as @xmath255 by two matrices @xmath256 , @xmath257 .
since @xmath258 is expressed as @xmath259 , @xmath260 , we have @xmath261 hence , multiplying by @xmath262 we obtain @xmath263 setting @xmath264 , we obtain the schlesinger system @xmath265 associated with @xmath266 .
we call the system the _ middle convolution _ of with parameter @xmath228 .
we note that these matrices @xmath267 @xmath108 are realizations of the linear transformations on @xmath268 induced from @xmath269 .
we next recall the middle convolution of monodromy matrices .
let @xmath213 be a fundamental solution matrix of the schlesinger system associated with an @xmath16-tuple of matrices @xmath214 .
then the monodromy matrices @xmath215 ( @xmath35 ) defined as @xmath270 give rise to an @xmath16-tuple @xmath92 of invertible matrices .
the multiplicative middle convolution @xmath271 we are going to explain below provides a way to construct the monodromy matrices for a certain fundamental solution matrix of the schlesinger system associated with the middle convolution @xmath272 with @xmath273 .
let @xmath274 be an arbitrary @xmath16-tuple of invertible matrices .
the multiplicative middle convolution @xmath275 of @xmath276 with parameter @xmath277 is constructed through three steps .
* step 1 * ( convolution ) : we define an @xmath16-tuple of invertible matrices @xmath278 as follows : @xmath279 then we call @xmath280 the _ convolution _ of @xmath276 with parameter @xmath281 , and denote it by @xmath282 . as we will see later , @xmath282 represents the @xmath16-tuple of monodromy matrices for a fundamental solution matrix of the schlesinger system @xmath227 .
* step 2 * ( @xmath283-reduction ) : setting @xmath284 , we decompose @xmath285 as @xmath286 with two matrices @xmath287 . taking a right inverse @xmath288 of @xmath289 for each @xmath35 so that @xmath290 , we define the block matrices @xmath291 and @xmath292 by @xmath293 where @xmath294 .
since the @xmath295-block of @xmath296 is given as @xmath297 by @xmath298 we obtain @xmath299 , i.e. @xmath300 , where @xmath301 which gives a @xmath302-reduction of the middle convolution @xmath303 . * step 3 * ( @xmath304-reduction ) : to construct monodromy matrices corresponding to the middle convolution @xmath305 , we set @xmath306 , and decompose @xmath307 , as @xmath308 by two matrices @xmath309 , @xmath310 with @xmath311 , and take a right inverse @xmath312 so that @xmath313
. then we obtain an @xmath16-tuple of monodromy matrices @xmath314 as @xmath315 we call the @xmath16-tuple @xmath316 the _ middle convolution _ of @xmath276 with parameter @xmath281 , and denote by @xmath305 .
we remark that these matrices @xmath317 ( @xmath22 ) are realizations of the linear transformations on @xmath318 induced from @xmath319 , where @xmath320 .
let @xmath213 be a fundamental solution matrix of the schlesinger system of and @xmath321 the @xmath16-tuple of invertible matrices defined by the monodromy representation as in .
we summarize below how one can obtain fundamental solution matrices whose monodromy representations correspond to the convolution @xmath282 and the middle convolution @xmath305 of @xmath276 . following the construction by @xcite , for a complex parameter @xmath322 we consider the @xmath323 block matrices @xmath324 called the _ euler transforms _ of @xmath213 . here
@xmath325 denotes the double loop in the @xmath326-plane @xmath327 encircling @xmath328 and @xmath329 for each @xmath35 .
we use the symbols @xmath330 and @xmath331 for the generators of @xmath332 encircling @xmath333 and @xmath329 in the positive direction such that @xmath334 .
then the double loops @xmath325 ( @xmath35 ) are expressed as @xmath335 = \alpha_k^{-1}\alpha_x^{-1}\alpha_k\alpha_x$ ] ( see figure 2 ) .
we also set @xmath336=\alpha_\infty^{-1}\alpha_x^{-1}\alpha_\infty\alpha_x$ ] .
= 1.2pt ( 80,80)(0,-14 ) ( 34,46 ) ( 73,46 ) ( 153,46 ) ( 192,46 ) ( 105,43)@xmath68 ( 30,56)@xmath69 ( 73,56)@xmath70 ( 148,56)@xmath71 ( 190,56)@xmath337 ( 110,0)@xmath48 ( 50,20)@xmath338 ( 95,35)@xmath339 ( 123,35)@xmath340 ( 170,20)@xmath331 ( 67,36)(-2,1)2 ( 95,32)(-2,3)2 ( 147,33)(3,2)2 ( 175,31)(3,2)2 we define the @xmath341 block matrix @xmath342 by @xmath343 & = \bigg ( \int_{l_j}(x - u)^\mu\,y(u)\frac{du}{u - t_i } \bigg)_{1\le i , j\le r } \in { \rm mat}(nr;\mathcal{o}(\widetilde{d } ) )
. \end{split}\ ] ] it is known by @xcite that @xmath342 is a solution matrix of the okubo system associated with @xmath344 , and that the monodromy of @xmath342 is described as @xmath345 in terms of the convolution @xmath346 of @xmath276 with @xmath273 .
[ thm : iy ] suppose that the following three conditions are satisfied .
* @xmath347 .
* for each @xmath35 and @xmath348 , any pair of distinct eigenvalues of @xmath238 has no integer difference .
* @xmath238 @xmath224 and @xmath349 have no eigenvalue in @xmath350 .
then the @xmath341 block matrix @xmath351 is a fundamental solution matrix of the okubo system associated with @xmath344 .
we prove that @xmath342 is a fundamental solution matrix .
the fundamental solution matrix @xmath213 of is written of the form @xmath352 around @xmath52 for each @xmath35 , and @xmath353 around @xmath53 .
we investigate the local behavior of @xmath354 at @xmath184 .
changing the integration variable by @xmath355 , for @xmath356 we have @xmath357}\frac{(1-v)^{\mu}}{(x - t_j)v-(t_i - t_j ) } f^{(j)}((x - t_j)v+t_j)v^{a_j}(x - t_j)^{a_j+\mu+1}c^{(j)}dv \nonumber \\ & = & \int_{l[0,1]}\frac{(1-v)^{\mu}}{t_j - t_i } f^{(j)}(t_j)v^{a_j}dv\,(x - t_j)^{a_j+\mu+1}c^{(j)}\big(1+o(x - t_j)\big ) \nonumber \\ & = & \frac{\widetilde b(a_j+1,\mu+1 ) } { t_j - t_i } ( x - t_j)^{a_j+\mu+1}c^{(j)}\big(1+o(x - t_j)\big ) , \end{aligned}\ ] ] where @xmath358}u^{a-1}(1-u)^{\beta-1}du \\ & = ( e(a)-1)(e(\beta)-1)b(a,\beta ) \qquad(a\in{\rm mat}(n;\mathbb{c}),\ \beta\in\mathbb{c } ) \end{split}\ ] ] denotes the regularized beta function with a matrix argument . in this
integral , as the base point we take a point in the interval @xmath359 with @xmath360 .
similarly we have @xmath361 since @xmath362 has no eigenvalue in @xmath350 by the assumption , we have @xmath363 , and hence @xmath364 for @xmath22 . similarly , @xmath365 for @xmath22 .
consider the vector space @xmath366 where we regard square matrices as linear transformations acting on column vectors .
[ lem : keriy ] under the assumption of theorem [ thm : iy ] , we have @xmath367 .
assume that a column vector @xmath368 belongs to @xmath369 , namely , @xmath370 .
then we have @xmath371 for @xmath35 .
this implies that the @xmath82-th row of @xmath342 satisfies @xmath372 since @xmath373 , we have @xmath374 hence we find that @xmath375 @xmath224 . to complete the proof , we assume @xmath376
. then by lemma [ lem : keriy ] , @xmath377 is expressed as @xmath378 for some @xmath379 . in the following we set @xmath380 for each @xmath381 .
note that by @xmath382 we have @xmath383 in this notation @xmath384 is computed as follows : @xmath385 by @xmath386 we obtain @xmath387 since @xmath388 ( @xmath22 ) , @xmath370 implies @xmath389 , and hence @xmath390 .
this completes the proof of theorem [ thm : iy ] .
we set @xmath391 for @xmath35 , and @xmath294 .
[ thm : iykpart ] in addition to the conditions @xmath392i@xmath393 , @xmath392ii@xmath393 , @xmath392iii@xmath393 of theorem [ thm : iy ] , suppose that the following condition is satisfied .
* for each @xmath35 , @xmath238 has non - integer eigenvalues @xmath394 where @xmath391 . using the constant matrices @xmath395 and @xmath396 as in and respectively , we set @xmath397 , @xmath398
. then the @xmath399 matrix @xmath400 is a fundamental solution matrix of the @xmath236-reduction of the convolution , and the monodromy matrices for @xmath400 are given by the @xmath283-reduction @xmath401 of @xmath402 .
[ thm : iylpart ] in addition to the conditions @xmath392i@xmath393 , @xmath392ii@xmath393 , @xmath392iii@xmath393 , @xmath392iv@xmath393 above , suppose that the following condition is satisfied .
* @xmath349 has non - integer eigenvalues @xmath403 where @xmath404 . using @xmath262 and @xmath405 as in and respectively , set @xmath406 .
then @xmath407 is a fundamental solution matrix of the schlesinger system corresponding to the middle convolution , and the monodromy matrices for @xmath407 are given by the middle convolution @xmath408 . as we already verified
, @xmath409 is a solution matrix of the @xmath236-reduction .
also , under the assumption @xmath410 , we have @xmath411 for @xmath35 . for @xmath289 , @xmath412 in ,
we take @xmath413 , @xmath414 such that @xmath415 since @xmath416 , we have @xmath417 we consider solution matrix @xmath418 of defined by @xmath419 noting that @xmath420 we compute the analytic continuation by @xmath49 as @xmath421 i_{l_k}^{\mu}y(x)e(\mu)s'_k\quad & ( j = k ) \end{cases}\ ] ] by @xmath422 . hence we have @xmath423 however , under the assumption @xmath410 , @xmath424 is _ not _ an eigenvalue of @xmath49 on the solution space of the @xmath236-reduction .
this means that @xmath425 @xmath224 , namely , @xmath426 since @xmath427 , the matrix @xmath428 for any regular point @xmath429 .
hence we have @xmath430 .
this completes the proof of theorem [ thm : iykpart ] .
theorem [ thm : iylpart ] can be proved in a similar way .
we consider the operation @xmath431 for an okubo system with residue matrices @xmath432 , where two additions are applied at @xmath52 ( @xmath35 ) .
the okubo systems of types @xmath1 , @xmath2 and @xmath3 are included in the class of the differential systems obtained from @xmath433 by a finite iteration of operations in the form .
we show that the schlesinger system corresponding to is an okubo system if the parameters @xmath434 and @xmath435 are generic in the sense that @xmath436 and @xmath437 . with the notations of and
, we first decompose the components of @xmath438 as @xmath439 for the decomposition of @xmath440 , we use the following lemma . [
lem : xxieta ] let @xmath441 be an @xmath87 block matrix of type @xmath18 , @xmath19 , where @xmath442 @xmath443 .
suppose that @xmath444 and @xmath445 for an index @xmath35 and set @xmath446 .
then there exist matrices @xmath447 such that @xmath448 furthermore , the @xmath449 matrix @xmath450 and the @xmath451 matrix @xmath452 are of maximal rank . without loss of generality
we assume @xmath453 and set @xmath454 noting that @xmath46 is invertible , we decompose this matrix into the form @xmath455 since @xmath456 , there exist matrices @xmath457 and @xmath458 of rank @xmath459 such that @xmath460 .
from @xmath461 we obtain @xmath462 where @xmath463 and @xmath464 .
if we apply lemma [ lem : xxieta ] to the matrix @xmath440 with the invertible block @xmath465 , then formula implies the decomposition @xmath466 of @xmath440 where @xmath467 in this case , the linear mapping @xmath468 induces the isomorphism from the quotient space @xmath469 for the middle convolution to @xmath470 therefore we redefine @xmath395 as @xmath471 & \ddots&\vdots \\[-4pt ] & & q_0 \\[-4pt ] & & \vdots & \ddots\\[-4pt ] & & -q_r & & q_r\ , \end{pmatrix}.\ ] ] then , from the convolution @xmath472 by the transformation @xmath473 we obtain the schlesinger system @xmath474 corresponding to the operation , where @xmath475 a_{l1 } & \cdots & q_lp_0 & \cdots & a_{lr } \\[4pt ] & & { \large o } & & \\
\end{pmatrix } -(\rho + c)e_{ll } \ \ ( l\neq k),\quad \widehat b_k= \begin{pmatrix } -q_1 \\[-3pt ] \vdots \\[-2pt ] q_0 \\[-3pt ] \vdots \\[-2pt ] -q_r \end{pmatrix } \begin{pmatrix } p_1 & \cdots p_0 & \cdots & p_r \end{pmatrix}\ ] ] with @xmath476 in the @xmath477 block of @xmath478 .
then the schlesinger system is transformed into the okubo system @xmath479 for @xmath480 , where @xmath481 \ a_{ij}-(\rho+c)\delta_{ij}\ & \vdots & \vdots & a_{ij } \\[-3pt ] & a_{k-1,k}(a_{kk}+c)(a_{kk}-\rho)^{-1 } & ( \rho+c)\xi_{k-1 } & & \\[4pt ] a_{k1}\ \ \ldots\ \ a_{k , k-1 } & a_{kk } & 0 & a_{k , k+1}\ \ \ldots \
\ a_{kr } \\[2pt ] \eta_{1}\ \ \ \ldots\ \ \ \eta_{k-1 } & 0 & \rho & \eta_{k+1}\ \ \
\ \eta_{r } \\[2pt ] & a_{k+1,k}(a_{kk}+c)(a_{kk}-\rho)^{-1 } & ( \rho+c)\xi_{k+1 } & \\[-4pt ] a_{ij } & \vdots & \vdots & a_{ij}-(\rho+c)\delta_{ij}\\[-3pt ] & a_{rk}(a_{kk}+c)(a_{kk}-\rho)^{-1 } & ( \rho+c)\xi_{r } & \end{pmatrix}\ ] ] and @xmath482 & \ddots & & \vdots&\vdots\\[-4pt ] & & 1 & a_{k-1k}(a_{kk}-\rho)^{-1 } & \xi_{k-1}\\[2pt ] & & & 1 & 0\\ & & & 0 & 1\\ & & & a_{k+1k}(a_{kk}-\rho)^{-1 } & \xi_{k+1}&1\\[-4pt ] & & & \vdots&\vdots & & \ddots\\[-4pt ] & & & a_{rk}(a_{kk}-\rho)^{-1 } & \xi_{r}&&&1 \end{pmatrix}.\ ] ] we remark that the transformation from the convolution to the okubo system is given by @xmath483 : @xmath484 [ lem : mcadd ] if the okubo system satisfies the conditions @xmath485 and @xmath437 for an index @xmath35 , then the schlesinger system with residue matrices is equivalent to the okubo system .
let @xmath10 be the canonical solution matrix of the okubo system .
the analytic continuation @xmath10 by @xmath49 @xmath224 gives an @xmath16-tuple of monodromy matrices @xmath276 as defined in section @xmath486
. then @xmath276 is _ of okubo type _ in the sense that @xmath487 are expressed as @xmath488 in terms of matices @xmath489 ( @xmath490 ) .
we show that the operation @xmath491 gives rises to monodromy matrices of okubo type as in the case of okubo system .
we decompose @xmath492 and take @xmath493 as @xmath494 also , setting @xmath495 , we take the decomposition @xmath496 for @xmath497 as in lemma [ lem : xxieta ] , and define @xmath498 , @xmath499 , @xmath405 so that @xmath500 , @xmath501 . as in the case of the differential system , the linear mapping @xmath502 @xmath503 induces the isomorphism from the quotient space @xmath504 for the middle convolution to @xmath505 therefore , taking block matrices @xmath506 & \ddots & & \vdots & & & \\[-4pt ] & & { \mathcal{q}}_{k-1 } & s{\mathcal{q}}_{k-1 } m_k & & & \\ & & & { \mathcal{q}}_0(m_{k+1}\cdots m_r)^{-1 } & & & \\ & & & { \mathcal{q}}_{k+1}m_{k+1}^{-1 } & { \mathcal{q}}_{k+1 } & & \\[-4pt ] & & & \vdots & & \ddots & \\[-4pt ] & & & { \mathcal{q}}_r(m_{k+1}\cdots m_r)^{-1 } & & & { \mathcal{q}}_r \end{pmatrix } , \quad { \mathcal{s}}= \begin{pmatrix } { \mathcal{s}}_1 & & & sm_{2}\cdots m_r{\mathcal{s}}_0 & & \\[-4pt ] & \ddots & & \vdots & & \\[-4pt ] & & { \mathcal{s}}_{k-1}&s m_k\cdots m_r{\mathcal{s}}_0&\\ & & & m_{k+1}\cdots m_r{\mathcal{s}}_0 & & \\ & & & m_{k+2}\cdots m_r{\mathcal{s}}_0 & { \mathcal{s}}_{k+1 } & \\[-4pt ] & & & \vdots & & \ddots & \\[-4pt ] & & & { \mathcal{s}}_0 & & & { \mathcal{s}}_r \end{pmatrix}},\ ] ] we obtain the monodromy matrices @xmath507 corresponding to @xmath508 as @xmath509 & \ddots & & & & \\
\lambda s^{-1}m_{i1 } & \cdots & \lambda s^{-1}m_{ii } & \cdots & m_{ir}\\[-4pt ] & & & \ddots & & \\ & & & & 1 \end{pmatrix}\quad(i\ne k)\ ] ] with the @xmath510-block replaced by @xmath511 , and @xmath512 we remark here that the @xmath513-th rows of @xmath514 and @xmath515 are given respectively as follows : @xmath516 ( \lambda m_{i}\cdots m_{k})_{i}\quad & ( 1\leq i\leq k ) , \end{cases}\\[2pt ] ( m_{0}^{(k)})^{-1}_i&=\begin{cases } ( \lambda^{-1}m_{i}^{-1}\cdots m_{k+1}^{-1})_{i}\quad & ( k+1\leq i\leq r ) \\[2pt ] ( \lambda^{-1 } m_{i}^{-1}\cdots m_{1}^{-1}m_{r}^{-1}\cdots m_{k+1}^{-1})_{i}\ \ & ( 1\leq i\leq k ) . \end{cases } \end{split}\ ] ] using these relations , we rewrite @xmath517 as @xmath518 -\big(\delta_{ij}-\lambda \widetilde{m}_{ij}^{(k)}\big ) _ { i = k+1,\ldots , r\atop j=1\ldots r\hfill } \end{pmatrix } \begin{pmatrix } \lambda s^{-1}{\mathcal{p}}_1 & \cdots & { \mathcal{p}}_0 & \cdots & { \mathcal{p}}_r \end{pmatrix},\ ] ] where @xmath519 . noting that @xmath520
, we set @xmath521 & \ddots & & \vdots \\[-2pt ] & & 1 & \lambda^{-1}s({\mathcal{p}}_0^{(k)})_{k-1}\\[2pt ] & & & 1 \\ & & & \lambda ( m_0^{(k)})^{-1}_{k+1 } { \mathcal{p}}_0^{(k ) } & 1 \\[-2pt ] & & & \vdots & & \ddots \\[-2pt ] & & & \lambda ( m_0^{(k)})^{-1}_{r } { \mathcal{p}}_0^{(k ) } & & & 1 \end{pmatrix},\ ] ] where @xmath522 stands for the @xmath513-th row of @xmath523 . using this matrix @xmath524 for the conjugation , from @xmath525
we obtain the following monodromy matrices of okubo type @xmath526 : @xmath527 & & \ddots & & & & \\ \lambda s^{-1 } m_{i1 } & & \cdots & \lambda s^{-1}m_{ii}&\cdots & \begin{pmatrix } m_{i(k1)}^{mc } & m_{i(k2)}^{mc } \end{pmatrix } \cdots & m_{ir}\\ & & & & \ddots&&\\[-4pt ] & & & & & & 1 \end{pmatrix}\,(i < k)\\ m_{i}^{mc}&= \begin{pmatrix } 1&&&&&&\\[-4pt ] & & \ddots & & & & \\
\lambda s^{-1}m_{i1 } \cdots & \begin{pmatrix } m_{i(k1)}^{mc } & m_{i(k2)}^{mc } \end{pmatrix } & \cdots & \lambda s^{-1}m_{ii } & & \cdots&m_{ir}\\ & & & & & \ddots&\\[-4pt ] & & & & & & 1 \end{pmatrix}\,(k < i)\\ m_k^{mc}&= \begin{pmatrix } 1 & & & & & \\
& \ddots&&&&\\ \lambda s^{-1}m_{k1 } & \cdots & m_{kk } & 0 & \cdots & m_{kr}\\ s^{-1}\eta_{1 } & \cdots & 0 & \lambda^{-1 } & \cdots & \lambda^{-1}\eta_{r}\\ & & & & \ddots&\\ & & & & & 1 \end{pmatrix } \end{split}\ ] ] where @xmath528 and @xmath529 are computed as follows : @xmath530 for @xmath531 , and @xmath532 for @xmath533 .
note that @xmath534 is the @xmath16-tuple of monodromy matrices for the fundamental solution matrix @xmath535 , where @xmath536 .
therefore we need to specify the right inverse @xmath537 of @xmath538 to solve the connection problem . in this case the matrices @xmath538 and @xmath537 are given as @xmath539 @xmath540 we remark that one can take @xmath541 in the form @xmath542 summarizing these arguments , we conclude : [ prop : mcadd ] the operation gives rise to the @xmath16-tuple of monodromy matrices of okubo type @xmath534 as in , .
furthermore @xmath534 is realized as the monodromy matrices for the fundamental solution matrix @xmath535 , where @xmath543 .
we consider the okubo system obtained from the okubo system by the operation .
let @xmath544 be the canonical solution matrix of .
then the monodromy matrices @xmath545 @xmath108 corresponding to are represented as @xmath546 where @xmath547 , and @xmath548 we now investigate the relation between @xmath187 and @xmath549 . from the forms of , the fundamental solution matrix @xmath535 is transformed into @xmath550 by some block diagonal matrix @xmath551 .
therefore we have recurrence relations @xmath552 the new connection coefficients @xmath553 are also given as @xmath554 looking at the diagonal blocks of the solution matrix @xmath555 and @xmath556 , we can compute @xmath557 as follows : @xmath558 using these formulas , the recurrence relations for the connection coefficients are rewritten as @xmath559 \displaystyle \left(\frac{t_i - t_k}{t_j - t_k}\right)^{\rho+c } \frac{e(\frac{1}{2}(\rho+c))\gamma(\rho+c - a_{ii})}{\gamma(-a_{ii } ) } c_{ij}\frac{\gamma(a_{jj}-\rho - c+1)}{\gamma(a_{jj}+1 ) } \ & ( i < j;\,i , j\neq k ) , \end{cases } \end{split}\ ] ] where @xmath560 if @xmath156 and @xmath561 otherwise . similarly ,
@xmath562 \displaystyle ( t_i - t_k)^{\rho+c}e(\tfrac{-1}{2}(\rho+c ) ) \frac{\gamma(\rho+c - a_{ii})}{\gamma(-a_{ii } ) } c_{ik}\frac{\gamma(a_{kk}-\rho)}{\gamma(a_{kk}+c ) } \quad & ( k <
i ) \end{cases } \end{split}\ ] ] @xmath563 \displaystyle -e(\tfrac{1}{2}(\rho+c))(t_j - t_k)^{-\rho - c } \frac{\gamma(1+\rho - a_{kk})}{\gamma(1-a_{kk}-c)}c_{kj } \frac{\gamma(a_{jj}-\rho - c+1)}{\gamma(a_{jj}+1 ) } \quad & ( k < j ) \end{cases } \end{split}\ ] ] [ connection ] the connection coefficients @xmath564 and @xmath565 , @xmath566 @xmath567 for the canonical solution matrix @xmath568 are determined from the connection coefficients @xmath187 for @xmath10 by the recurrence formulas and , respectively .
it seems that the matrices @xmath569 and @xmath570 can not be written as products of gamma functions in general , while they can be described as the limits as @xmath571 of certain explicit integrals . in this paper
we do not go into the detail of the description of @xmath569 and @xmath570 , since they can be determined by symmetry arguments in the context of the okubo systems in yokoyama s list that will be discussed below .
in what follows , we use the symbols @xmath121 , @xmath122 , @xmath123 , and @xmath124 to refer to the corresponding tuples @xmath572 of residue matrices as in . as for the nontrivial eigenvalues
, we also use the notations @xmath573 , @xmath574 , @xmath575 in order to specify the rank @xmath576 of the okubo system .
* case @xmath1 * : we first construct @xmath577 by @xmath578 starting from the differential equation @xmath579 of rank @xmath580 . the resulting system is given by @xmath581 the canonical form of type @xmath582 is obtained as the conjugation of @xmath30 by the matrix @xmath583 starting with @xmath582 , the okubo systems @xmath584 can be constructed inductively by the katz operations @xmath585 [ lem : xietaeqi ] for the system of type @xmath584 , @xmath586 is a matrix of rank 1 .
it is expressed as @xmath587 with @xmath588 and @xmath589 defined by @xmath590 we show how the canonical form of the okubo system @xmath591 is obtained from the canonical form of @xmath132 by the operation of . by lemma [ lem : mcadd ] the katz operation for @xmath132 gives rise to the okubo system @xmath592 where @xmath593 this matrix @xmath594 is precisely the canonical form of okubo system @xmath595 with characteristic exponents @xmath596 specified by @xmath597 * case @xmath9 * : the system of type @xmath122 is obtained by @xmath598 with a generic parameter @xmath228 from the schlesinger system @xmath599 of rank 1 , where @xmath600 @xmath601 .
the resulting system is given by @xmath602 then characteristic exponents @xmath603 and @xmath604 are determined as @xmath605 * cases @xmath2 and @xmath3 * : the okubo system @xmath606 as in is obtained from the system by @xmath607 with @xmath608 the okubo systems @xmath609 and @xmath610 can be constructed inductively by the following katz operations : @xmath611 [ lem : xietaeq ] @xmath612 for the system of type @xmath609 , @xmath613 is a matrix of rank 1 .
it is expressed as @xmath614 with a column vector @xmath588 and a row vector @xmath589 defined by @xmath615 @xmath616 for the system of type @xmath610 , @xmath617 is a matrix of rank 1 .
it is expressed as @xmath618 with a column vector @xmath588 and a row vector @xmath589 defined by @xmath619 this lemma is essentially used in @xcite for the construction of canonical forms .
the explicit forms of @xmath588 and @xmath589 described above can also be derived by the property of cauchy matrices as in @xcite . * from @xmath620 to @xmath621 * we show how the canonical form of the okubo system @xmath622 is obtained from the canonical form of @xmath623 by the operation of . by lemma [ lem : mcadd ] the katz operation for @xmath623
gives rise to the okubo system @xmath592 where @xmath624 this matrix @xmath594 is precisely the canonical form of okubo system @xmath123 with characteristic exponents @xmath625 , @xmath626 specified by @xmath627 * from @xmath621 to @xmath628 * similarly one can construct the canonical form of the okubo system @xmath629 from @xmath124 by the middle convolution . from lemma [ lem : mcadd ] ,
the schlesinger system of for @xmath124 is given by the okubo system @xmath592 of type @xmath630 , where @xmath631 the characteristic exponents @xmath632 are determined as @xmath633 note that the passage from @xmath634 to @xmath635 is also included in this procedure with @xmath636 replaced by @xmath435 . in what follows
we prove our main theorems in section @xmath637 using recurrence relations for the connection coefficients .
* case @xmath1 * : we determine the connection coefficients @xmath638 and @xmath639 for the canonical solution matrix @xmath640 of the okubo system @xmath121 in the canonical form . recall that the okubo system of type @xmath121 is constructed inductively by the katz operation .
hence , by proposition [ prop : mcadd ] the canonical solution matrix @xmath640 of type @xmath132 is obtained inductively in the form @xmath641 where @xmath642 .
also , by theorem [ connection ] we obtain the recurrence relations for the connection coefficients @xmath643 and @xmath644 @xmath645 as follows : @xmath646 therefore , using these relations we can completely determine @xmath647 and @xmath648 in terms of the gamma function , if we know the initial data @xmath649 and @xmath650 .
the other connection coefficients @xmath651 and @xmath652 are obtained from @xmath653 and @xmath648 by using symmetry of the okubo system .
let @xmath654 be the operation which exchanges @xmath655 and @xmath656 @xmath657 .
then the canonical form of the okubo system given by theorem [ eq : ci ] satisfies @xmath658 where @xmath659 is the permutation matrix corresponding the transposition @xmath660 of matrix indices . for the canonical solution matrix @xmath661 ,
the solution matrix @xmath662 is also canonical solution matrix .
therefore we obtain the other connection coefficients as @xmath663 the connection coefficients @xmath649 and @xmath650 for the okubo system of type @xmath664 are computed as follows .
firstly , the equation and the solution of the rank@xmath580 case are given by @xmath665 then the canonical form of okubo system @xmath664 is given as a conjugation of .
the monodromy matrices @xmath666 are given by @xmath667 then the coefficients @xmath668 and @xmath669 are computed as @xmath670 from @xmath668and @xmath669 , we obtain the connection coefficients @xmath649 and @xmath650 : @xmath671 therefore we can obtain @xmath653 from @xmath649as follows : @xmath672 the connection coefficient @xmath648 is obtained in the same way .
furthermore , exchanging @xmath673 , @xmath655 , we obtain @xmath651 and @xmath652 .
this completes the proof of theorem [ maini ] .
* case @xmath9 * : we consider the schlesinger system of rank @xmath580 and its solution defined by @xmath674 then the okubo system of type @xmath122 is constructed by @xmath675 , and its canonical solution matrix is expressed in the form @xmath676 , where @xmath677 . the components of the diagonal matrix @xmath678 are computed as @xmath679 therefore the connection coefficients @xmath187 are obtained as follows : @xmath680 & { \displaystyle}e(\tfrac{-\rho_1}{2 } ) \frac{\prod_{k\neq i,1\leq k\leq n}(t_i - t_k)^{\alpha_k-\rho_1 } } { \prod_{k\neq j,1\leq k\leq n}(t_j - t_k)^{\alpha_k-\rho_1 } } \frac{\gamma(-\alpha_i)\gamma(\alpha_j+1 ) } { \gamma(\alpha_j-\rho_1)\gamma(1+\rho_1-\alpha_i ) } \quad(i > j ) \end{cases } \end{split}\ ] ] this completes the proof of theorem [ maini * ] . * cases @xmath2 and @xmath3 * : keeping the notation of section 2 , we consider the okubo system of type @xmath123 or @xmath124 , and set @xmath681 or @xmath682 respectively .
we denote by @xmath683 the canonical solution matrix of rank @xmath684 . taking account of the operation ,
the canonical solution matrix is constructed as follows [ th : sol23 ] @xmath612 let @xmath685 be the canonical solution matrix of type @xmath623 . then the solution matrix @xmath686 is the canonical solution matrix of type @xmath123,where @xmath687 .
@xmath616 let @xmath688 be the canonical solution matrix of type @xmath123 .
then the solution matrix @xmath689 is the canonical solution matrix of type @xmath124 , where @xmath642 .
we determine the connection coefficients @xmath690 and @xmath691 for @xmath692 . from theorem
[ connection ] , we obtain the recurrence relations @xmath693 @xmath694 therefore using these relations , we can completely determine @xmath695 and @xmath696 in terms of the gamma function , if we know the initial data @xmath697 and @xmath698 .
the other connection coefficients @xmath699 and @xmath700 are obtained from @xmath695 and @xmath696 by exchanging @xmath338 , @xmath701 and @xmath702 , @xmath703 .
the connection coefficients @xmath697 and @xmath698 for the okubo system of type @xmath433 are given as follows .
we remark that we use characteristic exponent @xmath704 instead of @xmath705 .
@xmath706 therefore we can obtain @xmath695 from @xmath697as follows : @xmath707 the connection coefficients @xmath708,@xmath709 , and @xmath710 are obtained in the same way .
furthermore , exchanging @xmath711 , @xmath712 and @xmath713 , @xmath714 in @xmath695 , @xmath696 , we obtain @xmath699 and @xmath700 .
this completes the proof of theorem [ mainii , iii ] .
the author is grateful to professors toshio oshima , masatoshi noumi and yoshishige haraoka for their valuable comments and suggestions .
50 y. haraoka : canonical forms of differential equations free from accessory parameters siam * 25 * ( 1994),12031226 y. haraoka : monodromy representations of systems of differential equations free from accessory parameters siam * 25 * ( 1994),15951621 n.m .
katz : _ rigid local systems _ , annals of mathematics studies * 139 * , princeton university press , princeton , 1996
. m. kohno : _ global analysis in linear differential equations _ , mathematics and its applications , 471 .
kluwer academic publishers , 1999 .
xvi+526 pp .
k. okubo : _ on the group of fuchsian equations _ ,
seminar reports of tokyo metropolitan university , 1987 .
k. okubo , k. takano and s. yosida : a connection problem for the generalized hypergeometric equation , funkcial . ekvac .
* 31 * ( 1988 ) , 483295 .
m. dettweiler and s. reiter : an algorithm of katz and its application to the inverse galois problem , j. symbolic comput .
* 30 * ( 2000 ) , 761798 . m. dettweiler and s. reiter : middle convolution of fuchsian systems and the construction of rigid differential systems , j. algebra * 318 * ( 2007 ) , 124 .
t. oshima : fractional calculus of weyl algebra and fuchsian differential equations , msj memoirs * 28 * , mathematical society of japan , 2012 .
t. oshima : katz s middle convolution and yokoyama s extending operation , opuscula math . * 35 * ( 2015 ) , 665688 d. yamakawa : middle convolution and harnad duality , math
* 349 * ( 2011 ) , 215262 .
t. yokoyama : on an irreducibility condition for hypergeometric systems , funkcial .
* 38 * ( 1995 ) , 1119 t. yokoyama : construction of systems of differential equations of okubo normal form with rigid monodromy , math . nachr .
* 279 * ( 2006),no .
3 327348 t. yokoyama : recursive calculation of connection formulas for systems of differential equations of okubo normal form , j. dyn control syst . *
20 * ( 2014 ) , 241292 | explicit connection coefficients and monodromy representations are constructed for the canonical solution matrices of a class of okubo systems of ordinary differential equations as an application of the katz operations . |
recently the clas collaboration at jlab has published new data on @xmath3- and @xmath4-meson photoproduction at large momentum transfers @xcite .
the main goal of these data is to provide understanding for the mechanism responsible for the violation of the ozi rule in strong interactions @xcite .
the usual approach to the description of exclusive reactions consists in the use of some effective non - perturbative reaction model at small momentum transfers and hard perturbative gluonic exchange mechanisms at large momentum transfers @xcite .
these approaches are unable to explain the large spin effects at large @xmath5 in exclusive reactions @xcite .
one example is the large @xmath6 asymmetry in proton - proton elastic scattering which can be understood only if some helicity violating contribution is present at such momentum transfers .
the pqcd exchanges in these approaches lead only to quark helicity conserving contributions and therefore can not be the dominant mechanisms for exclusive reactions in the few gev region .
we have proposed a non - perturbative mechanism @xcite for exclusive reactions at large momentum transfers , whose relevance we have studied in @xmath4 meson electromagnetic production at large momentum transfer , which we here generalize to @xmath3- and @xmath7-meson photoproduction off the nucleon .
the new ingredient of our model is the gluonic contribution arising from the qcd mechanism associated with the flavor singlet axial anomaly and describable in terms of an additional gluonic pole in the amplitudes .
the gluonic pole was introduced in @xcite to describe features of ozi violation in the pseudoscalar meson nonet .
it has been shown to provide a natural explanation of proton spin @xcite .
recently , applications to radiative decays of pseudoscalar mesons , the muon anomalous magnetic moment and the determination of the photon structure @xmath8 function in polarized deep inelastic scattering have been discussed @xcite .
the origin for this pole is the periodicity of qcd the potential as a function of the topological charge and the existence of instantons tunneling between the various classical vacua of the theory @xcite . by showing the relevance of the gluonic pole exchange in vector meson photoproduction at large momentum transfers ,
we unveil a non - perturbative qcd effect , which is neither contained in the low energy effective theories nor in the pqcd ingredients of previous calculations .
it is very well known that at large energy and small momentum transfer the main contribution to the photoproduction of the vector mesons comes from the pomeron exchange . in the low energy region
one can expect the dominance of the pomeron at small @xmath5 only for @xmath4 meson electromagnetic photoproduction due to ozi rule . for @xmath3 and @xmath7-meson production
the contribution from secondary reggeon exchanges might be important . in the case of photoproduction the leading regge trajectory with appropriate quantum numbers
is the @xmath9 meson trajectory .
subleading regge exchanges , which might give a significant contribution , are @xmath10-meson exchange for @xmath3- and @xmath7-meson production and @xmath11-meson exchange for @xmath4-meson production @xcite ( fig.1a ) .
we use the donnachie - landshoff ( dl ) model @xcite to describe the soft pomeron contribution to the vector meson production differential cross section , @xmath12 where @xmath13 is the electromagnetic nucleon form factor .
the pomeron - quark couplings @xmath14 gev@xmath15 , @xmath16 gev@xmath15 are obtained from a fit to the total @xmath17 , @xmath18 and @xmath19 cross sections @xcite , @xmath20 , @xmath21 and the pomeron trajectory is @xmath22 , with @xmath23 and @xmath24 gev@xmath25 .
the effect of the @xmath9 trajectory can be taken into account by multiplying the equation for the pomeron contribution eq.([cross3 ] ) by the factor @xmath26 with @xmath27 where the @xmath9 trajectory is @xmath28 , and the coupling to quarks @xmath29 was taken also from the dl fit to the total hadron cross sections @xcite . the contribution of the pseudoscalar exchange to the cross section can be calculated by using the same procedure as in the dl approach to the pomeron contribution @xcite - meson with the @xmath30 and @xmath31 quarks because its contribution to @xmath3 and @xmath7-meson production is much smaller than the contribution of the @xmath10-meson . ]
direct calculation of the triangle diagram in fig.2 @xcite since one has large uncertainties in the estimates due to the poor knowledge of the @xmath32 form - factors .
] at the quark level with the quark - meson interactions @xmath33 gives the result @xmath34 where @xmath35 , @xmath36 , @xmath37 and we have used the constituent quark relations @xmath38 , @xmath39 and @xmath40 . for numerical estimates the value @xmath41
@xcite for pion - nucleon and the @xmath42 value @xmath43 for @xmath11-nucleon coupling @xcite have been used . in eq.([ps ] ) the pseudoscalar - nucleon form - factor is taken of monopole form @xmath44 with @xmath45 @xcite .
the reggeization of the pseudoscalar meson exchange can be perform in the standard way ( see for example @xcite ) .
it is done by substituting the pseudoscalar propagator in the following way , @xmath46 the pseudoscalar trajectories are taken as @xmath47 with slope @xmath48 .
it can be shown that the effect of reggeization leads to the multiplication of eq.([ps ] ) by the regge factor @xmath49 the pomeron contribution to vector meson photoproduction is shown in figs . 3 , 4 and 5 by long - dashed lines .
it underestimates the low @xmath5 cross - section for @xmath3 and @xmath7 production at jlab energies but describes rather well the low @xmath5 @xmath4-meson production .
the total contribution of the dl pomeron and @xmath9 exchange to the cross section for @xmath3 and @xmath7 photoproduction is presented in figs . 3 and 4 by dashed - dot lines . the contribution of the pion exchange in figs . 3 and 4 and @xmath11-meson exchange in fig .
5 is shown by short - dashed line .
it is apparent that pseudoscalar exchange is important only for @xmath7 production at very small momentum transfers .
our result is in contradiction with the result of ref .
@xcite where a nonreggeized pion exchange contribution to vector meson production was calculated . in our approach the suppression of the pseudoscalar contribution at large @xmath5 is not very sensitive to the @xmath5 dependence of the @xmath50 and @xmath51 form - factors and the absolute value of the pseudoscalar coupling with quarks .
this suppression is due to the smallness of the regge factor @xmath52 in the cross - section for large @xmath5 .
the effect of the reggeization for the case of the @xmath10 contribution to @xmath3 production is shown in fig .
this effect is very large and can not be neglected .
the pomeron and regge contributions describe the data for all mesons rather well in the small @xmath53 gev@xmath54 region .
however , for large @xmath55 gev@xmath54 , all these contributions deviate strongly from the data .
in our previous work @xcite a new gluonic , so - called @xmath56-pole , exchange was introduced to explain the large @xmath5 electromagnetic @xmath4 meson production at jlab .
this exchange , not related to the exchange of any mesonic state , arises from the nontrivial topological structure of the qcd vacuum . in some sense
this exchange represents the interaction of the quarks in the photon with the quarks in the proton generated by the excitations of the qcd vacuum . in qcd the anomalous gluonic
axial current @xmath57 has a massless pole @xmath58 which is related to the topological susceptibility of the qcd vacuum @xmath59 where @xmath60 is the density of topological charge .
this pole has been called the gluonic ghost pole @xcite in the literature and @xmath56-pole by us @xcite .
the diagonalization of the propagator matrix for the ghost operator @xmath61 and quark nonet states @xmath62 and @xmath63 leads to the following propagator @xmath64 which we called @xmath56-propagator .
@xmath65 in eq.([ghost ] ) is determined by the topological susceptibility of qcd vacuum @xcite @xmath66 we should stress that the @xmath56-propagator eq.([ghost ] ) does not depend on the momentum transfer @xmath67 and therefore the effective interaction induced by the ghost exchange is _ point - like_. this property is responsible for the large @xmath56-pole contribution to vector meson photoproduction at large momentum transfers as will be shown next . let us estimate the contribution of @xmath56-pole exchange to vector meson photoproduction ( fig.1b ) . to calculate this diagram the values of couplings @xmath68 and @xmath69 are needed .
we will use the fact that the contribution of the @xmath56-pole to the physical amplitudes in the flavor singlet channel leads to results which are different from the predictions of the ozi rule .
this mechanism produces , for the isosinglet axial - vector nucleon form factor at zero momentum transfer , the following generalized @xmath70 golberger - treiman relation @xcite @xmath71 where @xmath72 , @xmath73 mev and @xmath68 is @xmath56-nucleon coupling constant .
the experimental value of @xmath74 extracted from experimental data on nucleon spin - dependent structure function @xmath75 @xcite , allows us to estimate the value of @xmath56-nucleon coupling constant as @xmath76 the @xmath56-pole contribution to the @xmath77 decay has been discussed in @xcite where a modified formula for the effective coupling of the @xmath63 meson with photons has been obtained , @xmath78 let us consider the @xmath56-pole contribution to the @xmath79 coupling , where @xmath80 stands for @xmath81 , @xmath7 and @xmath4 .
the interaction of the vector mesons with the quarks is assumed to be photon - like @xmath82 due to this vector meson - photon analogy the generalization of eq.([ggg ] ) to the case of the @xmath83 amplitude is straightforward @xmath84 where @xmath85 , @xmath86 and @xmath87 .
the values of @xmath88 , @xmath89 and @xmath90 can be extracted from the experimental widths @xmath91 , @xmath92 and @xmath93 . the vector meson couplings with quarks are not well known . in order to estimate them we use a constituent quark model calculation for the @xmath3 and @xmath7 couplings
@xmath94 where @xmath95 is taken from @xcite and for the @xmath4 coupling we use the njl model prediction @xcite @xmath96 our final estimate for the @xmath56-vector meson couplings is @xmath97 gev@xmath98 , @xmath99 @xmath100 and @xmath101 @xmath100 .
we are now ready to calculate @xmath56-pole contribution to vector meson photoproduction .
the absolute value of the @xmath56-contribution strongly depends on its couplings with the nucleon , photon and vector meson and is given by @xmath102 where @xmath103 is the flavor singlet axial form factor of the nucleon with the value of @xmath104 equals to the mass of the flavor singlet @xmath105-meson @xcite .
it should be noted that the @xmath56-exchange induces a nucleon spin - flip , which produces a factor @xmath106 in eq.([gcross ] ) , and leads to an additional enhancement of the @xmath56-pole contribution at large @xmath5 , with respect to the pomeron contribution , which is nonspin - flip eq.([cross3 ] ) .
moreover the energy dependence of the @xmath56 contribution corresponds to that of a fixed pole with zero regge slope .
therefore the large @xmath106 regge suppression in vector meson production given by @xmath107 with the slope @xmath108
gev@xmath109 for the usual regge trajectories , e.g. @xmath110 and @xmath111 , is absent for the @xmath56-exchange .
the result of the calculation of the @xmath56-pole contribution is presented in fig . 3 , 4 and 5 by the dotted lines .
the sum of the pomeron , @xmath9 , @xmath110 , @xmath11 and @xmath56-pole contributions are shown by the solid lines .
it is evident that @xmath56-exchange determines the behavior of the cross section at large @xmath5 and our model based only on pomeron , secondary regge exchange and @xmath56-pole contributions reproduces the main features of a data .
the deviation from data at intermediate values of @xmath5 might be related to additional contributions from the pomeron and regge cuts or to non - linear @xmath106 dependences of pomeron and @xmath9 trajectories@xcite .
the gluonic degrees of freedom play a very important role in the @xmath3- @xmath7- and @xmath4-meson photoproduction at jlab energies . at small @xmath5 the cross - section
is described rather well by the donnachie - landshoff soft pomeron and secondary regge exchanges .
at large @xmath5 an extremely interesting phenomena , related to the complex structure of qcd vacuum , takes place .
we have shown that at large momentum transfers , the point - like interaction induced by the @xmath56- pole exchange , related to ghost pole in the correlator of the anomalous gluonic axial currents , gives the dominant contribution .
we should stress that vector meson photoproduction is only one of the possible exclusive processes where the @xmath56-pole contributes .
we plan to extend our considerations here to the study of other reactions at large momentum transfers .
as has been mentioned the g - pole exchange leads to a nucleon helicity flip and therefore can be separated from the helicity conserving pqcd hard gluonic exchanges in processes with polarized particles .
recently we have shown that at large energy and large momentum transfer a new anomalous trajectory with very large intercept @xmath112 and very small slope @xmath113 called @xmath114 is needed to explain elastic @xmath17 , @xmath115 and vector meson production at hera @xcite .
it would be interesting to find a relation between this @xmath114 trajectory and the @xmath116 exchange similarly as was discussed recently for the relation between pomeron and s - channel multi - glueball @xmath117 exchange @xcite .
we are grateful to s.b.gerasimov , w .- d.nowak and v.l.yudichev for useful discussion .
we would like to thank m. battaglieri and r. schumacher for providing us with the clas data .
this work was partially supported by dgicyt pb97 - 1227 , rfbr-01 - 02 - 16431 , intas-00 - 00366 grants and by the heisenberg - landau program . 99 clas collaboration , m.battaglieri et al .
* 87 * ( 2001 ) 172002 .
clas collaboration , e.anciant et al .
* 85 * ( 2000 ) 4682 .
a.i.titov , t .- s.h.lee , h.toki , and o.strelsova phys.rev . *
c60 * ( 1999 ) 035205 ; + y.oh , a.i.titov , t .- s.h.lee , phys.rev . *
c63 * ( 2000 ) 025201 .
r.williams , phys.rev . *
c57 * ( 1998 ) 223 .
q.zhao , b.saghai and j.s.al-khalili , nucl - th/0004055 .
laget , hep - ph/0107208 ; + phys . lett .
* b489 * ( 2000 ) 313 .
krisch , proc .
9th int . symp . on high energy spin physics , bonn , ( 1990 ) 20
; + g. bunce et al .
part . world*3 * ( 1992 ) 1 .
n.i.kochelev and v.vento , phys.lett .
* b515 * ( 2001 ) 375 .
shore , hep - ph/0111165 .
g.veneziano , nucl . phys .
* b159 * ( 1979 ) 213 .
g.veneziano , mod .
phys . lett . * a4 * ( 1989 ) 1605 .
d.i.diakonov and m.i.eides , sov.phys .
jetp * 54 * ( 1981 ) 2 .
a.donnachie and p.v.landshoff , phys . lett . *
b185 * ( 1987 ) 403 .
a.donnachie and p.v.landshoff , phys . lett .
* b296 * ( 1992 ) 227 .
v.stoks , r.timmermans , and j.j.de swart , phys.rev .
* c47 * ( 1993 ) 512 .
w.koepf , l.l.frankfurt , and m.strikman , phys.rev .
* d53 * ( 1996 ) 2586 .
m.guidal , j .- m.laget and m.vanderharghen , nucl . phys . * a627 * ( 1997 ) 645 .
d.p.barber et .
al . z.phys . *
c26 * ( 1984 ) 343 . g.m.shore and g.veneziano , nucl
. phys . * b381 * ( 1992 ) 3 .
m.anselmino , a.efremov and e.leader , phys.rep .
* 261 * ( 1995 ) 1 .
d.ebert , m.k.volkov and v.l.yudichev , j.phys . *
g25 * ( 1999 ) 2025 .
v.a.khoze , a.d.martin , m.g .
ryskin , eur.phys.j . *
c18 * ( 2000 ) 167 .
n.i.kochelev,d.-p.min , y.oh,v.vento and a.v.vinnikov , phys.rev.*d61 * ( 2000 ) 094008 ; nucl.phys.proc.suppl.*99a * , ( 2001 ) 24 ; + y.oh , n.i.kochelev , d .-
p.min , v.vento and a.v.vinnikov , phys.rev . * d62 * ( 2000 ) 017504 .
shuryak , phys.lett . * b486 * ( 2000 ) 378 . | non - perturbative qcd mechanisms are of fundamental importance in strong interaction physics . in particular
, the flavor singlet axial anomaly leads to a gluonic pole mechanism which has been shown to explain the @xmath0 mass , violations of the ozi rule and more recently the proton spin .
we show here that the interaction derived from the gluonic pole exchange explains the high momentum transfer behavior of the photoproduction cross sections of vector mesons at jlab energies . * gluonic effects in vector meson photoproduction at large momentum transfers * n.i .
kochelev@xmath1 and v. vento@xmath2 . |
the top quark is the heaviest fermion of the standard model .
its large mass allows to probe deeply into the qcd potential for nonrelativistic @xmath0 system produced near energy threshold .
such a system will provide a unique opportunity for a variety of novel qcd studies .
the lifetime of the top quark is shorter than the formation time of top mesons and toponium resonances .
therefore top decays intercept the process of hadronization at an early stage and practically eliminate associated nonperturbative effects .
the analysis of polarized top quarks and their decays has recently attracted considerable attention , see @xcite and references cited therein .
the reason is that this analysis will result in determination of the top quark coupling to the @xmath2 and @xmath3 bosons either confirming the predictions of the standard model or providing clues for physics beyond .
the latter possibility is particularly intriguing because @xmath4 plays an exceptional role in the fermion mass spectrum .
the polarization fourvector @xmath5 of the top quark can be determined from the angular - energy distributions of the charged leptons in semileptonic @xmath6 decays . in the @xmath6 quark rest frame
this distribution is in born approximation the product of the energy and the angular distributions@xcite : @xmath7\,{{\rm d}\gamma\over{\rm d}e_\ell } \label{eq : elec1}\ ] ] where @xmath8 , @xmath9 and @xmath10 is the angle between @xmath11 and the direction of the charged lepton .
qcd corrections essentially do not spoil factorization@xcite .
thus , the polarization analyzing power of the charged lepton energy - angular distribution remains maximal .
there is no factorization for the neutrino energy - angular distribution which is therefore less sensitive to the polarization of the decaying top quark . on the other hand
it has been shown @xcite that the angular - energy distribution of neutrinos from the polarized top quark decay will allow for a particularly sensitive test of the v - a structure of the weak charged current .
a number of mechanisms has been suggested that will lead to polarized top quarks .
however , studies at a linear electron - positron collider are particularly clean for precision tests .
moreover , close to threshold and with longitudinally polarized electrons one can study decays of polarized top quarks under particularly convenient conditions : large event rates , well identified rest frame of the top quark , and large degree of polarization . at the same time , thanks to the spectacular success of the polarization program at slc @xcite , the longitudinal polarization of the electron beam will be an obvious option for a future linear collider has been reviewed in @xcite and the top quark polarization has been recently considered in @xcite . ] . in this article some results are presented of a recent calculation @xcite of top quark polarization for the reaction @xmath12 in the threshold region . in sect.2
we discuss the dependence of the top quark polarization on the longitudinal polarizations of the beams . due to restricted phase space
the amplitude is dominantly @xmath13 wave and the electron and positron polarizations are directly transferred to the top quark . for a quantitative study
this simple picture has to be extended and the modifications originating from @xmath14 wave interference should be taken into account . in sect.3
all these corrections are calculated from numerical solutions of lippmann - schwinger equations .
we adopt the conventions of ref.@xcite and describe the longitudinal polarization of the @xmath1 system in its center - of - mass frame as a function of the variable @xmath15 where @xmath16 denote the polarizations of @xmath17 with respect to the directions of @xmath18 and @xmath19 beams , respectively collider @xmath20 , @xmath21 and then @xmath22 . ] . in the absence of phases from final state interaction , which can be induced by higher orders in @xmath23 and will be considered elsewhere @xcite , the top quark polarization is in the production plane . @xmath24 and @xmath25 denote the longitudinal and the transverse components of top polarization vector ( in its rest frame ) with respect to the electron beam .
the angle @xmath26 denotes the angle between @xmath19 and the top quark . in the threshold region
the top quark is nonrelativistic ( with velocity @xmath27 ) and the kinetic energy of the @xmath0 system @xmath28 is of the order @xmath29 . retaining only the terms up to @xmath30 one derives the following expressions for the components of the polarization vector : @xmath31 -0.5 cm = 5.5truein -0.2 cm the coefficients
@xmath32 , @xmath33 and @xmath34 depend on the polarization @xmath35 , the electroweak coupling constants , the @xmath3 mass and the center - of - mass energy @xmath36 .
they are plotted in fig.[fig - clchi ] for @xmath37 gev . @xmath32 and @xmath33 are shown in fig.[fig - clchi]a as the solid and the dashed lines , respectively , and @xmath34 as the solid line in fig.[fig - clchi]b .
the function @xmath38 describes the complicated dynamics of the @xmath0 system near threshold .
in particular it includes effects of the would - be toponium resonances and coulomb enhancement .
nevertheless , it is possible to calculate this function using the green function method .
the same function @xmath38 also governs the forward - backward asymmetry in @xmath12 @xmath39 where @xmath40 is shown as the dashed line in fig.[fig - clchi]b .
eqs.([eq:2 ] ) and ( [ eq:3 ] ) extend the results of @xcite into the threshold region .
the green function method has become a standard tool for studying @xmath1 annihilation in the threshold region @xcite .
we follow the momentum space approach of @xcite and solve the lippmann - schwinger equations numerically for the @xmath13-wave and @xmath41-wave green functions @xmath42 where @xmath43 is the momentum of the top quark in @xmath0 rest frame , @xmath44 @xmath45 denotes the top width and @xmath46 is the qcd potential in momentum space ; see @xcite for details .
the function @xmath38 is related to @xmath47 and @xmath48 : @xmath49 where @xmath50 has been introduced in order to cut off a logarithmic divergence of the numerator .
the denominator remains finite for @xmath51 . in experimental analyses
the contributions of very large intrinsic momenta will be automatically suppressed by separation of @xmath0 events from the background . in our calculation
we use @xmath52 . the function @xmath38 is plotted in fig.[fig - fir]a for @xmath37 gev .
the qcd potential depends on @xmath53 and our results have been obtained for @xmath54 . for a comparison in fig.[fig - fir]b the annihilation cross section @xmath55 is shown in units of the annihilation cross section @xmath56 .
= 4.5truein -0.5 cm -0.8 cm
mj would like to thank the stefan batory foundation for a travel grant and the local organizing committee for an additional support which enabled his participation in spin94 .
this work was supported in part by kbn grant 2p30225206 and by dfg contract 436pol173193s .
khn , `` top quark at a linear collider '' , in _ physics and experiments with linear @xmath1 colliders _ ,
harris et al . , ( world scientific , singapore , 1993 ) , p.72 .
m. jeabek , _ top quark physics _ , in proceedings of zeuthen workshop
_ physics at lep 200 and beyond _ , to appear in nucl .
b(1994 ) suppl . ;
karlsruhe preprint ttp94 - 09
. m. jeabek and j.h .
khn , nucl .
phys . b320 ( 1989 ) 20 .
a. czarnecki , m. jeabek and j.h .
khn , nucl .
b351 ( 1991 ) 70 ; + a. czarnecki and m. jeabek , nucl . phys .
b427 ( 1994 ) 3 .
m. jeabek and j.h .
khn , phys .
b329 ( 1994 ) 317 .
m. woods , _ polarization at slac _
, in these proceedings .
p.m. zerwas ( ed . ) , @xmath1 collisions at 500 gev : the physics potential , desy orange reports desy 92 - 123a , desy 92 - 123b and desy 93 - 123c .
fadin , v.a .
khoze and m.i .
kotsky , z. phys .
c64 ( 1994 ) 45 .
r. harlander , m. jeabek , j.h .
khn and t. teubner , to be published .
g. alexander et al .
( eds . ) , _ polarization at lep _ ,
cern 88 - 06 , geneva 1988 .
khn , a. reiter and p.m. zerwas , nucl .
b272 ( 1986 ) 560 .
fadin and v.a .
khoze , jetp lett .
46 ( 1987 ) 525 ; sov .
48 ( 1988 ) 309 .
strassler and m.e .
peskin , phys .
rev d43 ( 1991 ) 1500 .
y. sumino , k. fujii , k. hagiwara , h. murayama and c .- k .
ng , phys .
d47 ( 1993 ) 56 ; h. murayama and y. sumino , phys .
d47 ( 1993 ) 82 .
m. jeabek , j.h .
khn and t. teubner , z. phys . c56 ( 1992 ) 653 ; + m. jeabek and t. teubner , z. phys .
c59 ( 1993 ) 669 . | theoretical results are presented for top quarks produced in annihilation of polarized electrons on positrons .
polarization studies for @xmath0 pairs near threshold are free from hadronization ambiguities .
this is due to the short lifetime of the top quark .
semileptonic decays are discussed as well as their applications in studying polarization dependent processes involving top quarks .
the green function formalism is applied to @xmath0 production at future @xmath1 colliders with polarized beams .
lippmann
schwinger equation is solved numerically for the qcd chromostatic potential given by the two - loop formula at large momentum transfers and richardson ansatz at intermediate and small ones .
the polarization dependent momentum distributions of top quarks and their decay products are calculated .
-1.5 in = 5.5truein = 8.5 in * ttp94 - 24 * + *
november 1994 * |
the progenitors of type ia supernovae ( sne ia ) remain unknown , though the idea that sn ia progenitors belong to at least two distinct populations ( see e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) has recently been gaining ground .
a picture is emerging which supports populations of both quickly - evolving ( prompt ) progenitors with short delay times less than @xmath7 myr , as well as more slowly - evolving progenitors with ( sometimes rather ) long delay times spanning up to a hubble time .
the most favoured sn ia progenitor scenarios involve the double degenerate scenario ( dds ; @xcite ) , and the single degenerate scenario ( sds ; @xcite ) . in the dds ,
the merger of two carbon - oxygen ( co ) white dwarfs ( wds ) with a total mass exceeding the chandrasekhar mass limit , @xmath8 m@xmath9 , can lead to explosive carbon - burning which causes a sn ia explosion . in the sds ,
a co wd accretes from a non - degenerate stellar companion via stable roche - lobe overflow ( rlof ) , enabling the wd to accumulate mass toward @xmath10 until carbon is ignited explosively in the centre leading to a sn ia .
the companion filling its roche - lobe can be hydrogen - rich or helium - rich ; in this work we delineate between sds ( hydrogen - rich donors ) and helium - rich donor cases , as in the latter case the donor may be a non - degenerate helium - burning star , or a degenerate helium - rich wd .
for some time , population synthesis calculations ( e.g. , * ? ? ?
* ) have predicted that the number of merging co wds ( dds ) with a total mass exceeding @xmath10 is sufficient to match , and thus possibly account for , the rate of sne ia ( @xmath11 per century for the galaxy , * ? ? ?
at the same time , despite the fact that potential sds progenitors have been observed @xcite , the theoretically - predicted sn ia rate from the sds channel is usually unable to explain the observed rates of sne . in the majority of population studies , the relative frequency of sds events
is often found to be well below those of the dds ( * ? ? ?
* ; * ? ? ?
* see also @xcite ) being about an order of magnitude too low compared to observations of sne ia .
detailed wd merger simulations have shown that some wd mergers can successfully lead to a sn ia explosion @xcite .
however , the main argument against the dds is that the merging process leads to physical conditions in which a thermonuclear explosion is unlikely @xcite .
it is more probable that a merger between two co wds with a total mass @xmath12 will collapse and form a neutron star ; an accretion induced collapse ( aic , * ? ? ?
such events could be detectable in modern transient surveys , though would have an observational signature unlike those of sne ia @xcite .
recently , @xcite carried out a population synthesis study showing rates and delay times for sne ia involving wds whose mass at the time of sn ia had reached the chandrasekhar mass limit : dds , sds and helium - rich donor scenarios ( ` am cvn ' channel in that work ) .
we extend our investigation of progenitors and focus on the sub - chandrasekhar mass model ( detailed results are presented in ruiter et al .
2010 , in preparation ) . since these calculations are based on the work which was performed for @xcite , the reader is referred to that paper for a more detailed description of the dds , sds and helium - rich ( chandrasekhar mass wd ) donor scenario
. sub - chandrasekhar sne ia
sne ia which take place in wds whose mass is below @xmath10 likely consist of a ( probably co ) wd accreting via stable rlof from a helium - rich companion @xcite .
these systems have thus far been regarded as an unlikely model for sne ia owing to the fact that most synthetic light curves and spectra of these objects from previous studies did not match those of observations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , likely owing to the thickness of the helium shell ( @xmath13 ) .
more recently , @xcite have shown that conditions suitable for a detonation in the wd might be achieved for even lower helium shell masses than were assumed in previous studies .
double detonations in sub - chandrasekhar mass wds with low - mass helium shells have shown to be a robust explosion mechanism that can produce sne ia of normal brightness , provided that a detonation in the helium shell is successfully triggered @xcite .
recent results involving hydrostatic sub - chandrasekhar mass exploding wds with subsequent nucleosynthesis and radiative transfer calculations @xcite indicate that the sub - chandrasekhar mass scenario should be considered as a likely sn ia progenitor candidate .
additionally , population synthesis studies @xcite predict that there are a sufficient number of these binaries to explain the observed rate of sne ia .
we use the startrack population synthesis binary evolution code @xcite to evolve our stellar population .
the initial distributions for binary orbital parameters ( orbital periods , mass ratios , etc . ) are the same as described in @xcite . in @xcite
, it was assumed that the ejection of the envelope of the mass - losing star during a ce phase came at the expense of removing the orbital energy of the binary , as dictated by the ` energy - balance ' equation ( or ` @xmath14-formalism ' ) @xcite @xmath15 where @xmath16 is the gravitational constant , @xmath17 is the initial mass of the ( giant ) donor star just prior to the ce , @xmath18 is the ejected mass ( assumed to be the mass of the giant s envelope ) , @xmath19 is the mass of the companion ( assumed to be unchanged during the ce ) , @xmath20 is the final mass of the donor once the envelope has been ejected , @xmath21 is the initial radius of the donor star when it fills its roche - lobe , @xmath22 is the initial orbital separation , @xmath23 is the final orbital separation ( if @xmath23 is too small to accommodate the roche - lobes of the stars , the ce results in a merger ) , @xmath24 is the efficiency with which the binary orbital energy can unbind the ce , and @xmath25 is a parameterization of the structure of the donor star ; both @xmath24 and @xmath25 are fairly uncertain .
here we employ @xmath0 and @xmath1 . in an upcoming paper
we investigate the impact of the ce parameterization on sn ia delay times and rates ( ruiter et al . 2010 , in preparation ) .
it has been shown previously that a wd accumulating helium - rich material may be capable of exploding in a type ia supernova if the correct conditions are satisfied ; even if the wd is below the chandrasekhar mass limit @xcite .
we adopt the prescription of @xcite , applied to accretion from helium - rich companions only , to determine when a particular binary undergoes a sub - chandrasekhar mass sn ia ( * ? ? ?
* see section 5.7.2 for equations ) .
in short , we consider three different accretion rate regimes for accumulation of helium - rich material on co wds , adopting the input physics of @xcite . at high ( e.g. , initial rlof ) accretion rates ,
helium burning is stable and thus mass accumulation on the wd is fully efficient ( @xmath26 ) . at somewhat lower accretion rates ,
helium burning is unstable and the binary enters a helium - flash cycle , thus accumulation is possible but is not fully efficient ( @xmath27 ) . in both of these aforementioned accretion regimes ,
the co wd is allowed to accrete ( and burn ) helium , and its total mass may reach the chandrasekhar mass limit within a hubble time and explode as a type ia supernova through the helium - rich donor channel . however , for low accretion rates compressional heating at the base of the accreted helium layer plays no significant role , and a layer of unburned helium can be accumulated on the wd surface . following @xcite
, we assume that if such a co wd accumulating helium enters this ` low ' accretion rate regime and accumulates @xmath28 of helium on its surface , a detonation is initiated at the base of the helium shell layer .
consequently , a detonation in the core of the co wd is presumed to follow , and we assume that a sub - chandrasekhar type ia supernova takes place . in our model , only accreting wds with a _ total mass _ @xmath29 are considered to lead to potential sub - chandrasekhar sne ia , since lower mass cores , while they may detonate , are unlikely to produce enough radioactive nickel and hence will not be visible as sne ia .
thus in all future discussions we refer to sub - chandrasekhar systems whose total wd mass ( co core @xmath30 helium shell ) is at least @xmath31 at the time of sn ia .
for an instantaneous burst of star formation at @xmath3 , we have calculated the dtd and rates of sne ia arising from different formation channels : the dds , sds , helium - rich donor , and the sub - chandrasekhar scenario .
we find that only two sn ia formation scenarios are capable of matching the observed sne ia rates : the dds and the sub - chandrasekhar channels .
the adopted sub - chandrasekhar scenario ( @xmath32 ) is dominant at nearly all epochs @xmath33 gyr .
as found in @xcite , the dds dtd follows a continuous power - law with the largest number of events occurring starting at @xmath34 myr .
we present for the first time the dtd of the sub - chandrasekhar mass channel , which is clearly divisible into prompt and a delayed components . in figure
[ fig : dta1a125g1 ] we show the dtd of the four progenitor channels investigated in this work .
we show the delay times in units of sne ia per year per unit stellar mass born in stars ( snum @xmath35 ; snum = sne ia per century per @xmath36 ) . for our dtd normalization ,
we have assumed a binary fraction of 50% across the entire initial stellar mass function .
we note that the bumps in the smoothed plot are due to monte carlo noise .
-axis from @xcite , the shapes of the dtds from the three chandrasekhar ( or above ) mass models are the same as in that work ( model 1 ) .
the newly - calculated sub - chandrasekhar sn ia dtd clearly shows two distinct populations : the helium star channel ( spike at delay times less than @xmath37 myr ) and the helium wd channel ( from @xmath38 myr to a hubble time ) . ] as was found in @xcite model 1 , the dds dtd follows a power - law shape with @xmath39 @xmath40 , while the sds distribution is somewhat flat with no events with delay times less than @xmath41 myr .
the reason why the sds does not harbour prompt events is directly linked to the donor star s zams mass ; all sds progenitor donors are found to have zams masses below @xmath42 . when the secondary zams mass is @xmath43 , the binary will enter a ce phase when the secondary fills its roche - lobe , rather than a stable rlof phase . in such a case , the binary will not become an sds sn ia , though may under the right circumstances evolve to sn ia from the helium - rich donor channel .
the helium - rich donor dtd consists mostly of systems with relatively short ( @xmath39100 myr @xmath44 gyr ) delay times , with very few events at long delay times .
we refer the reader to @xcite for a description of these dtds ; the overall dtd shapes remain the same as in that work .
the sub - chandrasekhar mass systems can easily by eye be grouped into two classes : those which occur with delay times less than @xmath45 myr , and those with delay times above @xmath38 myr . not surprisingly , these two classes of sne ia stem from two very different evolutionary scenarios . those with short delay times consist of progenitors which involve a helium - burning star donor , where as the rest consist of helium or hybrid ( co core , helium mantle ) wd donors .
the prompt component ( delay times @xmath4 myr ) accounts for @xmath46 % of all sub - chandrasekhar sne ia that explode within 13 gyr of star formation .
nearly all of these systems ( @xmath47% of the total sub - chandrasekhar sne ) have helium star donors , with a small fraction of the prompt component originating from the hybrid wd channel . the delay time is governed by the main sequence ( ms ) lifetime of the donor star .
the companions with zams masses @xmath48 evolve off of the ms within @xmath49 myr .
after the first ce , which leaves behind a co primary wd and a ms secondary star , the ( slightly evolved ) secondary will fill its roche - lobe and mass transfer is once again unstable leading to a second ce phase .
the ce brings the co wd and newly - formed naked helium star on a close orbit ( @xmath50 min . orbit ) . within @xmath39 a few myr the helium star fills its roche - lobe . however , initial mass transfer rates for the helium star channel fall within the ` low accretion rate regime ' : typically such systems have initial mass transfer rates @xmath51yr@xmath52 .
the delayed component ( delay times @xmath53 myr ) comprise the other @xmath54 % of the sub - chandrasekhar mass progenitors ; binaries with helium wd donors make up @xmath55 % ( the rest being hybrid wd donors ) .
these binaries also evolve through two ce phases , as is expected for the evolution of am cvn binaries .
similar to the dds , the timescale governing the dtd for the helium wd channel is primarily set by the gravitational radiation timescale .
however unlike the dds , these wds do not merge upon contact , but enter a stable phase of rlof . in table 1
, we show the sn ia rates for a spike of star formation at @xmath3 , and a binary fraction of 50% .
as was determined in @xcite , model 1 , dds rates are able to ( just ) account for the observed galactic rate of sne ia , where as both the sds and helium - rich donor channels fall short by over an order of magnitude .
we find that the rate of our adopted sub - chandrasekhar sn ia model exceeds all other progenitor channels between @xmath56 gyr , and these systems are enough to account for the galactic sn ia rate , with a calculated galactic rate of @xmath57 sn ia yr@xmath52 . for comparison ,
the dds galactic rate is @xmath58 sn ia yr@xmath52 ( see * ? ? ? * for discussion of galactic rate calculation ) .
both of these values are within the galactic rate estimate from @xcite of @xmath59 sn ia yr@xmath52 ( see also * ? ? ?
* for a modern sn rate calculation in the local universe ) .
.rates of sne ia ( snum ) following a starburst at t=0 . [ cols= " < , > , > , > , > " , ]
hydrodynamic simulations of sub - chandrasekhar mass sne ia exhibit observational features which are characteristically similar to those of sne ia @xcite .
motivated by these findings , as well as population synthesis rate estimates , we have investigated sub - chandrasekhar mass sn ia formation channels and have calculated and presented the dtds and rates of their progenitors . within the framework of our adopted model , we find that only the dds and sub - chandrasekhar mass channels can potentially explain the observed rates of sne ia . however , while the predicted rates of the dds are not in conflict with observations , these systems are theoretically expected to produce events which lead to aic .
if this is the case , the galactic aic rate from this aic evolutionary ( wd merger ) channel alone would be as high as @xmath60 per year .
an absence of dds progenitors from the sn ia dtd would leave the sub - chandrasekhar mass model as the only evolutionary scenario able to produce a high enough rate of sne ia , as well as a bimodal dtd .
however , if one can say for certain that aics which are formed from the merger of co wds produce very neutron - rich ejecta yr@xmath52 from population synthesis , though note that this rate does not include mergers of co - co wds .
a galactic rate of @xmath61 aic yr@xmath52 is the upper limit estimate derived from solar system abundances of neutron - rich isotopes , which are expected to be produced in aics @xcite . ] , then this provides a potentially strong constraint on the outcome of these mergers namely that a non - negligible fraction of sne ia must be formed through the dds channel .
@xcite find that about half of sne ia explode within 330 myr of star formation , with a significant fraction occurring within 1 gyr of star formation .
if both the dds and sub - chandrasekhar mass channels contribute to sne ia at short delay times , this could potentially explain both the bimodality in the observed dtd , as well as increase the predicted number of prompt sne ia ( see fig .
1 ) . the sub - chandrasekhar model is the first model which demonstrates both a sufficient number of sne ia events to account for all ( or at least some substantial fraction ) of sne ia , as well as _ two distinct evolutionary channels with their own characteristic delay times _ : a prompt ( @xmath62 myr ) helium - star channel originating from binaries with more massive secondaries , and a more delayed ( @xmath63 myr ) double wd channel originating from am cvn - like progenitor binaries with lower mass .
considering the recent observational studies which have found evidence for such a dtd @xcite , further theoretical investigation of the sub - chandrasekhar mass sn ia channel is now strongly warranted .
scannapieco , e. , & bildsten , l. 2005 , apjl , 629 , l85 mannucci , f. , della valle , m. , & panagia , n. 2006 , mnras , 370 , 773 maoz , d. , & badenes , c. 2010 , mnras , 968 brandt , t. d. , tojeiro , r. , aubourg , . ,
heavens , a. , jimenez , r. , & strauss , m. a. 2010 , aj , 140 , 804 webbink , r. f. 1984 , apj , 277 , 355 whelan , j. , & iben , i. j. 1973 , apj , 186 , 1007 yungelson , l. r. , livio , m. , tutukov , a. v. , & saffer , r. a. 1994 , apj , 420 , 336 cappellaro , e. , evans , r. , & turatto , m. 1999 , a&a , 351 , 459 hachisu , i. , & kato , m. 2001 , apj , 558 , 323 yungelson , l. , & livio , m. 1998 , apj , 497 , 168 ruiter , a. j. , belczynski , k. , & fryer , c. 2009 , apj , 699 , 2026 mennekens , n. , vanbeveren , d. , de greve , j. p. , & de donder , e. 2010 , a&a , 515 , a89 piersanti , l. , gagliardi , s. , iben , i. , jr . , & tornamb , a. 2003 , apj , 598 , 1229 pakmor , r. , kromer , m. , rpke , f. k. , sim , s. a. , ruiter , a. j. , & hillebrandt , w. 2010 , nature , 463 , 61 yoon , s .- c . , podsiadlowski , p. , & rosswog , s. 2007 , mnras , 380 , 933 miyaji , s. , nomoto , k. , yokoi , k. , & sugimoto , d. 1980 , pasj , 32 , 303 darbha , s. , metzger , b. d. , quataert , e. , kasen , d. , nugent , p. , & thomas , r. 2010 , mnras , 1254 iben , i. , jr .
, & tutukov , a. v. 1991 , apj , 370 , 615 hoeflich , p. , & khokhlov , a. 1996 , apj , 457 , 500 nugent , p. , baron , e. , branch , d. , fisher , a. , & hauschildt , p. h. 1997 , apj , 485 , 812 shen , k. j. , & bildsten , l. 2009 , apj , 699 , 1365 fink , m. , rpke , f. k. , hillebrandt , w. , seitenzahl , i. r. , sim , s. a. , & kromer , m. 2010 , a&ap , 514 , a53 sim , s. a. , rpke , f. k. , hillebrandt , w. , kromer , m. , pakmor , r. , fink , m. , ruiter , a. j. , & seitenzahl , i. r. 2010 , apjl , 714 , l52 kromer , m. , sim , s. a. , fink , m. , rpke , f. k. , seitenzahl , i. r. , & hillebrandt , w. 2010 , apj , 719 , 1067 belczynski , k. , kalogera , v. , rasio , f. a. , taam , r. e. , zezas , a. , bulik , t. , maccarone , t. j. , & ivanova , n. 2008 , apjs , 174 , 223 taam , r. e. 1980 , apj , 242 , 749 woosley , s. e. , & weaver , t. a. 1994 , apj , 423 , 371 ivanova , n. , & taam , r. e. 2004 , apj , 601 , 1058 kato , m. , & hachisu , i. 1999 , apjl , 513 , l41 kato , m. , & hachisu , i. 2004 , apjl , 613 , l129 li , w. , chornock , r. , leaman , j. , filippenko , a. v. , poznanski , d. , wang , x. , ganeshalingam , m. , & mannucci , f. 2010 , arxiv:1006.4613 fryer , c. , benz , w. , herant , m. , & colgate , s. a. 1999 , apj , 516 , 892 dessart , l. , burrows , a. , ott , c. d. , livne , e. , yoon , s .- c .
, & langer , n. 2006 , apj , 644 , 1063 metzger , b. d. , piro , a. l. , & quataert , e. 2009 , mnras , 396 , 1659 maoz , d. , mannucci , f. , li , w. , filippenko , a. v. , della valle , m. , & panagia , n. 2010 , arxiv:1002.3056 | using the population synthesis binary evolution code startrack , we present theoretical rates and delay times of type ia supernovae arising from various formation channels .
these channels include binaries in which the exploding white dwarf reaches the chandrasekhar mass limit ( dds , sds , and helium - rich donor scenario ) as well as the sub - chandrasekhar mass scenario , in which a white dwarf accretes from a helium - rich companion and explodes as a sn ia before reaching the chandrasekhar mass limit .
we find that using a common envelope parameterization employing energy balance with @xmath0 and @xmath1 , the supernova rates per unit mass ( born in stars ) of sub - chandrasekhar mass sne ia exceed those of all other progenitor channels at epochs @xmath2 gyr for a burst of star formation at @xmath3 .
additionally , the delay time distribution of the sub - chandrasekhar model can be divided in to two distinct evolutionary channels : the ` prompt ' helium - star channel with delay times @xmath4 myr , and the ` delayed ' double white dwarf channel with delay times @xmath5 myr spanning up to a hubble time .
these findings are in agreement with recent observationally - derived delay time distributions which predict that a large number of sne ia have delay times @xmath6 gyr , with a significant fraction having delay times @xmath4 myr .
we find that the dds channel is also able to account for the observed rates of sne ia .
however , detailed simulations of white dwarf mergers have shown that most of these mergers will not lead to sne ia but rather to the formation of a neutron star via accretion - induced collapse .
if this is true , our standard population synthesis model predicts that the only progenitor channel which can account for the rates of sne ia is the sub - chandrasekhar mass scenario , and none of the other progenitors considered can fully account for the observed rates .
address = max planck institute for astrophysics , karl - schwarzschild - str . 1 , 85741 garching , germany address = los alamos national laboratory , ccs-2/isr-1 group , p.o .
box 1663 , los alamos , nm 87545 usa ; oppenheimer fellow , altaddress = astronomical observatory , university of warsaw , al .
ujazdowskie 4 , 00 - 478 warsaw , poland address = max planck institute for astrophysics , karl - schwarzschild - str . 1 , 85741 garching , germany address = max planck institute for astrophysics , karl - schwarzschild - str .
1 , 85741 garching , germany address = max planck institute for astrophysics , karl - schwarzschild - str . 1 , 85741 garching , germany address = max planck institute for astrophysics , karl - schwarzschild - str . 1 , 85741 garching , germany |
the double - exchange ( de ) model @xcite is one of the basic ones in the theory of magnetism .
magnetic ordering appears in this model due to exchange coupling between the core spins and the conduction electrons .
the hamiltonian of the model is @xmath1 where @xmath2 and @xmath3 are the electrons annihilation and creation operators , @xmath4 is the operator of a core spin , @xmath5 is the electron hopping , @xmath6 is the exchange coupling between a core spin and n electrons , @xmath7 is the vector of the pauli matrices , and @xmath8 are spin indices .
the model ( the core spins being treated as classical vectors ) began to be studied several decades ago @xcite . because of a recent general interest in manganites , it reappeared in the focus of attention ( see reviews @xcite and references therein ) . in particular , application of the dynamical mean field approximation ( dmfa )
@xcite substantially advanced our understanding of the properties of the de model @xcite .
most of the papers dealing with the model , starting from classical paper by de gennes @xcite , considered the de hamiltonian with infinite exchange ( and with the addition of the antiferromagnetic superexchange , which is crucial for the explanation of magnetic properties of manganites ) .
however lately many properties of the model for arbitrary strength of the exchange were analyzed @xcite . in this paper
we calculate the temperature of a paramagnetic - ferromagnetic transition @xmath0 in a double - exchange model for arbitrary electron dispersion law , concentration and relation between the exchange coupling and the electron band width by formulating and solving the dmfa equations .
we treat the core spins as classical vectors .
spins being considered as classical vectors @xmath9 ( with the normalization @xmath10 ) , the problem becomes a single electron one .
the de hamiltonian in a single electron representation can be presented as @xmath11 let us introduce green s function @xmath12 and local green s function @xmath13 in the last equation , the averaging is with respect to random configurations of the core spins . in the framework of the dmfa approach to the problem ( see @xcite and references therein ) the local green s function is expressed through the the local self - energy @xmath14 by the equation @xmath15 where @xmath16 is the bare ( in the absence of the exchange interaction ) local green s function .
the self - energy satisfies equation @xmath17 where @xmath18 , and @xmath19 is a probability of a given spin orientation ( one - site probability ) .
the quantities @xmath20 and @xmath14 are @xmath21 matrices in spin space .
the difference between eq .
( [ cpa ] ) and similar equation appearing in the coherent potential approximation @xcite , is due to the fact that in our problem the disorder is annealed ( and not quenched ) . that is the probability of a given ( random ) core spin configuration
is determined by the energy of electron subsystem interacting with this configuration . to understand the dmfa assumption allowing to find the probability @xmath19 self - consistently with the solution of eq .
( [ cpa ] ) , notice that eq .
( [ cpa ] ) reduces the problem of electron scattering by many spins , to scattering by a single spin with the effective potential @xmath22 .
the spin is embedded in an effective medium , described by the hamiltonian @xmath23 , and , hence , by the local green s function @xmath24 .
consider the change in the number of electron states in such system due to a single spin .
we get @xcite @xmath25,\end{aligned}\ ] ] where the argument of both @xmath26 and @xmath27 is @xmath28 .
so the change in thermodynamic potential is @xcite @xmath29 where @xmath30 is the fermi function .
the dmfa approximation for the one - site probability @xmath19 is : @xmath31.\end{aligned}\ ] ] eqs .
( [ cpa ] ) and ( [ prob2 ] ) are the system of non - linear ( integral ) equations .
however , the system is grossly simplified in paramagnetic ( pm ) phase , where the macroscopic magnetization @xmath32 . in this case
@xmath33 , the averaging in eq . (
[ cpa ] ) can be performed explicitly , and we obtain @xmath34 where @xmath35 , @xmath36 , where @xmath37 is a unity matrix . in the ferromagnetic ( fm ) phase near the curie temperature , eqs .
( [ cpa ] ) and ( [ prob2 ] ) can be linearized @xcite with respect to @xmath38 .
thus we reduce the dmfa equations to a traditional mf equation @xcite @xmath39 the parameter @xmath0 is formally introduced as a coefficient in the expansion of @xmath40 with respect to @xmath38 .
non - trivial solution of the mf equation @xmath41 can exist only for @xmath42 , hence @xmath0 is the ferromagnetic transition temperature . in all cases
@xmath0 turns out to be much less the chemical potential , so we can consider electron gas as degenerate . in this case @xmath43 where the chemical potential
( in paramagnetic phase ) is found from the equation @xmath44 and @xmath45 is the number of electrons per site . for the @xmath0 , after straightforward , though lengthy algebra , we obtain @xmath46de , \end{aligned}\ ] ] where @xmath27 and @xmath47 are determined by the properties of the system in the pm phase ; they are the solutions of eqs .
( [ local ] ) and ( [ cpa2 ] ) ; @xmath48 . eq .
( [ theta ] ) is the main result of our paper . in the case of weak exchange @xmath49 eq .
( [ theta ] ) takes the form @xmath50de.\end{aligned}\ ] ] in fact , this equation is the mf approximation @xcite for the rkky hamiltonian @xcite , to which the original hamiltonian ( [ hamdxm ] ) can be reduced in the case @xmath51 . in the opposite case @xmath52 we have to consider only one spin sub - band .
( because of the symmetry of the problem with respect to transformation @xmath53 , everywhere further on we consider only the case @xmath54 . )
after we shift @xmath55 , @xmath27 and @xmath56 by @xmath6 , we obtain @xmath57de , \end{aligned}\ ] ] where @xmath47 is found from the equation @xmath58
let us apply eq .
( [ theta ] ) to the case of semi - circular ( sc ) bare density of states ( dos ) @xmath59 then @xmath60.\end{aligned}\ ] ] from eq .
( [ local ] ) follows @xmath61 and eq .
( [ cpa2 ] ) can be written as @xmath62 eq .
( [ theta ] ) takes the form @xmath63de .
\end{aligned}\ ] ] in the case of weak exchange @xmath64 , we obtain @xmath65 where @xmath66 in the case of strong exchange @xmath67 , eq .
( [ theta2 ] ) takes the form @xcite
@xmath68de\nonumber\\ = \frac{w\sqrt{2}}{4\pi}\left [ \sqrt{1-y^2}-\frac{1}{\sqrt{3}}\tan^{-1 } \sqrt{3(1-y^2)}\right],\end{aligned}\ ] ] where @xmath69 .
if , in addition , @xmath70 , we obtain @xmath71 notice , that virtual crystal approximation ( for @xmath72 ) @xcite gives @xmath73 thus being imprecise by a factor of 4 . for arbitrary exchange , the integral in eq .
( [ thetasc ] ) can be calculated only numerically .
the results of such calculation are presented on fig .
1 . = 3truein [ fig.1 ]
note that that our main result ( equation for the @xmath0 ) indicates it s own limits of validity . in part of the @xmath74 plane , eq .
( [ thetasc ] ) gives @xmath75 .
negative value of @xmath0 means , that at any temperature , including @xmath76 , the paramagnetic phase is stable with respect to appearance of small spontaneous magnetic moment , and strongly suggests that the ground state in this part of the phase plane is nonferromagnetic @xcite .
the mechanisms leading to the destruction of ferromagnetism can be easily understood in the extreme cases of weak and strong exchange .
it is well known , that for @xmath49 the de hamiltonian can be reduced to to the rkky hamiltonian ( see review @xcite and references to the original papers therein ) @xmath77 thus weak exchange between the electrons and the core spins generates effective long - range and sign oscillating effective exchange interaction between the core spins . in the opposite case @xmath52 , we can reduce the de hamiltonian ( [ generic ] ) to effective ( classic ) @xmath78 model ( see review @xcite and references therein ) . to do that , first diagonalize the exchange part of the de hamiltonian by choosing local spin quantization axis on each site in the direction of @xmath79 , the latter being determined by polar angle @xmath80 and azimuthal angle @xmath81 . in this representation
the hamiltonian is @xmath82 where @xmath83 the next step is to apply a canonical transformation @xmath84+\frac{1}{2}[s[s , h]]+\dots\end{aligned}\ ] ] which excludes all band - to - band transitions .
this can be achieved if we chose the operator @xmath85 in the form @xmath86 we have @xmath87_{nn'}=-t_{nn'}\left(\begin{array}{cc}0 & b_{nn'}\\ b_{n'n}^ * & 0\end{array}\right)\nonumber\\ \left[s,[s , h^{ex}]\right]_{nn } = 2\sum_{n''}i_{nn''}\left(\begin{array}{cc } -|b_{nn''}|^2 & 0\\ 0 & |b_{nn''}|^2\end{array}\right),\end{aligned}\ ] ] where @xmath88 . keeping terms up to the second order with respect to @xmath89 ( and only site - diagonal part of the second order terms ) we obtain @xmath90 in the second quantization the hamiltonian obtained has the form ( ignoring the constant term ) @xmath91 d_n^{\dagger}d_{n'}\nonumber\\ + \sum_{nn'}i_{nn'}{\bf m}_n\cdot{\bf m}_{n'}d_n^{\dagger}d_{n},\end{aligned}\ ] ] where @xmath92 is the operator of creation ( annihilation ) of a spinless electron .
thus strong exchange between the electrons and the core spins generates effective short - range antiferromagnetic exchange between the core spins . in conclusion
, we formulated the dynamical mean field approximation equations for the double - exchange model with classical spins for arbitrary electron dispersion law , concentration and relation between the exchange and the electron bandwidth . in the vicinity of the paramagnetic - ferromagnetic transition critical temperature @xmath0 , these equations were reduced to a mf equation , describing a single spin in an effective field , proportional to the macroscopic magnetization , and we obtained explicit formula for the @xmath0 . for the semi - circular electron density of states we explicitly calculated the transition temperature as a function of the exchange interaction and electron concentration | 54.8pt 54.8pt we study paramagnetic - ferromagnetic transition due to exchange interaction between classical localized magnetic moments and conduction electrons .
we formulate the dynamical mean field approximation equations for arbitrary electron dispersion law , concentration and relation between exchange coupling and the electron band width . solving these equations
we find explicit formula for the transition temperature @xmath0 .
we present the results of calculations of the @xmath0 for the semi - circular electron density of states . 2 |
the properties of hadronic bound states containing heavy quarks are characterized by a large separation of energy scales .
effects associated with the heavy - quark mass @xmath2 are perturbative and can be controlled once they have been separated from other , long - distance effects .
this separation is most conveniently done using an effective low - energy theory . in the case of hadrons containing a single heavy quark ,
the relevant effective theory is the heavy - quark effective theory ( hqet ) ( see @xcite for a review ) .
its lagrangian is @xmath3 where @xmath4 is the velocity of the hadron containing the heavy quark , @xmath5 the velocity - dependent heavy - quark field , and @xmath6 the covariant derivative orthogonal to @xmath4 . the operators arising at order @xmath7 in the effective lagrangian correspond to the kinetic energy of the heavy quark and its chromomagnetic interaction @xcite .
hadronic matrix elements of these operators appear in many applications of hqet .
the coefficient of the kinetic - energy operator is fixed by reparametrization invariance , an invariance under infinitesimal changes of the velocity @xcite .
hence , the only non - trivial short - distance coefficient in the effective lagrangian at next - to - leading order in @xmath7 is the coefficient @xmath8 of the chromomagnetic operator .
the dependence of the wilson coefficient on the large scale @xmath2 can be factorized using the renormalization group . in general , @xmath9 where @xmath10 \ , , \qquad \beta_0 = \frac{11}{3 } c_a - \frac{4}{3}\,t_f n_f \label{beta}\ ] ] is the @xmath11-function , and @xmath12 \label{gamma2}\ ] ] the anomalous dimension of the chromomagnetic operator ( given in the @xmath13 scheme ) .
the two - loop coefficient @xmath14 of the anomalous dimension was calculated in @xcite , and the two - loop initial condition @xmath15 in @xcite .
the corresponding one - loop expressions were obtained much earlier in @xcite .
the factor @xmath16 in ( [ rgesol ] ) compensates the scheme and scale dependence of the chromomagnetic operator in the effective lagrangian ( [ leff ] ) .
all short - distance physics relevant to physical observables is contained in the renormalization - scheme invariant function @xmath17 which , up to an @xmath2-independent normalization , is given by @xmath18^{\gamma_0/2\beta_0}\ , \big [ 1 + c(m ) \big ] \
, , \qquad c(m ) = \sum_{n=1}^\infty c_n \left ( \frac{\alpha_s(m)}{4\pi } \right)^n \ , .
\label{hatc}\ ] ] in the @xmath13 scheme , the next - to - leading correction is given by @xcite @xmath19 in the above equations , @xmath20 , @xmath21 , and @xmath22 are the colour factors for an su@xmath23 gauge group .
as an application , consider the mass splitting between the ground - state pseudoscalar and vector mesons containing a heavy quark .
its leading contribution comes from the chromomagnetic operator in the effective lagrangian , which is the first term breaking the heavy - quark spin symmetry . including terms of order @xmath1 , one obtains @xcite is missing there . ]
@xmath24 \ , , \label{split}\end{aligned}\ ] ] where @xmath25 is the matrix element of the chromomagnetic operator between ground - state mesons .
similarly , the hadronic parameters @xmath26 are defined in terms of the matrix elements of operators appearing at order @xmath1 in the effective lagrangian @xcite : @xmath27 parametrizes the spin - orbit interaction , while @xmath28 and @xmath29 are bilocal matrix elements of kinetic - chromomagnetic and chromomagnetic - chromomagnetic insertions .
( we omit the contributions from 4-quark operators @xcite , which are irrelevant to our discussion . )
these parameters are independent of the heavy - quark mass , but they depend on the renormalization scale in such a way that the scale dependence of the wilson coefficients is compensated . the coefficient of the spin - orbit term is related by reparametrization invariance to that of the chromomagnetic operator @xcite@xcite ( see also the appendix of @xcite ) : @xmath30 . introducing then the renormalization - scheme invariant parameters @xmath31\,\rho_{ls}^3(\mu ) \
, , \nonumber\\ \widehat\rho_a^3 & = & k^2(\mu)\,\rho_a^3(\mu ) \
, , \qquad \widehat\rho_{ls}^3 = \rho_{ls}^3(\mu ) \,,\end{aligned}\ ] ] we find @xmath32 \ , . \label{mvmp}\ ] ] all short - distance effects in this relation are contained in the single coefficient @xmath17 , which is the object of our study .
the perturbation series for @xmath8 can be arranged as @xmath34 where @xmath35 is the number of loops , and @xmath33 the leading coefficient of the @xmath11-function in ( [ beta ] ) . in this letter , we sum the terms of order @xmath36 to all orders of perturbation theory .
in other words , we consider the limit of large @xmath33 for fixed @xmath37 and calculate the coefficient @xmath8 to order @xmath38 , neglecting terms of order @xmath39 and higher . strictly speaking
, there is no sensible limit of qcd in which @xmath33 may be considered a large parameter ( except , maybe , @xmath40 ) ; however , retaining only the leading @xmath33 terms often gives a good approximation to exact multi - loop results ( see , e.g. , @xcite ) , in particular in cases when there is a nearby infrared renormalon @xcite . at the least
, it will provide us with some information about the summability of the perturbation series .
the coefficients @xmath41 of the terms with the highest degree of @xmath33 in ( [ series ] ) are determined by diagrams with @xmath42 light - quark loops , which are rather straightforward to calculate .
we work in dimensional regularization with @xmath43 space - time dimensions and adopt the @xmath13 subtraction scheme . at first order in @xmath38 ,
multiplicative renormalization simply amounts to a subtraction of the @xmath44 poles , and coupling - constant renormalization is given by ( @xmath45 ) @xmath46 the perturbation series for @xmath8 can then be written as @xmath47 the function @xmath48 is regular at @xmath49 . following the methods of @xcite ( used also in
@xcite ) , we now expand @xmath48 in powers of @xmath50 and @xmath51 , and @xmath52^l$ ] in powers of @xmath53 , to obtain a quadruple sum in ( [ struct ] ) .
combinatoric identities relate the @xmath54 terms , and hence the anomalous dimension of the chromomagnetic operator , to the taylor coefficients of @xmath55 @xcite : @xmath56 the finite terms , which determine the wilson coefficient itself , receive contributions from the taylor coefficients of both @xmath55 and @xmath57 @xcite : @xmath58 in this expression , all dependence on the heavy - quark mass resides in the function @xmath59 , while a dependence on the renormalization scale also enters through the coupling @xmath60 . by separating in ( [ finite ] ) terms depending on the two scales @xmath2 and @xmath61
, we find that the next - to - leading logarithmic correction @xmath62 in ( [ hatc ] ) is given by @xmath63}\,s(u ) + o(1/\beta_0 ^ 2 ) \
, , \nonumber\\ s(u ) & = & e^{\kappa u } \left .
\frac{f(0,u)-f(0,0)}{u } \right|_{\mu = m } \ , .
\label{delc}\end{aligned}\ ] ] the same result can also be derived by evaluating ( [ rgesol ] ) in the large-@xmath33 limit .
here @xmath64 , and the constant @xmath65 is introduced to compensate the scheme dependence of @xmath66 .
the most natural choice is to have @xmath67 in the @xmath13 scheme
. then the combination @xmath68 defines the inverse coupling in the so - called v scheme @xcite , with @xmath69 .
the function @xmath70 is the borel transform of the perturbation series for @xmath62 in that scheme .
the coefficient @xmath8 is obtained by matching scattering amplitudes of an on - shell heavy quark in an external field in qcd and hqet , including terms of order @xmath7 @xcite . as mentioned above , our focus is on @xmath35-loop diagrams with @xmath42 light - quark loops .
all hqet diagrams vanish because they contain no mass scale .
the relevant qcd diagrams are shown in fig .
[ fig : loop ] . they must be supplemented by the wave - function renormalization of the external quark fields @xcite .
we find that in the @xmath13 scheme the result for the function @xmath48 in ( [ struct ] ) has the form @xmath71 \ , , \label{feu}\ ] ] where @xmath72 is related to the contribution of a light - quark loop to the gluon self - energy , and @xmath73 \ , .
\label{nres}\end{aligned}\ ] ] these formulae reproduce the known @xmath74 @xcite and @xmath75 @xcite results , where @xmath76 . having obtained an explicit result for the function @xmath48 , we are ready to derive all - order results for the anomalous dimension and wilson coefficient of the chromomagnetic operator in the large-@xmath33 limit . from ( [ gamma ] ) , we obtain @xmath77 \ , , \label{gamma1}\end{aligned}\ ] ] which reproduces correctly the leading ( in @xmath33 ) term of the exact two - loop result ( [ gamma2 ] ) .
the radius of convergence of the perturbation series in ( [ gamma1 ] ) is @xmath78 .
next , setting @xmath79 in ( [ finite ] ) , we find that the perturbative expansion of the wilson coefficient at the scale @xmath79 reads @xmath80 \frac{\beta_0\alpha_s(m)}{4\pi } \nonumber\\ & & \quad + \left [ \left ( \frac{317}{9 } + \frac 43\pi^2 \right ) c_f + \left ( \frac{3535}{162 } + \frac{25}{9}\pi^2 + \frac{14}{3}\zeta(3 ) \right ) c_a \right ] \left ( \frac{\beta_0\alpha_s(m)}{4\pi } \right)^2 + \dots \bigg\ } \ , , \label{cm2}\end{aligned}\ ] ] which again reproduces the leading term of the exact two - loop result @xcite . unlike the case of the anomalous dimension , here retaining only the leading term in @xmath38 gives a reasonable approximation ; for @xmath81 colours , the exact two - loop result is @xmath82 showing that the sub - leading term is about one third of the leading one .
finally , we quote the result for the next - to - leading logarithmic correction @xmath62 in ( [ hatc ] ) .
expanding the integral in ( [ delc ] ) in powers of the coupling constant , we find the expansion coefficients @xmath83 \beta_0 + o(\beta_0 ^ 0 ) \ , , \nonumber\\
c_3 & = & \left [ \left ( \frac{317}{9 } + \frac 43 \pi^2 \right ) c_f + \left ( \frac{3535}{162 } + \frac{25}{9}\pi^2 + 4\zeta(3 ) \right ) c_a \right ] \beta_0 ^ 2 + o(\beta_0 ) \,.\end{aligned}\ ] ] the result for @xmath84 reproduces the leading term in ( [ c1 ] ) . for @xmath81 colours , the numerical values of the coefficients are @xmath85 , @xmath86 , @xmath87 , @xmath88 , @xmath89 , etc . they exhibit a ( same - sign ) factorial divergence @xmath90 , which renders the perturbation series for @xmath62 divergent and not borel summable . on the contrary ,
the coefficients in the expansion of the anomalous dimension in ( [ gamma1 ] ) appear to be well behaved ; in fact , this series is rapidly convergent .
although we do not expect that the higher - order coefficients obtained in the large-@xmath33 limit accurately represent the exact values , we do trust in the qualitative features of the results .
this allows us to derive an approximation for the next - to - next - to - leading order ( nnlo ) correction @xmath91 in ( [ hatc ] ) .
expanding ( [ rgesol ] ) to second order in @xmath92 , and inserting the known expansion coefficients for the three - loop @xmath11-function @xcite , the two - loop anomalous dimension @xcite , and the two - loop matching coefficient @xmath15 @xcite , we obtain an exact expression for @xmath91 depending only on the unknown three - loop coefficient of the anomalous dimension , for which we write @xmath93 , where @xmath94 and @xmath95 are unknown .
this parametrization takes into account that the anomalous dimension vanishes in the abelian limit @xmath96 @xcite .
we then expand @xmath97 and find @xmath98 where the numerical values refer to @xmath81 colours . because of the good convergence of the perturbation series for the anomalous dimension indicated by our analysis of the large-@xmath33 limit , it is conservative to assume that @xmath99 , which is equivalent to the statement that the true value of the three - loop anomalous dimension differs from its value in the large-@xmath33 limit by less than 100% . under this assumption , we obtain @xmath100 the uncertainty due to the unknown terms in the three - loop anomalous dimension is negligible compared with the overall size of the coefficient . using this result together with the exact one - loop coefficient @xmath84 in ( [ c1 ] ) , we find for @xmath101 @xmath18^{9/25}\ , \left [ 1 + 0.672\alpha_s(m ) + ( 1.33\pm 0.04)\alpha_s^2(m ) + \dots \right ] \,.\ ] ] in the context of the heavy - quark expansion , this is the most precisely known wilson coefficient to date .
as a consequence of its divergent behaviour , the perturbation series for @xmath62 must be truncated close to its minimal term , and the perturbative result for the wilson coefficient is intrinsically ambiguous . although we have explored this feature only in the large-@xmath33 limit , it is believed to be of a rather general nature ( see , e.g. , @xcite ) . in regularization schemes without an explicit infrared cutoff ( such as dimensional regularization with minimal subtraction ) , the perturbative calculation of the coefficients @xmath102 involves an integration over all gluon momenta , including long - distance contributions from soft gluons .
high orders in the expansion probe the region of increasingly smaller gluon momenta , a regime where perturbation theory is bound to fail .
remarkably , the perturbation series knows about its deficiency and signals it through the divergent behaviour of the expansion coefficients .
we will now investigate this phenomenon in more detail . a convenient tool to study the ambiguities encountered in the attempt to resum an asymptotic perturbation series is provided by the borel image of that series , given by the function @xmath70 in the integral representation ( [ delc ] ) .
using ( [ feu ] ) , we find that @xmath103 - e^{-5u/3}\,\frac{c_a}{u } \ , . \label{f0u}\ ] ] if it existed , the laplace integral in ( [ delc ] ) would define the borel sum of the perturbation series for @xmath62 in the large-@xmath33 limit .
however , the presence of singularities along the integration contour ( i.e. for positive values of @xmath51 ) makes the integral ill - defined . in the large-@xmath33 limit ,
the function @xmath70 has pole singularities at half - integer values of @xmath51 called infrared renormalons @xcite . any attempt to resum
perturbation theory involves an arbitrary choice of how to deal with these singularities .
a measure of the ambiguity in the value of the borel sum , which is of the same size as the minimal term in the series , is provided by the residue of the nearest singularity , in the present case located at @xmath104 . using ( [ av ] ) , we obtain @xmath105 = \frac{1}{\beta_0 } \left ( 2 c_f + \frac 74 c_a \right ) \frac{\lambda_{\mathrm{v}}}{m } \ , , \label{delren}\ ] ] with @xmath69 .
the situation is similar to the well - known case of the infrared renormalon ambiguity in the definition of the pole mass of a heavy quark @xcite .
the perturbation series relating the pole mass to a short - distance mass is also affected by an infrared renormalon at @xmath104 , and the corresponding ambiguity is @xmath106 . taking into account that the chromomagnetic operator appears in the effective lagrangian ( [ leff ] ) multiplied by a power of the inverse pole mass
, we find that @xmath107 = \frac{7c_a}{8c_f}\,\frac{\delta m}{m^2 } + o(1/\beta_0 ^ 2 ) \,.\ ] ] hence , the infrared renormalon ambiguity in the product @xmath108 is purely non - abelian , and commensurate with the contributions of higher - dimensional operators in the @xmath7 expansion . indeed , in all predictions for physical quantities , the infrared renormalon ambiguities must cancel against corresponding ambiguities in the long - distance matrix elements of some higher - dimensional operators .
when the heavy - quark expansion is applied to calculate a physical quantity , e.g. the mass splitting in ( [ split ] ) , the resulting expressions contain short - distance wilson coefficients and long - distance hadronic matrix elements .
as mentioned above , in regularization schemes without a hard momentum cutoff the wilson coefficients also contain contributions from large distances , where perturbation theory is ill - defined , and these contributions produce infrared renormalon ambiguities .
likewise , the hadronic matrix elements ( which are not calculable perturbatively ) contain contributions from small distances , which lead to ultraviolet renormalon ambiguities .
in other words , in such schemes the separation of short- and long - distance contributions into wilson coefficients and matrix elements is intrinsically ambiguous .
only when all contributions are combined to form a physical quantity , an unambiguous result is obtained . in the context of the hqet
, the cancelations between infrared and ultraviolet renormalon ambiguities have been traced in detail in @xcite .
here we consider the particular case of the mass splitting .
the ultraviolet contributions to the hadronic parameters @xmath109 are independent of the nature of the external states and thus may be calculated using quark and gluon rather than hadron states . at one - loop order ,
the relevant diagrams are shown in fig .
[ fig : uv ] . on dimensional grounds , they are linearly divergent in the ultraviolet region . in dimensional
regularization such power divergences are not seen ( by definition ) , but they reflect themselves in a factorial divergence of the perturbation series when higher - order diagrams are taken into account
. we shall again analyse these contributions in the large-@xmath33 limit , by inserting an arbitrary number of light - quark loops in all possible ways into the diagrams of fig .
[ fig : uv ] . the relevant matrix elements
are then expanded to linear order in the gluon momentum @xmath110 and projected onto the structure of the chromomagnetic operator .
the external quarks are kept off - shell in order to provide for an infrared regulator .
the borel transforms of these matrix elements contain ultraviolet renormalon poles at @xmath104 .
for dimensional reasons , the residues are proportional to @xmath111 times the matrix element @xmath112 of the chromomagnetic operator , which is the only lower - dimensional operator contributing to the mass splitting .
therefore , the leading ultraviolet renormalon ambiguities of the @xmath109 parameters are proportional to @xmath113 .
note that diagrams in which the external gluon is attached to a heavy - quark line , or to a two - gluon kinetic or chromomagnetic vertex , do not contribute .
the diagram ( c1 ) does not contribute to the chromomagnetic structure since the spin - orbit vertex only couples to the chromoelectric field .
moreover , with a suitable choice of the external momenta one can achieve that the diagrams ( c2 ) and ( c3 ) vanish .
( their sum vanishes with any choice of momenta . )
a subtlety that needs to be taken into account is that there exists an ultraviolet renormalon at @xmath104 in the @xmath7 suppressed contribution to the quark wave - function renormalization with a kinetic - energy insertion , shown in diagrams ( d1 ) and ( d2 ) ; the corresponding ambiguity @xmath114 in the normalization of @xmath112 contributes to @xmath115 and cancels the @xmath116 terms in the diagrams ( a1 ) and ( a2 ) . from a direct evaluation of the non - vanishing diagrams , using the method described above
, we obtain for the ultraviolet renormalon ambiguities in the large-@xmath33 limit @xmath117 with coefficients @xmath118 , @xmath119 , and @xmath120 .
the sum of the ultraviolet ambiguities in ( [ split ] ) is thus given by @xmath121 it precisely cancels the infrared renormalon ambiguity in the leading term .
it is interesting that the requirement of a cancelation of renormalon ambiguities in physical quantities allows us to derive further information about the asymptotic behaviour of the expansion coefficients @xmath102 in ( [ hatc ] ) without any additional dynamical input @xcite .
the point is that , beyond the large-@xmath33 limit , the different terms in ( [ mvmp ] ) contain different powers of @xmath122^{\gamma_0/2\beta_0}$ ] , since they contain different powers of the wilson coefficient @xmath17 .
these leading logarithms are exactly known to all orders of perturbation theory .
they multiply the renormalon ambiguities of the various terms . in order to maintain the cancelation between infrared and ultraviolet renormalon ambiguities in the presence of the leading logarithms , the borel transform for the wilson coefficient in ( [ f0u ] )
has to be modified . in the vicinity of @xmath104 ,
the simple pole in @xmath123 must be replaced by a sum of branch points , @xmath124 \,,\ ] ] where @xmath125 is the one - loop coefficient of the anomalous dimension of the chromomagnetic operator given in ( [ gamma2 ] ) , and @xmath126 denotes the two - loop coefficient of the @xmath11-function .
the normalization constants @xmath127 are undetermined beyond the large-@xmath33 limit @xcite . up to corrections of order
@xmath128 , the corresponding asymptotic behaviour of the expansion coefficients @xmath102 is @xmath129 \,.\ ] ] for very large @xmath130 , the last term gives the dominant behaviour ( since @xmath131 ) ; however , for moderate values of @xmath130 all contributions are of similar importance . in order to evaluate the borel integral ( [ delc ] ) , it is useful to rewrite it in the form @xcite @xmath132 the function @xmath133 which is the inverse mellin transform of the borel image @xmath70 , describes the distribution of gluon virtualities in the one - loop diagrams in fig .
[ fig : loop ] , which contribute to the calculation of @xmath62 . using the methods developed in @xcite
, we obtain @xmath134 these functions are shown in fig . [
fig : plot ] .
note that the integral in ( [ distr ] ) runs over the landau pole in the running coupling constant , located at @xmath135 .
this is how infrared renormalons make their appearance . as in the case of the original borel integral
, we must specify an arbitrary prescription of how to deal with the landau singularity .
the renormalon ambiguity is given by the residue of the pole , i.e. @xmath136 = w(\tau_l)/\beta_0 $ ] . from an expansion of the distribution function in the region @xmath137
, we readily recover the result ( [ delren ] ) .
the representation ( [ distr ] ) makes explicit that perturbative calculations contain long - distance contributions from the region of low momenta in feynman diagrams . at the same time
, it provides for a convenient way to separate these long - distance contributions from the short - distance ones by introducing a hard separation scale @xmath138 .
we may then define the short - distance coefficient @xmath139 as @xmath140 as long as @xmath138 is chosen large enough , this coefficient can be reliably calculated in perturbation theory and is free of renormalon ambiguities . the long - distance contribution eliminated by this procedure must be combined with the non - perturbative power corrections in the heavy - quark expansion .
our goal was to study the large - order behaviour of the wilson coefficient of the chromomagnetic operator in the hqet lagrangian .
the renormalization - group invariant coefficient @xmath141 defined in ( [ hatc ] ) is known exactly at next - to - leading order @xcite . here
, we have derived an excellent next - to - next - to - leading order approximation by combining the exact two - loop matching condition obtained in @xcite with an approximation for the three - loop anomalous dimension .
this estimate has been obtained by summing fermion - loop contributions to all orders in perturbation theory , thus deriving the leading term in an expansion in powers of @xmath38 .
we have also derived an all - order result for the wilson coefficient in the large-@xmath33 limit , finding that the perturbation series for @xmath17 exhibits a divergent behaviour already in low orders , caused by a nearby infrared renormalon singularity .
the resulting ambiguity is commensurate with terms of order @xmath1 in the effective lagrangian , whose corresponding ultraviolet renormalons we have identified .
let us now investigate the implications of our results by comparing the various approximations for the wilson coefficient . to this end , we evaluate the partial sums @xmath142^{\gamma_0/2\beta_0}\ , \left [ 1 + \sum_{n=1}^n c_n \left ( \frac{\alpha_s(m)}{4\pi } \right)^n \right]\ ] ] for @xmath101 light quark flavours , using @xmath143 and @xmath144 in the @xmath13 scheme . the ratio of the two coefficient functions at these scales ( for @xmath101 ) gives the perturbative correction to the ratio of mass splitting between the ground - state pseudoscalar and vector mesons in the bottom and charm systems @xcite : @xmath145 \approx 0.89 \ , , \label{splitratio}\ ] ] where @xmath146 is a combination of the hadronic parameters @xmath109 introduced in ( [ split ] ) , and for simplicity we neglect short - distance corrections in the @xmath7 terms .
& 0.580 & 0.580 & 0.692 & 0.692 & 0.838 & 0.838 + 1 & 0.734 & 0.659 & 0.993 & 0.847 & 0.739 & 0.778 + 2 & 0.801 & 0.697(1 ) & 1.207 & 0.967(4 ) & 0.664 & 0.721(3 ) + 3 & 0.850 & & 1.463 & & 0.581 & + 4 & 0.898 & & 1.875 & & 0.479 & + 5 & 0.958 & & 2.714 & & 0.353 & + our predictions for the wilson coefficients @xmath151 and @xmath152 are given in table [ tab:1 ] . in the upper part of the table , we show the partial sums obtained at different orders in perturbation theory , both in the large-@xmath33 limit and exactly , so far as the exact results are known . in the large-@xmath33 limit , the divergent behaviour of the perturbation series sets in already in low orders . at the scale @xmath153 , the minimal term in the series
is reached around @xmath154 , whereas it is reached around @xmath155 at the scale @xmath156 . as a consequence ,
the values of the coefficient functions at the two scales become more and more different for larger @xmath147 , and their ratio drifts away from the experimental value of the ratio of mass splittings .
the reason for this behaviour lies in the nearby location of the first infrared renormalon .
this can be understood intuitively by considering the distribution functions shown in fig .
[ fig : plot ] , which are broad and extend far into the infrared region .
note that the point @xmath79 corresponds to @xmath157 in the @xmath13 scheme .
hence , there are essentially no contributions from scales above @xmath2 ( in the case of the @xmath158 terms those are subtracted by @xmath13 renormalization ) , but on the other hand the distribution functions decrease very slowly ( @xmath159 ) in the infrared region @xmath160
. the strength of this decrease is determined by the position of the nearest infrared renormalon singularity . in general ,
if the nearest infrared renormalon is located at @xmath161 , then @xmath162 for @xmath160 @xcite . as a consequence ,
the dominant contributions come from scales significantly below the heavy - quark mass .
the average value of @xmath163 is @xmath164 , corresponding to the scale @xmath165 in the @xmath13 scheme .
this is precisely the scale obtained in the blm scale - setting prescription @xcite . in view of the low value of the average virtuality ,
the bad convergence of the perturbation series is not surprising . ] in the lower portion of table [ tab:1 ] , we show the results for the wilson coefficients obtained by taking the principal value of the borel integral .
the quoted errors reflect the renormalon ambiguities , which are sizeable , in particular , at the scale @xmath156 .
the borel resummation gives values close to those obtained when the series is truncated at the minimal term ( in the latter case the error is taken to be the size of the minimal term ) , as it is expected on general grounds . in the last row
we show the short - distance contributions to the coefficient functions arising from virtualities above @xmath166gev ( in the v scheme ) , as defined in ( [ csd ] ) .
the portion of infrared contributions is much larger at the scale @xmath156 than it is at the scale @xmath153 .
this is clearly seen in fig .
[ fig : plot ] , where the arrows indicate the location of the separation scale @xmath167gev ( corresponding to 435 mev in the @xmath13 scheme ) .
if the infrared contributions are cut away , the resulting values of the coefficients are again close to their values obtained from low - order calculations
. although the all - order results are instructive , we must not forget that they are obtained in a very questionable approximation scheme ( the large-@xmath33 limit ) .
indeed , the few exact results available indicate that the perturbation series may be much better behaved than indicated by the large-@xmath33 limit . still , we trust in the qualitative observations that the series start to diverge at some ( low ) order , and that the onset of the divergence is reached early the lower the scale @xmath2 is
. we will now explore the implications of these results for the phenomenology of the heavy - quark expansion . for a consistent inclusion of power - suppressed effects the perturbative coefficients of the leading terms
must be known with sufficient accuracy .
different truncation or resummation schemes imposed on the short - distance coefficients imply different definitions of the hadronic parameters appearing at higher order in the heavy - quark expansion , such as the parameter @xmath146 in ( [ splitratio ] ) . only when the perturbative coefficients are truncated close to their asymptotic value , the hadronic parameters are @xmath2 independent as they should . when applied to the particular case of the chromomagnetic operator ,
the conclusion is that there is not much to be gained by calculating @xmath17 beyond the first few orders of perturbation theory .
taking into account that the series for @xmath168 diverges earlier than that for @xmath169 , we may argue that the optimal perturbative prediction for the ratio based on exact information is obtained by combining the nnlo result at the bottom scale with the nlo result at the charm scale .
this gives @xmath170 in the large-@xmath33 limit , the corresponding ratio equals 0.81 and is indeed very close to the principal value of the borel integral . using the result ( [ opt ] )
, we conclude that the power corrections in ( [ splitratio ] ) give a contribution of about 11% , corresponding to a scale @xmath171mev .
thus , if the asymptotic behaviour of perturbation theory is carefully taken into account , it appears that power corrections in the heavy - quark expansion are well under control . | the leading @xmath0 terms in the wilson coefficient and anomalous dimension of the chromomagnetic operator in the heavy - quark effective lagrangian are summed to all orders of perturbation theory .
the perturbation series for the anomalous dimension is well behaved , while that for the wilson coefficient exhibits a divergent behaviour already in low orders , caused by a nearby infrared renormalon singularity .
the resulting ambiguity is commensurate with terms of order @xmath1 in the effective lagrangian , whose corresponding ultraviolet renormalons are identified .
an excellent approximation for the scheme - invariant wilson coefficient at next - to - next - to - leading order in renormalization - group improved perturbation theory is obtained .
cern - th/97 - 102 + budker inp 97 - 60 + hep - ph/9707318 * higher - order estimates of the chromomagnetic + moment of a heavy quark * a.g .
grozin + _ budker institute of nuclear physics , novosibirsk 630090 , russia _
+ m. neubert + _ theory division , cern , ch-1211 geneva 23 , switzerland _ ( submitted to nuclear physics b ) cern - th/97 - 102 + july 1997 |
in the past several observations were made which could not be understood under the given knowledge of nuclear physics .
first @xcite , evidence for the production , via secondary reactions in cern w targets , of a long - lived superheavy isotope @xmath5112 has been obtained .
fission fragments were observed in hg sources chemically separated from the w targets , and the measured masses undergoing fission ( like 308 , 315 and 318 ) were consistently interpreted @xcite as due to 5 different common molecules ( like -cl , -n@xmath6 and -no@xmath7 ) of a superheavy isotope with z=112 ( eka - hg ) , n@xmath8160 .
there was however no understanding of the long deduced half - life of several weeks and of the large deduced fusion cross section of several mb @xcite . in a study of actinide fractions from the same w target
, long - lived isomeric states were found in the neutron - deficient @xmath9am and @xmath9bk nuclei with half - lives of 0.6 y and @xmath1030 d , respectively @xcite .
their character was not clear , being far from closed shell nuclei , where high spin isomers are known , and living much longer than the known fission isomers .
in addition , several unidentified @xmath4-particle groups were found in some actinide sources .
thus , 5.14 mev ( t@xmath11 = 3.8@xmath121 y ) , 5.27 mev ( t@xmath11 = 625@xmath1284 d ) and 5.53 mev ( t@xmath11 = 26@xmath127 d ) groups were respectively found in the bk , es and lr - no sources @xcite ( see fig .
again , one could not understand their relatively low energies ( e.g. , 5.53 mev in lr - no as compared to g.s . to g.s .
transitions around 8 mev , which have about 13 orders of magnitude larger penetrability factors ) , and very short half - lives ( 10@xmath13 - 10@xmath14 shorter than predicted from the systematics @xcite ) .
the deduced evaporation - residue cross sections @xcite , in the mb region , are also several orders larger than expected .
a number of unexplained phenomena have also been observed in nature .
po halos , produced by @xmath4-particles from @xmath15po ( t@xmath11 = 138.4 d ) , @xmath16po ( t@xmath11 = 164 @xmath17s ) and @xmath18po ( t@xmath11 = 3.0 m ) , have been seen in mica @xcite .
hulubei and cauchois @xcite saw induced po x - rays in the mineral petzite ( ag@xmath6aute@xmath7 ) . in both cases
the precursors th and u were not present .
evidence for unidentified @xmath4-particles around 4.5 mev have been seen in several minerals @xcite . according to the chemistry
used they were thought @xcite to be due to eka - os ( element 108 ) .
it was also suggested @xcite that @xmath19 cm may be a descendant of this superheavy element .
these results were not substantially convincing .
the predicted @xcite half - life for such low energy @xmath4-particles in z = 108 is around 10@xmath0 y , implying the existence of an impossibly large amount of material of about 0.1 g in the studied samples . in the following a consistent possible interpretation for these phenomena
will be given , based on the recently discovered @xcite long - lived high - spin super- and hyper - deformed isomeric states .
the @xmath0o + @xmath1au and @xmath2si + @xmath3ta reactions have been studied using beams from the pelletron accelerator in rehovot , catcher foil technique and long - period off - line measurements in jerusalem . in the first reaction at e@xmath20 = 80 mev @xcite , long - lived high - spin super - deformed isomeric states have been found .
they exhibited themselves by decaying with relatively low energy ( 5.2 mev ) and very enhanced ( t@xmath11
@xmath8 90 m , 3x10@xmath13 enhancement ) @xmath4-particles in coincidence with @xmath21-rays of a superdeformed band ( sdb ) , and also by long - lived proton decays with half - lives of about 6 and 70 h. in the @xmath2si + @xmath3ta reaction at e@xmath20 = 125 mev @xcite , about 10% below the coulomb barrier , a high energy ( 7.8 - 8.6 mev ) and long - lived ( 40 d @xmath22 t@xmath11 @xmath22 2.1 y ) @xmath4-particle group was found in coincidence with sdb @xmath21-rays ( fig .
2 ) , and consistently interpreted as due to production of a high - spin long - lived hyper - deformed isomeric state in @xmath23hg , which decays by a factor of 10@xmath24 retarded @xmath4-particles to sdb states in @xmath25pt .
( @xmath23hg may be produced via 1p1n evaporation reaction and 3 consecutive @xmath4-decays .
the probability that almost all the measured @xmath21-ray energies will accidentally fit the predicted energies of a sdb transitions is 4.1x10@xmath26 .
( the background is zero ) .
the high energy @xmath21-rays are most probably sum events rather than single photo - peak events .
therefore the number of compton events is relatively low [ 20 ] ) . - @xmath21 coincidence plot from one measurement of the @xmath2si + @xmath3ta reaction .
e@xmath20 = 125 mev , with 200 @xmath17g/@xmath27 c catcher foil , taken for 76.8 d , starting 77.4 d after the end of irradiation .
the @xmath21 - ray energies of the encircled events fit with sdb transitions .
the squared events fit with known characteristic x - rays and the events in triangles are identified with known @xmath21 - ray transitions ( see ref . [ 20 ] ) . *
right : * e@xmath21 versus j(j+1)-(j-1)j for the @xmath21 - rays seen in coincidence with 7.8 - 8.6 mev @xmath4 - particles .
( the encircled events in the left figure plus similar events obtained in a second measurement [ 20]).,title="fig:",scaledwidth=45.0% ] - @xmath21 coincidence plot from one measurement of the @xmath2si + @xmath3ta reaction .
e@xmath20 = 125 mev , with 200 @xmath17g/@xmath27 c catcher foil , taken for 76.8 d , starting 77.4 d after the end of irradiation .
the @xmath21 - ray energies of the encircled events fit with sdb transitions .
the squared events fit with known characteristic x - rays and the events in triangles are identified with known @xmath21 - ray transitions ( see ref . [ 20 ] ) . *
right : * e@xmath21 versus j(j+1)-(j-1)j for the @xmath21 - rays seen in coincidence with 7.8 - 8.6 mev @xmath4 - particles .
( the encircled events in the left figure plus similar events obtained in a second measurement [ 20]).,title="fig:",scaledwidth=41.0% ]
based on these results and on the predicted @xcite excitation energies of the second and third minima in various actinide nuclei , the unidentified @xmath4-particle groups from the actinide fractions ( section 1 ) can be consistently interpreted @xcite , both from the low energy and the enhanced lifetime points of view , as due to ii@xmath28 @xmath29 ii@xmath28 or iii@xmath28 @xmath29 iii@xmath28 transitions ( see table 1 ) . the 5.14 mev group in bk grew at the beginning ( fig . 1 , left , center ) .
it therefore may be due to a ii@xmath28 @xmath29 ii@xmath28 transition from @xmath30am or a iii@xmath28 @xmath29 iii@xmath28 transition from @xmath30 cm , which were produced from @xmath30bk by the ec/@xmath31@xmath32 process @xcite .
( fig . 6 of ref .
@xcite indicates that the first case , where a rather deep second minimum is predicted @xcite , is more likely ) .
the 5.27 and 5.53 mev groups in the es and lr - no sources are consistent with iii@xmath28 @xmath29 iii@xmath28 transitions from @xmath19es and @xmath33no , respectively .
llllll + mother & e@xmath4(mev ) & e@xmath4(mev)@xmath34 & e@xmath4(mev)@xmath34 & t@xmath11@xmath35 ( @xmath31@xmath7 ; @xmath31@xmath6 ; @xmath31@xmath36 ) & t@xmath37 + isotope & [email protected] . &
ii@xmath28@xmath29ii@xmath28 & iii@xmath28@xmath29iii@xmath28 + + @xmath30am@xmath38 & 5.94 & * 5.13 * & 4.53 & 10.8 y ( 0.71 ; 0.0 ; 0.09 ) & 3.8@xmath121 y + @xmath30cm@xmath38 & 6.51 & 5.90 & * 5.24 * & 0.5 y ( 1.05 ; 0.17 ; 0.0 ) & 3.8@xmath121 y + @xmath19es & 7.37 & 7.47 & * 5.27 * & 382 d ( 1.05 ; 0.19 ; 0.0 ) & 625@xmath1284 d + @xmath33no & 8.42 & @xmath87.8 & @xmath8 * 5.6 * & 81 d ( 1.20 ; 0.19 ; 0.0 ) & 26@xmath127 d + @xmath34 deduced from ref . [ 21 ]
. @xmath35 calculated according to formulas given in refs .
@xmath31@xmath7 and @xmath31@xmath36 values were deduced from the @xmath39@xmath7 and @xmath39@xmath36 values given in ref .
[ 21 ] using fig . 2 from w. nazarewicz and i. ragnarsson , handbook of nuclear properties .
d. poenaru and w. greiner ( clarendon press , oxford , 1996 ) p. 80 .
the value of @xmath31@xmath6 was taken equal to @xmath39@xmath6 .
@xmath38 since the intensity of the 5.14 mev group in bk grew at the beginning ( fig . 1 , left , center )
, it may be due to a decay of an isotope of cm or am which is produced from bk by the ec/@xmath40 process .
( similarly to the @xmath9bk and @xmath9am cases which decay to @xmath9pu [ 4 ] ) .
+ \a ) i@xmath28@xmath29i@xmath28 .
normal @xmath4 s .
+ b ) ii@xmath28@xmath29i@xmath28 .
retarded @xmath4 s : + @xmath41ir@xmath29@xmath42re ( ref . [ 20 ] ) . + c ) ii@xmath28@xmath29ii@xmath28 .
enhanced @xmath4 s : + @xmath15fr@xmath29@xmath43at ( ref . [ 18 ] ) .
+ @xmath8@xmath30am@xmath29@xmath44np ( ref . [ 8 ] ) .
+ d ) iii@xmath28@xmath29ii@xmath28 .
retarded @xmath4 s : + @xmath23hg@xmath29@xmath25pt ( ref . [ 20 ] ) .
+ e ) iii@xmath28@xmath29iii@xmath28 .
enhanced @xmath4 s : + @xmath45es@xmath29@xmath46bk ( ref . [ 8 ] ) .
+ @xmath8@xmath33no@xmath29@xmath47fm ( ref . [ 8 ] ) .
+ f ) ii@xmath28@xmath29i@xmath28 .
retarded protons : + @xmath48tl@xmath29@xmath1hg ( ? ) ( ref .
+ @xmath49fr@xmath29@xmath50rn ( ? ) ( ref .
+ figure 3 .
summary of abnormal particle decays seen in various experiments . + the long - lived isomeric states in the @xmath9am and @xmath9bk nuclei can be interpreted as due to the existence of super- or hyper - deformed isomers in these nuclei which decay by the ec or @xmath51 processes eventually to the normal states in @xmath9pu @xcite .
3 summarizes the various new and abnormal particle decays which were observed in different experiments .
it should be mentioned that a high - spin long - lived super - deformed isomeric state has been predicted back in 1969 @xcite .
the experiments @xcite concerning element 112 can now be understood as due to its production in a long - lived super- or hyper - deformed isomeric state , rather than in the ground state .
the large deduced fusion cross section is due to two effects .
a ) very little extra - push energy is needed in order to produce the compound nucleus in such a state .
b ) the fusion probability in the secondary reactions , where the projectile is itself a deformed fragment produced just within 2x10@xmath52 s before interacting with another w nucleus in the target , is much larger as compared to fusion with a normal projectile , due to the much reduced coulomb repulsion @xcite .
the discovered high spin super- and hyper - deformed isomeric states may be the source for the po halos @xcite ( either in the @xmath15po , @xmath16po and @xmath18po isotopes which decay by @xmath21 s to the ground states , or in the appropriate pb or bi nuclides which decay by @xmath31@xmath53 to the corresponding po isotopes ) , and also for the observed po induced x - rays in petzite @xcite .
predictions @xcite , and extrapolations from these predictions , of the iii@xmath28 @xmath29 iii@xmath28 @xmath4-particle energies .
the black dots are the predictions for various isotopes of z=96 and z=100 .
the straight lines are extrapolations from these predictions for z=108 and z=112 .
the open triangle shows the position of the 6.73 mev @xmath4-particles seen with the hg sources from the cern w targets @xcite .
the open circle shows the position of 4.5 mev @xmath4-particles in z=108 .
the low energy @xmath4-particles around 4.5 mev @xcite may consistently be interpreted as due to a very enhanced iii@xmath28 @xmath29 iii@xmath28 transition in z@xmath8108 and a@xmath8271 .
the predicted @xcite half - life in this case is around 10@xmath54 y , which implies an amount of material of about 10 ng .
this resolves the main difficulty in understanding these data where an impossible amount of 0.1 g was deduced assuming a normal @xmath4-transition ( see the introduction ) .
in addition , an extrapolation of the predictions @xcite of @xmath4-energies of iii@xmath28 @xmath29 iii@xmath28 transitions ( see fig .
4 ) , shows that for z=108 , e@xmath4 of about 4.5 mev corresponds to n@xmath8162 .
this is consistent with the suggestion @xcite that @xmath19 cm may be a descendent of the superheavy element with z=108 which decays by the 4.5 mev @xmath4-particles , since @xmath19 cm can be obtained from @xmath55hs@xmath56 by successive 6 @xmath4-decays .
it should however be mentioned that in principle the above 4.5 mev @xmath4-particles may also be due to a very retarded ii@xmath28 @xmath29 i@xmath28 or iii@xmath28 @xmath29 ii@xmath28 transition in the region of os itself .
( for normal 4.5 mev @xmath4-particles in os the expected @xcite half - life is about 1 y. such short - lived nuclide can not exist in nature ) .
we would like to thank j. l. weil for taking part in the measurements with the actinide sources , a. pape for bringing the work of ref .
@xcite to our attention and n. zeldes for very valuable discussions .
+ a. marinov , s. gelberg , d. kolb and j. l. weil , proc .
int . conf . on fission and netron - rich nuclei ,
andrews , scotland , 1999 .
j. h. hamilton , w. r. phillips and h. k. carter ( world scientific , 2000 ) p. 231 . | long - lived high - spin super- and hyper - deformed isomeric states , which exhibit themselves by abnormal radioactive decays , have been observed using the @xmath0o + @xmath1au and @xmath2si + @xmath3ta reactions .
they make it possible to understand the production , via secondary reactions , of the long - lived superheavy element with z = 112 and of the abnormally low energy and very enhanced @xmath4-particle groups seen in various actinide sources .
they might also explain some puzzling phenomena seen in nature . |
this paper continues the development of the variational approach to multisymplectic field theory introduced in marsden , patrick , and shkoller @xcite . in that paper , only first - order field theories were considered .
herein , we shall focus on second - order field theories , i.e. , those field theories governed by lagrangians that depend on the spacetime location , the field , and its first and second partial derivatives .
multisymplectic geometry and its applications to covariant field theory and nonlinear partial differential equations ( pde ) has a rich and interesting history that we shall not discuss in this paper ; rather , we refer the reader to @xcite and the references therein .
the covariant multisymplectic approach is the field - theoretic generalization of the symplectic approach to classical mechanics .
the configuration manifold @xmath2 of classical lagrangian mechanics is replaced by a fiber bundle @xmath3 over the @xmath4@xmath5@xmath6 dimensional spacetime manifold @xmath7 , whose sections are the physical fields of interest ; the lagrangian phase space is @xmath8 in lagrangian mechanics , whereas for kth - order field theories , the role of phase space is played by the kth - jet bundle of @xmath0 , @xmath9 , thus reflecting the additional dependence of the fields on spatial variables . for a given smooth lagrangian @xmath10
, there is a distinguished symplectic @xmath11-form @xmath12 on @xmath8 , whose hamiltonian vector field is the solution of the euler - lagrange equations of lagrangian mechanics .
lagrangian field theories , on the other hand , governed by covariant lagrangians @xmath13 , can be completely described by the multisymplectic @xmath4@xmath5@xmath11-form @xmath14 on @xmath15 , the field - theoretic analogue of the symplectic @xmath11-form @xmath12 of classical mechanics . in the case that @xmath7 is one dimensional , @xmath14 reduces to the usual time - dependent @xmath11-form of classical nonautonomous mechanics ( see @xcite ) . traditionally , the symplectic @xmath11-form @xmath12 as well as the multisymplectic @xmath4@xmath5@xmath11-form @xmath14 are constructed on the lagrangian side , using the pull - back by the legendre transform of canonical differential forms on the dual or hamiltonian side .
recently , however , marsden , patrick , and shkoller @xcite have shown that for first order field theories wherein @xmath16 , @xmath17 arises as the boundary term in the first variation of the action @xmath18 for smooth mappings @xmath19 .
this method is advantageous to the traditional approach in that * a complete geometric theory can be derived while staying entirely on the lagrangian side , and * multisymplectic structure can be obtained in non - standard settings such as discrete field theory .
the purpose of this paper , is to generalize the results of @xcite to the case that @xmath20 . in section 2 ,
we prove in theorem [ thm_main ] , that a unique multisymplectic @xmath4@xmath5@xmath11 form arises as the boundary term of the first variation of the action function .
we then prove in theorem [ thm_mff ] , the multisymplectic form formula for second - order field theories , a covariant generalization of the fact that in conservative mechanics , the flow preserves the symplectic structure .
we then obtain the covariant noether theorem for second - order field theories , by taking the first variation of the action function , restricted to the space of solutions of the covariant euler - lagrange equations . in section 3
, we use our abstract geometric theory on the camassa - holm ( ch ) equation , a model of shallow water waves that simultaneously exhibits solitary wave interaction and wave - breaking .
we show that the multisymplectic form formula produces a new conservation law ideally suited to study wave instability , and connect our intrinsic theory with bridges theory of multisymplectic structures ( see @xcite and @xcite ) .
section 4 is devoted to the discretization of second - order field theories .
we are able to use our general theory to produce numerical algorithms for nonlinear pde that are governed by second - order lagrangians which naturally respect a discrete multisymplectic form formula and a discrete noether theorem .
again , we demonstrate this methodology on the ch equation .
in this section , we review some aspects of multisymplectic geometry following @xcite , @xcite , @xcite , and @xcite .
let @xmath7 be an orientable ( @xmath4@xmath5@xmath6)-dimensional manifold ( which in applications is usually spacetime ) and let @xmath21 be a fiber bundle over @xmath7 .
sections @xmath22 of this _ covariant configuration bundle _ will be the physical fields . the space of sections of @xmath23 will be denoted @xmath24 or by @xmath25 .
the vertical bundle @xmath26 is the subbundle @xmath27 of @xmath28 , where @xmath29 denotes the tangent map of the @xmath23 . if @xmath7 has local coordinates @xmath30 , @xmath31 , adapted coordinates on @xmath0 are @xmath32 , @xmath33 , along the fibers @xmath34 , where @xmath35 and @xmath36 is the fiber dimension of @xmath0 .
@xmath9 denotes the @xmath37 jet bundle of @xmath0 , and this bundle may be defined inductively by @xmath38 .
recall that the first jet bundle @xmath39 is the affine bundle over @xmath0 whose fiber over @xmath40 consists of those linear mappings @xmath41 satisfying @xmath42 coordinates @xmath43 , @xmath44 on @xmath23 induce coordinates @xmath45 on the fibers of @xmath39 . given @xmath46 ,
its tangent map at @xmath35 , denoted @xmath47 , is an element of @xmath48 .
therefore , the map @xmath49 defines a section of @xmath39 regarded as a bundle over @xmath7 .
this section is denoted @xmath50 and is called the first jet of @xmath51 , or the first prolongation of @xmath51 . in coordinates , @xmath50 is given by @xmath52 where @xmath53 .
a section of the bundle @xmath54 which is the first prolongation of the section of @xmath55 is said to be holonomic .
the first jet bundle @xmath39 is the appropriate configuration bundle for first - order field theories , i.e. , field theories governed by lagrangians which only depend on the spacetime position , the field , and the first partial derivatives of the field .
herein , we shall focus on second - order field theories that are governed by lagrangians which additionally depend on the second partial derivatives of the fields ; thus , in second - order field theories , the lagrangian is defined on @xmath56 .
let us be more specific .
the second jet bundle is the affine bundle over @xmath39 whose fiber at @xmath57 consists of linear mappings @xmath58 satisfying @xmath59 one can define the second jet prolongation of a section @xmath60 , @xmath61 , as @xmath62 , that is a map @xmath63 , where @xmath50 is regarded as a section of @xmath39 over @xmath7 .
this map defines a section of @xmath64 regarded as a bundle over @xmath7 with @xmath65 being a linear map from @xmath66 into @xmath67 . in coordinates , @xmath61 is given by @xmath68 we shall also use the notation @xmath69 for second partial derivatives .
a section @xmath70 of @xmath71 is said to be @xmath11-holonomic if @xmath72 .
continuing inductively , one defines the @xmath37 jet prolongation of @xmath73 , @xmath74 , as @xmath75 .
consider a second - order lagrangian density defined as a fiber - preserving map @xmath76 , where @xmath77 is the bundle of @xmath78@xmath5@xmath79-forms on @xmath7 . in coordinates , we write @xmath80 where @xmath81 . for any @xmath82th - order lagrangian field theory , the fundamental geometric structure is the cartan form @xmath83 ; this is an @xmath84-form defined on @xmath15 ( see @xcite ) . for second - order field theories
, the cartan form is defined on @xmath1 , the covariant analogue of the phase space in mechanics .
the euler - lagrange equations may be written intrinsically as @xmath85 where @xmath86 denotes the interior product . traditionally , the cartan form is defined using the pull - back by the covariant legendre transform of the canonical multisymplectic @xmath4@xmath5@xmath6-form on the affine dual of @xmath15 ( see @xcite ) . in local coordinates , the cartan form on @xmath1 is given by @xmath87 where @xmath88 and @xmath89 , etc . for a @xmath37-order function @xmath90 , the _ formal _ partial derivative of @xmath91 in the direction @xmath92 , denoted by @xmath93 ,
is defined by @xmath94 for all @xmath95 , and is a smooth function on @xmath96 . in jet charts , @xmath97 in the next section
, we shall prove that the cartan form arises as the boundary term in the lagrangian variational principle . in this subsection
we show that a multisymplectic structure is obtained by taking the derivative of an action functional , and use this structure to prove the multisymplectic counterpart of the fact that in conservative mechanics , the flow of a mechanical system consists of symplectic maps .
let @xmath98 be a smooth manifold with ( piecewise ) smooth closed boundary .
define the set of smooth maps @xmath99 for each @xmath100 set @xmath101 and @xmath102 so that @xmath103 is a diffeomorphism .
let @xmath104 denote the closure of @xmath105 in some hilbert or banach space norm .
the choice of topology is not crucial in this paper , and one may assume all fields are smooth .
the tangent space to the manifold @xmath104 at a point @xmath106 is given by @xmath107 consider @xmath108 , the lie group of @xmath23-bundle automorphisms @xmath109 covering diffeomorphisms @xmath110 .
the group action @xmath111 is given by @xmath112 note that @xmath113 , and if @xmath114 , then @xmath115 .
the fundamental problem of the classical calculus of variations is to extremize the action functional over the space of sections of @xmath116 .
the @xmath117 @xmath118 is given by @xmath119 [ extremal ] @xmath120 is said to be * an extremum * of @xmath121 if @xmath122 for all smooth paths @xmath123 in @xmath108 , where for each @xmath124 , @xmath125 covers @xmath126 .
one may associate to each @xmath127 , the section of @xmath0 given by @xmath128 ; namely @xmath128 maps @xmath129 into @xmath130 .
if we choose the curve @xmath131 such that @xmath132 and @xmath133 , then we have that @xmath134 and @xmath135 .
this will be used in the following : @xmath136 where @xmath137 stands for the pull - back , and @xmath138 denotes the lie derivative .
now , let @xmath139 be the vertical subbundle ; this is the bundle over @xmath0 whose fibers are given by @xmath140 for each @xmath57 there exists a natural splitting @xmath141 . for example , for a vector @xmath142 , let @xmath143 , @xmath144 , and @xmath145 .
then @xmath146 on the other hand , by definition , @xmath147 .
therefore , @xmath148 which confirms that any vector @xmath149 may be decomposed into its horizontal component @xmath150 and its vertical component @xmath151 notice that @xmath152 for all @xmath153 , while @xmath154 and @xmath155 are vector fields on @xmath156 .
next , we define prolongations of automorphisms @xmath157 of @xmath0 and of elements @xmath158 .
given an automorphism @xmath157 of @xmath159 , its first prolongation @xmath160 is defined via @xmath161 if @xmath162 , then @xmath163 , with local coordinate expression @xmath164 to define the first prolongation of a vector @xmath158 , denoted @xmath165 , let @xmath166 be a flow of a vector field @xmath167 on @xmath0 with @xmath168 .
the first prolongation @xmath165 of @xmath169 is a vector field on @xmath39 given by @xmath170 if in a coordinate chart @xmath171 ; identifying @xmath169 with @xmath172 , we see that ( [ 1prolong ] ) yields the following local expression for @xmath173 : @xmath174 using induction , one can define the @xmath37 prolongation of an automorphism @xmath157 and the @xmath37 prolongation of a vector @xmath142 for all @xmath175 , and these will be denoted by @xmath176 and @xmath177 , respectively .
[ dervariational ] for a @xmath37-order function @xmath178 , the variational derivative of @xmath91 is the function on @xmath179 given by @xmath180 in particular , for a second order function @xmath181 , the variational derivative of @xmath91 is the function on @xmath182 given by @xmath183 let @xmath184 .
[ thm_main ] given a smooth lagrangian density @xmath185 , there exist a unique @xmath186 given by @xmath187 a unique map @xmath188 given by @xmath189 and a unique differential form @xmath190 given by @xmath191 such that @xmath192 for any @xmath106 , and the variation of the action functional @xmath121 is expressed by the following formula : for any @xmath158 and any open subset @xmath193 of @xmath7 such that @xmath194 , @xmath195\ , .\ ] ] furthermore , @xmath196\
\textit{in}\ u_x \ , \ ] ] where @xmath197 is the multisymplectic form on @xmath1 .
the variational principle ( [ var ] ) yields the euler - lagrange equations ( [ euler - lagrange ] ) on the interior of the domain , which in coordinates are given by @xmath198 while the form @xmath199 naturally arises in the boundary term and matches the definition of the cartan form given in ( [ cartan ] ) .
the proof proceeds in three steps .
we begin by computing the first variation using ( [ var1 ] ) .
then we show that the boundary term yields the cartan form .
lastly , we verify the statements related to the interior integral
. choose @xmath156 small enough so that it is contained in a coordinate chart .
if in these coordinates @xmath200 , then along @xmath201 , the coordinate expressions for @xmath202 are written as @xmath203 using the cartan formula we first compute the second term on the right - hand - side of ( [ var1 ] ) : @xmath204 using ( [ vertical ] ) , and the local expression for the vertical vector field @xmath155 , we have that @xmath205 \omega\nonumber\ , .\end{aligned}\ ] ] in the following , we shall use @xmath206 for the formal partial derivative of a function @xmath91 ( see ( [ derformal ] ) ) , and @xmath207 will denote @xmath208 . integrating ( [ var2 ] ) by parts , we obtain that @xmath209(v^v)^a\omega \\ & + \int_{u_x}\left(\frac{\partial l}{\partial y^a_\nu}(v^v)^a\right)_{,\nu } \omega + \int_{u_x}\left(\frac{\partial l}{\partial y^a_{\nu\mu } } ( v^v)^a_{,\nu}\right)_{,\mu}\omega \\ & -\int_{u_x}\left ( \frac{\partial}{\partial x^\mu}\left(\frac{\partial l } { \partial y^a_{\nu\mu}}\right)(v^v)^a\right)_{,\nu}\omega\ , .\end{aligned}\ ] ] using the fact @xmath210 , applying the stoke s formula @xmath211 , and combining the last calculation with ( [ easyterm ] ) , we obtain @xmath212(v^v)^a\omega \\ & + \int_{\partial u_x}\left(\frac{\partial l}{\partial y^a_\nu}-\frac{\partial } { \partial x^\mu}\left(\frac{\partial l } { \partial y^a_{\nu\mu}}\right)\right)(v^v)^a\omega_\nu + \frac{\partial
l } { \partial y^a_{\nu\mu}}(v^v)^a_{,\nu}\omega_\mu + l\ , v^\theta\omega_\theta \ , .\nonumber\end{aligned}\ ] ] a form @xmath213 on @xmath9 is contact , if @xmath214 for all @xmath215 .
[ boundary ] for a smooth lagrangian density @xmath185 there exist a unique differential form @xmath190 defined by ( [ cartanagain ] ) such that the boundary integral in ( [ var3 ] ) is equal to @xmath216\ , .\ ] ] furthermore , @xmath199 can be written as a sum of @xmath217 and a linear combination of a system of contact forms on @xmath64 with coefficients being functions on @xmath1 .
_ proof of lemma [ boundary ] _
let @xmath218 be an arbitrary vector field on @xmath1 , and let @xmath219 , a map from @xmath193 to @xmath1 .
then one computes @xmath220 using the formula @xmath221
one finds that @xmath222 similarly , @xmath223 thus if we let @xmath224 , use ( [ vprolong ] ) , and recall the local expression ( [ vert ] ) for @xmath225 , we obtain that @xmath226 next , observe that @xmath227 . also , @xmath228 these observations together with the previous identities imply the following important formulas @xmath229= ( v^v)^a_{,\nu}\,\omega_\mu\ , , \\ j^3(\phi\circ\phi_x^{-1})^\ast [ j^3(v)\intprod ( dy^a\wedge\omega_\nu - y^a_{\nu}\omega)]= ( v^v)^a\omega_\nu\ , . \end{array}\ ] ] substituting these formulas into the boundary integral of the variational principle ( [ var3 ] ) , we obtain that @xmath230
\\ = \int_{\partial u_x}j^3(\phi\circ\phi_x^{-1})^\ast [ j^3(v)\intprod & \left\ { \left ( \frac{\partial l}{\partial y^a_\nu}-d_\mu \left(\frac{\partial l}{\partial y^a_{\nu\mu } } \right ) \right)dy^a\wedge\omega_\nu + \frac{\partial l}{\partial y^a_{\nu\mu}}dy^a_\nu\wedge\omega_\mu \right .
\\ & + \left .
\left ( l-\frac{\partial l}{\partial y^a_\nu}y^a_\nu + d_\mu \left(\frac{\partial l}{\partial y^a_{\nu\mu } } \right)y^a_\nu - \frac{\partial l}{\partial y^a_{\nu\mu}}y^a_{\nu\mu } \right)\omega \right\ } \right ] \\
= \int_{\partial u_x}j^3(\phi\circ\phi_x^{-1})^\ast [ j^3(v)\intprod \theta_\mathcal{l } ] \ , .\end{aligned}\
] ] this proves the existence of a unique differential form @xmath231 and demonstrates how this form naturally arises in the boundary integral of the variational principle .
integration by parts yields the boundary integral with terms that involve partial derivatives of @xmath225 of all orders up to @xmath232 ( in our case @xmath233 ) . equation ( [ vv ] ) show that each partial derivative of @xmath225 has an associated @xmath84-form on @xmath64 , and substitution of these forms yields a unique differential @xmath84-form as desired .
since @xmath234 and its partial derivatives are functions on @xmath64 , then by ( [ derformal ] ) , @xmath235 is a function on @xmath1 , and therefore @xmath231 is a @xmath84-form on @xmath1 .
it is easy to show that @xmath236 for all integers @xmath237 and for all @xmath238 .
therefore , @xmath239 and @xmath240 are contact forms on @xmath64 .
hence the last statement of the lemma follows .
@xmath241 a simple computation then verifies that @xmath231 is the cartan form so that @xmath242 next , consider the interior integral of the variational principle ( [ var3 ] ) . since @xmath243 for all integers @xmath175
, we obtain that @xmath244(v^v)^a\omega \nonumber \\ & = \int_{u_x}j^4(\phi\circ\phi_x^{-1})^\ast\mathbf{i}_{j^4(v ) } \left [ \frac{\partial l}{\partial y^a } -d_\nu\left ( \frac{\partial l}{\partial y^a_\nu}\right)+d_\nu d_\mu\left ( \frac{\partial l}{\partial y^a_{\nu\mu}}\right)\right]dy^a\wedge\omega \nonumber \\ & = \int_{u_x}j^4(\phi\circ\phi_x^{-1})^\ast\mathbf{i}_{j^4(v ) } \left(\frac{\delta l}{\delta y^a}dy^a\wedge\omega\right)\ , , \end{aligned}\ ] ] where @xmath245 is the variational derivative of @xmath234 in the direction @xmath32 ( see definition [ dervariational ] ) .
since @xmath234 is a function of second order by hypothesis , then its variational derivative is a function on @xmath182 .
therefore the form @xmath246 is an @xmath247-form on @xmath182 .
moreover , the integrand in ( [ interior ] ) written as @xmath248 defines a unique smooth section @xmath249 as desired in the statement of the theorem .
now we shall prove the following [ criticalpoint ] the forms @xmath250 and @xmath251 satisfy the following relationship : @xmath252 for all @xmath106 and all vectors @xmath253 .
furthermore , a necessary condition for @xmath238 to be an extremum of the action functional @xmath121 is that @xmath254 for all vector fields @xmath255 on @xmath1 , which is equivalent to @xmath256 for all vector fields @xmath169 on @xmath182 .
_ proof of lemma [ criticalpoint]_. the proof will involve some lengthy computations that we partially present below . to compute @xmath257 ,
let us write @xmath199 as @xmath258 then , for @xmath259 , we obtain @xmath260(dy^a\wedge\omega_\nu - y^a_\nu\omega ) \\ & - d \left ( \frac{\partial l}{\partial y^a_\nu}-d_\mu \left(\frac{\partial l}{\partial y^a_{\nu\mu } } \right ) \right)\wedge ( w^a\omega_\nu - w^\theta dy^a\wedge\omega_{\nu\theta } - y^a_\nu w^\theta\omega_\theta ) \\ & + w \left [ \frac{\partial l}{\partial y^a_{\nu\mu } } \right ] ( dy^a_\nu\wedge\omega_\mu - y^a_{\nu\mu } \omega ) \\ & - d \left ( \frac{\partial l}{\partial y^a_{\nu\mu } } \right ) \wedge ( w^a_\nu\omega_\mu - w^\theta dy^a_\nu\wedge\omega_{\mu\theta } - y^a_{\nu\mu } w^\theta\omega_\theta ) \\
& + d_\mu \left ( \frac{\partial l}{\partial y^a_{\nu\mu } } \right ) ( w^a_\nu\omega - w^\theta dy^a_\nu\wedge\omega_\theta ) + \frac{\partial l}{\partial y^a}(w^a\omega - w^\theta dy^a\wedge\omega_\theta ) \ .\end{aligned}\ ] ] the last step is to pull back @xmath261 by @xmath262 ; this eliminates the terms with the contact forms .
in addition , using the fact that the pull back commutes with the exterior derivative , and applying formulas such as ( [ formula ] ) , we obtain that @xmath263{l } \left\ { d \left ( \frac{\partial l}{\partial y^a_\nu } - \frac{\partial } { \partial x^\mu } \left ( \frac{\partial l}{\partial y^a_{\nu\mu } } \right ) \right ) \wedge \left ( \frac{\partial(\phi\circ\phi_x^{-1})^a}{\partial x^\theta } \omega_\nu - \frac{\partial(\phi\circ\phi_x^{-1})^a } { \partial x^\nu } \omega_\theta \right . \right .
\\ \left .
+ \frac{\partial(\phi\circ\phi_x^{-1})^a } { \partial x^\nu } \omega_\theta \right ) + d \left ( \frac{\partial l}{\partial y^a_{\nu\mu } } \right ) \wedge \left ( \frac{\partial^2(\phi\circ\phi_x^{-1})^a}{\partial x^\theta \partial x^\nu}\omega_\mu - \frac{\partial^2 ( \phi\circ\phi_x^{-1})^a } { \partial x^\mu \partial x^\nu } \omega_\theta \right .
\\ \left .
+ \frac{\partial^2(\phi\circ\phi_x^{-1})^a } { \partial x^\mu \partial x^\nu } \omega_\theta \right ) \left .
- \frac{\partial}{\partial x^\mu } \left ( \frac{\partial l}{\partial y^a_{\nu\mu } } \right ) \frac{\partial^2(\phi\circ\phi_x^{-1})^a}{\partial x^\theta \partial x^\nu}\omega - \frac{\partial l}{\partial y^a } \frac{\partial(\phi\circ\phi_x^{-1})^a}{\partial x^\theta } \omega \right\ } \end{array}\\ & + & w^a_\nu \left\ { \frac{\partial}{\partial x^\mu } \left ( \frac{\partial l}{\partial y^a_{\nu\mu } } \right ) \omega - d \left ( \frac{\partial l}{\partial y^a_{\nu\mu } } \right ) \wedge\omega_\mu \right\ } \ .
\ ] ] some cancellation and further rearrangement yields @xmath264 letting @xmath224 , we have that @xmath265 where the right hand side equals @xmath266 by ( [ interior ] ) . hence , the relation ( [ psiomega ] ) is proved . a necessary condition for @xmath106 to be an extremum of the action
functional @xmath121 is that the interior integral in ( [ var3 ] ) vanish for all vectors @xmath253 . from the calculation
above , one may readily see that it is equivalent to the condition ( [ omegacondition ] ) . now if we let @xmath169 be a vector field on @xmath182 , then @xmath267 hence , @xmath268 thus , the condition @xmath269 for all vector fields @xmath169 on @xmath182 is equivalent to the condition ( [ omegacondition ] ) .
this completes the proof of the lemma .
lemma [ criticalpoint ] contains two equivalent conditions for @xmath106 to be extremal .
both conditions yield the same coordinate expression of the euler - lagrange equations given by @xmath270 which is the final statement of the theorem .
as one may see the proof we have presented can be generalized to lagrangian densities on @xmath9 .
one has to modify the labeling of variables to reflect the general case .
for example , @xmath271\ , \ ] ] where @xmath272 .
then the cartan form shall arise in the boundary integral as a linear combination of the forms above .
we shall call critical points @xmath73 of @xmath121 , solutions of the euler - lagrange equations .
we let @xmath273 denote the space of solutions of the euler - lagrange equations .
we are now ready to prove the multisymplectic form formula , a covariant generalization of the symplectic flow theorem to second - order field theories .
if @xmath131 is a smooth curve of solutions of the euler - lagrange equations in @xmath274 ( when such solutions exist ) , then differentiating with respect to @xmath124 at @xmath275 will give a tangent vector @xmath169 to the curve at @xmath276 . by differentiating @xmath277=0 $ ] , we obtain @xmath278=0\ ] ] for all vector fields @xmath255 on @xmath1
. therefore if @xmath274 is a submanifold of @xmath104 , then for any @xmath279 we may identify @xmath280 with the set of vectors @xmath169 that satisfy the above condition .
however , we do not require @xmath274 to be a submanifold . for any @xmath281 , @xmath282=0
\quad \mbox{for all vector fields } v\ \mbox{on}\ j^3y\}\ ] ] defines a set of solutions of the first variation equations of the euler - lagrange equations . *
( multisymplectic form formula)*[thm_mff ] if @xmath279 , then for all @xmath169 and @xmath255 in @xmath283 , @xmath284=0.\ ] ] we follow theorem 4.1 in @xcite .
define the @xmath6-forms @xmath285 and @xmath286 on @xmath104 by @xmath287\ ] ] and @xmath288\ , \ ] ] so that by ( [ var ] ) and ( [ 211 ] ) , @xmath289 furthermore , @xmath290 since @xmath291 , we have that @xmath292 given vectors @xmath293 we may extend them to vector fields @xmath294 on @xmath104 by fixing vector fields @xmath295 on @xmath0 such that @xmath296 and @xmath297 , and letting @xmath298 and @xmath299 . if @xmath166 covering @xmath300 is the flow of @xmath167 , then @xmath301 is the flow of @xmath302 .
notice that @xmath303 and @xmath304 , hence the equation ( [ star ] ) becomes @xmath305 recall that for any @xmath6-form @xmath306 on @xmath104 and vector fields @xmath294 on @xmath104 , @xmath307 - \mathcal{w}[\alpha ( \mathcal{v})]-\alpha([\mathcal{v } , \mathcal{w}])\ .\ ] ] also recall that for a vector field @xmath302 on @xmath104 and a function @xmath91 on @xmath104 , @xmath308=df\cdot\mathcal{v}$ ] .
we now use the latter and ( [ bracket ] ) on @xmath286 .
we have that @xmath309(\phi ) - \mathcal{w}[\alpha_2(\mathcal{v})](\phi ) - \alpha_2([\mathcal{v } , \mathcal{w}])(\phi ) \nonumber \\ & = [ d(\alpha_2(\mathcal{w}))\cdot\mathcal{v}](\phi ) - [ d(\alpha_2 ( \mathcal{v}))\cdot\mathcal{w}](\phi ) - \alpha_2(\phi ) \cdot [ v , w ] \nonumber \\ & = d(\alpha_2(\mathcal{w}))(\phi ) \cdot v - d(\alpha_2 ( \mathcal{v}))(\phi ) \cdot w - \alpha_2(\phi ) \cdot [ v , w]\ .\end{aligned}\ ] ] similarly , @xmath310\ .\end{aligned}\ ] ] let @xmath279 and @xmath311 be a curve in @xmath104 through @xmath73 such that @xmath312 now we restrict @xmath313 to @xmath283 .
we shall give a detailed computation of the first term on the right hand side of ( [ doublestar ] ) .
we have that @xmath314\ \\ & = \left.\frac{d}{d\lambda}\right|_{\scriptscriptstyle \lambda = 0 } \int_{\partial u_x}j^3(\phi\circ\phi_x^{-1})^\ast j^3(\eta_y^\lambda ) ^\ast [ j^3(w)\intprod \theta_{\mathcal{l } } ] \\ & = \int_{\partial u_x}j^3(\phi\circ\phi_x^{-1})^\ast \pounds_{j^3(v ) } ( j^3(w)\intprod \theta_{\mathcal{l}})\\ & = \int_{\partial u_x}j^3(\phi\circ\phi_x^{-1})^\ast d[j^3(v)\intprod j^3(w)\intprod \theta_{\mathcal{l}}]\\ & + \int_{\partial u_x}j^3(\phi\circ\phi_x^{-1})^\ast [ j^3(v)\intprod d(j^3(w)\intprod \theta_{\mathcal{l}})]\ , \end{aligned}\ ] ] where the last equality was obtained using cartan s formula .
we have also used the fact that @xmath315 and @xmath297 have the same @xmath82-th prolongation .
furthermore , using stoke s theorem , noting that @xmath316 is empty , and applying cartan s formula once again to @xmath317 , we obtain that @xmath318 \nonumber \\ & -\int_{\partial u_x}j^3(\phi\circ\phi_x^{-1})^\ast [ j^3(v)\intprod j^3(w)\intprod \omega_{\mathcal{l}}]\ .\end{aligned}\ ] ] similarly , @xmath319 \nonumber \\ & -\int_{\partial u_x}j^3(\phi\circ\phi_x^{-1})^\ast [ j^3(w)\intprod j^3(v)\intprod \omega_{\mathcal{l}}]\ .\end{aligned}\ ] ] now , @xmath320)= [ j^3(v ) , j^3(w)]$ ] ; hence , @xmath321=
\int_{\partial u_x}j^3(\phi\circ\phi_x^{-1 } ) ^\ast ( [ j^3(v ) , j^3(w)]\intprod\theta_{\mathcal{l}})\ .\ ] ] recall that for a differential form @xmath213 on a manifold @xmath322 and for vector fields @xmath323 on @xmath322 , @xmath324}\alpha = \pounds_x\mathbf{i}_y\alpha - \mathbf{i}_y \pounds_x\alpha\ .\ ] ] therefore , @xmath325 & = \int_{\partial u_x}j^3(\phi\circ\phi_x^{-1 } ) ^\ast [ \pounds_{j^3(v)}(j^3(w)\intprod\theta_{\mathcal{l}})- j^3(w)\intprod\pounds_{j^3(v)}\theta_{\mathcal{l}}]\nonumber
\\ = \int_{\partial u_x}j^3(\phi\circ & \phi_x^{-1})^\ast [ j^3(v)\intprod \pounds_{j^3(w)}\theta_{\mathcal{l}}-j^3(v)\intprod j^3(w ) \intprod\omega_{\mathcal{l}}-j^3(w)\intprod\pounds_{j^3(v)}\theta_{\mathcal{l}}],\end{aligned}\ ] ] where we have again used stoke s theorem and cartan s formula twice . substituting ( [ one ] ) ,
( [ two ] ) , and ( [ three ] ) into ( [ doublestar ] ) , we obtain that @xmath326\ .\ ] ] we now compute ( [ triplestar ] ) .
similar computations as above yield @xmath327 which vanishes for all @xmath279 and @xmath328
. similarly , @xmath329 for all @xmath330 and @xmath331 . finally , @xmath332 for all @xmath279 .
therefore , the equation ( [ triplestar ] ) vanishes for all @xmath279 and @xmath333 .
using the latter and ( [ four ] ) , equation ( [ star ] ) becomes @xmath334=0\ ] ] for all @xmath279 and all @xmath335 , as desired .
suppose that @xmath121 is invariant under the action @xmath336 of a lie group @xmath108 on @xmath104 .
this implies that for each @xmath337 , @xmath338 whenever @xmath281 .
we restrict the action to elements of @xmath274 .
for each element @xmath339 of the lie algebra @xmath340 of @xmath108 , let @xmath341 be the corresponding infinitesimal generator on @xmath104 restricted to elements of @xmath274 . by the invariance of @xmath121 , @xmath342 differentiating with respect to @xmath343 at @xmath344 , and using the fundamental property of the cartan form that @xmath345 , we find that @xmath346 then by theorem [ thm_main ] and the invariance of @xmath121 we have that @xmath347\nonumber \\ & = - \int_{u_x}j^3(\phi\circ\phi_x^{-1})^\ast[j^3(\xi_\mathcal{c}(\phi ) ) \intprod \omega_\mathcal{l}]\ , .\end{aligned}\ ] ] let @xmath348 satisfy @xmath349\ ] ] for all @xmath350 and @xmath351
. then the map @xmath352 defined by @xmath353 is the * covariant momentum map of the action*. with this definition , ( [ 101 ] ) becomes @xmath354 = 0 $ ] , and since this holds for any @xmath355 , the integrand must also vanish ; thus , @xmath356 = 0\ , .\ ] ] on the other hand , by stoke s theorem we may also conclude that @xmath357 last two statements are equivalent , and we refer to them as the covariant noether theorem .
the completely integrable bi - hamiltonian camassa - holm ( ch ) equation @xmath358 is a model for breaking shallow water waves that admits peaked solitary traveling waves as solutions ( see @xcite , @xcite ) .
such solutions , termed _ peakons _
, develop from any initial data with sufficiently negative slope , and because of the discontinuities in the first derivative , these solutions are difficult to numerically simulate , particularly in the case of a _ peakon antipeakon collision _ ( see @xcite ) .
the multisymplectic framework for the ch equation is intended to provide a foundation for numerical discretization schemes that preserve the hamiltonian structure of this model , even at the discrete level . after developing the multisymplectic framework for ( [ ch ] )
, we shall follow @xcite and develop the entire discrete multisymplectic approach to second - order field theories , concentrating on the discrete ch equation as our model problem . although we shall only produce the simplest _ multisymplectic - momentum _ conserving algorithm for this equation , our construction is completely general and
will allow for the creation of @xmath82th - order accurate schemes for arbitrarily large @xmath82 .
the ch equation ( [ ch ] ) is usually expressed in terms of the eulerian , or spatial velocity field @xmath359 , and is the euler - poincar equation for the reduced lagrangian @xmath360 alternatively , one may express ( [ ch ] ) in terms of the lagrangian variable @xmath361 arising from the solution of @xmath362 the lagrangian approach to the ch equation is ideally suited to the multisymplectic variational theory , and we begin by specifying our fiber bundle @xmath363 .
let @xmath364 , and @xmath365 .
we coordinatize @xmath7 by @xmath366 ( or @xmath367 ) and @xmath0 by @xmath368 ( or @xmath369 ) .
a smooth section @xmath215 represents a physical field and is expressed in local coordinates by @xmath370 , where @xmath371 is the lagrangian flow solving ( [ flow ] ) .
the material or lagrangian velocity @xmath372 is an element of @xmath373 , where @xmath374 . using ( [ flow ] ) together with @xmath375 , the lagrangian representation for the action may be expressed as @xmath376 the second jet bundle @xmath64 is a nine - dimensional manifold and @xmath11-holonomic sections of @xmath71 have local coordinates @xmath377 where for smooth sections @xmath378 .
the lagrangian density @xmath379 is expressed as @xmath380 for the camassa - holm equation the lagrangian density evaluated along the second jet of a section @xmath51 is given by @xmath381 dx\wedge dt \ , .\ ] ] as our lagrangian ( [ density ] ) depends only on @xmath382 , and @xmath383 , the euler - lagrange equation ( [ coordinateel ] ) simply becomes @xmath384 so that we have the lagrangian version of the ch equation ( [ ch ] ) given by @xmath385 by differentiating @xmath386 three times , one may verify that ( [ elch ] ) is indeed equivalent to ( [ ch ] ) .
now , using ( [ cartanagain ] ) we have that the cartan form @xmath231 is given by @xmath387 or @xmath388 if written in terms of the system of contact forms .
marsden , patrick , and shkoller in their @xcite paper have demonstrated how the multisymplectic form formula for first - order field theories when applied to nonlinear wave equations generalizes the notion of symplecticity given by bridges in @xcite .
using the example of the ch equation , we present below a simple interpretation of the multisymplectic form formula for the second - order field theories .
we show that the mff formula is an intrinsic generalization of the conservation law analogous to the one in appendix d of @xcite .
bridges has introduced the notion of a hamiltonian system on a multi - symplectic structure .
a multi - symplectic structure @xmath389 consists of a manifold @xmath390 , the phase space , and a family of pre - symplectic forms .
the phase space @xmath390 is a manifold modeled on @xmath391 . a hamiltonian system on a multi - symplectic structure
is then represented symbolically by @xmath392 with governing equation @xmath393 for all vector fields @xmath167 on @xmath390 where @xmath394 is an inner product on @xmath395 and @xmath396 is a curve in @xmath390 .
bridges has shown that this formulation is natural for studying wave propagation in open systems .
bridges , in particular , has obtained the following conservation law in the case of the wave equation : @xmath397 this law generalizes the notion of symplecticity of classical mechanics .
let us make an appropriate choice of the phase space @xmath390 for the camassa - holm equation .
our choice is entirely governed by the coefficients in the cartan form ( [ chcartan ] ) .
since the lagrangian ( [ density ] ) does not explicitly depend on time and space variables , that is the system is autonomous , we identify sections @xmath73 of @xmath0 with mappings @xmath398 from @xmath399 into @xmath400 , and similarly , sections of @xmath1 with mappings from @xmath399 into @xmath401 .
the cartan form ( [ chcartan ] ) suggests to introduce the following momenta : @xmath402 since @xmath231 is horizontal over @xmath39 , the covariant configuration bundle is really @xmath54 , and one should think of @xmath403 as field variables with each field variable having conjugate multi - momenta .
for example , @xmath404 function as conjugate spatial and temporal momenta for the field component @xmath371 .
then the transformation @xmath405 defines a mapping from the space of vertical sections of @xmath406 into the phase space @xmath407 modeled over @xmath408 .
we denote this transformation by @xmath409 .
let us now state the result that connects our paper to bridges theory : the multisymplectic form formula ( mff ) yields a multi - symplectic structure @xmath410 such that the mff formula becomes an intrinsic generalization of the following conservation law : for any @xmath313 in @xmath283 that are @xmath23-vertical @xmath411 moreover , the ch equation in both eulerian form ( [ ch ] ) and lagrangian form ( [ elch ] ) is equivalent to the hamiltonian system of equations on the multi - symplectic structure with the hamiltonian defined by @xmath412 consider two @xmath23-vertical vectors @xmath169 and @xmath255 in @xmath283 . then @xmath413 and @xmath414 are vertical - over-@xmath7 vector fields on @xmath415 whose components we shall denote via @xmath416 or just numerate by @xmath417 .
thinking of the components of the transformation @xmath409 as functions on @xmath1 , we immediately see that @xmath418\\ v^{p^t}&=dp^t\cdot j^3(v)\equiv j^3(v)[p^t]\\ v^{p^{tx}}&=dp^{tx}\cdot j^3(v)\equiv j^3(v)[p^{tx } ] \end{array}\ ] ] using expressions ( [ legendre ] ) we express @xmath419 as : @xmath420 next , combining with ( [ 111 ] ) we obtain that @xmath421 so that @xmath422 = \\ & \int_{\partial u_x}(v^4w^1- w^4v^1 + v^6w^3-w^6v^3)dt - ( v^5w^1-w^5v^1)dx\ .\nonumber\end{aligned}\ ] ] the integral on the right - hand - side of the above equation leads us to introduce two degenerate skew- symmetric matrices @xmath423 ,
@xmath424 on @xmath425 : @xmath426 \quad b_0=\left [ \begin{array}{cccccc } 0&0&0&0&1&0\\ 0&0&0&0&0&0\\ 0&0&0&0&0&0\\ 0&0&0&0&0&0\\ -1&0&0&0&0&0\\ 0&0&0&0&0&0 \end{array}\right]\ .\ ] ] to each matrix @xmath427 , we associate the @xmath11-form @xmath428 on @xmath425 given by @xmath429 , where @xmath430 . with the definition of @xmath428 and the use of ( [ 110 ] ) , the multisymplectic form formula ( [ mff ] ) becomes , for @xmath355 , @xmath431 hence by the stoke s theorem , @xmath432 dx\wedge dt = 0\ .\ ] ] since @xmath193 is arbitrary , we obtain the desired conservation law ( [ wavelawgeneral ] ) . in the special case ,
when the components @xmath433 and @xmath434 , one may verify that @xmath435 so that , letting @xmath436 denote an element @xmath437 , the formula ( [ wavelawgeneral ] ) takes the special form @xmath438 which is the complete analogue of bridges conservation law ( [ wavelaw ] ) for the wave equation .
next , since the inner product @xmath439 is independent of @xmath440 , the hamiltonian system of equations ( [ bridges ] ) on the multi - symplectic structure @xmath410 may be written as @xmath441 which results in @xmath442 with the choice of the hamiltonian ( [ hamiltonian ] ) the last four equations yield identities , and the first equation becomes @xmath443 using the legendre transformation expressions ( [ legendre ] ) for @xmath444 , the latter equation recovers the euler - lagrange equation ( [ generall ] ) , hence ( [ elch ] ) .
in other words , the euler - lagrange equations on @xmath1 are equivalent to hamilton s equations on the multi- symplectic structure @xmath445 .
we shall now generalize the veselov - type discretization of first - order field theory given in @xcite to second - order field theories , using the camassa - holm equation as our example .
we discretize @xmath7 by @xmath446 and the fiber bundle @xmath0 by @xmath447 .
elements of y over the base point @xmath448 are written as @xmath449 and the projection @xmath23 acts on @xmath0 by @xmath450 .
the fiber over @xmath451 is denoted @xmath452 . for the general case of a second - order lagrangian one must define the discrete second jet bundle of @xmath0 , and
this discretization depends on how one chooses to approximate the partial derivatives of the field .
for example , using central differencing and a fixed timestep @xmath82 and spacestep @xmath453 , we have that @xmath454 where @xmath455 and @xmath456 form a uniform grid in continuous spacetime .
-0.5 in 4.in 4.in 1.25 in -0.5 in we observe that an @xmath457-tuple @xmath458 is sufficient to approximate @xmath459 , where @xmath460 is in the center of the cell @xmath461{ccc } ( i-1 , j-1 ) , & ( i-1 , j ) , & ( i-1 , j+1),\\ ( i , j-1 ) , & ( i , j ) , & ( i , j+1),\\ ( i+1 , j-1 ) , & ( i+1 , j ) , & ( i+1 , j+1)\ ) \ , .
\end{array}\end{aligned}\ ] ] let @xmath462 denote the set of cells , i.e. @xmath463 . components of a cell are called vertices , and are numbered from first to ninth .
a point @xmath451 is * touched * by a cell if it is a vertex of that cell .
if @xmath464 , then @xmath451 is an * interior point * of @xmath98 if @xmath98 contains all cells touching @xmath448 .
the * interior * @xmath465 of @xmath98 is the set of all interior points of @xmath98 .
the * closure * @xmath466 of @xmath98 is the union of all cells touching interior points of @xmath98 .
a * boundary point * of @xmath98 is a point in @xmath98 and @xmath466 which is not an interior point .
the * boundary * of @xmath98 is the set of boundary points , so that @xmath467 . a * section * of the configuration bundle
@xmath55 is a map @xmath468 such that @xmath469 .
we are now ready to define the discrete multisymplectic phase space .
the discrete * second jet bundle * of @xmath0 is given by @xmath470 the fiber over @xmath451 is denoted @xmath471 .
we define the * second jet extension * of a section @xmath73 to be the map @xmath472 given by @xmath473{ccc } \phi ( i-1 , j-1 ) , & \phi ( i-1 , j ) , & \phi ( i-1 , j+1),\\ \phi ( i , j-1 ) , & \phi ( i , j ) , & \phi ( i , j+1),\\ \phi ( i+1 , j-1 ) , & \phi ( i+1 , j ) , & \phi ( i+1 , j+1)\ ) \ , .
\end{array}\end{aligned}\ ] ] given a vector field @xmath167 on @xmath0 the * second jet extension * of @xmath167 is the vector field @xmath474 on @xmath64 defined by @xmath475{ccc}(v(y_{i-1j-1}),&\ , v(y_{i-1j}),&\ , v(y_{i-1j+1}),\ , \\
v(y_{ij-1}),&\ , v(y_{ij}),&\ , v(y_{ij+1}),\ , \\
v(y_{i+1j-1}),&\ , v(y_{i+1j}),&\ , v(y_{i+1j+1}))\ , .
\end{array}\end{aligned}\ ] ] of course , this may easily be generalized to more accurate differencing schemes that require more than nine grid points to define second partial derivatives . restricting our attention to the ch equation and noting that its lagrangian depends only on @xmath476 , we may significantly simply our discretization of the second jet bundle @xmath64
; this will substantially reduce our calculations and simplify the exposition . to approximate @xmath459 we choose the forward difference evaluations of @xmath476 : @xmath477 for this particular choice
, our cell reduces to a rectangle . a * rectangle * @xmath241 of @xmath7 is an ordered @xmath478-tuple of the form @xmath479 for each rectangle , @xmath480 stand for the first , second , third , and fourth vertices respectively .
if @xmath481 is the first vertex , we shall denote the rectangle by @xmath482 .
the set of all rectangles in @xmath7 is denoted by @xmath483 .
the set - theoretical definitions of subsection [ general ] apply here . for example
, a point @xmath484 is * touched * ( see fig .
4.2 ) by four rectangles @xmath485 , etc .
3.in 3.in 0.25 in -0.5 in again as ( [ density ] ) does not depend on @xmath486 , we may restrict ourselves to a subbundle @xmath487 of the continuous @xmath64 defined via @xmath488 .
then the discrete analogue @xmath489 ( see fig .
4.3 ) of @xmath487 is identified with @xmath490 -0.1 in 6.in 4.in 1.25 in -0.5 in for a section @xmath491 , we define the * second jet extension * of @xmath51 to @xmath489 to be the map @xmath492 via @xmath493 given a vector field @xmath167 on @xmath0 we extend it to a vector field @xmath474 on @xmath489 by @xmath494 a * discrete lagrangian * on @xmath489 is then a function @xmath495 of five variables @xmath496 , where the @xmath497-variables are labeled in the order they appear in a @xmath478-tuple .
let @xmath98 be a * regular * subset of @xmath7 , i.e. @xmath98 is exactly the union of its interior and boundary .
let @xmath498 denote the set of sections of @xmath0 on @xmath98 , so @xmath498 is the manifold @xmath499 .
the * discrete action * is a real valued function on @xmath498 defined by the rule @xmath500 given a section @xmath51 on @xmath98 acting as @xmath501 , one can define an element @xmath502 to be a map @xmath503 acting as @xmath504 , where @xmath505 is thought as a vector emanating from @xmath506 .
given an element @xmath502 one can always extend it to a vector field @xmath167 on @xmath0 . on the other hand , given a vector field @xmath167 on @xmath0
, @xmath507 is an element of @xmath508 .
thus , it is sufficient to work with vector fields @xmath167 on @xmath0 alone .
if @xmath167 is a vector field on @xmath0 , consider its restriction @xmath509 to the fiber @xmath452 .
let @xmath510 be the flow of @xmath509 .
then by definition of the flow , @xmath511 therefore there is the @xmath6-parameter family of sections on @xmath98 defined by @xmath512 such that @xmath513 and @xmath514 .
thus , the * variational principle * is to seek those sections @xmath51 for which @xmath515 for all vector fields @xmath167 on @xmath0 . with our choice of @xmath489 , the discrete lagrangian for the camassa - holm equation
is @xmath516 the variational principle yields the * discrete euler - lagrange field equations * ( del equations ) as follows .
choose an arbitrary point @xmath517 .
henceforth , with a slight abuse of notation , we shall write @xmath449 for @xmath518 .
the action ( [ actionsum ] ) , written with its summands containing @xmath449 explicitly , is ( see figure 3.2 ) @xmath519 differentiating with respect to @xmath449 yields the del equations : @xmath520 for all @xmath521 .
equivalently , these equations may be written as @xmath522 for all @xmath521 .
computing and evaluating @xmath523 along rectangles touching an interior point @xmath448 , and substituting these expressions into ( [ delf ] ) , we obtain the discrete euler - lagrange equations for the ch equation : @xmath524 where @xmath525 to see that ( [ delfch ] ) are indeed approximating the continuous euler - lagrange equation ( [ elch ] ) , notice that the first two terms combine to approximate @xmath526\ , .\end{aligned}\ ] ] as to the third and fourth terms of ( [ delfch ] ) , @xmath527\ , .\ ] ] next , the fifth , sixth , seventh , and eighth terms combine as @xmath528\ , .\end{aligned}\ ] ] finally , the last two terms of ( [ delfch ] ) approximate @xmath529 the numerical scheme ( [ delfch ] ) proceeds as follows : suppose that @xmath530 as known from the two previous time steps ; then ( [ delfch ] ) may be written as @xmath531 these are implicit equations which must be solved for @xmath532 , where @xmath36 is the size of the spatial grid .
we consider arbitrary variations which are in no way constrained on the boundary @xmath533 .
for each @xmath534 there is at least one rectangle in @xmath98 touching @xmath448 since @xmath535 and @xmath98 is regular . on the other hand ,
not all four rectangles touching @xmath448 are in @xmath98 since @xmath536 .
therefore , each @xmath537 occurs as the @xmath538 vertex for either one , two , or three of the @xmath539 and the corresponding @xmath538 boundary expressions are given by @xmath540 where @xmath506 .
the sum of all such terms is the contribution to @xmath541 from the boundary @xmath533 .
we thus define the _ four _ @xmath6-forms on @xmath542 by @xmath543 we regard the @xmath478-tuple @xmath544 as being the discrete analogue of the multisymplectic form @xmath231 . given a vector field @xmath167 on @xmath0 such that @xmath545 , the first expression from the list ( [ boundarylist ] ) becomes @xmath546((i , j))$ ] , the others written similarly . with this notation , @xmath541 may be expressed as @xmath547(\square^1 ) \right ) \nonumber \\ & + \sum_{(i , j)\in\partial u } \left ( \sum_{\stackrel{\square\subseteq u ; l;}{(i , j)=\square^l } } [ ( j^2\phi ) ^\ast(j^2v\intprod\theta_l^l)](\square^1 ) \right)\ , .\end{aligned}\ ] ] for a rectangle @xmath241 in @xmath7 , define the projection @xmath548 by @xmath549 calculating the form @xmath550 on @xmath498 gives @xmath551 this immediately implies that the variation ( [ dds ] ) can be written as @xmath552 define the @xmath6-forms @xmath285 and @xmath286 on the space of sections @xmath498 to be the first and the second terms on the right hand side of ( [ dss ] ) , respectively . as in the previous section
we would like to derive the discrete analogue of symplecticity of the flow in mechanics .
let @xmath131 be a curve of solutions of ( [ delf ] ) that passes through @xmath73 at zero with @xmath553 . then for each interior point @xmath448 and each @xmath124 , the following holds : @xmath554 differentiating these equations with respect to @xmath124 at @xmath555
, we obtain the if @xmath73 is a solution of the discrete euler - lagrange equations ( [ delf ] ) , then a * first - variation * equation solution at @xmath73 is a vector @xmath556 such that for each @xmath557 , @xmath558 by definition of the forms @xmath285 and @xmath286 , @xmath559 . since @xmath291 , @xmath560 . using ( [ dss ] ) and denoting the vertices of @xmath241 by @xmath561 , we have that @xmath562 , which implies that for all @xmath563 @xmath564 therefore , @xmath565 substitution of ( [ pullbackomega ] ) into the exterior derivative of the right hand side of ( [ dss ] ) yields @xmath566 when specialized to two first - variation solutions @xmath169 and @xmath255 at @xmath51 , @xmath567 vanishes , because for each interior point @xmath448 all four rectangles touching it are contained in @xmath98 , and @xmath568 and @xmath569 .
therefore , @xmath570 and the equation @xmath291 becomes @xmath571 , which in turn is equivalent to @xmath572(\square^1 ) \right ) = 0\ ] ] for all vector fields @xmath295 on @xmath0 .
this is the discrete analogue of the multisymplectic form formula for the continuous spacetime .
we observe that @xmath573 , which shows that @xmath574 which in turn implies that only three of the @xmath11-forms @xmath575 , @xmath563 are in fact independent .
in addition , this implies that for a _ given and fixed _
rectangle @xmath576 , @xmath577 for all sections @xmath73 and all vectors @xmath313 .
we would like to derive the discrete version of the noether theorem for second order field theories .
this is not the most general form possible as we are working with a particular example .
however it is such as to facilitate the derivation of any other case without significant effort .
suppose that a lie group @xmath108 with a lie algebra @xmath340 acts on @xmath0 by vertical symmetries such that the lagrangian @xmath234 is invariant under the action .
vertical action simply means that the base elements from @xmath7 are not altered under the action , hence the action restricts to each fiber of @xmath0 .
let @xmath578 denote the action of @xmath108 on @xmath0 .
for every @xmath337 let @xmath579 be given by @xmath580 .
we also use the notation @xmath581 for the action .
then there is an induced action of @xmath108 on @xmath489 defined in a natural way : @xmath582 recall that the infinitesimal generator of an action ( of a lie group @xmath108 on a manifold @xmath322 ) corresponding to a lie algebra element @xmath350 is the vector field @xmath583 on @xmath322 obtained by differentiating the action with respect to @xmath584 at the identity in the direction @xmath339 . by the chain rule , @xmath585\ , , \ ] ] where @xmath586 is the lie algebra exponential map . using this formula
, we immediately see that @xmath587 the invariance of the lagrangian under the action implies that @xmath588 which , for a given @xmath241 , is equivalent to @xmath589 for all @xmath590 and all @xmath591 . for each @xmath592 , let us denote by @xmath593 the projection onto the @xmath538 component . using this projection the four components of the infinitesimal generator @xmath594 are expressed as @xmath595 hence , the equation ( [ linvariance ] ) becomes @xmath596 we observe that for each @xmath592 , @xmath597 is a function on @xmath489 which we denote by @xmath598 .
notice that @xmath599 is the discrete multisymplectic analogue of @xmath600 in classical mechanics so that @xmath583 is the global hamiltonian vector field of @xmath601 .
many symmetry groups act by special canonical transformations , i.e. @xmath602 , in which case @xmath603 . in such case ,
@xmath601 is uniquely defined . since @xmath604 is linear in @xmath339 , so are the functions @xmath605 , and we can replace the lie group action by a lie algebra action @xmath606 . finally , we are ready to define the momentum maps
. there are four @xmath607-valued * momentum mappings * @xmath608 on @xmath489 defined by @xmath609 for all @xmath350 and @xmath610 , where @xmath439 is the duality pairing .
the equation ( [ linvariance2 ] ) implies that @xmath611 so , as in the case of the lagrangian @xmath11-forms , only three of the four momenta are essentially distinct .
the discrete version of the noether theorem for second - order field theories now follows .
define the action of the lie group @xmath108 on @xmath498 by @xmath612 since the lagrangian is @xmath108-invariant , then @xmath613 once again letting @xmath614 and differentiating with respect to @xmath343 at @xmath344 , we obtain that @xmath615 for all @xmath616 .
one can readily verify that @xmath617 , which is an element of @xmath618 .
thus , @xmath619 for all @xmath350 and @xmath616 .
since @xmath121 is @xmath108-invariant , then @xmath108 sends critical points of @xmath121 to themselves , or in other words , the action restricts to the space of solutions of the euler - lagrange equations . therefore , if @xmath73 is a solution , so is @xmath620 , where @xmath621 and @xmath622 . substituting @xmath623 into the discrete euler - lagrange equations and differentiating with respect to @xmath343 at @xmath344
, we obtain that for any @xmath339 and @xmath73 , @xmath624 is a first - variation equation solution . using ( [ dds ] ) , ( [ sinvariance ] ) becomes @xmath625 for all @xmath51 from the solution space and all @xmath626 .
thus , the discrete version of the noether theorem is @xmath627(\square^1 ) \right ) = 0\ ] ] for all @xmath73 from the solution space .
the authors would like to thank the center for nonlinear studies at los alamos , where most of this work was completed , for providing a wonderful working environment .
ss was partially supported by nsf - kdi grant atm-98 - 73133 and the doe . , _ a multisymplectic framework for classical field theory and the calculus of variations i : covariant hamiltonian formulation _ , mechanics , analysis and geometry : 200 years after lagrange , m. francaviglia , ed . , ( north holland ) , 203235 . | this paper presents a geometric - variational approach to continuous and discrete _ second - order _ field theories following the methodology of @xcite . staying entirely in the lagrangian framework and letting @xmath0
denote the configuration fiber bundle , we show that both the multisymplectic structure on @xmath1 as well as the noether theorem arise from the first variation of the action function .
we generalize the multisymplectic form formula derived for first order field theories in @xcite , to the case of second - order field theories , and we apply our theory to the camassa - holm ( ch ) equation in both the continuous and discrete settings .
our discretization produces a multisymplectic - momentum integrator , a generalization of the moser - veselov rigid body algorithm to the setting of nonlinear pdes with second order lagrangians . |
interest in the relativistic harmonic oscillator has either focused on its mathematical and physical characteristics @xcite or on analytic approaches of solving its non - linear equation of motion , given non - dimensionally as @xmath4 the latter is motivated by the general significance of the study of strongly non - linear oscillators , particularly in developing efficient approximation methods to treat them @xcite .
recently , the non - perturbative methods of harmonic balance @xcite and homotopy @xcite have been succesfully applied to give analytic approximations of the oscillation period and periodic solution of eq.[harmonic ] . unlike traditional perturbation theory
, these asymptotic techniques do not rely on any linear term or physically small parameter .
hence their results are unrestricted by a radius of convergence .
for both methods , the approximations had minimal error over the entire domain of the particle s maximum velocity , @xmath5 . , where eq.[harmonic ] reduces to the linear oscillator .
while in the limit @xmath6 , the oscillator is energetic enough to reach relativistic speeds , which describes a highly non - linear system .
] however , the applicability of these approaches has thus far been restricted to the linear potential . in this work
we contribute a novel method for computing analytic approximations of the period of strongly anharmonic oscillations described by @xmath7 for an even polynomial potential @xmath8 of general form .
we adopt a non - perturbative approach that has been developed and applied by amore et al . in computing periods of non - linear oscillators @xcite and formally equivalent systems such as the deflection angle in gravitational lensing @xcite .
the method operates by converting the relevant integral into a series carrying a variational parameter @xmath9 , which can be chosen using the principle of minimal sensitivity ( pms ) @xcite . in our case the standard procedure for choosing @xmath9 is modified by treating the existing harmonics as free parameters . these are then assigned values by imposing the ultrarelativistic limit of the period as a boundary condition
. although the necessary computations are simple , the resulting approximations are compact and accurate over the whole solution domain .
as applied to the relativistic harmonic oscillator @xcite , the qualitative analysis of ordinary differential equations detailed in ref.@xcite can be used to show that the solution to eq.[coordtime ] is indeed periodic .
that is , the phase space trajectories given by @xmath10 where @xmath11 , are closed and have the strip structure : @xmath12 the period of this system can be derived from the generalized energy - momentum relation @xmath13 where @xmath14 is the total energy and @xmath15 is the relativistic momentum . in all our equations
the rest mass and speed of light are set to unity . by noting that the total energy is the sum of the potential energy at the amplitude @xmath16 and the rest mass , i.e. @xmath17 , it can be shown from eq.[energymom ] that @xmath18 which leads to the following expression for the period : @xmath19[\pa-\px+2]}}. \label{coordperiod}\end{aligned}\ ] ]
since the period is a function of the oscillation amplitude , it would be useful to express the physical regimes described by @xmath20 in terms of this parameter . setting @xmath21 in eq.[rel1 ] gives @xmath22}}{1+\pa}.\end{aligned}\ ] ] this relation maps the domain @xmath5 unto @xmath23 . for an even potential @xmath8 , the period from eq.[coordperiod ] can be written as @xmath24}}. \label{coordperiod2}\end{aligned}\ ] ] we approximate @xmath25 with the solvable harmonic potential @xmath26 , in a manner that is essentially a linear delta expansion @xcite : @xmath27[\w^2+\delta(\rx-\w^2)]}}. \label{lde}\end{aligned}\ ] ] in the above equation , when @xmath28 we have the case of simple harmonic motion , while for @xmath29 the system reduces to the original potential of eq.[coordperiod2 ] . with a change of variable @xmath30 ,
the above equation becomes @xmath31 where the substitution @xmath32 has been made .
a binomial expansion of the integrand in powers of @xmath33 then gives @xmath34 where we have set @xmath29 .
evaluating the above series to a finite order will leave an explicit @xmath9-dependence that should be eliminated .
indeed this is the basis of the pms , requiring that @xmath35 .
as shown in ref.@xcite , the condition is equivalent to choosing @xmath9 such that @xmath36 for an nth order optimization . to first order ,
the above criterion is equivalently @xmath37 where @xmath38 denotes the mean value of @xmath39 . for classical oscillators eq.[pms2 ]
is implemented easily , yielding remarkable results @xcite . in the case of relativistic oscillators
however , the expressions not only become cumbersome but are divergent at large oscillation amplitudes even with higher order computations .
for example , we find that the resulting @xmath9 computed for the relativistic quartic oscillator approaches unity as @xmath40 .
this marks an ineffective variation since the choice @xmath41 gives the exact untreated series .
hence for any order of calculation , the error of approximation increases rapidly beyond @xmath42 ( @xmath43 ) . the alternative criterion given by eq.36 of ref.@xcite is also found to be problematic because the resulting @xmath9 goes to infinity as @xmath40 .
since all terms in eq.[series ] are proportional to powers of @xmath44 , then any finite order approximation incorrectly goes to zero in the ultrarelativistic limit @xmath40 .
the correct large amplitude behavior can be verified by noting from eq.[rel1 ] that @xmath45 in this regime .
hence , eq.[coordperiod ] simplifies to what is expected for a photon bouncing in a box of width @xmath46 : @xmath47 the preceding limit will serve as the asymptotic boundary condition that determines our variational parameter .
we begin constructing our solution by noting that the zeroth order truncation of the series in eq.[series ] is @xmath48 we then assume the form of @xmath9 that directly satisfies eq.[pms ] , namely @xmath49 in the above equation @xmath9 is still dependent on the variable of integration @xmath50 since harmonics of the form @xmath51 are present . to remove this , we treat the harmonics as free parameters , denoting @xmath51 as @xmath52 .
next , we look at the limit of @xmath9 as @xmath40 . it can be shown that for an even potential with leading order @xmath53 , the large amplitude behavior is @xmath54 substituting eqs.[uperiod ] and [ wlarge1 ] in the limit @xmath40 of eq.[approx ] will yield a relation between all orders of free parameters that exist in @xmath55 .
this relation can be used recursively to provide @xmath3 equations .
however for a non - linear potential , the system of equations is under - determined since there are a total of @xmath56 free parameters .
we remedy this problem by factoring out @xmath57 from the numerator and denominator in eq.[wlarge1 ] , giving us @xmath58 thus the large amplitude limit of eq.[approx ] yields the relation @xmath59 under the ansatz of setting the lowest order free parameter to be its trigonometric average , i.e. @xmath60 , the @xmath61 case of the above equation determines @xmath62 .
these values are then used in the @xmath63 equation to solve for @xmath64 and so on , effectively building the whole set of @xmath3 free parameters from the bottom up .
this iterative computation yields that for integers @xmath65 the free parameters are given by @xmath66 we use these values in @xmath55 and determine @xmath9 by eq.[pms3 ] . then by eq.[approx ] the resulting approximation is @xmath67
as a first example we apply the method to the linear potential . plugging in @xmath68 in eq.[coordperiod ] and following the general derivation previously outlined yields @xmath69 in calculating the free parameters present in @xmath55 we utilize the relation given by eq.[wlarge1 ] since for @xmath70 the system is not under - determined .
accordingly , the ultrarelativistic limit of eq.[approx ] gives the relation @xmath71 by using @xmath60 in the above equation , we determine @xmath72 .
we then substitute these values into eq.[harmr ] and use eq.[approx2 ] , giving us the compact expression for the period @xmath73 we compare this with the exact period given by @xmath74 - 8 \text{k}\left[\frac{a^2}{4+a^2}\right]}{\sqrt{4+a^2 } } , \label{exact}\end{aligned}\ ] ] where @xmath75 $ ] and @xmath76 $ ] are the elliptic integrals of the first and second kind , respectively .
figure [ fig1 ] shows the relative error of eq.[harmperiod ] along with those of the results of harmonic balance @xcite and homotopy @xcite . despite the simplicity of our calculation , the third order approximations of both these methods are only more accurate by @xmath77 over a finite range of @xmath16 .
our solution is the only approximation whose error goes to zero at the ultrarelativistic limit .
an advantage is also apparent in the resulting expression itself , that it is compact and mathematically elementary , while those of harmonic balance and homotopy are lengthy and even require the evaluation of elliptic integrals .
the most note - worthy point of our method is its extensive applicability . for our second example
we calculate a general approximation for the periods of relativistic oscillators with the potential @xmath78 .
as we have done previously , we obtain the function @xmath55 by following the derivation starting from eq.[coordperiod ] .
the result is @xmath79 where @xmath80 note that this expression has been factored in a manner compatible with eq.[wlarge2 ] so that the existing free parameters are defined by eq.[free ] .
then the approximation simply follows from eq.[approx2 ] .
explicitly , the result is @xmath81\ } } } , \label{gent}\end{aligned}\ ] ] where we have evaluated the geometric sum appearing in eq.[genr ] . to give specific examples of this general result , the approximations for the periods of the relativistic quartic ( @xmath61 ) and sextic ( @xmath63 ) oscillators are respectively @xmath82}{\sqrt{(4 + 3 a^2)[6 a^4+(8+a^2 ) \pi ^2 ] } } \label{duffperiod}\end{aligned}\ ] ] and @xmath83[24 \pi ^4 + 3 a^2 \pi ^4+a^6 ( 20 \pi ^2 - 48)]}}. \label{qsperiod}\end{aligned}\ ] ] the approximations for the cases @xmath84 and the harmonic potential are fitted with the numerically integrated exact results in figure [ fig2 ] .
the relative errors shown in the inset are remarkably low even for the highly non - linear case of @xmath85 .
for such large values of @xmath3 the period exhibits a sharp jump from what is characteristically harmonic at amplitudes @xmath86 to the ultrarelativistic limit for @xmath87 . as a third example
, we derive a general approximation for the periods of relativistic oscillators with the potential @xmath88 .
again , the crux of the calculation is deriving a general form of @xmath55 and factoring it compatible with eq.[wlarge2 ] so that the free parameters are determined by eq.[free ] .
then by using eq.[approx2 ] , the following result is obtained @xmath89\ } } } , \label{gent2}\end{aligned}\ ] ] where @xmath90 for @xmath61 this reduces to eq.[duffperiod ] , the period of a relativistic quartic oscillator . for @xmath63
, it provides the following approximation to the period of a relativistic particle oscillating under the potential @xmath91 @xmath92}{\sqrt{[-24 a^4+(12 + 9 a^2 + 10 a^4 ) \pi ^2][-48 a^6 + 2 a^4 ( 9 + 10 a^2 ) \pi ^2 + 3 ( 8+a^2 ) \pi ^4 ] } } .
\label{period246}\end{aligned}\ ] ] figure [ fig3 ] shows the approximations for the cases @xmath93 and @xmath94 , fitted with the numerically integrated exact results .
notably , the inset shows low relative errors whose maximum do not increase with increasing @xmath3 .
lastly , we treat relativistic oscillators with the intrinsically non - linear potential @xmath2 . using the same procedure of deriving a general expression for @xmath55 and using eqs.[free ] and [ approx2 ] ,
the result is @xmath95}{a^{m-1}\sqrt{[4m+a^{2m}(1-\lambda_{2 m } ) ] \{\frac{3}{2}+\frac{\pi^2 - 6}{4}[1-(\frac{\pi^2 - 4}{\pi^2})^{m-2 } ] \ } } } .
\label{gent3}\end{aligned}\ ] ] for example , setting @xmath61 in the equation above gives the following approximation for the period of a relativistic oscillator with potential @xmath96 @xmath97 the approximations for the cases @xmath98 and @xmath99 and their numerically integrated exact results are shown in figure [ fig4 ] .
the method is able to capture the infinities at both regimes of the oscillation period .
our modification of a pms - based non - perturbative method offers an economical procedure of calculating compact and accurate analytic expressions for the period of an oscillating relativistic particle .
it is able to attack even polynomial potentials of general form , including those that are intrinsically non - linear .
the mechanics of the method itself guarantees the accuracy of the solution even at the ultrarelativistic limit . | the authors modify a non - perturbative approach based on the principle of minimal sensitivity to calculate the periods of relativistic oscillators with even polynomial potentials .
the optimization of the variational parameter is adapted by using the ultrarelativistic limit of the period as a boundary condition , which determines the values of additional free parameters introduced .
general approximations for the potentials @xmath0 , @xmath1 and @xmath2 prove to be accurate over the whole solution domain and even for large values of @xmath3 . |
the convection near solar surface develops on the different space and time scales .
numerical simulation provides useful information about evolution of convective structures and helps to construct consistent models of the physical processes underlying of observed solar phenomena .
we conduct an investigation of the temporal evolution and growth of convective modes on scales supergranulation in a three - dimensional computational box with imposed initially uniform weak magnetic field . in previous work by the author [ @xcite ] it was shown that collective motions of small convective cells of granulation expels weak magnetic field on the edges cells at mesogranular scales .
average size of such cells is 15 - 20 mm and the lifetime is 8 - 10 solar hours . in article [ @xcite ]
it was suggested that development of large spatial and long temporal mesogranular and supergranular scales naturally arise as result the collective advective interaction of many small - scale and short - lived granular plumes .
simulation of solar photosphere convection [ @xcite ] in a computational domain of size 48 mm in horizontal plane and 20 mm in depth showed that the sizes of convective cells increase with depth .
the purpose of this work is to investigate the effect magnetic field on the development of convection on scale supergranulation in a region of size 60 mm in the horizontal plane and 20 mm in depth .
as initial state we take the distribution of the main thermodynamic variables with radius from the standard solar model [ @xcite ] with parameters @xmath2 , where @xmath3 and @xmath4 are hydrogen and helium abundances by mass , and @xmath5 is the ratio of mixing length to pressure scale height in the convection region .
we used the opal opacity tables and the equation of state for solar matter [ @xcite ] .
the compressible nonideal magnetohydrodynamics equations are solved : @xmath6 @xmath7 = \rho \vec g + \nabla \cdot \hat \tau\ ] ] @xmath8 = \frac { 1}{4 \pi}\nabla \cdot \left(\vec b \times \eta \nabla \times \vec b\right)\nonumber \\ + \nabla \cdot \left(\vec v \cdot \hat
\tau \right ) + \rho \left(\vec g \cdot \vec v \right ) + q_{rad}\end{aligned}\ ] ] @xmath9 here @xmath10 is the density , @xmath11 is the pressure , @xmath12 is vector of the velocity , @xmath13 is vector of the magnetic field , @xmath14 is vector of the gravitation , @xmath15 is the total energy , @xmath16 is the energy transferred by radiation and @xmath17 is the viscous stress tensor .
the influence of small scales on large scale flows was evaluated in terms of viscous stress tensor and the rate of dissipation was defined from buoyancy and shear production terms [ @xcite ] .
the evolution of all variables in time was found using an explicit tvd[total variation diminishing ] conservative difference scheme ( [ @xcite ] ) @xmath18 where @xmath19 and the operator @xmath20 is @xmath21 the flux along each direction was defined by the local - characteristic method as @xmath22 \ .\ ] ] @xmath23 is the matrix whose columns are right eigenvectors of @xmath24 evaluated as a generalized roe average of @xmath25 and @xmath26 for real gases .
@xmath27 is the matrix of numerical dissipation .
the term @xmath28 represents the effect of gravitational forces and radiation .
time step integration is by third - order runge - kutta method [ @xcite ] .
this scheme is second - order in space and time .
central differences were used for the viscous term , and the diffusion approximation was applied for the radiative term in the energy equation .
we used a uniform grid in the @xmath29 and @xmath30 directions and a nonuniform grid in the vertical @xmath31 direction .
periodic boundary conditions were used in the horizontal directions and the top and bottom boundary conditions were choosen to be @xmath32 @xmath33 @xmath34 that is , reflection for the @xmath35 component of velocity , outflow for the @xmath29 and @xmath30 components .
the pressure and density were derived from the solution of the hydrostatic equation , using the equation of state by constant value of the internal energy .
we imposed initially the homogeneuos vertical magnetic field with strength of 50 g. for the magnetic diffusivity we take constant value @xmath36@xmath37sec@xmath38 .
we carried out calculations in the region of @xmath39 mm on the grid with @xmath1 cells during of 24 solar hours . on the figures 1 - 3
the results of the mhd numerical simulation of development of the convection in the horizontal plane near solar surface are presented .
we find that the magnetic field concentrates in extended regions similar sunspot with diameter about 5 mm and in thin sheets on the boundary of supergranular cells . in these regions
we reveal the increase of strength of the magnetic field to values 700 - 800 g , the decrease of temperature on few thousand kelvin , the small values of the vertical component of velocity in average about 0.05 km / sec and as result full absence of the convective motion .
the magnetic pressure prevents the inflow of the material from outside of sunspot and thus the transfer of radiation energy is suppressed . inside of supergranular cell
the value of vertical component of the magnetic field is very small about 1 g and less . diverging convective flows from the center supergranular expels weak magnetic field to the edges of convective cell .
an average size of the supergranular cells is 20 - 30 mm .
a material moves from the center of supergranular with the velocity of about 1 - 1.5 km / sec .
the maximum value of the magnetic field in the computational domain is 2000 g. inside of supergranular we have a typical picture of evolution of convection on the scale of granulation with average sizes of the cells about 1.5 mm and with lifetime about 4 - 5 minutes [ @xcite ] .
here we see wider upflows of warm , low density , and entropy neutral matter and downflows of cold , converging into filamentary structures , dense material .
we observe a continuous picture of formation and destruction of granules . the granules with highest pressure grow and push matter against neighboring granules , that then shrink and disappear .
ascending flow increases the pressure in the center of granule and upflowing fluids decelerates motion .
this process reduces heat transport to surface and allows the material above the granule to cool , become denser , and by action of gravity to move down .
we observe a formation of new cold intergranule lane splitting the original granule .
from figure 4 we can see properties of turbulent convection on depth 0 - 4 mm . in this region
the cold blobs of the matter moves down by maximum of velocity of about 4 km / sec and the mach number m=0.6 .
the downdrafts has different and very complicate vertical structure .
one part of the downdrafts travels on small distance from the surface and becomes weak enough to be broken up by the surrounding fluid motion .
another part of the downdrafts conserves a motion with high velocity and moves on the distance about 6 mm .
we observe that different nature of such behavior is provided by the initial conditions of downdrafts formation .
there are places on solar surface where the material moves from different sides to one point and here we find powerful vorticity motions of matter . due to action of strong compressibility effects
we observe quick output of internal energy and formation of downdrafts with maximum of values of the vertical velocity .
we detect extended regions with size about 5 mm in a diameter where turbulent convection is fully disappeared . on the depthes from 5 mm to 8 mm we reveal more quiet character of the convective flow than in the turbulent zone .
below 8 mm we see clearly the separate large scale density fluctuations and the streaming flow of the material which similar to jets with the largest value of the average velocity equal to 1 km / sec . in these places
the magnetic field has value about 300 g. besides we find on depthes bigger than 10 mm the separate regions of localization of the huge magnetic energy with values of magnetic field in range from 800 to 1000 g ( fig.5 ) .
distances between these places with strong concentration of the magnetic field is comparable with scales development of the local solar supergranulation .
the convective motion in different parts of photosphere leads to continuous changes of topology of the magnetic field .
we find inside of the supergranular cell many magnetic loops with big negative values of the vertical component of the magnetic field . | three - dimensional numerical simulations of solar surface magnetoconvection using realistic model physics are conducted .
the thermal structure of convective motions into the upper radiative layers of the photosphere , the main scales of convective cells and the penetration depths of convection are investigated .
we take part of the solar photosphere with size of @xmath0 mm in horizontal direction and by depth 20 mm from level of the visible solar surface .
we use a realistic initial model of the sun and apply equation of state and opacities of stellar matter .
the equations of fully compressible radiation magnetohydrodynamics with dynamical viscosity and gravity are solved .
we apply : 1 ) conservative tvd difference scheme for the magnetohydrodynamics , 2 ) the diffusion approximation for the radiative transfer , 3 ) dynamical viscosity from subgrid scale modeling . in simulation
we take uniform two - dimesional grid in gorizontal plane and nonuniform grid in vertical direction with number of cells @xmath1 .
we use 512 processors with distributed memory multiprocessors on supercomputer mvs-100k in the joint computational centre of the russian academy of sciences . |
in the cases of at least ten nearby ( redshift @xmath3 ) lgrbs , observations have revealed a sn ic - bl spectrum superimposed on the power - law continuum of the fading optical afterglow ( @xcite ; see @xcite and @xcite for reviews ) .
the spectra are characterized by wide features consistent with high ejecta velocities ( @xmath420,00030,000kms@xmath5 ) , and an absence of hydrogen and helium .
the other principal classes of core - collapse sn , in contrast , exhibit spectroscopic features consistent with more slowly moving ejecta .
the most common core - collapse sn are type ii ( @xmath6% ; @xcite ; @xcite ) , which exhibit hydrogen ( h ) in their spectra ; they are the final eruptions of stars that have retained their outer h shell .
when the progenitor sheds , transfers to a companion , or internally mixes its outer h envelope during pre - sn evolution , the explosion will produce an h - deficient sn ib , or an h- and he - deficient sn ic ( e.g. , * ? ? ? * and references therein ) . while the absence of h and he in the spectra of sn ic - bl indicates that their progenitors have lost their envelopes prior to core collapse
, simulations additionally suggest that the progenitors of sn ic - bl with associated lgrbs may also have rapid speeds of rotation . in models , only quickly rotating stars without an h envelope produce the outflowing jets that yield @xmath7-ray emission @xcite after core collapse to a black hole @xcite .
close massive binary systems that experience mass transfer , common - envelope evolution , or a merger @xcite are possible progenitors of lgrbs , because they can likely produce the required rapidly rotating massive stars without h or he envelopes .
recent observations of galactic o - type stars show that , in fact , @xmath8% of massive stars experience mass transfer with a companion and @xmath9% undergo a merger @xcite .
alternative progenitor candidates include quickly rotating , metal - poor stars that internally mix their outer envelopes @xcite .
single stars with high abundances , however , are considered improbable progenitors , because their comparatively strong winds @xcite are expected to reduce their angular momentum .
spectroscopy of @xmath10 host galaxies shows , in fact , that nearby sn ic - bl with an associated grb prefer more metal - poor environments than nearby type ic - bl sn having no obvious @xmath7-ray emission @xcite .
the latter , in turn , prefer more metal - poor @xcite and blue @xcite environments than do sn ic without broad features .
in several cases , however , the positions of low- and moderate - redshift lgrbs are spatially coincident or closely associated with massive , metal - rich galaxies ( e.g. , @xcite ) .
absorption spectroscopy also finds evidence for high - redshift @xmath11 systems along the line of sight to lgrbs with metallicities exceeding the @xmath11 cosmic average @xcite .
@xcite has inferred a supersolar abundance from @xmath12 absorption features consistent with a pair galaxies in the grb 090323 afterglow spectrum .
grb 130702a @xcite occurred , however , in a metal - poor faint satellite of a @xmath13 massive galaxy @xcite , raising the possibilty of superpositions or associations for some lgrbs that would be difficult to resolve at high redshift . here
we measure the host - galaxy properties of nearby core - collapse sn explosions using the imaging and photometry of the sloan digital sky survey ( sdss ) . for @xmath2 lgrb hosts ,
we estimate these host properties from published photometry and archival _ hst _ images . we show that sn ic - bl and lgrbs exhibit a strong preference for galaxies that have high stellar - mass density and star - formation - rate density for their stellar mass .
we also use sdss spectra to show that the gas kinematics of sn ic - bl hosts are exceptional . in [ sec : data ] , we describe the core - collapse sn and lgrb samples as well as the sdss and _ hst _ galaxy data that we use in this analysis .
section [ sec : methods ] presents our techniques to analyze the galaxy imaging and spectroscopy , and the statistical methods we employ . in [ sec : results ] , we describe the results of our analysis , while [ sec : discussion ] discusses the interpretation of the observed patterns .
our conclusions are presented in [ sec : conclusions ] .
we study the host galaxies of both nearby ( @xmath1 ) core - collapse sn discovered by `` galaxy - untargeted '' transient searches ( e.g. , the palomar transient factory ; ptf ; @xcite ) which do not target specific potential hosts , and @xmath2 lgrbs detected by @xmath7-ray satellites .
we use the sdss spectroscopic sample to build a control sample of low - redshift star - forming galaxies , and sdss photometry and spectroscopy to measure properties of both the sample of low - redshift star - forming galaxies and the host galaxies of the nearby sn . for the host galaxies of @xmath14 lgrbs , we estimate host properties using published photometry and _ hst _ imaging .
the sdss galaxy photometry and fiber spectra are from data release 10 ( dr10 ; * ? ? ?
* ) , and they were collected with the 2.5 m telescope at apache point , new mexico .
the imaging survey , which spans 14,555 square degrees , consists of 53.9s integrations through the sdss @xmath15 filters , and the typical limiting @xmath16-band ab magnitude is 22.2 .
the typical sensitivity of available sdss imaging makes possible detection of @xmath17 galaxies with absolute vega magnitudes @xmath18 or @xmath19 brighter than about @xmath20 .
each sloan @xmath21 pixel ccd array records a @xmath22 field of view .
the sdss spectroscopic survey acquired approximately 45min of total integration in clear conditions , split into a series of three successive exposures .
the spectrograph comprises @xmath23 ( @xmath24 for boss ) diameter fiber - optic cables placed at the positions of targets on the focal plane .
adjacent fibers can be no closer than @xmath25 in a single fiber mask because of engineering constraints @xcite , and the sdss spectrographs record light with wavelengths 38009200 . the targets selected for sdss fiber spectroscopy consist of three primary `` legacy '' categories of objects , and were also taken from several dozen ancillary programs , some of which were limited to specific parts of the survey ( e.g. , stripe 82 ) .
objects detected with @xmath26 significance in the imaging survey , with an extended light distribution and having @xmath16-band magnitude brighter than 17.77 , as well as qso candidates and luminous red galaxies ( lrgs ) , formed the legacy samples @xcite .
more limited special - program science targets included , for example , a @xmath27-band galaxy sample selected to investigate the properties of blue , faint galaxies .
table [ tab : selection ] shows the construction of our sample of sdss galaxy spectra .
the core - collapse sn sample is constructed from @xmath28 discoveries by sn surveys that do not target specific galaxies .
the `` galaxy - untargeted '' sn search technique is akin to that of wide - field @xmath7-ray satellite ( e.g. , _ swift _ ) searches for lgrbs .
the sn in our sample consist of discoveries by galaxy - untargeted searches reported to the international astronomical union ( iau ) , as well as those published @xcite or reported via astronomical telegrams by the ptf , from 1990 january 1 through 2012 may 10 .
the sn searches that we considered to be galaxy - untargeted are identical those listed by @xcite .
our sample consists of only those nearby sn whose host galaxies have sdss @xmath15 imaging and , when appropriate , a fiber spectrum .
the probability of detecting a sn in an image depends on the limiting magnitude of the transient search , as well as on the sn distance , luminosity , light - curve shape , and dust attenuation along the line of sight .
comparative analysis of the redshifts of core - collapse sn discovered by galaxy - untargeted searches found no significant evidence that variation among the principal species strongly affects their discovery rate @xcite .
@xcite showed that sn ic are more closely associated with the highest surface brightness regions of their host galaxies than sn ib , but here we combine these two sn types to assemble a larger sample of stripped - envelope sn that do not show high ejecta velocities .
we also show , for comparison , the host of ptf12gzk , a peculiar sn ic that did not exhibit broad spectroscopic features @xcite but in which radio observations found evidence for @xmath29 ejecta speeds @xcite .
lists of the nearby sn and the lgrbs are presented in tables 2 and 3 ( full tables are available in the electronic version ) .
our lgrb sample includes both objects with a detected optical afterglow , and `` dark '' bursts without a luminous optical afterglow @xcite whose position was determined using their x - ray flux @xcite .
analysis of the host galaxies of dark grbs shows that they are both more massive @xcite and dust - obscured @xcite than the hosts of grbs with detected optical afterglows .
the fraction of dark bursts in our sample ( ) is approximately representative of the fraction of dust - obscured grbs below the @xmath30 redshift upper limit of our lgrb sample reported by @xcite .
we assemble @xmath2 lgrb host galaxies having archival _ hst _ images from the unobscured lgrbs assembled by @xcite and dust - obscured lgrbs assembled by @xcite , after rejecting several datasets that showed evidence of residual lgrb light . in many cases ,
_ hst _ images were acquired at least a year after the explosion ; these are the data we use when available to measure galaxy sizes .
when only images taken closer to the time of the grb are available , we inspect the host galaxy to determine whether evidence for a point source at the explosion site exists . lgrbs exhibit a strong association with the brightest pixels of their hosts @xcite ,
so a potential concern is that we could reject data where the lgrb coincided with a bright star - forming region . in practice ,
possible confusion was minimal for the data taken within @xmath31 months after the grb trigger . while we do not expect any significant contamination , here we are interested in the integrated properties of the host galaxy and not the specific region where the lgrb occurred .
even when an optical afterglow can be detected , identifying a coincident sn in the fading afterglow light curve , or by identifying a sn spectrum superimposed on the power - law afterglow continuum , requires high signal - to - noise ratio ( s / n ) data .
sufficient follow - up observations have only been possible for lgrbs with @xmath32 .
our grb sample includes events with and without a detected optical afterglow or coincident sn .
since a representative fraction of the lgrbs in our sample have no optical counterpart , our findings should be robust to the effects of dust obscuration on lgrb detection .
lc ( 1 ) full catalog & 948,205 + ( 2 ) reliable line measurements & 910,532 + ( 3 ) star forming ( low or high s / n ) & 377,763 + ( 4 ) h@xmath33 s / n @xmath34 20 & 302,865 + ( 5 ) fiber offset @xmath35 & 289,474 [ tab : selection ]
for both the @xmath1 core - collapse sn and @xmath2 lgrb samples , as well as the sdss star - forming population , we estimate host - galaxy stellar masses @xmath36 and photometric star - formation rates ( sfrs ) by fitting pegase2 @xcite stellar population synthesis models to broadband photometry ; see @xcite for detailed information on the star - formation histories and initial - mass functions ( imfs ) used . for nearby core - collapse sn host galaxies and the sdss star - forming population , we fit sloan @xmath15 magnitudes .
the multi - band photometry of the host galaxies of lgrbs was assembled from the ghosts database .
we describe a second , complementary set of sfr estimates in the following section that is available only for the sdss star - forming spectroscopic sample and which uses both the fiber spectrum and broadband @xmath15 photometry .
when comparing among samples , however , we only compare sfr values estimated using the same method .
several teams have performed and made available detailed measurements of sdss galaxy properties from the sloan photometry and spectroscopy . for the @xmath28 samples of core - collapse sn host galaxies and the sdss star - forming population , we use sfrs estimated from fitting both spectra and photometry , and gas velocity dispersions @xmath37 measured from the h@xmath33 emission - line profile .
we use the star - forming classifications for sdss galaxies and hybrid spectroscopic and photometric sfr estimates made available by a collaboration that was both at the max planck institute for astronomy ( mpa ) and johns hopkins university ( jhu ) ( s. charlot , g. kauffmann , s. white , t. heckman , c. tremonti , and j. brinchmann ; mpa - jhu ) .
sdss fiber apertures generally do not cover the entire light distribution of each target galaxy .
the mpa - jhu team therefore estimates the total sfr of each galaxy as the sum of the sfr within the fiber aperture determined from fitting the spectrum , and from a fit to the @xmath15 light outside the fiber opening .
we adopt the gas velocity dispersions estimated from sdss emission - line profiles by the portsmouth group @xcite , which apply the public penalized pixel fitting @xcite ( ppxf ) and the gas and absorption line fitting @xcite ( gandalf v1.5 ) codes .
the velocity dispersion of the gas is estimated from the widths of emission lines ( e.g. , h@xmath33 , [ o@xmath38iii ] ) , and here we take the dispersion for h@xmath33 .
the sdss photo pipeline performs separate fits of a de vaucouleurs @xmath39 law and an exponential profile to the light distribution of each extended object .
the pipeline next finds the linear combination of the two models ( holding all parameters except flux fixed ) that minimizes the @xmath40 statistic . to obtain an estimate of the half - light radius @xmath41 , we compute the weighted average of the two components @xmath41 parameters and weight each by its fractional contribution to the total model flux .
we use the galfit @xcite program to perform the same surface - brightness fitting analysis on archival _ hst _ images of lgrb host galaxies .
we apply source extractor ( sextractor ; * ? ? ?
* ) to drizzled and cosmic - ray - rejected images to estimate object positions , ellipticities , and magnitudes , and these are used as input galfit parameters . extended sources except the lgrb hosts are modeled with sersic profiles , and the instrument point - spread function ( psf ) is used to model stars .
the host - galaxy @xmath41 estimates show good agreement with existing estimates of @xmath41 from profile fitting @xcite , and with published @xmath42 sextractor measurements @xcite .
the host galaxy of grb 020903 is part of a complex association of interacting clumps , so the galaxy is not well approximated by a simple surface - brightness model .
an additional complication is that much of the host - galaxy photometry was taken from the ground , where the galaxy components can not be resolved .
we therefore have excluded this host galaxy from the analysis .
the angular diameter of eso184-g82 , the host galaxy of the nearby grb 980425 ( @xmath43 ) , approximately spans the available _ hst _ images .
the physical resolution of the _ hst _ data is also substantially higher than that of the images of the sdss sn host galaxies , or other lgrb hosts . measurement of the host effective radius from model fitting requires significant imaging area without galaxy light to be able to fit robustly for the background level .
we therefore do not include grb 980425 , which does not have sdss exposures , in the lgrb sample .
the host galaxy of grb 051022 is also a system with two peaks that may possibly correspond to two strongly interacting galaxies , or else may instead be a galaxy with irregular morphology .
we expect that the sdss photo pipeline would be most likely to model the host as a single system , so we construct a galfit model that consists of a single galaxy .
we calculate the projected stellar - mass density @xmath44 and the projected star - formation density @xmath45 from the galaxy stellar mass @xmath36 and star - formation rate sfr , respectively , and the parameters of the model @xmath16-band isophotal ellipse that encloses half of the galaxy light .
we compute , for example , @xmath46 , where @xmath36 is the stellar mass ( in @xmath47 ) , while @xmath48 and @xmath49 are the semimajor and semiminor axes ( in kpc ) of the isophotal ellipse that contains half of the galaxy @xmath16-band flux , determined from fitting its surface - brightness distribution .
photometric magnitudes measured through an appropriate aperture designed to enclose a specific percentage of the host light ( e.g. , a petrosian aperture ; @xcite ) can be expected to sample an approximately consistent fraction of galaxy light with increasing distance to sources .
physical properties estimated from broadband fluxes ( e.g. , stellar mass @xmath36 ) should therefore not exhibit strong aperture biases with increasing redshift in our sample , although the effects of surface - brightness dimming may become important at high redshifts . to study the properties of galaxies derived from photometric measurements
, we therefore find the best - fitting @xmath36@xmath44 , @xmath36@xmath45 , @xmath36@xmath41 , and sfr@xmath41 relations for the complete @xmath50 sdss star - forming catalog .
we fit a second - degree polynomial to the median ordinate value ( e.g. , @xmath45 ) across at least five bins in galaxy stellar mass .
we include an additional point at ( log @xmath36 , @xmath41 ) = ( 0,0 ) and ( log sfr , @xmath41 ) = ( @xmath51,0 ) , respectively , when fitting for the sdss @xmath36sfr relation so that @xmath41 approaches zero for galaxies with negligible stellar mass @xmath36 or sfr .
we use the main sdss spectroscopic sample legacy `` galaxy '' targets ( i.e. , where the 64 bit of the primtarget bitmask was set ) to measure median relations of the properties of the low - redshift galaxy population . the fixed angular size of sdss @xmath23 fibers ( or @xmath24 for the baryon acoustic oscillation survey ; boss ) , in contrast to ( for example ) a petrosian aperture , samples a percentage of the light of each extended target that depends on the redshift and intrinsic size of the galaxy . to reach conclusions about spectroscopic measurements that are not sensitive to aperture effects
, we apply a separate predictive algorithm that we have developed that yields robust comparisons .
to predict the expected value of an observable ( e.g. , @xmath37 ) given other host - galaxy properties ( e.g. , mass @xmath36 ) , we perform , for each @xmath28 sn host galaxy , a locally weighted multiple linear least - squares fit to the sdss catalog . we model the observable of interest @xmath52 as @xmath53 where @xmath54 indexes the sdss galaxies , and each @xmath55 is a galaxy property , a combination ( e.g. , multiplicative product ) of galaxy properties , or unity . for example , we model the gas velocity dispersion @xmath37 as @xmath56 where @xmath36 is stellar mass and @xmath57 is galaxy ellipticity .
each row @xmath54 of the design matrix @xmath58 and @xmath59 is multiplied by the weight @xmath60 , where @xmath61 is @xmath62 and @xmath63 is galaxy redshift . to be able to predict a specific property @xmath64 of each host galaxy ( e.g. , @xmath37 ) independent of its observed value , @xmath65
, we exclude the host galaxy itself when fitting the predictive model to the properties of sdss galaxies .
this approach assigns greater weight to galaxies with similar observables , and mitigates any possible selection effect with galaxy redshift or fiber fraction .
the best - fit parameters @xmath66 for each fit are used to compute a prediction for the value of the galaxy property , @xmath64 , which can then be compared to the observed value , @xmath65 , as a residual , @xmath67 here the fitting analysis requires only modest numbers of nearby datapoints from the sdss spectroscopic catalog .
in figures [ fig : massmassdensity ] and [ fig : masssfrdensity ] , we plot galaxy stellar - mass density @xmath44 and star - formation density @xmath45 , estimated from broadband magnitudes , against stellar mass @xmath36 . these show that @xmath1 sn ic - bl and @xmath14 lgrb host galaxies have high stellar - mass density and star - formation density for their stellar masses , compared with the low - redshift ( @xmath1 ) sdss star - forming galaxy population .
in contrast , sn ib / ic ( with slower ejecta speeds ) and sn ii show no preference for galaxies that have high stellar - mass density or high star - formation density for their stellar masses .
we compute the residuals of each host galaxy from the sdss @xmath36@xmath44 and @xmath36@xmath45 relations .
our statistical method is to determine the probability that each pair of residual distributions is identical using the kolmogorov - smirnov ( ks ) two - sample test .
we find significant evidence for differences between the sn ic - bl ( @xmath68 ) and the sn ii ( @xmath68 ; and for @xmath44 and @xmath45 , respectively ) host distributions .
comparison between the sn ic - bl and sn ib / ic ( @xmath68 ) host distributions likewise finds evidence for distinct underlying distributions ( and ) .
the residuals of lgrb ( @xmath68 ) host galaxies from the sdss @xmath36@xmath44 and @xmath36@xmath45 relations are more positive than ( i.e. , @xmath44 and @xmath45 are greater than ) those of sn ib / ic ( and , respectively ) and sn ii ( and ) hosts , while we find no statistically significant difference with the sn ic - bl host residual distribution ( and ) .
the galaxy @xmath36@xmath44 relation shows no significant change with increasing redshift to @xmath69 ( e.g. , @xcite ) , so comparisons involving the lgrb host @xmath36@xmath44 relation should not be strongly affected by evolution in the galaxy population with redshift .
we study the @xmath1 sdss star - forming population to investigate whether galaxies with relatively high stellar - mass and star - formation densities have comparatively low chemical abundances . in figure
[ fig : sdsspop ] , we show the average @xcite oxygen abundance of @xmath1 sdss star - forming galaxies as a function of @xmath36 and @xmath44 as well as @xmath45 .
the star - forming sdss galaxies with the highest stellar - mass or star - formation - rate densities ( in each stellar - mass bin ) are not , on average , comparatively metal poor .
the galaxies with the highest stellar - mass densities are , on average , more metal - rich by @xmath40.2 dex .
an important question is whether the high stellar - mass and star - formation - rate densities of sn ic - bl and @xmath14 lgrb host galaxies can be attributed entirely to the preference for low metal abundance observed among low - redshift sn ic - bl @xcite and lgrbs @xcite .
we can represent such a preference as a progenitor formation efficiency @xmath70 per unit stellar mass created that depends solely on metallicity @xmath71 and diminishes with increasing @xmath71 .
if @xmath70 is a function of only metallicity @xmath71 , then the rate @xmath16 of fast - ejecta transients in a galaxy with a given sfr is @xmath72 half of the star formation in each galaxy mass @xmath36 bin plotted in figure [ fig : sdsspop ] occurs in galaxies below the @xmath73 and @xmath74 relations marked by thick blue lines .
a transient population whose @xmath70 has no environmental dependence would be expected to explode equally in galaxies above and below these sfr - weighted relations .
if @xmath70 instead increases at low abundance ( and depends only on @xmath71 ) , we would expect greater numbers of sn ic - bl and @xmath14 lgrbs on the side of the relation , either above or below , that is comparatively metal poor . inspection of figure [ fig : sdsspop ] shows that , on average , higher stellar - mass density corresponds to higher metallicity , while metallicity does not vary strongly with star - formation density among galaxies in each bin in stellar mass .
consequently , if @xmath70 increases at low abundance ( and depends only on @xmath71 ) , we would expect a plurality of sn ic - bl and @xmath14 lgrbs to be below the sfr - weighted @xmath73 relation , and approximately equal numbers above and below the sfr - weighted @xmath74 relation .
since this prediction is very different from the pattern we observe , the production of sn ic - bl and lgrb progenitors much be enhanced by an additional factor other than low metallicity that varies with @xmath44 as well as @xmath45 .
high gas dispersion velocities provide additional , independent evidence for distinct physical conditions in the host galaxies of low - redshift sn ic - bl . in figure
[ fig : halphaveldisp ] , the sn ic - bl hosts describe an @xmath36@xmath37 relation with an offset to high gas velocity dispersion measured from the h@xmath33 emission lines in the sdss fiber spectra .
the sn ic - bl host distribution is significantly different from the sn ib / ic ( ) and the sn ii ( ) host distributions .
the comparatively high gas velocity dispersions of sn ic - bl host galaxies may have one or more physical explanations .
rotationally supported galaxies having more compact and dense mass configurations are expected to have higher velocity dispersions .
alternatively , recent analysis of the gas kinematics of a sample of @xmath750.3 compact , highly star - forming galaxies finds complex profiles for the strong emission lines , with several narrow ( @xmath76120kms@xmath5 ) and broad ( @xmath77250kms@xmath5 ) components @xcite that may be due to strong stellar winds , or emission from expanding sn remnants . in figures [ fig : icblmosaic ] and
[ fig : othermosaic ] , we show images of core - collapse sn host galaxies and the measured h@xmath33 velocity dispersions juxtaposed with control samples of sdss star - forming galaxies having similar parameters .
these demonstrate the high velocity dispersions of sn ic - bl host galaxies . in figures [ fig : masshl ] and [ fig : sfrhl ] , we plot galaxy half - light radius @xmath41 ( in kpc ) against @xmath36 and sfr estimated from their broadband magnitudes . these show that @xmath28 sn ic - bl and @xmath14 lgrb host galaxies are compact for their stellar masses and sfrs , compared with the low - redshift ( @xmath28 ) sdss star - forming galaxy population . in contrast , sn ib / ic ( with slower ejecta speeds ) and sn ii show no preference for galaxies that are relatively compact in size . in figure [ fig : sfrspecdensity ] , we plot @xmath45 against @xmath36 for the sample of nearby ( @xmath28 ) core - collapse sn galaxies with sdss spectra . here
we calculate @xmath45 using the sfr estimated by the mpa - jhu group from the fiber spectrum and by modeling the galaxy _ ugriz _ light outside of the 3@xmath78 fiber .
these sfr measurements may be expected to be more accurate and precise than the sfr estimates that are based on galaxy broadband photometry alone , and provide evidence that the preference among sn ic - bl and lgrbs for high @xmath45 seen in figure [ fig : masssfrdensity ] galaxies is robust .
the sn ic - bl ( @xmath68 ) and sn ii ( @xmath68 ) distributions show significantly different ( @xmath79 ) residual distributions from their predicted @xmath45 values .
figure [ fig : msfr ] shows the relationship between @xmath36 and sfr for the core - collapse host galaxies and the sdss star - forming population .
this shows that host galaxies of sn ic - bl are not substantially more strongly star forming for their stellar masses than other core - collapse hosts , and suggests that relatively compact host sizes may primarily account for their high star - formation - rate densities .
given the existing evidence that @xmath10 sn ic - bl @xcite and lgrbs @xcite prefer metal - poor environments , a reasonable question is whether the high @xmath44 and @xmath45 host galaxies in our sample having sdss spectra describe a metal - poor mass - metallicity ( @xmath36@xmath71 ) relation .
kelly et al .
( 2014 , in prep . )
report no significant evidence that the sn ic - bl offset distribution ( @xmath68 ) differs from the sn ib / ic (; @xmath68 ) or the sn ii (;
@xmath68 ) distributions .
we have found that low - redshift sn ic - bl and @xmath2 lgrb host galaxies show stellar - mass and star - formation densities high compared with those of low - redshift galaxies having similar stellar masses .
core - collapse sn with more slowly expanding ejecta , however , exhibit no preference for galaxies having overdense stellar - mass distributions and star formation .
sn ic - bl host galaxies exhibit high gas velocity dispersions for their stellar masses , providing evidence for exceptional conditions in their hosts . given the effect of stellar metallicity to strengthen line - driven winds and remove the angular momentum of massive stars , studies of sn and lgrb host galaxies have often attempted to explain environmental patterns in terms of chemical abundance .
however , across the stellar - mass range populated by sn ic - bl and lgrb hosts , sdss star - forming galaxies with high stellar - mass and star - formation - rate densities are not , on average , more metal poor than less dense galaxies having similar stellar masses ( see figure [ fig : sdsspop ] ) .
therefore , a preference other than for low metallicity environments must be responsible for the overrepresentation of sn and lgrbs in galaxies with high stellar - mass and star - formation - rate densities .
furthermore , the @xmath36@xmath71 relation of sn ic - bl host galaxies , which have high stellar - mass and star - formation densities , is not significantly metal poor compared to the sdss @xmath36@xmath71 relation ( kelly et al .
2014 , in prep . ) .
in fact , a preference of high - velocity explosions for rapidly star - forming or overdense regions may help explain the observed association of lgrbs with the brightest regions of their host galaxies @xcite .
an initial suggestion was that a strong association with the brightest regions would be expected for lgrbs if they occur preferentially in low - mass , metal - poor galaxies .
if the most massive stars form in ob associations , then these will be more likely to correspond to the peaks of the light distribution of low - mass , metal - poor galaxies .
ob associations in more massive , metal - rich spirals may instead be outranked in brightness by , for example , the nucleus .
while such a low - metallicity preference may contribute , we suggest that an affinity to high star - formation density is important .
a plausible explanation for the association of lgrbs with regions of higher star - formation density is the formation efficiency of young , bound star clusters .
observations of extragalactic star clusters have found evidence that bound - cluster formation efficiency increases with the star - formation density @xcite .
binary systems may be created more frequently in bound clusters , and those that form are expected to become progressively tighter through dynamical interactions with other members of the cluster @xcite .
therefore , massive stars that form in dense star clusters are more likely to be in tight binary systems . interacting massive binaries are candidate progenitor systems for sn ic - bl and lgrbs , because mass transfer or common - envelope evolution leading ( in some cases ) to a merger can yield a rapidly rotating star whose outer envelope is not composed of h.
if sn ic - bl and lgrbs explode from stars that are more massive than the progenitors of sn ib / ic and sn ii , an imf that becomes top - heavy in dense , highly star - forming regions provides an alternative explanation for the patterns we observe .
indirect evidence from the absorption features of @xmath80 stars suggests instead that the imf may be bottom - heavy in elliptical galaxies with high stellar velocity dispersions and [ mg / fe ] , which are thought to have had high star formation densities during their formation epoch @xcite .
this suggests that the imf may be bottom heavy in dense , highly star - forming regions , while the opposite trend would be required to explain our observations .
an improved census of massive stars across low - redshift environments may help to address more directly whether the patterns we have found can be explained by variation in the upper end of the imf .
the median redshift ( @xmath81 ) of the @xmath14 lgrb sample is substantially higher than those of the core - collapse sn sample ( @xmath82 ) and the sdss sample of galaxy spectra ( @xmath83 ) .
the and relations among star - forming galaxies may evolve toward lower @xmath44 and @xmath45 densities from @xmath84 to the low - redshift universe . as we have shown , however , the low - redshift sn ic - bl host galaxy population exhibits the same preference for high stellar - mass and star - formation density galaxies as lgrbs . additionally ,
if stellar - mass and star - formation - rate density are proxies for the star - forming conditions that promote the formation of sn ic - bl progenitors , then we expect that these same physical conditions exist in high - redshift lgrb host galaxies having similar positions in the and planes .
given the evidence that the progenitors of high - velocity ejecta core - collapse explosions form more efficiently in galaxies that have high stellar - mass and star - formation densities , a reasonable expectation is that lgrbs will be improved tracers of star formation at earlier epochs where galaxies have higher densities of stellar mass and ongoing star formation than in the low - redshift universe @xcite .
lgrbs may be the most powerful probes of star formation at the highest redshifts .
the extreme luminosities of lgrbs allow detection at the earliest epochs of star formation ( @xmath85 ) , where optical and infrared instruments are not yet able to detect sub-@xmath86 galaxies .
the evidence we have found that the formation of lgrb progenitors depends sensitively on star - forming conditions suggests that they will be critical tools for investigating early star formation in detail .
we have studied the host galaxies of a sample of 245 low - redshift core - collapse sn , including 17 sn ic - bl discovered by galaxy - untargeted searches and 15 optically luminous and dust - obscured @xmath2 lgrbs .
we have used the uniform fiber spectra and photometry of the sdss to measure the properties of the core - collapse sn host - galaxy sample , as well as characterize the low - redshift star - forming galaxy population . published multi - band photometry as well as _
hst _ imaging were used to study the @xmath14 lgrb host galaxies .
the outflowing ejecta of sn ic - bl , from the wide features of their spectra , and of lgrbs , through their association with sn ic - bl and @xmath7-ray emission , are inferred to expand with high velocities ( @xmath420,00030,000kms@xmath5 ) .
we have found that these core - collapse explosions in which a significant fraction of the ejecta is moving at high velocity prefer galaxies with high stellar - mass and star - formation densities , when compared to galaxies having similar stellar masses .
the core - collapse sn in our sample having typical velocities of their ejecta , in contrast , are found approximately equally in galaxies with comparatively low and high stellar - mass and star - formation densities . from the widths of h@xmath33 emission lines , we find that the hosts of sn ic - bl have exceptionally high gas velocity dispersions when compared with star - forming galaxies having similar stellar masses . while early lgrb host environment analyses helped to establish the connection between lgrbs and massive stars ( e.g. , @xcite ) , subsequent analysis has focused on the effect of metal abundance on the creation of their progenitors ( e.g. , @xcite ) , or sought to explain environmental patterns in terms of a preference for metal - poor galaxies ( e.g. , @xcite ; @xcite ; @xcite ) .
here we have shown that the preference of low - redshift sn ic - bl and @xmath14 lgrbs for galaxies with high stellar - mass and star - formation - rate densities , compared to star - forming galaxies of similar stellar mass , is separate from a preference for low metal abundances .
observations of extragalactic star formation have suggested that dense stellar clusters may form with greater efficiency in regions of dense star formation @xcite .
a prospective explanation for the host - galaxy patterns we find is the efficient formation of tight massive binary progenitor systems in such densely star - forming environments . alternatively ,
if sn ic - bl and lgrb progenitors have greater stellar masses than those of core - collapse sn with typical ejecta velocities , a top - heavy imf in galaxies with high stellar - mass density and star - formation density could account for these same patterns .
however , such a top - heavy imf would need to be reconciled with evidence instead for a bottom - heavy imf in massive galaxies that were densely star forming while assembling their mass @xcite .
the construction of ground - based telescopes with large aperture ( e.g. , the thirty meter telescope ; giant magellan telescope ) and the james webb space telescope will make possible observations of galaxies formed shortly after reionization , but the faintest galaxies will remain beyond their sensitivity .
lgrbs , whose @xmath7-ray emission is currently visible to @xmath87 @xcite , may provide an effective approach to probing the detailed star - forming conditions in these early galaxies .
we thank sandra savaglio who , as referee , provided insightful comments and suggestions .
we additionally acknowledge useful discussions about measurements and host galaxies with jarle brinchmann , ori fox , david elbaz , john graham , matt lehnert , steven stahler , paul crowther , and josh bloom .
a.v.f.s group at uc berkeley has received generous financial assistance from the christopher r. redlich fund , the tabasgo foundation , weldon wood , and nsf grant ast-1211916 , as well as from nasa/_hst _ grant ar-12850 from the space telescope science institute , which is operated by the association of universities for research in astronomy ( aura ) , inc .
, under nasa contract nas5 - 26555 .
this research has made use of the ghosts database ( www.grbhosts.org ) , which is partly funded by nasa/_spitzer _ rsa agreement no . | spectra of broad - lined type ic supernovae ( sn ic - bl ) , the only kind of sn observed at the locations of long - duration gamma - ray bursts ( lgrbs ) , exhibit wide features indicative of high ejecta velocities ( @xmath0 ) .
we study the host galaxies of a sample of 245 low - redshift ( @xmath1 ) core - collapse sn , including 17 sn ic - bl , discovered by galaxy - untargeted searches , and 15 optically luminous and dust - obscured @xmath2 lgrbs .
we show that , in comparison with sdss galaxies having similar stellar masses , the hosts of low - redshift sn ic - bl and @xmath2 lgrbs have high stellar - mass and star - formation - rate densities .
core - collapse sn having typical ejecta velocities , in contrast , show no preference for such galaxies .
moreover , we find that the hosts of sn ic - bl , unlike those of sn ib / ic and sn ii , exhibit high gas velocity dispersions for their stellar masses .
the patterns likely reflect variations among star - forming environments , and suggest that lgrbs can be used as probes of conditions in high - redshift galaxies .
they may be caused by efficient formation of massive binary progenitors systems in densely star - forming regions , or , less probably , a higher fraction of stars created with the initial masses required for a sn ic - bl or lgrb .
finally , we show that the preference of sn ic - bl and lgrbs for galaxies with high stellar - mass and star - formation - rate densities can not be attributed to a preference for low metal abundances but must reflect the influence of a separate environmental factor . |
data from national sex surveys @xcite provide quantitative information on the number of sexual partners , the degree @xmath0 , of an individual . usually , surveys involve a random sample of the population stratified by age , economical and cultural level , occupation , marital status , etc .
the respondents are asked to provide information on sexual attitudes such as the number of sex partners they have had in the last 12 months or in their entire life .
although in most cases the response rate is relatively small , the information gathered is statistically significant and global features of sexual contact patterns can be extracted . in particular
, it turns out that the number of heterosexual partners reported from different populations is well described by power - law sf distributions .
table [ tab : scn ] summarizes the main results of surveys conducted in sweden , united kingdom , zimbabwe and burkina faso @xcite . the first thing to notice is the gender - specific difference in the number of sexual acquaintances @xcite .
this is manifested by the existence of two different exponents in the sf degree distributions , one for males ( @xmath9 ) and one for females ( @xmath10 ) .
interestingly enough , the predominant case in table [ tab : scn ] , no matter whether data refers to time frames of 12 months or to entire life , consists of one exponent being smaller and the other larger than 3 .
this is certainly a borderline case that requires further investigation on the value of the epidemic threshold .
the differences found in the two exponents @xmath10 and @xmath9 have a further implication for real data and mathematical modeling . in an exhaustive survey , able to reproduce the whole network of sexual contacts , the total number of female partners reported by men should equal the total number of male sexual partners reported by women .
mathematically , this means that the number of links ending at population @xmath11 ( of size @xmath12 ) equals the number of links ending at population @xmath13 ( of size @xmath14 ) , which translates into the following closure relation : @xmath15 assuming that the degree distributions for the two sets are truly scale - free , then @xmath16 , with the symbol @xmath17 standing for the gender @xmath18 , and @xmath19 being the minimum degree .
moreover , if @xmath20 and @xmath21 for any @xmath17 , eq .
( [ closure ] ) gives the relation between the two population sizes as @xmath22 which implies that the less heterogeneous ( in degree ) population must be larger than the other one .
[ cols="^,^,^,^,^,^,^,^,^",options="header " , ] although both epidemic thresholds , @xmath23 and @xmath24 , tend to zero as the population goes to infinity , the scaling relations , @xmath25 and @xmath26 , are characterized by two different exponents , @xmath27 and @xmath28 .
table [ tab : exp ] reports the expression of these two exponents as a function of @xmath10 and @xmath9 , showing that @xmath28 is always smaller than @xmath27 .
in particular , for the most common case ( see table [ tab : scn ] ) , i.e. when one degree distribution exponent is in the range @xmath292,3]$ ] , and the other one is larger than @xmath30 , the value of @xmath28 found for bipartite networks is two times smaller than @xmath27 . as a consequence , the results show that in finite bipartite populations the onset of the epidemic takes place at larger values of the spreading rate .
in other words , it could be the case that for a given transmission probability , in the unipartite representation shown in fig .
[ fig1 ] ( b ) the epidemic would have survived infecting a fraction of the population , while when only crossed infections are allowed , as in fig .
[ fig1 ] ( a ) , the same disease would not have produced an endemic state .
moreover , the difference between the epidemic thresholds predicted by the two approaches increases with the system size .
this dependency is shown in fig .
[ fig2 ] , where we have reported , as a function of the system size , the critical thresholds obtained by numerically solving eqs .
( [ threshold ] ) and ( [ threshold2 ] ) with the values of @xmath9 and @xmath10 found for the lifetime distribution of sexual partners in sweden @xcite and u.k @xcite .
to check the validity of the analytical arguments and also to explore the dynamics of the disease above the epidemic threshold , we have conducted extensive numerical simulations of the sis model in bipartite and unipartite computer - generated networks .
bipartite and unipartite graphs of a given size are built up ( see methods section ) having the same degree distributions , @xmath31 and @xmath32 , and thus they only differ in the way the nodes are linked .
a fraction of infected individuals is initially randomly placed on the network and the sis dynamics is evolved : at each time step susceptible individuals get infected with probability @xmath33 if they are connected to an infectious one , and get recovered with probability @xmath34 ( hence , the effective transmission probability is @xmath35 ) .
after a transient time , the system reaches a stationary state where the total prevalence of the disease , @xmath36 , is measured ( see methods ) .
the results are finally averaged over different initial conditions and network realizations .
[ fig3 ] shows the fraction of infected individuals as a function of @xmath37 for several system sizes and for the bipartite ( ( a ) and ( b ) ) and unipartite ( ( c ) and ( d ) ) graphs . in this figure
, the infection probability @xmath38 has been rescaled by the theoretical value @xmath39 given by eq .
( [ threshold ] ) .
the purpose of the rescaling is twofold .
first it allows to check the validity of the theoretical predictions and , at the same time , it provides a clear comparison of the results obtained for bipartite networks with those obtained for the unipartite case .
again we have used the values of @xmath9 and @xmath10 extracted from the lifetime number of sexual partners reported for sweden and u.k .
@xcite . fig .
[ fig3 ] indicates that the analytical solution , eq .
( [ threshold ] ) , is in good agreement with the simulation results for the two - gender model formulation .
conversely when the bipartite nature of the underlying graph is not taken into account , the epidemic threshold is underestimated , being @xmath40 smaller than @xmath41 .
in addition to this , the error in the estimation grows as the population size increases , in agreement with our theoretical predictions .
the inclusion of the bipartite nature of contact networks to describe crossed infections in the spread of stds in heterosexual populations is seen to affect strongly the epidemic outbreak and leads to an increase of the epidemic threshold .
our results show that , even in the cases when the epidemic threshold vanishes in the infinite network size limit , the epidemic incidence in finite populations is less dramatic than actually expected for unipartite scale - free networks .
the results also point out that the larger the population , the greater the gap between the epidemic thresholds predicted by the two models , therefore highlighting the need to accurately take into account all the available information on how heterosexual contact networks look like .
our results also have important consequences for the design and refinement of efficient degree - based immunization strategies aimed at reducing the spread of stds .
in particular , they pose new questions on how such strategies have to be modified when the interactions are further compartmentalized by gender and only crossed infections are allowed .
we finally stress that the present approach is generalizable to other models for disease spreading ( _ e.g. _ the `` _ susceptible - infected - removed _ '' model ) and other processes where crossed infection in bipartite networks is the mechanism at work .
synthetic bipartite networks construction starts by fixing the number of males , @xmath42 and the two exponents @xmath9 and @xmath10 of the power - law degree distributions corresponding to males and females respectively .
the first stage consists of assigning the connectivity @xmath43 ( @xmath44, ... ,@xmath12 ) to each member of the male population by generating @xmath45 random numbers with probability distribution @xmath46 ( @xmath47 , with @xmath48 ) .
the sum of these @xmath45 random numbers fixes the number of links @xmath49 of the network .
the next step is to construct the female population by means of an iterative process . for this purpose
we progressively add female individuals with a randomly assigned degree following the distribution @xmath50 ( @xmath51 , with @xmath48 ) .
female nodes are incorporated until the total female connectivity reaches the number of male edges , @xmath52 . in this way one
sets the total number of females @xmath14 .
once the two sets of @xmath12 males and @xmath14 females with their corresponding connectivities are constructed each one of the @xmath49 male edges is randomly linked to one of the available female edges avoiding multiple connections .
finally those few female edges that did not receive a male link in the last stage are removed and the connectedness of the resulting network is checked .
synthetic unipartite networks has been constructed in two ways .
the simplest one consists of taking the two sets of @xmath12 males and @xmath14 females constructed for the bipartite network and apply a rewiring process to the entire population , _ i.e. _ allowing links between individuals of the same sex . in the second method , a set of @xmath53 individuals whose connectivities
are randomly assigned following the degree distribution @xmath54 is generated before applying a wiring process between all pairs of edges . in both methods the wiring process avoids multiple and self connections and those isolated edges that remain at the end of the network construction are removed .
the connectedness of the networks is also checked .
montecarlo simulations of sis dynamics are performed using networks of sizes ranging from @xmath55 to @xmath56 .
the initial fraction of infected nodes is set to @xmath57 of the network size .
the sis dynamics is initially evolved for a time typically of @xmath58 time - steps and after this transient the system is further evolved over consecutive time windows of @xmath59 steps . in these time windows
we monitor the mean value of the number of infected individuals , @xmath60 .
the steady state is reached if the absolute difference between the average number of infected individuals of two consecutive time windows is less than @xmath61 .
we thank k.t.d .
eames and j.m .
read for their useful suggestions .
is supported by mec through the ramn y cajal program .
this work has been partially supported by the spanish dgicyt projects fis2006 - 12781-c02 - 01 and fis2005 - 00337 , and by the italian to61 infn project .
fenton , k. a. , korovessis , c. , johnson , a. m. , mccadden , a. , mcmanus , s. , wellings , k. , mercer , c. h. , carder , c. , copas , a. j. , nanchahal , k. , macdowall , w. , ridgway , g. , field , j. & erens , b. ( 2001 ) _ the lancet _ * 358 * , 1851 - 1854 . | the spread of sexually transmitted diseases ( _ e.g .
chlamydia , syphilis , gonorrhea , hiv _ ) across populations is a major concern for scientists and health agencies . in this context
, both data collection on sexual contact networks and the modeling of disease spreading , are intensively contributing to the search for effective immunization policies . here , the spreading of sexually transmitted diseases on bipartite scale - free graphs , representing heterosexual contact networks , is considered .
we analytically derive the expression for the epidemic threshold and its dependence with the system size in finite populations .
we show that the epidemic outbreak in bipartite populations , with number of sexual partners distributed as in empirical observations from national sex surveys , takes place for larger spreading rates than for the case in which the bipartite nature of the network is not taken into account .
numerical simulations confirm the validity of the theoretical results .
our findings indicate that the restriction to crossed infections between the two classes of individuals ( males and females ) has to be taken into account in the design of efficient immunization strategies for sexually transmitted diseases .
disease spreading has been the subject of intense research since long time ago @xcite .
on the one hand , epidemiologists have developed mathematical models that can be used as a guide to understanding how an epidemic spreads and to design immunization and vaccination policies @xcite .
on the other hand , data collections have provided information on the local patterns of relationships in a population . in particular , persons who may have come into contact with an infectious individual
are identified and diagnosed , making it possible to contact - trace the way the epidemic spreads , and to validate the mathematical models . however , up to a few years ago , some of the assumptions at the basis of the theoretical models were difficult to test .
this is the case , for instance , of the complete network of contacts -the backbone through which the diseases are transmitted . with the advent of modern society ,
fast transportation systems have changed human habits , and some diseases that just a few years ago would have produced local outbreaks , are nowadays a global threat for public health systems .
a recent example is given by the severe acute respiratory syndrome ( sars ) , that spread very fast from asia to north america a few years ago @xcite . therefore , it is of utmost importance to carefully take into account as much details as possible of the structural properties of the network on which the infection dynamics occurs .
strikingly , a large number of statistical properties have been found to be common in the topology of real - world social , biological and technological networks @xcite .
of particular relevance because of its ubiquity in nature , is the class of complex networks referred to as scale - free ( sf ) networks . in sf networks ,
the number of contacts or connections of a node with other nodes in the system , the degree ( or connectivity ) @xmath0 , follows a power law distribution , @xmath1 .
recent studies have shown the importance of the sf topology on the dynamics and function of the system under study @xcite .
for instance , sf networks are very robust to random failures , but at the same time extremely fragile to targeted attacks of the highly connected nodes @xcite . in the context of disease spreading , sf contact networks lead to a vanishing epidemic threshold in the limit of infinite population when @xmath2 @xcite .
this is because the exponent @xmath3 is directly related to the first and second moment of the degree distribution , @xmath4 and @xmath5 , and the ratio @xmath6 determines the epidemic threshold above which the outbreak occurs .
when @xmath7 , @xmath4 is finite while @xmath5 goes to infinity , that is , the transmission probability required for the infection to spread goes to zero .
conversely , when @xmath8 , there is a finite threshold and the epidemic survives only when the spreading rate is above a certain critical value .
the concept of a critical epidemic threshold is central in epidemiology .
its absence in sf networks with @xmath7 has a number of important implications in terms of prevention policies : if diseases can spread and persist even in the case of vanishingly small transmission probabilities , then prevention campaigns where individuals are randomly chosen for vaccination are not much effective @xcite .
our knowledge of the mechanisms involved in disease spreading as well as on the relation between the network structure and the dynamical patterns of the spreading process has improved in the last several years @xcite .
current approaches are either individual - based simulations @xcite or metapopulation models where network simulations are carried out through a detailed stratification of the population and infection dynamics @xcite . in the particular case of sexually transmitted diseases ( stds ) , infections occur within the unique context of sexual encounters , and the network of contacts @xcite is a critical ingredient of any theoretical framework . unfortunately , ascertaining complete sexual contact networks in significatively large populations is extremely difficult .
however , here we show that it is indeed possible to make use of known global statistical features to generate more accurate predictions of the critical epidemic threshold for stds . |
the isospin constraints on the @xmath1 hadronic decay modes are known for the @xmath2 and @xmath3 modes@xcite . for the @xmath4 final states ,
only the simplest decay mode @xmath5 has been studied . using the formalism of symmetry classes introduced by pais@xcite
, we generalize the relations for an arbitrary value of @xmath6 and give a geometrical representation of the constraints .
the @xmath7 system produced by a @xmath1 decay has isospin 1 ; the possible values of the @xmath8 isospin @xmath9 are 0 and 1 and the isospin @xmath10 of the @xmath6 pion system is 1 for @xmath11 and 0 , 1 or 2 for @xmath12 .
since there is no second - class current in the standard model , interferences between amplitudes with @xmath13 and @xmath14 vanish in the partial widths @xcite .
therefore we have the relation @xmath15 which is true for each charge configuration of the @xmath2 system , and , since @xmath16 , @xmath17 using the most easily observable states .
the amplitudes are classified by the values of @xmath9 and @xmath10 . to complete the classification
, we use the isospin symmetry class@xcite of the @xmath2 system i.e. the representation of the permutation group @xmath18 to which belongs the state .
it is characterized by the lengths of the three rows of its young diagram @xmath19 . due to the pauli principle ,
the momentum and isospin states have the same symmetry .
thus integration over the momenta kills the interferences between amplitudes in different classes and there is no contribution from them in the partial widths .
since @xmath20 and @xmath21 amplitudes belong to different symmetry classes@xcite , their interferences vanish .
the presence of @xmath22 amplitudes makes the problem more intricate since they share symmetry properties with some @xmath21 or @xmath20 amplitudes @xcite .
for instance , in the case @xmath23 the symmetry class @xmath24 is shared by @xmath20 and @xmath22 ; in the case @xmath25 , the symmetry class @xmath26 is shared by @xmath21 and @xmath22 .
therefore the allowed domains in the space of the charge configuration fractions ( @xmath27 ) must be determined separately for each symmetry class and @xmath9 , taking interferences into account when necessary .
the complete allowed domain is the convex hull of the sub - domains corresponding to the different @xmath9 and symmetry classes and its projections are the convex hulls of their projections . for @xmath28 , which is always true in a @xmath1 decay , the isospin values and the symmetry class characterize unambiguously the amplitude properties @xcite , therefore there is , at most , one interference term per class .
the sub - domain , for such a symmetry class associated with two different @xmath10 values , is then a two - dimensional one since the partial widths @xmath29 are linear functions of three quantities : the sums of squared amplitudes for the two values of @xmath10 and the interference term .
its boundary is determined by the schwarz s inequality @xcite .
this boundary is an ellipse ; it can be parametrized by writing the sums of squared amplitudes for the two values of @xmath10 as @xmath30/2\,$ ] and the largest interference term allowed by the schwarz s inequality as @xmath31 , where the coefficient @xmath32 depends on the coupling coefficients .
the most general domain is hence the convex hull of a set of points corresponding to the symmetry classes without @xmath22 and a set of ellipses .
the cases @xmath23 and @xmath25 are presented in detail in the following sections .
they both have the property that only one symmetry class is associated with two isospin values .
higher values of @xmath6 are not expected , for some time , to be of experimental interest .
the possible states that can be observed for a @xmath34 decay are the following : @xmath35 + @xmath36 + @xmath37 + @xmath38 + @xmath39 + @xmath40 + @xmath41 .
as mentioned before , not all the corresponding partial widths are independent and we can use the four fractions : @xmath42 , @xmath43 , @xmath44 and @xmath45 , whose sum is equal to 1 , to describe the possible charge configurations in a three - dimensional space .
the ratio @xmath46 is a free parameter independent of the charge configuration fractions .
the partial widths for all the charge configurations can be expressed as functions of the positive quantities @xmath47}$ ] which are the sums of the squared absolute values of the amplitudes with the given values of the isospins and the interference term @xmath48 of the @xmath49 and @xmath50 amplitudes . with only two pions
the coefficients are readily obtained from a clebsch - gordan table and we get @xmath51}+{{\textstyle}1{\over\textstyle}2}s_{[1,1]}+{{\textstyle}3{\over\textstyle}10}s_{[1,2]}\nonumber\\ \gamma_{k^+\bar k^0\pi^-\pi^-}&=&{{\textstyle}6{\over\textstyle}10}s_{[1,2]}\\ \gamma_{k^0 k^-\pi^+\pi^-}&=&{{\textstyle}2{\over\textstyle}3}s_{[1,0]}+{{\textstyle}1{\over\textstyle}2}s_{[1,1]}+ { { \textstyle}1{\over\textstyle}30}s_{[1,2]}+{\cal i}\nonumber\\ \gamma_{k^0 k^-\pi^0\pi^0}&=&{{\textstyle}1{\over\textstyle}3}s_{[1,0]}+{{\textstyle}2{\over\textstyle}30}s_{[1,2]}-{\cal i}.\nonumber \end{aligned}\ ] ] the partial width @xmath52 is the sum over the charge configurations : @xmath53 the schwarz s inequality bounding the interference term @xmath48 is @xmath54}s_{[1,2]}}.\ ] ] since the partial width @xmath55 is independent of the interference , the above equations are also true for the fractions @xmath56 , replacing the @xmath57 by the weights @xmath58 and normalizing the interference term .
for @xmath59 $ ] equal to @xmath60 $ ] and @xmath61 $ ] , the sub - domains are merely points we will refer to as @xmath60 $ ] and @xmath61 $ ] .
for the two interfering classes @xmath14 , @xmath62 ( @xmath63}=w_{[1,1]}=0 $ ] ) , the sub - domain is the two - dimensional domain in the plane @xmath64 bounded by the ellipse of equation @xmath65 : @xmath66 as explained before , the complete domain is the convex hull of this ellipse and the two points @xmath60 $ ] and @xmath61 $ ] .
calling @xmath67 the point of the ellipse for which @xmath68 , the domain is made of the tetrahedron having the points i , @xmath60 $ ] , @xmath61 $ ] and @xmath69 $ ] for vertices and of the two half - cones whose bases are the halves of the ellipse delimited by the points @xmath67 and @xmath69 $ ] and whose vertices are the points @xmath61 $ ] and @xmath60 $ ] respectively . + [ cols="^,^ " , ] for practical purposes it is useful to draw the projections of the domain on two - dimensional planes .
the method is very simple : we first draw the projection of the ellipse on the plane and then the tangents to the projected ellipse from the projections of the points @xmath60 $ ] and @xmath61 $ ] .
a first simple example is the projection on the plane @xmath70 , with @xmath71 and @xmath72 . here
the ellipse projection is a mere segment and the projected domain is the polygon whose vertices are the ( projected ) points @xmath60 $ ] , @xmath61 $ ] , @xmath69 $ ] and @xmath73 $ ] .
more interesting is the projection @xmath71 , @xmath74 , since the two final states @xmath39 and @xmath40 have the same topology : one @xmath75 and three charged hadrons .
the complement @xmath76 is the fraction @xmath45 of decays with two @xmath77 s .
the projected ellipse has vertical tangents at the points @xmath69 $ ] and @xmath73 $ ] .
it is also tangent to the line @xmath78 at the point @xmath67 ( @xmath79 ) , for which @xmath80}=w_{[1,2]}/5 $ ] since @xmath45 can be 0 , because of the interference , only when the two contributions have the same modulus .
the second tangent from @xmath60 $ ] touch the ellipse at the point of coordinates @xmath81 and @xmath82 .
the allowed domain is shown on fig.1 .
the main constraint is the inequality @xmath83 corresponding to the second tangent .
one can see from the plot that a large value of the ratio @xmath84 implies the dominance of @xmath12 and a small value the dominance of @xmath11 .
with one dominant isospin for the @xmath8 system , the ratio @xmath46 measures the proportions of the two g - parities i.e. the contributions of axial and vector currents .
the final states for the decay @xmath86 are : @xmath87 + @xmath88 + @xmath89 + @xmath90 + @xmath91 + @xmath92 + @xmath93 .
the relations between the @xmath94 and @xmath95 final states are the same as in the decay .
thus the charge configurations are described in a four - dimensional space by the five fractions : @xmath96 , @xmath97 , @xmath98 , @xmath99 and @xmath100 .
we shall label the amplitudes by the two isospin values and the symmetry class : @xmath101 $ ] .
the relations between the partial widths for the charge configurations and the amplitudes use both clebsch - gordan coefficients and the similar coefficients for the symmetry classes @xcite . with the notations defined in the previous section
, they can be written @xmath102}+{{\textstyle}1{\over\textstyle}2}s_{[0,(210)^1]}+{{\textstyle}1{\over\textstyle}10}s_{[1,(300)^1]}+ { { \textstyle}1{\over\textstyle}4}s_{[1,(210)^1]}+{{\textstyle}3{\over\textstyle}20}s_{[1,(210)^2]}+{\cal i}\nonumber\\[.2 cm ] 2\gamma_{k^+ k^-\pi^-\pi^-\pi^+}&=&\gamma_{k^0\bar k^0\pi^-\pi^-\pi^+}+\gamma_{k^+ k^-\pi^-\pi^-\pi^+}\nonumber\\ & = & { { \textstyle}4{\over\textstyle}5}s_{[0,(300)^1]}+{{\textstyle}1{\over\textstyle}2}s_{[0,(210)^1]}+{{\textstyle}2{\over\textstyle}5}s_{[1,(300)^1]}+ { { \textstyle}1{\over\textstyle}4}s_{[1,(210)^1]}+{{\textstyle}3{\over\textstyle}20}s_{[1,(210)^2]}-{\cal i}\nonumber\\[.2 cm ] \gamma_{k^0k^-\pi^+\pi^-\pi^0}&=&s_{[1,(111)^0]}+{{\textstyle}1{\over\textstyle}5}s_{[1,(300)^1]}+ { { \textstyle}1{\over\textstyle}2}s_{[1,(210)^1]}+{{\textstyle}1{\over\textstyle}10}s_{[1,(210)^2]}+{{\textstyle}2{\over\textstyle}\sqrt{3}}\,{\cal i}\nonumber\\ \gamma_{k^0k^-\pi^0\pi^0\pi^0}&=&{{\textstyle}3{\over\textstyle}10}s_{[1,(300)^1]}\\ \gamma_{k^+\bar k^0\pi^-\pi^-\pi^0}&=&{{\textstyle}3{\over\textstyle}5}s_{[1,(210)^2]}.\end{aligned}\ ] ] here the partial width @xmath103 is a function of the interference term : @xmath104 and the schwarz s inequality reads @xmath105 $ ] , @xmath106 $ ] , @xmath107 $ ] and @xmath108 $ ] . for the interfering classes
@xmath109 $ ] and @xmath110 $ ] , the plane of the two - dimensional sub - domain is determined by the two relations : @xmath111 and @xmath1122(1 + 1)_k^+k^-^-^0 ^ 0 + 2(1 - 1)_k^+k^-^+^-^--_k^0k^-^+^-^0 - 13_k^+k^0 ^ -^-^0=0.@xmath112 the boundary is given by the saturation of the schwarz s inequality @xmath113 .
the complete , four - dimensional domain is the convex hull of this two - dimensional sub - domain and the four points .
two - dimensional projections can be constructed by the same method that we used in the previous section .
the example shown on fig .
2 takes @xmath114 fraction of decays without neutral hadron and @xmath115 fraction of decays with one @xmath77 and one @xmath75 . the complement @xmath76 is the fraction of decays with two or three @xmath77 s .
the ellipse bounding the projection of the two - dimensional sub - domain goes through the two points @xmath109 $ ] and @xmath110 $ ] .
it is tangent to the lines @xmath116 and @xmath78 , since , due to the interference , the contributions of the two classes of amplitudes @xmath109 $ ] and @xmath110 $ ] can cancel out in @xmath117 or in @xmath118 and there is no contribution to @xmath119 from @xmath109 $ ] and @xmath110 $ ] amplitudes .
the boundary of the projected domain is made of two arcs of the curve , four tangents and a segment of the line @xmath120 . due to the interference ,
the ratio of the number of decays with two or three @xmath77 s over the number of decays without @xmath77 is only bounded by 0 and 1 .
the fraction of decays with two or three @xmath77 s is always lower than 1/2 . to distinguish two @xmath77 from three @xmath77 decays ,
we can use a third coordinate @xmath121 .
the three - dimensional domain is the cone having for basis the above described contour in the @xmath70 plane and , for vertex , the point @xmath107 $ ] .
we have presented the complete isospin constraints on the @xmath0 decay modes in the space of charge configurations with some details in the cases @xmath23 and @xmath25 .
gilman and s.h .
rhie , phys .
d 31 ( 1985 ) 1066 a. roug , z. phys .
c70 ( 1996 ) 109 a. pais , ann .
( 1960 ) 548 a. pais , ann .
( 1963 ) 274 h. pilkuhn , nucl .
( 1961 ) 168 l. michel , nuovo cim . 22 ( 1961 ) 203 | the construction of the complete isospin relations and inequalities between the possible charge configurations of a @xmath0 decay mode is presented .
detailed applications to the cases of two and three pions are given .
= -11 mm = -1.5 mm = -1.5 mm + x - lpnhe 97/08 + |
the evidence suggest the problem of a single model explaining both galactic scale and cosmological scale observations is fundamental @xcite . among a variety of models that have been suggested to link cosmological scale and
galactic scale observations are models using a scalar field .
a scalar field has been linked to dark matter @xcite , a cosmological model @xcite , the rotational curves of spiral galaxies @xcite , and axisymmetric galaxies @xcite .
the great majority of elliptical galaxies are observed to be much poorer in cool gas and hydrogen than spiral galaxies of comparable luminosity @xcite .
the bulk of the interstellar matter ( ism ) in spiral galaxies is h and hydrogen . in elliptical galaxies ,
the bulk of the ism consists of hot plasma distributed approximately spherically rather than in a thin disk @xcite .
a characteristic of elliptical galaxies not found in spiral galaxies is that the x - ray surface brightness is nearly proportional to the optical surface brightness @xcite .
the study of dust lanes suggests that gas and dust are falling into elliptical and lenticular galaxies @xcite and are formed internally in spiral galaxies @xcite .
some evidence has been presented that suggests irregular galaxies will settle down to being a normal elliptical galaxy @xcite . in low surface brightness ( lsb )
spiral galaxies , the outer rotation curve ( rc ) generally rises ( * ? ? ?
* and references therein ) .
in contrast , `` ordinary '' elliptical galaxies , with luminosities close to the characteristic @xmath6 ( = 2.2 @xmath7 in b band solar units for a hubble constant @xmath8 km s@xmath9 mpc@xmath9 ) show a nearly keplerian decline with radius outside @xmath10 , where @xmath11 is the galaxy s `` effective radius '' enclosing half its projected light @xcite . @xcite and @xcite
provides an overview of the current state of knowledge of rcs of spiral galaxies .
the rcs of spiral galaxies have a high rotation velocity of over 1000 km s@xmath9 near the nucleus @xcite .
@xcite and @xcite have observed keplerian motion to within one part in 100 in elliptical orbits of stars that are from a few 100 pc to a few 1000 pc from the center of spiral galaxies .
the keplerian characteristic decline in rotation velocity is sometimes seen in h@xmath12 rcs .
this is followed by gradual rise to a knee or sharp change of slope at a rotation velocity of less than 300 km s@xmath9 .
the outer rc is beyond the knee .
interacting galaxies often show perturbed outer rcs .
the outer part of an rc is often relatively flat with rising and falling rcs occasionally being found .
the particles most often measured in the disk region of a galaxy are hydrogen gas by h observation and stars by observing the h@xmath12 line .
the particles being measured in the inner bulge region are stars by observation of h@xmath12 , co , and other spectral lines .
also , the rc differs for different particles .
for example , the h and h@xmath12 rcs for @xcite differ in the outer bulge and approach each other in the outer disk region .
@xcite found that the mass in the central region of spiral galaxies is 0.0012 of the mass of the bulge .
@xcite reported that about 0.1% of a spiral galaxy s luminous mass is at the center of galaxies and that the density of supermassive black holes sbh s in the universe agrees with the density inferred from observation of quasars .
@xcite found similar results in their study of the relationship of the mass @xmath13 of the central supermassive black hole and the velocity dispersion @xmath14 ( @xmath15 relation ) .
@xcite found a tight relation between rotation velocity @xmath16 in the outer disk region and bulge velocity dispersion @xmath17 ( @xmath18 ) which is strongly supporting a relationship of a center force with total gravitational mass of a galaxy .
@xcite showed @xmath13 of agn galaxies and their bulge luminosity follow the same relationships as their ordinary ( inactive ) galaxies , with the exception of narrow line agn .
@xcite found correlations between @xmath13 and structural parameters of elliptical galaxies and bulges .
either the dynamics of many spiral galaxies are producing the same increase of mass in the center at the same rate , a feedback controlled mechanism exists to evaporate the mass increase that changes as the rate of inflow changes as suggested by @xcite , or sufficient mass is ejected from spiral galaxies to maintain the correlations .
the rcs imply the dynamics of spiral galaxies differ so the former is unlikely .
the study of dust lanes suggests mass is ejected from spiral galaxies .
the following is taken from @xcite .
galaxies in groups and clusters ( `` clusters '' ) are much more likely to be elliptical or lenticular than in the field .
spiral galaxies lie farther from the center of clusters than do elliptical galaxies .
the fraction @xmath19 of galaxies that are elliptical galaxies in clusters varies from 15% to 40% .
clusters with a large value of @xmath19 tend to have a regular , symmetric appearance , often with a large cd galaxy at its center .
clusters with a low value of @xmath19 generally have a ratty appearance .
the fraction @xmath20 of spiral galaxies in centrally - concentrated clusters increases with radius @xmath21 from the center of the cluster .
the observations are consistent with the model of their being no spiral galaxies in the cores of regular clusters .
the lenticular ( s0 ) galaxies become increasingly dominant at small radii . nearer
the core the fraction @xmath22 of s0 galaxies declines sharply as @xmath19 increases sharply . also , the @xmath19 increases and @xmath20 decreases as the projected number @xmath23 density of galaxies increases .
there appears to be a close relation between @xmath23 and @xmath21 .
the morphology of galaxies appears strongly correlated with the current surrounding density .
a galaxies radius within a cluster appears to be the primary factor that dictates its morphology except where a spiral galaxy has a nearby companion .
the nearby companion has a disproportionate probability of being elliptical .
studies of h in spiral galaxies show the average h deficiency varies monotonically with @xmath21 .
most rich clusters of galaxies are filled with plasma with temperature @xmath24 k ( intracluster medium ) .
the intracluster medium mass exceeds the total mass in the luminous parts of cluster galaxies .
the intracluster medium is composed of hydrogen - helium plasma and significant numbers of heavy - element ions , which were probably formed in galaxies .
the hot plasma may be condensing into galaxies ( s0 and e ) that are found at the center of clusters .
that the redshift @xmath0 of emitted photons from galaxies generally increases with the extragalactic distance @xmath2 ( mpc ) between the emitting galaxy and observer is well known ( @xmath25 relationship ) .
currently fashionable , cosmological models attributes @xmath0 to a doppler shift of a light wave .
the assumption of homogeneity in cosmological scale volumes is compatible with either a static galaxy distribution or with a very special velocity field obeying the hubble law @xcite @xmath26 where @xmath27 ( km s@xmath9 mpc@xmath9 ) is the hubble constant and @xmath28 ( km s@xmath9 ) is the speed of light .
the @xmath27 occupies a pivotal role in current cosmologies .
the methods of calculating supernova distances , the cosmological microwave background ( cmb ) power spectrum , weak gravitational lensing , cluster counts , baryon oscillation , expansion of the universe , and the fundamental aspects of the big bang model depend on @xmath27 and the hubble law . however , the determination of @xmath27 has a large uncertainty and different researchers calculate different values .
figure [ fig:1 ] shows the calculated redshift @xmath29 using eq .
( [ eq:1 ] ) , @xmath30 km s@xmath9 mpc@xmath9 , and the distance @xmath31 calculated using cepheid variable stars for 32 galaxies @xcite versus the measured galactocentric redshift @xmath32 .
the correlation coefficient of @xmath29 versus @xmath32 is 0.80 . using eq .
( [ eq:1 ] ) and @xmath2 calculated using cepheid variable stars for 32 galaxies @xcite versus the measured redshift @xmath32 .
the straight line is a plot of @xmath33 .
the circles indicate the data points for galaxies with ( l , b ) = ( 290@xmath34,75@xmath35).,scaledwidth=80.0% ] the deviation of the recession velocity of a galaxy from the straight line of the hubble law is ascribed to a `` peculiar velocity '' of the photon emitting galaxy relative to earth @xcite .
the deviation from the hubble law is the only means of determining the peculiar velocity of a galaxy . the average peculiar velocity for all galaxies
is assumed to be zero on the scale that the universe appears homogenous .
the 2dfgrs @xcite suggests this scale is @xmath36 .
the circles in fig .
[ fig:1 ] denote data for galaxies in the general direction of @xmath37 .
@xcite found the peculiar velocity field in the local supercluster is directed toward ( ) with @xmath38 at a speed of 331@xmath3941 km s@xmath9 .
this has been called the `` virgocentric infall '' .
is a peculiar , large , elliptical galaxy with strong x - ray emissions .
in addition , @xcite detected a quadrupolar tidal velocity field from spiral galaxy data in addition to the virgocentric infall pointing toward ( l , b ) = ( 308@xmath40 , 13@xmath41 ) at a speed of @xmath42 km s@xmath9 .
( ) at ( l , b ) = ( 310@xmath43 , 19@xmath43 ) and at a distance of @xmath44 mpc @xcite is a galaxy with properties @xcite similar to . @xcite found elliptical galaxies at distances in the 2000 - 7000 km s@xmath9 range are streaming toward a `` great attractor '' at @xmath45 . at ( l ,
b ) = ( 310@xmath43 , 2@xmath43 ) is a galaxy with properties similar to . in a more recent analysis
, @xcite suggested a bulk flow of 225 km s@xmath9 toward @xmath46 .
however , the total mass in these directions appears to be insufficient to account for the peculiar velocity fields using newtonian dynamics .
@xcite found a @xmath47 periodicity @xmath48 for approximately 70 qsos .
@xcite confirmed these claims of periodicity using various statistical tests on @xmath32 of 2164 qsos .
however , another claim @xcite that @xmath49 is periodic with a period of 0.206 was found to be unsupported .
* and references therein ) offered further evidence of periodic @xmath32 in qsos .
@xcite found `` discrete velocity periods '' of @xmath32 of spiral galaxies in clusters .
the `` discrete velocity periods '' of @xmath32 showed an octave , or doubling , nature .
* and references therein ) confirmed the existence of the `` discrete velocity periods '' of @xmath47 for 83 sci galaxies .
* and references therein ) presented evidence that redshifts from normal spiral galaxies have an `` intrinsic '' component that causes the appearance of peculiar velocities in excess of 5000 km s@xmath9 . in this
paper , galaxy and cluster observations are used to derive the characteristics of a scalar potential model ( spm ) .
the spm suggests spiral galaxies are sources of the scalar potential field and early type galaxies are sinks of the scalar potential field .
the cluster observations support the movement of matter from source galaxies to sink galaxies .
an equation is derived that recovers eq .
( [ eq:1 ] ) for cosmological distances from data that is anisotropic and inhomogeneous .
the resulting model is used to calculate redshift of particle photons for a sample of 32 galaxies with @xmath31 .
the calculated redshift @xmath50 has a correlation coefficient of 0.88 with @xmath32 .
further , the spm suggests the discrete variations in @xmath32 @xcite are consistent with the spm .
the object of this article is to derive a spm based on galaxy and cluster observations , to apply the spm by developing an equation for the calculation of @xmath0 , and to test the equation using observations of galaxies with a @xmath31 .
sink and source galaxies are discussed in section [ sec : sink ] . in section [ sec : model ] , the spm @xmath0 calculation equation is developed . the resulting model is used to calculate @xmath50 in section [ sec : results ] .
section [ sec : xfactor ] discusses the factors of the @xmath50 calculation .
section [ sec : tiff ] discusses the discrete variations in @xmath32 relative to the model of section [ sec : xfactor ] .
the discussion and conclusion is in section [ sec : disc ] .
the spm postulates the existence of a scalar potential @xmath51 ( erg ) field with the characteristics to cause the observed differences in spiral and elliptical galaxies .
the gradient of @xmath51 is proportional to a force @xmath52 ( dyne ) that acts on matter , @xmath53 where the arrow over a parameter denotes a vector , the @xmath54 is a proportionality constant analogous to the gravitational constant @xmath55 , and @xmath56 is a property of the test particle on which the @xmath52 acts . because the @xmath56 property of matter is currently unidentified ,
call the units of @xmath56 `` cs '' .
the spm suggests @xmath52 exerts a force to repel matter from spiral galaxies and to attract matter to early type galaxies .
therefore , the @xmath51 is highest at the center of spiral galaxies as a point source and lowest at the center of early type galaxies as a point sink .
call spiral galaxies sources and early type galaxies sinks . because the scalar field @xmath57 due to sources is highest at point sources , it obeys the inverse square law as does gravity .
therefore , @xmath58 where @xmath59 is a number representing the strength of the source analogous to the central mass of the gravitational potential , @xmath60 is the proportionality constant , and @xmath61 ( mpc ) is the distance from a source to the point at which @xmath51 is calculated . because the nature of @xmath59 is currently unidentified , call the units of @xmath59 `` q '' . the unit of @xmath60 is erg mpc q@xmath9 . because the scalar field @xmath62 due to sinks is lowest at point sinks
, it obeys the inverse square law as does gravity .
therefore , @xmath63 where @xmath64 ( q ) is a number representing the strength of the sink analogous to the central mass of the gravitational potential , @xmath65 ( erg mpc q@xmath9 ) is the proportionality constant , and @xmath66 ( mpc ) is the distance from a sink to the point at which @xmath51 is calculated .
the @xmath51 is the sum of the effects of all galaxies , @xmath67 where the subscript , lower case , italic , roman letters are indices ; `` p '' is the point where @xmath68 is evaluated ; @xmath69 ( mpc ) and @xmath70 ( mpc ) are the distances to the point where @xmath68 is evaluated from the @xmath71 source and @xmath72 sink , respectively ; and @xmath73 and @xmath74 are the number of source and sinks , respectively . in the universe , @xmath73 and @xmath74
are very large , perhaps effectively infinite .
the boundary condition of a very large universe produces considerable ambiguity in eq .
( [ eq:13b ] ) .
one way to resolve this `` olber s paradox '' type condition is to postulate that @xmath75 and @xmath76 .
call the volume with a simply connected surface containing equal source and sink strengths a `` cell '' .
this is analogous to the electric charges in an atom .
as the universe may be considered a collection of clusters , the spm considers the universe to be a collection of cells . ignoring the effect of neighboring sources and sinks ( the environment ) and the radial movement of a test particle ,
the tangential rotation velocity @xmath77 ( km s@xmath9 ) of a test particle in a galaxy becomes @xmath78 where in a source galaxy , @xmath79 in a sink galaxy , @xmath80 where , for simplicity , the mass of the test particle is assumed to be constant over time , @xmath81 ( kpc ) is the radial distance of the test particle , @xmath82 ( m@xmath83 ) is the mass inside the sphere of radius @xmath81 , @xmath84 indicates absolute value , and the inertial mass @xmath85 equals gravitational mass @xmath86 of the test particle @xcite .
because the outer rcs of elliptical galaxies are keplerian @xcite , the total force on mass must be centrally directed acting along the radius of the elliptical galaxy .
the @xmath52 and the gravitational force @xmath87 of the surrounding mass of the galaxy are directed radially inward .
thus , the @xmath88 of a sink galaxy appears greater than newtonian dynamical expectation as the `` virgocentric infall '' and `` great attractor '' phenomena require . in spiral galaxies
the various rotation velocity curves can result if @xmath52 is directed radially outward and @xmath87 is directed radially inward with a @xmath89 ratio varing with @xmath81 . the @xmath59 and @xmath64 terms of eqs .
( [ eq:52 ] ) and ( [ eq:52a ] ) mimic a hidden mass .
however , because @xmath82 is the total mass , the @xmath59 and @xmath64 terms are massless . because the h and h@xmath12 rcs differ , the @xmath89 factor must be different for different matter ( elemental ) types .
the @xmath56 can not be linearly related to the mass .
another characteristic of matter must be the characteristic of @xmath56 .
thus , the @xmath82 of a source galaxy is greater than newtonian dynamical expectation .
the nature of the @xmath59 may ( 1 ) be derived from a property of matter , ( 2 ) be independent of matter , or ( 3 ) be the cause both of @xmath51 and of matter . because the @xmath51 field must be repulsive of matter in source galaxies and attractive of matter in sink galaxies ,
the first possibility is unlikely .
if matter falls into a galaxy before a source is present , the galaxy is a sink .
if a @xmath51 field of a source exists before matter forms around it , the @xmath51 field prevents matter from forming a galaxy .
hence the second possibility is unlikely .
because sources exist only with spiral galaxies around them , the @xmath51 field and the @xmath82 of a spiral galaxy are formed concurrently .
therefore , @xmath90 , where @xmath91 is the total mass of the source galaxy .
similarly , @xmath92 , where @xmath93 is the total mass of the sink galaxy .
because sinks acquire their mass from neighboring source galaxies , if the sink is in a low galaxy population density environment , then local source galaxies yield little matter .
such sinks have little matter , have low luminosity , and have low @xmath94 . because of the low density of galaxies , the @xmath95 of neighboring galaxies is insignificant .
the extent of the @xmath95 field is limited to nearly the extent of the galaxy mass . because @xmath96 , measurements of lensing @xcite and velocities of satellites outside the sink galaxy measure more mass using newtonian dynamics than @xmath97 @xcite .
this is a good description of @xmath6 elliptical galaxies @xcite . if the sink is in a dense environment , then the sink has more matter , is bright , and has high @xmath98 .
the higher values of @xmath99 and @xmath100 compresses the matter ( stars ) well inside the extent of the @xmath95 field and the mass profile is steeper than low luminosity sinks . because of the high density of galaxies , the @xmath95 of neighboring galaxies is significant .
thus , the @xmath95 field is spatially more extended than the mass .
data from lensing , stellar dynamics @xcite , and velocities of satellites @xcite outside the mass concentration of the sink indicates much more mass than @xmath97 and the presence of a mass - like component more extended than the luminous component using newtonian dynamics .
this is a good description of bright elliptical galaxies .
these observations are consistent with the premise that irregular , elliptical , and lenticular galaxies are sinks and galaxies with spiral morphology are sources . because spiral galaxies form around sources and because cd galaxies are often found in the center of large clusters with elliptical galaxies , cd galaxies
were also considered sinks .
the spm suggests the photon is a particle .
the derivation of planck s black body radiation equation includes the proposition that the energy of a photon is discrete and composed of a number @xmath101 of basic energy / particle packets .
the calculated redshift @xmath50 is @xmath102 where @xmath103 is the wavelength , @xmath104 is the emitted @xmath101 , and @xmath105 is the observed @xmath101 .
because the energy of the photon is changed as it traverses the @xmath51 field , @xmath101 is changed by the amount of @xmath51 through which the photon travels . the change of @xmath101 per unit volume @xmath106 of @xmath51 through which the photon travels is posited to be proportional to @xmath107 entering @xmath106 , @xmath108 where @xmath109 is the proportionality constant .
the @xmath106 is @xmath110 where @xmath111 is the average @xmath51 in @xmath106 and @xmath112 is the cross section of @xmath106 which is the cross section of the photon particle .
posit that a photon has a minimum number @xmath113 of energy packets .
combining eqs .
( [ eq:3 ] ) and ( [ eq:4 ] ) yields , @xmath114 for the @xmath2 and change of @xmath101 considered herein , @xmath112 is considered a constant . for greater distances , where the total change of @xmath101 is relatively larger than considered herein , the @xmath112 is a function of @xmath101 and @xmath51 at the position of the photon .
the @xmath115 where d@xmath116 is the incremental distance @xmath116 ( mpc ) traveled by the photon .
the emitted luminosity of a galaxy is proportional to the flux of photons from the galaxy and is assumed to be isotropic .
other factors such as the k correction were considered too small for the @xmath2 of the sample . for a spiral galaxy to exist , @xmath100 must balance the @xmath99 .
matter is ejected until this balance is achieved .
therefore , the mass and , hence , the luminosity of a source galaxy is a first order approximation of @xmath59 .
therefore , @xmath117 where @xmath118 @xmath119 ( q s erg@xmath9 ) is a proportionality constant ; and @xmath120 ( mag . ) , @xmath121 ( mag . ) , and @xmath122 ( mag . ) are the absolute magnitude , apparent magnitude , and extinction , respectively , in a given band ( @xmath123 ) of the galaxy .
similarily , for sink galaxies , @xmath124 where @xmath125 ( q s erg@xmath9 ) is a proportionality constant .
the conservation of matter / energy implies for each source galaxy @xmath126 where @xmath127 ( @xmath128 s@xmath9 ) , @xmath129 ( @xmath128 s@xmath9 ) , @xmath130 ( @xmath128 s@xmath9 ) , @xmath131 ( erg s@xmath9 ) , and @xmath132 ( @xmath128/erg ) are the matter per second ( exclusive of photons ) entering the galaxy from other galaxies , the matter per second ( exclusive of photons ) emitted by the galaxy , the matter per second increase in the galaxy , the luminosity due to photons into the galaxy , and the proportionality constant to convert @xmath59 units to @xmath128 s@xmath9 , respectively .
matter is `` in '' a galaxy when it is in orbit around the source , where the dynamics of the matter is principally determined by the galaxy parameters .
similarly for sink galaxies , @xmath133 sum eqs .
( [ eq:9 ] ) and ( [ eq:9a ] ) over all source and sink galaxies , respectively . because the mass ejected from source galaxies go to sink galaxies , the combined @xmath134 . because the measured cmb radiation is nearly an ideal black body radiation @xcite , the sum of the @xmath135 terms for all sources equals the sum of the @xmath135 terms for all sinks . equating the sum of all matter emitted from source galaxies to the sum of all matter entering sink galaxies yields , @xmath136 where @xmath137 . because @xmath106 is a function of @xmath116 from the observer to a point along the line of sight and time @xmath138 , @xmath139 combining eqs .
( [ eq:7 ] ) , ( [ eq:6 ] ) , ( [ eq:8 ] ) , and ( [ eq:14 ] ) yields , @xmath140 where @xmath141 where : ( 1 ) relatively small terms such as terms involving the relative @xmath59 of the emitter and observer were ignored . ( 2 ) the @xmath142 is a constant for a given @xmath104 .
( 3 ) the @xmath143 . ( 4 ) the @xmath144 \mathrm{d } x$ ] to a first order approximation . ( 5 ) the @xmath145 ( erg@xmath9 mpc@xmath146 ) , @xmath147 ( mpc@xmath9 ) , @xmath148 ( erg@xmath9 mpc@xmath9 ) , @xmath149 ( erg@xmath9 ) , @xmath150 ( erg mpc@xmath9 ) , and @xmath151 ( erg@xmath9 mpc@xmath9 deg.@xmath9)are constants . ( 6 ) the @xmath51 is posited to be constant over time .
( 7 ) the relative velocity of the emitting and observing galaxies causes a change in @xmath106 , hence @xmath101 , and has three possible causes .
one is the expansion of our universe .
this component is linearly related to @xmath152 .
the second cause is due to the possible peculiar velocity of the milky way relative to the reference frame derived by summing over all sources and sinks .
another cause derives from the inaccuracy of defining the reference frame because the galaxies directly on the other side of the milky way center from earth are unobservable from earth .
the component @xmath153 deriving from the second and third causes is proportional to the cosine of the angular difference between the direction of the target galaxy and the direction of @xmath153 .
thus , @xmath154 where @xmath155 ( degrees ) and @xmath156 ( degrees ) are the galactic latitude and longitude , respectively , of the emitting galaxy ; and @xmath157 ( degrees ) and @xmath158 ( degrees ) are the galactic latitude and galactic longitude , respectively , of the direction of @xmath153 . setting @xmath159 considers @xmath160
was incorporated into the constants of eq .
( [ eq:27 ] ) and yields , @xmath161
the sample galaxies were selected from the ned database .
the selection criteria were that the heliocentric redshift @xmath162 be less than 0.03 and that the object be a galaxy .
the parameters obtained from the ned database included the galaxy name , @xmath156 , @xmath155 , @xmath162 , morphology , the @xmath163 was the b - band apparent magnitude @xmath164 ( mag . ) as defined by ned , and the galactic extinction @xmath165 ( mag . ) .
the @xmath166 was calculated from @xmath162 .
the 21-cm line width @xmath167 ( km s@xmath9 ) at 20 percent of the peak and the inclination @xmath168 ( arcdegrees ) between the line of sight and polar axis were obtained from the leda database when such data existed .
the constants to be discovered are the constants of eq .
( [ eq:26 ] ) , the cell limitation of @xmath69 and @xmath70 , and @xmath169 . calculating the constants
was done by making the following simplifications : ( 1 ) estimate the @xmath2 to the sample galaxies ( see appendix [ sec : initial ] ) .
( 2 ) galaxies with an unlisted morphology in the ned database were considered to have negligible effect .
an alternate method may be to assign a high value of @xmath170 to such galaxies .
this option was rejected because a large number of such galaxies were from the 2dfgrs and 2mass .
these surveys include only limited areas of the sky .
therefore , including the 2dfgrs and 2mass galaxies with unlisted morphology in the calculation would introduce a selection bias into the sample .
( 3 ) galaxies with an unlisted @xmath164 were assigned @xmath171 mag .
( 4 ) objects with @xmath172 in ned were assigned an @xmath173 mag .
( 5 ) all the sample galaxies were considered mature and stable .
( 6 ) galaxies with a spiral ( non - lenticular ) , barred , or ringed morphology in the ned database were considered sources .
all other galaxies were considered sinks .
the result was a sample of 22,631 source galaxies and 7,268 sink galaxies .
define the luminosity ratio @xmath174 as the ratio of the sum of the luminosities of all sources within a limiting distance @xmath175 divided by the sum of the luminosities of all sinks within @xmath175 , @xmath176 figure [ fig : lum ] shows a plot of @xmath174 versus @xmath175 for the 29,899 sample galaxies .
because the sample galaxies were limited to @xmath177 , the selection of galaxies for @xmath178 mpc was incomplete .
therefore , from eq .
( [ eq:28 ] ) . for galaxies with a distance @xmath2 less than the limiting distance @xmath175 versus @xmath175.,scaledwidth=80.0% ]
form the linear relation , @xmath179 where @xmath180 is the least squares slope and @xmath181 is the least squares intercept of the presumed linear relationship between @xmath50 and @xmath182 .
the constants of eq .
( [ eq:26 ] ) were adjusted to maximize the correlation coefficient of eq .
( [ eq:35 ] ) with @xmath183 and with @xmath184
. figure [ fig : ccvsdlc ] shows a plot of the correlation coefficient @xmath185 of eq .
( [ eq:35 ] ) using the best values of the constants to be discovered for each data point versus the distance @xmath186 ( mpc ) limitation of @xmath69 and @xmath70 .
peaks of @xmath187 were obtained at @xmath188 mpc and @xmath189 mpc .
therefore , the @xmath69 and @xmath70 were limited to 15 mpc . of the sample galaxies , 3,480 source galaxies and
1,604 sink galaxies were within 15 mpc of at least one of the category a galaxies .
note the distances to the close , major clusters ( virgo and fornax ) vary between 15 mpc and 20 mpc .
the distance of the next farther clusters ( pisces , perseus , and coma ) vary between 40 mpc and 60 mpc . of eq .
( [ eq:35 ] ) using the best values of the constants to be discovered versus the distance @xmath186 ( mpc ) limitation of @xmath190 and @xmath191 .
, scaledwidth=80.0% ] figure [ fig:7 ] shows a plot of @xmath50 versus @xmath166 for the 32 category a galaxies .
tables [ tab:1 ] and [ tab:2 ] lists the data for the 32 category a galaxies .
table [ tab:4 ] lists the calculated constants of eq .
( [ eq:26 ] ) .
the @xmath192 and @xmath193 at 1@xmath194 with a correlation coefficient of 0.88 . versus the measured redshift @xmath32 for 32 category a galaxies @xcite .
the straight line is a plot of @xmath195 .
the circles indicate the data points for galaxies with ( l , b ) = ( 290@xmath34,75@xmath35).,scaledwidth=80.0% ] the error bars indicate the @xmath50 for 1.05@xmath31 and 0.95@xmath31 , which is consistent with the error cited in @xcite .
for some galaxies such as both of the recalculations yield a higher @xmath50 calculation than the calculation using @xmath31 .
[ cols= " < , < , > , > , > , > , > , > , > , > , > " , ] @xmath196rounded to the nearest degree .
@xmath197the data point at @xmath198 is 6.4x .
@xmath199note the source galaxies closer to earth . @xmath200note
the presence of a number of sinks at @xmath201 causing the lapse in the discrete ratio . @xmath202note the presence of additional sinks at 1x and 2x and the odd source at approximately 3.5x .
@xmath203note the presence of additional sinks at 2x causing the nearby sources to have lower @xmath0
. figures [ fig:2 ] and [ fig:4 ] show @xmath155 versus @xmath156 of the galaxies within approximately six arcdegrees surrounding the identified sink .
the angular location of the identified sink is marked by the crosshairs .
the filled circles denote the galaxies within one arcdegree of the identified sink of figs .
[ fig:1z ] and [ fig:3 ] , respectively .
source galaxies surround the sink galaxies .
figure [ fig:1z ] shows the @xmath204 effect around the sink .
the @xmath0 value of galaxies closer than the identified sink is increased due to @xmath205 .
the @xmath0 value of galaxies farther than the identified sink is decreased due to @xmath206 .
the overall effect is the range of @xmath0 values of galaxies around the identified sink are tightened toward the @xmath0 value of the identified sink .
figure [ fig:3 ] shows the additional effect of @xmath207 in the @xmath0 values of more distant target galaxies that also have nearby sinks and whose light must pass through another , closer cluster .
the nearer cluster causes an incremental increase in @xmath0 and the farther sink in the cluster of the target galaxy causes the distribution of @xmath0 values to tighten .
therefore , the total effect is of the two clusters .
if the @xmath207 effect of clusters are approximately equal , then the @xmath0 value of the farther group is @xmath208 , where @xmath209 is an integer count of the clusters the light passes through .
the net effect is a near scarcity of @xmath0 values half way between integer @xmath0 variation of the @xmath0 of the first cluster from earth .
figure [ fig:5 ] of the constellation shows the situation may become very complex .
the previous plots used situations with few sink galaxies for clarity . versus the angle @xmath210 ( arcdegrees ) subtended from ( e in , @xmath211 mag . )
( @xmath212 ) = ( 150@xmath43,-13.34@xmath43,7.4727@xmath213 ) .
the open diamonds indicate the data points for source galaxies .
the filled diamonds indicate the data points for sink galaxies .
the right plot is the distance @xmath2 ( mpc ) from earth versus @xmath210 .
the open squares indicate the data points for galaxies with the @xmath2 calculated herein .
the filled squares indicate the data points for galaxies with the @xmath2 calculated using the tully - fisher relationship.,scaledwidth=80.0% ] the calculation of @xmath0 from eq .
( [ eq:26 ] ) presents a formidable task . also , as distance increases , a bias develops resulting from the omission of undetected galaxies .
the effects of @xmath214 , @xmath207 , @xmath215 , and @xmath204 relative to @xmath216 , @xmath217 , and @xmath218 suggest the @xmath0 may be calculated by adding the effects of the local group to the @xmath0 at @xmath219 mpc , fitting a linear @xmath220 relation to the @xmath221 curve for 3 mpc@xmath222 mpc , applying the linear relation to the @xmath221 curve for @xmath223 mpc , and modifying the @xmath51 calculation by counting the number of clusters the light passes near or through .
this method may be more appropriate than the hubble law for situations such as gravitational lensing in which the homogeneous assumption is invalid .
the evidence is increasing that the discrete or periodic nature of redshift is genuine .
@xcite found the observations of @xmath6 elliptical galaxies @xcite are in full concordance with the modified newtonian dynamics model ( mond ) .
however , @xcite using velocities of satellites and @xcite using lensing and stellar dynamics of bright elliptical galaxies found results that contradict mond . if @xmath99 and @xmath100 of only the target galaxy act on a test particle in a galaxy , the @xmath99 appears as a modification to @xmath100 .
the spm allows the @xmath99 of neighboring galaxies to also influence test particles .
the addition of @xmath99 from neighboring galaxies , which is not radial in general , in a dense environment is unaccounted in mond . therefore , because of the differing environmental densities , the spm expects mond to apply to @xmath6 elliptical galaxies and to fail for bright elliptical galaxies .
this suggests mond may be an application of the spm for low to medium galaxy densities and the modified acceleration of mond is another force ( @xmath99 ) .
the calculation of @xmath51 involved only the luminosity and position of galaxies .
the spm posits @xmath51 and the variation in @xmath51 is the major influence on photons redshift .
the model in this paper considered gravitational forces or potentials as being linearly related to the @xmath59 or @xmath64 of the galaxies .
therefore , as done herein , the spm suggests the strength ( luminosity ) and position of galaxies are necessary and sufficient to determine intergalactic effects .
the @xmath99 derives from @xmath224 .
therefore , the intergalactic @xmath99 may be directed at any angle relative to our line of sight .
if the universe is finite and flat , then there is an edge where @xmath225 and where @xmath226 is directed away from us .
therefore , the objection to a newtonian flat , finite universe is overcome . at larger distance ( say @xmath227 mpc ) the @xmath228 and @xmath229 terms approach an average per unit distance because the total number of encounters with sources and sinks per unit distance becomes nearly constant .
therefore , if the volumes considered are greater than the spacing between several clusters , the mass distribution of the universe appears homogeneous and the hubble law is recovered for cosmological scale distance .
however , the error ( peculiar velocity ) in the hubble law s proportionality calculation is much larger than currently accepted .
indeed , the constant of proportionality may vary with direction . examining
the @xmath229 and @xmath228 data of the more distant sample galaxies herein suggests the error is at least @xmath230 , which is consistent with @xcite .
the purpose of the present investigation was to define the structure of a scalar potential field to be consistent with various galaxy observations .
the special focus was on observations that are inconsistent with other currently popular cosmological models .
this paper has not addressed the physical mechanism of sources ( @xmath59 ) and sinks ( @xmath64 ) , the nature of @xmath51 , or the nature of the @xmath56 property of matter except to exclude the mass property . however , this paper has found a physical effect of these quantities .
a speculative example of a possible physical mechanism is galactic wind @xcite .
the shells of outward flowing , shocked gas around the galactic center @xcite@xcite and the presents of heavy - element ions in the intracluster medium suggests the emission of the ubiquitous , corporeal wind media is from the center of spiral galaxies .
if all stars emit matter in addition to radiation , newton s spherical property and , hence , the inverse square law may be applied to the effect of the wind .
the gradient of the energy of the wind media ( @xmath51 ) exerts a repulsive force ( @xmath99 ) on the cross - sectional area of particles ( @xmath56 ) .
the rate of emission by sources and of absorption by sinks may be identified with @xmath59 and @xmath64 , respectively .
the @xmath231 ratio of eq .
( [ eq:49 ] ) produce differing @xmath99 for differing particle density .
such a model is consistent with the radial distribution of black holes , quark stars , neutron stars , high metallicity stars , and lighter hydrogen stars . in the outer bulge h@xmath232 rcs of varying metallicity ,
hot stars have peculiar shapes and differ from h rcs of hydrogen gas .
however , in the outer disk h@xmath232 rcs of hot , hydrogen stars approach the h rcs ( see for example @xcite for ) . because the source of the @xmath59 is in the center of the galaxy ,
the @xmath51 variation across our solar system is insignificant . for @xmath233
au , @xmath234 .
however , a gravitational experiment must carefully consider the differing @xmath56 effects of particles .
the development of @xmath214 in section [ sec : xfactor ] included a central source and @xmath235 .
these conditions yielded the hubble law and @xmath27 . in the part of a spiral galaxy where neighboring galaxies have an insignificant effect
, the @xmath51 has the same , inverse square law form .
therefore , the spm suggests the appearance of @xmath27 in galaxy observations is sufficient to indicate that an action of @xmath99 is present and is insufficient to indicate that cosmological scale , general relativity effects are present .
the latter is consistent with the finding of @xcite that the expansion of the universe is essentially negligible for scales up to galactic clusters . for a sample of 32 galaxies with distances calculated using cepheid variable stars ,
the redshift @xmath50 calculated using the spm @xmath51 variation had a 0.88 correlation coefficient to the measured redshift @xmath32 .
the calculation argued spiral galaxies are sources and elliptical and other galaxies are sinks . because the variation of @xmath51 appears to influence @xmath32 , the variation of @xmath51
was qualitatively related to the observed , discrete variation in @xmath32 .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
this research has made use of the leda database ( http://leda.univ-lyon1.fr ) .
i acknowledge the helpful comments by the reviewer .
i acknowledge and appreciate the financial support of maynard clark , apollo beach , florida , while i was working on this project .
category a galaxies consisted of 32 galaxies .
the distances @xmath31 to category a galaxies were calculated using cepheid variable stars as listed by @xcite and @xcite .
category b galaxies consisted of 5967 spiral galaxies in the sample that were not category a galaxies with @xmath167 , @xmath168 , and @xmath236 values listed in the databases .
the distance @xmath237 ( mpc ) for each of the category b galaxies were calculated following the method of tully - fisher @xcite as follows : ( 1 ) for the category a galaxies , the absolute magnitude @xmath238 was calculated , @xmath239 ( 2 ) a plot of @xmath238/mag . versus @xmath240 , where @xmath241 is the inclination corrected @xmath167 , is presented in fig . [ fig:2a ] .
ic 1613 ( an irregular galaxy classified as a sink ) , ngc 5253 ( an irregular galaxy classified as a sink ) , and ngc 5457 ( @xcite noted the @xmath31 was calculated differently ) were omitted from the plot . the straight line in fig .
[ fig:2a ] is a plot of @xmath242 where @xmath243 , and @xmath244 at one standard deviation ( 1@xmath194 ) .
the correlation coefficient is -0.90 .
@xcite calculated a similar relation with @xmath245 and @xmath246 .
the circles indicate the data points for galaxies with ( l , b ) = ( 290@xmath34,75@xmath35 ) as in fig .
[ fig:1 ] . however , unlike fig .
[ fig:1 ] , the data in fig .
[ fig:2a ] for the outlier galaxies appears consistent with the other sample galaxies data . versus the inclination corrected 21 cm .
line width @xmath241 at 20% of peak value for 29 of the 32 category a galaxies @xcite .
the straight line is a plot of eq .
( [ eq:31 ] ) .
the circles indicate the data points for galaxies with ( l , b ) = ( 290@xmath34,75@xmath35).,scaledwidth=80.0% ] category c galaxies consisted of galaxies in the sample that were not in the previous categories with @xmath249 . a plot of @xmath31 versus @xmath182 for the category a galaxies in the category c range is presented in fig .
[ fig:3a ] . because the goal of this section is to arrive at the initial distance estimation , ngc 2541 and ngc 4548 are outlier galaxies and
were omitted from the plot . the straight line in fig
. [ fig:3a ] is a plot of @xmath250 where @xmath251 km s@xmath9 mpc@xmath9 , and @xmath252 mpc at 1@xmath194 .
the correlation coefficient is -0.93 . category e galaxies are all other galaxies in the sample not in the previous categories .
the distance @xmath257 ( mpc ) to these galaxies was calculated using the hubble law with @xmath258 km s@xmath9 mpc@xmath9 .
m. aaronson et al . , _ apj _ * 258 * ( 1982 ) , p. 64 .
j. e. m. aguilar et al . , preprint ( http://www.arxiv.org/abs/gr-qc?0411074 ) , ( 2004 ) .
e. battaner and e. florido , preprint http://www.arxiv.org/abs/astro-ph?0010475v1 ( 2000 ) .
m. b. bell , _ apj _ * 616 * ( 2004 ) , p. 738 .
m. b. bell , s. p. comeau and d. g. russel , preprint http://www.arxiv.org/abs/astro-ph?0312132 ( 2003 ) .
m. b. bell , s. p. comeau and d. g. russel , preprint http://www.arxiv.org/abs/astro-ph?0407591 ( 2004 ) .
j. binney and m. merrifield , galactic astronomy , princeton , nj : princeton university press ( and references therein ) ( 1998 ) .
g. burbidge , _ apj _ * 154 * ( 1968 ) , p. l41 .
f. i. cooperstock , v. faraoni and d. n. vollick , _ apj _ * 503 * ( 1998 ) , p. 61 . w. j. g. de block , s. s. mcgaugh and v. c. rubin , _ aj _ * 122 * ( 2001 ) , p. 2396 .
d. duari , p. d. gupta and j. v. narlikar , _ apj _ * 384 * ( 1992 ) , p. 35 .
s. fay , _ a&a _ * 413 * ( 2004 ) , p. 799 .
l. ferrarese , _ apj _ * 578 * ( 2002 ) , p. 90 .
l. ferrarese and d. merritt , preprint ( http://www.arxiv.org/abs/astro-ph?0206222 ) ( 2002 ) .
w. l. freedman et al .
, _ apj _ * 533 * ( 2001 ) , p. 47 .
a. m. ghez et al . , _ nature _ * 407 * ( 2000 ) , p. 349 .
a. w. graham et al . , _ apj _ * 563 * ( 2001 ) , p. l11 .
a. w. graham , preprint http://www.arxiv.org/abs/astro-ph?0206248 ( 2002 ) .
a. w. graham et al .
, preprint http://www.arxiv.org/abs/astro-ph?0306023 ( 2003 ) .
j. guzik and u. seljak , _ mnras _ * 335 * ( 2002 ) , p. 311 . m. hudson et al . , _ mnras _ * 352 * ( 2004 ) , p. 61 . f. p.
israel , _ aarv _ * 8 * ( 1998 ) , p. 237 .
k. g. karlsson , _
a&a _ * 58 * ( 1977 ) , p. 237 .
a. knigl , preprint http://www.arxiv.org/abs/astro-ph?0302109 ( 2003 ) .
b. lilje , a. yahil and b. j. t. jones , _ apj _ * 307 * ( 1986 ) , p. 91 .
d. lynden - bell et al . , _ apj _ * 326 * ( 1988 ) , p. 19 .
r. j. mclure and j. s. dunlop , _ mnras _ * 331 * ( 2002 ) , p. 795 .
l. m. macri et al . , _ apj _ * 559 * ( 2001 ) , p. 243 .
j. c. mather et al .
, _ apj _ * 354 * ( 1990 ) , p. l37 . j. c. mather et al . , _ apj _ * 512 * ( 1999 ) , p. 511 . j. p.
mbelek , _ a&a _ ,
* 424 * ( 2004 ) , p. 761 .
d. merritt and l. ferrarese , _ mnras _ * 320 * ( 2001a ) , p. l30 .
d. merritt and l. ferrarese , the central kpc of starbursts and agns , eds .
j. h. knapen , j. e. beckman , i. shlosman and t. j. mahoney http://www.arxiv.org/abs/astro-ph?0107134 ( 2001b ) . m. milgrom and r. h. sanders , _ apj _ * 599 * ( 2003 ) , p. l25 . t. paumard , j. p. maullard , m. morris and f. rigaut , _ a&a _ * 366 * ( 2001 ) , p. 466 . j. a. peacock et al . ,
_ nature ( london ) _ * 410 * ( 2001 ) , p. 169 .
y. f. pirogov , preprint ( http://www.arxiv.org/abs/gr-qc?0505031 ) ( 2005 ) .
f. prada et al .
, _ apj _ * 598 * ( 2003 ) , p. 260 . m. rejkuba ,
_ a&a _ * 413 * ( 2004 ) , p. 903 .
m. a. rodriguez - meza et al .
, preprint http://www.arxiv.org/abs/gr-qc?0504111 ( 2005 ) .
a. j. romanowsky et al .
, _ science _ * 301 * ( 2003 ) , p. 1696 . d. g. russell , preprint http://www.arxiv.org/abs/astro-ph?0503440 ( 2005 ) . j. a. sellwood and a. kosowsky , _ aspc _ * 240 * ( 2001 ) , p. 311 .
m. j. sempere et al . , _
* 296 * ( 1995 ) , p. 45 .
c. shu , h. j. mo , and s. mao , _ chjaa _ * 5 * ( 2005 ) , p. 327 .
j. silk and m. j. rees , _
a&a _ * 331 * ( 1998 ) , p. l1 .
y. sofue et al . , _ apj _ * 523 * ( 1999 ) , p. 136 .
t. takamiya and y. sofue , _ apj _ * 534 * ( 2000 ) , p. 670 . w. g. tifft , _ apj _ * 468 * ( 1996 ) , p. 491 . w. g. tifft , _ apj _ * 485 * ( 1997 ) , p. 465 . t. treu and l. v. e. koopmans , _ apj _ * 611 * ( 2004 ) , p. 739 .
r. b. tully and j. r. fisher , _
a&a _ * 54 * ( 1977 ) , p. 661 .
s. veilleux , g. cecil and j. bland - hawthorn , preprint http://www.arxiv.org/abs/astro-ph?0504435 ( 2005 ) .
a. wandel , preprint http://www.arxiv.org/abs/astro-ph?0303562 ( 2003 ) .
c. will , theory and experiment in gravitational physics , cambridge , united kingdom : cambridge university press ( 1993 ) . | on the galactic scale the universe is inhomogeneous and redshift @xmath0 is occasionally less than zero .
a scalar potential model ( spm ) that links the galaxy scale @xmath0 to the cosmological scale @xmath0 of the hubble law is postulated .
several differences among galaxy types suggest that spiral galaxies are sources and that early type , lenticular , and irregular galaxies are sinks of a scalar potential field . the morphology - radius and
the intragalactic medium cluster observations support the movement of matter from source galaxies to sink galaxies . a cell structure of galaxy groups and clusters is proposed to resolve a paradox concerning the scalar potential like the olber s paradox concerning light . for the sample galaxies , the ratio of the luminosity of source galaxies to the luminosity of sink galaxies approaches @xmath1 .
an equation is derived from sample data , which is anisotropic and inhomogeneous , relating @xmath0 of and the distance @xmath2 to galaxies .
the calculated @xmath0 has a correlation coefficient of 0.88 with the measured @xmath0 for a sample of 32 spiral galaxies with @xmath2 calculated using cepheid variable stars .
the equation is consistent with @xmath3 observations of close galaxies . at low cosmological distances
, the equation reduces to @xmath4 , where @xmath5 is a constant , positive value .
the equation predicts @xmath0 from galaxies over 18 gpc distant approaches a constant value on the order of 500 .
the spm of @xmath0 provides a physical basis for the @xmath0 of particle photons .
further , the spm qualitatively suggests the discrete variations in @xmath0 , which was reported by w. g. tifft , 1997 , astrophy .
j. 485 , 465 and confirmed by others , are consistent with the spm .
keyword galaxies : distances and redshifts , galaxies : elliptical and lenticular , cd , galaxies : spiral , cosmology : theory 98.62.py , 98.52.eh , 98.52.nr , 98.52.lp |
both scalar fields and perfect fluids are pervasive in our present models for the evolution of the universe since its birth until the current epoch of accelerated expansion . both ( early time ) inflation and the present dark energy epoch
are commonly supposed to be supported by scalar fields .
the intermediate epochs in the evolution of the universe usually consist of a radiation - dominated era followed by a matter - dominated era .
these can be well described by barotropic perfect fluids with equations of state @xmath0 equal to one third and zero , respectively .
a barotropic perfect fluid can be defined as a perfect fluid where the pressure is a function of the energy density only ( the same function for the background and the perturbations ) .
because of the importance of scalar fields and perfect fluids in present - day cosmological models , the theory of cosmological perturbations in these models has been intensively studied and is well developed . for instance , cosmological perturbations for perfect fluids at linear order have been studied in @xcite , at second order in @xcite and more recently at third order @xcite .
some fully nonlinear results have also been obtained ; see for example @xcite .
similarly , linear cosmological perturbations in canonical scalar field systems have been studied in @xcite , and in @xcite at second and third order in perturbations .
more recently , cosmological perturbations in noncanonical scalar field models have attracted much attention .
the reason for such interest is that it was realized that inflationary models in string theory @xcite naturally contain scalar fields with nonstandard kinetic terms and these models can also give rise to large non - gaussianity of the primordial curvature perturbation .
@xcite studied linear perturbations . for second - order perturbations ,
see for instance @xcite and for third - order , see @xcite .
perturbations up to third order of inflationary models with multiple noncanonical scalar fields have also been studied ; see for instance @xcite .
because of their importance in cosmology , it is natural to ask when , if in any case , can we describe a scalar field by a perfect fluid and vice versa . if there are some models where this equivalence is realized one may be able to use the known results of perturbation theory mentioned above to study models of perfect fluids or scalar fields where the calculation in one side of the duality has not been done .
a recent explicit example of such a case can be found in @xcite .
when both results of perturbation theory for scalar fields and perfect fluids exist , in the dual models , one can use the equivalence as an extra consistency check for the calculations . in this brief report , we will discuss the equivalence of a barotropic perfect fluid with a so - called k - essence / k - inflation scalar field @xcite . this paper is organized as follows . in the next section , we will introduce a general class of k - essence / k - inflation models under study . we will also present the background equations of motion . in section [ sec : main ] , we will briefly discuss linear perturbations ; we shall _ obtain _ and _ solve _ the second - order partial differential equation that gives us the class of scalar field models that is dual to a barotropic perfect fluid , under the assumption of an irrotational fluid flow .
finally , section [ sec : conclusion ] is devoted to conclusion .
we consider a fairly general class of single scalar field models described by the action @xmath1 , \label{action}\ ] ] where @xmath2 is the scalar field , @xmath3 is the planck mass that will be set to unity hereafter , @xmath4 is the ricci scalar , and @xmath5 is the scalar field kinetic term .
@xmath6 is the metric tensor .
we label the scalar field lagrangian by @xmath7 and we assume that it is a well behaved function of two variables , the scalar field @xmath2 and its kinetic term @xmath8 .
this general field lagrangian includes as particular cases the common canonical scalar field model , dirac - born - infeld inflation @xcite and k - inflation / k - essence @xcite .
the energy - momentum tensor is defined as @xmath9 which gives @xmath10 if @xmath11 is timelike ( i.e. @xmath12 ) , then this is of the perfect fluid form @xmath13 with the pressure @xmath7 and the energy density @xmath14 where @xmath15 denotes the derivative of @xmath7 with respect to @xmath8 .
the four - velocity is subject to the constraint @xmath16 and reads @xmath17 the scalar field equation of motion is @xmath18 where @xmath19 denotes a covariant derivative .
we are interested in flat , homogeneous , and isotropic friedmann - lematre - robertson - walker background universes described by the line element @xmath20 where @xmath21 is the scale factor .
the friedmann equation and the continuity equation read @xmath22 where the hubble rate is @xmath23 , and @xmath24 is the background energy density of the scalar field given by @xmath25 where @xmath26 denotes the derivative of @xmath7 with respect to @xmath8 evaluated at the background value @xmath27 .
in this section , we will first discuss linear perturbations and the condition for the equivalence between a scalar field model and a barotropic fluid .
then we will obtain and solve the second - order partial differential equation that determines the class of scalar field models dual to an irrotational barotropic perfect fluid . at the end of the section
, we will briefly discuss the behavior of the nonadiabatic pressure perturbation . on the comoving time - slices ,
the scalar field fluctuations vanish , @xmath28 , and the three - dimensional spatial metric @xmath29 is perturbed as @xmath30\,,\end{aligned}\ ] ] where tensor perturbations have been neglected and @xmath31 denotes the curvature perturbation on comoving slices . @xmath32 and @xmath33 can be put to zero by an appropriate choice of spatial coordinates , where @xmath33 denotes an intrinsic vector perturbation .
the linear equation of motion for @xmath31 is @xmath34 where @xmath35 and @xmath36 is the speed of propagation of scalar perturbations ( speed of sound " ) given by @xcite @xmath37 in the case of a barotropic perfect fluid , scalar perturbations propagate with speed @xmath38 given by @xmath39 which is often called the adiabatic sound speed . the comoving slicing condition for a fluid is @xmath40 , and the curvature perturbation on comoving slices @xmath31 follows eq .
( [ calrceq ] ) with @xmath41 replaced by @xmath42 . in general these two speeds
are different @xcite .
but one can consider under which conditions they are the same . requiring @xmath43 , and after using @xmath44 and the background equations of motion , one finds the following equation : @xmath45 this equation can be integrated once to give @xmath46 where @xmath47 is an arbitrary function of @xmath2 .
using the method of characteristics , one can further integrate eq .
( [ intonce ] ) to find ) and the solution ( [ solution ] ) for a rather different problem ; the necessary condition for the existence of exact stationary field configurations .
the nontrivial fact that the condition for the existence of stationary configurations perfectly coincides with the requirement of exact equivalence between a scalar theory and a barotropic perfect fluid may deserve further study . ] @xmath48 where @xmath49 and @xmath50 are arbitrary functions .
for this lagrangian , the adiabatic sound speed , which is equal to the speed of sound , is given by @xmath51 where we define @xmath52 as @xmath53 .
this is the most general class of scalar field models @xmath54 that is exactly equivalent to a barotropic perfect fluid , under the assumption of the velocity being described by a single scalar potential , i.e. the fluid flow is irrotational .
figure [ diagram ] is a graphical description of this result .
the fact that any solution to the einstein equations in the presence of a barotropic perfect fluid can be interpreted as a solution in the presence of a purely kinetic k - essence scalar field is known in the literature @xcite ( see also for example @xcite ) . here
we show that all the other k - essence models that depend on the values of the scalar field and satisfy eq .
( [ mastereq ] ) are also equivalent to a barotropic perfect fluid .
the lagrangian of these other models has to be of the form ( [ solution ] ) ) and the purely kinetic k - essence models possess the same equation of state parameter and speed of sound and thus are also equivalent to barotropic perfect fluids .
here we show that these are the only scalar field models that are equivalent because only these models satisfy the necessary and sufficient condition ( [ mastereq ] ) . ] . a standard canonic scalar field , i.e. @xmath55 , where @xmath56 is the potential , is not included in this class of models .
the speed of sound is @xmath57 and the adiabatic sound speed is @xmath58 all purely kinetic k - essence models , i.e. the field s lagrangian is of the form @xmath59 , are included in this class of models .
these fields have been used to drive inflation @xcite and to be the dark energy @xcite . if @xmath60 , where @xmath53 , then eq . ( [ energy ] ) can be written as @xmath61 and one has that the energy density @xmath62 is a function of @xmath52 only .
so if @xmath63 is invertible one can in principle find the equation of state @xmath64 . assuming @xmath65 so that @xmath66 , the previous equation can also be put in the form @xmath67 where @xmath68 is an integration constant . for a given invertible equation of state one
has @xmath69 , and if @xmath60 , eq .
( [ energy ] ) can be written as @xmath70 and integrated to give @xmath71 from where one finds the lagrangian given an equation of state @xmath72 .
it may be worth mentioning the expression for the conserved ( `` baryon '' ) number density @xmath73 . for @xmath74 and @xmath75 , it is determined by @xmath76 this may be readily integrated to give @xmath77 where @xmath78 is a constant of integration , which would be a function of entropy for a perfect fluid . for the scalar field lagrangian ( [ solution ] ) , because both the pressure and the energy density are functions of one variable only , i.e. @xmath52 , it can be shown that the nonadiabatic pressure perturbation defined for instance in @xcite vanishes exactly to all orders in perturbations and on all scales .
this is not surprising because all models of the form ( [ solution ] ) are dual to barotropic ( i.e. adiabatic ) perfect fluids where the nonadiabatic pressure perturbation vanishes by definition . in other words ,
the equivalence between the dual models is independent of the background because the derived second - order differential equation ( [ mastereq ] ) is independent of the background but dependent only on the form of @xmath7 as a function of @xmath8 and @xmath2 , despite the fact that it was derived from the condition for the equivalence in linear perturbation theory .
in this brief report , we have studied under which conditions a general k - essence scalar field is equivalent to an irrotational barotropic perfect fluid .
we have found that the condition can be written as a second - order partial differential equation for the lagrangian of the field , eq .
( [ mastereq ] ) , that simply states that the sound speed @xmath36 at which scalar perturbations propagate has to be equal to the adiabatic sound speed @xmath38 .
we have found the most general solution for that equation , eq .
( [ solution ] ) .
the lagrangian of the solutions we found depends explicitly on both the kinetic term @xmath8 and the scalar field value @xmath2 .
however , this dependence is of the restrictive form @xmath79 , where @xmath49 and @xmath50 are arbitrary functions . after a suitable field redefinition , @xmath80 where @xmath81
, the lagrangian can be cast in the form of a purely kinetic k - essence model .
we note , however , that in reality there may exist plural fields or fluids that interact with each other , which may lead to small violations of the perfect fluid / adiabaticity condition . in such a case
, the field redefinition may or may not be useful depending on the form of interactions as well as on the situation of interest . one can show that , for the lagrangian ( [ solution ] ) , the nonadiabatic pressure perturbation vanishes exactly to all others in perturbations and on all scales .
this is because both the pressure and the energy density are functions of one variable only .
this result is less surprising if we note that these scalar field models are equivalent to barotropic perfect fluids where the nonadiabatic pressure perturbation vanishes by definition .
we are grateful to david wands for useful discussions and for drawing our attention to reference @xcite .
fa is supported by the japanese society for the promotion of science ( jsps ) .
ms is supported in part by jsps grant - in - aid for scientific research ( a ) no .
21244033 , and by jsps grant - in - aid for creative scientific research no .
this work was also supported in part by mext grant - in - aid for the global coe program at kyoto university , `` the next generation of physics , spun from universality and emergence . ''
j. m. bardeen , phys . rev .
* d22 * , 1882 ( 1980 ) .
h. kodama and m. sasaki , prog .
. suppl . * 78 * , 1 ( 1984 ) .
v. f. mukhanov , h. a. feldman , and r. h. brandenberger , phys . rept . * 215 * , 203 ( 1992 ) .
n. bartolo , e. komatsu , s. matarrese , and a. riotto , phys
. rept . * 402 * , 103 ( 2004 ) , astro - ph/0406398 .
k. a. malik and d. wands , phys . rept . *
475 * , 1 ( 2009 ) , 0809.4944 .
n. bartolo , s. matarrese , and a. riotto , ( 2010 ) , 1001.3957 .
g. damico , n. bartolo , s. matarrese , and a. riotto , jcap * 0801 * , 005 ( 2008 ) , 0707.2894 .
a. j. christopherson and k. a. malik , jcap * 0911 * , 012 ( 2009 ) , 0909.0942 .
d. h. lyth , k. a. malik , and m. sasaki , jcap * 0505 * , 004 ( 2005 ) , astro - ph/0411220 .
d. langlois and f. vernizzi , phys .
lett . * 95 * , 091303 ( 2005 ) , astro - ph/0503416 . v. f. mukhanov and g. v. chibisov , jetp lett . * 33 * , 532 ( 1981 ) .
m. sasaki , prog .
. phys . * 70 * , 394 ( 1983 ) .
m. sasaki , prog .
. phys . * 76 * , 1036 ( 1986 ) .
j. m. maldacena , jhep * 05 * , 013 ( 2003 ) , astro - ph/0210603 . k. a. malik , jcap * 0703 * , 004 ( 2007 ) , astro - ph/0610864 .
d. seery , j. e. lidsey , and m. s. sloth , jcap * 0701 * , 027 ( 2007 ) , astro - ph/0610210 .
d. seery , m. s. sloth , and f. vernizzi , jcap * 0903 * , 018 ( 2009 ) , 0811.3934 .
e. silverstein and d. tong , phys
. rev . * d70 * , 103505 ( 2004 ) , hep - th/0310221 .
x. chen , phys . rev . *
d71 * , 063506 ( 2005 ) , hep - th/0408084 .
x. chen , jhep * 08 * , 045 ( 2005 ) , hep - th/0501184 .
j. garriga and v. f. mukhanov , phys . lett . *
b458 * , 219 ( 1999 ) , hep - th/9904176 .
d. seery and j. e. lidsey , jcap * 0506 * , 003 ( 2005 ) , astro - ph/0503692 .
x. chen , m .- x .
huang , s. kachru , and g. shiu , jcap * 0701 * , 002 ( 2007 ) , hep - th/0605045 .
x. chen , m .- x .
huang , and g. shiu , phys . rev .
* d74 * , 121301 ( 2006 ) , hep - th/0610235 . f. arroja and k. koyama , phys . rev . * d77 * , 083517 ( 2008 ) , 0802.1167 .
x. chen , b. hu , m .- x .
huang , g. shiu , and y. wang , jcap * 0908 * , 008 ( 2009 ) , 0905.3494 .
f. arroja , s. mizuno , k. koyama , and t. tanaka , phys . rev . *
d80 * , 043527 ( 2009 ) , 0905.3641 .
d. seery and j. e. lidsey , jcap * 0509 * , 011 ( 2005 ) , astro - ph/0506056 .
d. langlois and s. renaux - petel , jcap * 0804 * , 017 ( 2008 ) , 0801.1085 .
d. langlois , s. renaux - petel , d. a. steer , and t. tanaka , phys .
* 101 * , 061301 ( 2008 ) , 0804.3139 .
d. langlois , s. renaux - petel , d. a. steer , and t. tanaka , phys . rev . *
d78 * , 063523 ( 2008 ) , 0806.0336 . f. arroja , s. mizuno , and k. koyama , jcap * 0808 * , 015 ( 2008 ) , 0806.0619 .
s. renaux - petel and g. tasinato , jcap * 0901 * , 012 ( 2009 ) , 0810.2405 . s. mizuno , f. arroja , k. koyama , and t. tanaka , phys .
rev . * d80 * , 023530 ( 2009 ) , 0905.4557 . s. mizuno , f. arroja , and k. koyama , phys . rev . *
d80 * , 083517 ( 2009 ) , 0907.2439 .
s. renaux - petel , jcap * 0910 * , 012 ( 2009 ) , 0907.2476
. l. boubekeur , p. creminelli , j. norena , and f. vernizzi , jcap * 0808 * , 028 ( 2008 ) , 0806.1016 . c. armendariz - picon , t. damour , and v. f. mukhanov , phys .
. * b458 * , 209 ( 1999 ) , hep - th/9904075 . m. alishahiha , e. silverstein , and d. tong , phys . rev . *
d70 * , 123505 ( 2004 ) , hep - th/0404084 .
a. j. christopherson and k. a. malik , phys .
b675 * , 159 ( 2009 ) , 0809.3518 .
s. unnikrishnan and l. sriramkumar , ( 2010 ) , 1002.0820 .
r. akhoury , c. s. gauthier , and a. vikman , jhep * 03 * , 082 ( 2009 ) , 0811.1620 .
a. diez - tejedor and a. feinstein , int .
j. mod . phys .
* d14 * , 1561 ( 2005 ) , gr - qc/0501101 .
s. matarrese , proc .
lond . * a401 * , 53 ( 1985 ) . c. quercellini , m. bruni , and a. balbi , class .
quant . grav .
* 24 * , 5413 ( 2007 ) , 0706.3667 .
d. bertacca , n. bartolo , a. diaferio , and s. matarrese , jcap * 0810 * , 023 ( 2008 ) , 0807.1020 .
e. j. copeland , m. sami , and s. tsujikawa , int .
d15 * , 1753 ( 2006 ) , hep - th/0603057 .
d. wands , k. a. malik , d. h. lyth , and a. r. liddle , phys .
* d62 * , 043527 ( 2000 ) , astro - ph/0003278 .
k. a. malik and d. wands , class .
* 21 * , l65 ( 2004 ) , astro - ph/0307055 . | in this brief report , we obtain the necessary and sufficient condition for a class of noncanonical single scalar field models to be exactly equivalent to barotropic perfect fluids , under the assumption of an irrotational fluid flow
. an immediate consequence of this result is that the nonadiabatic pressure perturbation in this class of scalar field systems vanishes exactly at all orders in perturbation theory and on all scales .
the lagrangian for this general class of scalar field models depends on both the kinetic term and the value of the field .
however , after a field redefinition , it can be effectively cast in the form of a purely kinetic k - essence model . |
the extent to which gluons are polarized in the nucleon , and consequently the origin of the nucleon spin , has persisted as an elusive question for almost two decades in spite of strenuous experimental efforts and theoretical activity @xcite .
although the spin dependent gluon density in principle can be sized in inclusive deep inelastic scattering measurements , mainly through the scale dependence of the measured asymmetries , this dependence is rather mild in the kinematical range accessed by experiments , and conclusions about it are also veiled by our ignorance regarding the polarization of the other partonic species , which also contribute to the scale dependence , specially that of sea quarks .
therefore , even in the most ambitious scenario , inclusive deep inelastic scattering data can at most suggest mild constraints on the gluon polarization . in a recent article @xcite
, we have shown that the enduring efforts to measure less inclusive observables in deep inelastic scattering have finally begun to yield , allowing combined next to leading order global qcd fits to inclusive and semi - inclusive deep inelastic scattering data where sea quarks and gluons are much more definitely constrained . in the mean time , independent measurements of other less inclusive observables , such as pion production in polarized proton - proton collisions @xcite , and high transverse momentum hadron pair production in deep inelastic scattering have begun to provide more direct assessments of the gluon polarization with competing precision @xcite .
it is therefore of great interest to compare the gluon polarization estimates coming from both the global analysis of deep inelastic scattering data and from the more direct measurements . in the following
we perform such comparison and we find that although preliminary direct measurements still have a moderate impact in the fits , there is a remarkable agreement and complementarity between both approaches , what encourage us to incorporate the forthcoming data in future global analysis .
in the case of inclusive and semi - inclusive polarized deep inelastic scattering ( dis ) , the next to leading order ( nlo ) qcd framework required to compute the respective observables have been available for some time @xcite , and indeed recent global analyses have demonstrated both the relevance of these corrections and also the non negligible impact of most recent semi - inclusive data @xcite .
specifically , in reference @xcite it has been found that the best global qcd fits to combined dis data constrain the gluon polarization to be moderately positive with a first moment of this density @xmath1 of @xmath2 at @xmath3 with an uncertainty range given by @xmath4 $ ] for a one - unit increase in @xmath5 and by @xmath6 $ ] allowing a more conservative @xmath7 variation of @xmath5 . in these constraints , both the requirement of positivity of the polarized parton densities relative to a modern set of unpolarized parton densities @xcite , and the correlation between gluon and sea quark polarization , are found to be crucial . fits with a wide variation in the gluon polarization reproduce inclusive data equally well , however they are clearly differentiated because of their sea quark polarization by semi - inclusive data .
the cross section for single inclusive large @xmath0 pion production in longitudinally polarized proton - proton collisions , which is right now being measured at brookhaven national laboratory relativistiv heavy ion collider ( bnl rhic ) @xcite have also been computed at nlo accuracy , and have been found to be significantly dependent on @xmath8 @xcite .
recently , the phenix collaboration has presented preliminary results with considerably reduced errors @xcite which clearly disfavors scenarios with large gluon polarization and are in nice agreement with estimates of updated polarized fits . in figure
[ deltag]a we show the expectation for the double spin asymmetry computed with the best fit of reference @xcite , together with the data reported by phenix @xcite .
we also plot the uncertainty band associated to a @xmath9 variation .
@xcite compared with the estimate coming from the the nlo fit of reference @xcite and the uncertainty band allowing @xmath9 ; ( b ) the same but for estimates for @xmath10 from @xcite .
[ deltag],title="fig:",width=336 ] @xcite compared with the estimate coming from the the nlo fit of reference @xcite and the uncertainty band allowing @xmath9 ;
( b ) the same but for estimates for @xmath10 from @xcite .
[ deltag],title="fig:",width=336 ] the @xmath5 for phenix data obtained with the best fit of reference @xcite , which is previous to the latest set of phenix data , and computed including the 40% scaling uncertainty in the nondiagonal covariance matrix , results to be @xmath11 , for @xmath12 data points , which is well within the @xmath13 range expected for a given subset of data in global fit and therefore should be considered as consistent .
the value for @xmath5 is comparable to the one obtained ( @xmath14 ) in a recent fit @xcite to both inclusive dis data and the phenix measurement , although not including the scaling error in the computation .
similar agreement is found comparing the expectation of the fit for @xmath15 at @xmath16 gev@xmath17 against preliminary data from compass @xcite , and previous measurements @xcite , as shown in figure [ deltag]b . in this case
we include both the leading order ( lo ) and the nlo expectation because the reported values for @xmath15 correspond to a lo extraction improved with montecarlo higher order corrections .
we include in the plot the uncertainty band coming from a 2% variation in @xmath5 , plus that coming from varying @xmath18 up to @xmath19 gev@xmath17 , what again highlights the nice consistency between independent data set and the frameworks implemented for the corresponding analyses .
further insight on the interplay between dis data and that coming from phenix can be obtained analyzing the profile of @xmath5 function for the different subsets of data in a combined fit , against the range of variation of the net gluon polarization , as it was done in @xcite . in figure [ profile ] we show the profile of the total @xmath5 of a global fit to inclusive and semi - inclusive data along the lines of that of reference @xcite but also including phenix data as a solid line .
the curve has similar shape to the one found in @xcite but shifted upwards between eleven and thirteen units , which is essentially the partial contribution of @xmath20 to @xmath5 . in order to see the relevance of @xmath20 data , this last contribution
is also plotted as a dashed - dotted line with an offset 430.91 units , which the partial contribution of dis data @xmath21@xmath22 for the best fit .
notice the partial contribution of @xmath20 is almost flat around the minimum of @xmath5 and reaches its own minimum for slightly lower values of @xmath23 but within the one - unit variation of @xmath5 , what highlights the consistency between both data sets . of different sets of data against the gluon net polarization .
[ profile],width=432 ] in figure [ profile ] we have plotted also the profile obtained for dis data only in @xcite as a dashed line , with an offset given by the partial contribution of @xmath20 at its minimun @xmath21@xmath24 , in order to see the net effects on the fit of phenix data . in table [ tab : table1 ] we show the partial contributions to @xmath5 at both minima .
.[tab : table1 ] partial contributions to @xmath5 values and first moment of @xmath8 at @xmath25 gev@xmath17 [ cols="^,^,^,^,^",options="header " , ] clearly , these effects are almost imperceptible around the minimum but can be noticed for @xmath5@xmath26 variations between the one - unit and the @xmath7 increase .
close to the minimum of @xmath5 , the decrease of @xmath5@xmath24 for decreasing @xmath23 is overpowered by the increase of @xmath5@xmath22 and consequently , the position of the minimum remains that found for dis data .
given the large number of dis data included in the fit ( 478 ) compared to the rather limited set of @xmath20 available at present , the small impact in the fit is not surprising , nevertheless the consistency shown , and the possibility of increasing considerably the statistics in the future is encouraging .
we have compared the constraints on the gluon polarization in the nucleon obtained in next to leading order global qcd fits to polarized deep inelastic scattering data with those coming the double spin asymmetry measured by phenix at rhic .
although the relative statistical weight of @xmath20 data in a global nlo fit including also dis data is rather limited , we find a remarkable agreement and moderate improvement when combining both data sets . in the case of other direct measurements such as high-@xmath0 hadron production studied by compass at cern , the lack of a nlo framework for the computation of the corresponding asymmetries does not allow to include them yet in a combined nlo global analysis
however we find preliminary agreement which hopefully will be checked at nlo accuracy in the near future .
we warmly acknowledge daniel de florian for comments and suggestions and y. fukao for discussions relative to phenix data .
b. lampe and e. reya , phys . rep .
332 , 1 ( 2000 ) ; + e. leader et al . , phys , rep . 261 , 1 ( 1995 ) .
d. de florian , g. a. navarro and r. sassot , _ phys .
d _ * 71 * , 094018 ( 2005 ) .
collaboration , k. boyle , talk given at the xviith particles and nuclei international conference ( panic05 ) , http://www.panic05.lanl.gov/ e. s. ageev _ et al .
_ , phys .
b * 633 * , 25 ( 2006 ) .
s. procureur , arxiv : hep - ex/0605043 .
a. airapetian _ et al .
_ , phys .
lett . * 84 * , 2584 ( 2000 ) .
b. adeva _ et al .
_ , phys .
d * 70 * , 012002 ( 2004 ) .
r. mertig , w. l. van neerven , z.phys.*c70 * , 637,(1996 ) ; w. vogelsang , phys.rev.*d54 * , 2023 , ( 1996 ) .
d. de florian , c. a. garcia canal and r. sassot , nucl .
b * 470 * , 195 ( 1996 ) [ arxiv : hep - ph/9510262 ] .
d. de florian and r. sassot , phys .
d * 62 * , 094025 ( 2000 ) .
b. jger , a. schafer , m. stratmann and w. vogelsang , phys . rev .
d * 67 * , 054005 ( 2003 ) [ arxiv : hep - ph/0211007 ] .
d. de florian , phys .
d * 67 * , 054004 ( 2003 ) [ arxiv : hep - ph/0210442 ] .
a. d. martin et al . , _ eur .
j. c _ * 28 * 455 ( 2003 ) .
s. s. adler _ et al .
_ [ phenix collaboration ] , phys .
* 93 * ( 2004 ) 202002 [ arxiv : hep - ex/0404027 ] . m. hirai , s. kumano and n. saito , arxiv : hep - ph/0603213 . | we compare constraints on the gluon polarization in the nucleon obtained in next to leading order global qcd fits to polarized deep inelastic scattering data with those coming from observables more directly linked to the gluon polarization , such as the double spin asymmetry measured by phenix at rhic , and high-@xmath0 hadron production studied by compass at cern . |
the feasibility of processing graphs in the data stream model was one of the early questions investigated in the streaming model @xcite .
however the results were not encouraging , even to decide simple properties such as the connectivity of a graph , when the edges are streaming in an arbitrary order required @xmath3 space . in comparison to the other results in the streaming model ,
@xcite which required polylogarithmic space , graph alogithms appeared to difficult in the streaming context and did not receive much attention subsequently .
however in recent years , with the remergence of social and other interaction networks , questions of processing massive graphs have once again become prominent .
technologically , since the publication of @xcite , it had become feasible to store larger quantities of data in memory and the semi - streaming model was proposed in @xcite . in this model
we assume that the space is ( near ) linear in the number of vertices ( but not necessarily the edges ) .
since its formulation , the model has become more appealing from the contexts of theory as well as practice . from a theoretical viewpoint
, the model still offers a rich potential trade - off between space and accuracy of algorithm , albeit at a different threshold than polylogarithmic space . from a practical standpoint , in a variety of contexts involving large graphs , such as image segmentation using graph cuts ,
the ability of the algorithm to retain the most relevant information in main memory has been deemed critical .
in essence , an algorithm that runs out of main memory space would become unattractive and infeasible .
in such a setting , it may be feasible to represent the vertex set in the memory whereas the edge set may be significantly larger . in the semi - streaming model ,
the first results were provided by @xcite on the construction of graph spanners .
subsequently , beyond explorations of connectivity @xcite , and ( multipass ) matching @xcite , there has been little development of algorithms in this model . in this paper
we focus on the problem of graph sparsification in a single pass , that is , constructing a small space representation of the graph such that we can estimate the size of any cut .
graph sparsification @xcite remains one of the major building blocks for a variety of graph algorithms , such as flows and disjoint paths , etc . at the same time
, sparsification immediately provides a way of finding an approximate min - cut in a graph .
the problem of finding a min - cut in a graph has been one of the more celebrated problems and there is a vast literature on this problem , including both deterministic @xcite as well as randomized algorithms @xcite see @xcite for a comprehensive discussion of various algorithms .
we believe that a result on sparsification will enable the investigation of a richer class of problems in graphs in the semi - streaming model .
in this paper we will focus exclusively on the model that the stream is adversarially ordered and a single pass is allowed . from the standpoint of techniques , our algorithm is similar in spirit to the algorithm of alon - matias - szegedy @xcite , where we simultaneously sample and estimate from the stream .
in fact we show that in the semi - streaming model we can perform a similar , but non - trivial , simultaneous sampling and estimation .
this is pertinent because sampling algorithms for sparsification exist @xcite , which use @xmath4 edges .
however these algorithms sample edges in an iterative fashion that requires the edges to be present in memory and random access to them .
[ [ our - results ] ] our results : + + + + + + + + + + + + our approach is to recursively maintain a summary of the graph seen so far and use that summary itself to decide on the action to be taken on seeing a new edge . to this end , we modify the sparsification algorithm of benczur and karger @xcite for the semi streaming model .
the final algorithm uses a single pass over the edges and provides @xmath5 approximation for cut values with high probability and uses @xmath6 edges for @xmath7 node and @xmath8 edge graph .
let @xmath9 denote the input graph and @xmath7 and @xmath8 respectively denote the number of nodes and edges .
@xmath10 denotes the value of cut @xmath11 in @xmath9 .
@xmath12 indicates the weight of @xmath13 in graph @xmath9 .
@xcite a graph is * k - strong connected * if and only if every cut in the graph has value at least @xmath14 .
* k - strong connected component * is a maximal node - induced subgraph which is k - strong connected .
the * strong connectivity * of an edge @xmath13 is the maximum @xmath14 such that there exists a @xmath14-strong connected component that contains @xmath13 . in @xcite ,
they compute the strong connectivity of each edge and use it to decide the sampling probability .
algorithm [ alg : sparsify ] is their algorithm .
we will modify this in section [ sec : algdesc ] .
[ alg : sparsify ] * benczur - karger*(@xcite ) + compute the strong connectivity of edge @xmath15 for all @xmath16 @xmath17 here @xmath18 is a parameter that depends on the size of @xmath9 and the error bound @xmath19 .
they proved the following two theorems in their paper .
[ thm : benczur_errorbound ] @xcite given @xmath19 and a corresponding @xmath20 , every cut in @xmath21 has value between @xmath22 and @xmath1 times its value in @xmath9 with probability @xmath23 .
[ thm : benczur_spacebound ] @xcite with high probability @xmath21 has @xmath24 edges . throughout this paper ,
@xmath25 denotes the input sequence .
@xmath26 is a graph that consists of @xmath27,@xmath28,@xmath29,@xmath30 .
@xmath31 is the strong connectivity of @xmath13 in @xmath9 and @xmath12 is weight of an edge @xmath13 in @xmath9 .
each edge has weight 1 in @xmath33 .
@xmath34 where scalar multiplication of a graph and addition of a graph is defined as scalar multiplication and addition of edge weights .
in addition , @xmath35 if and only if @xmath36 . @xmath37 is a sparsification of a graph @xmath26 , i.e. , a sparsified graph after considering @xmath30 in the streaming model .
we can not use algorithm [ alg : sparsify ] in the streaming model since it is not possible to compute the strong connectivity of an edge in @xmath9 without storing all the data .
the overall idea would be to use a strongly recursive process , where we use an estimation of the connectivity based on the current sparsification and show that subsequent addition of edges does not impact the process .
the modification is not difficult to state , which makes us believe that such a modification is likely to find use in practice .
the nontrivial part of the algorithm is in the analysis , ensuring that the various dependencies being built into the process does not create a problem . for completeness
the modifications are presented in algorithm [ alg : streamsparsify ] .
[ alg : streamsparsify ] * stream - sparsification * + @xmath38 we use @xmath39 given @xmath40 ; once again @xmath41 is a constant which determines the probability of success .
we prove two theorems for algorithm [ alg : streamsparsify ] .
the first theorem is about the approximation ratio and the second theorem is about its space requirement . for the simplicity of proof
, we only consider sufficiently small @xmath19 .
[ thm : correctness ] given @xmath40 , @xmath21 is a sparsification , that is @xmath35 , with probability @xmath42 .
[ thm : space ] if @xmath35 , @xmath21 has @xmath43 edges .
we use a sequence of ideas similar to that in benczur and karger @xcite .
let us first discuss the proof in @xcite .
in that paper , theorem [ thm : benczur_errorbound ] is proved on three steps .
first , the result of karger @xcite , on uniform sampling is used .
this presents two problems .
the first is that they need to know the value of minimum cut to get a constant error bound .
the other is that the number of edges sampled is too large .
in worst case , uniform sampling gains only constant factor reduction in number of edges . to solve this problem , benczur and karger @xcite decompose a graph into @xmath14-strong connected components . in a @xmath14-strong connected component ,
minimum - cut is at least @xmath14 while the maximum number of edges in @xmath14-strong connected component(without @xmath44-strong connected component as its subgraph ) is at most @xmath45 .
they used the uniform sampling for each component and different sampling rate for different components . in this way
, they guarantee the error bound for every cut .
we can not use karger s result @xcite directly to prove our sparsification algorithm because the probability of sampling an edge depends on the sampling results of previous edges .
we show that the error bound of a single cut by a suitable bound on the martingale process . using that we prove that if we do not make an error until @xmath46 edge , we guarantee the same error bound for every cut after sampling @xmath47 edge with high probability . using union bound , we prove that our sparsification is good with high probability .
we prove theorem [ thm : correctness ] first .
first , we prove the error bound of a single cut in lemma [ thm : errorbound_singlecut ] . the proof will be similar to that of chernoff bound @xcite .
@xmath48 in lemma [ thm : errorbound_singlecomponent ] is a parameter and we use different @xmath48 for different strong connected components in the later proof .
[ thm : errorbound_singlecut ] let @xmath49 with @xmath50 be a cut in a graph @xmath9 such that @xmath51 and @xmath52 .
the index of the edges corresponds to the arrival order of the edges in the data stream .
let @xmath53 be an event such that @xmath54 for all @xmath55 .
let @xmath21 be a sparsification of @xmath9 .
then , @xmath56 < 2\exp(-\beta^2pc/4)$ ] for any @xmath57 .
let @xmath58 and @xmath59=pw_g(e_{i_j})$ ] .
then , @xmath60 if and only if @xmath61 .
as already mentioned , we can not apply chernoff bound because there are two problems : 1 .
@xmath62 are not independent from each other and 2 .
values of @xmath62 are not bounded .
the second problem is easy to solve because we have @xmath53 .
let @xmath63 be random variables defined as follows : @xmath64 if @xmath53 happens , @xmath65 .
thus , @xmath66 & = & { \mathbb{p}}[a_{c}\wedge(|\sum_j x_j-\sum_j \mu_j|>\beta pc ) ] \nonumber \\ & = & { \mathbb{p}}[a_{c}\wedge(|\sum_j y_j-\sum_j \mu_j|>\beta pc ) ] \nonumber \\ & \leq & { \mathbb{p}}[|\sum_j y_j-\sum_j \mu_j|>\beta pc ] \label{eqn : conclusion}\end{aligned}\ ] ] the proof of ( [ eqn : conclusion ] ) is similar to chernoff bound @xcite . however , since we do not have independent bernoulli random variables , we need to prove the upperbound of @xmath67 $ ] given @xmath68 .
we start with @xmath69 $ ] .
[ thm : yibound ] @xmath70\leq \exp(\mu_j(e^t-1))$ ] for any @xmath68 and @xmath71 .
there are two cases . given @xmath71 , @xmath72 or @xmath73 . at the end of each case , we use the fact that @xmath74 .
case 1 : if @xmath73 , @xmath75 .
@xmath76 & = & \exp(t\mu_j ) \nonumber \\ & < & \exp(\mu_j(e^t-1 ) ) .
\nonumber\end{aligned}\ ] ] case 2 : if @xmath77 , @xmath65 .
so @xmath70={p_{e_{i_j}}}\exp(t\mu_j/{p_{e_{i_j}}})+(1-{p_{e_{i_j}}})$ ] .
let @xmath78 .
observe that @xmath79 for @xmath80 .
so @xmath81 is decreasing function .
also we have @xmath82 since @xmath51 .
hence , @xmath83 therefore , @xmath76 & \leq & \mu_j(\exp(t)-1)+1 \nonumber \\ & \leq & \exp(\mu_j(e^t-1 ) ) .
\nonumber\end{aligned}\ ] ] from case 1 and 2 , @xmath70\leq\exp(\mu_j(e^t-1))$ ] for any @xmath71 .
now , we prove the upperbound of @xmath67 $ ] .
[ thm : sibound ] let @xmath84 . for any @xmath68 and @xmath71 , @xmath85\leq\exp(\sum_{k = j}^{l}\mu_j(e^t-1))$ ] .
we prove by induction . for @xmath86 , @xmath85={\mathbf e}[\exp(ty_l)|{h_{i_j-1}}]\leq\exp(\mu_l(e^t-1))$ ] by lemma [ thm : yibound ] .
assume that @xmath87\leq\exp(\sum_{k = j+1}^{l}\mu_k(e^t-1))$ ] for any @xmath88 .
then , @xmath89 & = & \sum_{y } { \mathbb{p}}[y_j = y|{h_{i_j-1 } } ] \sum_{{h_{i_{j+1}-1 } } } { \mathbf e}[\exp(t(y+s_{j+1}))|{h_{i_{j+1}-1}}]{\mathbb{p}}[{h_{i_{j+1}-1}}|y_j = y,{h_{i_j-1 } } ] \nonumber \\ & = & \sum_{y } \exp(ty ) { \mathbb{p}}[y_j = y|{h_{i_j-1 } } ] \sum_{{h_{i_{j+1}-1 } } } { \mathbf e}[\exp(ts_{j+1})|{h_{i_{j+1}-1}}]{\mathbb{p}}[{h_{i_{j+1}-1}}|y_j = y,{h_{i_j-1 } } ] \nonumber \\ & \leq & \sum_{y } { \mathbb{p}}[y_j = y|{h_{i_j-1 } } ] \exp\left(\sum_{k = j+1}^{l}\mu_k(e^t-1)\right ) \nonumber \\ & = & \exp\left(\sum_{k = j+1}^{l}\mu_k(e^t-1)\right ) { \mathbf e}[y_j|{h_{i_j-1 } } ] \nonumber \\ & \leq & \exp\left(\sum_{k = j}^{l}\mu_k(e^t-1)\right ) \nonumber\end{aligned}\ ] ] therefore , @xmath85\leq\exp(\sum_{k = j}^{n}\mu_k(e^t-1))$ ] for any @xmath71 and @xmath68 .
now we prove lemma [ thm : errorbound_singlecut ] . remember that we only need to prove @xmath90 < 2\exp(-\beta^2pc/4)$ ] by ( [ eqn : conclusion ] ) .
let @xmath91 and @xmath92 .
we prove in two parts : @xmath93\leq \exp(-\beta^2\mu/4)$ ] and @xmath94\leq \exp(-\beta^2\mu/4)$ ] .
we prove @xmath93<\exp(-\beta^2\mu/4)$ ] first . by applying markov s inequality to @xmath95 for any @xmath96
, we obtain @xmath97}{\exp(t(1+\beta)\mu ) } \nonumber \\ & \leq & \frac{\exp(\mu(e^t-1))}{\exp(t(1+\beta)\mu)}. \nonumber\end{aligned}\ ] ] the second line is from lemma [ thm : sibound ] . from this point , we have identical proof as chernoff bound @xcite that gives us bound @xmath98 for @xmath99 . to prove that @xmath94<\exp(-\beta^2pc/4)$ ] we applying markov s inequality to @xmath100 for any @xmath96 , and proceed similar to above . using union bound to these two bounds , we obtain a bound of @xmath101 .
now we prove the following lemma given a @xmath14-strong connected component and parameter @xmath48 .
this corresponds to the proof of uniform sampling method in @xcite .
[ thm : errorbound_singlecomponent ] let @xmath102 be a @xmath14-strong component such that each edge has weight at most 1 .
@xmath103 is its sparsified graph .
let @xmath104 for some constant @xmath105 .
suppose that @xmath106 be an event such that every edge in @xmath102 has sampled with probability at least @xmath48 .
then , @xmath107={\mathcal{o}}(1/n^{2+d}m)$ ] .
consider a cut @xmath11 whose value is @xmath108 in @xmath102 .
if @xmath106 holds , every edge in @xmath11 is also sampled with probability at least @xmath48 . by lemma [ thm : errorbound_singlecut ] , @xmath109\leq 2\exp(-\beta^2p\alpha k/4)=2(n^{4+d}m)^{-\alpha}$ ] .
let @xmath110 .
let @xmath111 be the number of cuts with value less or equal to @xmath108 .
by union bound , we have @xmath112\leq p(1)f(1)+\int_1^{\infty}p(\alpha)\frac{df}{d\alpha}d\alpha.\ ] ] the number of cuts whose value is at most @xmath113 times minimum cut is at most @xmath114 . since the value of minimum cut of @xmath102 is @xmath14 , @xmath115 . since @xmath116 is a monotonically increasing function , this bound
is maximized when @xmath117 .
thus , @xmath118 & \leq & f(1)p(1)+\int_1^{\infty}p(\alpha)\frac{df}{d\alpha}d\alpha \nonumber \\ & \leq & n^2p(1)+\int_1^{\infty}p(\alpha)(2n^{2\alpha}\ln n)d\alpha \nonumber \\ & \leq & \frac{2}{n^{2+d}m}+\int_1^{\infty}\frac{\ln n}{n^{\alpha ( 2+d)}m^{\alpha}}d\alpha \nonumber \\ & = & { \mathcal{o}}(1/n^{2+d}m ) .
\nonumber\end{aligned}\ ] ] [ thm : errorbound_singlestep ] the probability of @xmath119 being the first integer such that @xmath120 is @xmath121 .
if @xmath122 for all @xmath123 , @xmath124 . remember that @xmath31 denotes the strong connectivity of @xmath13 in graph @xmath9 .
@xmath125 @xmath126 is a sparsification of @xmath127 .
@xmath128 consists of @xmath129-strong connected components . for every @xmath130 , @xmath131 .
so it is sampled with probability at least @xmath132 .
if we consider one @xmath129-strong connected component and set @xmath39 , by lemma [ thm : errorbound_singlecomponent ] , every cut has error bound @xmath133 with probability at least @xmath134 . since there are less than @xmath135 such distinct strong connected components , with probability at least @xmath136 , @xmath137 for every @xmath138 .
hence , @xmath139 therefore , @xmath140={\mathcal{o}}(1/n^dm)$ ] . from lemma [ thm : errorbound_singlestep ] , theorem [ thm : correctness ] is obvious .
j < i.h_j \in ( 1\pm\epsilon)g_j)\wedge(h_i\notin ( 1\pm\epsilon)g_i)]={\mathcal{o}}(1/n^d)$ ] . for the proof of theorem [ thm : space ] , we use the following property of strong connectivity .
[ thm : connectivity_bound ] @xcite if the total edge weight of graph @xmath9 is @xmath142 or higher , there exists a @xmath14-strong connected components .
@xmath35 , total edge weight of @xmath21 is at most @xmath143 .
let @xmath144 be a cut @xmath145 .
since @xmath35 , @xmath146 .
total edge weight of @xmath21 is @xmath147 since each edge is counted for two such cuts .
similarly , @xmath9 has @xmath148 edges . therefore , if @xmath35 , total edge weight of @xmath21 is at most @xmath143 .
let @xmath149 .
@xmath150 is a set of edges that sampled with @xmath151 .
we want to bound the total weight of edges in @xmath150 .
@xmath152 .
let @xmath153 be a subgraph of @xmath21 that consists of edges in @xmath150 .
@xmath153 does not have @xmath154-strong connected component .
suppose that it has .
then there exists the first edge @xmath13 that creates a @xmath154-strong connected component in @xmath153 . in that case
, @xmath30 must be in the @xmath154-strong connected component .
however , since weight @xmath13 is at most @xmath155 , that component is at least @xmath44-strong connected without @xmath13 .
this contradicts that @xmath156 .
therefore , @xmath153 does not have any @xmath154-strong connected component . by lemma [ thm : connectivity_bound ] , @xmath152 .
now we prove theorem [ thm : space ] . if the total edge weight is the same , the number of edges is maximized when we sample edges with smallest strong connectivity .
so in the worst case , @xmath157 in that case , @xmath14 is at most @xmath158 .
let this value be @xmath159 . then
, total number of edges in @xmath21 is @xmath160
first , we prove a simple space lowerbound for weighted graphs , where the lowerbound depends on @xmath19 .
[ thm : lowerbound ] for @xmath161 , @xmath162 bits are required in order to sparsify every cut of a weighted graph within @xmath163 factor where @xmath11 is maximum edge weight and @xmath164 is minimum edge weight .
let @xmath165 be a set of graphs such that there is a center node @xmath166 and other nodes are connected to @xmath166 by an edge whose weight is one of @xmath167 .
then , @xmath168 .
for @xmath169 , they must have different sparsifications .
so we need @xmath170 bits for sparsfication .
it is easy to show that @xmath171 .
now we use the same proof idea for unweighted simple graphs .
since we can not assign weight as we want , we use @xmath172 nodes as a center instead of having one center node . in this way
, we can assign degree of a node from @xmath164 to @xmath172 .
[ thm : unweighted_lowerbound ] for @xmath161 , @xmath173 bits are required in order to sparsify every cut of a graph within @xmath163 .
consider bipartite graphs where each side has exactly @xmath172 nodes and each node in one side has a degree @xmath174 or @xmath172 .
for each degree assignment , there exists a graph that satisfies it .
let @xmath165 be a set of graphs that has different degree assignments .
then , @xmath175 .
@xmath169 can not have the same sparsification .
so we need at least @xmath176 bits . another way of viewing the above claim is a direct sum construction , where we need to use @xmath177 bits to count upto a precision of @xmath1 .
we presented a one pass semi - streaming algorithm for the adversarially ordered data stream model which uses @xmath178 edges to provide @xmath19 error bound for cut values with probability @xmath179 .
if the graph does not have parallel edges , the space requirement reduces to @xmath180 . we can solve the minimum cut problem or other problems related to cuts with this sparsification . for the minimum cut problem , this provides one - pass @xmath181-approximation algorithm .
andrs a. benczr and david r. karger . approximating s - t minimum cuts in o(n2 )
time . in _
stoc 96 : proceedings of the twenty - eighth annual acm symposium on theory of computing _ , pages 4755 , new york , ny , usa , 1996 .
chandra s. chekuri , andrew v. goldberg , david r. karger , matthew s. levine , and cliff stein .
experimental study of minimum cut algorithms . in _
soda 97 : proceedings of the eighth annual acm - siam symposium on discrete algorithms _ , pages 324333 , philadelphia , pa , usa , 1997 .
society for industrial and applied mathematics .
jianxiu hao and james b. orlin . a faster algorithm for finding the minimum cut in a graph . in _ soda 92 : proceedings of the third annual acm - siam symposium on discrete algorithms _ , pages 165174 , philadelphia , pa , usa , 1992 . society for industrial and applied mathematics .
david r. karger .
global min - cuts in rnc , and other ramifications of a simple min - out algorithm . in _ soda 93 : proceedings of the fourth annual acm - siam symposium on discrete algorithms _ , pages 2130 , philadelphia , pa , usa , 1993 .
society for industrial and applied mathematics .
david r. karger .
random sampling in cut , flow , and network design problems . in _ stoc 94 : proceedings of the twenty - sixth annual acm symposium on theory of computing _ , pages 648657 , new york , ny , usa , 1994 .
acm .
daniel a. spielman and nikhil srivastava .
graph sparsification by effective resistances . in _
stoc 08 : proceedings of the 40th annual acm symposium on theory of computing _ , pages 563568 , new york , ny , usa , 2008 . | analyzing massive data sets has been one of the key motivations for studying streaming algorithms . in recent years
, there has been significant progress in analysing distributions in a streaming setting , but the progress on graph problems has been limited .
a main reason for this has been the existence of linear space lower bounds for even simple problems such as determining the connectedness of a graph .
however , in many new scenarios that arise from social and other interaction networks , the number of vertices is significantly less than the number of edges .
this has led to the formulation of the semi - streaming model where we assume that the space is ( near ) linear in the number of vertices ( but not necessarily the edges ) , and the edges appear in an arbitrary ( and possibly adversarial ) order .
however there has been limited progress in analysing graph algorithms in this model . in this paper
we focus on graph sparsification , which is one of the major building blocks in a variety of graph algorithms .
further , there has been a long history of ( non - streaming ) sampling algorithms that provide sparse graph approximations and it a natural question to ask : since the end result of the sparse approximation is a small ( linear ) space structure , can we achieve that using a small space , and in addition using a single pass over the data ?
the question is interesting from the standpoint of both theory and practice and we answer the question in the affirmative , by providing a one pass @xmath0 space algorithm that produces a sparsification that approximates each cut to a @xmath1 factor .
we also show that @xmath2 space is necessary for a one pass streaming algorithm to approximate the min - cut , improving upon the @xmath3 lower bound that arises from lower bounds for testing connectivity . |
throughout this paper we let @xmath0 be a convex function with @xmath1 on @xmath2 and let @xmath3 denote its induced monge ampre measure @xmath4 associated to @xmath5 there are three , typically degenerate / singular , elliptic operators @xmath6 , @xmath7 , and @xmath8 defined as @xmath9 where @xmath10 stands for the matrix of cofactors of @xmath11 , that is , @xmath12 from the fact that the columns of @xmath10 are divergence - free , it follows that @xmath13 the elliptic equation @xmath14 is the linearization of the monge ampre equation @xmath15 at the function @xmath5 . the first identity in implies that @xmath7 admits both nondivergence ( trace ) and divergence ( variational ) forms . in their seminal works @xcite , l. caffarelli and
c. gutirrez developed a real analysis associated to @xmath5 leading to their groundbreaking proof of a harnack inequality for nonnegative solutions to @xmath16 .
as a crucial feature of their approach stands the description of the intrinsic geometry to study the linearized monge ampre equation .
this geometry is given by the monge
ampre sections of @xmath5 defined as @xmath17 where @xmath18 and @xmath19 are called the center and the height of the section @xmath20 , respectively , and @xmath21 denotes the dot product in @xmath2 .
notice that the case of @xmath22 accounts for the laplacian and the euclidean balls , since @xmath23 and @xmath24 for every @xmath25 and @xmath26 .
the caffarelli
gutirrez regularity theory for @xmath7 was originally motivated by its applications to fluid dynamics ( see ( * ? ? ?
* section 1 ) and references therein ) .
further applications have emerged , for instance , in relation to the affine plateau problem in affine geometry and the prescribed affine mean curvature equation , as exposed in the work of n. trudinger and x .- j wang in @xcite , n. q. le in @xcite , and references therein .
after @xcite , both the regularity theory and the associated real analysis for the linearized monge ampre equation have seen further progress .
the caffarelli gutirrez harnack inequality has been proved to hold under minimal geometric conditions on @xmath5 in @xcite . it has been later extended as to allow for lower - order terms in @xcite and by n. q. le in @xcite .
interior @xmath27- , @xmath28- and @xmath29-estimates for solutions to @xmath30 have been established by c. gutirrez and t. nguyen in @xcite and c. gutirrez and f. tournier in @xcite , respectively .
global ( up to the boundary ) @xmath27- and @xmath29-estimates have been proved by n. q. le and o. savin in @xcite and n. q. le and t. nguyen @xcite , respectively .
estimates for green s functions on monge
ampre sections have been established in @xcite and by n. q. le in @xcite .
a liouville - type theorem for entire solutions to @xmath16 in @xmath31 has been proved by o. savin in @xcite .
sobolev and poincar - type inequalities associated to @xmath7 have been proved in @xcite and by g. tian and x .- j .
wang in @xcite . in this paper
we develop a nonlocal version of the linearized monge ampre equation and establish a harnack inequality on monge ampre sections .
more precisely , our purpose is to accomplish the following goals .
a. [ goal : define : ls : lls ] to define the fractional powers @xmath32 and @xmath33 on arbitrary monge
ampre sections ( within their natural nondivergence / divergence form contexts ) and to prove existence and uniqueness of solutions to the nonlocal equations @xmath34
b. [ goal : ls = lls ] to show that the interplay between the ( local ) nondivergence and divergence structures in persists on the ( nonlocal ) operators @xmath32 and @xmath33 . c. [ goal : h : l ] to prove , under minimal geometric assumptions on @xmath5 , a harnack inequality for nonnegative solutions to on the sections of @xmath5 , showing in particular that the monge ampre geometry carries over to our nonlocal equations .
the goals above could be regarded as a linear counterpart to the current efforts to provide a correct definition of a fractional nonlocal monge ampre equation by l. caffarelli and f. charro @xcite and l. caffarelli and l. silvestre @xcite .
our results will hold true for every @xmath35 .
regarding , we should first observe that @xmath6 is an operator in nondivergence form . in section [ section : extension nondivergence ]
, we show how to define the fractional powers @xmath36 on arbitrary monge
ampre sections via the semigroup generated by @xmath6 . in section [ sec
: example : ls ] we illustrate the definition of @xmath32 by computing an explicit example of its action on the monge ampre quasi - distance . using the corresponding eigenfunctions , in section [ section : extension divergence ] we define the fractional powers @xmath37 .
then , in terms of existence and uniqueness of solutions to , we have the following result ( see subsections [ subsec : proof : exist : l ] and [ subsec : proof : exist : ll ] ) .
[ thm : existence ] fix a section @xmath38 . 1 . [ exist : l ] given any @xmath39 there exists a unique solution @xmath40 to @xmath41 2 .
[ exist : ll ] given any @xmath42 there exists a unique solution @xmath43 to @xmath44 as far as is concerned , we show that the fractional powers @xmath32 and @xmath33 do preserve the dual nondivergence / divergence nature of @xmath6 and @xmath8 from .
the following equality is proved in section [ sec : ls = lls ] .
[ thm : l = ll ] fix a section @xmath38 .
then @xmath45 for every @xmath46 . regarding ,
let us mention that by minimal geometric assumption on @xmath5 we mean a doubling condition for @xmath3 on the sections of @xmath5 known as the @xmath47-doubling condition , see . throughout the article , a _
geometric constant _ will be a constant depending only on the @xmath47 constants , dimension @xmath48 , and @xmath49 .
[ h : nondiver ] assume @xmath50 .
there exist geometric constants @xmath51 and @xmath52 such that for every section @xmath53 , every @xmath54 , every @xmath55 solution to @xmath56 and every section @xmath57 , the following harnack inequality holds true @xmath58 furthermore , there exist geometric constants @xmath59 and @xmath60 such that @xmath61 , where @xmath62 denotes the intrinsic monge
ampre quasi - distance defined in the harnack inequality makes a case for the central role of the monge ampre geometry , based on the sections @xmath63 , also in the study of the nonlocal operators @xmath32 .
in addition , from the definition of @xmath64 in , we have @xmath65 and , by the sobolev embedding , @xmath66 for every @xmath67 , where @xmath68 is the local @xmath69-hlder class with respect to the euclidean distance .
now , inequality says that @xmath70 with respect to the intrinsic monge
ampre quasi - distance @xmath62 . for the particular choice @xmath22 , theorem [ h : nondiver ]
complements , by also including a nonhomogeneous right hand side @xmath71 , theorem a from @xcite .
indeed , such a result implies a harnack inequality for nonnegative solutions to the fractional nonlocal equation @xmath72 in a ball @xmath73 . here
@xmath74 stands for the dirichlet laplacian in the unit ball @xmath75 .
an essential tool for the proof of our main results is the extension problem characterization of the fractional nonlocal operators @xmath32 and @xmath33 . the celebrated extension problem for the fractional laplacian on @xmath2
was first explored from the pde point of view in the pioneering work of l. caffarelli and l. silvestre @xcite .
this is a far reaching technique that allows to handle nonlocal problems for @xmath76 in a local way through a degenerate elliptic equation in @xmath77-dimensions .
later on , the semigroup language approach and the extension problem for fractional powers of positive operators was developed in @xcite . in @xcite a harnack inequality for fractional nonlocal elliptic equations admitting variational form was proved . the most general extension problem so far has been obtained in @xcite .
it includes fractional powers of closed operators in banach spaces allowing , in particular , to deal with nonvariational equations .
thus the results in @xcite apply to our nondivergence form elliptic operator @xmath6 .
theorems [ thm : existence ] and [ thm : l = ll ] will be consequences of the semigroup language approach , the localization provided by the extension problem of @xcite and the variational structure of @xmath6 given by .
the main steps in the proof of theorem [ h : nondiver ] are as follows .
first , given @xmath54 and the nonnegative solution @xmath78 , we will establish the equivalence between the fractional nonlocal equation and the local degenerate / singular extension problem in one more variable @xmath79 as shown in @xcite .
here @xmath80 is an explicit constant defined in .
there is a unique nonnegative solution @xmath81 , vanishing at infinity in the sense of , such that @xmath82 second , by setting @xmath83 for every @xmath84 in the cylinder @xmath85 we have that @xmath86 solves @xmath87 equation can be recast as the linearized monge ampre equation @xmath88 where @xmath89 is the strictly convex function ( recall that @xmath90 ) defined as @xmath91 in section [ sec : s(q ) ] we set @xmath92 and define a functional class @xmath93 that contains @xmath94 .
then , theorem [ h : nondiver ] will result from and the following harnack inequality ( see section [ sec : s(q ) ] for notation ) .
[ thm : h : vtilde ] assume @xmath50 .
then , there exist geometric constants @xmath51 and @xmath95 such that for every nonnegative @xmath96 solution to @xmath97 pointwise in @xmath98 and every section @xmath99 with @xmath100 the following harnack inequality holds true @xmath101 here @xmath102 stands for the @xmath103-norm of the normal derivative of @xmath104 on the intersection @xmath105 , as defined in . consequently
, there exist geometric constants @xmath59 and @xmath106 such that @xmath107 .
the major obstacles in the proof of theorem [ thm : h : vtilde ] are the following : a. on the hyperplane @xmath108 , the matrix @xmath109 becomes degenerate ( if @xmath110 ) or singular ( if @xmath111 ) .
this degeneracy / singularity prevents the direct application ( even under stronger assumptions than @xmath50 ) of known harnack inequalities for the linearized monge ampre equation including ( * ? ? ?
* theorem 5 ) , ( * ? ? ?
* theorem 1 ) , ( * ? ? ?
* theorem 1.4 ) , and ( * ? ? ?
* theorem 1 ) , which require continuous second - order derivatives as well as positive - definite hessians for the underlying convex function .
b. furthermore , the fact that @xmath112 may vanish implies that the caffarelli gutirrez proof of the `` passage to the double section '' ( that is , ( * ? ? ?
* theorem 2 ) , which is essential for the weak - harnack inequality in ( * ? ? ?
* theorem 4 ) ) can not be applied to .
c. also , since @xmath113 becomes degenerate or singular , the solvability of the dirichlet problem for @xmath114 in can not be taken for granted and then the argument from @xcite used to prove the critical - density estimate ( see ( * ? ? ?
* theorem 2 ) ) can not be carried out for .
d. the function @xmath94 does not belong apriori to @xmath115 , which prevents its direct use in the aleksandrov
pucci maximum principle .
notice that the hyperplane @xmath108 is central to our approach because it is precisely there where the connection between @xmath94 and @xmath116 takes place , see .
in addition to the obstacles above , we pursue ( and prove ) theorem [ thm : h : vtilde ] under minimal geometric conditions , that is , under @xmath50 only .
this is in contrast with @xcite . in order to overcome those obstacles , in section [ sec : cd ]
we show that , despite the degeneracy or singularity of @xmath113 and under the hypothesis @xmath50 only , a critical - density estimate for @xmath94 ( theorem [ thm : cd ] below ) can be established , for _ some _ @xmath117 , in terms of the @xmath118-norm of the normal derivative @xmath119 .
that is , a suitable control on the normal derivative will allow for a critical - density estimate in the absence of the hypothesis of continuous second - order derivates and positive - definite hessian .
this represents a novelty in the study of the linearized monge ampre equation . with a critical - density estimate at hand
, mean - value inequalities ( see theorems [ thm : lb : q=1 ] and [ thm : lb ] ) can be obtained by fairly standard arguments , as described in section [ sec : lb ] .
the next step is to prove a weak - harnack inequality for @xmath94 .
as mentioned , the methods from ( * ? ? ?
* sections 2 and 3 ) are not applicable .
instead , we rely on the variational side of the linearized monge ampre equation .
the idea is the following . by
, we have @xmath120 , and we know that @xmath121 comes from the extension problem associated to @xmath6 .
now , by developing the extension problem associated to the divergence - form operator @xmath8 from @xcite ( namely , below ) we find a solution @xmath122 . certainly , after contrasting the extension problems associated to @xmath6 and @xmath8 , one can not expect that @xmath123 .
however , in section [ sec : ls = lls ] we prove that , modulo a change of variables ( more precisely , a change in the extension variables @xmath124 and @xmath125 ) , the equality @xmath126 does hold true .
a key consequence of this equality is the energy estimate for @xmath81 given by in remark [ rmk : finite : e ] . although @xmath81 solves the pde in which can also be written in divergence form , we were able to arrive at the energy estimate only by going through @xmath127 . with this insight on the variational side of @xmath81 , in section [ sec : pi : f ]
we develop a weak poincar inequality associated to @xmath128 .
one advantage of working in the variational context ( which is based on energy estimates ) is that the pointwise singularity of @xmath129 will become inconsequential because @xmath130 for every @xmath49 . in turn , in section [ sec : every : cd ] we use the poincar inequality to prove a critical - density estimate for _ every _ @xmath131 ( see theorem [ thm : every : cd ] ) . in section [ sec : weak : h ] we use the fact that the critical density can be taken small enough and combine it with a covering lemma to prove a weak - harnack inequality for @xmath94 ( see theorem [ thm : weak : h ] ) . finally , in section
[ sec : proofs : main ] we bring all the previous results together to prove theorems [ h : nondiver ] and [ thm : h : vtilde ] .
we close this introduction by mentioning that no use of the normalization technique from ( * ? ? ?
* section 1 ) or of the local monge ampre-@xmath132 space ( which dominated the variational approach in @xcite ) is made in this article .
fix any section @xmath133 . on @xmath134
we consider the linearized monge ampre equation with homogeneous dirichlet boundary condition : @xmath135 our first goal is to introduce the semigroup generated by @xmath6 .
the matrix of coefficients @xmath136 is symmetric and positive definite , with entries in @xmath137 .
since @xmath138 is a compact set , such a matrix is uniformly elliptic on @xmath134 .
notice that we use this fact without ever resorting to estimates depending on the size of the eigenvalues of @xmath139 for @xmath140 .
let the domain of @xmath6 be the space @xmath141 where @xmath142 is the banach space @xmath143 endowed with the @xmath144 norm .
here as usual @xmath145 denotes the space of continuous functions on @xmath138 under the @xmath144 norm .
it is clear that @xmath146 depends on the section @xmath134 .
also , let us define @xmath147 as @xmath148 it follows from ( * ? ? ?
* theorem 4.1 ) that @xmath149 generates a bounded holomorphic semigroup @xmath150 on @xmath145 . for convenience
we recall the relevant definitions , see for example @xcite .
the family @xmath150 is a _ semigroup _ on @xmath145 ( see ( * ? ? ?
* section 1.1 ) ) if the following conditions hold : a. for each @xmath151 , @xmath152 is a bounded linear operator from @xmath145 into itself ; b. the semigroup property holds : for every @xmath153 and for any @xmath154 , @xmath155 c. for every @xmath154 we have @xmath156 .
the semigroup @xmath152 is _ bounded holomorphic _ if the operator valued function @xmath157 from @xmath158 into the algebra of bounded linear operators on @xmath145 has a holomorphic extension to an open sector of the complex plane contained in @xmath159 , which is bounded on proper subsectors , see @xcite .
it is shown in ( * ? ? ?
* proposition 4.4 ) that @xmath149 is dissipative , see also ( * ? ? ?
* lemma 3.4.2 ) . then , as a consequence of ( * ? ? ?
* proposition 3.7.16 ) , we obtain that the semigroup is a _
contraction_. in other words , for every @xmath154 , @xmath160 the fact that @xmath149 is the _ generator _ of the semigroup @xmath152 means that for every @xmath161 the function @xmath162 is the unique solution to the parabolic equation @xmath163 see ( * ? ? ?
* sections 1.1 and 1.2 ) .
observe that the semigroup @xmath152 , though a contraction , is not a @xmath164-semigroup on @xmath145 because @xmath146 is not dense in @xmath145 , see ( * ? ? ?
yosida teorem in section 1.3 ) .
in particular , in order for @xmath165 above to converge to the initial data @xmath166 uniformly in @xmath134 we need to take @xmath161 . as @xmath152 is a bounded holomorphic semigroup , from ( * ? ? ? * theorem 3.7.19 ) we have that @xmath167 finally , the following _ decay estimate _ holds : there are constants @xmath168 such that , for every @xmath161 , @xmath169 see ( * ? ? ?
* theorem 1 ) , also ( * ? ? ?
* theorem 4.1 , proposition 4.7 ) .
[ rem : positive ] it is important to notice that the semigroup @xmath152 is _ positive _ on @xmath146 .
namely , if @xmath170 and @xmath171 in @xmath134 then @xmath172 , for every @xmath173 and @xmath151 . indeed , this follows from the well known weak minimum principle for parabolic equations in nondivergence form .
the semigroup generated by @xmath6 allows us to define the fractional powers @xmath175 as in ( * ? ? ?
* , ( 5 ) ) and ( * ? ? ?
* theorem 4.1 ) .
[ definition : fractional nondivergence ] let @xmath49 .
the fractional operator @xmath175 is defined for any @xmath161 and every @xmath173 as @xmath176 the integral in is absolutely convergent in the sense of bochner .
indeed , the integrand @xmath177 is a function of @xmath178 with values in @xmath142 and , since @xmath161 and @xmath179 is a contraction , we have ( see ( * ? ? ?
* , ( 2.5)(2.6 ) ) ) , @xmath180 on the other hand , by contractivity , @xmath181 , for every @xmath151 .
therefore , for any @xmath182 , @xmath183 in particular , the following fractional sobolev - type interpolation inequality holds @xmath184 for any @xmath161 , @xmath182 and @xmath49 , and , as a consequence , @xmath185 we finally notice that , unlike the local differential operator @xmath6 that in general has values in @xmath145 , see , the fractional nonlocal operator @xmath174 has range in @xmath142 , namely , @xmath186 the semigroup expression for @xmath187 in yields the following maximum principle .
let @xmath171 and suppose that @xmath188 for some @xmath189 .
then @xmath190 in @xmath134 .
moreover , @xmath191 if and only if @xmath192 in @xmath134 .
indeed , since the semigroup is positive ( see remark [ rem : positive ] ) we have @xmath193 and the first conclusion follows from by noticing that @xmath194 . furthermore ,
@xmath195 if and only if @xmath196 for all @xmath197 , which in this case is equivalent to @xmath192 in @xmath134 by the usual weak maximum principle for parabolic equations . given @xmath39 , it follows from general theory ( see ( * ? ? ?
* section 3.8 ) and ( * ? ? ? * chapter ix , section 11 ) ) that @xmath198 is a solution to @xmath199 in @xmath134 with @xmath200 on @xmath201 . to prove uniqueness , assume that @xmath202 is a non trivial solution to @xmath203 in @xmath134 .
let @xmath189 be a point such that @xmath204 .
by changing @xmath116 by @xmath205 we can always assume that @xmath206 .
then , from and we conclude that @xmath207 . the case @xmath208 is excluded by hypothesis .
thus , @xmath209 for every @xmath197 , contradicting .
the proof of uniqueness is complete . using the language of semigroups we can characterize @xmath174 with the extension problem established in ( * ? ? ?
* theorem 1.1 , theorem 2.1 ) .
observe that the main results in @xcite apply in principle to generators of bounded @xmath164-semigroups .
nevertheless , it is easy to follow the proofs there and conclude that we can extend them to our present case .
let @xmath161 . for @xmath173 and @xmath210
we define @xmath211 then @xmath212 and , for each @xmath210 ,
@xmath213 as a function of @xmath125 , @xmath214 is @xmath215 , for every @xmath173 .
furthermore , @xmath81 is a classical solution to the extension problem @xmath216 indeed , by usual elliptic regularity , @xmath217 for every @xmath218 .
moreover , @xmath219 and @xmath220 uniformly in @xmath134 , where @xmath221 hence , @xmath222 is a solution to the extension problem @xmath223 for all the details see @xcite .
it is clear that the even extension of @xmath214 to @xmath224 given by @xmath225 satisfies @xmath226 for any @xmath218 , and solves @xmath227 given a function @xmath39 , let @xmath161 be the solution to @xmath199 in @xmath134 .
the solution @xmath81 in can also be written as @xmath228 then , for every @xmath173 we readily get @xmath229 for the details see ( * ? ? ?
* theorems 1.1 , theorem 2.1 ) , also @xcite .
[ rem : uniqueness ] by the weak maximum principle for elliptic equations , it is easy to see that there is at most one solution to the extension problem such that @xmath230 using the semigroup decay it is readily checked that @xmath214 as defined in satisfies , so this is indeed the unique solution .
in this section we give two explicit examples on the action of @xmath32 .
our examples are inspired by the identity @xmath231 [ thm : ls : example ] given an arbitrary section @xmath133 introduce the function @xmath232 then , for every @xmath49 , @xmath233 notice that @xmath234 and that @xmath235 in @xmath134 .
also , from the definition of @xmath236 in , we have @xmath237 in @xmath134 . in order to show we first need to find the unique solution @xmath81 to the extension equation @xmath238 that satisfies in remark [ rem : uniqueness ] .
we do so by pursuing a solution @xmath81 of the form @xmath239 where @xmath240 has to be found .
notice that @xmath241 hence , by using the equation @xmath242 , the function @xmath243 must be a solution to @xmath244 that decays to zero as @xmath245 .
the equation is a bessel equation and in order to find its unique solution we follow the analysis in ( * ? ? ?
* ; * ? ? ?
* section 3.1 ) , see also @xcite for details about bessel functions . to simplify the notation , fix @xmath246 and set @xmath247 thus
, the general solution to is @xmath248 where @xmath249 denotes a general cylinder function , see @xcite . by using the boundary condition @xmath250 as @xmath245 ( see the asymptotic expansions of bessel functions in ( * ? ? ?
* ; * ? ? ?
* section 3.1 ) ) we obtain the modified bessel function of the second kind @xmath251 : @xmath252 where @xmath253 is an arbitrary constant .
recall that @xmath254 as @xmath255 .
hence , @xmath256 and , in order to satisfy the initial condition @xmath257 , we impose @xmath258 therefore , @xmath259 is the unique bounded solution to .
it is clear that @xmath243 .
moreover , from the asymptotic behavior @xmath260 , as @xmath261 , it follows that @xmath262 exponentially , as @xmath245 . in conclusion ,
@xmath263 with @xmath264 as in and @xmath265 as in , is the unique positive bounded solution to that satisfies , see remark [ rem : uniqueness ] . on the other hand ,
from we know that if @xmath81 is the unique solution to the extension problem then @xmath266 again , to simplify the computation , let us put @xmath267 so that @xmath268 by using the chain rule , the properties of the derivatives of bessel functions and the fact that @xmath269 , we can compute @xmath270 whence , from , and the asymptotic behavior of @xmath271 as @xmath255 , we get @xmath272 recalling the definition of @xmath265 from and the identity , we arrive at . as a consequence of theorem [ thm : ls : example ] we obtain the following result on the fractional dirichlet laplacian .
for every @xmath49 we have @xmath273 where @xmath274 is the fractional dirichlet laplacian in the unit ball @xmath275 .
use theorem [ thm : ls : example ] with @xmath276 for every @xmath277 and notice that @xmath278 and @xmath279 , which gives @xmath280 .
this section is devoted to the definition of the fractional powers of the divergence form operator @xmath8 subject to the homogeneous dirichlet boundary condition . as in section [ section : extension nondivergence ] , fix any section @xmath133 . for @xmath281 , consider the following dirichlet problem for @xmath8 : @xmath282 observe that the right hand side @xmath71 in appears multiplied by the monge
ampre measure @xmath283 .
this can always be assumed by considering @xmath284 .
the fractional powers @xmath33 , @xmath49 , will be defined by using the dirichlet eigenfunctions and eigenvalues along the lines of @xcite .
let @xmath286 denote the completion of @xmath287 with respect to the norm @xmath288 here @xmath289 stands for the monge
ampre gradient , which is defined as @xmath290 by the sobolev inequality for the monge
ampre quasi - metric structure , see ( * ? ? ?
* theorem 1 ) , an equivalent norm in @xmath286 is @xmath291 . a weak solution @xmath292 to is a function @xmath293 such that @xmath294 notice that the matrix of coefficients @xmath295 is symmetric and uniformly elliptic in the compact set @xmath138 and that @xmath296 is isometrically embedded in @xmath297 .
therefore , by standard techniques ( see for example ( * ? ? ?
* chapter 6 ) or ( * ? ? ?
* section 8.12 ) ) , there exist a sequence of eigenvalues @xmath298 and a corresponding family of eigenfunctions @xmath299 such that @xmath300 in the weak sense . in other words , for every @xmath301 and every @xmath302 , @xmath303 moreover , @xmath304 forms an orthonormal basis of @xmath296 . for @xmath305
, we consider the hilbert space @xmath306 endowed with the inner product @xmath307 where @xmath308 .
observe that @xmath309 . from it
is readily verified that @xmath310 as hilbert spaces and @xmath311 we read the right - hand side in as the definition of @xmath312 for @xmath313 .
we are now in position to define the fractional power @xmath33 in @xmath314 .
let @xmath49 .
the fractional operator @xmath315 is defined for any @xmath316 as the unique element @xmath37 in the dual space @xmath317 acting as @xmath318 given @xmath319 in @xmath320 , the unique solution @xmath321 is given by @xmath322 .
in particular , if @xmath323 then @xmath324 .
let us introduce @xmath325 and , for @xmath173 , @xmath326 by reasoning as in @xcite , see also @xcite , the extension problem characterization of can now be obtained as follows .
let @xmath316 .
we say that a function @xmath327 , defined for @xmath173 and @xmath328 , is a weak solution to the extension problem @xmath329 if @xmath127 , @xmath330 and @xmath331 belong to the weighted space @xmath332 , @xmath333 on @xmath334 in the sense of traces , @xmath335 as @xmath336 in @xmath337 , and for every test function @xmath338 such that @xmath339 in @xmath134 , we have @xmath340 by proceeding as in ( * ? ? ?
* ; * ? ? ?
* section 3.3.1 ) or ( * ? ? ?
* ; * ? ? ?
* section 3.1 ) , see also ( * ? ? ?
* section 2.3 ) , the unique weak solution @xmath127 that weakly vanishes as @xmath341 can be written using the fourier coefficients @xmath342 of @xmath292 and the eigenfunctions @xmath343 as @xmath344 where the coefficients @xmath345 are given by @xmath346 here @xmath347 is the modified bessel function of the second kind and parameter @xmath348 .
moreover , @xmath349 where @xmath350 that is , if @xmath351 is a test function then reads @xmath352 by using @xmath127 as a test function we obtain the energy identity @xmath353 where we used that @xmath354 .
[ rem : semigroup definition ] as in @xcite we could have used the semigroup generated by @xmath8 to obtain an equivalent expression for @xmath355 which explicitly shows that the fractional operator @xmath285 is a nonlocal integro - differential operator in divergence form . it is also possible to write the solution @xmath127 above by using the semigroup generated by @xmath8 ( see @xcite ) . instead , by means of a change of variables , in section [ sec : ls = lls ] we will directly relate the solution @xmath127 of the extension problem in divergence form to the solution of the extension problem _ in nondivergence form _ .
in this section we establish the connection between the nondivergence form and divergence form extension problems and , which will ultimately leads us to the proof of theorem [ thm : l = ll ] .
let us start with the following proposition .
[ prop : dl : in : dll ] for every section @xmath133 the following inclusion holds true @xmath356 recall from definition [ definition : fractional nondivergence ] that @xmath357 as previously defined in . on the other hand , @xmath358 ,
see . let us remark that , since the eigenvalues @xmath359 from increase towards @xmath360 , we have @xmath361 for all sufficiently large @xmath362 , and then @xmath363 . also , the fact that the hessian @xmath364 is a positive definite matrix with entries in @xmath365 implies that @xmath366 and @xmath367 as hilbert spaces . here
@xmath368 denotes the usual sobolev space with respect to lebesgue measure .
now , given @xmath369 we have @xmath370 as well as @xmath371 , and by the @xmath372-solvability theorem for the dirichlet problem ( see , for instance , theorem 9.15 on @xcite and notice that @xmath373 ) , we obtain @xmath374 , which , combined with @xmath370 , gives @xmath375 .
hence , @xmath376 which proves .
the connection between the fractional powers @xmath32 and @xmath33 will materialize through the change of variables @xmath377 see @xcite .
define @xmath378 for @xmath173 and @xmath379 , where @xmath81 is the unique solution to the extension equation satisfying , see section [ section : extension nondivergence ] .
then , @xmath380 and @xmath381 therefore , @xmath382 also , from , @xmath383 where in the second identity we noticed that @xmath8 acts only in the variable @xmath173 for each fixed @xmath210 and @xmath379 .
therefore , from and , since @xmath81 is the solution to the extension equation , @xmath384 where @xmath385 is as in .
therefore , @xmath127 defined by is a solution to with @xmath386 for @xmath369 .
moreover , by , we see that @xmath387 , as @xmath341 , weakly in @xmath337 .
hence @xmath127 in is the unique solution to . from , for every @xmath173 , @xmath388 on the other hand
, it is readily verified that @xmath389 thus , @xmath390 hence , as @xmath391 , for @xmath369 we get @xmath392 therefore , @xmath393 and both the divergence and nondivergence structures occur simultaneously .
notice that , in order to keep the equality in consistent with , the righthand sides of @xmath394 and @xmath395 in must differ by @xmath3 , as stemming from the equations and .
[ rmk : finite : e ] the following finite - energy estimate will play a key role in section [ sec : s(q ) ] : @xmath396 we prove by using the change of variables .
indeed , @xmath397 on the other hand , from , @xmath398 therefore , by the energy identity , @xmath399\\ & = ( 2s)^{2s -1 } c_s\int_{s } |{\mathcal{l}}_{\varphi}^{s/2}u(x)|^2 { \ , d{\mu_{\varphi}}}(x ) < \infty .
\end{aligned}\ ] ]
throughout the article , the function @xmath400 will denote a generic convex function which is used as a placeholder for the functions @xmath5 , @xmath128 , and @xmath401 to be introduced in section [ sec : def : phi ] .
let also @xmath402 denote a generic dimension that will take the values @xmath48 , @xmath403 , or @xmath404 . for a strictly convex function @xmath405 ( strictly convex in the sense that its graph contains no line segments ) , its associated monge
ampre measure @xmath406 acts on a borel set @xmath407 as @xmath408 where @xmath409 denotes the lebesgue measure of a subset @xmath410 . given @xmath411 and @xmath412 , its monge ampre section @xmath413 is defined as the open , convex set @xmath414 where @xmath415 we write @xmath416 if there exists a constant @xmath417 such that @xmath418 where , for a convex set @xmath134 , @xmath419 denotes its @xmath420-contraction with respect to its center of mass ( with respect to the lebesgue measure ) . the condition @xmath416 is equivalent to the structure of space of homogeneous type for the triple @xmath421 ( see @xcite and references therein ) , which we understand as the minimal structure to carry out real analysis .
it is in this sense that we refer to @xmath416 as a minimal geometric condition .
if @xmath422 is twice differentiable at point @xmath423 with @xmath424 , then the matrix of cofactors of @xmath425 at @xmath426 is given by @xmath427 if @xmath400 is three times differentiable at a point @xmath423 with @xmath428 , then @xmath429 has divergence - free columns ( see , for instance , @xcite )
. then , given @xmath430 twice differentiable at @xmath426 we have @xmath431 suppose that @xmath432 with @xmath433 .
if @xmath416 then @xmath434 is a continuously differentiable homeomorphism and the convex conjugate of @xmath400 , denoted by @xmath435 , satisfies @xmath436 in particular , we have @xmath437 in addition , @xmath416 implies @xmath438 ( with doubling constants depending on the ones for @xmath406 and dimension @xmath402 ) and the monge ampre sections of @xmath400 and @xmath439 are related as follows @xmath440 for constants @xmath441 depending only on @xmath442 in and dimension @xmath402
. see ( * ? ? ?
* section 5 ) for these and related results .
associated to a strictly convex @xmath432 with @xmath433 and @xmath416 , poincar inequalities with respect to its monge ampre sections and gradient @xmath443 have been proved in ( * ? ? ?
* theorem 1.3 ) .
namely , there exists a constant @xmath444 , depending only on the @xmath445 constant and @xmath402 , such that for every section @xmath446 and every @xmath447 we have @xmath448 where @xmath449 and @xmath450 . from now on , for a borel measure @xmath451 , which will be either a monge ampre measure @xmath406 or the lebesgue measure , and a measurable set @xmath407 we put @xmath452 here we record the following consequence of lemma 3.1 from @xcite : given any strictly convex function @xmath455 ( no doubling assumptions required ) and any section @xmath456 , we have @xmath457 by recalling the definition of @xmath454 from , for each fixed @xmath423 , we have @xmath458 and then @xmath459 which makes an estimate on the @xmath453-energy , with respect to the monge ampre gradient @xmath460 , of the mapping @xmath461 .
henceforth , fix @xmath0 with @xmath462 in @xmath2 and @xmath50 . given @xmath49 introduce @xmath463 and set @xmath464 observe that both @xmath401 and @xmath465 are strictly convex ( in the sense that their graphs do not contain line segments ) , continuously differentiable functions . from , for every @xmath466 we have @xmath467 by defining for @xmath128 as in , for @xmath468 , we have @xmath469 notice that , since @xmath470 , the function @xmath471 is a muckenhoupt @xmath472-weight on the real line for some @xmath473 if and only if @xmath474 and @xmath475 if and only if @xmath476 ( see example 9.1.7 on @xcite ) . hence , @xmath477 for every @xmath49 , which makes it a doubling weight on the real line ( equivalently , @xmath478 , since in dimension 1 the @xmath479 doubling property coincides with the usual doubling property ) whose doubling constant depends only on @xmath348 .
in particular , there exists @xmath480 , depending only on @xmath49 , such that @xmath481 for every @xmath482 .
now , since @xmath50 and @xmath478 , from ( * ? ? ?
* lemma 6 ) it follows that @xmath465 , being the tensor sum of @xmath5 and @xmath401 , satisfies @xmath483 with constants depending only on the @xmath479 constants for @xmath3 , dimension @xmath48 , and @xmath348 .
in addition , the condition @xmath484 is quantitatively equivalent to the existence of @xmath485 such that @xmath486 for every @xmath487 . by ( * ? ? ? *
lemma 6 ) the sections of @xmath128 are related to the ones of @xmath5 and @xmath401 by @xmath488 for every @xmath489 and @xmath412 .
we recall that constants depending only on the @xmath479 constants for @xmath3 in , @xmath49 and dimension @xmath48 will be called _
geometric constants_. by ( * ? ? ?
* corollary 3.3.2 ) , the condition @xmath484 implies the following doubling property for @xmath490 : there exists a geometric constant @xmath491 such that @xmath492 iterations of yield @xmath493 where @xmath494 .
also , by @xmath484 ( as well as the hypotheses @xmath89 and its strict convexity ) , there exists a geometric constant @xmath495 such that for every section @xmath496 we have @xmath497 see for instance ( * ? ? ?
* theorem 1 ) . from , for every @xmath466 @xmath499 where we have excluded the value @xmath500 to avoid the singularities of @xmath501 or @xmath502 . from and , for every @xmath466
the matrix of cofactors of @xmath503 equals @xmath504 important features are that @xmath505 and @xmath506 is differentiable for ( lebesgue ) a.e .
@xmath507 . notice that the first @xmath48 columns of @xmath508 are differentiable with respect to @xmath509 and the last column is differentiable with respect to @xmath125 . also , since the columns of @xmath510 are divergence free , so are the columns of @xmath508 .
if @xmath511 is differentiable at a point @xmath512 , the monge
ampre gradient of @xmath513 at @xmath514 is then given by @xmath515 which implies @xmath516 for @xmath465 as in , we will next prove the following counterpart to : for every section @xmath99 it holds true that @xmath519 notice that @xmath520 , so we can not directly apply with @xmath521 . instead
, we will use the tensorial nature of @xmath465 .
given a section @xmath522 , with @xmath523 , by means of and we can write @xmath524 now , from the first inclusion in and ( used with @xmath525 , since @xmath0 ) , @xmath526 where for the last two inequalities above we used the second inclusion in and the doubling property . on the other hand , by ,
@xmath527 let us write the one - dimensional section @xmath528 as @xmath529 , where @xmath530 satisfy @xmath531 from the first inclusion in we have @xmath532 as @xmath533 is increasing , @xmath534 at this point we split the last integral above as @xmath535 where the last equality is due to .
therefore , @xmath536 and follows .
our goal in this section is to prove a version of the poincar inequality with lebesgue measure being replaced by the monge
ampre measure @xmath490
. we will reason along the lines of ( * ? ? ?
* section 4 ) , where the change in the opposite direction ( i.e. from monge ampre to lebesgue measure ) was made by means of convex conjugation . in the case of @xmath128 , however , an approximation argument will be used to circumvent the fact that @xmath537 .
[ thm : poincare : phi ] let @xmath128 be as in . then there exist geometric constants @xmath538 and @xmath539 , such that for every section @xmath540 and @xmath541 with @xmath542 we have @xmath543 where @xmath544
let @xmath545 be nonnegative and compactly supported in @xmath546 $ ] with @xmath547 . for @xmath548 set @xmath549 and define @xmath550 and @xmath551 then , for every @xmath548 , we have that @xmath552 with @xmath553 . in addition , @xmath554 converges to @xmath555 in @xmath556 and , consequently , @xmath557 converges to @xmath490 in @xmath558 as @xmath559 tends to @xmath560 .
also , @xmath561 converges to @xmath562 uniformly on compact sets of @xmath563 .
the matrix of cofactors of @xmath564 is given by @xmath565 now , a simple computation shows that , for each @xmath548 , the measure @xmath554 is a doubling measure on the real line with doubling constant smaller than or equal to the doubling constant for @xmath555 , which , in turn , depends only on @xmath348 .
indeed , let @xmath566 denote the doubling constant for @xmath555 as measure on the real line . given @xmath567 and @xmath568 ,
let @xmath569 $ ] and @xmath570 $ ] .
then , @xmath571 for each @xmath572 , by changing variables @xmath573 we get @xmath574 if and only if @xmath575 .
hence , using that @xmath555 is even and doubling with constant @xmath576 , @xmath577 thus , @xmath578 uniformly in @xmath548 . by (
* lemma 6 ) , it follows that the @xmath479 constants for @xmath579 are controlled by the one for @xmath128 uniformly in @xmath548 .
next , let @xmath580 denote the convex conjugate of @xmath579 .
by we have @xmath581 with @xmath582 for every @xmath548 .
fix @xmath583 , @xmath412 , and set @xmath584 .
let @xmath585 be the geometric constants from applied to @xmath579 and define @xmath586 and @xmath587 .
hence , the inclusions yield @xmath588 assume first that @xmath589 and define @xmath590 as @xmath591 which , by putting @xmath592 , yields @xmath593 now , by the poincar inequality applied to @xmath580 on the section @xmath594 and @xmath590 , we get @xmath595 with @xmath596 a geometric constant , and by changing variables @xmath597 in and using that the identity gives @xmath598 and @xmath599 we obtain @xmath600 where we have put @xmath601 and used that @xmath602 , due to the definition of @xmath603 in .
next , the inclusions and the doubling property give @xmath604 consequently , the inclusions , the inequality , and the fact that @xmath605 converges to @xmath606 in @xmath607 , imply with @xmath608 in the case @xmath589 . then , in the case of @xmath542 follows by approximation in @xmath609 .
just notice that smooth functions are dense in @xmath610 for every open , bounded , convex subset @xmath611 and every @xmath612 and that on the compact set @xmath613 we have @xmath614 .
next we record a simple consequence of the weak poincar s inequality known as `` fabes lemma '' .
[ cor : fabes ] let @xmath128 be as in and let @xmath615 and @xmath616 be the geometric constants from theorem [ thm : poincare : phi ] . then ,
for every @xmath131 , every section @xmath617 and every @xmath541 with @xmath542 and @xmath618 we have that @xmath619 setting @xmath620 and @xmath621 , for @xmath622 we have @xmath623 so that @xmath624 and then follows from the weak - poincar s inequality .
from here on we fix an arbitrary @xmath625 and @xmath626 and define the cylinder @xmath627 also , let us put @xmath628 and define @xmath629 henceforth , all of our sub- and super - solutions will belong to a class @xmath93 modeled after some key properties of @xmath630 where @xmath81 , as in , is the solution to the extension problem in @xmath53 for a nonnegative @xmath631 .
we write @xmath632 if the following conditions hold .
a. [ z : symmetry ] @xmath633 for every @xmath634 ; b. [ c : c2 ] @xmath635 , in particular , @xmath636 , and then @xmath637 , exist a.e . in @xmath638 ; c. [ finite : l2:e ] for every section @xmath99 with @xmath639 we have @xmath640 d. [ normal : derivative ] for every @xmath641 the limit @xmath642 exists and the function @xmath643 is continuous in @xmath644 . notice that if the limit in exists , then we have @xmath645 given @xmath632 and an open subset @xmath647 , define @xmath648 the continuity of the function @xmath643 in @xmath644 ensures that @xmath649 for every open subset @xmath647 .
let @xmath81 be the solution to the extension problem in the section @xmath644 and set @xmath630 .
then @xmath94 satisfies that @xmath651 , @xmath94 is even with respect to @xmath125 , and for every @xmath641 , @xmath652 with @xmath653 for every @xmath369 . also , given @xmath654 and @xmath412 such that @xmath639 , the inclusions yield @xmath655 now , by recalling , to get an expression for @xmath656 , and by the energy estimate from remark [ rmk : finite : e ] , we obtain @xmath657 hence , @xmath650 .
notice that @xmath93 is a vector space that contains the constant functions .
also , no vanishing condition on @xmath658 is prescribed in the definition of @xmath632 , although this is the case for @xmath659
for @xmath660 introduce @xmath661 where @xmath480 is the quasi - triangle constant in , which depends only on @xmath348 . the main result in this section
is the following critical - density estimate .
[ thm : cd ] let @xmath128 be as in . there exist geometric constants @xmath662 such that for every section @xmath663 with @xmath664 and every @xmath96 with @xmath665 pointwise in @xmath98 , the inequalities @xmath666 and @xmath667 imply @xmath668 as mentioned in the introduction , due to the degeneracy of @xmath669 on the hyperplane @xmath670 , one can not directly resort to the proof of theorem 1 in @xcite or the one for theorem 2 in @xcite . in order to address the degeneracy or singularity of @xmath669 on the hyperplane @xmath670 , the proof of theorem [ thm : cd ] will be broken down into three cases according to the position of the sections of @xmath128 with respect to @xmath670 .
these three cases are illustrated in figure [ fig : cdsections ] . in the case of sections intersecting @xmath670
, the @xmath103-norm of normal derivatives , in the sense of , of super - solutions will play a key role in addressing the degeneracy or singularity of @xmath114 .
( within @xmath638 ) to be considered in the proof of theorem [ thm : cd ] . when the section intersects @xmath670 , the control on the size of the normal derivative of a solution , as defined by , will counteract the degeneracy or singularity of @xmath671 at @xmath670.,width=528 ] let us start by stating a version of the aleksandrov bakelman
pucci maximum principle on which the cases are built .
recall that given a domain @xmath672 and @xmath673 , the normal mapping of @xmath292 at @xmath674 , denoted by @xmath675 , is defined as the set @xmath676 and , given @xmath677 , @xmath678 also , let @xmath679 denote the convex envelope of @xmath292 in @xmath680 and let @xmath681 denote the contact set of @xmath292 and @xmath682 in @xmath680 .
in particular ( see @xcite ) , @xmath683 the next lemma allows to replace @xmath684 in ( * ? ? ?
* theorem 1.4.5 ) with @xmath685 when @xmath680 is bounded and convex ( its proof is a combination of the one for theorem 1.4.5 from @xcite and the normalization technique in ( * ? ? ?
* section 1 ) ) .
[ lem : abp1 ] let @xmath672 be open , convex , bounded and let @xmath686 satisfy @xmath687 on @xmath688 .
let @xmath689 and @xmath690 denote the convex envelope of @xmath691 in @xmath680 and the contact set of @xmath691 with @xmath689 in @xmath680 , respectively .
then , there exists a dimensional constant @xmath692 such that @xmath693 we will be using the following consequence of lemma [ lem : abp1 ] . [ cor : abp2 ] let @xmath672 be open , convex , and bounded .
suppose that @xmath694 satisfies the following conditions a. @xmath695 on @xmath688 , b. there is an open set @xmath696 such that @xmath697 and @xmath698 , where @xmath699 .
then , @xmath700 lemma [ lem : abp1 ] applied to @xmath701 yields @xmath702 by with @xmath703 , @xmath704 now , for @xmath705 the fact that @xmath706 on @xmath688 yields @xmath707 ( with @xmath708 unless @xmath709 ) .
let us see that this implies the inclusion @xmath710 indeed , given @xmath711 and @xmath712 the definition of normal mapping gives @xmath713 and since @xmath714 and @xmath715 for every @xmath716 , the inequality implies that @xmath717 , thus proving . by combining and , we have @xmath718 so that , from , @xmath719 finally , by the assumptions @xmath698 and @xmath697 , the inequality @xmath720 follows from the usual formula of change of variables and the proof is complete .
[ thm : cd : case1 ] let @xmath128 be as in .
let @xmath721 and @xmath412 such that @xmath722 .
set @xmath723 , @xmath724 and suppose that @xmath725
. then , there exists a geometric constant @xmath726 such that for every @xmath96 with @xmath727 and @xmath728 in @xmath729 , the inequality @xmath730 implies @xmath731 introduce the auxiliary function @xmath732 where the function @xmath733 is convex with @xmath734 . by
, there is @xmath735 such that @xmath736 , and then @xmath737 now , the abp maximum principle in corollary [ cor : abp2 ] applied to the function @xmath701 on the convex set @xmath738 ( notice that @xmath739 on @xmath740 ) yields @xmath741 in particular , on the contact set @xmath742 we have that @xmath513 is negative , and then @xmath743 also , @xmath744 on @xmath742 so that , recalling that @xmath745 is positive in @xmath746 , on @xmath742 we have @xmath747^{n+1 } ( \det d^2 { \phi})^{-n}\\ & = \left[\frac { { \operatorname{trace}}(a_{\phi}d^2 w ) + \frac{2}{r } { \operatorname{trace}}(a_{\phi}d^2{\phi } ) } { n+1 } \right]^{n+1 } ( \det d^2 { \phi})^{-n}\\ & \leq \left(\frac { 2 \det d^2 { \phi}}{r } \right)^{n+1 } ( \det d^2 { \phi})^{-n } = \left(\frac { 2 } { r } \right)^{n+1 } \det d^2 { \phi}.\end{aligned}\ ] ] therefore , @xmath748 then , by recalling the definition of @xmath749 in , @xmath750 and follows with @xmath751 . for each @xmath660 , set @xmath752 [ thm : cd : center : z0 ] let @xmath128 be as in .
let @xmath25 and @xmath412 such that @xmath753 .
put @xmath754 and @xmath755 .
suppose that @xmath96 satisfies @xmath756 and @xmath728 pointwise in @xmath757 .
then , there exist geometric constants @xmath758 such that the inequalities @xmath759 and @xmath760 imply @xmath761 notice that the expression for @xmath518 in and the fact that @xmath762 give @xmath763 which makes all sections centered at @xmath670 symmetric with respect with @xmath125 .
let @xmath764 for @xmath765 , introduce the auxiliary function @xmath766 as @xmath767 which makes @xmath513 symmetric in @xmath125 as well ( that is , @xmath768 for every @xmath769 .
now , by the inclusions , @xmath770 so that @xmath765 implies @xmath771 and then @xmath772 ( because @xmath762 ) .
consequently , from the definition of @xmath773 , @xmath765 implies @xmath774 ( where it holds that @xmath775 ) and using that @xmath728 and @xmath776 we have @xmath777 on the other hand , for every @xmath778 ( and using @xmath779 again ) , @xmath780 as a consequence , the convex envelope of @xmath781 in @xmath746 can not touch @xmath781 on @xmath670 ; moreover , the contact set @xmath742 lies at a positive distance from @xmath670 .
therefore , @xmath782 . by corollary [ cor : abp2 ]
applied to @xmath513 with @xmath783 and @xmath784 ( notice that @xmath785 ) we obtain @xmath786 for @xmath514 in the contact set @xmath782 we have @xmath787 in @xmath757 we have @xmath788 and @xmath789 .
so that for @xmath790 , @xmath791 which combined with gives @xmath792 now , by , there exists @xmath793 such that @xmath794 and then @xmath795 where we have used that , from , @xmath796 and @xmath797 by the definition of @xmath798 in we get @xmath799 and then @xmath800 .
then , from , the fact that @xmath800 and , we deduce @xmath801 next , on the contact set we have @xmath802 , so that by using we get ( recalling ) @xmath803 on the other hand , @xmath804 so that implies @xmath805 and by choosing @xmath806 such that @xmath807 , the inequality follows with @xmath808 when a section @xmath99 , not necessarily centered somewhere at @xmath809 , satisfies @xmath810 , the first step will be to relate it to a section centered at @xmath809 of comparable height .
more precisely , we have [ lem : betas ] given a section @xmath99 centered at @xmath811 with @xmath810 , there exists @xmath812 such that @xmath813 where @xmath480 , depending only on @xmath348 , is the quasi - triangle constant for @xmath401 in and @xmath814 , also depending only on @xmath348 , is as in .
given @xmath99 with @xmath810 , the first inclusion in gives @xmath815 the fact that @xmath810 implies that the closed interval @xmath816 contains @xmath560 , that is , @xmath817 with @xmath818 . without loss of generality ,
let us assume that @xmath819 . by putting @xmath820
we have @xmath821 ( see figure [ fig : sectionshs ] ) and then @xmath822 .,width=432 ] also , from the quasi - triangle inequality for @xmath401 in along with @xmath823 and @xmath824 , we get @xmath825 and , using that @xmath826 , @xmath827 next , we claim that @xmath828 indeed , given @xmath829 and using the triangle inequality along with @xmath824 and , @xmath830 which proves . consequently , @xmath831 and the lemma is proved .
[ thm : cd : case3 ] let @xmath128 be as in and @xmath814 be as in ( which is the same constant as in lemma [ lem : betas ] ) .
let @xmath832 and @xmath412 such that @xmath833 .
put @xmath663 , @xmath834 , and suppose that @xmath835 . then , there exist geometric constants @xmath836 such that for every @xmath96 with @xmath837 and @xmath728 in @xmath838 the inequalities @xmath839 and @xmath840 imply @xmath841 let us take @xmath842 with @xmath806 as in theorem [ thm : cd : center : z0 ] . by lemma [ lem : betas ] used with @xmath843 and setting @xmath844 ,
we have @xmath845 since , by definition , @xmath846 . in particular , by hypothesis and the definition of @xmath847 , @xmath848 now , inequality and the first inclusion in imply that @xmath849 , which combined with allows us to use theorem [ thm : cd : center : z0 ] with the sections @xmath850 and @xmath851 to obtain @xmath852 by the doubling property and the first inclusion in we get @xmath853 and then follows , with @xmath854 , from , , and the last inclusion in .
the proof follows , with the geometric constants @xmath855 and @xmath856 , from theorems [ thm : cd : case1 ] and [ thm : cd : case3 ] , and the fact that @xmath857 and @xmath858 .
let us define @xmath859 where @xmath860 is as in and @xmath814 is as in . with the critical - density estimate from theorem [ thm : cd ] at hand
, we can deduce the following local - boundedness results .
[ lem : lemma6 ] let @xmath128 be as in . there exist geometric constants @xmath861 such that for every @xmath832 and @xmath412 with @xmath862 and every @xmath96 with @xmath863 and @xmath728 pointwise in @xmath98 the inequalities @xmath864 imply @xmath865 [ thm : lb : q=1 ] let @xmath128 be as in .
there exist geometric constants @xmath866 such that for every @xmath832 and @xmath412 with @xmath862 and every @xmath96 with @xmath863 and @xmath728 pointwise in @xmath98 we have @xmath867 [ thm : lb ] let @xmath128 be as in .
then there are geometric constants @xmath51 and @xmath868 such that for every @xmath869 there exist constants @xmath870 , depending only on geometric constants and @xmath871 , such that for every section @xmath872 and every @xmath96 with @xmath863 and @xmath728 pointwise in @xmath98 , we have @xmath873 the proof of lemma [ lem : lemma6 ] follows as the proof of lemma 6 in @xcite .
essentially , the only modification is to have the expression @xmath874 ( in the proof of ( * ? ? ?
* lemma 6 ) ) replaced with @xmath875 ( and , of course , replacing the function @xmath5 with @xmath128 ) . a key point in the proof
is that the sub - solution be locally bounded , which is the case in lemma [ lem : lemma6 ] since @xmath96 .
similarly , theorem [ thm : lb : q=1 ] follows from lemma [ lem : lemma6 ] as in the proof of theorem 3 on @xcite and theorem [ thm : lb ] follows from theorem [ thm : lb : q=1 ] as in the proof of theorem 7 on @xcite .
by relying on the divergence - form side of the linearized monge ampre operator , in this section we extend theorem [ thm : cd ] by proving that every @xmath131 can work as a critical - density parameter ( see theorem [ thm : every : cd ] and corollary [ cor : every : cd ] below ) .
statements are referred to lebesgue measure , which is equivalent to a.e . with respect to @xmath490 due the hypothesis @xmath0 with @xmath462 in @xmath2 and
the fact that @xmath876 .
[ lem : energy : log : h ] let @xmath128 be as in . fix a section @xmath877 and suppose that @xmath878 satisfies the following conditions : a. @xmath879 exists a.e . in @xmath877 , b. @xmath880 a.e . in @xmath877 , c. there exists @xmath881 such that @xmath882 for every @xmath883 , d. @xmath884 , that is , @xmath885 then , @xmath886 where @xmath491 is the doubling constant from .
multiply the inequality @xmath887 by @xmath490 to obtain , a.e . in @xmath877 , @xmath888 for @xmath889 ,
supported in @xmath890 $ ] , with @xmath891 on @xmath892 $ ] and @xmath893 define @xmath894 now multiply by @xmath895 and integrate by parts on @xmath877 to obtain @xmath896 from young s inequality , @xmath897 hence , @xmath898 where we have used that @xmath899 , @xmath900 , and to guarantee that @xmath901 on the other hand , since @xmath902 we get @xmath903 where for the last inequality we used the energy estimate . therefore , @xmath904 and follows from the doubling property .
[ thm : every : cd ] let @xmath128 be as in . fix @xmath96 with @xmath905 and @xmath665 in @xmath98 . fix @xmath832 and @xmath906 with @xmath872 and put @xmath907 . then , given @xmath908 , the inequalities @xmath909 and @xmath910 imply that @xmath911 where @xmath51 and @xmath912 are the geometric constants from theorem [ thm : lb ] and @xmath913 where @xmath914 are the geometric constants from the weak poincar inequality and @xmath915 are the geometric constants from theorem [ thm : lb ] corresponding to @xmath916 . for @xmath917 define the function @xmath918 as @xmath919 which is a convex , lipschitz function with @xmath920 and @xmath921 , and put @xmath922 let @xmath923 be a smooth approximation of @xmath924 such that @xmath925 and introduce @xmath926
let us see that @xmath927 for every @xmath928 by checking the conditions - from the definition of @xmath93 in section [ sec : s(q ) ] .
conditions and are immediate .
condition follows from @xmath929 and the fact that @xmath930 , uniformly in @xmath928 .
similarly , condition , follows from @xmath931 consequently , @xmath927 for every @xmath928 . on the other hand , @xmath932 and then , always in @xmath98 , @xmath933 that is , @xmath927 satisfies @xmath934 and @xmath935 in @xmath98 . by theorem [ thm : lb ]
applied to @xmath936 with @xmath916 we have @xmath937 and , by and , @xmath938 and then , by taking limits as @xmath939 in , it follows that @xmath940 now , since @xmath941 the hypothesis says that @xmath942 and , by corollary [ cor : fabes ] , @xmath943 at this point we use lemma [ lem : energy : log : h ] with @xmath944 in the section @xmath945 to get @xmath946 and then @xmath947 by the definition of @xmath948 in , we have @xmath949 and follows from and .
[ cor : every : cd ] let @xmath128 be as in . fix @xmath96 with @xmath905 and @xmath665 in @xmath98 . fix @xmath832 and @xmath906 with @xmath872 and put @xmath907 . then , for every @xmath131 there exists @xmath917 , depending only on @xmath559 and geometric constants , such that the inequalities @xmath950 and @xmath951 imply that @xmath952 where @xmath51 is the geometric constant from theorem [ thm : lb ] .
given @xmath131 let @xmath917 be defined by @xmath953 and apply theorem [ thm : every : cd ] to @xmath954 .
in this section we prove a weak - harnack inequality for nonnegative super - solutions in @xmath93 ( see theorem [ thm : weak : h ] ) .
the proof relies on establishing a `` non - homogeneous '' version of theorem 7.1 in @xcite . towards that end , we next adapt a result known as the `` crawling ink spots lemma '' to the elliptic monge ampre context .
the `` crawling ink spots lemma '' has been developed by krylov - safonov in ( * ? ? ?
* section 2 ) ; the lemmas in ( * ? ? ?
* section 2 ) correspond to the parabolic case ; see ( * ? ? ?
* lemma 1.1 ) for the elliptic case in @xmath2 , and ( * ? ? ?
* lemma 7.2 ) , for instance , for a version in doubling metric spaces ) .
[ lem : cis ] let @xmath128 be as in . fix any @xmath955 , where @xmath485 is the quasi - triangle constant from . given a section @xmath956 , a measurable subset @xmath957 , and
@xmath958 define the open set @xmath959 then either @xmath960 or @xmath961 where @xmath962 is as in .
the proof follows as the one for ( * ? ? ?
* lemma 7.2 ) by means of vitali s covering lemma for monge ampre sections . in turn , vitali s covering lemma for monge ampre sections follows as in the proof of theorem 1.2 in @xcite for general spaces of homogeneous type . in the monge
ampre case , the dilation constant in vitali s lemma can be any @xmath963 with @xmath964 . [ thm : weak : h ] let @xmath128 be as in .
there exist geometric constants @xmath965 and @xmath966 such that for every @xmath96 with @xmath665 and @xmath905 in @xmath98 and every @xmath967 with @xmath968 , we have @xmath969 where @xmath51 is the geometric constant from theorem [ thm : lb ] .
let us start by defining @xmath970 as @xmath971 with @xmath868 being the geometric constant from theorem [ thm : lb ] .
let us fix @xmath958 such that @xmath972 where @xmath973 is the product of geometric constants from in lemma [ lem : cis ] , and for @xmath974 as in choose @xmath131 such that @xmath975 with this choice of @xmath559 , and recalling the definition of @xmath976 in , define @xmath917 by means of and put @xmath977 , which makes @xmath978 , @xmath974 , and @xmath979 geometric constants . for @xmath980 and @xmath981 set @xmath982 and let @xmath983 satisfy @xmath984 which yields @xmath985 where @xmath986 is a geometric constant .
we will show that @xmath987 where @xmath988 is the geometric constant defined as @xmath989 indeed , given @xmath990 we consider two cases : when @xmath991 and when @xmath992 if holds true for some @xmath990 , then is immediate from the definition of @xmath993 in and the fact that @xmath994 and @xmath990 imply @xmath995 .
suppose then that holds true for every @xmath990 .
we will prove by repeatedly applying corollary [ cor : every : cd ] and lemma [ lem : cis ] .
if for some @xmath996 and @xmath997 we have @xmath998 ( with @xmath999 as in ) by the doubling property ( and recalling that @xmath51 ) it follows that @xmath1000 where for the last equality we used the definition of @xmath559 in .
next , let us see that @xmath1001 and @xmath996 imply the inclusion @xmath1002 indeed , given @xmath1003 , by the @xmath860-quasi - triangle inequality @xmath1004 where for the last equality we used the definition of @xmath1005 in .
now , from the fact that @xmath1006 , the inclusion , and the hypothesis , we get @xmath1007 then , corollary [ cor : every : cd ] applied to @xmath1008 on the section @xmath1009 with @xmath559 as in yields @xmath1010 and , consequently , @xmath1011 . by lemma [ lem : cis ] applied to the section @xmath1012 and the set @xmath1013 it follows that either @xmath1014 or @xmath1015 where @xmath1016 is as in .
now , if @xmath1014 for some @xmath990 , then ( due to the inclusion @xmath1017 ) we have @xmath1018 , which means @xmath1019 and the inequality follows .
hence , we can assume that holds for every @xmath1020 and then write @xmath1021 where for the last inequality we used the definition of @xmath1022 in .
in particular , @xmath1023 and , using that @xmath1024 and @xmath1025 , by applied with @xmath1026 , we obtain @xmath1027 so that corollary [ cor : every : cd ] applied to @xmath1028 on the section @xmath1012 with @xmath559 as in yields @xmath1029 and follows .
now , by setting @xmath1030 , from and we obtain @xmath1031 and then , for any @xmath1032 , @xmath1033 which proves with @xmath1034 .
we are now position to prove theorems [ h : nondiver ] and [ thm : h : vtilde ] .
let us start with by theorem [ thm : weak : h ] applied to @xmath96 , we get @xmath1035 now , by theorem [ thm : lb ] applied to @xmath104 with @xmath1036 , @xmath1037 and notice that , from , we have @xmath1038 .
hence , the harnack inequality follows with @xmath1039 .
next , for @xmath1040 , consider the functions @xmath1041 a standard argument ( see for instance ( * ? ? ?
* section 8.9 ) ) implies the existence of geometric constants @xmath59 and @xmath106 such that @xmath1042 which , in turn , implies .
given a section @xmath53 , @xmath54 , and a nonnegative @xmath78 solution to @xmath199 in @xmath1043 , let @xmath81 be the solution to the extension problem . in particular , we have @xmath1044 and @xmath1045 uniformly in @xmath1043 .
now , let us set @xmath1046 so that given a section @xmath133 with @xmath57 , by the first inclusion in we get @xmath1047 and then the harnack inequality for @xmath94 on the section @xmath1048 gives @xmath1049 on the other hand , the second inclusion in yields @xmath1050 and , since @xmath1051 for every @xmath1052 , it follows that @xmath1053 as well as @xmath1054 which , along with implies @xmath1055 but , by the inclusion @xmath1056 , we have @xmath1057 so that from the definition of @xmath1058 in and the limit in , we get @xmath1059 then , the harnack inequality follows from and with @xmath1046 and @xmath1060 . also by a restriction argument , the monge
ampre hlder estimate follows from applied to @xmath94 and then restricting to @xmath670 .
notice that in principle we can not prove directly from the harnack inequality , as we did to go from to via .
this is due to the fact that if @xmath369 and @xmath1061 , then it is not true that @xmath1062 because @xmath1063 does not vanish on @xmath201 . | the fractional nonlocal linearized monge
ampre equation is introduced . a harnack inequality for nonnegative solutions to the poisson problem on monge
ampre sections is proved . |
the abundance of the super volatile molecule n@xmath1 in cometary nuclei is still under debate . searches for uv fluorescence from n@xmath1 with the far ultraviolet spectroscopic explorer ( fuse ) have been unsuccessful ( bockele - morvan et al . 2004 ) .
measurements of the n@xmath1 abundance in 1p / halley by giotto were not feasible : owing to the low resolution of the mass spectrometer , it was not possible to discriminate n@xmath1 from co ( eberhardt et al .
ground - based observation of the n@xmath2 band at 3914 is a difficult challenge due both to the contamination of the @xmath3 ( 0,2,0)-(0,0,0 ) band and the presence of telluric n@xmath2 emission lines .
different positive detections of this feature have been claimed in the ionic tail of comets humason ( 1962 viii ) ( greenstein & arpigny 1962 ) , 1p / halley ( lutz et al .
1993 ; wyckoff & theobald 1989 ) , c/1987 p1 bradfield ( lutz et al . 1993 ) , and c/2002 vq94 ( linear ) ( korsun et al . 2006 ) .
arpigny , from photographic spectra and other archival data estimated a positive n@xmath2/co@xmath4 intensity ratio for 12 comets ( cochran et al .
all these positive detections are based on low - resolution spectra .
so far similar searches with high - resolution spectra have been unsuccessful .
this was the case for observations of comets 122p / de vico , c/1995 o1 ( hale - bopp ) , and 153p/2002 c1 ( ikeya - zhang ) that did not detect n@xmath2 band , yielding upper limits of @xmath5 10@xmath610@xmath7 on the abundance of n@xmath1 relative to co ( cochran et al . 2000 ; cochran 2002 ) .
an interpretation of the possible n@xmath1 deficiency in comets has been proposed by iro et al .
these authors assumed that comets were made of clathrate hydrates ( hereafter clathrates ) and argued that co was preferentially incorporated in clathrates compared to n@xmath1 after having investigated the competition between the trapping of these two molecules in clathrates .
in order to explain the n@xmath1 deficiency in comets , they argued that the nebula s temperature never cooled down below @xmath045 k in their formation region , impeding the formation of n@xmath1-bearing ice ( in clathrate or pure condensate form ) that would have been agglomerated by comets .
however , this mechanism is not consistent with the fact that pluto and triton * possess thick nitrogen ice covers * ( lellouch et al .
2011 ) while they are both expected to have been formed in the same region of the primitive nebula as ecliptic comets ( kavelaars et al .
in addition , iro et al .
( 2003 ) did not consider the competition between the trapping of n@xmath1 in clathrates with that of the other abundant volatiles present in the nebula ( co , co@xmath1 , ch@xmath8 , h@xmath1s , ... ) . indeed , due to their various propensities for trapping , the consideration of a larger set of molecules can drastically change the calculated fraction of n@xmath1 incorporated in clathrates formed in the nebula .
here we propose a scenario that could explain the apparent n@xmath1 deficiency in comets in a way consistent with the presence of this molecule * on the surfaces of pluto and triton*. we use a statistical thermodynamic model to investigate the composition of the successive multiple guest clathrates ( hereafter mg clathrates ) that may have formed during the cooling of the primordial nebula from the most abundant volatiles present in the gas phase ( see section 2 ) .
these clathrates agglomerated with the other ices and formed the building blocks of comets ( hereafter cometesimals )
. the major ingredient of our model is the description of the guest - clathrate interaction by a spherically averaged kihara potential with a nominal set of potential parameters .
our model allows us to report that n@xmath1 is a poor clathrate former when considering a plausible gas phase composition of the primordial nebula , implying that its trapping into cometesimals requires a low disk s temperature ( @xmath020 k ) in order to allow the formation of its pure condensate .
we find that it is possible to explain the lack of nitrogen in comets as a consequence of their postaccretion internal heating engendered by the decay of radiogenic nuclides ( see section 3 ) .
in our model , the volatile phase incorporated in cometesimals is composed of a * mixture * of pure ices , stoichiometric hydrates ( such as * nh@xmath9-h@xmath1o hydrate * ) and mg clathrates that crystallized in the form of microscopic grains at various temperatures in the outer part of the disk . our model is based on the assumption that cometesimals have grown from the agglomeration of these icy grains due to collisional coagulation , implying no loss of volatile during their growth phase ( weidenschilling 1997 ) . here
, the clathration process stops when no more crystalline water ice is available to trap the volatile species and then only pure condensates form at lower temperature in the disk .
the process of volatiles trapping in icy grains is calculated using the equilibrium curves of hydrates and pure condensates , our model determining the equilibrium curves and compositions of mg clathrates , and the thermodynamic paths detailing the evolution of temperature and pressure between 5 and 30 au in the protoplanetary disk .
we refer to the work of alibert et al .
( 2005 ) for a full description of the turbulent model of accretion disk used here .
the composition of the initial gas phase of the disk is defined as follows : we assume that the abundances of all elements ( c , n , o , s , p , ar , kr and xe ) are protosolar ( asplund et al . 2009 ) and that o , c , and n exist only under the form of h@xmath1o , co , co@xmath1 , ch@xmath9oh , ch@xmath8 , n@xmath1 , and nh@xmath9 .
the abundances of co , co@xmath1 , ch@xmath9oh , ch@xmath8 , n@xmath1 and nh@xmath9 are then determined from the adopted co : co@xmath1:ch@xmath9oh : ch@xmath8 and n@xmath1:nh@xmath9 gas phase molecular ratios .
once the abundances of these molecules are fixed , the remaining o gives the abundance of h@xmath1o .
concerning the distribution of elements in the main volatile molecules , we set co : co@xmath1:ch@xmath9oh : ch@xmath8 = 70:10:2:1 in the gas phase of the disk , values that are consistent with the ism measurements considering the contributions of both gas and solid phases in the lines of sight ( mousis et al .
in addition , s is assumed to exist in the form of h@xmath1s , with h@xmath1s : h@xmath1 = 0.5 @xmath10(s / h@xmath1)@xmath11 , and other refractory sulfide components .
we also consider n@xmath1:nh@xmath9 = 1:1 in the nebula gas - phase , a value compatible with thermochemical calculations in the solar nebula that take into account catalytic effects of fe grains on the kinetics of n@xmath1 to nh@xmath9 conversion ( fegley 2000 ) .
note that ch@xmath9oh is considered to be formed only as a pure condensate in our calculations since , to the best of our knowledge , no experimental data concerning the equilibrium curve of its associated clathrate have been reported in the literature .
the equilibrium pressure for the successive mg clathrates formed in the solar nebula can be expressed as ( hand et al .
2006 ) : @xmath12^{-1}\ ] ] where @xmath13 is the mole fraction of the component @xmath14 in the fluid phase .
the equilibrium pressure curves of each species are determined by fitting the available theoretical and laboratory data ( lunine & stevenson 1985 ) .
their equations are of the form @xmath15 , where @xmath16 and @xmath17 are the partial equilibrium pressure ( bars ) and temperature ( k ) of the considered species @xmath14 , respectively .
the relative abundances of guest species incorporated in mg clathrate formed at a given temperature and pressure from the solar nebula gas phase are calculated following the method described by lunine & stevenson ( 1985 ) and mousis et al . (
2010 ) , which uses classical statistical mechanics to relate the macroscopic thermodynamic properties of clathrates to the molecular structure and interaction energies .
it is based on the original ideas of van der waals & platteeuw ( 1959 ) for clathrate formation , which assume that trapping of guest molecules into cages corresponds to the three - dimensional generalization of ideal localized adsorption .
this approach is based on four key assumptions ( lunine & stevenson 1985 ; sloan 1998 ) : 1 .
the host molecules contribution to the free energy is independent of the clathrate occupancy .
this assumption implies in particular that the guest species do not distort the cages . 2 .
\(a ) the cages are singly occupied .
( b ) guest molecules rotate freely within the cage .
3 . guest molecules do not interact with each other .
classical statistics is valid , i.e. , quantum effects are negligible . in this formalism
, the fractional occupancy of a guest molecule @xmath18 for a given type @xmath19 ( @xmath19 = small or large ) of cage can be written as @xmath20 where the sum in the denominator includes all the species which are present in the initial gas phase .
@xmath21 is the langmuir constant of species @xmath18 in the cage of type @xmath19 , and @xmath22 is the partial pressure of species @xmath18 .
this partial pressure is given by @xmath23 ( we assume that the sample behaves as an ideal gas ) , with @xmath24 the mole fraction of species @xmath18 in the gas phase , and @xmath25 the total gas pressure , which is dominated by h@xmath1 .
the langmuir constant depends on the strength of the interaction between each guest species and each type of cage , and can be determined by integrating the molecular potential within the cavity as @xmath26 where @xmath27 represents the radius of the cavity assumed to be spherical , @xmath28 the boltzmann constant , and @xmath29 is the spherically averaged kihara potential representing the interactions between the guest molecules @xmath18 and the h@xmath1o molecules forming the surrounding cage @xmath19 . for a spherical guest molecule , this potential @xmath30 can be written as ( mckoy & sinanolu 1963 ) : @xmath31 , \ ] ] with @xmath32.\ ] ] in eq .
( [ pot_kihara ] ) , @xmath33 is the coordination number of the cell .
this parameter depends on the structure of the clathrate ( i or ii ) and on the type of the cage ( small or large ) .
the kihara parameters @xmath34 , @xmath35 and @xmath36 for the molecule - water interactions employed in this work are listed in table [ kihara ] .
note that our parameters are different from those used by iro et al .
( 2003 ) , who instead of using potential parameters describing the guest - clathrate interaction in their statistical model , employed potential parameters corresponding to guest guest interactions ( case of pure solutions ) . when comparing different data sets determined for the same x - h@xmath1o interaction ( with x = ch@xmath8 , n@xmath1 and xe )
, we opted for the most recent ones because they are fitted to a larger range of experimental data . in the case of ar ,
kr , co and ph@xmath9 , the listed sets are the unique ones that we found in the literature . for the h@xmath1s - h@xmath1o and co@xmath1-h@xmath1o interaction parameters we used those of parrish and prausnitz ( 1972 ) and thomas
( 2009 ) ( the set of co@xmath1-h@xmath1o interaction parameters provided by thomas et al .
( 2009 ) is an update of the one determined by parrish and prausnitz ( 1972 ) and very close to the one recently derived by herri and chassefire ( 2012 ) ) .
we also investigated the interaction parameters of sloan and koh ( 2008 ) but these parameters yield unphysical results when used in our models .
we suspect this is because the code of sloan and koh ( 2008 ) is optimized for industrial purposes and the parameters are not appropriate for our temperature - pressure regime .
we have tested our code against the one of herri and chassefire ( 2012 ) so we do not believe the problem resides in our code .
finally , the mole fraction @xmath37 of a guest molecule @xmath18 in a clathrate can be calculated with respect to the whole set of species considered in the system as @xmath38 where @xmath39 and @xmath40 are the number of small and large cages per unit cell respectively , for the clathrate structure under consideration , and with @xmath41 @xmath42 .
values of @xmath33 , @xmath27 , @xmath39 and @xmath40 are taken from parrish & prausnitz ( 1972 ) .
each time a species is fully trapped in a mg clathrate , its gaseous abundance is set to zero in the disk .
this change of the disk s gas phase composition induces the formation of a new mg clathrate at different temperature and pressure conditions and with a distinct composition . as a consequence
, the compositions and formation conditions of successive mg clathrates are calculated @xmath43 times in the nebula , with @xmath43 corresponding to the total number of fully enclathrated species . once all the water budget has been used for clathration , the volatiles remaining in the gas phase form pure condensates at lower temperatures .
the equilibrium curves of these pure condensates derive from the compilation of laboratory data given in the crc handbook of chemistry and physics ( lide 2002 ) .
the composition of the volatile phase incorporated into cometesimals formed at a given distance from the sun is finally given by the intersection of the disk s thermodynamic path at this location with the equilibrium curves of the different formed ices .
the amount of volatile , @xmath14 ( relative to water ) , used to form the ice @xmath44 in the nebula ( either in the form of stoichiometric hydrate , mg clathrate or of pure condensate ) , can be determined by the relation : @xmath45 where @xmath46 is the mass mixing ratio with respect to h@xmath1 of the volatile @xmath14 used to form ice @xmath44 in the nebula .
@xmath47 is the mass mixing ratio of h@xmath1o with respect to h@xmath1 in the nebula . @xmath48 and @xmath49 are the surface density of the disk at distance @xmath50 from the sun at the epoch of formation of ice @xmath44 , and at the epoch of condensation of water , respectively .
the global volatile , @xmath14 , to water mass ratio incorporated in cometesimals is then given by @xmath51 , with @xmath52 corresponding to the number of solid phases in which volatile @xmath14 is incorporated .
note that a thermodynamic path has been arbitrarily selected at the heliocentric distance of 5 au to determine the composition of the formed ices .
the adoption of any other distance range between @xmath05 and 30 au in the nebula would not affect the composition of the ices because it remains almost identical irrespective of i ) their formation distance and ii ) the input parameters of the disk , provided that the initial gas phase composition is homogeneous ( marboeuf et al .
2008 ) .
figure [ mole ] represents the composition of the volatile phase incorporated in cometesimals agglomerated from ices condensed in the outer nebula and figure [ ratios ] displays the molar ratios between incorporated volatiles of interest .
both figures show calculations expressed as a function of the formation temperature of cometesimals in the primordial nebula .
observations of active comets when their activity is driven by the sublimation of water ice
i.e. for heliocentric distances smaller than about 2 au suggest that the mean co production rate relative to water is typically a few percents , with extreme values comprised between 0.4 and 30% ( bockele - morvan et al .
2004 ) . assuming that the production rates relative to water measured in comets are representative of the bulk composition of cometesimals , figure [ ratios ] shows that their formation temperature must be lower than @xmath047 k in the nebula to get co / h@xmath1o @xmath53 1% in their interior .
the figure also shows that a formation temperature lower than @xmath025k would increase the co / h@xmath1o ratio up to @xmath0@xmath54 in cometesimals , as a result of the incorporation of pure co condensate at this lower nebular temperature .
formation temperatures of ices lower than @xmath022 k in the nebula allow trapping of n@xmath1 with n@xmath1/h@xmath1o of order of 1% in the interior of planetesimals , a value far above the highest estimates ( @xmath00.02% ) inferred from observations ( wyckoff et al .
1991 ; cochran et al .
2000 ) . on the other hand , assuming a formation temperature of cometesimals in the 2247 k range in the nebula would allow n@xmath1/h@xmath1o @xmath00.01@xmath55 and co / h@xmath1o @xmath06.6@xmath55 in their interior , values that are consistent with observations . a simple explanation for the formation of cometesimals would then argue that they need to be agglomerated from grains formed in the 2247 k range in order to match the observations of the cometary production rates .
however , this mechanism is not consistent with the fact that pluto and triton possess * n@xmath1-rich ice covers * while they are expected to be formed in the same region as comets .
an alternative possibility would be that comets accreted from planetesimals formed at very low temperature in the disk , but that subsequent radiogenic heating enabled their partial devolatilization through their pores .
in the next section , we explore how radiogenic heating can induce important losses of n@xmath1 in comets after their formation .
we use a fully three dimensional model of heat transport described by @xcite to compute the time evolution of the temperature distribution within cometary nuclei , accounting for the heating due to the radioactive decay of short - lived nuclides @xmath56al and @xmath57fe .
we assume that these objects formed by cold accretion , i.e. no accretional heating is accounted for .
the model computes the temperature distribution and its evolution with time , by solving the heat diffusion equation : @xmath58 with @xmath59 [ kg m@xmath60 the bulk density ,
@xmath61 [ j kg@xmath62 k@xmath62 ] the heat capacity , @xmath64 [ w m@xmath62 k@xmath62 ] the thermal conductivity and @xmath65 [ w m@xmath60 the internal power production per unit volume due to the decay of short - lived radioactive nuclides . as an example , we show in this paper the results for an idealized hale - bopp " with a size of 35 km @xcite .
we consider this object as a sphere , made of a porous mixture of crystalline water ice and dust , homogeneously distributed within the ice matrix .
we use generic parameters to describe the bulk material : we consider it can be described with a dust to water ice mass ratio @xmath66 = 1 ( with @xmath67 and @xmath68 the mass fractions of dust and water respectively ) , a porosity @xmath69 = 50% , and a bulk density @xmath59 = 700 kg m@xmath70 ( typical values of the comet literature ; see @xcite ) .
the heat capacity and thermal conductivity of such a mixture are @xmath61 = 800 j kg@xmath62 k@xmath62 and @xmath64 = 0.09 w m@xmath62 k@xmath62 respectively .
the thermal evolution is computed with the formation time @xmath71 time for an object to grow to its final size as a reference for time zero , from an initial temperature of 20 k. heating due to radioactive decay is described by : @xmath72 with @xmath73 the dust bulk density , @xmath74 the mass fraction of the given radioactive isotope after a formation time @xmath71 , @xmath75 the heat released per unit mass upon decay ( 4.84 @xmath10 10@xmath76 j kg@xmath62 and 5.04 @xmath10 10@xmath76 j kg@xmath62 for @xmath56al and @xmath57fe respectively ( see * ? ? ? * ) , @xmath77 its mean - lifetime ( using 1.05 myr for @xmath56al and 3.78 myr for @xmath57fe @xcite , and @xmath19 the time .
the object formation time @xmath71 is accounted for in the model through the decay of radioactive isotopes that occurs during accretion .
no additional accretional heating is accounted for , though .
decay during the formation time @xmath71 results in a decrease of each nuclide initial abundance in the simulations : @xmath78 with @xmath79 the initial mass fraction of each nuclide .
we consider the initial nuclide ratios @xmath56al/@xmath80al = 5 @xmath10 10@xmath6 @xcite and @xmath57fe/@xmath81fe = 1 @xmath10 10@xmath82 @xcite , and mass fractions from @xcite for cv chondrites .
although the model has many free parameters , only a few can actually strongly affect the resulting thermal history , such as the formation time of the object , its composition or the thermal conductivity .
we also point out that the object s size plays an important role in its thermal evolution , owing to the surface to volume ratio .
figure [ formation ] represents the influence of the formation time , through the abundance of radiogenic nuclides .
it shows the evolution of the object s central temperature as a function of time after formation , for different formation times .
the central temperature corresponds to the maximum temperature achieved inside the object : since the surface is in thermal equilibrium with the surrounding nebula at these stages of evolution , the heating due to radioactive decay is balanced by a * cold load * coming from the surface .
figure [ profile ] displays the time evolution of the temperature profile within the body , calculated from a formation delay of 3 myr .
this figure shows that , at a given epoch of its evolution and for the adopted set of input parameters , the condition that the temperature profile must be in the 2247 k range can be satisfied in almost the whole body , thus enabling the release of n@xmath1 from this zone .
only the very close subsurface ( down to a few hundreds meters deep ) remains at lower temperatures , but this region should be rapidly depleted in volatiles once the comet follows subsequent orbits closer to the sun .
the combined effects of composition and porosity are presented in figure [ composition ] . for those simulations
, we also considered a plausible formation time of 3 myr and a fixed density of 700 kg m@xmath70 .
figure [ conductivity ] represents the influence of thermal conductivity for the standard parameters previously mentioned .
all together , these simulations show that it is possible to heat planetesimals at temperatures high enough to simultaneously imply the loss of n@xmath1 and the preservation of co , between 22 and 47 k.
* our computations show that , under certain sets of conditions and evolutionary paths , initial nitrogen - rich cometesimals similar to triton and pluto may have evolved into nitrogen - depleted comets .
the proposed scenario of selective trapping in clathrates in the nebula followed by outgassing due to short - lived radiogenic heating thus reconciles present - day observations with solar system formation models . * in our scenario ,
the composition of comets would have evolved as a result of radiogenic decay of short - lived nuclides . in the case of large objects like pluto or triton
, the thermal evolution is more complex and involves partial or total physical differentiation .
although this significant processing could involve a substantial loss of volatiles , pluto and triton are actually large enough for most volatiles to be retained at the surface and in the atmosphere by gravity , as shown by schaller & brown ( 2007 ) .
moreover , if accretion of these bodies were occurring in a cold environment and were slow
so that radiative removal of the impact heat were effective then triton , pluto and other large kbos could have retained their volatiles and accreted fairly cold ( see discussion by canup & ward ( 2002 ) in the context of the accretion of ganymede and callisto ) .
our model predicts that comets should contain xe / h@xmath1o ( @xmath03.6 @xmath10 10@xmath83 ) and kr / h@xmath1o ( @xmath05.0 @xmath10 10@xmath84 ) , ratios close to solar xe / o and kr / o , respectively .
in contrast , the ar / h@xmath1o ratio is predicted to be equal to @xmath01.7 @xmath10 10@xmath6 in comets , a value about 300 times lower than solar ar / o . in our model , the ar depletion ( compared to solar ) in comets results from the poor propensity of this element to be trapped in different mg clathrates formed in the nebula .
because of this property , ar was accreted by cometesimals only at low nebular temperature ( @xmath020 k ) and in pure condensate form . as a consequence , *
the radiogenic heating above 22 k triggers the dissociation of clathrates and the outgassing of n@xmath1 and sublimation of ar . *
measuring the noble gas abundances in comets might require in situ measurements via a high sensitivity mass spectrometer or ultraviolet spectroscopy .
the alice ultraviolet spectrometer and the rosina instrument should put upper limits on the noble gas abundances in the coma of comet 67p / churyumov - gerasimenko ( balsiger et al . 2007 ; stern et al .
2007 ) * that will probably help understand * the formation conditions of comets in the primordial nebula .
however , the measurement of solar or subsolar noble gas abundances in comets will require a new generation of instruments such as the cryotrap , which is part of the mass spectrometer maspex developed in the context of the prime spacecraft project ( young et al .
2010 ) .
huebner , w. f. , benkhoff , j. , capria , m. t. et al . , 2006 , in : heat and gaz diffusion in comet nuclei , published for the int .
bern , switzerland , by esa publ .
noordwijk , the netherlands parrish , w. r. , prausnitz , j. m. 1972 , dissociation pressures of gas hydrates formed by gas mixtures .
chem . process design dev .
11 ( 1 ) , 26 ( erratum : parrish , w. r. , prausnitz , j. m. 1972 , ind .
chem . process design dev .
11(3 ) , 462 ) ( pp72 ) .
top panel represents a global view of the body s heating due to radionuclides and then cooling over a timespan of 85 myr after formation .
bottom panel represents a zoom of the temperature evolution of the body s subsurface over 2.5 myr after formation . ] | we propose a scenario that explains the apparent nitrogen deficiency in comets in a way consistent * with the fact that the surfaces of pluto and triton are dominated by nitrogen - rich ice*. we use a statistical thermodynamic model to investigate the composition of the successive multiple guest clathrates that may have formed during the cooling of the primordial nebula from the most abundant volatiles present in the gas phase .
these clathrates agglomerated with the other ices ( pure condensates or stoichiometric hydrates ) and formed the building blocks of comets .
we report that molecular nitrogen is a poor clathrate former , when we consider a plausible gas phase composition of the primordial nebula .
this implies that its trapping into cometesimals requires a low disk temperature ( @xmath020 k ) in order to allow the formation of its pure condensate .
we find that it is possible to explain the lack of molecular nitrogen in comets as a consequence of their postformation internal heating engendered by the decay of * short - lived radiogenic nuclides*. this scenario is found consistent with the presence of * nitrogen - rich ice covers on pluto and triton*. our model predicts that comets should present xenon - to - water and krypton - to - water ratios close to solar xenon - to - oxygen and krypton - to - oxygen ratios , respectively .
in contrast , the argon - to - water ratio is predicted to be depleted by a factor of @xmath0300 in comets compared to solar argon - to - oxygen , as a consequence * of poor trapping efficiency and radiogenic heating*. |
the process of star formation , particularly the origin of the stellar initial mass function ( imf ) , is a fundamental problem in astrophysics .
photometric properties and chemical evolution of galaxies depend on their stellar content .
the process of galaxy formation can not be described independently of the process of star formation , since galaxies are partly made of stars .
stars are also an important energy source for the interstellar medium of galaxies
. stars are formed in molecular clouds ( mcs ) , which have been the focus of the research on star formation for more than two decades .
currently , there is no generally accepted theory of star formation , capable of predicting the star formation rate and the stellar imf based on the physical properties of mcs .
this is hardly surprising , since turbulent motions are ubiquitously observed in mcs , and the physics of turbulence is poorly understood , due to the great mathematical complexity of the fluid equations .
magnetic field , self gravity and high mach numbers further increase the complexity .
the steady growth of computer performance has now made large three dimensional numerical simulations of super sonic magneto - hydrodynamic ( mhd ) turbulence feasible ( padoan & nordlund 1997 , 1999 ; stone , ostriker & gammie 1998 ; mac low et al .
1998 ; padoan , zweibel & nordlund 2000 ; klessen , heitsch & mac low 2000 ; mac low & ossenkopf 2000 ; ostriker , stone & gammie 2000 ; heitsch , mac low & klessen 2000 ; padoan et al .
2001a , b ) .
comparisons of numerical experiments with observational data have shown that super
sonic turbulence can explain the morphology and kinematics of mcs , and the formation of dense cores , provided that the motions are also super - alfvnic ( padoan , jones & nordlund 1997 ; padoan et al .
1998 ; padoan & nordlund 1997 , 1999 ; padoan et al .
1999 ; padoan , rosolowsky & goodman 2001 ; padoan et al .
2001a , b ) .
we refer to this process of formation of dense cores in mcs by super sonic turbulence as _
turbulent fragmentation_. since protostars evolve from the collapse of gravitationally unstable cores in mcs , even the stellar imf could then be the result of turbulent fragmentation , with the power law shape of the imf ultimately being the consequence of the self
similar nature of turbulence .
in this paper we do not use these increasingly sophisticated numerical simulations directly , but instead develop an analytic model .
the assumptions of the model are _ inspired _ by the qualitative properties of the numerical model , but the current work does not _ depend _ on any particular set of numerical models or results .
previous analytic models by larson ( 1992 ) , henriksen ( 1986 , 1991 ) and elmegreen ( 1993 , 1997 , 1999 , 2000b ) have derived the stellar imf on the basis of the self similar structure of mcs .
larson , assuming one dimensional accretion , predicted a rather steep imf slope , equal to the mc fractal dimension , while henriksen found the imf slope to depend on both the mc fractal dimension and the relation between density and linear size for structures inside mcs .
elmegreen pointed out that the imf that results from random sampling of a self similar cloud has an exponent @xmath2 ( salpeter @xmath3 ) , independent of the cloud fractal dimension . in elmegreen s model
the imf is steeper than @xmath2 because the random sampling rate is assumed to be proportional to the square root of density , and because of `` mass competition '' .
these models are based on assumptions about the cloud geometry justified by the apparent fractal structure of mcs ( beech 1987 ; bazell & dsert 1988 , scalo 1990 ; dickman , horvath & margulis 1990 ; falgarone , phillips & walker 1991 ; zimmermann , stutzki & winnewisser 1992 ; henriksen 1991 ; hetem & lepine 1993 ; vogelaar & wakker 1994 ; elmegreen & falgarone 1996 ) , but the processes responsible for generating the assumed geometry are not discussed in detail . in a previous attempt to relate the stellar imf to the physical properties of super sonic turbulence ( padoan , nordlund & jones 1997 )
, we obtained the distribution of the local jeans mass in super sonic isothermal turbulence , from the probability distribution of the gas density .
that work also provides a prediction for the stellar imf , by identifying each local jeans mass with a protostar . although the prediction of the lower mass cutoff and the low mass portion of the imf might be roughly correct , this model under estimates the number of massive stars relative to low mass stars ( as any other log - normal imf ) .
the main reason , as pointed out by scalo et al .
( 1998 ) , is the unphysical assumption that the most massive stars originate from gas at relatively low density including only a small fraction of the total mass . in the present work we consider specific properties of mc turbulence and their consequences for the formation of protostellar cores . in particular , we i ) assume approximate self
similarity of the super sonic and super alfvnic velocity field , expressed by a power law shape of the power spectrum of turbulence , and ii ) assume that the jump conditions for isothermal mhd shocks determine the scaling of the typical size of protostellar cores .
based on these fundamental assumptions , we are able to derive a power law mass distribution of gravitationally unstable cores , with a slope consistent with the salpeter stellar imf ( salpeter 1955 ) . in order to derive the mass distribution of gravitationally unstable cores below one solar mass we also make use of the probability density function ( pdf ) of the mass density in super sonic turbulence , following padoan , nordlund & jones ( 1997 ) .
the formation of dense cores and the relation between the size of cores and the thickness of the postshock gas are discussed in the next section . in
3 the power spectrum of super sonic turbulence is discussed .
results on the pdf of the gas density in super sonic turbulence are summarized in 4 .
the mass distribution of dense cores formed by the process of turbulent fragmentation is derived in 5 , and the mass distribution of gravitationally unstable cores
is computed in
the relevance of these results for the origin of the stellar imf is discussed in 7 and 8 .
conclusions are summarized in 9 .
the qualitative properties of the local density maxima that form in super sonic turbulence follow from first principles and numerical experiments have confirmed such properties .
dense cores are formed in roughly isothermal super sonic turbulent flows as the densest parts of sheets or filaments of shocked gas .
their typical size is therefore comparable to the sheet thickness , @xmath4 . assuming that the magnetic pressure in the postshock gas exceeds the thermal pressure ( this is the case for a large range of values of the preshock magnetic field strength , even when the flow is super alfvnic ) , the isothermal shock jump conditions are : @xmath5 @xmath6 @xmath7 where @xmath8 , @xmath9 and @xmath10 , @xmath11 are the values of the gas density and magnetic field strength before and after the shocks , respectively .
@xmath12 is the linear extension of the gas before the shock ( measured in the direction perpendicular to the shock surface ) and @xmath4 is the thickness of the postshock gas .
@xmath13 is the alfvnic mach number of the shock , that is the ratio of the flow velocity and the alfvn velocity measured in the preshock gas : @xmath14 + here the relevant magnetic field components are the ones parallel to the shock surface , because the perpendicular component does not provide pressure support against the compression and it is not amplified by the compression . since only the parallel components of @xmath15 are amplified , the field in the postshock sheets is nearly parallel to the sheet surface , and elongated in the direction of dense filaments ( real ones , or two dimensional sections of sheets ) . since dense cores are formed predominantly in corrugations of sheets , the magnetic field within dense cores can occasionally show a strong curvature , depending on the orientation relative to the line of sight .
this may considerably complicate the interpretation of dust polarization measurements ( ward thompson et al .
2000 ) and should be taken into account .
the power spectrum of turbulence in the inertial range ( below the energy injection scale and above the dissipation scale ) may be assumed to be a power law , @xmath16 where @xmath17 is the wave number , and the spectral index is @xmath18 for incompressible turbulence ( kolmogorov power spectrum ) ( kolmogorov 1941 ) , and @xmath19 for pressureless turbulence ( burgers power spectrum ) ( burgers 1974 ; gotoh & kraichnan 1993 ) . the @xmath20 power spectrum has often been assumed to be the correct power spectrum of super - sonic turbulence , at least for the compressional component of the velocity field . the purpose of the present work is primarily that of establishing analytically a relation between the power spectrum of turbulence and the stellar imf and we only need to assume that the power spectrum is a power law in the inertial range .
our analytic model does not assume a specific value of the spectral index , as obtained for example in numerical simulations .
however , the predictions of our analytic model will be tested using the power spectrum of super sonic turbulence computed from numerical simulations or estimated from observational data .
for that reason we briefly summarize in this section some numerical and observational results .
the most detailed study of the power spectrum of numerical compressible turbulence has been presented by porter , pouquet & woodward ( 1992 , 1994 ) and porter , woodward & pouquet ( 1998 ) . from the 1992 paper to the 1998 paper the largest numerical resolution increased from @xmath21 to @xmath22 .
these works are limited to decaying turbulence , with mach numbers close to unity initially , and below unity at later times .
the runs are therefore sub sonic , except for an initial period of time . a magnetic field
is not included .
the velocity field is usually decomposed into its solenoidal @xmath23 and compressional @xmath24 components as @xmath25 with @xmath26 and @xmath27 .
the velocity fourier spectrum , @xmath28 , is also separated into its solenoidal and compressional parts : @xmath29 .
porter , pouquet & woodward ( 1992 ) found that the compressional modes have a power spectrum @xmath30 , and the solenoidal modes @xmath31 . in the later works , after the larger numerical resolution runs were performed , the same authors concluded instead that both compressional and solenoidal modes develop a kolmogorov power spectrum , @xmath32 , with @xmath33 ( porter , pouquet & woodward 1994 ; porter , woodward & pouquet 1998 ) .
this conclusion is not strongly supported by the plots of power spectra presented by the authors , since the largest resolution runs ( @xmath34 and @xmath22 ) are consistent with a kolmogorov power spectrum for the solenoidal modes only over a very limited range of wave numbers , approximately @xmath35 . at larger wave numbers , the power spectrum is flatter , approximately @xmath31 .
an interpretation of the shallower power spectrum of solenoidal modes at large wave numbers is provided , in terms of a `` near dissipation range '' .
the discrepancy between the steeper ( burgers ) power spectrum of compressional modes in the early @xmath21 runs ( porter , pouquet & woodward 1992 ) and the kolmogorov power spectrum in the latest @xmath21 runs ( porter , pouquet & woodward 1994 ; porter , woodward & pouquet 1998 ) is not discussed .
we have recently started to perform a large number of numerical experiments with the purpose of computing the spectral index of the inertial range of driven super
sonic mhd turbulence as a function of the sonic and alfvnic rms mach numbers of the flow .
the experiments use an isothermal equation of state , uniform initial density and magnetic fields , random initial velocity and random large scale forcing ( both solenoidal ) , as described in previous works ( padoan et al .
1998 ; padoan & nordlund 1999 ; padoan , zweibel & nordlund 2000 ) .
the results of these new experiments will be reported and discussed elsewhere .
results on power spectra and structure functions are reported in boldyrev , nordlund & padoan ( 2002 ) , where the scaling relations are also predicted on the basis of a new analytic model of super sonic turbulence ( boldyrev 2002 ) .
a power spectrum intermediate between the burgers and the kolmogorov power spectra is found , @xmath36 .
the observed velocity dispersion size larson relation ( larson 1979 , 1981 ) , @xmath37 , should reflect the power spectrum of turbulence in the interstellar medium ( @xmath38 ) , although it is often obtained from a combination of different clouds and cores inside the same cloud , or even different molecular transitions ( see goodman et al .
1998 for a discussion of the line width size relation ) .
larson finds @xmath39 in the range of scales @xmath40 pc ( larson 1979 ) ; and @xmath41 in the range of scales @xmath42 pc ( larson 1981 ) ; leung , kutner & mead ( 1982 ) obtain @xmath43 for @xmath44 pc ; myers ( 1983 ) gets @xmath45 for @xmath46 pc ; sanders , scoville & solomon ( 1985 ) find an unusually large value @xmath47 for @xmath48 pc , which they use to rule out any relation between a turbulent power spectrum and the line width
size relation ; dame et al .
( 1986 ) obtain @xmath45 for @xmath49 pc ; finally falgarone , puget & prault ( 1992 ) use a compilation of data from the literature together with their own new data , in order to sample a very large range of scales , @xmath50 pc , and include also a significant number of unbound objects ( velocity dispersion larger than the virial velocity ) , which are usually not included in earlier studies .
they find a correlation consistent with @xmath51 , and a large total scatter of almost one order of magnitude in line width .
the value @xmath51 for the exponent of the line width
size relation corresponds to a power spectrum of turbulence @xmath52 .
miesch & bally ( 1994 ) have estimated the power spectrum of turbulence in molecular clouds by computing the autocorrelation and structure functions of emission line centroid velocities .
their results correspond to an average exponent for the line width
velocity relation @xmath53 , or a power spectrum with @xmath54 .
previous attempts to measure the power spectrum of turbulence in molecular clouds with the same method had provided much shallower spectra ( kleiner & dickman 1987 ; hobson 1992 ) . a new method to estimate
the power spectrum of turbulence in molecular clouds has also been recently proposed by brunt & heyer ( 2002 ) , using the principal component analysis by heyer & schloerb ( 1997 ) .
the method has already been applied to 23 molecular clouds in the outer galaxy , and the result is a power spectrum with exponent varying from cloud to cloud , in the range @xmath55 , with a typical error of 0.08 .
the average exponent is @xmath56 .
this method has been calibrated using stochastic fields ( stutzki et al .
1998 ) , with purely random phases and no correlation between density and velocity .
correlations in real turbulent flows are likely to be important , and could affect the calibration of this method .
the study of probability density functions ( pdf ) in turbulent flows has received increasing attention over the last few years .
pdfs can provide important information complementary to power spectra .
a well known example of a combined use of power spectrum and pdf , limited to linear density fluctuations , is the press
schechter model of the galaxy mass distribution ( press & schechter 1974 ) .
a number of numerical studies have established that the pdf of mass density in isothermal turbulent flows is well approximated by a log normal distribution ( vzquez - semadeni 1994 ; padoan , nordlund & jones 1997 ; scalo et al .
1998 ; passot & vzquez - semadeni 1998 ; nordlund & padoan 1999 ; ostriker , gammie & stone 1999 ) , which can be understood analytically ( nordlund & padoan 1999 ) .
a highly radiative turbulent flow develops a complex system of interacting shocks that are able to fragment the mass distribution into a random network of dense cores , filaments and sheets and low density `` voids '' , with a large density contrast .
the intermittent nature of the log normal pdf of mass density means that most of the mass concentrates in a small fraction of the total volume of the simulation . the log normal distribution may be written as : @xmath57d\ln n ' \label{pdf1}\end{aligned}\ ] ] where @xmath58 is the number density in units of the average density @xmath59 , @xmath60 the mean of the logarithm of density , @xmath61 , is determined by the standard deviation of the logarithm of density , @xmath62 : @xmath63 which is found to be a function of the rms mach number of the flow @xmath64 : @xmath65 or , for the linear density : @xmath66 where @xmath67 , from numerical experiments .
the standard deviation of the linear density distribution thus grows linearly with the rms mach number of the flow ( nordlund & padoan 1999 ; ostriker , gammie & stone 1999 ) .
in this section we derive the mass distribution of dense cores based on the two following assumptions : i ) the power spectrum of turbulence is a power law ; ii ) the typical size of a dense core scales as the thickness of the postshock gas .
if the typical size of a dense core is comparable to the thickness of the postshock gas , @xmath4 ( 2 ) , its mass @xmath68 is : @xmath69 where we have used the jump conditions ( [ mhdn ] ) and ( [ mhdl ] ) . equation ( [ m ] ) shows that on a given scale @xmath12 , the mass of dense cores is proportional to the total mass available ( @xmath70 ) divided by the second power of the mach number on that scale ( @xmath71 ) .
given the power spectrum ( [ spectrum ] ) , the rms velocity , @xmath72 on the scale @xmath12 is @xmath73 where @xmath74 the typical shock velocity on the scale @xmath12 is therefore @xmath75 , and the shock mach number is given by ( [ ma ] ) , where @xmath76 is replaced by @xmath75 .
substituting this scale dependent expression of @xmath13 into ( [ m ] ) one obtains @xmath77 where @xmath78 is the ( large ) scale where the turbulent velocity is @xmath79 and the rms mach number is @xmath80 . the smallest scale where significant density fluctuations may be expected is approximately the scale where the mach number is of order unity : @xmath81 in mcs with a mass @xmath82 m@xmath83 a typical value of the mach number is @xmath84 .
we then obtain @xmath85 m@xmath83 , for @xmath86 .
if @xmath78 in equation ( [ m2 ] ) is defined as the largest scale of the turbulent flow ( the scale of turbulent energy injection ) , and we take @xmath87 , we obtain an estimate of the mass of the largest cores formed by turbulent fragmentation , @xmath88 where @xmath80 is the rms mach number on the largest turbulent scale . in mcs with a mass @xmath82 m@xmath83 and mach number @xmath89 , @xmath90 m@xmath83 .
to arrive at the distribution of core masses we first consider the mass distribution of a completely self similar case and then consider the modification arising from the mach number dependence .
the completely self similar case leads to equal mass contributions from each logarithmic interval ( imf slope -1 ) , as shown by elmegreen ( 1997 ) , who also demonstrated that this result does not depend on the fractal dimension of the self similar distribution . to illustrate this result , and to derive the modification of it due to the mach number dependence , it is useful to consider the setup and interpretation of numerical experiments .
numerical turbulence experiments are in a certain sense scale free ; the density is usually rescaled so that the average density @xmath91 and the size of the box is scaled so that @xmath92 . the only distinguishing , scale dependent properties that remain after such scalings are the ( sonic and alfvnic ) mach numbers ; from larson s relations ( larson 1981 ) we expect larger scales to correspond to larger rms mach numbers . to recover the completely self similar result we consider two experiments with identical initial conditions ( also with respect to mach numbers and average gas density ) , but interpreted at different scales @xmath93 , and @xmath94 .
the total mass in the `` large scale '' experiment is obviously @xmath95 larger than in the `` small scale '' experiment .
the cores in the large scale experiment would be equal in number , but heavier by the ratio @xmath95 than cores in the small scale experiment .
on the other hand , the total number of cores in the small scale experiment is @xmath95 larger than in the large scale experiment , if the same total mass is used in the two cases ( that is if cores from a number @xmath95 of small scale experiments are counted together ) .
the result is therefore a total number of cores that depends on scale as : @xmath96 in agreement with elmegreen s ( 1997 ) result .
when the mach number dependence on scale is taken into account the result is that the larger scales contribute relatively less massive cores , because of the scaling relation ( [ m2 ] ) .
we assume that the number of cores per scale @xmath12 still scales as @xmath97 . combining the relations ( [ m2 ] ) and ( [ nl ] ) we obtain : @xmath98 if the spectral index is consistent with the observed velocity dispersion size larson relation ( larson 1981 ) and with our numerical and analytical results ( boldyrev , nordlund & padoan 2001 ) , then @xmath0 and the mass distribution is @xmath99 which is almost identical to the salpeter stellar imf ( salpeter 1955 ) .
the mass distribution of dense cores has been computed assuming that the pre shock density is @xmath59 and the postshock density @xmath100 , where @xmath101 is scale dependent .
a more precise computation should include the effect of the probability distribution of the value of @xmath101 at each scale , or the overall effect of the statistics of the turbulent velocity field , which is the generation of a log normal pdf of mass density ( see 4 ) .
this is necessary to compute the fraction of dense cores that are gravitationally unstable and collapse into protostars , since dense cores can be significantly denser than their average density predicted by the scaling laws . while most of the large cores will be dense enough to collapse , the probability that small cores are dense enough to collapse is determined by the pdf of mass density .
because of the intermittent nature of the log normal pdf , even very small ( sub stellar ) cores have a finite chance to be dense enough to collapse .
we write the thermal jeans mass as : @xmath102 where : @xmath103 is the jeans mass at the mean density @xmath59 .
the distribution of the jeans mass is obtained from the pdf of density assuming constant temperature as in padoan , nordlund & jones ( 1997 ) : @xmath104 d\ln m_j , \label{}\end{aligned}\ ] ] where @xmath105 is in solar masses , and : @xmath106 the fraction of cores of mass @xmath68 with gravitational energy in excess of their thermal energy is given by the integral of @xmath107 from 0 to @xmath68 .
the mass distribution of collapsing cores is therefore @xmath108\,d\ln m \label{imfpdf}\ ] ] = 1 the mass distribution is plotted in figure 1 , for @xmath86 . in the top panel
the mass distribution is computed for three different values of the largest turbulent scale @xmath78 , assuming larson type relations ( larson 1981 ) to rescale @xmath59 and @xmath109 according to the value of @xmath78 .
the mass distribution is a power law , determined by the power spectrum of turbulence , for masses larger than approximately 1 m@xmath83 . at smaller masses
the mass distribution flattens , reaches a maximum at a fraction of a solar mass , and then decreases with decreasing stellar mass . collapsing sub
stellar masses are found , thanks to the intermittent density distribution in the turbulent flow .
the middle and bottom panel of figure 1 show the dependence of the mass distribution on the rms mach number of the flow and on the average gas density respectively .
the magnetic critical mass is derived in the next section .
we have not used it here to obtain the mass distribution of collapsing cores because the thermal jeans s mass is a more strict condition for collapse .
the magnetic critical mass depends on the core morphology in relation to the field geometry and on the magnetic field strength that correlates with the gas density with a very large scatter ( see below ) .
it is possible therefore that magnetic pressure support against the gravitational collapse affects the shape of the mass distribution , but only as a secondary effect .
observations show that the stellar imf is a power law above 12 m@xmath83 , with exponent around the salpeter value @xmath3 , roughly independent of environment ( elmegreen 1998 , 2000 ) , gradually flattens at smaller masses , and peaks at approximately 0.20.6 m@xmath83 ( hillenbrand 1997 ; bouvier et al .
1998 ; luhman 1999 ; luhman & rieke 1999 ; luhman 2000 ; luhman et al .
the shape of the imf below 12 m@xmath83 , and particularly the relative abundance of brown dwarfs , may depend on the physical environment ( luhman 2000 ) .
the scalings discussed above result in a mass distribution of dense cores consistent with the stellar imf for masses larger than 1 m@xmath83 , without invoking a sampling rate proportional to the free fall time , or `` competition for mass '' as in elmegreen ( 1997 , 1999 ) .
two conclusions are possible ; either there are effects in addition to those considered by elmegreen , and they all happen to cancel each other , or else additional effects are not important in the first place . in the spirit of occam ,
let s consider the latter possibility .
if , as argued elsewhere by elmegreen ( 2000a ) , star formation essentially happens in a crossing time , then we may indeed see only one generation of stars being produced at each scale , rather than the repeated process implied by scaling with the local dynamical time .
the picture thus is one where a particular mc forms as a consequence of the random intersection of counter - streaming , super sonic motions ( ballesteros paredes , hartmann & vzquez
semadeni 1999 ; hartmann , ballesteros
paredes & bergin 2001 ) , internal turbulence creates the distribution of core masses derived above , and the cores are then grabbed by gravitation to form one generation of stars .
energy feedback from stars subsequently disperses the cloud before the process has time to repeat . in the process
envisaged above , turbulent fragmentation is responsible for creating the core mass distribution , while gravity is only responsible for the collapse of each protostar .
the flattening and the turn around of the imf is also easily accounted for in such a model .
while scale free turbulence generates a power law mass distribution down to very small masses , only cores with a gravitational binding energy in excess of their magnetic and thermal energy can collapse .
the shape of the stellar imf is then determined by the pdf of gas density , that is by the probability of small cores to be dense enough to collapse .
the mass distribution of collapsing cores derived in the previous section and based on the log normal pdf of mass density is indeed consistent with the observed imf .
the scaling of the mass where the imf peaks can be derived without a knowledge of the pdf of mass density , using the scaling laws and the definition of the critical mass for collapse .
we first consider the magnetic critical mass , @xmath110 where @xmath111 is the magnetic critical mass at the average number density @xmath59 , @xmath112 ( mckee et al .
padoan & nordlund ( 1999 ) have shown that super sonic and super alfvnic turbulence generates a correlation between gas density and magnetic field strength , consistent with the observational data .
the two most important properties of such a @xmath113@xmath114 relation are the very large scatter , and the power law upper envelope ( @xmath115 ) .
more recently , padoan et al . ( 2001b ) have computed the magnetic field strength in dense cores produced in numerical simulations of self gravitating , super sonic and super alfvnic turbulence .
they found typical field strength as a function of column density in agreement with new compilations of observational samples by crutcher ( 1999 ) and bourke et al .
( 2000 ) . here
we adopt the following empirical @xmath113@xmath114 relation consistent with our previous works : @xmath116 where the exponent is @xmath117 , and not @xmath118 as reported above , because we now refer to the average values of @xmath113 inside bins of @xmath114 , and not to the upper envelope of the @xmath113@xmath114 relation , as above .
the slight steepening is due to the fact that the lower envelope of the @xmath113@xmath114 relation is steeper than the upper envelope .
note that ( [ bn ] ) is a statistical relation between the average magnetic field strength as a function of core density , while eqs .
( [ mhdn ] ) and ( [ mhdb ] ) are fundamental shock jump relations .
flows along magnetic field lines are able to alter the ratio of mass to magnetic flux in regions upstream of shocks , resulting in the ensemble that ( [ bn ] ) characterizes .
both relations are valid , and enter into the derivation of the critical mass .
we find the critical mass by imposing @xmath119 , where @xmath68 is given by equation ( [ m2 ] ) , and @xmath120 by equation ( [ mb ] ) , @xmath121 for @xmath86 , we get @xmath122 the critical mass is therefore typically a few times smaller than the critical mass at the average density . for @xmath89 , @xmath123 @xmath124 and @xmath125 @xmath126 g , the critical mass is approximately @xmath127 m@xmath83 .
@xmath126 g provides a normalization of the @xmath113@xmath114 relation consistent with the results of our super - alfvnic numerical simulations discussed in padoan & nordlund ( 1999 ) and in padoan et al .
( 2001b ) . ]
the probability that cores smaller than this mass are larger than their critical mass decreases with decreasing core mass , which could produce some flattening of the stellar imf .
because of the large scatter in the @xmath113@xmath114 relation the magnetic critical mass does not define a sharp cut - off in the imf , but rather a gradual flattening .
a sharper mass scale is defined by the jeans mass in the cores , @xmath128 where @xmath129 is the jeans mass at the average density defined in ( [ mj ] ) . for @xmath89 , @xmath130 @xmath124 ,
@xmath131 k and @xmath86 , the thermal critical mass is approximately @xmath118 m@xmath83 .
therefore , using physical parameters typical of nearby molecular clouds , the present model predicts that the imf should gradually flatten below approximately 2 m@xmath83 and peak in a ( logarithmic ) neighborhood of 0.4 m@xmath83 , due to increasing thermal pressure support at smaller masses ( cores smaller than their jeans mass are not included in the mass distribution ) . this result is consistent with the analytic expression of the imf derived in the previous section based of the density pdf .
as can be seen in figure 1 , the imf peaks at approximately @xmath118 m@xmath83 , for @xmath132 , @xmath130 @xmath124 , @xmath131 k and @xmath86 . in
5 we have estimated the largest mass of dense cores formed by the process of turbulent fragmentation , given by the expression ( [ mmax ] ) .
the largest stellar mass should also be of the order of @xmath133 . assuming that the turbulent velocities are of the order of the virial velocities in the parent molecular cloud , and adopting the larson relation @xmath134 ( larson 1981 )
, we obtain @xmath135 .
this is close to the empirical relation @xmath136 ( larson 1982 ) , which is found for a cloud sample known to follow the above size density relation .
it is likely that the true exponent of this relation is slightly larger than the value found by larson ( 1982 ) , because the lifetime of the most massive stars is comparable to , or shorter than , the lifetime of their parent molecular clouds , and therefore the probability of observing the most massive stars decreases with increasing mass .
elmegreen ( 1993 , 1997 ) has argued that the largest stellar mass is related to the mass of the parent cloud for purely statistical reasons : the larger the cloud mass is , the higher the probability of populating the high mass tail of the imf .
such a statistical argument is correct only if the normalization of the imf is independent of the total cloud mass .
the scalings derived above show that this is not the case ; larger clouds in general have larger velocities and therefore form relatively fewer stars of a given mass .
this is also indicated by observations , since it is commonly found that larger clouds have lower star formation efficiency than smaller clouds .
however , since the star formation efficiency also depends on the low mass cut - off of the imf , and since age differences also may enter , it would have been hard to draw firm conclusions from observations alone .
in section [ coremass ] we found that the mass distribution of dense cores formed by turbulent fragmentation has a power law shape , with a slope consistent with the stellar imf at intermediate and large masses , and in section [ collapse ] we found the distribution of the mass of collapsing cores , which has a shape consistent with the observed imf also at small masses .
three assumptions were essential for the derivation , and we would here like to emphasize what these assumptions are , and point out that they are independent of each other .
the first assumption is that approximate selfsimilarity holds for supersonic turbulence of varying mach number , to the extent that the scaling of the number of cores with upstream scale @xmath12 scales as @xmath97 , also when the mach number varies with scale @xmath12 .
this is a reasonable assumption because the compressed cores occupy only a tiny fraction of the volume where their mass was collected from .
the second assumption is that the mass of the compressed cores scales as the cube of the thickness of the shocked gas , which again scales as the inverse of the mach number .
this assumption is independent of the particular density and scale of the upstream flow , and rests only on the fundamental shock jump condition and on the core shape being independent of mach number .
the third assumption is that the mach number of the upstream gas depends on the scale @xmath12 of the upstream flow as @xmath137 , where @xmath138 is essentially the index in larson s ( 1979 , 1981 ) velocity size relation .
it is remarkable that , under these assumptions , an imf slope consistent with salpeter s ( 1955 ) value is obtained as a direct result of the observed velocity dispersion
size relation of ism turbulence and the fundamental jump conditions for isothermal mhd shocks .
provided that some general criteria are met , the slope of the stellar imf is thus independent of the physical conditions in the star forming clouds , as also indicated by the observations ( elmegreen 1998 , 2000 ) .
we have also interpreted the gradual flattening of the imf around 12 m@xmath83 and its peak at approximately 0.30.5 m@xmath83 as the effect of thermal ( possibly also magnetic ) support against the gravitational collapse .
the mass distribution of collapsing cores , computed on the basis of the pdf of mass density and the thermal jeans mass , is consistent with the observed stellar imf , down to sub
stellar masses .
we have shown that the imf at low stellar masses is sensitive to the average physical properties of the star forming gas , such as the rms mach number and the average gas density , also consistent with the observations ( luhman 2000 ) .
an even larger number of sub
stellar objects would be predicted if a significant fraction of them originate as close companions of more massive stars and later on separate as individual brown dwarfs . in the present work we do not study the evolution of density fluctuations smaller than their jeans mass within the collapsing background of an unstable larger core , and therefore this mechanism for the formation of giant planets and brown dwarfs is not discussed here .
the assumption behind the result is that the upstream density is sampled from a log - normal density distribution , and that the distribution of the densities of cores therefore also is approximately log - normal .
an exact log - normal density distribution requires exactly isothermal conditions , but even for non - isothermal conditions the central part of the density pdf is still approximately log - normal ( scalo et .
al 1998 , nordlund & padoan 1999 ) .
given the fact that turbulent fragmentation is unavoidable in super sonic turbulence , and given the success of the present model in predicting the correct slope of the stellar imf , it is difficult to argue that super sonic turbulence does not play a dominant role in the generation of the stellar imf .
other processes such as gravitational fragmentation ( larson 1973 ; elmegreen & mathieu 1983 ; zinnecker 1984 ) , opacity limited fragmentation ( hoyle 1953 ; gaustad 1963 ; yoneyama 1972 ; suchkov & shchekinov 1976 ; low & lynden - bell 1976 ; rees 1976 ; yoshii & saio 1985 ; silk 1977a , b ) , protostar interactions and coagulation ( nakano 1966 ; arny & weissman 1973 ; silk & takahashi 1979 ; bastien 1981 ; yoshii & saio 1985 ; lejeune & bastien 1986 ; allen & bastien 1995 , 1996 ; price & podsiadlowski 1995 ; murray & lin 1996 ) , stellar winds and outflows ( silk 1995 ; nakano , hasegawa & norman 1995 ; adams & fatuzzo 1996 ) , competitive accretion ( larson 1978 ; tohline 1980 ; bonnell et al . 1997 ; myers 2000 ) must be relatively unimportant . this conclusion is supported by a recent computation of the mass distribution of dense self gravitating cores in numerical simulations of self gravitating super sonic mhd turbulence ( padoan et al .
the result is a power law mass distribution consistent with the stellar imf .
mass distributions of cores from numerical simulations of super sonic turbulence , roughly consistent with the stellar imf , are also reported by klessen ( 2000 ) .
it should be noted that even the formation of molecular clouds , a process that some authors ascribe to the random collisions of gas streams ( e.g. ballesteros - paredes , hartmann & vzquez - semadeni 1999 ; hartmann , ballesteros
paredes & bergin 2001 ) , may be a manifestation of a process analogous to the one discussed above , just operating on a larger scale .
if motions on larger scales are also characterized by power laws , one would then expect the distribution of mc masses to obey a salpeter - like scaling
. however , such scaling would be rather difficult to probe observationally , due to the fact that mcs are dispersed by the process of star formation before they are able to collapse as a whole , which prevents them from being well defined as individual objects . for almost twenty years estimates of the mass distribution of dense cores in molecular clouds ,
based on molecular line studies , found a shallow power law mass distribution with a single exponent in the range @xmath139 , with a typical value @xmath140 .
the main reasons for the shallow mass distribution obtained in these works are i ) the relatively low density traced by the molecular emission lines normally used ( @xmath141 @xmath124 for @xmath142co , and @xmath143 @xmath124 for c@xmath144o ) ; ii ) the limited density range probed by each molecular emission line ; iii ) the use of inappropriate `` clump find '' algorithms ; iv ) the relatively low resolution that allows only the selection of large cores , with typical mass ranging from @xmath145 m@xmath83 to hundreds or thousands m@xmath83 .
clearly , cores selected in this way can not be identified with single protostellar cores , as discussed in padoan ( 1995 ) , where it was predicted that the exponent of the intrinsic mass distribution of protostellar cores should have been @xmath146 , based on the stellar imf .
recently , onishi et al .
( 1999 ) have obtained a sample of dense cores in the taurus molecular cloud complex using a higher density tracer , h@xmath142co@xmath147 , which probes a density of approximately @xmath148 @xmath124 .
they found a power law mass distribution with exponent @xmath149 , in the mass range between @xmath150 m@xmath83 and @xmath151 m@xmath83 . while most previous determinations of core mass distributions using molecular line maps are affected by the arbitrary definition of an individual core , which is far from trivial in the hierarchical cloud structure , the dense h@xmath142co@xmath147 cores found by onishi et al .
( 1999 ) in taurus are all isolated and therefore unambiguously defined . recent dust continuum emission surveys , which also probe relatively high densities ( @xmath148@xmath152 @xmath124 ) , have provided more support to the idea that the stellar imf reflects the mass distribution of dense cores .
the mass distribution of dense cores in @xmath153 ophiuchi ( motte et al .
1998 ; johnstone et al .
2000 ) , in the serpens core ( testi & sargent 1998 ) and in orion ( motte et al.2001 ; johnstone et al .
2001 ) are found to be consistent with the stellar imf .
motte et al .
( 1998 ) , for example , obtained a power law mass distribution with exponent @xmath154 ( in logarithmic units such that salpeter s exponent is @xmath3 ) in the range of masses between @xmath117 m@xmath83 and @xmath155 m@xmath83 , and @xmath140 in the range between @xmath156 m@xmath83 and @xmath117 m@xmath83 .
the results by motte et al .
( 1998 , 2001 ) are confirmed by johnstone et al .
( 2000 , 2001 ) .
testi & sargent ( 1998 ) found @xmath157 , in the range of masses between @xmath117 m@xmath83 and @xmath158 m@xmath83 .
detailed studies of pre
stellar cores selected in this way could provide accurate volume density measurements and therefore a test for the density size relation implied by our equation ( [ m ] ) .
note , however , that because all quantities involved have wide distribution functions it is important to distinguish between average or typical values , extreme values , and observed values . the latter may often be more representative of extreme values than of average values .
one example is larson s density size relation , which by its nature tends to refer to the densest cores of a given size rather than the average cores of that size ; many low density cores may exist that are unobservable , or are ignored . in the context of our derivation and assumptions
one may understand larson s density size relation as the result of combining eq .
[ mhdn ] and [ mhdb ] and then taking the extreme with respect to @xmath12 ; i.e. , the relation represents those extreme cases where roughly the same initial density gas has been compressed by shocks , perhaps repeatedly , to various degrees .
if each compression is essentially one dimensional one retains the approximately inverse relation between core size and core density .
in conclusion , we have related the stellar imf to the mass distribution of dense cores formed by the process of turbulent fragmentation , assuming that only cores with gravitational energy in excess of their magnetic and thermal energy can collapse as protostars .
critical cores disperse back into the turbulent flow , and are therefore irrelevant for the process of star formation , as we have argued in other recent works ( padoan et al .
2001a , b ) .
previous theories of star formation ( see shu , adams & lizano 1987 ) assume that stars of small and intermediate mass are formed from sub critical cores .
in such theories , sub
critical cores are in static equilibrium and evolve quasi statically , on the time
scale of ambipolar drift .
we have argued that such a scenario is inconsistent with the turbulent nature of mcs ( padoan et al . 2001a , b ) .
we have derived the mass distribution of dense cores generated by the process of turbulent fragmentation , and have found a power law mass distribution consistent with the salpeter stellar imf , for stellar masses larger than 12 m@xmath83 .
this result is a direct consequence of fundamental physical properties of super sonic turbulence in mcs , such as the power spectrum of turbulence and the jump conditions for isothermal mhd shocks .
we have also shown that another fundamental physical property of turbulent flows , namely the pdf of the gas density , can explain the shape of the imf at smaller stellar masses , and the formation of gravitationally unstable cores of sub stellar mass .
the main results of this work are : i ) the power law stellar imf at masses larger than 12 m@xmath83 is the result of the near self
similar nature of inertial range super sonic turbulence ; ii ) the low mass roll - over and cut - off of the imf is caused by the combined effects of the thermal support of the smallest cores against gravitational collapse and the density pdf of super sonic turbulence ; iii ) the mass scaling of the peak of the imf may be expressed as a function of the physical parameters of turbulent star forming clouds ; primarily their rms turbulent velocity , temperature , magnetic field strength and average density ; iv ) for physical parameters typical of nearby mcs the peak of the imf is predicted to be around 0.30.5 m@xmath83 ; v ) the slope of the imf at masses larger than 12 m@xmath83 is determined by the inertial range spectral index of super sonic turbulence and the jump conditions for isothermal mhd shocks ; vi ) for magnetically dominated jump conditions , which are applicable to typical molecular cloud conditions , the imf spectral index is equal to @xmath159 , where @xmath160 is the inertial range spectral index of super sonic turbulence ; vii ) for a value of the inertial range spectral index @xmath0 , consistent with larson s relations and with new numerical results , the imf spectral index is @xmath161 , almost identical to salpeter s slope .
we conclude that the process of turbulent fragmentation is essential to the origin of the stellar imf , in support of the thesis that star formation can be viewed as the main consequence of the dissipation of super sonic turbulence in molecular clouds .
we thank the anonymous referee for useful comments .
we are grateful to bruce elmegreen , alyssa goodman , chris mckee , phil myers and ralph pudritz for valuable discussions .
this work was performed while pp held a national research council associateship award at the jet propulsion laboratory , california institute of technology . .n .
acknowledges partial support by the danish national research foundation through its establishment of the theoretical astrophysics center . ,
t. , mizuno , a. , kawamura , a. , fukui , y. 1999 , in star formation 1999 , proceedings of star formation 1999 , held in nagoya , japan , june 21 - 25 , 1999 , editor : t. nakamoto , nobeyama radio observatory , p. 153 - 158 , p. 153 , k. , umemoto , t. , kameya , o. , hirano , n. , hasegawa , t. , hayashi , m. , iwata , t. , kaifu , n. , mikami , h. , murata , y. , nakano , m. , nakano , t. , ohashi , n. , sunada , k. , takaba , h. , yamamoto , s. 1993 , , 404 , 643 * figure captions : * + * figure [ fig3 ] : * mass distributions of gravitationally unstable cores from equation ( [ imfpdf ] ) .
top panel : mass distribution for different values of the largest turbulent scale @xmath78 , assuming larson type relations ( for rescaling @xmath59 and @xmath109 with @xmath78 ) , @xmath131 k and @xmath86 .
middle panel : mass distribution for different values of @xmath109 , assuming @xmath130 @xmath124 , @xmath131 k and @xmath86 .
bottom panel : mass distribution for different values of @xmath59 , assuming @xmath162 , @xmath131 k and @xmath86 .
the mass distribution peaks at approximately @xmath118 m@xmath83 , for the values @xmath162 , @xmath130 @xmath124 , @xmath131 k and @xmath86 , typical of nearby molecular clouds . | the morphology and kinematics of molecular clouds ( mcs ) are best explained as the consequence of super sonic turbulence . super sonic turbulence fragments mcs into dense sheets , filaments and cores and large low density `` voids '' , via the action of highly radiative shocks .
we refer to this process as _ turbulent fragmentation_. in this work we derive the mass distribution of gravitationally unstable cores generated by the process of turbulent fragmentation .
the mass distribution above one solar mass depends primarily on the power spectrum of the turbulent flow and on the jump conditions for isothermal shocks in a magnetized gas . for a power spectrum index @xmath0 ,
consistent with larson s velocity dispersion
size relation as well as with new numerical and analytic results on super sonic turbulence , we obtain a power law mass distribution of dense cores with a slope equal to @xmath1 , consistent with the slope of the stellar imf .
below one solar mass , the mass distribution flattens and turns around at a fraction of a solar mass , as observed for the stellar imf in a number of stellar clusters , because only the densest cores are gravitationally unstable .
the mass distribution at low masses is determined by the probability distribution of the gas density , which is known to be approximately log
normal for an isothermal turbulent gas .
the intermittent nature of the turbulent density distribution is thus responsible for the existence of a significant number of small collapsing cores , even of sub stellar mass . since turbulent fragmentation is unavoidable in super
sonically turbulent molecular clouds , and given the success of the present model in predicting the observed shape of the stellar imf , we conclude that turbulent fragmentation is essential to the origin of the stellar imf .
= 1 epsf 2ns^/n |
warm absorbers reveal their presence by imprinting absorption edges on the soft x - ray spectra of active galaxies ( agn ) .
many were found and studied on the basis of observations by ( e.g. , nandra & pounds 1992 , komossa 1997 ) and ( e.g. , mihara et al .
1994 , reynolds 1997 ) . within the last year evidence has accumulated that some warm absorbers contain significant amounts of dust .
this possibility was first suggested by brandt et al .
( 1996 ) to explain the lack of excess x - ray _ cold _ absorption despite strong optical reddening of the quasar iras 13349 + 2438 .
models that explicitly include the dust - gas and dust - radiation interaction using the photoionization code _ cloudy _ ( ferland 1993 ) were calculated by komossa & fink ( kofi hereafter ; 1996 , 1997a - d ) .
depending on the warm absorber parameters , the presence of dust turned out to have a strong influence on the x - ray absorption structure .
the models were applied to several seyfert galaxies .
ngc 3227 ( kofi 1996 , 1997b ) and ngc 3786 ( kofi 1997c ) were shown to be very good candidates for dusty warm absorbers as judged from optical - uv reddening properties as well as _ successful x - ray spectral fits_. on the other hand , the bulk of the warm material in ngc 4051 was found to be dust - free ( kofi 1997a ) .
recently , reynolds et al .
( 1997 ) also presented a _ cloudy_-based dusty absorber model which they successfully applied to the outer warm absorber in mcg 6 - 30 - 15 .
indirect evidence for the association of some warm absorbers with dust was discussed in reynolds ( 1997 ) . iras 17020 , which is part of the present study ,
is a narrow line seyfert1 galaxy ( nlsy1 ; moran et al .
1996 , wisotzki & bade 1997 ) at redshift @xmath1=0.06 ( de grijp et al .
wisotzki & bade stressed the heavy reddening of the optical spectrum .
they also presented a powerlaw fit to the pspc spectrum and pointed to the discrepancy between the cold column density derived from optical reddening and the one from the x - ray fit .
leighly et al .
( 1997 ; l97 ) detected an oxygen ovii edge in the spectrum of iras 17020 as well as high optical polarization .
they also confirmed the strong optical reddening and derived a corresponding gaseous column density of @xmath2 = 4 @xmath3 @xmath4 . on this basis
they suggested the presence of a warm absorber with internal dust in iras 17020 .
to test whether the model of a warm absorber that _ includes _ the presence of dust consistently fits the x - ray spectrum of iras 17020 , we apply such a model to the ( archival ) pspc spectrum of this source ( sect .
2 ) . motivated by increasing indirect evidence for the presence of dusty warm absorbers , in the main part ( sect .
3 ) we point out some important general properties of the dusty material that can result in strong modifications of the x - ray absorption structure as compared to the dust-_free _ case , and study consequences .
since the presence of dust influences the x - ray spectrum most strongly at soft energies , one signature being a pronounced carbon edge at 0.28 kev outside the sensitivity range , ( trmper 1983 ) is particularly well suited for such a study . to search for time variability of iras 17020 , and check for the potential contribution to the x - ray spectral complexity of this source from the closeby optically bright star sao 46462 , we also present ( pi ) hri observations
_ pspc_. a 2.6 ksec pointed observation centered on iras 17020 was performed on aug .
28 , 1992 .
the background corrected source photons were extracted and vignetting and dead - time corrections were applied using exsas ( zimmermann et al .
the mean source count rate is 0.80@xmath5 cts / s and there are hints for 25% variability during the observation ( fig .
[ light ] ) . _
iras 17020 was observed with the hri on march 4 ( 1.4 ksec ) , and april 4 - 5 , 1994 ( 4.1 ksec ) .
we find mean source count rates of 0.29@xmath5 and 0.22@xmath6 cts / s , respectively .
again , there are hints for short - timescale variability ( fig .
[ light ] ; we note that there are no strong variations in the background during the observations ) .
the f8 star sao 46462 in 108 distance from iras 17020 is detected in x - rays . however , in the pspc ( hri ) , the star is a factor of @xmath750 ( 80 ) weaker than the target source and thus does not significantly confuse the spectrum of iras 17020 . the conversion of all count rates to pspc rates ( assuming constant spectral shape ) yields values of @xmath8 = 0.91 , 0.80 , 0.71 , 0.93 and 1.1 cts / s for the survey , pspc , hri-1 , hri-2 and ( l97 ) observation , respectively , revealing small - amplitude long - term variability .
for the spectral analysis source photons in the channels 11 - 240 were binned according to a signal / noise ratio of 10@xmath9 . a single powerlaw ( pl ) with cold absorption as a free parameter gives a good fit to the pspc x - ray spectrum ( @xmath10 = 0.8 ) with a photon index @xmath11 , absorption of = 0.3510@xmath12 @xmath4 and a 1-kev photon flux @xmath13 = 2.7110@xmath14 ph/@xmath15/s / kev .
the error ellipses for this model are displayed in fig .
[ chi ] .
application of other spectral models , like a black body or a raymond - smith model , yielded no acceptable fits .
a pl plus black body and a double pl fit was performed , fixing the cold absorption to the value @xmath2 derived from optical reddening .
this was to test whether a low - energy soft excess component could compensate for the stronger cold absorption thus explaining the data without warm absorber .
again , no acceptable fit could be achieved .
we have carefully searched for remaining systematic residuals in the pl fit around the locations of the oxygen absorption edges none are found .
to check whether a dusty warm absorber may nevertheless be present , with a sequence of absorption edges not individually resolved ( cf .
3 ) , we have fit our _ cloudy_-based dusty warm absorber models to the x - ray spectrum of iras 17020 .
the models and the chosen intrinsic continuum are described in more detail in komossa & fink ( 1997a , b , c ) . in brief , the ionized material was assumed to be of constant density , ionized by a ` mean seyfert ' ir to gamma - ray continuum , and in photoionization equilibrium ( for cautious notes on the latter assumption see krolik & kriss 1995 ) .
the dust composition and grain size distribution were chosen like in the galactic diffuse interstellar medium ( mathis et al .
1977 ) if not mentioned otherwise , as incorporated in _ cloudy _ , and the metal abundances were depleted correspondingly ( see ferland 1993 , baldwin et al .
1991 for details ) . the fit parameters of the dusty warm absorber are its column density @xmath16 and the ionization parameter @xmath17 . in a first step ,
the photon index was fixed to the value derived from the fit of a single pl , = 2.4 .
fitting a warm absorber with free @xmath18 , @xmath16 and @xmath19 , we are led back to the limit of no warm absorption . a fit with @xmath16 fixed to @xmath2 yields cold absorption consistent with the galactic value but @xmath10 = 1.6 .
the fit becomes acceptable if non-`standard ' dust ( i.e. , silicate only ) is chosen ( @xmath10 = 1 ) . allowing for steeper intrinsic powerlaw spectra
, we find satisfactory fits also for dust including the graphite species ( cf .
[ sedx ] ) .
fixing @xmath16 = @xmath2 , the spectrum can be successfully described by parameters ( , @xmath18 , @xmath20 ) @xmath0 ( 2.6 , 0.310@xmath12 , 1.0 ) and ( 3.0 , 0.4510@xmath12 , 0.4 ) and values inbetween for @xmath21 1 .
finally , we checked whether the spectrum could be reconciled with the canonical photon index of = and a _ dust - free _ warm absorber . in this case
we find @xmath22 23.2 and @xmath23 0.9 , but @xmath10 = 1.2 .
although highly ionized high column density material may provide another explanation for the observed optical polarization via electron scattering , this model predicts a strong oviii edge ( not observed by _
asca _ ; l97 ) and leaves systematic residuals in the fit .
\(i ) gas - dust interactions influence the thermal conditions in the gas and change its ionization state ( e.g. draine & salpeter 1979 ) . in particular , for high ionization parameters dust effectively competes with the gas in absorbing photons ( laor & draine 1993 ) . adding dust to large column density warm material ( cf .
[ dust_seq]b ) results in strongly increased soft x - ray absorption , partially due to the stronger temperature gradient across the absorber , with more gas in a cold state .
+ ( ii ) dust scatters and absorbs the incident radiation ( e.g. martin 1970 ) and x - ray absorption edges are created by inner - shell photoionization of metals bound in the dust ( cf .
figs 4,5 of martin & rouleau 1991 ) .
strong edges are those of neutral carbon and oxygen ( fig .
[ dust_seq]a , b ) .
the shift in edge energies due to chemical and solid state effects is only of the order of a few ev ( greaves et al .
1984 ) .
the modifications of the observed x - ray spectrum as compared to the dust - free case can be drastic and are important to take into account when interpreting the observed x - ray properties of agn .
consequences of the presence of dust are \(a ) ` flattening effect ' : the sequence of individual edges ( dust - ones adding to the gas - ones ) and the shift to lower ionization lead to a ` smoothing ' and apparent ` flattening ' of the observed x - ray spectrum ( fig . [ dust_seq]a ) .
individual edges are less pronounced as in the dust - free case ( where ovii and/or oviii often dominate ) .
this ` flattening effect ' of dust ( in fact opposite to that of a dust - free warm absorber , which effectively steepens the x - ray spectrum in the band ) has consequences for nlsy1s in general ( sect .
another consequence is the possibility to ` hide ' dusty warm absorbers in the spectra ( as shown by the successful fit to iras 17020 ) if the column densities are not too large , although effect ( b ) has to be taken care of ( or else non - standard dust properties allowed ; dust in other galaxies may be different from galactic one , but standard dust properties are also usually assumed to estimate optical reddening ) .
\(b ) carbon edge : often , the strongest dust - created edge is that of carbon at 0.28 kev , stemming from the graphite species of dust . its presence may prevent a successful spectral fit and it is important to actually apply the model of a dusty warm absorber to the data .
whereas dust including the graphite species is well consistent with the x - ray spectra of ngc 3227 ( kofi 1997b ) and ngc 3786 ( kofi 1997c ) and pure graphite was favoured by reynolds et al .
( 1997 ) for mcg 6 - 30 - 15 , it hampered a successful fit in other cases .
\(c ) increased sensitivity to radiation pressure : material of high ionization parameter and particularly dust particles are strongly subject to radiation pressure ( e.g. , laor & draine 1993 , chang et al .
1987 , binette et al .
1997 ) , causing the gas to outflow ( if not otherwise confined ) and leading to temporal changes of the absorber properties .
indeed , strong variability in x - ray count rate ( of factors 10 - 15 ) has been recently detected for the dusty - warm - absorber candidates ngc 3227 and ngc 3786 which could be explained by variability in column density or ionization state of the warm absorber ( kofi 1997b , c ) .
given the potentially strong modifications of the x - ray spectrum in the presence of _ dusty _ warm material , the successful x - ray spectral fit lends further support to the suggestion ( l97 ) of the presence of a dusty warm absorber in iras 17020 .
although there is evidence for small excess cold absorption ( e.g. fig .
1 ) no description of the soft x - ray spectrum that involves the _ large cold _ column inferred from optical observations could be found .
nlsy1s like iras 17020 are generally characterized by narrow broad lines , strong feii emission and steep x - ray spectra ( see brandt et al .
1997 for a recent discussion ) .
several models were suggested to explain their apparently steep soft x - ray spectra , like a strong soft excess on top of a flat powerlaw ( e.g. puchnarewicz et al . 1992 ) or a ( dust - free ) warm - absorbed flat powerlaw ( e.g. kofi 1997d ) .
recent observations reveal increasing x - ray spectral complexity , with often both , a warm absorber and a soft excess present .
the presence of _ dusty _ warm material adds further to this complexity .
in particular , the effective flattening of the observed spectrum implies a steeper intrinsic one , thus exaggerating the problem of producing the strong feii emission ( that in standard models requires hard x - ray photons ) in nlsy1s .
the possibility of a warm - absorber contribution to the optical high - ionization lines in seyferts and nlsy1s was studied by kofi ( e.g. 1997d , fig .
, we note that in case the absorber is _ dusty _ , weaker contribution to line emission is found .
in particular , due to the binding of fe in dust , negligible contribution to high - ionization fe coronal lines is expected .
although we find hints for x - ray variability of iras 17020 of the order of 25% on timescales from several hundred seconds to years , this amplitude is remarkably small as compared to some other reported cases of seyfert and nlsy1 variability .
we have shown how the presence of dust in warm absorbers can lead to strong modifications of the observed x - ray spectrum .
the model of a dusty warm absorber was successfully fit to the x - ray spectrum of the nlsy1 galaxy iras 17020 .
this corroborates the suggestion of the presence of dusty warm material in this galaxy made on the basis of optical properties ( wisotzki & bade 1997 , l97 ) and the detection of an oxygen edge in the spectrum ( l97 ) . this first good case for a _ dusty _ warm absorber in a nlsy1-type galaxy , together with the other good candidates in the sy1.8 galaxy ngc 3786 ( kofi 1997c ) , the sy1.5 ngc 3227 ( kofi 1996,1997b ) , the sy1 mcg 6 - 30 - 15 ( reynolds et al .
1997 ) , and the quasar iras 13349 ( brandt et al . 1996 ) suggests this component to be common in all types of agn .
given the strong x - ray absorption edges of neutral dust - bound metals , dusty warm absorbers will play an important role not only in probing components of the active nucleus , like the dusty torus that is invoked in unification schemes , but also are they a very useful diagnostic of the ( otherwise hard to determine ) dust properties and dust composition in other galaxies .
current and future x - ray missions with sensitivity at soft energies and high spectral resolution ( like sax , axaf and spektrum - x ) will play an important role in studying these issues .
the project is supported by the german bundesministerium fr bildung , wissenschaft , forschung und technologie ( bmbf / dlr ) and the max - planck - society .
we thank gary ferland for providing _ cloudy _ , and hartmut schulz , jochen greiner and dirk grupe for a critical reading of the manuscript . | we present a study of the properties of _ dusty _ warm absorbers and point to some important consequences for the resulting x - ray absorption spectra of agn .
pronounced effects of the presence of _ dust _ in _ warm _ material are ( i ) an apparent ` flattening ' of the observed x - ray spectrum in the band due to a sequence of absorption edges and a shift to lower gas ionization , ( ii ) the presence of a strong carbon edge at 0.28 kev , and ( iii ) the expectation of increased time variability of the warm absorber parameters .
the first two effects can be drastic and may completely hamper an x - ray spectral fit with a dusty warm absorber even if a dust-_free _ one was successfully applied to the data . in order to demonstrate facets of the dusty warm absorber model and test the
recently reported important , albeit indirect , evidence for dusty warm material in the narrow - line seyfert 1 galaxy iras 17020 + 4544 we have analyzed pspc and hri observations of this galaxy .
the x - ray spectrum can be successfully described by a single powerlaw with index = 2.4 plus small excess cold absorption , or alternatively by a steeper intrinsic powerlaw ( @xmath0 2.8 ) absorbed by a _ dusty _
warm absorber .
the findings are discussed in light of the nlsy1 character of iras 17020 and consequences for nlsy1s in general are pointed out .
in particular , the presence of dusty warm gas results in a steeper intrinsic powerlaw than observed , thus exaggerating the ` feii problem ' .
it also implies weaker potential warm - absorber contribution to high - ionization fe coronal lines .
= 1 2.7 cm _ _ _ _ |
since the discovery of high - temperature superconductors ( htscs ) , their engineering applications at liquid nitrogen temperatures have drawn much attention .
the yba@xmath1cu@xmath2o@xmath3 ( ybco ) is one of the promising htscs materials for various technical applications .
however , in ceramic materials the critical current density ( @xmath10 ) is , due to the weak link effects , very low , rendering them unsuitable for applications . in order to enhance @xmath10 , the so - called melt - textured growth ( mg ) process
@xcite has been developed , which can significantly enhance @xmath10 .
one interesting phenomenon in the mg htscs is the fishtail effect ( fe ) in @xmath10 as well as in isothermal magnetic hysteresis loops ( @xmath11 curves ) @xcite . as for the fe origin
, some researchers propose the vortex ordered phase to disordered phase transition @xcite to explain this interesting phenomenon , while others attribute it to the bulk pinning @xcite . in the vortex phase transition explanation ,
the fe exists both on a single @xmath12 curve and on a single @xmath13 curve .
however , in the bulk pinning explanation , the fe appears on @xmath11 curves whereas on a single magnetization vs. temperature @xmath14 curve it is seldom reported . in the previous work , the critical state model is used to study the isothermal @xmath11 curves @xcite .
however , for htscs , due to their operation at higher temperatures and due to their small activation energy @xmath15 , the flux creep is significant .
hence , non - linear flux creep models were developed .
meanwhile , the surface barrier @xcite , which prevents vortex entering ( @xmath16 ) and exiting ( @xmath17 ) superconductors , has strong effects on the irreversibility of htscs .
the goal of this work was to contribute to the understanding of the fe origin in terms of significant influence of both bulk pinning and surface barriers .
mg samples yba@xmath1cu@xmath2o@xmath3 and yba@xmath1cu@xmath4ag@xmath5o@xmath3 were grown as was described in @xcite .
two samples , pure crystal ( s1 ) and ag - doped crystal ( s2 ) , were used in the present work .
physical properties of the samples are presented in table [ tab:1 ] .
.[tab:1]physical characteristics for samples yba@xmath1cu@xmath2o@xmath3 ( s1 ) and yba@xmath1cu@xmath4ag@xmath5o@xmath3 ( s2 ) [ cols="^,^,^,^,^",options="header " , ] field dependences of critical current density @xmath10 ( a ) and derivative @xmath18 ( b ) at temperature @xmath1970 k for sample s1 right after annealing . points a and b denotes positions of minimum and maximum respectively.,width=307,height=336 ] the measurements of isothermal hysteresis loops ( @xmath11 curves ) at various temperatures ( 20 k@xmath785 k ) were carried out in fields ( -6 t@xmath206 t ) parallel to the c - axis by means of a commercial vibrating sample magnetometer ( vsm ) .
the field sweep rate was @xmath21 t / min . to obtain an additional parameter @xmath22
the resistivity @xmath23 curves were measured on two samples ( the first one is pure and the second one is ag - doped ) which were obtained from the same single - domain bulks as s1 and s2 .
transport measurements were performed using a physical property measurement system ( ppms ) .
the experiment was divided into two steps .
@xmath11 curves were measured : a ) right after annealing of samples in oxygen ( during 240 hours at 400@xmath24c ) ; b ) after long - term aging at the room temperature ( six months ) , in which case almost all relaxation processes were finished .
the first value for @xmath25 in table [ tab:1 ] corresponds to a ) and the second value corresponds to b ) .
from the measured @xmath11 curves the critical current density was calculated using the critical state model @xcite that have been developed in @xcite : @xmath26 where @xmath27 ( @xmath280 , @xmath290 are the values of magnetic moment at field @xmath30 , @xmath31 is the density of sample , a , b and c are geometric dimensions of sample .
dependences @xmath13 ( fig . [ fig:1]a for @xmath32=70 k ) were calculated for the field range 0 t@xmath206 t. to obtain the exact position of maximum on @xmath13 curves the derivative @xmath18 was calculated for each temperature ( see example at fig . [
fig:1]b ) .
[ fig:1 ] shows positions of both minimum ( point a ) and maximum ( point b ) on j vs h curve , where derivative @xmath330 .
contour for @xmath34 for sample s1 right after annealing .
points a and b coincide with points a and b in the fig .
[ fig:1],width=307 ] then the 3d surface @xmath35 was obtained .
intermediate curves were defined by the 3rd degree polynomial ( spline - function ) .
the contour for @xmath34 is shown in fig .
[ fig:2 ] . points a and b correspond to those in fig .
[ fig:1 ] .
this method is useful to obtain position of maximum @xmath36 ( or ensemble of points b ) not only at given temperatures but also for intermediate temperatures .
temperature dependences of critical current density at zero magnetic field for samples s1 ( green lines ) and s2 ( blue lines ) , right after annealing ( curves 1,2 ) and after long - term aging ( curves 3,4),width=307 ] fig .
[ fig:3 ] shows the temperature dependences @xmath12 , on which fe appears .
however , according to @xcite , fe appears not due to the order - disorder phase transition of vortex lattice .
the possible reason of the fe existence is a significant influence of both bulk pinning and surface barriers @xcite . from the model in previous chapter
the @xmath37 dependences of @xmath36 were obtained ( see fig .
[ fig:4 ] ) .
temperature - field dependences of @xmath36-position for samples s1 ( green lines ) and s2 ( blue lines ) , right after annealing ( curves 1,2 ) and after long - term aging ( curves 3,4),width=307 ] as one can see from table [ tab:1 ] the value of the critical temperature @xmath25 for sample s1 changed from 90.2 k to 89.8 k and for sample s2 changed from 91.3 k to 91.1 k. this indicates the optimal oxygen annealing process and aging effects are not caused by changing in @xmath38 ( especially for sample s2 ) . according to the estimations performed in @xcite the surface barriers are more important at high temperatures while the bulk pining is more important at low temperatures .
as one can see from fig .
[ fig:4 ] the position of @xmath36 significantly evolves after long - term aging .
the maximum field @xmath9 for sample s1 decreases from 4.25 t to 3.8 t and the minimum temperature @xmath39 increases from 37.5 k to 48 k ( curves 1 and 3 in fig . [ fig:4 ] ) .
such changes can be induced by the redistribution of pinning centers in the crystal or , in another words , bulk pinning changes during aging .
constancy of the maximum temperature @xmath40 for s1 corresponds to the influence of surface barriers at high temperatures .
the aging effect for sample s2 is not so significant as for s1 .
it is related to the presence of considerable mechanical tensions in the sample ( large value of an additional parameter @xmath411.5 k ) that prevent the redistribution of pinning centers .
larger value of @xmath9 ( for s2 as compared with s1 ) corresponds to the presence of additional pinning centers on silver atoms @xcite .
a possible reason of the decrease of @xmath9 and increase of @xmath39 for s1 is the decreasing number of dislocations . in another words ,
the crystal s1 becomes more homogeneous after aging .
in summary , mg ybco samples were investigated by measuring @xmath0-curves . on these curves
fe was observed is in a wide temperature range .
the origin of the fe existence is a significant influence of both surface barrier and bulk pinning .
bulk pinning influences parameters @xmath9 and @xmath39 while surface barrier influences @xmath40 .
undoped ybco is characterized by a significant aging effect which is manifested in a considerable change of @xmath9 and @xmath39 .
this corresponds to the change of bulk pinning . for ag - doped ybco
there are no significant changes in @xmath9 , @xmath39 and @xmath40 .
such difference is caused by the redistribution of pinning centers in undoped ybco while ag - doped ybco is characterized by huge mechanical tensions which prevent redistribution .
possible origin of redistribution for s1 is a decrease in number of dislocations , i.e. sample becomes more homogenous .
the work was partly supported by the slovak research and development agency ( no .
apvv-0006 - 07 ) and the slovak grant agency vega ( no.1/0159/09 ) the financial support of u.s.steel - dz energetika koice is acknowledged .
one of author ( d.l . )
is very thankful to the national scholarship program of slovak republic ( saia ) for the financial support during his stay in slovakia . | @xmath0-curves were experimentally investigated .
fishtail - effect ( fe ) was observed in mg yba@xmath1cu@xmath2o@xmath3 and yba@xmath1cu@xmath4ag@xmath5o@xmath3 ( at @xmath6 0.02 ) crystals in a wide temperature range 40 k @xmath7 75 k at the orientation of magnetic field @xmath8 .
it was obtained that the influence of bulk pinning on fe is more effective at low temperatures while the influence of surface barriers is more effective at high temperatures .
the value @xmath9 for ag - doped crystals is larger than for a pure one that due to the presence of additional pinning centers , above all on silver atoms . |
the human bronchial tree can be seen as a set of dyadically connected pipes , which sums up to 23 bifurcation levels from the trachea to terminal branches , on which gas exchanges occur .
thanks to poiseuille s law for a pipe , which states a proportionalty relation between air flow rate and pressure jump , it behaves as a fluid conductor like a resistive network , where pressure at bifurcating nodes plays the role of the electric potential , and air flow the role of electric intensity .
schematically , air is driven to the alveoli ( zone of gas exchange with the blood ) by a negative pressure maintained on the outlets during inspiration .
in the situation where pressure is a constant @xmath1 and the tree is regular ( i.e. resistances are the same for all pipes of a given generation ) , the overall process follows a generalized poiseuille law ( fluid counterpart of ohm s law ) @xmath2 where @xmath3 is the global resistance . yet , as soon as pressures are not uniform , or if the tree is no longer regular ( which can happen for example during an asthma crisis ) , this scalar ohmic law has be to extended . for the idealized tree with @xmath4 outlets , this law takes the form of a linear ( if atmospheric pressure is set at @xmath5 ) relation between a collection of @xmath4 values for the pressure , and a vector of fluxes .
the corresponding matrix @xmath3 can be written explicitly as a function of the resistances ( see @xcite ) . in a recent paper @xcite , a model of the respiratory system as an infinite tree was proposed .
this extrapolation of the finite resistive tree to an infinite one is natural because the actual bronchial tree exhibits some geometric regularity .
indeed , it is shown in @xcite that dimensions of the pipes ( at least in the conducting , central part of the tree , i.e. between generations 6 and 17 ) progress in a geometric way , with a reduction factor @xmath6 close to @xmath7 . as the poiseuille resistance of a pipe scales like @xmath8 ( where @xmath9 is the length and @xmath10 the radius ) , resistance at generation @xmath11 scales like @xmath12 , with @xmath13 . for the extrapolated infinite tree ,
as generation @xmath11 contains @xmath14 resistances in parallel , one obtains a finite global resistance @xmath15 and this finiteness is stable under small perturbations of the geometry .
next step consists in replacing the collection of discrete pressures at the end of the actual tree by a `` continuous '' field ( a function defined over the uncountable set of ends ) .
a first way to identify the corresponding trace space is immediate .
denoting by @xmath16 the set of vertices of tree @xmath17 , @xmath18 the pressure at vertex @xmath19 , writing @xmath20 for connected vertices and @xmath21 the resistance of the corresponding edge , sobolev space @xmath22 over @xmath17 is defined as all those pressure fields such that @xmath23 note that the previous definition makes sense for any kind of network with bounded connectivity : the quantity above simply represents the instantaneous rate of dissipated energy by viscous effects . denoting by @xmath24
the closure of finitely supported fields , a first expression of the trace space is the quotient @xmath25 . as detailed in @xcite ( in a probabilistic framework ) , non triviality of this space , which is equivalent to @xmath26 , indicates that `` something happens '' at infinity , i.e. it does make sense to prescribe certain pressure fields at infinity to drive some fluid through the tree . in order to obtain a more explicit description of the space of trace functions ( pressure fields on the set of ends ) , an explicit construction of a hilbert basis of harmonic , finite energy functions
is proposed in @xcite .
such functions are in some sense piecewise constant at infinity , so that their trace can be defined canonically .
general trace theorems then follow in a standard way by density . following a suggestion by p. colmez
, we propose here an alternative approach , based on the identification of the boundary of the tree ( set of ends ) to the ring of 2-adic integers @xmath27 , and a different strategy to define traces . to any pressure field
@xmath28 defined over the tree , we shall consider its restriction @xmath29 to the @xmath11-th generation , consider the corresponding function @xmath30 in the schwartz space of @xmath0 ( constant over each bunch of leafs stemming from any of the @xmath11-th generation vertices ) , and show that this sequence converges in @xmath31 , and possibly in a stronger way under some condition on the resistance growth .
regularity of the corresponding trace function will be expressed in terms of behaviour near infinity of the fourier transform ( sobolev - like regularity ) .
the approach presented here may seem excessively and artificially abstract considering the actual object it aims at reproducing in some way .
indeed , we are aware that some parts of what we present can be ( and actually have been in some cases ) presented in a different and less abstract setting , in particular 1 .
regularity of the trace functions ; 2 .
expression of the neuman - dirichlet operator , which maps flow fields onto pressure fields . concerning @xmath32 , functions over the set of ends of the tree can be described in terms of @xmath33 regularity , as presented in @xcite , without any reference to 2-adic integers .
yet , as we will see , the @xmath27 approach allows a very natural and concise way to define sobolev spaces in the fourier setting , which enlights the deep similarities with trace operators in the partial differential equations context .
as for the neuman dirichlet operator , it can be represented by matrices for finite tree ( see @xcite ) , and by kernel operators for infinite tree ( see @xcite ) .
we will present here how the @xmath27 approach makes it possible to express them as a convolution by a riesz kernel with a exponent directly related to the geometric growth of resistances .
this paper is structured as follows : in section [ sec : ident ] , we present the identification between the set of ends of a dyadic tree and @xmath0 , and we give a first trace theorem based on this identification . in section [ sec : sob ] we give some regularity properties of trace functions ( sobolev regularity ) .
section [ sec : dn ] is dedicated to dirichlet- neuman and neuman - dirichlet operators , and in section [ sec : embed ] we investigate the possibility to imbed the end of the infinite tree onto a domain of @xmath34 ( actual domain occupied by a real lung ) .
finaly we gather in the appendices some facts on fourier analysis and sobolev spaces on @xmath27 . in what follows the ideal dyadic infinite geometric tree , with resistances following a geometric growth in @xmath12 ( with @xmath35 close to @xmath36 for a healthy lung , as stated before ) will play a central role . yet , we shall present results with maximal generality , allowing when it is possible non - regular trees ( non - uniform resistances within a generation ) .
we first gather some definitions and standard properties of 2-adic numbers ( see e.g. @xcite ) . for any @xmath37 , @xmath38 , with
@xmath39 odd , one defines valuation @xmath40 as @xmath41 .
one extends this definition to rational numbers by setting @xmath42 for any @xmath43 , @xmath44 , and @xmath45 . now setting @xmath46 , the 2-adic distance over @xmath47
is defined as @xmath48 this distance is ultrametric : it verifies a strong triangle inequality @xmath49 as a consequence , a ball is centered at any of its elements .
the complete closure of @xmath47 for this distance is called @xmath50 .
any element of @xmath50 can be identified to a series @xmath51 where @xmath52 , @xmath53 , and therefore written ( here in the case @xmath54 ) as @xmath55 the ring @xmath27 of 2-adic integers is defined as the closed unit ball of @xmath50 . in the above representation , it corresponds to the case where @xmath56 for all @xmath57 .
let now explain how the set of ends @xmath58 of a dyadic tree @xmath17 can be identified with @xmath27 .
the set of vertices can be seen as the disjoint union of the @xmath59 s , for @xmath60 , @xmath61 , @xmath62 , as illustrated by fig .
[ fig : tree ] .
we shall denote by @xmath63 the vertex @xmath64 at generation @xmath11 ( i.e. @xmath63 is @xmath64 considered as an element of @xmath59 ) . now denoting by @xmath65 , with @xmath66 , the canonical surjection from @xmath67 onto @xmath59 , the set of edges of @xmath17 , namely @xmath68 , consists of all those couples @xmath69 such that @xmath70 .
the corresponding edge is denoted by @xmath71 .
the set @xmath72 of ends of @xmath17 ( infinite paths toward infinity ) can be represented by the projective limit of the system @xmath73 : @xmath74 this set is naturally identified to @xmath27 : any sequence @xmath75 is uniquely associated to a sequence @xmath76 with @xmath77 such that @xmath78 and therefore to @xmath79 . note that , for two ends @xmath19 and @xmath80 , the 2-adic distance measures their proximity with regards to the tree , more precisely @xmath81 is the index of the generation at which the corresponding paths splitted . [
fig : tree ] as @xmath16 , the set of vertices of @xmath17 , identifies with the disjoint union of the @xmath82 s , any pressure field @xmath83 can be seen as a sequence @xmath84 , with @xmath85 .
we define a `` piecewise constant '' function @xmath30 over @xmath27 as @xmath86 as an example , consider a field @xmath28 which takes values @xmath87 and @xmath61 at @xmath88 and @xmath89 , respectively .
then @xmath90 is @xmath87 on @xmath91 and @xmath61 on @xmath92 .
note that @xmath30 lies in the schwartz space @xmath93 ( see definition [ def : s ] in the appendix ) .
the trace of @xmath28 on @xmath94 will be defined as the limit of @xmath30 as @xmath11 go to infinity , whenever it exists in some sense .
we consider from now on a dyadic resistive tree @xmath96 , where @xmath10 is the collection of edge resistances @xmath97 .
@xmath98 let us denote by @xmath99 the set of all those functions with finite dirichlet energy @xmath100 with the semi - norm @xmath101 note that in the case @xmath102 is constant in each generation ( @xmath103 at generation @xmath11 ) , the @xmath104 semi - norm reads @xmath105 where @xmath30 is defined by ( [ eq : deftp ] ) and where @xmath31 is defined with respect the haar measure @xmath106 , see definition [ def : haar ] in the appendix .
it is a direct consequence of the fact that @xmath107 .
the following proposition allows to define a trace of @xmath104 functions over @xmath17 as soon as some condition on the resistances is met : [ prop : tracel2 ] let @xmath108 be a resistive dyadic tree , with @xmath109 .
assume @xmath110 then @xmath111 ( defined by ( [ eq : deftp ] ) ) converges strongly to some @xmath112 .
the linear operator @xmath113 is in @xmath114 .
it suffices us to show that @xmath115 using cauchy - schwarz inequality with coefficients @xmath116 ( to be chosen later ) , we have : @xmath117 for the first term , it comes from the definition @xmath118 _
i=0 ^2^n-1 |p(x_n^i)-p ( _ n-1^n(x_n^i ) ) |^2 2 ^ -n.@xmath119 hence , @xmath120 we choose the coefficients @xmath116 such that this last quantity is bounded by @xmath121 , i.e. such that @xmath122 with this particular choice , the first term in ( [ cauchy ] ) is bounded by @xmath121 .
moreover the second term in ( [ cauchy ] ) is finite thanks to ( [ eq : condrl2 ] ) and @xmath123 from which we deduce inequality ( [ cont : trace ] ) and the proposition .
we have a similar result for the @xmath124 space instead of @xmath125 . for @xmath126 , if @xmath127^{p/2 } < + \infty \label{eq : condrlp}\ ] ] then @xmath128 converges strongly in @xmath124 .
operator @xmath129 belongs to @xmath130 .
condition ( [ eq : condrlp ] ) is met as soon as @xmath131 with @xmath132 .
as for the infinite version of the actual human lungs , resistances vary like @xmath12 with @xmath133 , so that such a trace operator can be defined properly in @xmath124 for @xmath134 .
as in the case of sobolev spaces over domains in @xmath34 , it is natural to expect some regularity of functions in @xmath135 . it will be expressed in terms of @xmath136 regularity .
as the functions we consider here are restricted to @xmath27 , standard fourier transform @xmath137 does not depend on the integer part of @xmath138 ( i.e. @xmath139 as soon as @xmath140 ) . as a consequence ,
the appropriate notion is a reduced fourier transform ( in the same way fourier transform is replaced by fourier series for periodic functions over @xmath141 ) .
this reduced fourier transform , or fourier series , is defined as @xmath142 where @xmath143 runs over @xmath144 which identifies to @xmath145 $ ] .
sobolev space of index @xmath146 is then defined as the set of all those functions such that @xmath147 is finite .
note that both notions are consistent : given @xmath148 , if we define @xmath149 as the standard fourier transform of the extension of @xmath150 by @xmath5 on @xmath151 , we have @xmath152 as soon as @xmath153 .
we refer the reader to appendix [ sec : fourier ] for more details on the underlying framework .
[ proptrhs ] let @xmath108 be a resistive dyadic tree , with @xmath109 .
assume @xmath154 for some @xmath155 .
then @xmath156 ( defined by prop .
[ prop : tracel2 ] ) maps continuously @xmath99 onto @xmath157 .
we follow the same ideas as for proposition [ prop : tracel2 ] , proving @xmath158 in order to check this claim , we have to estimate the sobolev norm of @xmath159{\bf 1}_{j+ 2^{n}\zz}.\ ] ] we fix @xmath11 and write @xmath160 . using @xmath161 ( see prop . [ prop : ex ] in appendix ) and a change of variable , we get @xmath162 hence for @xmath163 , we obtain @xmath164 moreover for @xmath165 , we directly have @xmath166 finally , it comes @xmath167\left[p(x_n^{j'})-p\left ( \varphi_{n-1}^n(x_n^{j'})\right ) \right ] \frac{4^{-n } } { |j - j'|_2^{1 + 2s } } \\ & + \sum_{j } \left[p(x_n^j)-p\left ( \varphi_{n-1}^n(x_n^j)\right ) \right]^2 2^{n(-1 + 2s)}.\end{aligned}\ ] ] splitting the first sum with @xmath168 , it comes two symmetrical sums . then using @xmath169 we prove that @xmath170^ 2 2^{n(-1 + 2s)}.\ ] ] then we conclude the proof by the same way as for proposition [ prop : tracel2 ] .
let us treat the particular case of a finite resistance sub - geometrical tree : the resistances satisfy @xmath171 for some parameter @xmath172 .
assumption ( [ eq : condrhs ] ) is satisfied if and only if we are in the sub - critical case : @xmath173 for @xmath174 , we can obtain a bound for the convergence velocity of @xmath30 to @xmath175 .
more precisely , @xmath176 the following theorem gives a positive result in the critical case @xmath177 for the sub - geometrical trees ( we have a positive result of convergence without a precise estimate of the convergence velocity ) let us consider a sub - geometrical tree @xmath178 with @xmath179
. then @xmath156 ( defined by prop .
[ prop : tracel2 ] ) maps continuously @xmath99 onto @xmath180 , with @xmath181 .
we claim first that for every function @xmath182 , the trace @xmath175 belongs to the limit space @xmath183 and let us first conclude . from that , the linear operator @xmath156 is acting from @xmath99 to @xmath183 . since the previous continuity from @xmath99 in @xmath31 , it is easy to check that the graph of @xmath156 is closed in @xmath184
. then banach s theorem of closed graph implies the desired continuity of @xmath156 .
+ it remains to prove that @xmath156 maps @xmath104 onto @xmath183 .
let us fix a function @xmath182 .
we are going to show that the sequence @xmath128 is cauchy in @xmath183 .
let @xmath185 be two integers and write @xmath186 } } { \bf 1}_{j+ 2^m{\mathbb z}_2}(x).\ ] ] [ fig : ij ] hence @xmath187 } } \left [ p(x_n^i ) -p(x_m^j ) \right ] { \bf 1}_{j+ 2^m{\mathbb z}_2}(x).\ ] ] as previously ( see fig . [
fig : ij ] for the meaning of notations ) , we get @xmath188 } } \sum_{i'=0}^{2^n - 1 } \sum_{\genfrac{}{}{0pt}{}{0\leq j'<2^m}{j'\equiv i ' \ ; [ 2^n ] } } \ \left [ p(x_n^i)-p(x_m^j ) \right]^ 2 \frac{4^{-m } } { |j - j'|_2^{1 + 2s_\alpha } } \\ & \hspace{-2 cm } \lesssim \sum_{i=0}^{2^n - 1 } \ \sum_{\genfrac{}{}{0pt}{}{0\leq
j < 2^m}{j\equiv i \ ; [ 2^n ] } } \left [ p(x_n^i)-p(x_m^j ) \right]^ 2 2^{m(-1 + 2s_\alpha)}.\end{aligned}\ ] ] to control @xmath189 , we use triangle inequality on the telescopic series along the path between @xmath190 and @xmath191 , as follows @xmath192^ 2 & = \left [ \sum_{k = n}^{m-1 } p(\varphi_k^m(x_m^j))-p(\varphi_{k+1}^m(x_m^j ) ) \right]^ 2 \\ & \lesssim \left [ \sum_{k = n}^{m-1 } \frac{\left| p(\varphi_k^m(x_m^j))-p(\varphi_{k+1}^m(x_m^j ) ) \right|^2}{r_{k+1}^{j } } \right ] \alpha^m.\end{aligned}\ ] ] we have used cauchy - schwarz inequality at the last step , the assumption of sub - geometric increasing of the resistances , the fact that @xmath193 .
note that @xmath194 , which is a priori not defined as @xmath195 might be larger than @xmath196 , represents obviously @xmath197 , where @xmath198 , with @xmath199 $ ] . by definition of @xmath200 , we deduce @xmath201 } } \left [ \sum_{k = n}^{m-1 } \
\frac { \left|p(\varphi_k^m(x_m^j))-p(\varphi_{k+1}^m(x_m^j ) ) \right| ^2}{r_{k+1}^{j } } \right ] \alpha^m 2^{m(-1 + 2s_\alpha ) } \\ & \hspace{-2 cm } \lesssim \sum_{i=0}^{2^n - 1 } \
\sum_{\genfrac{}{}{0pt}{}{0\leq j <
2^m}{j\equiv i \ ; [ 2^n ] } } \frac{\left|p(\varphi_k^m(x_m^j))-p(\varphi_{k+1}^m(x_m^j ) ) \right|^2}{r_{k+1}^{j } } \\ & \hspace{-2 cm }
\lesssim \| p \|_{h^1(t_n)}^2,\end{aligned}\ ] ] where @xmath202 is the subnetwork corresponding to the the set of generations @xmath64 with @xmath203 .
as @xmath204 goes to @xmath5 , we have also proved that @xmath128 is a cauchy sequence in @xmath183 so that @xmath175 belongs to this space .
given a pressure field on @xmath0 ( seen as the set of ends of the tree @xmath108 ) , we are interested in the fluxes which it drives through @xmath108 ( dirichlet problem ) , and in particular in the quantity of air which exits the tree through its boundary , which amounts to solve the dirichlet to neuman ( dn ) problem associated with the tree . the reciprocal ( nd )
mapping is straightforward to obtain , as detailed in @xcite . indeed
as soon as the global flux is known , it identifies with the flux through the first edge , which gives the pressure at generation @xmath5 ( pressure at the root is @xmath5 ) .
all pressures can be computed recursively in a similar manner , for the flux through any subtree is known . following this procedure in the case of a regular tree ( resistance @xmath103 at generation @xmath11 ) , given a flux field @xmath205
( the regularity of which will be addressed later ) , pressure at end @xmath206 writes formally @xmath207 where @xmath208 is the cumulated resistance .
if one assumes geometric growth of the resistances according to some @xmath209 , more precisely @xmath210 one obtains @xmath211 , and consequently @xmath212 where @xmath213 is the riesz kernel on @xmath0 ( see section [ sec : riesz ] for a brief presentation of these multipliers ) @xmath214 and @xmath215 is the local zeta function .
( neuman - dirichlet operator ) + the riesz operator @xmath216 ( corresponding to the convolution by @xmath217 on @xmath0 ) maps continuously @xmath218 onto @xmath219 , with @xmath220 . from a functional point of view
, we detail in section [ sec : riesz ] that the riesz operators @xmath216 corresponds to some power of a laplacian operator @xmath221 .
so it is natural to expect that it maps continuously @xmath218 onto @xmath222 for all exponent @xmath223 ( the desired result is a particular case of this property ) .
let us give a more detailed proof .
we refer the reader to section [ sec : sobolevzz ] for a presentation of sobolev spaces on @xmath0 .
thanks to ( [ eq : sobzz ] ) we have @xmath224 with by convention @xmath225 for @xmath226 . moreover , ( [ rel : fourier2 ] ) gives @xmath227 so it comes @xmath228 for @xmath229 , by convention @xmath225 so @xmath230 for @xmath231 then @xmath232 so @xmath230 we also conclude that as expected @xmath233 which for @xmath234 gives us the desired estimate . according to corollary [ cor : riesz ] , we know that the riesz multiplier @xmath216 is invertible in distributional sense or in @xmath95 sense and @xmath235 .
it gives an explicit expression of the dirichlet - neuman operator for the regular tree , i.e. the mapping @xmath236 which is the core of the ventilation process .
one expected outcome of this approach is to provide a sound functional framework for the coupling of a resistive tree with an elastic medium onto which it is embedded ( see @xcite for a first coupled model in one - dimension , and @xcite for an homogenized description of a foamy like medium ) .
we must say that the 2-adic viewpoint does not allow to improve significantly the results which are presented in @xcite on this matter . considering a mapping from @xmath0 onto a domain in @xmath34 ( which models the way our infinite tree is imbedded on the physical space , i.e. the actual lung )
, we simply give here a property which allows to describe how sobolev regularity of a function on the domain can be transported back to its @xmath0 counterpart , as soon as some hlder regularity of the mapping is verified .
let @xmath237 be an open set of @xmath34 ( equipped with its euclidean structure and lebesgue measure ) with @xmath238 , and @xmath239 a measure - preserving mapping from @xmath0 onto @xmath237 : for every measurable set @xmath240 @xmath241 we then define the following operator : @xmath242 [ iso ] for all exponent @xmath243 $ ] , @xmath244 continuously acts from @xmath245 to @xmath246 .
more precisely @xmath244 is an isometry : @xmath247 assume first that @xmath248 .
using @xmath249 , we obtain immediately @xmath250 the @xmath251 case follows by having @xmath28 go to infinity .
we are now looking for condition on @xmath239 such that @xmath244 keeps some regularity .
[ injreg ] if @xmath252 is @xmath253-hlderian , then @xmath244 is continuous from @xmath254 to @xmath255 for all @xmath256 .
we use the characterization of sobolev spaces , given by proposition [ prop : sob ] @xmath257 the previous proposition yields @xmath258 the @xmath253-hlderian regularity of @xmath239 gives @xmath259 which implies @xmath260 then , we use again the previous proposition , to deduce @xmath261 using the well - known characterization of local sobolev spaces , we get the desired estimate @xmath262
[ def : haar ] the set @xmath263 endowed with its @xmath264-adic distance @xmath265 is a locally compact group .
it owns a haar measure @xmath106 which satisfies for every @xmath266 and @xmath267 @xmath268 this measure defines a probability measure on @xmath0 .
[ def : s ] the schwartz space @xmath269 is defined as the space of all those functions which are compactly supported and locally constant .
it is spanned by characteristics functions of balls @xmath270 , @xmath271 , @xmath272 .
[ def : ft ] for any @xmath273 , its fourier transform is defined by @xmath274 characteristic functions of closed balls containing ( i.e. centered at ) @xmath5 play the role of central gaussian distributions in @xmath34 : [ prop : ex ] we have @xmath275 in particular @xmath276 .
[ prop : formule ] we have @xmath277 and @xmath278 the fourier transform on @xmath263 satisfies to the same properties than the ones on @xmath279 : [ thm : plancherel ] the fourier transform is an isometry on @xmath280 : for all @xmath281 @xmath282 consequently , fourier transform can be extended by density to a continuous operator over @xmath280 .
[ thm : inversion ] the fourier transform is invertible on @xmath280 and , for any function @xmath283 , @xmath284 we would like to finish this section by describing the theory of fourier series . on the euclidean space @xmath285 ,
it is more convenient to use fourier series for functions supported on @xmath286 $ ] , similarly we can define fourier series for functions supported on @xmath0 .
this new operation is denoted by @xmath287 , we follow the same scheme as in the euclidean framework with identifying @xmath285 to @xmath263 and @xmath286 $ ] to @xmath0 .
we shall denote the countable set @xmath288 by @xmath289 .
let @xmath290 and @xmath291 a function supported on @xmath0 .
we define @xmath292 we note that @xmath19 belonging to @xmath0 , the previous quantity is well - defined for @xmath293
. then we have the following properties ( coming from those of the whole fourier transform ) : for @xmath291 , we have @xmath294 and @xmath295
similarly to what is done on the euclidean space , we define sobolev spaces on @xmath0 and the equivalent of the @xmath296 approximate spaces used in article @xcite ( see @xcite for the definition and main properties of the @xmath33 spaces in the euclidean case ) .
[ def : sobolev ] for any @xmath297 , we define for a schwartz function @xmath298 @xmath299 then @xmath300 is a norm and we define the sobolev space @xmath219 as the completion of @xmath301 for this norm . then for @xmath256 , @xmath218 can be identified to the dual space @xmath302 .
the sobolev spaces are hilbert spaces . in the previous section
, we have seen that for functions defined on @xmath0 , we can use fourier series instead of fourier transform . for such functions ,
if @xmath303 then @xmath304 else the quantity @xmath305 depends only on the class of @xmath306 and @xmath307 consequently , we have the following representation of the sobolev norms , using fourier series . for a function @xmath150 supported on @xmath0 @xmath308 with by convention @xmath225 for @xmath226 .
regularity estimates in @xcite rely on @xmath33 norms ( see @xcite for a full description of this general setting ) . although it does not play a central role in the present approach , we present here an equivalent definition of sobolev spaces based on this definition of regularity .
let @xmath309 , @xmath310 and let @xmath311 be the projector onto @xmath312 for the scalar product of @xmath313 .
let @xmath314 .
we define the @xmath315 space by @xmath316 the norm associated to this space is given by @xmath317 the following proposition establishes the link between @xmath33 regularity and sobolev regularity ( fourier setting ) .
[ pr2 ] let @xmath314
. then the following identification holds @xmath318 let @xmath223 . using the plancherel s formula ( see theorem [ thm : plancherel ] ) , we obtain that for all @xmath319 @xmath320 but , if @xmath321 , then for all @xmath322 @xmath323 indeed , the first part of property ( [ 1 ] ) is a direct consequence of the explicit formula of @xmath324 ( see the proof of proposition [ proptrhs ] ) .
let @xmath309 and @xmath325 . to obtain the second part of property ( [ 1 ] ) , it is enough to prove that for all functions @xmath321 such that @xmath326 with @xmath327 we have @xmath328 let @xmath329 .
then for all @xmath330 @xmath331 and so @xmath332 which ends the proof of ( [ 1 ] ) .
applying ( [ 1 ] ) , we obtain @xmath333 we deduce that @xmath334 and so @xmath335 where @xmath336 making the change of variables @xmath337 , we obtain @xmath338 which ends the proof of proposition [ pr2 ] . without requiring fourier transform or frequential decomposition , we have a more geometrical characterization of sobolev spaces ( with positive index ) : [ prop : sob ] for all @xmath223 and @xmath291 , @xmath339 considering the double integral and properties of the haar measure @xmath106 , a change of variables yields @xmath340 then according to plancherel s inequality ( theorem [ thm : plancherel ] ) , we get @xmath341 consequently @xmath342 with @xmath343 it also suffices to prove @xmath344 first if @xmath303 , then @xmath345 ( as @xmath346 would be an integer in the integral ) ; ( [ eq : equiv ] ) holds since @xmath347 .
+ else for @xmath348 , we denote by @xmath28 the negative integer satisfying @xmath349 , which is equivalent to @xmath350 . for all @xmath351 with @xmath352 , @xmath353 and so @xmath354 .
hence @xmath355 where we used that for a non positive integer @xmath356 @xmath357 due to props .
[ prop : ex ] and [ prop : formule ] . that concludes the proof of ( [ eq : equiv ] ) .
we recall here some well - known properties concerning particular multipliers , namely riesz kernels ( see @xcite for more details ) . for @xmath358 , we consider the following function defined on @xmath263 @xmath359 and , for @xmath360 , @xmath361 where @xmath362 stands for the principal value based on the following representation formula : @xmath363 d\mu(x),\ ] ] and @xmath364 is the local zeta function : @xmath365 we denote by @xmath366 the multiplier defined as the convolution operator by @xmath367 .
we have a precise description in the frequency space of the symbols associated to these multipliers : in the distributional sense , we have @xmath368 we refer the reader to section 3 of @xcite for detailed proofs of such results .
it rests on two ideas : in the one hand holomorphic properties of the map @xmath369 and in the other hand the direct calculus for @xmath370 @xmath371 \quad \hbox{(thanks
to~(\ref{eq : intspheres}))}\\ & = |\xi|_2^{-\beta } \left[\frac{1}{2(1 - 2^{-\beta } ) } -\frac{2^\beta}{2}\right ] = |\xi|_2^{-\beta } \frac{1 - 2^{\beta-1}}{1 - 2^{-\beta } } = |\xi|_2^{-\beta } \frac{\zeta(\beta)}{\zeta(1-\beta)}. \end{aligned}\ ] ] by analogy with the euclidean case , we can define a positive laplacian operator @xmath372 on @xmath263 via the frequency space , as follows : @xmath373 then , we emphasize that the riesz operators can be also considered as a power of the laplacian : @xmath374 .
we refer the reader to @xcite for more details about @xmath28-adic pseudo - differential theory .
we deduce also the following properties ( which can be obtained by a direct approach , see @xcite ) : the riesz operators satisfy to the semigroup property ( also known as `` riesz reproduction formula '' ) : for exponents @xmath375 such that @xmath376 @xmath377 and in particular @xmath378 using the notion of fourier series ( developed in section [ sec : fourier ] ) , we can define multipliers on @xmath0 as follows : for @xmath358 , we define on @xmath0 @xmath379 and for @xmath360 @xmath380 we write @xmath381 the multiplier operator on @xmath0 defined as the convolution ( on @xmath0 ) operator by @xmath217 . we have the following fourier representation : for all @xmath290 @xmath382 by definition , we have @xmath383(\lambda ) = \widehat{k^\beta{\bf 1}_{\zz}}(\lambda).\ ] ] then ( [ rel : fourier2 ] ) follows from the proof of ( [ rel : fourier ] ) . [ cor : riesz ] we deduce also that for exponents @xmath375 with @xmath384 : @xmath385 in particular @xmath386 where all these operators are defined on @xmath0 .
we would like to thank albert cohen for his suggestions and fruitful remarks , and pierre colmez for bringing to our attention the natural identification between the set of ends of the tree and the ring of 2-adic integers .
s. albeverio , a. yu . khrennikov and v. m. shelkovich , _ harmonic analysis in the @xmath28-adic lizorkin spaces : fractional operators , pseudo - differential equations , @xmath28-adic wavelets , tauberian theorems _ , j. four .
anal . and appl .
* 12 * ( 2006 ) , no .
4 , 393425 . | we propose here a general framework to address the question of trace operators on a dyadic tree .
this work is motivated by the modeling of the human bronchial tree which , thanks to its regularity , can be extrapolated in a natural way to an infinite resistive tree .
the space of pressure fields at bifurcation nodes of this infinite tree can be endowed with a sobolev space structure , with a semi - norm which measures the instantaneous rate of dissipated energy .
we aim at describing the behaviour of finite energy pressure fields near the end .
the core of the present approach is an identification of the set of ends with the ring @xmath0 of 2-adic integers .
sobolev spaces over @xmath0 can be defined in a very natural way by means of fourier transform , which allows us to establish precised trace theorems which are formally quite similar to those in standard sobolev spaces , with a sobolev regularity which depends on the growth rate of resistances , i.e. on geometrical properties of the tree .
furthermore , we exhibit an explicit expression of the `` ventilation operator '' , which maps pressure fields at the end of the tree onto fluxes , in the form of a convolution by a riesz kernel based on the 2-adic distance .
frdric bernicot bertrand maury delphine salort |
supersymmetry is at present the only known framework in which the higgs sector @xcite of the standard model ( sm ) , so crucial for its internal consistency , is natural @xcite .
the minimal version of the supersymmetric standard model ( mssm ) contains two higgs doublets @xmath15 with opposite hypercharges : @xmath16 , @xmath17 , so as to generate masses for up- and down - type quarks ( and leptons ) , and to cancel gauge anomalies . after spontaneous symmetry breaking induced by the neutral components of @xmath18 and @xmath19 obtaining vacuum expectation values , @xmath20 , @xmath21 , @xmath22 , the mssm contains two neutral @xmath0-even ( @xmath1 , @xmath3 ) , one neutral @xmath0-odd ( @xmath5 ) , and two charged ( @xmath23 ) higgs bosons @xcite .
because of gauge invariance and supersymmetry , all the higgs masses and the higgs couplings in the mssm can be described ( at tree level ) in terms of only two parameters , which are usually chosen to be @xmath24 and @xmath25 , the mass of the @xmath0-odd higgs boson . in particular , all the trilinear self - couplings of the physical higgs particles can be predicted theoretically ( at the tree level ) in terms of @xmath25 and @xmath24 .
once a light higgs boson is discovered , the measurement of these trilinear couplings can be used to reconstruct the higgs potential of the mssm .
this will go a long way toward establishing the higgs mechanism as the basic mechanism of spontaneous symmetry breaking in gauge theories .
although the measurement of all the higgs couplings in the mssm is a difficult task , preliminary theoretical investigations by plehn , spira and zerwas @xcite , and by djouadi , haber and zerwas ( dhz ) @xcite , of the measurement of these couplings at the lhc and at a high - energy @xmath26 linear collider , respectively , are encouraging .
we have considered in detail @xcite the question of possible measurements of the trilinear higgs couplings of the mssm at a high - energy @xmath26 linear collider .
we assume that such a facility will operate at an energy of 500 gev with an integrated luminosity per year of @xmath27 @xcite .
( this is a factor of 10 more than the earlier estimate . ) in a later phase one may envisage an upgrade to an energy of 1.5 tev .
the trilinear higgs couplings that are of interest are @xmath9 , @xmath10 , and @xmath28 , involving both the @xmath0-even and @xmath0-odd higgs bosons .
the couplings @xmath9 and @xmath10 are rather small with respect to the corresponding trilinear coupling @xmath29 in the sm ( for a given mass of the lightest higgs boson @xmath30 ) , unless @xmath30 is close to the upper value ( decoupling limit ) .
the coupling @xmath28 remains small for all parameters . throughout ,
we include one - loop radiative corrections @xcite to the higgs sector in the effective potential approximation .
in particular , we take into account the parameters @xmath5 and @xmath31 , the soft supersymmetry breaking trilinear parameter and the bilinear higgs(ino ) parameter in the superpotential , respectively , and as a consequence the left right mixing in the squark sector , in our calculations .
we thus include all the relevant parameters of the mssm in our study @xcite .
related work has recently been presented by dubinin and semenov @xcite . for a given value of @xmath30
, the values of these couplings significantly depend on the soft supersymmetry - breaking trilinear parameter @xmath5 , as well as on @xmath31 , and thus on the resulting mixing in the squark sector .
since the trilinear couplings tend to be small , and depend on several parameters , their effects are somewhat difficult to estimate .
the dominant source of multiple production of the higgs ( @xmath1 ) boson , is through higgs - strahlung of @xmath3 , and through production of @xmath3 in association with the @xmath0-odd higgs boson .
this source of multiple production can be used to extract the trilinear higgs coupling @xmath9 .
the non - resonant fusion mechanism for multiple @xmath1 production , @xmath32 , involves two trilinear higgs couplings , @xmath9 and @xmath10 , and is useful for extracting @xmath10 .
at the tree level , the higgs sector of the mssm is described by two parameters , which can be conveniently chosen as @xmath25 and @xmath24 @xcite .
there are , however , substantial radiative corrections to the @xmath0-even neutral higgs masses and couplings @xcite .
they are , in general , positive , and they shift the mass of the lightest mssm higgs boson upwards .
the higgs mass falls rapidly at small values of @xmath24 .
since the lep experiments are obtaining lower bounds on the mass of the lightest higgs boson , they are beginning to rule out significant parts of the small-@xmath24 parameter space , depending on the model assumptions .
aleph finds a lower limit of @xmath33 gev , irrespective of @xmath24 , and a limit of @xmath34 gev for @xmath35 @xcite .
we take @xmath36 to be a representative value .
in units of @xmath37 , the relevant tree - level trilinear higgs couplings are given by @xmath38 with @xmath39 the mixing angle in the @xmath0-even higgs sector , which can be calculated in terms of the parameters appearing in the @xmath0-even higgs mass matrix .
the dominant one - loop radiative corrections are proportional to @xmath40 @xcite .
[ fig : lam - mh ] ( 15,8.5 ) ( -1.5,1 ) the trilinear couplings depend significantly on @xmath25 , and thus also on @xmath30 .
this is shown in fig .
[ fig : lam - mh ] , where we compare @xmath9 , @xmath10 and @xmath28 for three different values of @xmath24 , and the sm quartic coupling @xmath41 ( which also includes one - loop radiative corrections @xcite ) . at low values of @xmath30 ,
the mssm trilinear couplings are rather small .
for some value of @xmath30 the couplings @xmath9 and @xmath10 start to increase in magnitude , whereas @xmath28 remains small .
the values of @xmath30 at which they start becoming significant depend crucially on @xmath24 . for @xmath36 ( fig .
[ fig : lam - mh]a ) this transition takes place around @xmath42100 gev , whereas for @xmath43 the critical value of @xmath30 increases to 100110 ( see fig . [
fig : lam - mh]b ) . in this region ,
the actual values of @xmath9 and @xmath10 ( for a given value of @xmath30 ) change significantly if @xmath5 becomes large and positive .
a non - vanishing squark - mixing parameter @xmath5 is thus quite important .
also , for special values of the parameters , the couplings may vanish @xcite . to sum up the behaviour of the trilinear couplings , we note that @xmath9 and @xmath10 are small for @xmath44120 gev , depending on the value of @xmath24 .
however , as @xmath30 approaches its maximum value , which is reached rapidly as @xmath25 becomes large , @xmath45 gev , these trilinear couplings become large .
thus , as functions of @xmath25 , the trilinear couplings @xmath9 and @xmath10 are large for most of the parameter space .
we also note that , for large values of @xmath24 , @xmath9 tends to be relatively small , whereas @xmath10 becomes large , if also @xmath25 ( or , equivalently , @xmath30 ) is large .
the different mechanisms for the multiple production of the mssm higgs bosons in @xmath26 collisions have been discussed by dhz .
the dominant mechanism for the production of multiple @xmath0-even light higgs bosons ( @xmath1 ) is through the production of the heavy @xmath0-even higgs boson @xmath3 , which then decays via @xmath8 .
the heavy higgs boson @xmath3 can be produced by @xmath3-strahlung , in association with @xmath5 , and by the resonant @xmath46 fusion mechanism .
these mechanisms for multiple production of @xmath1 @xmath47 are shown in fig .
[ fig : feynman - resonant ] . all the diagrams of fig .
[ fig : feynman - resonant ] involve the trilinear coupling @xmath9 .
[ fig : feynman - resonant ] ( 16,7 ) ( 2,-12 ) a background to ( [ eq : res - hhh ] ) comes from the production of the pseudoscalar @xmath5 in association with @xmath1 and its subsequent decay to @xmath48 @xmath49 leading to @xmath50 final states .
a second mechanism for @xmath51 production is double higgs - strahlung in the continuum with a @xmath52 boson in the final state , @xmath53 we note that the non - resonant analogue of the feynman diagram of fig .
[ fig : feynman - resonant]b involves , apart from the coupling @xmath9 , the trilinear higgs coupling @xmath10 as well .
[ fig : feynman - nonres - ww ] ( 16,7 ) ( 1,-12 ) finally , there is a mechanism of multiple production of the lightest higgs boson through non - resonant @xmath46 fusion in the continuum ( see section 7 ) : @xmath54 it is important to note that all the diagrams of fig .
[ fig : feynman - resonant ] involve the trilinear coupling @xmath9 only . on the other hand , the non - resonant analogue of fig .
[ fig : feynman - resonant]b , and fig .
[ fig : feynman - nonres - ww]c involve both the trilinear higgs couplings @xmath9 and @xmath10 .
the dominant source for the production of multiple higgs bosons ( @xmath1 ) in @xmath26 collisions is through the production of the heavier @xmath0-even higgs boson @xmath3 either via higgs - strahlung or in association with @xmath5 , followed , if kinematically allowed , by the cascade decay @xmath8 .
the cross sections for these processes can be found in @xcite .
[ fig : sigma-500 - 1500 ] ( 15,8.5 ) ( -1.5,1 ) in fig .
[ fig : sigma-500 - 1500 ] we plot the cross sections for the @xmath55 centre - of - mass energies @xmath56 , as functions of the higgs mass @xmath57 and for @xmath58 . for large values of the mass @xmath25 of the pseudoscalar higgs boson ,
all the higgs bosons , except the lightest one ( @xmath1 ) , become heavy and decouple @xcite from the rest of the spectrum . at values of @xmath24 that are not too large , the trilinear @xmath59 coupling @xmath9 can be measured by the decay process @xmath8 , which has a width proportional to @xmath60 . however , this is possible only if the decay is kinematically allowed , and the branching ratio is sizeable . in fig .
[ fig : br - h - a ] we show the branching ratios ( at @xmath36 ) for the main decay modes of the heavy @xmath0-even higgs boson as a function of the @xmath3 mass .
apart from the @xmath51 decay mode , the other important decay modes are @xmath61 , @xmath62 .
we note that the couplings of @xmath3 to gauge bosons can be measured through the production cross sections for @xmath63 ; therefore the branching ratio @xmath64 can be used to measure the triple higgs coupling @xmath9 .
[ fig : br - h - a ] ( 15,8.5 ) ( -1.5,1 ) for increasing values of @xmath24 , the @xmath59 coupling gradually gets weaker ( see fig . [
fig : lam - mh ] ) , and hence the prospects for measuring @xmath9 diminish .
this is also indicated in fig .
[ fig : br - h - a ] , where we show the @xmath3 branching ratios for @xmath43 .
[ fig : hole ] ( 15,8.5 ) ( -1.5,1 ) there is actually a sizeable region in the @xmath25@xmath24 plane where the decay @xmath65 is kinematically forbidden .
this is shown in fig .
[ fig : hole ] , where we also display the regions where the @xmath65 branching ratio is in the range 0.10.9 . clearly , in the forbidden region , the @xmath9 can not be determined from resonant production .
for small and moderate values of @xmath24 , the study of decays of the heavy @xmath0-even higgs boson @xmath3 provides a means of determining the triple - higgs coupling @xmath9 . in order to extract the coupling @xmath10 , other processes involving two - higgs ( @xmath1 ) final states
must be considered .
the @xmath50 final states , which can be produced in the non - resonant double higgs - strahlung @xmath66 , could provide one possible opportunity , since it involves the coupling @xmath10 .
these non - resonant processes have also been investigated @xcite .
[ fig : sig - zll-2 ] ( 15,8.5 ) ( -1.5,1 ) we show in fig .
[ fig : sig - zll-2 ] the @xmath50 cross section , with @xmath67 .
the structure around @xmath68 ( in the case of no mixing ) is due to the vanishing and near - vanishing of the trilinear coupling . in the case of no mixing
, there is a broad minimum from @xmath69 to 90 gev , followed by an enhancement around @xmath70100 gev .
this structure is due to the vanishing of the branching ratio for @xmath65 , which is kinematically forbidden in the region @xmath6990 gev , see fig .
[ fig : hole ] ( this coincides with the opening up of the channel @xmath71 ) , followed by an increase of the trilinear couplings .
this particular structure depends considerably on the exact mass values @xmath57 and @xmath30 .
thus , it depends on details of the radiative corrections and on the mixing parameters @xmath5 and @xmath31 .
a two - higgs ( @xmath51 ) final state in @xmath26 collisions can also result from the @xmath46 fusion mechanism , which can either be a resonant process as in ( [ eq : res - hhh ] ) , or a non - resonant one like ( [ eq : ww - fusion ] ) .
since the neutral - current couplings are smaller than the charged - current ones , the cross section for the @xmath72 fusion mechanism in ( [ eq : res - hhh ] ) and ( [ eq : ww - fusion ] ) is an order of magnitude smaller than the @xmath46 fusion mechanism , and is here ignored .
the @xmath46 fusion cross section for @xmath73 can be written as @xcite ( see also @xcite ) @xmath74 ^ 2}\ { \cal f}(x , y)\right ] \cos^2(\beta - \alpha ) .
\label{eq : fusion - exact}\ ] ] this cross section is plotted in fig .
[ fig : sigma-500 - 1500 ] for the centre - of - mass energy @xmath75 gev , and for @xmath58 , as a function of @xmath57 .
the resonant fusion mechanism , which leads to @xmath76 $ ] + [ missing energy ] final states is competitive with the process @xmath77 $ ] + [ missing energy ] , particularly at high energies . since the dominant decay of @xmath1 will be into @xmath78 pairs , the @xmath3-strahlung and the fusion mechanism will give rise to final states that will predominantly include four @xmath79-quarks . on the other hand , the process @xmath6 will give rise to six @xmath79-quarks in the final state , since the @xmath80 final state typically yields three - higgs @xmath81 $ ] final states . besides the resonant @xmath46 fusion mechanism for the multiple production of @xmath1 bosons
, there is also a non - resonant @xmath46 fusion mechanism : @xmath82 through which the same final state of two @xmath1 bosons can be produced .
the cross section for this process , which arises through @xmath46 exchange as indicated in fig .
[ fig : feynman - nonres - ww ] , can be written in the `` effective @xmath46 approximation '' as @xmath83 where @xmath84 . here
, the cross section is written as a @xmath46 cross section , at invariant energy squared @xmath85 , folded with the @xmath46 `` luminosity '' @xcite : @xmath86 where @xmath87 .
the @xmath46 cross section receives contributions from several amplitudes , according to the diagrams ( a)(d ) in fig .
[ fig : feynman - nonres - ww ] .
we have evaluated these contributions @xcite .
our approach differs from that of dhz in that we do not project out the longitudinal degrees of freedom of the intermediate @xmath88 bosons .
instead , we follow the approach of ref .
@xcite , where transverse momenta are ignored everywhere except in the @xmath88 propagators .
[ fig : sig - ww-2 ] ( 15,8.0 ) ( -1,1 ) we show in fig .
[ fig : sig - ww-2 ] the @xmath46 fusion cross section , at @xmath89 , as given by eqs .
( [ eq : fusion - exact ] ) and ( [ eq : sigww - nonres ] ) , with @xmath67 .
the structure is reminiscent of fig .
[ fig : sig - zll-2 ] , and the reasons for this are same .
notice , however , that the scale is different .
in fig . [ fig : sensi-500 ] we have indicated in the @xmath25@xmath24 plane the regions where @xmath9 and @xmath10 might be measurable for @xmath90 .
we identify regions according to the following criteria @xcite : * regions where @xmath9 might become measurable are identified as those where @xmath91 ( solid ) , with the simultaneous requirement of @xmath92 [ see figs . [
fig : br - h - a][fig : hole ] ] . in view of the recent , more optimistic , view on the luminosity that might become available
, we also give the corresponding contours for 0.05 fb ( dashed ) and 0.01 fb ( dotted ) .
* regions where @xmath10 might become measurable are those where the _ continuum _ @xmath93 cross section [ eq .
( [ eq : sigww - nonres ] ) ] is larger than 0.1 fb ( solid ) . also included are contours at 0.05 ( dashed ) and 0.01 fb ( dotted ) . such regions are given for two cases of the mixing parameters @xmath5 and @xmath31 , as indicated .
we have excluded from the plots the region where @xmath94 , according to the lep lower bound @xcite .
this corresponds to low values of @xmath25 .
[ fig : sensi-500 ] ( 15,8.0 ) ( -1.5,1 ) with an integrated luminosity of 500 fb@xmath95 , the contours at 0.1 fb correspond to 50 events per year .
this will of course be reduced by efficiencies , but should indicate the order of magnitude that can be reached . at @xmath90 , with a luminosity of 500 fb@xmath95 per year ,
the trilinear coupling @xmath9 is accessible in a considerable part of the @xmath25@xmath24 parameter space : at @xmath25 of the order of 200300 gev and @xmath24 up to the order of 5 . with increasing luminosity
, the region extends somewhat to higher values of @xmath25 .
the `` steep '' edge around @xmath96 ( where increased luminosity does not help ) is determined by the vanishing of @xmath97 , see fig .
[ fig : hole ] .
the coupling @xmath10 is accessible in a much larger part of this parameter space , but with a moderate luminosity , `` large '' values of @xmath24 are accessible only if @xmath5 is small .
it should be stressed that the requirements discussed here are necessary , but not sufficient conditions for the trilinear couplings to be measurable .
we also note that there might be sizable corrections to the @xmath46 approximation , and that it would be desirable to incorporate the dominant two - loop corrections to the trilinear couplings .
we have presented the results of a detailed investigation @xcite of the possibility of measuring the mssm trilinear couplings @xmath9 and @xmath10 at an @xmath26 collider . where there is an overlap , we have confirmed the results of ref . @xcite . our emphasis has been on taking into account all the parameters of the mssm higgs sector .
we have studied the importance of mixing in the squark sector , as induced by the trilinear coupling @xmath5 and the bilinear coupling @xmath31 . at moderate energies ( @xmath90 ) the range in the @xmath25@xmath24 plane that is accessible for studying @xmath9 changes quantitatively for non - zero values of the parameters @xmath5 and @xmath31 .
as far as the coupling @xmath10 is concerned , however , there is a qualitative change from the case of no mixing in the squark sector .
if @xmath5 is large , then high luminosity is required to reach `` high '' values of @xmath24 . at higher energies ( @xmath89 ) , the mixing parameters @xmath5 and @xmath31 change the accessible region of the parameter space only in a quantitative manner .
j. ellis , g. ridolfi and f. zwirner , phys .
* b257 * , 83 ( 1991 ) ; y. okada , m. yamaguchi and t. yanagida , prog .
phys . * 85 * , 1 ( 1991 ) ; h. e. haber and r. hempfling , phys .
lett . * 66 * , 1815 ( 1991 ) .
j. ellis , g. ridolfi and f. zwirner , phys .
* b262 * , 477 ( 1991 ) ; + r. hempfling and a. h. hoang , phys . lett .
* b331 * , 99 ( 1994 ) ; m. carena , j.r .
espinosa , m. quirs and c.e.m .
wagner , phys .
* b355 * , 209 ( 1995 ) ; m. carena , m. quirs and c.e.m .
wagner , nucl .
* b461 * , 407 ( 1996 ) ; s. heinemeyer , w. hollik and g. weiglein , phys .
rev . * d58 * , 091701 ( 1998 ) . h. e. haber , in proceedings of the conference on _ perspectives for electroweak interactions in e@xmath98 e@xmath99 collisions _ , ringberg ( tegernsee ) , germany , 1995 ; ed .
b. a. kniehl ( world scientific , singapore , 1995 ) p. 219 .
r. n. cahn and s. dawson , phys .
b136 * , 196 ( 1984 ) ; s. dawson , nucl .
b249 * , 42 ( 1984 ) ; m. chanowitz and m. k. gaillard , phys .
* b142 * , 85 ( 1984 ) ; i. kuss and h. spiesberger , phys .
rev . * d53 * , 6078 ( 1996 ) . | trilinear couplings of the neutral @xmath0-even higgs bosons in the minimal supersymmetric standard model ( mssm ) can be measured through the multiple production of the lightest @xmath0-even higgs boson ( @xmath1 ) at high - energy @xmath2 colliders .
this includes the production of the heavier @xmath0-even higgs boson ( @xmath3 ) via @xmath4 , in association with the @xmath0-odd higgs boson ( @xmath5 ) in @xmath6 , or via @xmath7 , with @xmath3 subsequently decaying through @xmath8 .
these processes can enable one to measure the trilinear higgs couplings @xmath9 and @xmath10 , which can be used to theoretically reconstruct the higgs potential .
we delineate the regions of the mssm parameter space in which these trilinear higgs couplings could be measured .
university of bergen , department of physics + scientific / technical report no.1999 - 01 + issn 0803 - 2696 + hep - ph/9902270 + february 1999 + * measuring trilinear higgs couplings in the mssm * 0.5 cm p. osland@xmath11 and p. n. pandita@xmath12 + @xmath13 department of physics , university of bergen , n-5007 bergen , norway + @xmath14 department of physics , north eastern hill university , shillong 793 022 , india |
density functional theory ( dft ) is an important tool in computational quantum chemical research@xcite .
this is due to the fact that dft offers a good compromise between accuracy and computing time , which allows to investigate large molecules with hundreds of atoms .
the essential ingredient of dft , the exact exchange correlation functional , is not known . during the last decades
, efforts were made to develop better and better approximations to the exact exchange correlation functional ( by increasingly advanced analytic anstze and improved reference data sets for their parametrization)@xcite . as a result ,
a plethora of ( approximate ) density functionals is now at our disposal .
unfortunately , there is no truly rigorous way to systematically develop such approximations , and their reliability is assessed in statistical analyses by comparison to reference data sets . as a consequence , a given density functional might provide reasonable or even excellent results in some application , but completely fail in another one . comparing the values of one or several chemical properties calculated with different density functionals with accurate reference data , obtained either experimentally or by quantum chemical methods of controllable accuracy ( e.g. , coupled cluster@xcite ) , provides a means to measure the quality of these density functionals , at least with respect to the properties studied .
pople and coworkers were among the first to realize the importance of benchmark studies , which they applied to assess the quality of the so - called g1 and g2 composite approaches@xcite . in 1981 , they proposed a comparatively small database containing the atomization energies of 31 small molecules such as lih , ch@xmath7 , co@xmath8 , and h@xmath8nnh@xmath8@xcite .
this database has been continuously extended ; in 1997 , g2/97 was presented which comprised already 148 molecules@xcite , followed by the g3/99 test set with 222 molecules in 1999@xcite . in 2005 ,
the latest revision of this pople benchmarking database was presented
g3/05 includes a total of 454 reference data points@xcite .
more recently , truhlar and coworkers assembled a large collection of databases including atomization energies , ionization potentials , electron and proton affinities , barrier heights , anharmonic vibrational zero point energies , saddle point geometries , and a broad range of binding energies ( covering systems with hydrogen bonds , charge - transfer complexes , dipole interactions , and weakly interacting systems)@xcite .
this broad range of different properties shall assure a rather complete picture of the performance of a given density functional . in a similar spirit , grimme and korth proposed
`` mindless '' benchmarking relying on a large library of automatically generated `` artificial '' molecules , which should remove any chemical biases@xcite .
goerigk and grimme incorporated this test set into the gmtkn24 database covering main group thermochemistry , kinetics , and noncovalent interactions@xcite , which was recently extended to form the so - called gmtkn30 database@xcite .
a drawback of all of these databases is the fact that the coverage of large transition metal complexes is rather poor .
this deficiency can be significant for practical purposes as coordination chemistry in the condensed phase usually involves complex , spatially extended ligand environments .
by contrast , a compact small diatomic like cr@xmath8 with unsaturated valencies can be a true challenge for electronic structure methods , but it is no prototypical system for coordination chemistry .
while the grimme databases are specifically set up for main group chemistry and do therefore not cover any transition metals , the truhlar databases include some small transition metal compounds like ag@xmath8 , cuoh@xmath9 , fe(co)@xmath10 , ferrocene , and bis(benzene)chromium .
et al . _ developed a database with the heats of formation of 225 molecules containing 3d - transition metals@xcite .
however , most of these molecules are also rather small .
in addition , many of the experimental reference data have a rather large uncertainty , which might result in wrong conclusions as to the performance of a given density functional .
very recently , zhang _ et al .
_ therefore selected the 70 molecules which are estimated to have experimental uncertainties in their enthalpies of formation of at most 2kcalmol@xmath0@xcite and tested no less than 42 density functionals on that database .
however , this reduced dataset again contains only very small compounds ; furthermore , only a minority of them contains organic ligands .
hughes and friesner assembled a rather large database of spin - splittings of 57 octahedral first - row transition metal complexes@xcite .
furche and perdew also presented a database of 3d transition metals , including reaction energies .
again , also this database includes only very small compounds ( and reactions such as the dissociation of the vanadium dimer)@xcite .
by contrast , _ large _ organometallic compounds are key players in important fields such as homogeneous catalysis@xcite .
it is often not possible to draw conclusions from quantum chemical calculations on small complexes for larger compounds .
furthermore , dispersion effects have a much larger influence in bigger complexes .
therefore , a database of reliable values for important chemical properties like ligand dissociation energies for a variety of large transition metal complexes is highly desirable . for such large complexes , no truly reliable _ ab initio _ computational data are currently available .
therefore , for a functional investigation of large complexes that shall reach beyond a mutual comparison of popular functionals , we have to resort to experimental data .
ideally , these data should not only be accurate but also obtained in the gas phase , such that strong ( environmental ) intermolecular interactions like solvent effects can be excluded .
some of us accurately measured ligand dissociation energies of a broad range of large transition metal complexes in the gas phase by means of tandem mass spectrometry .
the experimental ligand dissociation energies were obtained by injecting the intact complex into a mass spectrometer and letting it subsequently collide with an inert noble gas such as argon or xenon .
this collision led to dissociation of a ligand , and the dissociation cross sections were measured as a function of the nominal collisional energy .
the actual value of the ligand dissociation energy was then determined by monte carlo simulations based on a rather involved model of the entire dissociation process involving the ligand dissociation energy as a fit parameter@xcite .
this procedure assures a sufficient and determinable accuracy of the ligand dissociation energies .
here , we present a database of ten selected ligand dissociation energies from refs . @xcite and investigate the performance of nine widely used density functionals .
this work is organized as follows : in section [ sec : compmet ] , we introduce our new database of coordination energies ( which we will abbreviate in the following as wccr10 ) , the density functionals investigated , and the computational benchmarking procedure .
the results are discussed in section [ sec : results ] before providing a conclusion together with an outlook in section [ sec : conclusion ] .
our database consists of the ten dissociation reactions depicted in fig .
[ fig : reactions ] .
these reactions were selected out of all such reactions that have been published so far with the experimental methodology mentioned in the introduction@xcite .
specific care was taken to achieve a good balance of different transition metals and different ligand types .
hence , duplication of similar ligand dissociation reactions has been avoided .
all molecules in this database feature a closed - shell electronic structure .
all transition metal complexes are charged , which is an implication of the fact that the ligand dissociation energies have been determined by mass spectrometry .
overview of the ten reactions of the wccr10 database .
the aromatic substituent abbreviated as `` ar '' is 2,6-c@xmath11h@xmath12cl@xmath8 . ]
table [ tab : reference ] lists the experimental ligand dissociation energies for the ten reactions depicted in fig .
[ fig : reactions ] .
the error bounds were estimated in the original literature from approximately 10 to 20 regressions with the l - cid program@xcite for three independent data sets .
furthermore , a lab - frame energy uncertainty of 0.15ev was assumed@xcite .
these experimental ligand dissociation energies are based on reaction rates derived within approximate rice
ramsperger kassel marcus ( rrkm ) theory and
can therefore be considered to be free energies at 0 kelvin .
therefore , these energies have to be compared to computed total electronic energies which are corrected for their zero - point vibrational energies . reactions 4 and 5 feature a twofold degeneracy of the reaction paths , i.e. , there are two equivalent pathways to product formation , involving either of the two phosphine ligands for reaction 4 and either of the two platinum - bound methyl groups for reaction 5 .
however , the reported experimental reaction energies@xcite were determined with a version of l - cid@xcite that did not take into account this degeneracy in the microcanonical rrkm treatment of the rates . as a result
, the reaction rates as a function of collision energy have been overestimated by a factor of two , which slightly affects both the kinetic shift treatment and the fitted reaction energies and indirectly , the other fitting parameters .
this typically leads to a minor underestimation of the reaction energies by @xmath133kjmol@xmath0 ( this corresponds well to the 2kjmol@xmath0 difference obtained from the eyring equation , which applies to canonical ensemble rate theory ) . whereas this error is within the experimental uncertainty of the reaction energies
, we decided for the wccr10 set to refit the original experimental data sets with a version of l - cid that does take into account the twofold reaction path degeneracy@xcite .
the new data obtained for reactions 4 and 5 are given in table [ tab : reference ] ( see the supporting information for further details ) .
.[tab : reference]experimental reference values for the ten ligand dissociation energies in the wccr10 database . [ cols="<,>,>",options="header " , ] @xmath5 note that for b3lyp - d3 , no data are available for reactions 2 , 3 , 4 , and 9 . for reactions 2 , 3 , and 4 , the other dispersion - corrected density functionals ( bp86-d3 and b97-d - d2 ) have particularly large errors .
we might therefore expect that the values presented in this table would be significantly worse for b3lyp - d3 if these data could be taken into account . by investigating the mean absolute deviations ( mads ) of the individual functionals ( see table [ tab : errors ] )
, we see that most functionals have mads of about 30kjmol@xmath0 .
the mad of the dispersion - corrected functionals bp86-d3 and b97-d - d2 is clearly larger , which is due to the above - mentioned significant errors in the case of reactions 2 , 3 and 4 .
we note again that there is no data available for these reactions for b3lyp - d3 , and we may expect that the mad of the results obtained with this functional would otherwise be significantly larger .
the largest absolute deviation ( lad ) from the experimental values is about 50kjmol@xmath0 for most functionals .
for bp86 the largest deviation is 67kjmol@xmath0 , and for b3lyp it is even as large as 83kjmol@xmath0 . excluding the dispersion - corrected functionals , b3lyp performs worst considering both the mad as well as the largest absolute deviation from experiment . bp86 has a comparable mad , but its lad is significantly better than in the case of b3lyp .
a considerable improvement can be obtained by using the non - empirical pbe functional .
with 32kjmol@xmath0 its mean absolute deviation is about 8kjmol@xmath0 lower than that of bp86 .
the lad is improved by roughly 13kjmol@xmath0 , being about 53kjmol@xmath0 . moving from this gga functional to the non - empirical meta - gga functional tpss
, no further improvement is obtained ; both the mad as well as the lad are almost identical to those of pbe . in the case of these two functionals , admixture of hartree
fock exchange can further decrease the mean absolute deviations .
in particular for pbe0 , the mad decreases to 27kjmol@xmath0 , which is the lowest value of all density functionals investigated in this work .
however , adding hartree fock exchange does not generally lead to a reliable description of the ligand dissociation energies , as exemplified by the b3lyp results .
note that in an earlier benchmark study by fuche and perdew@xcite , the tpss and tpssh functionals were found to be best functionals for dipole moments , reaction energies , and harmonic frequencies of 3d transition metal complexes .
still , the best - performing functionals in our study have errors of up to about 50kjmol@xmath0 for certain reactions and thus no functional performs well for our complete wccr10 set . even in a theory - only comparison
the scatter of the calculated results is significant : fig .
[ fig : energies2 ] provides a comparison of deviations in dissociation energy ( with the pbe0 functional chosen as a reference ) .
deviations of dissociation energies calculated with different functionals from those calculated with pbe0 , which has been chosen as a reference . ] considering again fig .
[ fig : energies ] , we note that the energies of reactions 2 , 3 , and 4 calculated with dispersion - corrected functionals deviate from both the experimental values and the values calculated with all other density functionals
. it is therefore appropriate to investigate this aspect in more detail .
grimme s dispersion corrections were parametrized for pairs of atoms in specific valence states within a reference molecule .
as the dispersion coefficients are calibrated against accurate results on neutral systems , this might lead to too large dispersion coefficients for cationic systems .
since there are no anionic counterions present in our study ( the dispersion coefficients of which might then be estimated to be too small ) , there can not be any cancellation of errors as discussed in ref .
@xcite .
one might argue that another possible cause for the deviation of dispersion - corrected energies observed here could be rooted in the large size of the complexes studied ( note that the dispersion - corrected functionals perform particularly badly for reactions 2 , 3 , and 4 , which involve the largest compounds of the wccr10 test set ) .
it has been found by tkatchenko and von lilienfeld that for large , bulky molecules ( and even more so for condensed - phase systems ) , three - body interactions must not be neglected@xcite .
this has also been confirmed by risthaus and grimme@xcite .
while it is in principle possible to account for these three - body dispersion effects by means of an axilrod teller muto term , this correction is not used by default in grimme s set of dispersion corrections@xcite . in order to assess their magnitude , we calculated the three - body correction term for the structures obtained with the standard d3 dispersion correction .
the resulting values for this three - body correction are very similar for the functionals bp86-d3 , b3lyp - d3 and b97-d - d2 , so that we may focus on the results obtained with bp86-d3 .
we find that the three - body interaction leads in all cases to a repulsive term , thus slightly destabilizing the molecule .
this destabilization is more pronounced the larger the molecules are . for example , in the case of the modified grubbs catalyst ( 174 atoms ; reaction 4 ) , the three - body term amounts to a contribution of + 14.0kjmol@xmath0 , while it is + 8.9kjmol@xmath0 for the azabox copper(i ) complex of reaction 3 ( 90 atoms , cf . ,
[ fig : dispersion ] ) , and only + 1.6kjmol@xmath0 for the aquo platinum(ii ) complex of reaction 1 ( 42 atoms ) . including the three - body correction thus leads to slightly smaller ligand dissociation energies . in the case of reaction 3 , the ligand dissociation energy is 7.5kjmol@xmath0 smaller , while the ligand dissociation energy of reaction 4 is lowered by 5.3kjmol@xmath0 . clearly , these corrections do not significantly improve the agreement with the experimental values and can therefore not explain the discrepancies . however , ligand dissociation energies could be sensitive to the spatial cut - off function , since the reactions studied involve the formation of a _ short _ metal ligand bond ( dispersion is difficult to model at short distances@xcite ) . in order to investigate this issue , we calculated the ligand dissociation energy of reaction 3 with bp86-d3 in combination with the becke
johnson ( bj ) damping function ( the standard damping function of the d3 correction is often called zero - damping , as the dispersion interaction between two atoms approaches zero if the internuclear distance approaches zero ) . with bp86-d3-bj
the reaction energy is calculated to be 298kjmol@xmath0 , while it is found to be 297kjmol@xmath0 with bp86-d3 .
therefore , the damping function has only a minor effect on the overall ligand dissociation energy of reaction 3 .
it is thus not likely that it can account for the discrepancies observed in this work .
this result is in line with the work by grimme and coworkers who found that the damping function in general has a negligible effect on the results ( such as ligand dissociation energies)@xcite . in order to better understand the role of dispersion interactions on the molecular structures and ligand dissociation energies , we calculated the reaction energy of reaction 8 also with bp86-d3 and b3lyp - d3 single - point calculations employing the bp86 and b3lyp structures , respectively . with bp86 ,
the reaction energy is found to be 194kjmol@xmath0 .
when calculating this energy with bp86-d3 at the bp86 molecular structures , a value of 208kjmol@xmath0 is found , i.e. , the ligand dissociation is more endothermic by 14kjmol@xmath0 .
when carrying out also the structure optimizations with bp86-d3 , the ligand dissociation energy is even more endothermic , namely 213kjmol@xmath0 .
we conclude from this that the structural changes due to dispersion interactions amount to a net effect of 5kjmol@xmath0 on the reaction energy . regarding b3lyp and b3lyp - d3 ,
a similar picture emerges .
b3lyp yields a reaction energy of 178kjmol@xmath0 for reaction 8 .
it is changed to 184kjmol@xmath0 when dispersion interactions are considered by means of single - point calculations on the b3lyp structures .
if these interactions are also taken into account during the structure optimizations , the reaction energy is 192kjmol@xmath0 .
we also analyzed the explicit values for the c@xmath11 and c@xmath14 dispersion coefficients for selected reactions .
for reactions 2 , 3 , and 4 , in which the dispersion corrections lead to particularly large discrepancies , data is available only for bp86-d3 and b97-d - d2 .
the dispersion coefficients of the d2 and d3 methods are rather different from each other , which is due to the fact that the d2 correction includes only the ( scaled ) c@xmath11 terms , whereas the d3 method includes ( unscaled ) c@xmath11 and c@xmath14 terms .
by contrast , considering the dispersion coefficients invoked for the calculation of the reactant complex of reaction 8 very similar values are applied in the bp86-d3 and b3lyp - d3 calculations .
still , the many small deviations between the dispersion coefficients for bp86-d3 and b3lyp - d3 add up to a difference of 35kjmol@xmath0 in the dispersion energies for this compound ( the total dispersion energy being @xmath15215kjmol@xmath0 in the case of bp86-d3 and @xmath15180kjmol@xmath0 for b3lyp - d3 ) .
the reaction energies of reaction 8 ( 213kjmol@xmath0 for bp86-d3 and 192kjmol@xmath0 for b3lyp - d3 ) agree well with the experimental reference of 194kjmol@xmath0in fact , with b3lyp - d3 the reference value is almost perfectly reproduced .
however , it is very intriguing that the structures deviate more from each other ( the bp86-d3 and b3lyp - d3 structures have a rmsd of 50pm ) , than those of the corresponding parent functionals bp86 and b3lyp ( the rmsd in this case being 17pm ) .
it is instructive to compare our results presented above with the findings obtained in previous studies . from the plethora of benchmark studies published so far we select here a few which are comparable to our work in the sense that they also focus on transition metal compounds .
the data obtained in these studies is included in table [ tab : errors ] . in the first study ( which we dub `` study 1 '' here ) , zhao and truhlar@xcite studied a model system for the grubbs ii catalyst .
they compared the energies of important stationary points in the catalytic cycle of the metathesis reaction obtained with different density functionals to accurate reference energies obtained with a ccsd(t)-based composite approach .
as the system studied in this work is very similar to reaction 4 of our work , one would expect a similar performance for the different density functionals .
in fact , we observe from the data in table [ tab : errors ] the same qualitative ordering , namely , pbe0 performs best , while pbe , tpss , and tpssh are only slightly worse .
b3lyp , however , is significantly worse compared to the other functionals .
nevertheless , we also observe some interesting differences compared to our study .
b3lyp performs 10kjmol@xmath0 worse in the work of zhao and truhlar , while the mads of other functionals are approximately 10kjmol@xmath0 better than in our benchmark .
this is in agreement with our study as we find a deviation between the experimental ligand dissociation energy of reaction 4 and the one calculated with b3lyp of 48.7kjmol@xmath0 , while pbe0 reproduces this energy rather accurately , having an error of only 4.5kjmol@xmath0 ( i.e. , this error is even smaller than the one found in study 1 ) .
however , the functionals tpss and pbe have an absolute error of roughly 30kjmol@xmath0 for reaction 4 , and are thus found to be significantly worse in our study than in study 1 . in a second , very recent
study ( hereafter called `` study 2 '' ) , zhang _ el al .
_ investigated 70 small 3d transition metal compounds and compared dft results to accurate experimental data@xcite .
these compounds are different compared to those in our wccr10 set .
nevertheless , the results of this study are in rather good agreement with our results , which might be an indication of a certain transferability of both results . the overall errors of pbe0 and pbe are approximately 7kjmol@xmath0 lower compared to our findings , and tpssh performs even roughly 14kjmol@xmath0 better in study 2 . however , it might come as a surprise that b3lyp has a mad of only 21kjmol@xmath0 in study 2 , which is in sharp contrast to our findings .
finally , furche and perdew@xcite also assessed the performance of different density functionals on a data set of small 3d transition metal compounds . here , we focus on their results from the analysis of 18 different reaction energies ( such as the dissociation of v@xmath8 , coh , and crf ) , for which experimental data are also available .
we denote this third study as `` study 3 '' in table [ tab : errors ] .
overall , furche and perdew found larger errors for the individual functionals than in our study , which might be due to the unsaturated valencies in the small molecules leading to peculiar electron - correlation effects .
the qualitative ordering is , however , roughly comparable to that found in our study . in summary
, we have seen that the actual numerical values for the absolute error of a given density functional can vary significantly , depending on the data set employed for the tests .
the qualitative ordering of different density functionals appears to be more robust in this respect , although also here , different test sets can lead to different conclusions .
therefore , when considering the performance of a given density functional one has to pay special attention to how this performance was assessed . while very large , diverse databases might give a good indication of the average performance of a density functional , they might under- or overestimate the errors of that density functional in a specific application .
likewise , a very specialized database covering only a certain class of compounds can hardly be considered representative for a broad area of chemistry .
we have presented the database wccr10 , a set of ten selected ligand dissociation energies for large transition metal complexes for which accurate , consistently measured experimental gas - phase data are available in the literature .
we then investigated the performance of nine different density functionals in a mutual comparison and with respect to the experimental ligand dissociation energies .
the density functionals selected encompass a broad range of different functional classes , namely gga , meta - gga , and hybrid functionals ; in addition , for some functionals also grimme s third generation dispersion correction was employed .
the calculations were set up such that any result observed has to be attributed to the density functional and not to other factors such as an incomplete basis set or numerical artifacts due to loose thresholds . in our benchmark study , we found for some reactions that grimme s dispersion correction yields reaction energies which overshoot those from experiment and from results obtained without dispersion corrections .
furthermore , we have identified cases in which different dispersion - corrected functionals can lead to significantly different structures , whereas the structures resulting from the uncorrected parent functionals were in much better mutual agreement .
this may not be taken as an argument against such dispersion corrections as the parent functional apparently plays a crucial role .
unfortunately , a definitive assessment of the quality of the structures is not possible since there are no reliable reference data to compare with , but our study indicates that care should be taken when applying dispersion corrections in structure optimizations . among all functionals
not corrected for dispersion , b3lyp performs worst with a mean absolute deviation of about 39kjmol@xmath0 , while bp86 is a little bit better , at least in terms of the largest absolute deviation .
the non - empirical density functionals pbe and tpss have a clearly smaller error both in terms of mean absolute deviation and largest absolute deviation .
this result is particularly appealing from a theoretical point of view , since these density functionals are constructed by choosing a mathematical expression that satisfies constraints which would also be fulfilled by the true exchange correlation functional , instead of by simply fitting it to a reference set of chemical data . adding a portion of hartree
fock exchange to these functionals improves their performance further .
we have compared our findings to selected benchmark studies from the literature and found that the actual numerical values for the errors of the individual density functionals depend heavily on the test set chosen .
the relative performance of different density functionals is at least from a qualitative point of view somewhat more robust with respect to the choice of the test set , but also here one finds astonishing differences .
therefore , care must be taken when utilizing such benchmark results to assess the performance of density functionals .
even the best functional identified for the wccr10 set , pbe0 , has a rather large mean absolute deviation . on average ,
the ligand dissociation energies deviate by more than 20kjmol@xmath0 , while the maximum deviation is even as large as 55kjmol@xmath0 .
as these values lie well within the range of chemically relevant energies ( which are on the order of a few kjmol@xmath0 to a few hundred kjmol@xmath0 ) , the accuracy of contemporary density functionals often can not be expected to be sufficient for a given application .
it is therefore of utmost importance to further analyze the density functionals available today in order to find possible cures to their weaknesses .
relevant for the development and benchmarking of new or improved functionals is the study of composite approaches@xcite as exemplified by the so - called correlation - consistent composite approach ( ccca method ) of wilson and coworkers@xcite , the gaussian4 method of curtiss and coworkers@xcite , the method of peterson and coworkers@xcite , the benchmarking efforts by martin and co - workers @xcite , and a very recent approach developed by bross _
et al._@xcite or a reinvestigation of ancient parts of present - day functionals by , for instance , explicit reconstruction from accurate ( spin ) densities@xcite .
this work has been supported by the swiss national science foundation snf ( projects 200020_144458 and 200020_137505 ) .
cartesian coordinates of all optimized structures investigated in this work , discussion of structural changes due to dispersion corrections of the reactant complex of reaction 4 , discussion on the mechanism of reactions 2 and 3 , and details of the redetermination of the energies of reactions 4 and 5 .
this material is available free of charge via the internet at http://pubs.acs.org . | we present the wccr10 data set of ten ligand dissociation energies of large cationic transition metal complexes for the assessment of approximate exchange
correlation functionals .
we analyze nine popular functionals , namely bp86 , bp86-d3 , b3lyp , b3lyp - d3 , b97-d - d2 , pbe , tpss , pbe0 , and tpssh by mutual comparison and by comparison to experimental gas - phase data measured with well - known precision .
the comparison of all calculated data reveals a large , system - dependent scattering of results with nonnegligible consequences for computational chemistry studies on transition metal compounds .
considering further the comparison with experiment , the non - empirical functionals pbe and tpss turn out to be among the best functionals for our reference data set .
the deviation can be lowered further by including hartree fock exchange .
accordingly , pbe0 and tpssh are the two most accurate functionals for our test set , but also these functionals exhibit deviations from experiment by up to 50kjmol@xmath0 for individual reactions . as an important result we found no functional to be reliable for all reactions .
furthermore , for some of the ligand dissociation energies studied in this work , dispersion corrections yield results which increase the deviation from experiment . this deviation increases further if structure optimization including dispersion corrections is performed .
finally , we compare our results to other benchmark studies and highlight that the performance assessed for different density functionals depends significantly on the reference molecule set chosen . * new benchmark set of transition - metal coordination reactions for the assessment of density functionals * thomas weymuth@xmath1 , erik p. a. couzijn@xmath2 , peter chen@xmath3 , + and markus reiher@xmath4 + @xmath5 eth zurich , laboratorium fr physikalische chemie , + vladimir - prelog - weg 2 , 8093 zurich , switzerland @xmath6 eth zurich , laboratorium fr organische chemie , + vladimir - prelog - weg 2 , 8093 zurich , switzerland date : = march 19 , 2014 + status : submitted to _ j. chem .
theory comput . _ |
the physics of holes moving in an af background has received much attention because of its possible connection to high - t@xmath6 superconductivity . a microscopic realization of this physics is given by the @xmath0-@xmath2 model which was first introduced as a conjecture @xcite , and then derived as a low - energy limit of a realistic model representing the electronic structure of cuprates @xcite .
recent comparison of the @xmath0-@xmath2 model results to the angle - resolved photoemission data @xcite showed that the overall shape of the experimental single - hole band can be fitted satisfactorily using the pure @xmath0-@xmath2 model . to account for
the detail line shapes , one needs to include more distant hopping terms ( @xmath7 , @xmath8 , etc . )
the presence of these terms also follows from a careful mapping study @xcite .
numerical studies of the @xmath0-@xmath2 model by means of ed and other methods suggest the presence of hole binding in the physical parameter range @xmath9 @xcite , and that the dominant symmetry of the pairing correlation is @xmath10 .
this finding also supports the common belief that some variant of the @xmath0-@xmath2 model is able to describe the physics of the real compounds . however , to show that this @xmath4-wave binding in the @xmath0-@xmath2 model is relevant to the physics of high-@xmath11 materials , one has to clarify how strong and generic the reasons behind this pairing are .
unfortunately , numerical studies alone are of limited use for this purpose , so that an analytical study is necessary to develop an insight into the problem .
in this paper we attempt such a study by investigating the ising limit of the @xmath0-@xmath2 model the @xmath0-@xmath1 model .
it is well - known to be a simplified limiting case of the @xmath0-@xmath2 model , but only recently has its properties been fairly well understood . in the pioneering work by bulaevskii _
et al _ @xcite the single - hole energy in the @xmath12 limit has been calculated .
subsequently , brinkmann and rice @xcite considered the @xmath13 limit and introduced the retraceable - path approximation .
more recent works @xcite used the self - consistent born approximation ( scba ) to study the @xmath0-@xmath1 , @xmath0-@xmath2 , and @xmath0-@xmath14-@xmath2 models . using this approximation , an analytical expression for the hole green s function in the @xmath0-@xmath1 model for arbitrary values of @xmath5 has been found by starykh and reiter @xcite .
the pairing of holes in the @xmath0-@xmath1 model has been considered in the works by trugman @xcite and by shraiman and siggia @xcite , where important results were obtained . some other studies used different modifications of the variational approach @xcite and obtained similar results . however , as we shall show , previous analytical treatments involved approximation which prevented one from obtaining the correct answers . the @xmath0-@xmath1 model has also been studied using numerical methods @xcite . barnes _
et al _ @xcite compared their ed results on a 16-site square lattice with some analytical results , and riera and dagotto @xcite developed a modified lanczos technique which enable them to study the one- and two - hole ground states of the model on lattices as large as 50-sites .
( for a review see ref .
it was shown that finite - size effects in the ed data of the @xmath0-@xmath1 model are much smaller than those of the @xmath0-@xmath2 model @xcite . in this paper we develop an analytical treatment of the @xmath0-@xmath1 model based on the results of ref .
comparison with our ed data allows us to justify the validity of this analytical approach .
furthermore , we use the results of this study to shed light on the problem of whether the pairing in the @xmath4-channel is a generic feature of @xmath0-@xmath2-like models , and what interaction defines the symmetry of the ground state of the system . our numerical ed results for the @xmath0-@xmath1 model are obtained on a 32-site square lattice using the same method as in refs .
this is up to now the only ed results on the largest lattice .
this numerically exact method allows us to analyze various properties of the single - hole ground state as well as the two - hole bound states with different symmetries .
our paper is organized as follows .
sections ii and iii describe the representation and feynman rules used in this paper . in sec .
iv we solve the single - hole problem and compare the results to ed data . in sec .
v the two - hole problem is considered in detail . in sec .
vi we summarize our results and draw conclusions .
the hamiltonian of the @xmath0-@xmath1 model is @xmath15 \end{aligned}\ ] ] where the summation runs over nearest - neighbor bonds @xmath16 , @xmath17 , @xmath18 , @xmath19 .
since we are modeling electrons on a 2d square lattice , the coordination number @xmath20 and the spin @xmath21 . assuming the presence of long - range antiferromagnetic order one can introduce a representation for the hubbard operators ( operators of spin and constrained fermion ) through the spinless fermion and boson operators .
the choice of the representation is , in general , arbitrary and is motivated by the problem to be solved . in our work
we use a generalization of the dyson - maleev ( dm ) representation for the hubbard operators . as usual , such a representation conserves the algebra of the operators but extends the hilbert space of the problem . in sublattice
@xmath22 the dm representation is given by @xmath23 in sublattice @xmath24 , the corresponding representation is given by @xmath25 thus , we have two types of fermions and bosons associated with the @xmath26 and @xmath27 sublattices .
the hamiltonian in eq.([h ] ) expressed in terms of the new variables eqs.([rep1],[rep2 ] ) is @xmath28\ , \nonumber\\ & & { \cal h}_1 = t\sum_{\langle ij\rangle}\bigl[f_i^{\dag}g_j ( b_j^{\dag}+a_i)+{\rm h.c.}\bigr]\ , \\ & & { \cal h}_2=-t\sum_{\langle ij\rangle}\bigl[f_i^{\dag}g_j b_j^{\dag}a_i^{\dag}a_i+g_j^{\dag}f_ia_i^{\dag}b_j^{\dag}b_j \bigr]- \frac{1}{2}j\sum_{\langle i , j \rangle}\bigl [ n_i^f b_j^{\dag}b_j + n_j^g a_i^{\dag}a_i + n_i^f n_j^g + 2a_i^{\dag}a_ib_j^{\dag}b_j\bigr ] \ . \nonumber\end{aligned}\ ] ] here we have divided the hamiltonian eq.([h1 ] ) into three parts : @xmath29 , consisting of the linear spin - excitation and hole terms , @xmath30 consisting of the bare hole - magnon interaction term , and @xmath31 , consisting of the nonlinear hole - magnon , static hole - magnon , direct hole - hole , and magnon - magnon interaction terms .
the third and fourth terms in @xmath29 take into account the energy of the four af bonds broken by a hole introduced into the system . in the following this energy
is included in the ground state energy and the corresponding terms are omitted .
note that we have omitted a constant term .
the same hamiltonian eq.([h1 ] ) in @xmath32-space is given by @xmath33 + 2j\sum_{\bf k}\bigl [ f_{\bf k}^{\dag}f_{\bf k}+g_{\bf k}^{\dag}g_{\bf k}\bigr]\ , \nonumber\\ & & { \cal h}_1 = 4t\sum_{\bf k , q } \gamma_{\bf k - q}\bigl[f_{\bf k - q}^{\dag}g_{\bf k}b_{\bf q}^{\dag } + g_{\bf k - q}^{\dag}f_{\bf k}a^{\dag}_{\bf q } + { \rm h.c.}\bigr]\ , \\ & & { \cal h}_2=-4t\sum_{\bf k , q , q_1,q_2}\gamma_{\bf k - q } \bigl[f_{\bf k - q - q_1+q_2}^{\dag}g_{\bf k}b_{\bf q}^{\dag } a_{\bf q_1}^{\dag}a_{\bf q_2}+g_{\bf k - q - q_1+q_2}^{\dag}f_{\bf k } a_{\bf q}^{\dag } b_{\bf q_1}^{\dag}b_{\bf q_2}\bigr ] \\ & & \phantom{{\cal h}_1=}- 2j\sum_{\bf k , q , q_1}\gamma_{\bf q - q_1}\bigl [ f_{\bf k - q+q_1}^{\dag}f_{\bf k } b_{\bf q}^{\dag}b_{\bf q_1 } + g_{\bf k - q+q_1}^{\dag}g_{\bf k } a_{\bf q}^{\dag}a_{\bf q_1 } \bigr ] \nonumber\\ & & \phantom{{\cal h}_1= } -2j\sum_{\bf k , k^{\prime},q}\gamma_{\bf q } f_{\bf k - q}^{\dag}g_{\bf k^{\prime}+q}^{\dag}g_{\bf
k^{\prime}}f_{\bf k } -4j\sum_{\bf q_1,q_2,q}\gamma_{\bf q } a_{\bf q_1-q}^{\dag}b_{\bf q_2+q}^{\dag}b_{\bf q_2}a_{\bf q_1}\ , \nonumber\end{aligned}\ ] ] where @xmath34 , and all summations are restricted to the first magnetic brillouin zone ( bz ) .
the feynman rules for the model in eq .
( [ h2 ] ) are : + ( i ) hole propagator : @xmath35^{-1}$ ] 1.2 cm = 0.9 cm ( 0,0.5 ) ( 0,0.5 ) + ( ii ) magnon propagator : @xmath36^{-1}$ ] 0.5 cm = 1.3 cm ( 1,0.5 ) ( 0,0.2 ) + ( iii ) hole - magnon vertex : + + 0.7 cm @xmath37 ( 2,2 ) ( 0,0.7 ) + the indices indicate that @xmath38 ( or @xmath39)-holes can emit only @xmath40 ( or @xmath41)-magnons and absorb only @xmath41 ( or @xmath40)-magnons .
this is an important feature of the hole - magnon interaction , and is a consequence of spin conservation . in the following
we omit the indices .
we use the following shorthand notations @xmath42 denotes @xmath43 and @xmath44 means @xmath45 .
+ ( iv ) hole - two - magnon vertex : + + 0.7 cm @xmath46 ( 2,2 ) ( 0,0.7 ) + note that @xmath38 ( or @xmath39)-holes interact with @xmath40 ( or @xmath41)-magnons only .
+ ( v ) hole - three - magnon vertex : + + 0.7 cm @xmath47 = 2.5 cm ( 2,2 ) ( 0,0.5 ) + note that there is no vertex diagram for the reversed order of emission - absorption because of the non - hermitian nature of the corresponding terms in the hamiltonian eq.([h2 ] ) .
+ ( vi ) hole - hole vertex : + + 0.7 cm @xmath48 ( 2,2 ) ( 0,0.7 ) + ( vii ) magnon - magnon vertex : + + 0.7 cm @xmath49 ( 2,2 ) ( 0,0.7 ) + where we use the notations @xmath50 and @xmath51 for the momentum and frequency of the hole and magnon respectively .
thus the model eq.([h2 ] ) involves fermion - fermion ( hole - hole ) , boson - boson ( magnon - magnon ) , and three types of fermion - boson ( hole - magnon ) interactions . in general the problem of finding the low - energy excitations in such a model
is very complicated .
however , as we will show below , in the case of the @xmath0-@xmath1 model it is possible to take into account the result of renormalization almost exactly using a regular diagrammatic treatment .
if we ignore @xmath31 , the hamiltonian in eq .
( [ h2 ] ) for a system with a single - hole is similar to the problem of a single electron interacting with the local phonon mode .
when the interaction ( @xmath0 in our case ) is small compared to the energy of the boson ( @xmath52 ) , we can use perturbation theory and consider only the lowest contribution to the hole energy , which corresponds to the virtual emission - absorption of a single magnon . at larger @xmath0 many - magnon intermediate states
must be considered .
such a problem , if solved self - consistently , must take into account the renormalization of the fermion - boson interaction , or crossing diagrams .
however , in the @xmath0-@xmath2 model it is well known that the lowest correction to the bare hole - magnon vertex @xmath53 is exactly zero since the hole - magnon interaction conserves the spin , or more exactly , the pseudospin associated with the sublattice type . therefore , the crossing diagrams can be omitted and
the renormalization of the hole energy is given only by the series of diagrams where the magnons are absorbed in exactly the reversed order in which they are emitted .
this is the so - called self - consistent born approximation ( scba ) , which in the case of the @xmath0-@xmath1 model is identical to the `` retraceable path approximation '' .
this coincidence is due to the local nature of the spin - excitations in the model . the corresponding diagrammatic equation for the hole green s function is shown in fig .
[ diag1 ] .
note that the vertices are unrenormalized .
= 1.5 cm ( 7,3 ) ( 1,1 ) the retraceable path approximation implies that the hole is confined by a `` string '' of spin excitations and the hole motion is completely incoherent , i.e. the hole form a localized state . the first non - zero contribution to the hole self - energy from the crossing diagrams
is shown in fig .
it is of sixth order in @xmath0 and is small compared to the contribution of the `` retraceable '' diagrams of the same ( sixth ) order due to the geometrical factor .
( 2,3 ) ( 7,0.5 ) this diagram is the first member of the family of the so - called `` trugman paths '' .
their contribution to the hole energy is small in a wide range of @xmath5 , but they are responsible for the finite coherent band of the hole .
real - space consideration shows that such a diagram corresponds to the hole s motion around an elementary square loop in which the hole makes one and a half turn . as a result
all spin excitations are cured and the hole is translated along the diagonal of the square plaquette , i.e. to a next - nearest neighbor . as we have noted , the corresponding correction to the energy is small and we will neglect these diagrams in the rest of this section .
since @xmath54 and @xmath55 are momentum - independent and the hole - magnon interaction depends only on @xmath56 , the full green s function @xmath57 is also @xmath32-independent .
thus the equation shown in fig .
[ diag1 ] can be simplified as @xmath58^{-1}\ .\end{aligned}\ ] ] the solution of eq .
( [ scba ] ) at all @xmath5 has been found in ref .
one can try an ansatz for @xmath57 of the form : @xmath59 this transforms eq .
( [ scba ] ) into a difference equation : @xmath60 which is the recursion relation for the bessel functions .
thus @xmath61 is given by @xmath62 where @xmath63 are the bessel functions .
such a form is the consequence of the continued fraction form of the green s function , which in turn is the result of the retraceable path approximation .
note that in ref .
@xcite @xmath64 .
thus the poles of the green s function in the scba approximation are defined by the zeros of the bessel function . since
all these zeros are real , the hole spectral function consists of the set of delta - function peaks corresponding to the energy levels of the quasiparticle states in the `` string '' potential well .
inclusion of the trugman processes does not change this result significantly because the resulting bandwidth is much smaller than the separation of the levels . although the solution ( [ g1 ] ) of eq .
( [ scba ] ) is an almost exact solution of @xmath65 , it is not a solution of the original @xmath0-@xmath1 hamiltonian . besides the crossing diagrams
, there are other contributions to the self - energy originating from @xmath31 . in order to compare with other analytical approaches and ed ,
it is necessary to account for these contributions . to make this statement evident
let us now consider the two limiting cases @xmath12 and @xmath66 , and compare the results of the scba with the known facts from the real - space approach and ed numerical data . for @xmath12
the ground - state energy for the hole is given by eqs .
( [ scba ] ) as @xmath67 this is inconsistent with the result of the real - space approach @xmath68
the reason for this discrepancy is that the spin - flip ( magnon ) has lower energy ( @xmath69 compared to @xmath52 ) if it is located in the neighborhood of the hole , and the hamiltonian @xmath65 does not take this difference into account . note that throughout this paper we assign zero energy level to the state with one _ static _ hole , i.e. @xmath70 .
according to ref .
@xcite , in the large @xmath0 limit the ground - state energy given by eq .
( [ g1 ] ) is @xmath71 where @xmath72 is the first zero of the airy function .
there is no alternative asymptotic expression in this limit which is accurate up to order @xmath2 , but the first two terms which are of order @xmath0 and @xmath73 can be compared to the results of the `` string '' approach by shraiman and siggia ( ss ) @xcite and the fitting of 16-site ed data by barnes _
@xcite @xmath74 this comparison makes one to suspect that scba ( eq.([w_0_2 ] ) ) overestimates both the absolute value of the depth of the `` band '' and the slope of the walls of the linear confining potential .
the origin of these discrepancies becomes evident when we consider the real - space picture of the retraceable hole movement .
[ mov ] shows the real - space images of three components of the spin - polaron wave function : the one- , two- , and three - magnon strings .
the first important observation is that the coordination number in the creation of the first magnon is indeed @xmath20 , but is @xmath75 in the creation of each next magnon .
this is why the factor in the first term of the ss expression eq .
( [ w_0_3 ] ) for the hole energy is @xmath76 ( instead of @xmath77 ) .
if @xmath0 is much larger than @xmath2 , the average length of the strings contributing to the spin - polaron wave function is large .
then the weight of the components with coordination number different from others becomes insignificant .
since this component is the `` bare '' hole , it also means that the quasiparticle residue of the state decreases as @xmath0 grows and in the @xmath78 limit the spectrum will be incoherent .
this result was first obtained by brinkmann and rice @xcite in the @xmath13 limit .
( 2,3 ) ( 3,0.5 ) having in mind that the energy of spin excitation is determined by the number of broken af bonds associated with its creation , one can see that the energy of the first magnon in the string is @xmath69 ( 3 bonds , @xmath79 each ) , but the energy of each subsequent magnon is just @xmath2 ( 2 bonds ) .
this is true for all strings of length @xmath80 and for most other longer strings except for those which have self - tangencies .
since the latter cases are relatively rare due to the geometrical factor , one can assign the same energy @xmath81 to every string of length @xmath82 . a more accurate account for the energy of the self - tangient strings
makes only tiny changes to the energy of the whole excitation .
thus , making the changes @xmath83 and @xmath84 in the expression for the scba hole energy will change the factor @xmath85 in the second term of eq .
( [ w_0_2 ] ) to @xmath86 , leaving it in reasonable agreement with eq .
( [ w_0_3 ] ) . summarizing the results of the comparison one can say that the reason for the failure of the scba is that @xmath65 is not the same as @xmath87 and that an improved formalism must take into account the following facts : ( i ) the energy of the first magnon in a string is @xmath69 , ( ii ) the energy of the @xmath88th magnon ( @xmath89 ) is @xmath2 , ( iii ) the coordination number is @xmath75 for @xmath90 and @xmath20 for @xmath91 , where @xmath82 is the length of the string . after some thinking using the above real - space arguments one can suggest the following modification to the equation for the green s function ( [ scba ] ) to meet all these requirements @xmath92^{-1}\ , \nonumber \\ & & g_a(\epsilon)=[\epsilon-3t^2g_a(\epsilon - j)]^{-1}\ , \\ & & g_a(\epsilon)=-\frac{1}{\sqrt{3}t}\frac{j_{-\epsilon / j}(2\sqrt{3}t / j ) } { j_{-\epsilon / j-1}(2\sqrt{3}t / j)}\ . \nonumber\end{aligned}\ ] ] such a modification gives the correct hole energy in the small @xmath0 limit . in the @xmath78
limit the resulting hole energy is @xmath93 .
let us now come back to the original problem with the full hamiltonian @xmath94 , eq .
( [ h2 ] ) and discuss how such an `` improvement '' of the green s function proposed in eq.([g2 ] ) can be done formally in the diagrammatic language .
the fact that magnons in the string have lower energy than free excitations is the result of magnon - magnon binding .
the magnon - magnon interaction ( last term of @xmath31 , eq .
( [ h2 ] ) ) is attractive and leads to a bound state .
a regular approach to the bound state problem is based on the bethe - salpeter equation shown in fig .
[ diag2 ] .
( 3,3 ) ( 2,0.5 ) the equivalent integral equation is @xmath95 it is easy to see that the momentum and frequency dependence of the vertex can be separated .
the suggestion @xmath96 transforms eq .
( [ bsm ] ) into @xmath97 here we have used the relation @xmath98 . thus the renormalized magnon - magnon vertex is given by @xmath99 the pole of this function , @xmath100 , corresponds to the bound state whose energy is lower than that of two free magnons by @xmath101 , in agreement with our expectations .
similarly the hole - two - magnon interaction ( second term in @xmath31 , eq .
( [ h2 ] ) ) lowers the energy of the magnons in the neighborhood of a hole .
the hole - three - magnon interaction ( first term in @xmath31 ) originates from the projection operator and projects out the unphysical states with the hole and magnon at the same site created by the hopping term in @xmath29 .
a consistent consideration of these interactions yields the following dyson equation for the single - hole propagator : @xmath102^{-1 }
\ , \nonumber\\ & & \sigma(\epsilon)=\sum_{\omega}g(\epsilon-\omega)d^0(\omega ) \sum_{\bf q } \gamma^0_{\bf k - q } \gamma_{\bf k , q}(\epsilon,\omega)\ , \end{aligned}\ ] ] where the renormalized hole - magnon vertex is @xmath103 a diagrammatic representation of eqs .
( [ dyson],[dyson1 ] ) is shown in fig .
[ diag3 ] .
= 4.5 cm ( 6,5 ) ( 0,0.5 ) note that the lowest order correction to _ all _ vertices are zero , and higher order corrections are neglected .
we have already discussed the absence of one - loop correction to the hole - magnon vertex .
it is less evident in cases of the hole - two- and hole - three - magnon vertices .
however , if one recalls the @xmath104 and @xmath105 indices for the hole and magnon lines in fig .
[ diag3 ] , it becomes evident that the emission of an extra magnon prior to the hole - two - magnon interaction alters the type of the hole and therefore prevents it from absorbing and reemitting the original magnon .
the same is true for the hole - three - magnon vertex .
it is interesting to note that the set of diagrams in fig .
[ diag3 ] is still in the scba , i.e. , the crossing diagrams are absent . again
the momentum and frequency dependence of @xmath106 can be separated .
the almost evident substitution @xmath107 simplifies eqs .
( [ dyson],[dyson1 ] ) which , after integrating over the internal frequencies and momenta , becomes @xmath108\ .
\nonumber\end{aligned}\ ] ] each term in the square bracket comes from the corresponding irreducible diagram in fig .
[ diag3 ] . after combining
similar terms @xmath109 becomes @xmath110^{-1 } \\ & & \phantom{\tilde{\gamma}(\epsilon,2j ) } = g(\epsilon-2j)^{-1 } \biggl[\epsilon-3j/2 - 3t^2g(\epsilon-3j)\tilde{\gamma}(\epsilon-2j , j ) \biggr]^{-1 } = \frac{g_a(\epsilon-3j/2)}{g(\epsilon-2j)}\ , \nonumber\end{aligned}\ ] ] where @xmath111 has a continued fraction form .
summarizing eqs .
( [ dyson]),([dyson2]),([dyson3 ] ) one finally obtains @xmath112^{-1}\ , \\ & & g_a(\epsilon)=[\epsilon-3t^2g_a(\epsilon - j)]^{-1 } = -\frac{1}{\sqrt{3}t}\frac{j_{-\epsilon / j}(2\sqrt{3}t / j ) } { j_{-\epsilon
/ j-1}(2\sqrt{3}t / j)}\ , \nonumber\end{aligned}\ ] ] which is exactly the same as the green s function suggested in eq .
( [ g2 ] ) .
the rest of this section is devoted to the comparison of the results from eq .
( [ dyson4 ] ) with our numerical ed results on a 32-site lattice .
[ comp1 ] shows the ed results together with analytical results from the scba eq .
( [ scba ] ) and from eq .
( [ dyson4 ] ) .
( 9,7.5 ) ( 5,-4 ) we recall that contributions from the trugman s paths are left out from eqs .
( [ dyson4 ] ) and ( [ scba ] ) .
these paths give rise to a small dispersion of the hole , @xmath113 , with @xmath114 .
this form of dispersion implies that the correction to the energy at the point @xmath115 due to the trugman paths is almost zero .
therefore , it is natural to compare the ed data at this @xmath32-point to the analytical results . from fig .
[ comp1 ] one can see a beautiful agreement of the numerical data at @xmath116 with the results of eq .
( [ dyson4 ] ) in the whole range of @xmath5 . for comparison purpose ,
[ comp1 ] also shows the numerical data at @xmath117 .
we also remark that our 32-sites ed - data at @xmath117 agree with the 50-sites results of the modified lanczos study by riera and dagotto @xcite ( not shown in fig .
[ comp1 ] ) up to the fourth digit at @xmath118 and up to the seventh digit at @xmath119 .
it is already clear from fig .
[ comp1 ] that our present approach works better than the scba . to further illustrate this point
, we plot the quasiparticle residue in figure [ comp2 ] .
three sets of data are shown .
they are the results from our present analytical approach , the scba , and our numerical ed study .
the ed results are calculated using the same method as in ref .
one can see that such a comparison unambiguously favor the present approach .
( 9,8.5 ) ( 5,-4 )
in this section we consider the problem of hole pairing in the @xmath0-@xmath1 model using the diagrammatic formalism introduced in previous sections .
our goal is to study in detail the bound state of two holes .
the role of different interactions and higher order corrections will be analyzed .
bound states with @xmath120- , @xmath3- , and @xmath4-symmetry and zero total momentum will be investigated . throughout this paper
we consider the @xmath121 bound states only .
consequently we are interested in the interactions of the particles originating from different sublattices ( @xmath38 and @xmath39 fermions ) .
if an exact two - particle green s function has a pole in the scattering channel at fixed total energy of the particles , then there exists a bound state whose energy is the total energy of the particles .
thus , to analyze the bound state problem of two holes one has to solve the bethe - salpeter equation , whose most general form is given by @xmath122 where @xmath123 and the indices are given in accordance with fig .
[ bsg_fig ] .
= 2.5 cm ( 3,3 ) ( 1,0.5 ) the bare vertex function @xmath124 includes all diagrams which can not be reduced to the second term of eq .
( [ bsg ] ) , i.e. , those which can not be cut by a vertical line between the ends @xmath125 , @xmath126 and @xmath127 , @xmath77 into two parts joined only by two hole lines .
the set of all such diagrams is called a `` compact '' vertex . at the pole ,
when @xmath128 ( @xmath129 is the energy of the bound state ) , the left - hand side and the second term in the right - hand side of eq .
( [ bsg ] ) are singular . therefore @xmath130 , and the first term in the right - hand side can be neglected .
hence the integral equation ( [ bsg ] ) becomes homogeneous in @xmath131 .
the variables @xmath132 and @xmath133 become parameters and are not defined by the equation itself . omitting them and the indices @xmath134 yields @xmath135 further simplifications of this general formula eq .
( [ bs1 ] ) are possible based on the specifics of the problem being considered .
rather general and frequently used simplification can be performed , for example , when the bare interaction @xmath106 is a potential - like term , i.e. independent of the frequency . in this case
@xmath136 becomes @xmath137-independent and plays only an auxiliary role .
furthermore , it is more convenient to use the bound state wave function @xmath138 . by changing the variables and integrating both sides of eq .
( [ bs1 ] ) over @xmath137 we obtain the schrdinger equation for the bound state in momentum representation @xmath139 if the green s functions are the bare ones , the sum over @xmath137 simply gives @xmath140 $ ] , where @xmath141 is the single - particle energy . in the @xmath0-@xmath1 model
the following characteristics allow one to simplify eq .
( [ bs1 ] ) considerably and obtain most of the results for the bound state problem in a transparent analytical form : + ( i ) the single - hole green s function is mostly @xmath32-independent , meaning that in a wide range of @xmath5 ratio the @xmath32-dependent contribution to the self - energy is insignificant or small .
+ ( ii ) the lowest order contributions to the compact vertex @xmath142 have a simple kinematic structure and , together with ( i ) , allow one to separate the momentum and frequency dependence of the exact vertex @xmath136 .
its @xmath32-dependent part can be classified in terms of @xmath120- , @xmath3- and @xmath4-wave harmonics and can be integrated out easily .
+ ( iii ) all one - loop and lowest order crossing corrections are exactly zero because of spin conservation in the hole - magnon interaction .
non - zero corrections are of higher order and are expected to be small .
therefore , the lowest order diagrams in @xmath106 can be left unrenormalized .
+ ( iv ) as we have shown in sec .
iii the single - hole spectral function consists of a set of well separated quasiparticle peaks .
therefore , the green s function of the hole can always be written in the form @xmath143 where @xmath144 and @xmath145 are the residue and the energy of the pole @xmath146 , respectively . then
, if the characteristic binding energy is much smaller than the separation between the quasiparticle peaks ( @xmath147 ) , i.e. one half of the bound state energy is close to the lowest peak ( @xmath148 ) , one can safely neglect the contribution of the higher poles and use the lowest pole approximation for the hole green s function : @xmath149 as a matter of fact , in the limiting case @xmath150 the separation of the levels is @xmath151 whereas the binding energy , as we will show , is @xmath152 .
thus , @xmath153 and the lowest pole approximation should work well . in the opposite limit
@xmath154 the binding energy is @xmath155 while the separation of the quasiparticle poles is @xmath156 . besides a factor of the order of @xmath157 , the contribution of the higher poles is strongly reduced because their quasiparticle residues are negligible ( @xmath158 ) , thus justifying the lowest pole approximation in this case .
one can then argue that the same is true for all @xmath5 . with these ideas in mind
let us look at the compact vertex @xmath106 .
= 6.5 cm ( 8,8 ) ( 0,0.7 ) the lowest order contributions to the hole - hole compact vertex @xmath106 are given in fig .
[ gam_comp ] .
@xmath159 , @xmath160 , and @xmath161 are of first , second , and third order in @xmath162 ( eq . ( [ h1 ] ) ) , respectively .
they can also be classified in terms of powers of @xmath2 and @xmath0 : @xmath163 , @xmath164 , and @xmath165 . if we consider @xmath160 only , the full vertex @xmath136 in fig .
[ bsg_fig ] is given by the sum of the ladder diagrams only .
to some extent , this approximation is analogous to the scba in the single - hole problem , and possesses an important property of the latter : all crossing diagrams and one - loop corrections are exactly zero . more importantly , such corrections to _ all _ the lowest order diagrams ( @xmath159 , @xmath160 , and @xmath161 ) are also zero because of the same spin - conservation rule . for the same reason ,
higher order corrections to fig .
[ gam_comp ] are strongly suppressed , so that the first dressing of @xmath159 , @xmath160 , and @xmath161 by magnons appears in the fifth , sixth , and seventh order , respectively .
the corresponding set of diagrams is given in fig .
[ gam_corr ] .
= 3.5 cm ( 4,4 ) ( 0,0.7 ) note that because all these corrections are of similar origin to the trugman s processes for the self - energy , one can expect that they play only a minor role in the effective interaction . however , there are other sources of correction to @xmath106 coming from the hole - two - magnon , hole - three - magnon , and magnon - magnon interactions .
most of them result in the renormalization of the hole - magnon vertices similar to those in fig .
[ diag3 ] .
while a formal account of such corrections is rather difficult , one can anticipate the result of such a renormalization using the real - space consideration of the interactions .
as in the single - hole case , the dressing of the vertex accounts for two facts .
first , the energy of a magnon in the neighborhood of a hole is @xmath166 per hole - magnon link lower than the energy of a free excitation .
second , the coordination number for hole hopping in a state with a string of magnons is @xmath167 instead of @xmath168 .
the first effect renormalizes the energy of the magnon lines in @xmath160 and @xmath161 , while the second effect renormalizes the inner hole line in @xmath161 . in what follows we will consider the compact vertex @xmath106 as given by @xmath169 in fig .
[ gam_comp ] and will include the renormalization later .
we first consider the @xmath170 limit when the solution of the bs equation ( [ bs1 ] ) is trivial . in this limit
only @xmath171 survives and the hole green s function is simply @xmath172 .
thus eq .
( [ bs1 ] ) becomes : @xmath173 evidently , the dependence of @xmath136 on @xmath174 is arbitrary and @xmath175 is a constant . integrating over
@xmath176 one finds : @xmath177 where @xmath136 plays the role of the wave function .
since the @xmath32-structure of the vertex is very simple , one can classify all solutions of eq .
( [ t_0_a ] ) in terms of the nearest - neighbor @xmath120 , @xmath3 , and @xmath4 waves : @xmath178 , @xmath179 , @xmath180 .
using the property : @xmath181 , one can see that these solutions are degenerate with energy @xmath182 .
this is expected if one remembers that while an individual hole breaks four af bonds , two nearest - neighbor holes save one bond .
a finite hopping constant @xmath0 has three effects : ( i ) the hole green s function becomes renormalized , ( ii ) @xmath160 and @xmath161 start to contribute to the interaction , ( iii ) the hole self - energy also acquire a small @xmath183-dependent contribution . in order to see their roles ,
let us incorporate these changes into the bs equation ( [ bs1 ] ) one by one .
consider the case when the hole green s function is renormalized as described in sec .
iii and is @xmath32-independent , but the interaction still comes from the @xmath159 term only .
since @xmath159 is @xmath174 and @xmath137 independent , the bs equation is similar to eq .
( [ t_0_a ] ) @xmath184 since the @xmath32-structure of the core is unchanged , all three harmonics remain degenerate .
hence @xmath185 in the lowest - pole approximation @xmath57 is given by eq .
( [ g_lp ] ) so the binding energy relative to the bottom of the hole band is : @xmath186 in the limit @xmath154 , @xmath187 . in the limit @xmath150 , @xmath188 so @xmath189 .
the binding energy ( in units of @xmath2 ) v.s .
@xmath5 is shown in fig .
[ loc_is1 ] .
this energy is always negative because we assume that the hole forms a completely localized excitation . in this case a bound state is always formed as long as the effective interaction remains attractive .
now we consider the magnon exchange interaction @xmath160 which is given by : @xmath190\ .\end{aligned}\ ] ] since @xmath160 depends on the total momentum of the pair @xmath174 , the solutions of the bs equation also depends on @xmath174 .
the energy of the bound state at different @xmath174 provides a dispersion law for the pair . in general , the lowest energy bound state
is always at zero or some finite momenta corresponding to high symmetry points in the @xmath32-space . in this work
we do not consider finite-@xmath174 bound states because their physics is unimportant for the purposes of this paper .
the bs equation ( [ bs1 ] ) at @xmath191 with interaction @xmath192 reads : @xmath193\bigg\}\hat{\gamma}_{\bf 0,p}(e,\epsilon_1)\ . \nonumber\end{aligned}\ ] ] evidently , the @xmath32 and @xmath137 dependencies of @xmath136 can be separated and one can see that the @xmath3- and @xmath4-wave solutions are not affected by the magnon - exchange interaction because @xmath194 .
this observation can be expressed in a different form : in the zero total momentum @xmath3- or @xmath4-wave state the amplitudes of the spin - flip exchange from different lobes of the wave function cancel each other . for the @xmath120-wave state , substituting @xmath195 and integrating over @xmath32
give : @xmath196\bigg\}\hat{\gamma}_e(\epsilon_1)\ .\end{aligned}\ ] ] in the so called static approximation , when the retardation due to the magnon propagator is neglected , @xmath160 is @xmath137-independent and eq .
( [ t_2_b ] ) becomes @xmath197 as in the previous case , @xmath198 .
the problem with retardation can be solved using a method introduced in ref .
one can assume that @xmath199 has no singularities @xcite and therefore the integral over @xmath176 in eq .
( [ t_2_b ] ) is defined by the poles of @xmath57 and the magnon propagator only . in the lowest - pole approximation , by using the obvious parity @xmath200 and by choosing the proper contour of integration one obtains : @xmath201\bigg\}\hat{\gamma}_e(\delta^s/2)\ .\end{aligned}\ ] ] multiplying both sides by @xmath202 and integrating over @xmath137 finally yields : @xmath203 this expression coincides with that given by the wigner - brillouin perturbation theory .
the resulting binding energies ( from @xmath192 ) v.s .
@xmath5 are shown in fig .
[ loc_is1 ] .
( 9,8.5 ) ( 0,-3 ) thus the @xmath120-wave binding energy vanishes at @xmath204 .
since the spin - polarons are assumed to be localized the disappearance of the bound state is not due to the competition between the delocalizing kinetic energy and an attraction , but rather due to the competition between different types of interaction .
one can see that the magnon exchange actually leads to an effective repulsion between the holes . in a naive point of view
the holes can be mutually attracted by following each other in a caterpillar - like process , because in this way the disturbance on the af background is only virtual . however
, this scenario does not take into account the fact that after an elementary hopping of a hole pair , the order of the holes is different from the initial one , i.e. , they are exchanged @xcite .
this in turn makes the corresponding energy correction positive , i.e. it leads to repulsion .
it is interesting to note that while the hole - magnon interaction plays a major role in the hole dressing , it has no direct effect on the @xmath3 and @xmath4 bound states and is destructive for the @xmath120-wave state .
let us now add the third part of the hole - hole interaction @xmath161 and see if the combination of the hole - magnon interaction with the two - magnon vertex causes an additional attraction .
the diagrams in fig .
[ gam_comp](c ) are equivalent to : @xmath205\ .\end{aligned}\ ] ] integration over the inner momentum @xmath206 and frequency @xmath207 yields : @xmath208\ .\end{aligned}\ ] ] one can see that the kinematic structure of this vertex is identical to @xmath159 , i.e. this term is effectively a contact nearest - neighbor interaction .
this is evident from the real - space picture of the process , which is shown in fig .
[ real_space ] .
( 2,2 ) ( 4,0.7 ) initially the holes are at nearest - neighbor sites .
one of them jumps to a neighboring site creating a spin - flip at its site of origin .
then the second hole virtually absorbs and reemits this magnon .
finally the first hole jumps back and the original state is restored .
naturally , such an interaction depends only on @xmath209 and does not depend on @xmath174 .
one can also guess the character of the resulting interaction .
being in the neighborhood of another hole , the spin excitation generated by the hopping of the first hole lowers its energy .
therefore the interaction between the holes via such a process must also be attractive . in the lowest - pole approximation @xmath161 eq . ( [ t_3_1 ] )
is given by : @xmath210\ .\end{aligned}\ ] ] one can consider the static limit of the problem by eliminating the frequency dependence of the propagators in eq .
( [ t_3_2 ] ) .
this gives @xmath211 in the @xmath150 limit this expression yields @xmath212 .
thus , in this limit the effective interaction constant of the contact interaction is renormalized from @xmath2 to @xmath0 . for the @xmath3- and
@xmath4-waves it gives a bound state energy @xmath213 , which is a factor of @xmath214 larger than the result obtained with @xmath159 only .
this result , while being quite simple , is somewhat unexpected and , to our knowledge , has never been discussed in the literature . a rigorous account for the retardation yields modified equations for the bound state energies , leaving the qualitative picture drawn in the static limit essentially unchanged : @xmath215 thus the energies of the @xmath3- and @xmath4-wave states are always negative . in the @xmath120-wave bound state
, @xmath161 can not overcome the magnon - exchange repulsion @xmath160 .
it also has little effect on @xmath216 because in the small @xmath0 region where @xmath217 goes to zero , @xmath161 is roughly 4 times smaller than @xmath160 .
we now consider the effect of the renormalization . as we noted before we have enough reasons to rule out the possibility that the trugman - like diagrams fig .
[ gam_corr ] can change the interaction constants significantly .
the anticipated smallness of these diagrams in the @xmath218 region is due to the geometrical factor , which is of the order of @xmath219 .
most of these corrections are relatively easy to evaluate using the same approximations as above .
the results of these calculations show that those terms with the same kinematic structure as in @xmath159 , @xmath160 , and @xmath161 are indeed suppressed by this small factor .
they also show that among the corrections there are terms with a more complicated kinematic structure than the original @xmath106 .
since these terms are not suppressed by the small factor , they might enhance the effective interaction .
some of these terms correspond to the contact interaction between the holes in states where both of them have `` tails '' of strings .
evidently , this leads to interaction of polarons at sites farther than nearest neighbors . for the bound state wave function it would mean that harmonics higher than the nearest - neighbor one should be considered
nevertheless , we do not expect these terms to make any qualitative change and in the following we will omit all corrections from fig .
[ gam_corr ] to avoid complications . at the same time , diagrams of the non - scba type ( fig .
[ diag3 ] ) are straightforward to include using the real - space language .
we have to take into account the lower energy of the magnons in the intermediate state in the diagrams fig .
[ gam_comp](b , c ) .
this simply changes @xmath220 in the denominators of eqs .
( [ t_3_4 ] ) to @xmath221 .
renormalization of the hole energy in the intermediate state in @xmath161 produces much weaker effect .
this renormalization finally yields : @xmath222 solutions of these equations are plotted in fig .
[ loc_is2 ] .
one can see that the energies of the @xmath3 and @xmath4 bound states are changed significantly from those in fig .
[ loc_is1 ] .
( 9,8.5 ) ( 0,-3 ) thus , the hierarchy of terms in the effective hole - hole interaction emerged from our study is as follows : + ( i ) the contact `` sharing common link '' attraction yields bound states with @xmath120- , @xmath3- , @xmath4-symmetry with energy @xmath223 at small @xmath5 ; + ( ii ) the magnon - exchange interaction has no effect on the @xmath191 @xmath3- and @xmath4-wave states because of symmetry reasons , and is strongly repulsive for the @xmath120-wave state ; + ( iii ) the hole - magnon attraction , which leads to an attraction between a hole and a string of another hole , is the strongest interaction in the present problem in the physical limit ( @xmath224 ) .
it leads to strong @xmath3 and @xmath4 bound states with energy @xmath225 .
now , after all essential interactions are taken into account and carefully analyzed , the last effect which has to be incorporated into the bs equation before we can make comparison with numerical results is the @xmath32-dependent contributions to the hole green s function . as noted in sec .
iii such contributions are small , but lead to a coherent hole band . therefore the ed data show some dispersion in the hole energy .
for instance , at @xmath118 the energy difference between the @xmath117 and @xmath116 points ( almost exactly half of the bandwidth ) is @xmath226 .
although this difference is small compared to the absolute value of the lowest pole energy , @xmath227 , or the separation between the energy levels , @xmath228 , it is as large as the hole - hole binding energy calculated above , @xmath229 . therefore , one would expect this dispersion to have strong influence on the pairing of holes .
one can also foresee that such an influence is likely to be destructive , because now the pairing is not only an interplay between repulsions and attractions , but also the question of how much kinetic energy is lost if the holes are in a bound state . for our purposes we will need the hole green s function in the lowest - pole approximation , which now reads :
@xmath230 where @xmath231 , @xmath232 , @xmath233 , and we have neglected the @xmath183-dependence of @xmath234 .
the role of the trugman s loops larger than an elementary square is negligible , so we consider the dispersion given by the next - nearest - neighbor hopping only .
the resulting band is isotropic and has a minimum at the center of the bz . now in the bs equation ( [ bs1 ] ) it is easier to integrate over @xmath176 first . neglecting the effect of the dispersion on the retardation and after some algebra one obtains : @xmath235 + \gamma_{\bf k}\gamma_{\bf p}\frac{16t^2}{j(\delta - j)}\bigg\ } \ .\end{aligned}\ ] ] where @xmath236 is a nearest - neighbor @xmath120- , @xmath3- , or @xmath4-wave . as before the magnon - exchange interaction ( the second term in the curly bracket )
is orthogonal to the @xmath3- and @xmath4-wave states .
( [ t_4 ] ) is an almost evident generalization of eqs .
( [ t_3_5 ] ) to the case of finite hole dispersion .
the only new parameter introduced is the hole bandwidth @xmath237 .
our numerical ed data for @xmath238 as a function of @xmath5 calculated on a 32-site lattice are given in fig .
we do not calculate this quantity analytically .
instead we treat it as a parameter and use the numerical values in fig [ w ] .
( 9,8.5 ) ( 0,-4 ) two statements can be made about the solutions of eq .
( [ t_4 ] ) . now all three symmetries have thresholds @xmath239 above which negative energy solution of the corresponding symmetry cease to exist .
also , the @xmath3- and @xmath4-waves are split with the @xmath3 wave state having lower energy .
the first statement is evident if one consider the threshold condition ( @xmath240 ) for the @xmath3- or @xmath4-wave state in eq .
( [ t_4 ] ) , assuming @xmath150 so that @xmath241 : @xmath242 the sum over @xmath132 gives a number of the order of unity .
thus , there is a value of @xmath237 ( @xmath243 ) at which the equality condition is satisfied . if @xmath237 grows with @xmath0 , as it in fact does , then above this point eq .
( [ t_4 ] ) has no solution . using the dependence of @xmath237 on @xmath5 ( fig .
[ w ] ) , the critical values of @xmath244 can be extracted from the same condition eq .
( [ t_4_1 ] ) .
the second statement is a consequence of the isotropic form of the dispersion law .
the sum over @xmath132 in eq .
( [ t_4_1 ] ) can be rewritten as : @xmath245=1+c_1+c_2+\cdots\end{aligned}\ ] ] in the case of @xmath4 wave the series is alternating with @xmath246 and @xmath247 . in the case of @xmath3 wave @xmath248 for odd @xmath88 , and @xmath249 for even @xmath88 .
numerical integration gives @xmath250 , @xmath251 , and therefore @xmath4 wave solution disappears at a lower value of @xmath5 .
numerical solutions of eq .
( [ t_4 ] ) for all three symmetries together with results of our 32-site ed calculations are shown in fig .
[ s_p_d ] . for the purpose of comparison
we also show the results obtained by a modified lanczos method on a 50-site cluster in ref .
@xcite .
( 9,8.5 ) ( 0,-3 ) one can see a very good _ quantitative _ agreement between our analytical and numerical results for the @xmath120-wave state . for the @xmath4- and @xmath3-states
, they have the same qualitative behaviors .
it is clear that our eq .
( [ t_4 ] ) , while involving different kinds of approximations , captures all essential effects of the pairing problem in the @xmath0-@xmath1 model in a very good quantitative level .
it also describes the problem in a very easy and transparent way .
namely , all essential tendencies in the hole binding are described by this equation : ( i ) the @xmath120-wave is pushed up to the continuum by the magnon - exchange process , ( ii ) the @xmath3- and @xmath4-waves are much more stable , ( iii ) the trugman s dispersion split these two levels and push them to the continuum at some larger critical values of @xmath5 .
the analytical results underestimate the pairing strength , especially in the region @xmath252 .
we believe that this is due to the contribution from the higher poles which are neglected in our study . in this region
a significant portion of the spectral weight is transfered to these poles while the separation between them is still of the order of @xmath52 .
since our lowest - pole approximation uses the ratio @xmath253 ( @xmath254 in this region ) as the small parameter , one would expect the energy correction from the higher poles to be of the same order . in the region where the binding energy is close to zero
we expect this approximation to be more accurate .
however , for the @xmath3 and @xmath4 waves this is also the region of @xmath255 where the omitted higher order diagrams with more complicated kinematic structure contribute to the interaction .
we refer the smaller splitting of the @xmath3- and @xmath4-wave states found in our analytical results compared to those in our ed data to the effect of these corrections .
the same higher order interactions cause a tiny splitting of the states at @xmath256 which is shown in fig .
[ s_p_d ] .
we remark that a @xmath5 dependence of the binding energies very similar to our ed results has been found recently by series - expansion calculations @xcite .
thus we believe that the agreement between the results can be improved by an accurate consideration of the contribution of higher poles and higher order diagrams of various types . however , this is out of the scope of our paper .
it is natural and instructive to consider the case where small transverse spin fluctuations @xmath257 ( with @xmath258 ) are included in the hamiltonian .
this helps one to see the effects of these fluctuations on the pairing problem and to infer the tendencies in the isotropic @xmath0-@xmath2 model . as noted in ref .
@xcite only two changes are necessary to account for these fluctuations up to first order in @xmath259 .
first , the holes acquire a dispersion law given by @xmath260 where @xmath26 is of the order of @xmath261 in the @xmath154 limit and of the order of @xmath2 in the @xmath150 limit @xcite .
such a band is strongly anisotropic , i.e. degenerate minima form a line along the magnetic bz boundary . in the full @xmath0-@xmath2 model ( @xmath262 )
this dispersion dominates over the trugman s terms at all @xmath5 .
second , the hole - magnon interaction becomes : @xmath263 \ .\end{aligned}\ ] ] corrections to the magnon energy are of higher order in @xmath259 : @xmath264 . as we shall see
both changes lead to the splitting of the @xmath3- and @xmath4-wave levels , and both favor the @xmath4-wave as the ground state of the system .
consider eq .
( [ t_4 ] ) with the dispersion from eq .
( [ t_5 ] ) and the hole - magnon exchange term coming from eq .
( [ t_5_1 ] ) . for the purpose of illustration
it is sufficient to consider the interactions in the static limit : @xmath265
-\frac{16t^2}{j^2}\bigg[\gamma_{\bf k}\gamma_{\bf p}-\frac{\alpha}{2 } \gamma_{\bf k+p}(\gamma_{\bf k}^2+\gamma_{\bf p}^2)\bigg ] \bigg\ } \ .\end{aligned}\ ] ] corrections to the bound state energies of the @xmath3 and @xmath4 wave up to first order in @xmath259 are : @xmath266 where the first term in the bracket comes from the dispersion and the second is from the magnon - exchange interaction .
the kinetic energy pushes up the energy of both states , but for the @xmath3-wave this shift is three times larger .
an important observation is that the magnon - exchange interaction is now attractive for the @xmath4-wave and repulsive for the @xmath3-wave state ! to show how subtle the selection between the @xmath3- and @xmath4-wave ground state is , let us consider a more `` realistic '' dispersion law : @xmath267 it is more realistic in the sense that it has minima at @xmath268 points which are isotropic and it agrees with the experimentally measured profile in the cuprates better than the one in eq.([t_5 ] ) .
it is now known that this difference between the @xmath0-@xmath2 model dispersion law ( eq.([t_5 ] ) ) and the one observed in reality ( eq.([t_5_4 ] ) ) is due to additional intrasublattice hoppings ( @xmath269 etc . ) in the real systems @xcite . with the dispersion law in eq .
( [ t_5_4 ] ) corrections to the binding energies are given by : @xmath270 now the kinetic energy pushes up the @xmath4-wave energy @xmath271 times more than that of the @xmath3-wave and @xmath272 $ ] is not necessarily positive .
therefore , @xmath3-wave can become a ground state again .
summarizing our observations in the small @xmath259 limit , we conclude that the transverse fluctuations create an additional attractive interaction in the @xmath4-symmetry state and that an anisotropic kinetic energy profile also favors @xmath4-wave .
thus , we suggest that the @xmath4-wave is most likely to be the ground state of the system in the @xmath0-@xmath2 model ( @xmath262 ) .
this conclusion is in agreement with many analytical and numerical studies .
nevertheless , recent 32-site ed study showed that at @xmath273 the ground state of the two - hole @xmath0-@xmath2 model is a @xmath3-wave state with total momentum @xmath274 @xcite .
( in our case momenta @xmath275 and @xmath274 are equivalent because of the existence of long - range order . )
furthermore , such a behavior is also suggested by recent series - expansion calculations @xcite . due to the apparently strong finite - size effects on the ed results ,
it is difficult to conclude whether this behavior will survive in the thermodynamic limit .
nevertheless , these new data show that even in the pure @xmath0-@xmath2 model the final word on the problem of the symmetry of the bound state is still to come .
besides effects arising from the @xmath0-@xmath2 model there are also concerns about the robustness of the bound states in the real system .
the characteristic bound - state energy found in the @xmath0-@xmath2 model is of the order of @xmath276 in the physical region of @xmath5 .
this energy is far below the accuracy of the model itself .
this means that in order to describe the real materials , the low - energy model must include at least additional hopping terms ( @xmath269 etc . ) .
it is worth noting again that because of these terms the realistic dispersion is isotropic , which , as we have just shown , can favor @xmath3-wave .
more important conjecture can be made using the results of ref .
@xcite and new ed data @xcite .
since the isotropic dispersion involves higher kinetic energy , the overall pairing tendency is weakened by these additional hopping terms .
it can push all the bound states up to the continuum , thus destroying the foundation of the preformed pair and phase separation scenarios well before the physical region of parameters is reached .
in this paper we attempt to address the issue of how generic @xmath4-wave pairing in @xmath0-@xmath2-like models is and to clarify the problem of how this symmetry is selected among others .
we have thoroughly examined the @xmath0-@xmath1 model by means of analytical diagrammatic study supplemented by our 32-site ed results .
this formalism is tested in the one - hole problem and perfect agreement with ed results is found .
the ground state of two holes in the physical region @xmath277 is found to be a @xmath3-wave bound state . from a careful analysis of the interactions involved in the pairing problem we established that the pairing in the @xmath120-wave channel is highly unfavorable because of the repulsive character of the magnon - exchange interaction in this state .
while this reason is rather generic , the problem of selecting between the @xmath3- and @xmath4-wave is much more subtle . in our analytical study
these states remain degenerate until the hole dispersion is included . in the @xmath0-@xmath1 model
such a dispersion comes from higher order hopping processes and is , in fact , very small .
however , it has important consequences .
it is responsible not only for the significant splitting between the @xmath3- and @xmath4-wave states , but also for the destruction of hole pairing at some critical @xmath278 .
the isotropic form of the dispersion favors @xmath3-wave as the ground state of the system .
we have also considered the role of transverse fluctuations in the perturbative limit @xmath258 .
they lead to an anisotropic hole dispersion and an additional hole - hole interaction .
we have shown that both features favor @xmath4-wave as the lowest energy state . however , in this case the selection between @xmath3- and @xmath4-wave strongly relies on the anisotropic form of the hole dispersion , and such a dispersion does not agree with the experimentally measured dispersion in real systems . moreover , since the characteristic energy of the bound states found in the @xmath0-@xmath2 models is small , we are skeptical that these states can survive when additional kinetic energy terms are included in the model
. our skepticism is supported by earlier studies @xcite and recent numerical data @xcite .
such conclusions for the pairing problem in @xmath0-@xmath2-like models drawn from our study create a framework in which it might be necessary to look for @xmath4-wave superconductivity unrelated to the bose - condensation of preformed @xmath4-wave pairs . furthermore , to make the phase separation scenario possible , one might need to invoke arguments other than an _ a priori _ existing collapsing tendency in the spin - hole system .
we are indebted to prof .
r. j. gooding for suggesting this work and for valuable discussions and comments .
we also thank f. marsiglio for illuminative discussion , and a. castro neto , m. zhitomirsky , and o. starykh for comments .
the work of a.l.c . was supported in part by the nserc of canada and the university of california and los alamos national laboratory under the auspices of the us department of energy .
the work of p.w.l . was supported by the hong kong rgc grant hkust6144/97p .
numerical diagonalizations of the 32-site systems were performed on the intel paragon at hkust .
j. h. jefferson , h. eskes , and l. f. feiner , phys .
b * 45 * , 7959 ( 1992 ) ; m. e. simon and a. a. aligia , _ ibid .
_ * 48 * , 7471 ( 1993 ) ; v. i. belinicher and a. l. chernyshev , _ ibid . _ * 49 * , 9746 ( 1994 ) ; v. i. belinicher , a. l. chernyshev , and l. v. popovich , _ ibid . _ * 50 * , 13768 ( 1994 ) ; v. i. belinicher , a. l. chernyshev , and v. a. shubin , _ ibid . _ * 53 * , 335 ( 1996 ) .
the true assumption is weaker .
one can require @xmath175 to be the sum of a constant term and singular terms , latter having poles far from the poles of the hole green s functions in the integral .
approximating @xmath175 by the constant term is valid provided that @xmath279 . | the @xmath0-@xmath1 model is the strongly anisotropic limit of the @xmath0-@xmath2 model which captures some general properties of the doped antiferromagnets ( af ) .
the absence of spin fluctuations simplifies the analytical treatment of hole motion in an af background and allows us to calculate the single- and two - hole spectra with high accuracy using regular diagram technique combined with real - space approach . at the same time
, numerical studies of this model via exact diagonalization ( ed ) on small clusters show negligible finite size effects for a number of quantities , thus allowing a direct comparison between analytical and numerical results .
both approaches demonstrate that the holes have tendency to pair in the @xmath3- and @xmath4-wave channels at realistic values of @xmath5 .
the interactions leading to pairing and effects selecting @xmath3 and @xmath4 waves are thoroughly investigated .
the role of transverse spin fluctuations is considered using perturbation theory .
based on the results of the present study , we discuss the pairing problem in the realistic @xmath0-@xmath2-like model .
possible implications for preformed pairs formation and phase separation are drawn . |
many signals for new physics at hadron colliders involve electroweak gauge bosons ( @xmath9 ) and jets containing heavy quarks ( @xmath10 ) .
the prime example is @xmath11 jets , with one or more jets containing a @xmath12 tag , which led to the discovery of the top quark @xcite .
other examples include signals for the higgs boson and the superpartners of the known particles @xcite .
it is therefore crucial to understand the standard - model background from the production of electroweak bosons and heavy - quark jets with good accuracy .
the simplest processes of this type are the production of a single electroweak gauge boson and one heavy - quark jet . in the case of the @xmath13 boson ,
the leading - order process is @xmath14 @xcite , which has been calculated at next - to - leading order ( with @xmath15 nonzero ) @xcite . for a photon ,
the leading - order process is @xmath16 ( @xmath2 ) , which has also been calculated at next - to - leading order ( with @xmath17 ) @xcite . in this paper
we consider the analogous leading - order process for the @xmath0 boson , @xmath1 , shown in fig .
[ fig : gqzq ] , which we calculate at next - to - leading order ( with @xmath17 ) .
an alternative calculational scheme is to regard @xmath18 as the leading - order process ( with @xmath19 nonzero ) , and to allow one heavy quark to be emitted collinear to the beam , yielding a @xmath3 final state .
this approach has two drawbacks .
first , the expansion parameter of this calculation is @xmath20 rather than @xmath21 , so perturbation theory is less convergent . using a heavy - quark distribution function sums these collinear logarithms to all orders , resulting in a perturbative expansion in @xmath21 and @xmath22 @xcite .
second , it is much more difficult to obtain the next - to - leading - order correction to @xmath18 than to @xmath1 , both because there is one more particle in the final state , and because the heavy - quark mass must be maintained in the calculation of @xmath18 to regulate the collinear region ( one may set @xmath17 in @xmath1 , provided that the heavy - quark transverse momentum is much larger than its mass ) . at present , @xmath18 is known at next - to - leading order only for @xmath17 @xcite .
in contrast , the next - to - leading - order calculation of @xmath1 ( with @xmath17 ) that we perform in this paper is sufficiently straightforward that the next - to - next - to - leading - order calculation may be available in the foreseeable future @xcite .
the process @xmath1 is a background to @xmath23 , where the @xmath0 boson and the higgs boson decay to the same final state ( @xmath24 , @xmath25 , or @xmath26 ) @xcite .
in addition , @xmath1 could potentially be used to measure the @xmath27 distribution function .
the @xmath12 distribution function is needed for the above - mentioned higgs - production process as well as for inclusive higgs production , @xmath28 @xcite .
it is also needed for the single - top production processes @xmath29 @xcite and @xmath30 @xcite , and the charged - higgs production process @xmath31 @xcite .
the process @xmath32 is much more sensitive to the charm distribution function than to that of bottom , due to the greater electric charge of the charm quark @xcite . at present , the @xmath12 distribution function is derived perturbatively from the gluon distribution function @xcite and there is no direct measurement of it . the @xmath33 distribution function is similarly derived from the gluon distribution function , and it agrees well with direct measurements .
thus we expect the same to be true of the @xmath12 distribution function .
the uncertainty in the @xmath12 distribution function derives from the uncertainty in the gluon distribution function .
another source of @xmath3 events is @xmath34 , shown in fig .
[ fig : qqzqq ] , where either one @xmath27 is missed , or the two @xmath27 s coalesce into a single jet .
we show that this process is much more significant at the tevatron than at the lhc , since light - quark distribution functions are relatively more important at large values of bjorken @xmath35 .
this calculation is carried out at leading order with a non - zero @xmath27 mass .
in addition , we also calculate @xmath36 production at next - to - leading order , where @xmath37 denotes a light - quark or gluon jet , as shown in fig .
[ fig : zj ] @xcite . using a silicon vertex detector to tag heavy quarks , the probability that such a jet fakes a heavy - quark jet is around @xmath38 . taking this probability into account
, we show that this source of fake @xmath3 events is comparable to genuine @xmath3 events at the tevatron , but is relatively less important at the lhc .
all of the above processes can also lead to final states with two jets , with varying numbers of heavy quarks . for completeness
, we also calculate these cross sections ( at leading order ) . when combined with the next - to - leading - order cross section for @xmath3 , one obtains the inclusive cross section for @xmath0 plus at least one heavy - quark jet ( at next - to - leading order ) .
the next - to - leading - order calculations in this paper were performed with the monte - carlo code mcfm @xcite .
the leading - order calculations were performed both with this code and with madevent @xcite .
the next - to - leading - order ( nlo ) calculation of @xmath1 parallels our nlo calculation of @xmath23 @xcite , and we refer the reader to that work for a detailed discussion of the calculational scheme .
the contributing subprocesses are * @xmath1 ( one loop ) * @xmath39 * @xmath40 * @xmath18 the subprocess @xmath34 is considered separately in the following section , and is not regarded as a correction to @xmath1 .
we work in the simplified acot scheme @xcite , which allows one to neglect the @xmath27 mass throughout .
this is a good approximation , and simplifies the calculation .
the error made by this approximation is proportional to @xmath41 @xcite .
we use the dipole - subtraction method @xcite as formulated in ref .
@xcite to isolate and subtract collinear divergences . in our nlo calculation of @xmath1
, we demand one and only one jet with transverse momentum @xmath42 gev within a rapidity range @xmath43 at the fermilab tevatron .
this jet must contain a heavy quark . at the cern large hadron collider ( lhc ) , the heavy - quark jet must have @xmath42 gev and @xmath44 .
if two partons lie within a cone of radius @xmath45 , we merge them into a single jet with four - momentum equal to the sum of the two partons four - momenta .
this is done before the cuts described above are applied to the jets .
the nlo process @xmath18 yields two heavy quarks in the final state .
if they are merged into a single jet , we record it as a double - heavy - quark jet .
this is only about 1% of the @xmath3 cross section .
similarly , the nlo process @xmath46 , and processes related by crossing , yield two heavy quarks in the final state .
however , these processes amount to a correction of less than 1% , so we neglect them @xcite .
we list in the first column of tables [ tab : tev ] and [ tab : lhc ] the lo ( in parentheses ) and nlo cross sections for @xmath47 and @xmath48 at the tevatron and the lhc .
the @xmath49 cross section exceeds that of @xmath50 by 70% at the tevatron and 35% at the lhc because the charm distribution function is larger than that of bottom .
this is partially compensated by the fact that the @xmath0 has stronger coupling to bottom than to charm .
the ratio of the @xmath50 and @xmath49 partonic cross sections is proportional to @xmath51 more importantly , a silicon vertex detector ( svx ) can tag a @xmath12 jet with an efficiency around 60% , and a @xmath33 jet with an efficiency of about 15% .
thus the majority of @xmath3 events tagged with an svx come from @xmath50 .
the nlo processes that contribute to @xmath3 also give rise to final states with more than one jet , and we list these cross sections in the remaining columns .
these events are grouped in three classes .
the second column , labeled @xmath52 , corresponds to events with a single jet that contains two heavy quarks . as mentioned earlier , this is only about 1% of the @xmath3 cross section .
the third column , labeled @xmath53 , corresponds to events with two jets , one of which contains a heavy quark .
this is about @xmath54 of the @xmath3 cross section at the tevatron , and about @xmath55 at the lhc ( for @xmath42 gev ) .
the fourth column , labeled @xmath56 , corresponds to events with two jets , both of which contain heavy quarks .
this is significantly less than @xmath53 .
the @xmath56 events arise from the nlo process @xmath57 ( as do the @xmath52 events ) .
the final column sums these various processes , to give the inclusive cross section for @xmath0 plus at least one heavy - quark jet at next - to - leading order .
we show in figs .
[ fig : mudep_tevb][fig : mudep_lhcc_lolum ] the scale dependence of the inclusive cross section for @xmath50 and @xmath49 production at the tevatron and the lhc at leading order ( lo ) and next - to - leading order ( nlo ) .
the dependence on the renormalization and factorization scales are shown separately .
both scale dependencies are reduced at nlo in comparison with lo .
we use @xmath58 , @xmath59 as our default value of the renormalization and factorization scales .
we also estimate the uncertainty in the nlo inclusive cross section .
the first uncertainty is due to varying the renormalization scale between half and twice its default value of @xmath58 .
the second uncertainty , obtained in the same manner , is due to the factorization scale .
the third uncertainty is from the parton distribution functions @xcite .
there is also an additional uncertainty of 4% due to @xmath60 .
@xcite we show in figs .
[ fig : pt_b_tev][fig : pt_b_lhc ] the transverse - momentum distributions of both the @xmath0 and the highest-@xmath61 @xmath12 jet at the tevatron and the lhc . at leading order
these distributions are identical , since the @xmath0 recoils against the @xmath12 jet . at next - to - leading order
the distributions for the inclusive cross section are slightly different , since the @xmath0 can recoil against two jets .
the distributions for @xmath49 production are qualitatively similar .
.cross sections ( pb ) for @xmath0-boson production in association with heavy - quark jets at the tevatron ( @xmath62 tev @xmath63 ) . a jet lies in the range @xmath42 gev and @xmath43 .
two final - state partons are merged into a single jet if @xmath64 .
no branching ratios or tagging efficiencies are included .
the labels on the columns have the following meaning : @xmath65 exactly one jet , which contains a heavy quark ; @xmath66 exactly one jet , which contains two heavy quarks ; @xmath67 exactly two jets , one of which contains a heavy quark ; @xmath68 exactly two jets , both of which contain a heavy quark . for the last set of processes , the labels mean : @xmath69 exactly one jet , which does not contain a heavy quark ; @xmath70 exactly two jets , neither of which contain a heavy quark . for @xmath3 and @xmath36 , both the leading - order ( in parentheses ) and
next - to - leading - order cross sections are given .
the last column is the next - to - leading - order inclusive cross section , which is the sum of four previous columns .
the cteq6 m parton distribution functions are used throughout , except for the lo cross sections in parentheses , where cteq6l1 is used @xcite . the factorization and renormalization scales
are chosen as @xmath71 .
the uncertainties are from the variation of the renormalization scale , the factorization scale , and the parton distribution functions , respectively . [
cols="^,^,^,^,^,^ " , ] [ tab : lhc ]
another contribution to @xmath3 production comes from @xmath34 ( @xmath72 ) , shown in fig .
[ fig : qqzqq ] , where one @xmath27 is outside the coverage of the detector .
the dominant contribution to the cross section comes from diagrams in which the @xmath0 is radiated from the initial - state quarks while the heavy quarks arise from gluon splitting , as shown in fig .
[ fig : qqzqq](a ) .
we maintain the heavy - quark mass throughout the calculation in order to regulate the singularity that would arise from a gluon splitting to massless collinear quarks .
although the nlo cross section for this process with massless quarks is available @xcite , the cross section with massive quarks is known only at lo .
the lo cross section for @xmath3 production via @xmath34 is given in the first column of tables [ tab : tev ] and [ tab : lhc ] .
at the tevatron , the cross section is about @xmath73 of @xmath1 . at the lhc ,
the contribution from @xmath34 is relatively much less , only about @xmath74 of @xmath1 .
this reflects the fact that this process is initiated by a @xmath75 collision . at the tevatron , where the typical values of bjorken @xmath35 are relatively large , the valence quark distribution functions are significant . in contrast , the typical values of @xmath35 are relatively small at the lhc , so the valence quark distribution functions are less important .
the second column in tables [ tab : tev ] and [ tab : lhc ] lists the contribution from @xmath34 when the @xmath27 and @xmath76 merge into a single jet .
this is comparable in size to the contribution of this process to @xmath3 production .
this is due to the aforementioned enhancement that arises when the heavy quarks are collinear .
this process is the dominant source of events where the @xmath0 is accompanied by a single jet that contains two heavy quarks .
the last column gives the contribution of @xmath34 when both jets are within the coverage of the detector .
the @xmath77 cross section is about 1/2 of the @xmath50 cross section from this process at both machines , and the @xmath78 cross section is about 1/3 of the @xmath49 cross section .
this is in contrast with the @xmath1 process , where the @xmath56 final state ( that arises at nlo ) is much less than @xmath3 .
a light - quark or gluon jet can fake a heavy - quark jet . with a silicon vertex detector ,
the mistagging rate is typically around 1% .
since the cross section for @xmath36 production is much greater than that for @xmath3 , it is potentially a large source of fake @xmath3 events .
we list in tables [ tab : tev ] and [ tab : lhc ] the lo ( in parentheses ) and nlo cross sections for @xmath36 production .
the lo processes are @xmath79 ( @xmath81 ) and @xmath80 ( @xmath72 ) , as shown in fig .
[ fig : zj ] .
the nlo processes also contribute to the @xmath82 final state , and we list that cross section in the tables as well .
the @xmath36 cross section is almost two orders of magnitude greater than that of @xmath1 at the tevatron .
taking into account the 1% mistagging rate , we see that @xmath36 is a significant source of fake @xmath3 events at the tevatron . in contrast , at the lhc the @xmath36 cross section is not nearly as significant .
thus there will be relatively fewer mistagged events at the lhc .
qualitatively similar results are obtained for the ratio of @xmath82 to @xmath53 .
the dominant contribution to @xmath50 production is @xmath47 . at the tevatron
, @xmath83 also makes a significant contribution .
combining these two contributions , the total inclusive cross section for @xmath50 production at the tevatron is about 20 pb .
thus there are about 2000 inclusive @xmath50 events produced within the coverage of the detector at the tevatron for every 100 pb@xmath84 of integrated luminosity delivered .
the cleanest decay mode of the @xmath0 boson is to leptons ( @xmath85 ) , with a branching ratio of 6.7% . including a @xmath12-tagging efficiency of 60% yields about 80 tagged @xmath86 events for every 100 pb@xmath84 .
there are slightly more tagged events when one accounts for the fact that some events contain two heavy quarks , either within the same jet or in separate jets . with data sets between 4000 and 8000 pb@xmath84 expected in run ii at the tevatron
, there will be between 3200 and 6400 tagged @xmath87 events . at the lhc ,
the contribution from @xmath83 is much less significant , so most of the cross section comes from @xmath47 .
the total inclusive cross section for @xmath50 production is about 1090 pb , a factor of about 50 larger than at the tevatron .
thus there will be an enormous number of @xmath50 events at the lhc .
the total inclusive cross section for @xmath49 production at the tevatron is about 70% greater than that of @xmath50 , while at the lhc , @xmath49 is about 35% greater than @xmath50 .
however , the svx tagging efficiency for charm is about 1/4 of the @xmath12-tagging efficiency , so there will be fewer tagged @xmath49 events than tagged @xmath50 events . in contrast , the number of @xmath88 events far outweighs the number of @xmath89 events at both machines , since charm has twice the electric charge of bottom and the charm distribution function is larger than that of bottom . even taking the greater tagging efficiency of bottom into account , the number of tagged @xmath88 events is larger than the number of tagged @xmath89 events @xcite .
@xmath36 events , where the jet is mistagged as a heavy quark , are a significant source of fake @xmath3 events at the tevatron , but much less so at the lhc .
thus a larger fraction of the tagged @xmath3 events at the lhc will be from genuine heavy - quark production .
for this reason , and also due to the relatively smaller contribution from @xmath34 , the lhc provides a cleaner environment for the extraction of the heavy - quark distribution functions via @xmath1 .
in addition to the decay @xmath90 , there will be many @xmath91 events ( @xmath92 ) , which will yield heavy - quark monojets .
this will also yield dijet events with large missing transverse momentum , with one or both jets containing heavy quarks .
regardless of the @xmath0 decay mode , the majority of @xmath93 jet events with a single heavy - quark tag at the lhc come from @xmath53 , not @xmath56 . at the tevatron , where @xmath34 is relatively more important , the @xmath53 and @xmath56 final states are comparable in size .
we are grateful for conversations and correspondence with l. christofek , t. junk , t. liss , k. pitts , d. stuart , m. whalley , and j. womersley .
f. m. warmly thanks the department of physics of the `` terza universit di roma '' for the kind hospitality and support .
s. w. thanks the aspen center for physics for hospitality .
this work was supported in part by the u. s. department of energy under contracts nos .
de - ac02 - 76ch03000 and de - fg02 - 91er40677 and by the national science foundation under grant no .
phy99 - 07949 .
99 f. abe _ et al . _
[ cdf collaboration ] , phys .
* 74 * , 2626 ( 1995 ) [ arxiv : hep - ex/9503002 ] .
s. abachi _ et al . _
[ d0 collaboration ] , phys .
lett . * 74 * , 2632 ( 1995 ) [ arxiv : hep - ex/9503003 ] .
d. acosta _ et al .
_ [ cdf collaboration ] , phys . rev .
d * 65 * , 052007 ( 2002 ) [ arxiv : hep - ex/0109012 ] . m. carena and j. lykken , `` physics at run ii : supersymmetry / higgs workshop , '' fermilab - pub-00 - 349 _ prepared for physics at run ii : workshop on supersymmetry / higgs : 1st general meeting , batavia , illinois , 14 - 16 may 1998 . _
u. baur , f. halzen , s. keller , m. l. mangano and k. riesselmann , phys .
b * 318 * , 544 ( 1993 ) [ arxiv : hep - ph/9308370 ] .
w. t. giele , s. keller and e. laenen , phys .
b * 372 * , 141 ( 1996 ) [ arxiv : hep - ph/9511449 ] .
e. l. berger and l. e. gordon , phys .
d * 54 * , 2279 ( 1996 ) [ arxiv : hep - ph/9512343 ] .
b. bailey , e. l. berger and l. e. gordon , phys .
d * 54 * , 1896 ( 1996 ) [ arxiv : hep - ph/9602373 ] .
m. a. aivazis , j. c. collins , f. i. olness and w. k. tung , phys .
d * 50 * , 3102 ( 1994 ) [ arxiv : hep - ph/9312319 ] . j. c. collins , phys .
d * 58 * , 094002 ( 1998 ) [ arxiv : hep - ph/9806259 ] .
t. stelzer , z. sullivan and s. willenbrock , phys .
d * 56 * , 5919 ( 1997 ) [ arxiv : hep - ph/9705398 ] . j. m. campbell and r. k. ellis , phys .
d * 62 * , 114012 ( 2000 ) [ arxiv : hep - ph/0006304 ]
. j. campbell and r. k. ellis , phys .
d * 65 * , 113007 ( 2002 ) [ arxiv : hep - ph/0202176 ] .
j. campbell , r. k. ellis and d. rainwater , arxiv : hep - ph/0308195 .
e. w. n. glover , nucl .
phys . proc .
suppl . * 116 * , 3 ( 2003 ) [ arxiv : hep - ph/0211412 ] .
d. choudhury , a. datta and s. raychaudhuri , arxiv : hep - ph/9809552 .
c. s. huang and s. h. zhu , phys .
d * 60 * , 075012 ( 1999 ) [ arxiv : hep - ph/9812201 ] .
j. campbell , r. k. ellis , f. maltoni and s. willenbrock , phys .
d * 67 * , 095002 ( 2003 ) [ arxiv : hep - ph/0204093 ] .
d. dicus , t. stelzer , z. sullivan and s. willenbrock , phys .
d * 59 * , 094016 ( 1999 ) [ arxiv : hep - ph/9811492 ] . c. balazs , h. j. he and c. p. yuan , phys .
d * 60 * , 114001 ( 1999 ) [ arxiv : hep - ph/9812263 ] .
f. maltoni , z. sullivan and s. willenbrock , phys .
093005 ( 2003 ) [ arxiv : hep - ph/0301033 ] . t. stelzer , z. sullivan and s. willenbrock , phys .
d * 58 * , 094021 ( 1998 ) [ arxiv : hep - ph/9807340 ] .
a. p. heinson , a. s. belyaev and e. e. boos , phys .
d * 56 * , 3114 ( 1997 ) [ arxiv : hep - ph/9612424 ] .
s. zhu , phys .
b * 524 * ( 2002 ) 283 [ erratum - ibid .
b * 537 * ( 2002 ) 351 ] .
v. d. barger , r. j. n. phillips and d. p. roy , phys .
b * 324 * , 236 ( 1994 ) [ arxiv : hep - ph/9311372 ] .
s. h. zhu , phys .
d * 67 * , 075006 ( 2003 ) [ arxiv : hep - ph/0112109 ]
. t. plehn , phys .
d * 67 * , 014018 ( 2003 ) [ arxiv : hep - ph/0206121 ] .
j. pumplin , d. r. stump , j. huston , h. l. lai , p. nadolsky and w. k. tung , jhep * 0207 * , 012 ( 2002 ) [ arxiv : hep - ph/0201195 ] .
f. abe _ et al .
_ [ cdf collaboration ] , phys . rev .
lett . * 83 * , 3124 ( 1999 ) [ arxiv : hep - ex/9810031 ] .
t. affolder _ et al . _
[ cdf collaboration ] , phys .
d * 65 * , 052006 ( 2002 ) [ arxiv : hep - ex/0106012 ] .
p. b. arnold and m. h. reno , nucl .
b * 319 * , 37 ( 1989 ) [ erratum - ibid .
b * 330 * , 284 ( 1990 ) ] .
r. j. gonsalves , j. pawlowski and c. f. wai , phys .
d * 40 * , 2245 ( 1989 ) .
w. t. giele , e. w. glover and d. a. kosower , nucl .
b * 403 * , 633 ( 1993 ) [ arxiv : hep - ph/9302225 ] .
r. k. ellis _ et al .
_ [ qcd tools working group collaboration ] , arxiv : hep - ph/0011122 . f. maltoni and t. stelzer , jhep * 0302 * , 027 ( 2003 ) [ arxiv : hep - ph/0208156 ] .
r. k. ellis , d. a. ross and a. e. terrano , nucl .
b * 178 * , 421 ( 1981 ) .
s. catani and m. h. seymour , nucl .
b * 485 * , 291 ( 1997 ) [ erratum - ibid .
b * 510 * , 503 ( 1997 ) ] [ arxiv : hep - ph/9605323 ] .
k. hagiwara _ et al .
_ [ particle data group collaboration ] , phys . rev .
d * 66 * , 010001 ( 2002 ) . | the leading - order process for the production of a @xmath0 boson and a heavy - quark jet at hadron colliders is @xmath1 ( @xmath2 ) .
we calculate this cross section at next - to - leading order at the tevatron and the lhc , and compare it with other sources of @xmath3 events .
this process is a background to new physics , and can be used to measure the heavy - quark distribution function .
* associated production of a @xmath0 boson and a + single heavy - quark jet * + + @xmath4high energy physics division , argonne national laboratory + argonne , il 60439 + @xmath5theoretical physics department , fermi national accelerator laboratory + p. o. box 500 , batavia , il
60510 + @xmath6centro studi e ricerche `` enrico fermi '' + via panisperna 89/a , 00184 rome , italy + @xmath7department of physics , university of illinois at urbana - champaign + 1110 west green street , urbana , il 61801 + @xmath8kavli institute for theoretical physics , university of california + santa barbara , ca 93106 + |
muon capture had been shown to be an excellent reaction for studying isovector magnetic transitions in light elements .
the weak interaction aspects of the reaction are well understood , at the level of a few percent ; this is far better than the nuclear structure uncertainties , which can be a factor of 2 , or even greater .
a recent survey was carried out by gorringe _
et al._@xcite for the targets of @xmath5na , @xmath6 mg , @xmath7si , @xmath8p , and @xmath9s .
they showed that the muon capture transition rates are similar to those observed in the ( n , p ) reaction , but can not be reproduced very well by calculations based on the oxbash code @xcite for nuclear structure in the 2s-1d shell .
brown and wildenthal @xcite obtained the universal sd interaction , used in this code , by fitting 440 energy levels in the 2s-1d shell . + for 1p shell nuclei the situation is better on the whole .
muon capture on @xmath10c has been reproduced by the recent calculations of hayes and towner @xcite and volpe _
et al._@xcite , and @xmath11o seems to be reasonably understood too , as indicated by the work of haxton and johnson@xcite and warburton _
. however these are recent calculations using quite complex descriptions of the a=12 and a=16 nuclei .
the case of @xmath1n remains one of the more puzzling examples amongst light nuclei .
the experiment of giffon _
et al._@xcite is the most recent experiment of muon capture on @xmath1n , and they observed only one @xmath0-ray , viz the one at 7017 kev , which is the ground state transition from the 7019 kev level ( the difference in energy is because the recoil energy is quite significant for such a high energy gamma and a light nucleus ) .
now giffon _
et al._@xcite obtained a capture rate to this level of 4640 @xmath12 700 s@xmath13 which is equivalent to a yield of ( 7.0 @xmath12 1.2 ) % per muon capture .
earlier measurements of this capture rate were slightly higher but of lower accuracy , and will be discussed later .
we take the total muon capture rate to be ( 66 @xmath12 5 ) @xmath14 10@xmath15 s@xmath13 , taking an average of three inconsistent measurements @xcite .
+ there are three calculations of the capture rate to the 7019 kev level ; mukhopadhyay @xcite obtained a value of 20,000 s@xmath13 ( i.e. a yield of 30 % ! ! ) .
he used the cohen - kurath model for 1 p shell nuclei , and harmonic oscillator radial wave functions .
mukhopadhyay realized that the calculated rate was unrealistic and so estimated that the ( 2s-1d)@xmath16 excitations in the 7019 kev level could reduce the rate to @xmath17 9700 s@xmath13 , which is better but still somewhat high .
a later calcuation by desgrolard _
_ @xcite confirmed that in the cohen - kurath model , the capture rate is @xmath17 23,000 s@xmath13 , and the addition of exchange currents reduces the rate by only 20 % which is far from sufficient @xcite .
thus it is clear that for @xmath1n , these calculations are in serious disagreement with experiment even though there is no major problem with other nuclei ( apart from @xmath18b which is also an ongoing problem @xcite ) .
these calculations were for the allowed transitions , i.e. the 1@xmath19 transitions which , from the 1@xmath19 ground state of @xmath1n , excite the @xmath1c ground state ( 0@xmath19 ) , the 6589 kev level ( 0@xmath19 ) and the 7019 kev level ( 2@xmath19 ) , under discussion .
two other unbound levels are also fed , a 2@xmath19 at 8.3 mev and a 1@xmath19 at 11.3 mev with a calculated rate of @xmath17 1.3 @xmath14 10@xmath15 ( yield of @xmath17 2 % ) .
+ many other reactions excite these same levels ; for example ( d,@xmath16he ) , ( @xmath20,@xmath0 ) , and ( n , p ) .
the best comparison is probably the ( d,@xmath16he ) reaction , but unfortunately no data exists . the ( @xmath20,@xmath0 ) reaction has been studied with an energy resolution of 720 kev @xcite and clear structure is observed at 7.0 , 8.3 , 10.7 , 15.4 , and 20 mev excitation energy .
only the first peak is bound and thus all the possible transitions to the 6 excited bound states are not resolved .
the ( @xmath20,@xmath0 ) reaction has been reviewed in general by gmitro _
et al._@xcite and the reaction @xmath1n(@xmath20,@xmath0 ) in particular has been analyzed by kissener _
now the ( @xmath20,@xmath0 ) reaction proceeds mainly from the 2p atomic state and only a small fraction from the 1s state , depending on the mass ; for @xmath1n r@xmath21:r@xmath22 = 5:1 . because the 2p state adds a 1@xmath23 to the initial state j@xmath24 conditions , 2@xmath19 and 3@xmath19 transitions are more prominent .
thus a direct comparison between the ( @xmath20,@xmath0 ) and ( @xmath3,@xmath4 ) reactions is hazardous , but if a calculation has addressed both reactions , one can have more confidence in that approach . similarly the ( n , p ) reaction can be used for another valuable comparison .
the reaction @xmath1n(n , p)@xmath1c has been studied at 59 mev by needham _
et al._,@xcite and at 280 mev by a group at triumf @xcite , both with an energy resolution of about 1 mev .
together these experiments give an excellent and consistent picture of the situation .
the triumf data at 0@xmath25 show levels at 7.0 mev ( 2@xmath26 , 46% ) ; 8.3 mev ( 2@xmath26,36% ) ; 10.4 mev ( 2@xmath26,8% ) ; and 11.3 mev ( 1@xmath26,10% ) .
the j@xmath24 of each level is given , as is the contribution to the 0@xmath25 strength ( roughly b(gt ) ) .
note that the group also has complementary and consistent data on the ( p , n ) and ( p , p@xmath27 ) reactions .
the needham _
_ results are compatible , with figures illustrating data at 16@xmath25 , 18@xmath25 , and 48@xmath25 , which show levels excited at 7.0 , 8.3 , 11.3 , 15.4 mev and a double peaked giant resonance at about 20.4 mev .
the angles of 16@xmath25 and 18@xmath25 correspond best to a momentum transfer of the ( @xmath3 , @xmath4 ) reaction , although 59 mev is a little low in bombarding energy .
the main conclusion from all these reactions is that the 1@xmath26 transition strength is much more distributed than that calculated in the cohen - kurath model .
+ we thus come to the final calculation of the @xmath1n(@xmath3,@xmath4)@xmath1c reaction , that by kissener _
et al._@xcite .
they were analyzing both the ( @xmath3,@xmath4 ) and the ( @xmath20,@xmath0 ) reactions , and hence were more aware of the pitfalls .
in particular they knew that ( e , e@xmath27 ) scattering had shown that the 1@xmath19 strength to 2@xmath19 levels was split about equally between the bound level at 7019 kev , and an unbound one at 8.3 mev . in the cohen - kurath model , the bound level takes almost all the strength .
however kissener _
_ split the strength , apparently phenomenologically , and not by introducing 2p-2h excitations in their calculation .
this calculation resulted in a yield of 11% to each of these levels , and explains at least one factor of 2 .
the other important feature of their calculation is that they address all the possible transitions , including high energy excitations in @xmath1c which decay via neutron emission to levels in @xmath2c and @xmath10c .
although the results of the calculations do not correspond in every detail to the observations in the present experiment , at least a good overall picture is given .
thus for example kissener _
show that the @xmath1c ground state is only weakly fed in direct transitions by the ( @xmath3,@xmath4 ) reaction , and this is related to the slow beta decay rate of @xmath1c .
_ also find that the sum of transitions to bound levels of @xmath1c is 15% ; for higher energy excitations in @xmath1c , they estimate that the yield is 42% to bound levels of @xmath2c , and 40% to bound levels of @xmath10c .
the rest is made up of charged particle emission ( @xmath17 3% ) , which is probably a low value ( @xmath17 15% is more common ) .
note that kissener _
et al . _ calculated the total muon capture rate to be 109 @xmath14 10@xmath15 s@xmath13 compared to the experimental value of ( 66 @xmath12 5 ) @xmath14 10@xmath15 s@xmath13 , but by presenting their results as yields per capture , they circumvent that problem .
+ for the reaction @xmath1n(@xmath3,@xmath4n)@xmath2c , one can envisage two reaction modes .
one is a direct knock - out term , similar to the pole - term seen in the ( @xmath20,@xmath0 ) reaction .
the levels observed in @xmath2c would thus correspond to levels excited in reactions such as ( d,@xmath15he),(p,2p ) or ( e , e@xmath27p ) .
the spectroscopic factors for these reactions are very similar .
an alternative description of muon capture would be in terms of an excitation of a spin - dipole giant resonance ( mainly 2@xmath23 and some 1@xmath23 ) , followed by a de - excitation similar to the ( @xmath0,p ) reaction .
of course the ( @xmath0,p ) reaction goes via the e1 giant dipole resonance , which is at a similar but not identical energy .
the comparison between the knock - out and excitation models has been discussed in some detail in a recent review of muon capture @xcite .
there is not a lot of difference between the observable effects , but as long ago as the muon capture work of miller _
@xcite in 1972 on @xmath7si , it was shown that the ( @xmath0,p ) reaction is a better analogue .
of course there will be some knock - out component , especially when a high energy neutron is produced in muon capture , but this represents only 10 % to 20 % of neutron production .
the data for the reaction @xmath1n(@xmath0,p ) are limited , but we shall make a fruitful comparison . +
one unresolved problem in muonic nitrogen is the hyperfine effect . for nuclei with spin , the 1s state in a muonic atom
is split into two hyperfine levels , which can be a few ev to @xmath17 1 kev apart in energy .
the m1 transition is too slow to be important , but telegdi and winston and co - workers at chicago @xcite showed that auger emission can speed up the transition rate to time scales of @xmath17 @xmath28s in light elements , and @xmath17 ns in heavy elements .
the best example is @xmath29f which was studied by the chicago group and has been confirmed several times since @xcite .
the case of @xmath1n is quite perplexing because a depolarization was observed in @xmath3sr by ishida _
et al._@xcite .
although some of the effect was attributed to external electromagnetic fields , the residual was ascribed to a hyperfine transition rate of ( 0.076 @xmath12 0.033 ) @xmath28s@xmath13 , i.e. @xmath30 @xmath17 13 @xmath28s .
now winston@xcite calculated the muonic hyperfine rate , via a simple atomic model of electron ejection , and his model fits every known example , especially @xmath29f . for @xmath1n
the prediction is that there should be no effect , because the separation of the hyperfine energy levels is 7.4 ev , whereas in the carbon atom , the least bound electron is bound by 11.3 ev , i.e. the ionization potential .
( the @xmath3 is well within the inner electrons , so the muonic @xmath1n system appears to be a @xmath1c nucleus to the atomic electrons ) .
if the pseudo @xmath1c atom has formed a `` c''n bond , the ionization potential is 14.3 ev , which makes auger emission even less likely .
thus it is possible that ishida _
et al._@xcite observed other depolarization mechanisms .
we note that wiaux @xcite , using a muon capture technique for @xmath18b , observed a slower hyperfine transition rate than that observed in a @xmath3sr experiment .
we have searched for hyperfine effects in @xmath1n , but because the supposed time constant is long , it is difficult to detect , and our observations are inconclusive
. thus we shall assume that our capture rates are for a statistical mixture of the two hyperfine states .
however even if there were a slow hyperfine transition , the effect on our results would be smaller than the errors .
+ thus muon capture in @xmath1n exhibits many interesting effects .
our experiment was proposed in order to obtain some more relevant observations .
one great advantage of @xmath1n is that there are only a few bound levels in @xmath1c , @xmath2c , and @xmath10c .
thus we are able to make a complete analysis of all possible transitions for these nuclei .
the experimental technique is well established , and has been described in several previous publications @xcite .
details can be found in the thesis of stocki@xcite .
we shall limit ourselves to the more important features .
+ the experiment was performed on the m9b channel at the triumf cyclotron .
the beam line includes a 6 m , 1.2 t superconducting solenoid in which 90 mev / c pions can decay .
the resulting backward muons are then selected by a bending magnet and pass through a collimator into the experimental area .
the collimator was made out of lead bricks , but faced with 13 mm of polyethylene to minimize background neutrons and @xmath0-rays .
the beam rate was about 2 @xmath14 10@xmath31 s@xmath13 , with a pion contamination of @xmath32 0.2 % and an electron contamination of @xmath17 20 % .
three plastic scintillators defined a muon stop , the defining counter being 51 mm in diameter .
+ the target was liquid nitrogen contained in a styrofoam box , internal dimensions 215 mm ( beam direction ) @xmath14 195 mm @xmath14 550 mm .
the length in the beam direction was more than sufficient to stop the muons , so the veto counter behind the target was effectively redundant .
a mu - metal shield was immersed in the liquid nitrogen to reduce the ambient magnetic field from 1.5 gauss to 0.1 gauss .
this minimized spin precession effects .
+ to detect the @xmath0-rays from muon capture , two high purity germanium detectors ( hpge ) were placed opposite each other at 90@xmath25 to the beam direction .
both detectors were n - type detectors to minimize radiation damage .
the largest was ge1 , a 44% detector , with an in - beam energy resolution of 2.53 kev at 1.3 mev and a timing resolution of 6 ns .
the other , ge2 , was a 21% detector , with an in - beam energy resolution of 2.19 kev at 1.3 mev , and a timing resolution of 7 ns .
their front windows were 27.5 cm and 33.1 cm away from the centre of the target , respectively .
in front of each was a plastic scintillator to tag electron events .
each hpge was surrounded by a compton suppressor , composed of a segmented nai(tl ) annulus . in this experiment
the suppressors were implemented in software mode , so single and double escape peaks were in fact observed , so we were able to take advantage of that feature . + the electronics consisted of spectroscopic amplifiers and timing filter amplifiers , followed by constant fraction discriminators ( cfd ) .
an event was defined by a pulse in a germanium detector , and then tags from the electron counters , and pulse height information from the compton suppressors were also recorded .
events from a delayed muon stop were then sent to a router for 10.8 @xmath28s .
this device can accept up to 4 pulses and route them to four distinct tdcs .
thus one can identify events with only one candidate muon , from those with two , three , or even four muon stops in the previous 10.8 @xmath28s .
typically 15% of events have two muon stops , which made selection of events problematic , for determination of lifetimes .
two methods were used ; one selecting events with only one muon stop , the other accepting all events . accepting the first muon , whether or not it was followed by another muon , created a clearly distorted time spectrum .
+ the energy calibration of the detectors was carried out by a variety of techniques .
offline @xmath33eu and rdth sources were useful .
the @xmath0-ray energies were known to a better than 10 ev which is far more accurate than is needed .
of course the beam on conditions are somewhat different , and calibrations at a much higher energy were required .
the spectra were divided into 3 regions ( low , medium , and high energy ) and it was found that , in each region , linear fits for the energy calibration were quite adequate .
the beam - on calibrations that were used are listed in table [ tab1 ] .
note that the ( n,@xmath0 ) lines are room background and not due to the liquid nitrogen target , but they are narrow lines and therefore are convenient for calibration purposes , especially as the energies are known to 0.1 kev in a region with few calibration lines . also noticeable is the @xmath0-ray from @xmath11n @xmath34-decay , coming from the irradiated cooling water in the nearby quadrupoles .
it has an energy of 6,129,140(30 ) ev@xcite whereas several compilations seem to use the erroneous excitation energy of e@xmath35 = 6,129,893(40 ) ev and a @xmath0-ray energy of 6,128,630(40 ) ev @xcite .
we note that we have used the single and double escape peaks with energy differences of 511.00 and 1022.00 kev respectively .
this is unreliable and the actual difference can be 0.1 to 0.3 kev less , depending on the detector @xcite . because our lines are doppler broadened , such small calibration errors were not studied .
higher precision work would need to take this effect into account .
+ the efficiency of the germanium detectors was determined offline with calibrated sources , and online using muonic x - rays .
an excellent target for this purpose is gold , and the emission probabilities are given in table [ tab2 ] .
the 2p-1s intensities were taken from hartmann _
et al._@xcite and the others from a muon cascade program @xcite .
the efficiency curve for ge2 is presented in figure [ fig1 ] .
the individual values were fitted to a curve of the form : @xmath36 where e is the photon energy and a@xmath37 are parameters for the fit .
this gives an adequate fit over the whole energy region .
because there are a limited number of @xmath0-rays which can be produced in muon capture in @xmath1n , it was not too difficult to identify the various lines , using the known transitions .
however , many other transitions were also observed including : muonic x - rays in c , n , fe , and ni ; ( n , n@xmath27 ) lines from n , al , fe , ge , and pb ; @xmath34-decay lines from @xmath11n , @xmath38na , @xmath39ar , @xmath40co , and as mentioned before ( n,@xmath0 ) capture @xmath0-rays from h , n , al , cl , fe , ge , and in . over 150 lines were identified , most with confidence .
this was done to ensure that we had not missed a candidate line from muon capture .
the reason that so many background lines occur is twofold .
first in @xmath1n only 12.7(8)% of muons capture , the rest decay .
secondly , because of the longer muon lifetime , the @xmath0-ray gate has to be kept open for @xmath17 10 @xmath28s , whereas for an element like ca a gate of only 1.7 @xmath28s is needed .
this makes the @xmath1n spectra about 50 times more vulnerable to background lines than those for ca . +
another difficulty in @xmath1n is that the muon capture lines are doppler broadened as they are often emitted while the @xmath1c recoils after the neutrino emission .
this broadening is box shaped and amounts to 14 kev per mev @xmath0-ray energy , i.e. about 100 kev for a 7 mev @xmath0-ray .
a few energy levels in @xmath1c and @xmath2c are long lived with respect to the slowing down time ( @xmath17 1 ps ) , and then the @xmath0-rays are narrow peaks . in @xmath2c
the 3684 kev line is doppler broadened by the neutrino emission and then again by neutrons of various energies .
empirically we found a triangular shape fitted the observed line . +
the box shaped spectra were fitted to the formula : @xmath41 + ax + b \label{e2}\ ] ] where e is the centroid of the peak , @xmath34 is the recoil velocity divided by the speed of light ( @xmath17 0.0075 ) , s is the parameter accounting for the hpge energy resolution , n is the overall amplitude , and a and b are background terms . + a typical fit to the full energy peak of the 7017 kev peak is shown in figure [ fig2 ] .
this line is fortunately free from background @xmath0-rays and the yield is easy to determine ( note that there should be a very weak line from cl(n,@xmath0 ) as predicted by the cl(n,@xmath0 ) branching ratios , the effect of this line is subtracted out of the 7017 kev @xmath0-ray yield ) .
the 6092 kev peak is much more complicated and two fits to this peak are illustrated in figure [ fig3 ] .
it is immediately clear that a spectrum with good statistics is required to fit such a complex structure .
we observe two background lines superimposed on the @xmath1c 6092 kev @xmath0 -ray , the @xmath42cl(n,@xmath0 ) at 6110.88 kev and the @xmath11n @xmath34 decay line at 6129.14 kev , both fortunately are narrow peaks . at the centre of the structure
is an interesting peak , due to @xmath0-rays from the same @xmath1c level at 6092 kev , but narrow because they come from cascades from higher , longer lived levels .
the single escape peak is illustrated in figure [ fig4 ] , and another background line is apparent , the 5592.2 kev line from @xmath43i(n , @xmath0 ) as well as the single escape peaks from all the other effects of figure 3 .
+ in figure [ fig5 ] is illustrated a fit to the @xmath2c peak at 3685 kev .
it is composed of a doppler broadened component , and a central peak , again coming from a cascade from the long lived level at 3854 kev .
the separation of these two components is more uncertain , as the shape of the doppler broadened component is purely a phenomenological assessment .
from the various fits to the three @xmath1c @xmath0-rays ( not all illustrated ) we obtain the energies observed in this experiment , listed in table [ tab3 ] .
the @xmath0-ray energies are corrected for the recoil correction ( 1.89 , 1.74 , 1.42 respectively ) to obtain the energy of the excited state .
this is compared with the compilation of ajzenberg - selove @xcite , and a study of the @xmath2c(d , p)@xmath1c reaction by piskor and schferlingova @xcite .
although this study was noted by ajzenberg - selove under reaction 16 , it was not integrated into her table of level energies which is the same as the 1986 compilation .
our results for the 6092 and 6730 kev levels agree , but the 7019 kev is inconsistent .
we note again that no correction has been made for the escape peaks being slightly less than the 511 kev difference ; our other errors are much larger . +
as there is concern about the hyperfine transition , we fitted the time dependence of several @xmath0-rays .
the 3684 kev line in @xmath2c has a high yield , but a complicated shape because of the cascading , see figure 5 .
however hyperfine asymmetry effects have been seen in such cases , so a study of the time dependence was attempted .
the yields of this @xmath0-ray and of electron events were selected . each time spectrum
was then fitted to obtain a flat background , which was then subtracted off .
finally the ratio of these two subtracted spectra is illustrated in figure [ fig6 ] . even though about 7000 @xmath0-ray events were obtained , it is clear that the study is inconclusive .
the results are compatible with no hyperfine transition with r(@xmath0/e ) = 4 @xmath14 10@xmath44 , but could also be fit with a significant asymmetry .
of course there is no guarantee that a particular transition , especially one from a ( @xmath3,@xmath4n ) reaction , will have a large asymmetry .
it is clear that better statistics are needed , and also results from more lines . for the rest of the analysis we shall assume no hyperfine transition , and a statistical population of the hyperfine levels .
+ the yields of many lines have been studied by using the fitting techniques described above , using gaussian , box , or triangular fitting functions , where appropriate . in the cases where a limit was set the possible doppler broadening of lines
was taken into account , where appropiate .
windows were set on the expected position , and the nearby background was subtracted .
no attempt was made to fit possible shapes as these are variable and uncertain .
however that technique has the potential of setting slightly lower limits . in table
[ tab4 ] we list the limits on various lines obtained using ge1 ; also included are the lifetimes of the original levels @xcite .
only 3 of these lines have positive identification .
the 4438 kev line in @xmath10c can be caused by background from target induced neutrons in the plastic scintillators .
we made a time study to eliminate background neutrons , but did not have enough information to estimate the contribution from target produced neutrons .
similarly @xmath45b could have come from capture on beam scintillators .
note that we have no positive identification of ( @xmath3,@xmath4p ) nor ( @xmath3,@xmath46 ) reactions ( i.e. @xmath2b or @xmath45be ) with quite useful limits .
+ in table [ tab5 ] we present the results for the @xmath2c and @xmath1c @xmath0-rays .
some of the @xmath2c results are an average of ge1 and ge2 , but those for @xmath1c are for ge1 only as the acceptance for ge2 was too low to be useful .
the fitting for the 3685 kev @xmath0-ray finds that ( 15.9 @xmath12 1.0)% of the @xmath0-ray yield is fed from the 3854 kev level via a 169 kev cascade ( b.r .
= 36.3% ) .
this @xmath0-ray was below threshold , but from the yield of the 3854 kev ground state transition , using the known branching ratio ( 62.5% ) , one can deduce that ( 19 @xmath12 8 % ) of the 3685 kev yield is from the cascade , in agreement with the fitting procedure .
+ equally well the @xmath1c 6092 kev @xmath0-ray has a complex structure . from the full energy peak we find that ( 8.6 @xmath12 5.3 ) % of the yield is from cascade feeding ; from the single escape peak
the value is ( 15 @xmath12 12 % ) .
furthermore there is no evidence for a second doppler broadened peak on the principal peak , which we use as evidence against feeding from the 7341 kev level .
equally well , the ground state transition for this 7341 kev level ( b.r .
= 16.7 % ) is not observed .
thus , for the 6092 kev @xmath0-ray we assume any contribution from the 7341 level to be zero , and average the two central gaussian yields to obtain that a weighted average of ( 10 @xmath12 6)% of the 6092 kev yield is cascading .
thus the direct feeding of the 6094 kev level is ( 1.2 @xmath12 0.6)% and the cascading ( 0.12 @xmath12 0.08)% per capture .
now we know the 6732 kev level is populated ( @xmath30 = 66 ps ) and would contribute ( 0.05 @xmath12 0.02)% per muon capture to the 6092 kev @xmath0-ray .
this leaves ( 0.07 @xmath12 0.08)% which could be feeding from the 6589 kev level which has a branching ratio of 98.9% for a cascade via 495.35 kev @xmath0-ray .
this @xmath0-ray is not observed , but the limit is only @xmath32 0.049% per muon capture .
thus our results are compatible with a feeding of @xmath17 0.04% which is expected in several calculations . + from these fitting procedures , and using other known branching ratios we obtain the direct feeding of all the levels .
table [ tab6 ] presents the limits for complex reactions ; table [ tab7 ] presents the results for @xmath2c and @xmath1c .
apart from the 7019 kev level in @xmath1c , these are all new results .
the sum of the yields to bound @xmath1c levels is ( 6.0 @xmath12 1.6)% which agrees well with the ( @xmath20,@xmath0 ) result of ( 6.22 @xmath12 0.40)% @xcite , + our absolute yield for the 7019 kev level is lower than all previous results .
most have not been published and we have to rely on conference proceedings , reports , and the review of mukhopadhyay @xcite . these results , our own , and a recommended average are presented in table [ tab8 ] . note that we have used the rates from other experiments to calculate the yields ; in an experiment it is done the other way but we do not have the original yields . as we had significant difficulty with our absolute normalization , we take the agreement to be satisfactory .
earlier unpublished results could have significant contributions from background lines and we know that bellotti _
_ had a serious contamination of unknown origin , see page 131 of mukhopadhyay@xcite ( we do not observe any line at 6315 kev ) .
none of these results had the statistical accuracy , nor the energy resolution of our own data . thus we combine our results with only the saclay result to obtain a recommended yield of ( 6.6 @xmath12 0.9)% , i.e. a rate of ( 4390 @xmath12 580 ) s@xmath13 . +
all the results for @xmath2c and @xmath1c are combined in table [ tab9 ] , and compared with a variety of reactions , and theoretical calculations .
the @xmath2c level feeding is compared with the results for the reaction @xmath1n(@xmath0,p ) @xcite .
we take the integrated cross section up to 29 mev ; the ground state and 7550 kev transitions are given by gellie et al .
@xcite , and the 3089 , 3685 , and 3854 kev levels are taken from thompson _
et al._@xcite who studied the @xmath0-rays emitted during irradiation by 29 mev bremsstrahlung .
the ( d,@xmath15he ) reaction has been studied by hinterberger _
et al._@xcite and the spectroscopic factor c@xmath16s was found to be 0.8 , 0.3 , and 1.4 for the ground state , 3685 , and 7550 kev states respectively ; the feeding of the 3089 kev level is small ( c@xmath16s @xmath17 0.05 ) and the feeding of the 3854 kev level is not even detected . thus the ( @xmath0,p ) reaction is a better analogue because the 3089 and 3854 kev levels are clearly detected in muon capture .
et al._@xcite predicted that the pattern of capture be close to the spectroscopic factors , with none feeding the 3089 and 3854 kev levels , in contradiction to experiment . + for @xmath1c , we compare with the calculation of kissener _
et al._@xcite for the ( @xmath28,@xmath4 ) reaction .
remember that the split between the 7019 kev level and the 8318 kev level is phenomenological .
the calculation of mukhopadhyay @xcite was for only allowed transitions , i.e. 1@xmath19 transitions to 0@xmath19 , 1@xmath19 , 2@xmath19 states .
the 28% feeding of the 7019 kev was confirmed by desgrolard _
et al._@xcite for the cohen - kurath model .
mukhopadhyay estimated that the ( 2p-2h ) contributions would roughly halve the transition rate ; the other half would go to another 2@xmath19 state and we have used this estimate in table 9 .
note that mukhopadhyay gives transition rates , so we converted to yield using a total capture rate of ( 66 @xmath12 5 ) @xmath14 10@xmath15 s@xmath13 .
however kissener _
et al . _ found a total capture rate of 109 @xmath14 10@xmath15 s@xmath13 and converted to yield themselves , so this difference probably explains why the sums for the yields of mukhopadhyay are higher than those of kissener _
_ it is clear , however , that even using the saclay yield of ( 7.0 @xmath12 1.0)% , the kissener calculation is still overestimating the yield for the 7012 kev level .
a major problem thus remains .
+ we also list in table 9 the experimental results of perroud _
et al._@xcite for the ( @xmath20,@xmath0 ) reaction at rest . as their resolution was only 720 kev all the bound levels
are in one unresolved peak with a yield of 6.2(4)% of all radiative capture .
we also give the calculated yields of kissener _
et al._@xcite for the ( @xmath20,@xmath0 ) reaction .
these values have been estimated from the height of the bars in a figure , so are approximate .
we can see that the calculation fits the experimental yields quite well , and both compare quite favourably with the ( @xmath3,@xmath4 ) reaction , even though only 17% of the ( @xmath20,@xmath0 ) captures are from the 1s state .
+ we can now estimate the overall situation for muon capture in @xmath1n , using empirical information .
if we renormalize our results up by 50% to approach the saclay yields , we suggest about 9(2)% of muon captures give bound states in @xmath1c and 14(3)% produce bound excited states in @xmath2c .
the ( @xmath0,p ) reaction then implies about 32(8)% direct feeding of the @xmath2c ground state and 27(8)% direct feeding of the @xmath10c ground state via the 7.55 level in @xmath2c .
if we add 4% feeding of the first excited state in @xmath10c and 13% for charged particles , we obtain a sum of 99(10)% , which is reasonable . note
that this empirical approach is fairly similar to the theoretical estimate of kissener _
15% to bound levels of @xmath10c , 42% to bound levels of @xmath2c and 40% to bound levels of @xmath10c with only 3% charged particle emission .
we have provided a lot of new information about muon capture in @xmath1n .
the direct transitions to @xmath1c are clearly detected with major feeding going to only three of the 7 bound levels .
these three levels , at 6094 , 6732 , and 7019 kev have similar yields , yet the calculation of kissener _
predicts that the 7019 kev should dominate .
( note that our relative yields are more dependable than the absolute values ) .
it thus seems that the nuclear structure of the a = 14 system is quite complex and 2p-2h excitations in the wave function make radical changes to the yields .
calculations along the lines of work in @xmath10c @xcite and @xmath11o @xcite would be really helpful .
+ three @xmath0-rays from @xmath2c are observed and their pattern follows the ( @xmath0,p ) reaction yields better than the knock out reactions and their spectroscopic factors .
even the ( @xmath0,p ) reaction is not a perfect analogue , and it would be interesting to have data on the integrated yield for lower energy bremsstrahlung ( @xmath17 24 mev or so ) .
+ other @xmath0-rays have been searched for but no convincing evidence found .
the only line clearly detected was from the 4439 kev state in @xmath10c , and it is certainly excited in muon capture , but it is also strongly excited by background mechanisms , so no useful estimate could be made ( one would need extensive runs with carbon and boron targets to separate the various effects ) . +
no @xmath0-rays were observed from the reaction @xmath1n(@xmath3,@xmath4p)@xmath2b which is surprising .
we obtained a limit of @xmath32 0.17% for the first excited state at 3483 kev .
in other light nuclei this reaction has been clearly detected ; for example miller _
et al._@xcite found that the reaction @xmath6mg(@xmath3,@xmath4p)@xmath5ne gives a yield of 0.20(5)% for the 980 kev @xmath0-ray and 0.5(1)% for the 1770 kev line ; similarly they observed for @xmath7si(@xmath3,@xmath4p)@xmath47 mg a yield of 1.9(2)% for the 984 kev line .
our only comment is that the high energy of the first excited state in @xmath2b may affect the yield adversely .
equally well kissener _
_ estimate a low yield for this reaction , but there is obviously some uncertainty in their estimate .
+ in heavy nuclei the reaction ( @xmath3,@xmath4pn ) is stronger than the ( @xmath3,@xmath4p ) reaction @xcite . in @xmath48ca
these reactions have equal @xmath0-ray yields of about 6% each @xcite .
miller observed a yield of 4.4(6)% in @xmath6 mg and 10(1)% in @xmath7si .
it is thus surprising that our limit for the first excited state of @xmath10b at 953 kev is @xmath32 0.27% + the mass a=14 system has retained its enigmatic status .
in particular there is more than sufficient evidence that the cohen - kurath model is inadequate and that the nuclear structure is quite complex .
a comprehensive theoretical study using modern computational technology would be most welcome .
we wish to thank ulrich giesen for his help with the ge detector acceptances .
we would also like to thank andrew macfarlane for his help with the design of the liquid nitrogen target vessel .
we wish to thank the natural sciences and engineering research council of canada and similarly the national science foundation in the u.s.a . for support and equipment ; we also thank the staff at triumf and the national research council of canada for providing the muon beam facilities .
et al . _ ,
@xmath49 ( 1999 ) 055501 .
b. johnson _
et al . _ ,
rev . @xmath50 ( 1996 ) 2714 .
brown , a. etchegoyen , w.d.m .
rae , and n.s .
the oxford - buenos - aires - msu shell - model code ( oxbash ) .
msucl report n@xmath51 524 , ( 1986 ) .
b.a . brown and b.h .
wildenthal , ann .
nucl . part .
sci . @xmath52 ( 1998 ) 29 .
hayes and i.s .
towner , phys .
@xmath53 ( 2000 ) 044603 . c. volpe _ et al .
@xmath54 ( 2000 ) 015501 .
haxton and c. johnson , phys .
@xmath55 ( 1990 ) 1325 .
warburton , i.s .
towner , and b.a .
brown , phys .
@xmath56 ( 1994 ) 824 .
m. giffon _
et al . _ ,
@xmath57 ( 1981 ) 241 .
t. suzuki _
et al . _ ,
@xmath58 ( 1987 ) 2212 .
mukhopadhyay , phys .
@xmath59 ( 1973 ) 33 .
p. desgrolard _
et al . _ , nuovo cim .
@xmath60 ( 1979 ) 120 .
p. desgrolard and p.a.n .
guichon , phys .
@xmath61 ( 1978 ) 475 .
v. wiaux , ph.d .
thesis , universit catholique de louvain , 1999 .
et al . _ ,
@xmath62 ( 1986 ) 542 .
m. gmitro _
j. particles nuclei @xmath63 ( 1982 ) 513 ; ibid @xmath64 ( 1983 ) 323 .
@xmath65 ( 1978 ) 523 .
et al . _ , nucl .
@xmath66 ( 1978 ) 394 .
et al . _ ,
@xmath67 ( 1982 ) 349 .
jackson , private communication .
et al . _ ,
phys . @xmath68 ( 1973 ) 424 .
measday , phys .
reports , to be published .
g.h . miller _
@xmath69 ( 1972 ) 487 .
g. culligan , j.f .
lathrop , v.l .
telegdi , r. winston and r.a .
lundy , phys .
@xmath70 ( 1961 ) 458 .
r. winston , phys .
@xmath71 ( 1963 ) 2766 .
stocki , ph.d .
thesis , university of british columbia 1998 ; to be published .
k. ishida _
et al . _ ,
@xmath72 ( 1986 ) 31 .
et al . _ ,
table of isotopes , 8@xmath73 edition , published by j. wiley , n.y . ,
march 1996 .
f. ajzenberg - selove , nucl .
@xmath74 ( 1991 ) 1 .
r. engfer _ et al .
_ , atomic data and nuclear data tables @xmath64 ( 1974 ) 509 .
et al . _ ,
@xmath75 ( 1974 ) 413 . g. fricke _
et al . _ , atomic data and nuclear data tables @xmath76 ( 1995 ) 177 .
world wide web site : thermal neutron capture gammas by energy : http://www.nndc.bnl.gov/wallet/tncngtblcontentbye.shtml m.r .
bhat , nuclear data sheets @xmath77 ( 1992 ) 195 ( see p230 ) .
r. vennink _ et al .
phys . @xmath78 ( 1980 ) 421 .
a. h. wapstra , nucl .
@xmath79 ( 1990 ) 671 .
world wide web site : http://www.nndc.bnl.gov d.r .
et al . _ ,
@xmath80 ( 1993 ) 1 .
e.k . warburton and d.e .
alburger , nucl .
@xmath67 ( 1982 ) 189 .
et al . _ , z. phys .
@xmath81 ( 1982 ) 189 .
akylas and p. vogel , computer physics communications @xmath82 ( 1978 ) 291 .
s. piskor and w. schferlingova , nucl .
@xmath83 ( 1990 ) 301 .
f. ajzenberg - selove , nucl .
@xmath84 ( 1988 ) 1 .
f. ajzenberg - selove , nucl .
@xmath85 ( 1990 ) 1 .
mukhopadhyay , physics reports @xmath86 ( 1977 ) 1 .
et al . _ ,
jinr report r-14 - 42 - 41 ( 1968 ) quoted by v.a .
et al . _ ,
@xmath87 ( 1970 ) 295 .
l. grenacs , quoted on page 57 of mukhopadhyay @xcite .
e. bellotti _
et al . _ , sin newsletter @xmath88 ( 1976 ) 41 and quoted on pages 58 and 131 of mukhopadhyay @xcite .
r.w . gellie _
@xmath89 ( 1972 ) 1689 .
@xmath90 ( 1970 ) 211 . f. hinterberger _
et al . _ ,
@xmath91 ( 1968 ) 161 .
a. wyttenbach _
et al . _ ,
@xmath92 ( 1978 ) 278 .
igo - kemenes _
et al . _ ,
@xmath93 ( 1971 ) 286 . | many new @xmath0-rays have been observed , following muon capture on @xmath1n .
one had been reported before , and the low yield is confirmed , indicating that the nuclear structure of @xmath1n is still not understood .
gamma - rays from @xmath2c resulting from the reaction @xmath1n(@xmath3,@xmath4n)@xmath2c compare favourably with states observed in the reaction @xmath1n(@xmath0,p)@xmath2c .
more precise energies are also given for the 7017 and 6730 kev @xmath0-rays in @xmath1c .
, , , , , muon capture , @xmath1n , gamma - rays 25.30.-c , 23.40.hc , 27.20.+n |
most spins in a magnetic nanoparticle are at the particle surface where the reduced coordination and the broken translational symmetry of the lattice are responsible for changes in some local physical properties of the material . among them
, there is the reduction in exchange interaction constants and the increased local anisotropy at the particle surface with respect to bulk values @xcite phenomenological model that described this last phenomenon was first introduced by nel in @xcite , but experimental determination of the microscopic structure and properties of the surface of nanoparticles is still a challenge to be solved .
for example , an estimate of the surface anisotropy constant can only extracted indirectly from magnetic measurements by fitting to an effective anisotropy constant @xcite .
however , there are increasing evidences from experiments that surface effects play a dominant role in the anomalous low temperature magnetic properties of nanoparticle systems @xcite . therefore , simulations of magnetic properties in which microscopic parameters extracted form experiments are used @xcite may serve as a benchmark to test the phenomenological models at hand . in this article , we use the monte carlo method to simulate the magnetization processes of individual nanoparticles modeled by classical heisenberg spins on the maghemite lattice with radial surface anisotropy @xmath0 different from that on the core .
the model considered is an extension to heisenberg spins of our previous mc simulation for an ising spin model of a @xmath1-fe@xmath2o@xmath3 ( maghemite ) particle @xcite .
the particles considered have spherical or ellipsoidal shape with diameter @xmath4 and elongations @xmath5 ( in units of the cell size @xmath6 ) , containing from @xmath7 to @xmath8 spins .
the surface spins of the particle are considered to be those on the outer unit cell and have anisotropy constant @xmath0 and radial local anisotropy axes along the directions @xmath9 , whereas core spins have uniaxial anisotropy @xmath10 pointing along the z - axis .
indirect estimations of the anisotropy constants yield values in the range @xmath11 k , @xmath12 k , so we will fix @xmath13 k and vary only @xmath0 .
the interaction hamiltonian in temperature units can be written as @xmath14+\sum_{i= 1}^{n_{surf}}\left[k_s(\vec{s}_i \cdot \hat n_i)^2 \right]\ .
\end{aligned}\ ] ] in the first term , @xmath15 are the nearest - neighbours exchange constants among the spins in the two sublattices with different coordination of maghemite , all of them are antiferromagnetic , being @xmath16 the strongest in magnitude and making maghemite a ferrimagnetic compound @xcite .
we have computed equilibrium configurations by starting from a disordered state at a high temperature and decreasing the temperature in steps @xmath17 .
hysteresis loops are started from a saturating field along the core easy - axis that is decreased in steps @xmath18 , during which the magnetization is averaged over @xmath19 mc steps . in both cases ,
the standard mc method for continuous spins has been used with a trial jump consisting in a combination of uniform and small angle around the current spin direction that allows to obtain reliable simulations for treat the weak and strong anisotropy cases with the same algorithm , see @xcite for the details .
in fig . [ fig_1 ] , we show snapshots of the spin configurations attained after the cooling process previously described , for several values of the surface anisotropy constant @xmath0 .
the pictures present a cut of one unit cell width along the equatorial plane and along the major axis of the particle . for values of @xmath0
smaller than a certain critical value @xmath20 ( which depends on the particle size and on the shape ) , the radial anisotropy at the surface is not strong enough to ca nt the core spins towards the radial direction .
therefore , the core spins are antiferromagnetically aligned with a small net magnetization component along the @xmath21 axis due a small number of non - compensated spins ( see fig .
[ fig_1 ] for @xmath22 ) .
however , at the particle surface , the spins present already a high degree of disorder due to the competition between the reduced local exchange field ( locally reduced with respect to that on the core spins ) and the anisotropy , inducing some degree of frustration at the surface . as @xmath0
is further increased ( see fig .
[ fig_1 ] for @xmath23 ) , the disorder of the surface spins is transmitted to the particle core and results in the departure of core spins from the local anisotropy axis towards the radial direction .
finally , when anisotropy dominates over exchange energy ( @xmath24 , see fig .
[ fig_1 ] for @xmath25 ) , the hedgehog - like structures are formed , with the spins pointing along the local radial direction and having antiparallel orientations with the nearest neighbors on the other sublattice .
we have also studied the influence of surface anisotropy and particle shape on the magnetization reversal processes . for this purpose ,
zero temperature hysteresis loops have been simulated using the procedure described previously for ellipsoidal particles with elongations @xmath5 and diameter @xmath4 .
some selected results are presented in fig .
[ fig_2 ] , where the total magnetization component along the field direction ( left panels ) is shown together with the contribution of the surface and core spins to @xmath26 , the sum of the projections of spins onto the local anisotropy axis ( central and right panels ) . by comparing the loops for @xmath0= 10 to those for @xmath0= 100 , we see that the coercive field increases and the remanence decreases with increasing surface anisotropy , independently of the elongation l of the particle .
moreover , the presence of disordered groups of spins at the surface induced by surface anisotropy makes the loops more elongated and increases the closure fields of the loops as found also in experiments on ferrimagnetic oxides @xcite .
the rounding of the loops near the coercive field clearly indicates a progressive departure from a uniform reversal mechanism with increasing @xmath0 . when looking only at the @xmath27 component , not much difference is appreciated between the loops for particles with different l because of the compensation of the spin components transverse to the field direction due to the symmetry of revolution of the particles around the @xmath21 axis .
however , upon further inspection of the @xmath28 and @xmath29 contributions and animated snapshots taken along the loops @xcite , the details of the reversal process can be better under - stood .
let us notice first that , when @xmath0 is increased from @xmath20 , the reversal mechanism changes from quasi - uniform ( induced by the core ) rotation at low @xmath0 values , to a process in which the formation of surface hedgehog - like structures induce the non - uniform switching of the whole particle . in the first regime ( @xmath22 case in fig .
[ fig_2 ] ) , the core and surface spins point mostly along the z axis ( @xmath30 , @xmath31 ) except near the coercive field , where they make short excursions towards the radial direction driven by the surface anisotropy ( see the dips in @xmath29 and the cusps in @xmath28 ) . however , for @xmath0 > @xmath20 ( @xmath25 case in fig .
[ fig_2 ] ) , surface spins remain near the local radial easy - directions ( @xmath32 ) during the reversal , while the core spins are dragged out the z local easy - axis by the surface spins during the reversal , except for values of h near the closure field .
this is indicated by the widening of the dips in @xmath29 and the global decrease of @xmath29 values as @xmath0 increases .
finally , let us remark also that , for all the @xmath0 considered , the @xmath28 values along the whole hysteresis loops increase with increasing @xmath33 , which indicates that the surface spins remain closer to the local radial direction during the reversal as particles become more elongated .
when increasing @xmath33 , the dips in @xmath29 become less profound for @xmath24 , and indication that reversal of core spins along the radial direction is suppressed by the elongation . however , for weak anisotropy ( @xmath34 ) , the more elongated the particles are , the greater the deviation of surface spins towards the radial direction during the reversal .
we acknowledge cesca and cepba under coordination of c@xmath35 for computer facilities . this work has been supported by the spanish seeuid through the mat2003 - 01124 project and the generalitat de catalunya through the 2001sgr00066 cirit project . | the influence of surface anisotropy on the magnetization processes of maghemite nanoparticles with ellipsoidal shape is studied by means of monte carlo simulations .
radial surface anisotropy is found to favor the formation of hedgehog - like spin structures that become more stable as the surface anisotropy constant at the surface @xmath0 is increased form the value at the core .
we have studied the change in the low temperature hysteresis loops with the particle aspect ratio and with @xmath0 , finding a change in the magnetization reversal mode as @xmath0 or the particle elongation is increased . |
it is known from the work of allouche , btrma , and shallit ( see @xcite ) that a square free sequence on 6 letters can be obtained by encoding the optimal solution to the standard hanoi towers problem on 3 pegs by an automaton .
roughly speaking , given an input word which is the binary representation if the number @xmath0 , the automaton ends in one of the 6 states .
these states represent the six possible moves between the three pegs and if the automaton ends in state @xmath1 , this means that the one needs to move the top disk from peg @xmath2 to peg @xmath3 in step @xmath0 of the optimal solution .
the obtained sequence over the 6-letter alphabet @xmath4 is an example of an automatic sequence .
we choose to work with a slightly different type of automata , which under a suitable interpretation , produce integer sequences in the output .
the difference with the above model , again roughly speaking , is that not only the final state matters , but the output depends on every transition step taken during the computation and both the input and the output words are interpreted as encodings of integers .
the integer sequences that can be obtained this way are called transducer integer sequences .
we provide some examples that illustrate the notion of a transducer integer sequence .
all provided examples are related to the hanoi towers problem on 3 pegs . in recent years
, a very fruitful line of research in group theory has led to the notion of a self - similar group @xcite ( also known as automata groups @xcite or state closed groups @xcite ) .
many challenging problems have been solved by using finite automata to encode groups of tree automorphisms with interesting properties , leading to solutions to outstanding problems . to name just a few ,
such examples are the first grigorchuk group @xcite , solving the problem of milnor on existence of groups of intermediate growth and the day - von neumann problem on existence of amenable but not elementary amenable groups , basilica group @xcite , providing an example of amenable but not subexponentially amenable group , wilson groups @xcite , solving the problem of gromov on existence of groups of non - uniform exponential growth , the realization of the lamplighter group @xmath5 by an automaton @xcite , leading to the solution of the strong atuyah conjecture on @xmath6-betti numbers @xcite , and the recent solution to hubbard s twisted rabbit problem in holomorphic dynamics @xcite .
the geometric language and insight coming from the interpretation of the action of the automata as tree automorphisms greatly simplifies the presentation and helps in the understanding of the underlying phenomena , such as self - similarity , contraction , branching , etc ( see @xcite for definitions , examples , and details ) .
the language of finite automata has been proved extremely suitable in working with self - similarity phenomena .
indeed , in the case of automatic sequences , it is known from cobham theorem @xcite that such sequences are precisely those that are obtained as images under codings of fixed points of uniform endomorphisms ( limits of iterations of endomorphisms ) .
the contracting self - similar groups have been related by nekrashevych to finite partial self - coverings of orbispaces @xcite . in the current article
we use automata in the sense of transducers . as such
, they generate self - similar groups ( or semigroups ) of tree automorphisms ( or endomorphisms ) . in the same time , the output words are interpreted as encodings of integers , thus bringing the topic closer to the topic of automatic sequences .
thus , it is not surprising that the concrete examples of transducer integer sequences that are exhibited here all gave high level of self - similarity and can be defined as limits of certain iterations of sequences .
for @xmath7 denote @xmath8 .
the free monoid @xmath9 has the structure of a @xmath10-ary _ rooted tree _
@xmath9 in which the _ empty word _ @xmath11 is the _ root _
, the words of length @xmath12 constitute the _ level @xmath12 _ and each vertex @xmath13 has @xmath10 _ children _ , namely @xmath14 , for @xmath2 a letter in @xmath15 ( see figure [ 3tree ] for the ternary tree ) . @xmath16 \ar@{-}[d ] \ar@{-}[rrrd ] & \\ & 0 \ar@{-}[ld ] \ar@{-}[d ] \ar@{-}[rd ] & & & 1 \ar@{-}[ld ] \ar@{-}[d ] \ar@{-}[rd ] & & & 2 \ar@{-}[ld ] \ar@{-}[d ] \ar@{-}[rd ] & & \\
01 & 02 & 03 & 10 & 11 & 12 & 20 & 21 & 22 \\
\dots & \dots & \dots & \dots & \dots & \dots & \dots & \dots & \dots & } \ ] ] the tree structure imposes order on @xmath17 , which is the well known _
prefix order_. namely , we say that @xmath18 if @xmath19 is a vertex on the unique geodesic from @xmath13 to @xmath11 in @xmath9 , which is equivalent to saying that @xmath19 is a _ prefix _ of @xmath13 .
a map @xmath20 is a _ tree morphism _ if it preserves the word length and the prefix relation , i.e @xmath21 for all words @xmath19 and @xmath22 over @xmath23 . in case @xmath24
, morphisms are called _ endomorphisms _ and bijective endomorphisms are called _
every tree morphism @xmath25 can be decomposed as @xmath26 where @xmath27 is a map called the _ root transformation _ of @xmath28 and @xmath29 , @xmath2 in @xmath23 , are tree morphisms called the _ sections _ of @xmath28 .
the root permutation and the sections of @xmath28 are uniquely determined by the recursive relation @xmath30 which holds for every letter @xmath2 and word @xmath22 over @xmath23 .
thus the sections describe the action of @xmath28 on the @xmath31 subtrees hanging below the root in @xmath32 and the root transformation @xmath33 describes the action of @xmath28 at the root .
a quite efficient way of defining tree morphisms is by using finite transducers .
finite @xmath31 to @xmath35 transducer _ is a 5-tuple @xmath36 , where @xmath37 is a finite set of _ states _ , @xmath23 and @xmath38 are the _ input and output alphabets _ , @xmath39 is a map called the _ transition map _
od @xmath40 , and @xmath41 is a map called the _ output map _ of @xmath40 .
every state @xmath42 of the finite transducer @xmath40 defines a tree morphism , also denoted @xmath42 by setting @xmath43 , for @xmath44 , and @xmath45 to be the restriction of @xmath46 defined by @xmath47 .
thus , for each state @xmath42 of @xmath40 we have @xmath48 for @xmath2 a letter in @xmath15 and @xmath22 a word over @xmath15 .
when started at state @xmath42 , the transducer reads the first input letter @xmath2 , produces the first letter of the output according to the transformation @xmath49 and changes its state to @xmath50 .
the state @xmath50 then handles the rest of the input and output .
the states of a @xmath10-ary transducer ( transducer in which @xmath51 ) define @xmath10-ary tree endomorphisms .
an invertible @xmath10-ary transducer is a transducer in which @xmath51 and the transformation @xmath49 is a permutation of @xmath15 , for each state @xmath42 in @xmath37 .
the states of an invertible @xmath10-ary transducer define @xmath10-ary tree automorphisms .
when @xmath52 and , for each state @xmath42 , the vertex transformation @xmath49 is injective then every state of the transducer @xmath40 is an embedding of the @xmath31-ary tree into the @xmath35-ary tree .
we call such a transducer an _ injective transducer_. the boundary @xmath53 of the @xmath10-ary tree @xmath9 consists of all infinite ( to the right ) words over @xmath15 .
the boundary has a structure of an ultrametric space homeomorphic to a cantor set .
the recursive definition applies to both finite and infinite words @xmath22 .
the action of a state @xmath42 of a @xmath10-ary transducer on the boundary @xmath53 is by continuous maps , while the action of an invertible @xmath10-ary transducer is by isometries .
more on these aspects of actions on rooted trees can be found in @xcite .
there are two common ways to represent finite @xmath31 to @xmath35 transducers by labeled directed graphs such as the ones in figure [ ah - al ] .
the graph on the left represents an invertible ternary transducer , denoted @xmath54 .
the vertices are the states , each state @xmath42 is labeled by its corresponding transformation ( in this case permutation ) @xmath49 , and the edges labeled by the letters from @xmath55 define the transition function @xmath56 ( for every @xmath42 in @xmath37 and @xmath2 in @xmath55 there exists an edge from @xmath42 to @xmath43 labeled by @xmath2 ) .
the graph on the right represents a 3 to 2 transducer .
the vertices are the states and for each pair @xmath57 in @xmath58 an edge labeled by @xmath59 connects @xmath42 to @xmath50 .
one can easily switch back and forth between the two formats .
we refer to the second form ( the one in which the output is indicated on the edges ) as the _ moore diagram _ of the automaton .
consider the transducer in the right half of figure [ at ] .
it is known @xcite ( see also @xcite ) that the group @xmath66 is the lamplighter group @xmath5 which is the wreath product of the cyclic group of order 2 ( representing a switch ) and the infinite cyclic group ( representing moves between consecutive lamps ) .
the realization of the lamplighter group @xmath5 by the automaton @xmath67 was used by grigorchuk and uk @xcite to calculate the spectrum of the markov operator on the cayley graph of @xmath5 , which then lead to the solution of strong atiyah conjecture in @xcite . for @xmath68 , the row @xmath0 word @xmath69 in the matrix @xmath70 ( the @xmath0-th gray code word of length @xmath12 ) is equal to @xmath71_2^r)^r$ ] , where @xmath72 denotes word reversion and
@xmath73_2 $ ] is the length @xmath12 representative of @xmath0 ( including leading zeros , if necessary ) .
we can define a variation on the notion of transducer integer sequences as sequences that can be obtained from transducers by reading the input starting from the most significant digit ( and interpreting the output as starting from the most significant digit ) .
call these sequences sf transducer sequences ( for significant first ) .
since the sequence of binary gray code words can be obtained by feeding the binary representations of integers , most significant digit first , into @xmath67 starting at @xmath74 , we see that the sequence a003188 of integers @xmath75 represented by the binary gray code words is a sf binary transducer sequence . on the other hand , this sequence is not a binary transducer sequence .
this is clear since the transformation @xmath76 does not preserve prefixes .
namely @xmath77 , while @xmath78 .
we offer two 2 to 3 transducers each of which generates all the configurations on the geodesic lines between the regular configurations @xmath79 , @xmath80 and @xmath81 ( depending on chosen initial state ) . the first one uses the order prescribed by the binary gray code , while the other uses the natural order .
the transducer @xmath82 in figure [ optimal ] generates the optimal configurations in hanoi towers problem .
more precisely , for @xmath83 , @xmath84 , starting at state @xmath85 , and feeding the reversal @xmath86 of the length @xmath12 row @xmath0 binary gray code word from @xmath70 into @xmath82 produces the reverse of the length @xmath12 ternary word representing the unique @xmath12 disk configuration at distance @xmath0 along the geodesic from @xmath87 to @xmath88 in @xmath89 .
@xmath90[f- ] { } \ar@/_1pc/[r]^{1/1 } \ar@/^1pc/[d]^{0/0 } \ar@{}[l]|<<<<{\textstyle t_{01 } } & * + + [ o][f- ] { } \ar@/_1pc/[r]^{1/2 } \ar@/^1pc/[d]^{0/1 } \ar@{}[l]_<<<{\textstyle t_{12 } } & * + + [ o][f- ] { } \ar@/_2pc/[ll]_{1/0 } \ar@/^1pc/[d]^{0/2 } \ar@{}[r]|<<<<{\textstyle t_{20 } } & \\ & * + + [ o][f- ] { } \ar@/_2pc/[rr]_{1/2 } \ar@/^1pc/[u]^{0/0 } \ar@{}[l]|<<<<{\textstyle t_{02 } } & * + + [ o][f- ] { } \ar@/_1pc/[l]^{1/0 } \ar@/^1pc/[u]^{0/1 } \ar@{}[l]^<<<{\textstyle t_{10 } } & * + + [ o][f- ] { } \ar@/_1pc/[l]^{1/1 } \ar@/^1pc/[u]^{0/2 } \ar@{}[r]|<<<<{\textstyle t_{21 } } & } \ ] ] it is well known that the unique geodesic path of length @xmath94 from @xmath87 to @xmath88 connects @xmath87 to @xmath95 in the first @xmath96 steps , then in the next step the largest disk is moved to get the configuration @xmath97 and then the last @xmath96 steps are used to connect @xmath97 to @xmath88 .
since we want to use gray code words to describe the configurations along the way , we observe that in the first part of the geodesic from @xmath87 to @xmath88 ( corresponding to the last digit in the gray code being 0 ) the last digit in the reached configurations is @xmath2 , while in the second part ( corresponding to the last digit in the gray code being 1 ) the last digit in the reached configuration is @xmath3 .
this explains the root transformations in the above decomposition . as for the sections , in the first part of the geodesic ( last digit 0 in the gray code ) the configurations corresponding to the first @xmath98 digits describe the path from @xmath99 to @xmath100 , while in the second part ( last digit 1 in the gray code ) the configurations corresponding to the first @xmath98 digits describe the path from @xmath100 to @xmath101 in the natural order and from @xmath101 to @xmath100 in the gray code word order ( because of the flip in the second half of the gray code ) .
thus the section at 0 is @xmath102 and the section at 1 is @xmath103 .
the transducer @xmath104 in figure [ optimal2 ] generates the optimal configurations in hanoi towers problem .
more precisely , for @xmath83 , @xmath84 , starting at state @xmath1 , and feeding the reversal @xmath73_2^r$ ] of the length @xmath12 binary representative of @xmath0 ( including leading 0 s if needed ) into @xmath104 produces the reverse of the length @xmath12 ternary word representing the unique @xmath12 disk configuration at distance @xmath0 along the geodesic from @xmath87 to @xmath88 in @xmath89 .
@xmath105[f- ] { } \ar@/^1pc/[rr]^{0/0 } \ar@{}[ddrr]|<<<<<{\textstyle q_{01 } } \ar@/^1pc/[dl]^{1/1 } & & * + + [ o][f- ] { } \ar@/^1pc/[dr]^{1/2 } \ar@{}[ddll]|<<<<<{\textstyle q_{02 } } \ar@/^1pc/[ll]^{0/0 } & \\ * + + [ o][f- ] { } \ar@/^1pc/[ur]^{1/1 } \ar@{}[r]|<<<<<<{\textstyle q_{21 } } \ar@/^1pc/[dr]^{0/2 } & & & & * + + [ o][f- ] { } \ar@/^1pc/[dl]^{0/1 } \ar@{}[l]|<<<<<{\textstyle q_{12 } } \ar@/^1pc/[ul]^{1/2 } \\ & * + + [ o][f- ] { } \ar@/^1pc/[ul]^{0/2 } \ar@{}[uurr]|<<<<<{\textstyle q_{20 } } \ar@/^1pc/[rr]^{1/0 } & & * + + [ o][f- ] { } \ar@/^1pc/[ll]^{1/0 } \ar@{}[uull]|<<<<<{\textstyle q_{10 } } \ar@/^1pc/[ur]^{0/1 } & } \ ] ] observe that , for any permutation @xmath91 of the three letters in @xmath55 the states of the transducer @xmath104 have ( as tree morphisms ) the decomposition @xmath106 where @xmath93 .
this is precisely the decomposition that corresponds to the natural order in the previous proof .
the automaton @xmath104 , started at @xmath107 , generates the sequence a055661 @xmath108 but only when all input words are adjusted by leading zeros to have odd length , and it gives the sequence @xmath109 which does not appear in the on - line encyclopedia of integer sequences ( as of december 2006 ) , when the input words are adjusted to have even length .
in fact , the former sequence records the integers whose ternary representations give the configurations in the hanoi towers problem on the geodesic line in the infinite schreier graph @xmath110 determined by applying repeatedly the automorphisms @xmath111 , @xmath112 and @xmath113 ( in that order ) and the latter records the integers whose ternary representations give the configurations on the geodesic line in @xmath110 determined by applying repeatedly the automorphisms @xmath112 , @xmath111 and @xmath113 ( in that order ) .
there is nothing strange in this split , since it is known that the optimal solution transferring disks to peg 1 follows different paths depending on the parity of the number of disks . by flipping the input and the output symbol in the automata @xmath82 and @xmath104 we obtain two automata that can be used to recognize the configurations on the geodesic lines between @xmath114 , @xmath115 and @xmath116 and
encode them either by using the gray code words or binary representations . more generally , when @xmath36 is injective transducer one can define a partial inverse transducer @xmath117 in which @xmath118 , and @xmath119 and @xmath120 are partial maps , defined by @xmath121 and @xmath122 whenever @xmath123 and @xmath124 .
the inverse transducer @xmath125 , recognizes the optimal configurations in hanoi towers problem .
more precisely , starting at the inverse state @xmath126 , @xmath127 , @xmath84 , and feeding ternary words of length @xmath12 into the inverse transducer @xmath125 , only reversals of ternary words representing configurations on the geodesic from @xmath87 to @xmath88 in @xmath89 are read entirely by the transducer and , for such configurations , the reversal of the corresponding binary gray code word of length @xmath12 is produced in the output .
the inverse transducer @xmath128 , recognizes the optimal configurations in hanoi towers problem .
more precisely , starting at the inverse state @xmath129 , @xmath127 , @xmath84 , and feeding ternary words of length @xmath12 into the inverse transducer @xmath128 , only reversals of ternary words representing the configurations on the geodesic from @xmath87 to @xmath88 in @xmath89 are read entirely by the transducer and , for such configurations , the output represents reversals of the binary representation of the distance to @xmath87 .
for example , the configuration @xmath130 is not accepted starting from the state @xmath131 ( after it is fed into @xmath128 as @xmath132 it stops after reading the first 4 symbols in state @xmath133 and it can not read the last symbol ) .
this simply means that this configuration is not on the geodesic between @xmath134 and @xmath135 . on the other hand , @xmath136 is read completely and it produces the output @xmath137 , which says that the configuration @xmath136 is on the geodesic between @xmath134 and @xmath135 and its distance to @xmath134 is @xmath138 .
if we read @xmath136 starting at state @xmath139 in @xmath128 we obtain the output @xmath140 , which confirms that the configuration @xmath136 is on the geodesic between @xmath135 and @xmath134 and that its distance to @xmath135 is @xmath141 . laurent bartholdi and zoran unik . some solvable automaton groups . in _ topological and asymptotic aspects of group theory _ , volume 394 of _ contemp .
_ , pages 1129 .
soc . , providence , ri , 2006 .
rostislav i. grigorchuk and andrzej uk . on the asymptotic spectrum of random walks on infinite families of graphs . in
_ random walks and discrete potential theory ( cortona , 1997 ) _ , sympos . math . ,
xxxix , pages 188204 .
cambridge univ . press ,
cambridge , 1999 . | the notion of transducer integer sequences is considered through a series of examples . by definition ,
transducer integer sequences are integer sequences produced , under a suitable interpretation , by finite automata encoding tree morphisms ( length and prefix preserving transformations of words ) .
transducer integer sequences are related to the notion of self - similar groups and semigroups , as well as to the notion of automatic sequences . * tree morphisms , transducers , and + .1 in integer sequences * zoran uni + department of mathematics + texas a&m university + college station , tx 77843 - 3368 , usa + [email protected] + .2 in 2000 _ mathematics subject classification _ : primary 11y55 ; secondary 11b85 , 20m20 , 20m35 .
+ _ keywords _ : transducers , integer sequences , automatic sequences , self - similar groups , self - similar semigroups . |
the formation of collapsed objects , both in the context of the collapse of molecular cloud cores and in the later context of the clumping of material in a circumstellar disk is very difficult to model numerically because of the huge range of spatial scales involved .
accurate simulation of the collapse usually requires that all of these scales be well resolved , if the result is not to be contaminated by numerically induced fragmentation . for three dimensional ( 3d ) simulations
, @xcite defined a minimum resolution condition for the numerical validity of a simulation that models the collapse of a molecular cloud core using a grid based hydrodynamic code .
contemporary work by ( * ? ? ?
* hereafter bb97 ) has examined necessary resolution conditions for the collapse of a similar system in the context of smoothed particle hydrodynamics ( sph ) simulations , and also discussed the influences that choices of the form of gravitational softening may have on the results . both of these works define minimum resolution criteria in the context of a jeans collapse of a gas cloud resolved in 3d , but while @xcite observe that fragmentation is enhanced by a failure of the criterion , bb97 only observe enhanced fragmentation if the gravitational softening used in their particle based simulations is smaller than the hydrodynamic smoothing .
otherwise , they observe that fragmentation can be unphysically delayed in a system where it is known to occur .
later work of @xcite explores the problem of jeans collapse in isolation .
they find that sph simulations of systems with initial conditions like that of the original linearized analysis do not fragment artificially , even when under resolved .
a number of recent works in the field of planet and brown dwarf formation @xcite have discussed the formation of planets and brown dwarfs via gravitational fragmentation in disks . in disks systems like those discussed ,
the jeans formalism used to develop the @xcite and bb97 criteria is not valid both because the disk scale height is usually small compared to the jeans wavelength and because of the existence of shear , which plays a role as important as self gravity for the growing structures .
as yet however , no analogous resolution criterion exists for disks , which may be modeled in either fully in two dimensions ( 2d ) , effectively integrating over the vertical coordinate , or in 3d , for which no vertical integration is assumed , but for which the jeans wavelength based criterion still may not apply .
in addition to requirements that simulations resolve the wavelengths of the relevant instabilities sufficiently , simulations must also satisfy a number of other criteria if they are to be believed .
approximations made outside the realm of the physical model , perhaps used to model the behavior of unresolved or poorly resolved phenomena , must not drive the results of the simulations themselves . in the context of simulations of disks ,
particularly those using particle based hydrodynamic methods such as sph , an important consideration will be the implementation of gravitational softening . at a more fundamental level ,
the algorithm used to solve the equations used to model physical system must do so accurately , without becoming unstable or generating large errors through some other sort of deficiency . a quick glance through the literature ( see e.g. textbooks of * ? ? ?
? * ; * ? ? ?
* ) will demonstrate to even a casual observer that the study of numerical methods in relation to their stability is one in which extensive studies on many topics have been performed .
verification of methods on physically relevant test problems however , is comparatively more widespread but may often suffer from insufficient detail or generality .
while many studies discuss the fidelity of the numerical solution on some simple or contrived test problem , often no test problems sufficiently similar to the physical system under study can even be devised . in section
[ sec : numerical - factors ] , we first extend the previous 3d work of @xcite and bb97 to self gravitating thin disk systems , then outline alternatives for gravitational softening in particle simulations , and the possible consequences each choice may have on results . in section [ sec : testing ] , we define a test problem with which we determine appropriate values for resolution for particle simulations .
we continue in section [ sec : softening ] with a discussion of the gravitational softening assumed in particle simulations , and the specific numerical issues encountered in a study of disks .
next , in section [ sec : disks-3d ] , we discuss the application of the criterion to simulations done in 3d , and define a test problem to validate the accuracy of numerical codes attempting to model disks in 3d . using this test problem
, we demonstrate that failure to resolve the vertical structure of disks will lead to large errors in the numerical solution for the evolution of the entire physical system .
finally , in section [ sec : remarks ] , we summarize our results and discuss them in the context of the models presented in the literature .
a numerical simulation of any physical system may suffer from inaccuracies from several different origins .
for example , an incorrect physical model or incorrect initial conditions will produce results irrelevant to the system being modeled .
on the side of numerics , a shortcoming in the numerical method may erroneously trigger some physical process to become active in the evolution , where in reality , no such physical process is important .
a shortcoming in the numerical method may also trigger effects of purely artificial origin . in this section ,
we first describe the conditions under which we may expect a numerically induced , but physically based instability leading to fragmentation to be present in simulations of disks .
secondly , we discuss alternative treatments of the hydrodynamics and gravity in particle simulations on the spatial scales of the smoothing and softening lengths in particle simulations .
we describe conditions under which we may expect their implementations to influence a simulation , possibly also leading to artificially induced fragmentation .
finally , we discuss the strengths and weaknesses of 2d and 3d treatments of a system , the meaning of physical quantities realized in a 2d approximation and consequences that may arise when the assumptions underlying one or the other treatment break down . a condition on the minimum resolution to ensure the collapse of a cloud is of physical rather than numerical origin was defined by @xcite , using the ratio of the local grid resolution , @xmath7 , and local jeans wavelength , @xmath2 , in the fluid : @xmath8 where @xmath7 is the size of a grid cell and @xmath2 is the local jeans wavelength : @xmath9 and @xmath10 is the sound speed , @xmath11 is the volume density and @xmath12 is the gravitational constant . to obtain a form more useful in particle based numerical methods ( e.g. sph ) where the resolution element is a unit of mass rather than of length , bb97 used the jeans mass as defined from energy considerations @xcite for a homogeneous sphere , to define an analogous criterion for the maximum resolvable density . generalizing their result to a gas with @xmath13 internal degrees of freedom we find : @xmath14^{1/2 } { { c_s^3}\over{(g^3\rho)^{1/2}}}\ ] ] where
the superscript ` energy ' is included to distinguish this definition of @xmath15 from two others defined in the next section , @xmath16 is the ratio of specific heats and @xmath13 is the number of degrees of freedom , equal to 3 for a monotonic ideal gas .
this equation yields a maximum resolvable density : @xmath17 { { c_s^6}\over{(m_p n_{\rm reso})^2}},\ ] ] where we equate the jeans mass @xmath15 to a sum of @xmath18 sph particle masses , @xmath19 , that are required to resolve it .
an exactly analogous stability condition can be made for rotationally supported ( i.e. disk ) systems using the local toomre wavelength , @xmath1 , @xmath20 where @xmath1 is the wavelength which defines neutral stability in disks .
we can derive @xmath1 from the dispersion relation for waves in disks , whose solution ( see e.g. , * ? ? ?
* ) has four branches : @xmath21 corresponding to leading and trailing , short and long wavelength spiral density waves .
the variables @xmath22 , @xmath4 and @xmath23 are defined by @xmath24 respectively .
physically , these variables are the wave number , the well known toomre @xmath4 parameter and the doppler shifted pattern frequency of wave of symmetry @xmath25 ( i.e. with @xmath25 spiral arms ) , normalized to the local epicyclic frequency , @xmath26 , in the disk .
they depend on the disk s surface density @xmath27 , its orbit frequency , @xmath28 , as well as the pattern frequency , @xmath29 and symmetry , @xmath25 , ( i.e. the number of spiral arms ) of the spiral waves .
neutral stability is defined by the condition that the term under the square root is zero , for which the wave number is @xmath30 .
the critical wavelength corresponding to this wave number is : @xmath31 @xcite derive an expression for the longest wavelength that will be unstable in a given disk as @xmath32 however this wavelength is of limited relevance to the question of numerical resolution because it is neither the most unstable nor the shortest unstable wavelength in the disk . as noted by them ,
when the disk first becomes unstable ( i.e. @xmath33 ) , @xmath34 is a factor of two longer than @xmath1 . due to our interest in determining a critical resolution requirement , which requires the shorter length scale
be used , we use the wavelength definition of eq .
[ eq : toom - wavel ] rather than that of eq .
[ eq : disk - critwavel ] in defining our criterion . as in the 3d case ,
a form useful for particle based simulations can be obtained , this time by defining a toomre mass . unlike the 3d case however , due to the more complicated geometry of the mass distribution , no easily derivable form of the toomre mass may be determined from energy considerations analogous to that in eq .
[ eq : mjeans ] .
we therefore use a circular volume element to define @xmath35 a maximum resolvable surface density follows directly as : @xmath36 where the symbols have the same meanings as in equation [ eq : max - voldens ] above .
determination of appropriate values for @xmath37 and for disk simulations will be the subject of section [ sec : testing ]
. it will frequently be useful to determine the resolution required for the same overall morphology but with varying stability , as required for a parameter study varying a disk s toomre @xmath4 value .
given the same overall morphology , @xmath4 becomes a function of sound speed only , which is also the only term in equation [ eq : max - surfdens ] that will vary .
a useful re - parameterization of this expression to illustrate the sensitivity of the resolution to the disk stability directly will therefore be to replace the sound speed with @xmath4 through its definition : @xmath38 the maximum resolvable surface density is thus proportional to the fourth power of @xmath4 . in order to obtain identical effective resolution , as quantified by identical values of @xmath39 , a simulation with @xmath33
must therefore be resolved with @xmath40 more particles than a simulation with @xmath41 . the criteria in equations [ eq : jeans - wavel ] and [ eq : max - voldens ]
can easily be shown to be equivalent up to constant factors by casting the former as a jeans mass .
the mass inside a sphere of radius @xmath42 is : @xmath43 so that comparison of equations [ eq : mjeans ] ( determined from a similar spherical volume assumption ) and equation [ eq : sphere - wavel - mjeans ] yields the desired identification . yet another definition of the jeans mass is given in @xcite using a cubic rather than spherical volume element as : @xmath44 in each of these definitions , the constant factors differ . specifically , the three forms @xmath45 , yield values of the jeans mass in a ratio of @xmath46 relative to each other , assuming a monatomic ideal gas .
analogously for the 2d case , using a square area element , @xmath47 rather than circular , would result in a definition of toomre mass or density that is a factor @xmath48 larger than stated in equations [ eq : mtoomre ] and [ eq : max - surfdens ] , respectively .
although these constant factors may indeed be regarded as insignificant in a global sense since the basic functional dependence is the same , their quantification is important because it allows us to interpret the results of simulations obtained using various versions of the criterion .
for example , @xcite found empirically that @xmath3 was sufficient to suppress numerical instabilities in a test problem , which implies a mass per grid zone of at most @xmath49 . for particle simulations
, bb97 showed that numerical instability could be suppressed in a similar test problem by resolving the local jeans mass with particles less massive than @xmath50 .
although ostensibly very similar , these criteria differ by a hidden factor of @xmath51 from each other due to the differing proportionality factors , with the particle based criterion being more conservative ( requiring more particles to obtain equivalent resolution for a given mass distribution ) . in other words , the @xmath52 particle ( @xmath53 )
condition of bb97 translates to a @xmath54 particle ( @xmath55 ) condition condition if we use the jeans mass defined by equation [ eq : sphere - wavel - mjeans ] or a @xmath56 particle ( @xmath57 ) condition if we use the jeans mass definition of equation [ eq : cube - wavel - mjeans ] . in general , for equivalent resolution of the physical parameters of the flow , a particle simulation will require roughly ten times as many fluid elements as a grid code like the one discussed in @xcite . with the same number of fluid elements ( particles or grid cells ) a grid simulation will be able to resolve higher mass densities before becoming numerically unstable than a particle simulation .
the critical wavelength for disks from equation [ eq : toom - wavel ] is linearly dependent on the temperature to mass ratio through the sound speed and mass density , while the jeans wavelength in equation [ eq : jeans - wavel ] is dependent only on its square root .
thus , the toomre criterion will be more strict on smaller spatial scales than the jeans criterion , but less strict on larger scales .
we can determine the relative sizes of the two wavelengths , and the crossover point at which both are equal , from their ratio : @xmath58 if we assume that the disk is near keplerian so that @xmath59 , that the volume density and the surface density are related by @xmath60 , where @xmath61 is a coefficient specifying the exact proportionality between surface and volume densities , and that the local disk scale height is @xmath62 ( but see section [ sec : bad - approx ] below ) , then we can combine equation [ eq : jeans - wavel ] and [ eq : toom - wavel ] to produce @xmath63 thus if the local value of @xmath4 falls below about one half , as it may in regions beginning to fragment , the toomre instability wavelength will be a more restrictive criterion for numerical simulations than the jeans criterion used by @xcite .
on first inspection , equation [ eq : tj - relative ] and the wavelength proportionalities that went into it would seem to be backwards . when the collapse is well underway ( @xmath64 ) , rotation ceases to matter so that the collapse should proceed according to a jeans prescription .
in other words , we would expect that the jeans wavelength should be smaller than the toomre wavelength .
indeed , this would be the case for a collapsing region that is much smaller than the disk scale height , but this equation shows that for the initial stages of collapse relevant to the analytic wavelength derivations , it is not the presence or absence of rotation that is relevant , but rather the dimensionality of the problem that plays the most important role .
the mathematical analysis leading to the jeans resolution criterion is fully three dimensional , while that leading to the toomre criterion is limited to two dimensions : the equations are integrated over the @xmath65 coordinate .
the difference is important because the jeans analysis may not be valid in a disk simulation ( even those evolved using fully 3d models ) because it assumes a homogeneous medium which is infinite in all three spatial coordinates . by definition , a disk structure will violate this assumption since the matter will always be condensed into a midplane above and below which relatively little matter lies .
the violation will be important if the jeans wavelength is comparable to or larger than the disk scale height , determined for the local conditions in a given disk . in this case
, even the initial conditions of a system may not satisfy the underlying assumptions of the analysis .
taking the specific example shown in figure [ fig : boss - crit]b of section [ sec : bad - approx ] below , we find that jeans wavelength is long compared to the disk scale height everywhere , so that the jeans analysis is indeed inapplicable . on the other hand ,
the analysis leading to the toomre wavelength is valid in the limit of a thin , rotating system for which a ` surface density ' is a meaningful concept , whether or not a given simulation is actually performed in two or three dimensions . by construction of the analysis itself ,
disks fall into this category , so at least we may construct initial conditions that satisfy the underlying model .
as in the case of the jeans wavelength , by examining figure [ fig : boss - crit]b , we see that the toomre wavelength is also long in comparison to the disk scale height . in this case , the large ratio means only that the analytic assumptions become more valid , not less .
violation of the model assumptions may still be important with wavelengths comparable to a disk scale height scale because of the neglect of vertical motion and structure in the analysis .
violation of the assumptions in the analytic derivations may also occur at times after simulations have evolved for some time because the analyses assume that variations of all quantities from their initial values are small , even if the initial conditions satisfy all requirements of the linearized analyses .
for example , a density perturbation may be only slightly enhanced from the local background in a fragmenting molecular cloud , while the background potential is characterized by a comparatively steep overall gradient due to some large structure nearby .
a second example may be that fluid velocities are large while other perturbations are small because the initial state was seeded with some spectrum of turbulent velocity perturbations .
finally , numerical simulations may not solve the equations of hydrodynamics accurately , or may do so with only poor fidelity for some problems .
unfortunately , we will find that in the most relevant range of parameter space for disk simulations , both the jeans and toomre wavelengths reach values comparable to the disk scale height and local perturbations reach amplitudes beyond those for which linearized analyses are strictly valid .
as with the original 3d analysis of collapsing clouds , we therefore resort to numerical simulations to determine the required ` safety factor ' ( i.e. the values of , ` @xmath37 ' and ` @xmath66 ' above for disk or cloud collapse analyses ) for which we can be reasonably confident that the basic features of that analysis are not violated , even though it may be applied in a region where its assumptions may be called into question .
although there is only one criterion for disk evolution and one for cloud collapse , there are effectively five implementations of them as applied to disk systems that could be used under different circumstances
. first the equations [ eq : jeans - cond ] or [ eq : toom - cond ] might be used directly : a 2d simulation could use the directly available surface density , or a 3d simulation could use the directly available volume density .
the criteria could also be used indirectly using the disk scale height to convert between volume density and surface density .
finally , a surface density could be obtained from a 3d simulation by directly integrating over the @xmath65 coordinate , for use in the toomre criterion .
we will show in section [ sec : bad - approx ] that using the approximate indirect forms ( i.e. making the volume or surface density conversion using the disk the scale height ) can yield seriously discrepant values of the stability wavelength , perhaps leading to erroneous conclusions regarding the veracity of a given simulation . in all numerical simulations , the modeler would like to resolve the largest range of spatial scales possible , so that both the smooth and highly inhomogeneous regions are accurately modeled . in this regard , a limit on the resolution will be the scale on which the gravitational and hydrodynamic forces can be resolved . for grid based simulations , resolution of both the hydrodynamic features in the flow and gravitational forces will be related to the local dimensions of the grid ( see e.g. * ? ? ?
* ; * ? ? ?
* ) . for particle simulations ,
the limits will be related to two length scales , one for gravity and one for hydrodynamics , each defining the spatial extent of the particles in different ways . in order to make clear that they are distinct from each other
, we will make a distinction between the terms used to refer to each .
specifically we will refer to gravitational ` softening ' lengths and hydrodynamic ` smoothing ' lengths for particles , to describe each effect .
fluid quantities in a particle based simulation using sph are reconstructed from the positions of the particles at a given time .
contributions of particles are weighted according to a smoothing function ( the ` kernel ' ) and all contributions are summed to define each of the hydrodynamic quantities at the current location of each particle @xcite .
quantities such as density are calculated using the kernel directly , while forces due to pressure are calculated using its derivative . in each case
our implementation follows the discussion in @xcite and , in the work presented here , we use the now standard spline kernel of @xcite given by : @xmath67 here , @xmath23 is the number of dimensions , @xmath68 , @xmath69 is the normalization with values of @xmath70 and @xmath71 in two or three dimensions respectively
. we will also study the effect of an important modification of that derivative described by @xcite , which takes the form : @xmath72 and acts to reduce unphysical particle clumping @xcite , by providing a net repulsive pressure force even at zero separation .
the modification affects the region @xmath73 where in its unmodified form , the derivative decreases monotonically to zero as @xmath74 itself approaches zero .
price ( 2005 , personal communication ) notes that this modification may cause changes in the effective sound speed determined from linear analyses , because normalization conditions on the kernel s first and second derivatives are not satisfied .
the normalization conditions will be exact only in the limits where the smoothing lengths approach zero , neighbor counts approach infinity and the distribution of neighbors over the kernel volume allows an accurate correspondence between a summed and an integral form of the equations . since none of these three conditions are realized in practice , it is difficult to evaluate what consequences may result in actual simulations . in section [ sec:2dtest - evo ]
we investigate the differences between using the standard and modified kernel derivatives on the outcomes of our simulations .
detailed analyses of the numerical stability and errors resulting from this modified formulation are beyond the scope of this work and we defer such discussions to the future . an important subtlety of particle simulations is to employ a gravitational softening length that removes the singularity in the force obtained when two particles in a simulation approach each other too closely , effectively modifying newton s law of gravitation .
such a modification is equivalent to defining a mass density distribution for the particles , so that they include an assumption of some spatial extent , rather than that they are point like objects .
modifying the gravitational force is acceptable in hydrodynamic and collisionless @xmath75-body simulations because particles do not in fact represent single particles in the underlying system , but only a statistical representation of the local distribution of gas or particles .
the question we face is how to implement an appropriate gravitational softening length in our simulations which , on one hand , prevents unphysically large forces from developing between particles , and on the other , is small enough to allow small scale features to develop in the flow .
two alternatives for implementing softening that are common throughout the literature are first to assume each particle has the mass distribution of a plummer sphere , so that the force law is modified to the form of a plummer force law @xcite , or to use a spline based kernel as discussed by @xcite , who interprets the smoothing kernel used to realize the hydrodynamic quantities as a mass distribution .
two variants are common in the latter case , first to use a softening that varies with the local conditions ( usually chosen to be identical to the sph smoothing length of each particle ) or second , to fix the softening to a single constant value at the beginning of the simulation . examples of variable kernel softening are found in codes used by @xcite , bb97 and our own previous work , while the treesph , gadget and gasoline codes of @xcite and @xcite , respectively , use fixed softening .
we implement the variable kernel softening variant using same kernel used for the sph quantities given by equation [ eq : sphkern ] . for gravitational softening ,
the masses used to compute the acceleration is modified from its newtonian form according to the prescription that a source particle s mass is reduced from its true value by a factor proportional to the volume enclosed by a sphere whose radius is the separation between the particles : @xmath76 where @xmath19 is the mass of the particle from which a force contribution is to be calculated and @xmath77 is the number of spatial dimensions . the sink particle , on which forces are calculated , is assumed to be a point mass , so that for two particles of mass , @xmath78 and @xmath79 , the force exerted by particle 1 on particle 2 is @xmath80 where @xmath81 is the separation between the particles one and two . except for a change of sign ,
this definition is manifestly invariant to exchanging the identities of the two particles , and therefore conserves momentum exactly . in 3d , and as the separation between two particles decreases , the gravitational force between them will also decrease , ultimately to zero at @xmath82 , because @xmath83 depends on @xmath84 . in 2d
, @xmath83 is proportional only to @xmath85 and the force instead approaches a _
non_-zero constant value as the separation decreases to zero , rather than zero .
the physical reason for the discrepancy is that while equation [ eq : mass - kernsoft ] implies a 2d structure for the mass distribution , equation [ eq : forc - kernsoft ] retains a 3d structure for the forces they cause . on the scale of the interparticle separation
, the scale height of the disk not be negligible , so that in fact it is truly three dimensional .
we are aware of only one treatment that attempts to account for the conflict between the 2d and 3d behaviors @xcite , by assuming a vertical structure and numerically integrating the contributions to the forces on a point mass over the @xmath65 coordinate .
thereafter koller applies the derived correction factor to the forces , tabulated as a function of distance .
we expect koller s treatment will not be generally suitable for particle simulations because it requires different tables for different vertical structures and because , even with the modification , he finds that a plummer - like softening term is still required , though it can be made much smaller in magnitude .
instead , we propose an alternative form of softening , still based on the 2d kernel softening discussed above , with the modification that the effective mass @xmath83 is multiplied by an additional modification factor so that @xmath86 the choice of this exact form of softening is arbitrary , but is motivated by three desirable conditions on the force within and outside the softened region .
we require that the force decreases linearly to zero at zero separation , yielding a truly collisionless form .
second , at separations @xmath87 , we require the force returns to its correctly normalized , perfectly newtonian form and , finally , the derivative of the force is continuous , so that the force varies smoothly at all separations . coincidentally , for @xmath73 , this form duplicates the algebraic form of the derivative of the standard kernel , a fact that will prove advantageous for obtaining ratios between gravitational and pressure forces near unity in section [ sec : softsmooth - limits ] . throughout this paper
we implement gravitational softening using equation [ eq : mass - kernsoft ] with the variable smoothing length , @xmath88 , as its length scale .
we will perform separate series of simulations employing either the original mass distribution defined by equation [ eq : mass - kernsoft ] and incorporating the kernel in equation [ eq : sphkern ] , or the modified mass distribution of equation [ eq : mhat-2d ] .
an important advantage of variable gravitational softening is that it allows a modeler to soften the gravitational forces on the same length scale used for generating the hydrodynamic quantities .
the variation is required because the hydrodynamic quantities are generated using an approximately ( or exactly ) fixed number of ` neighbors ' but the local particle density is not constant . in order to retain the approximately fixed neighbor count , the smoothing must be correspondingly smaller in high density regions than in low density regions . in order to retain the equality over the duration of a simulation
, the softening must also be allowed to vary according to local conditions .
the advantage of softening with the same length scale as smoothing is that large imbalances in the gravitational and hydrodynamic forces between pairs of particles can not develop due to particles being in range of one or the other of the lengths , but not both .
the disadvantage is that energy may no longer be conserved because no account is made of the change in the internal mass distribution of particles as their smoothing length changes .
contributions to gravitational potential energy dependent on such changes will therefore not be evaluated correctly . in the case of fixed softening ,
an advantage is that it conserves energy , with disadvantages including the fact that the smallest resolvable length scale is both fixed at the beginning of the simulation and is the same everywhere .
a collapsing or expanding body may quickly reach a size where the flow is dominated by the softening in one region , while in another the flow may become unphysically point - like .
several previous works have discussed the consequences of the alternatives in simulations of different types , but no clear consensus has emerged from the discussion . for example , bb97 note that the jeans wavelength , cast in the form of a jeans mass , must be well resolved by both the gravitational softening length and by the sph smoothing length , used respectively to ensure numerical stability in the code and to produce hydrodynamic quantities from the particle distribution . for simulations in which the jeans wavelength is much larger than either the smoothing or softening lengths ( i.e. , that the region is stable against gravitational collapse ) , bb97 claim that little difference in behavior should be expected . however , for a marginally stable mass distribution ( e.g. one jeans mass distributed over a volume of radius one jeans length ) , fragmentation could be artificially suppressed or enhanced by changing the ratio of the gravitational softening to sph smoothing length , because of the large force imbalances that develop .
to avoid such artificial results , they recommend a variable gravitational softening set to the same length scale as the hydrodynamic smoothing . in the alternative ,
when fixed softening must be used , they recommend a softening length no smaller than the value that would cause the resolution condition of equation [ eq : max - voldens ] to be violated
. other work @xcite fixes the softening length , but allows the smoothing to vary , down to a limit of either @xmath89 or @xmath90 .
they show that a simulation including cooling will produce much more fragmentation in the latter case , and conclude that smoothing lengths must be restricted to be greater than the softening length to avoid artificial fragmentation .
although implementing fixed softening rather than variable , their conclusion appears consistent with that of bb97 , but takes no account of artificially suppressed fragmentation . on the other hand , @xcite examine fixed softening models where the smoothing is allowed to vary with and without the constraint that it may not decrease below the softening length .
they conclude that the smoothing does not need to be equal to softening and that smoothing must not be constrained because hydrodynamic shocks can not be properly resolved on size scales smaller than the softening length . for sph , where hydrodynamic quantities are derived from interpolations between pairs of particles , a necessary mathematical condition for the interpolation kernel
is that it be continuous and have a continuous first derivative ( see e.g. * ? ? ?
* for additional details ) .
physically , the requirement is equivalent to the statement that contributions to hydrodynamic quantities like mass density and mutual pressure forces undergo no discontinuous jumps as pairs of particles approach each other .
an unfortunate consequence of the two continuity requirements is that for kernels like equation [ eq : sphkern ] , particles that approach each other experience pressure forces that first increase with decreasing separation , then decrease to zero as the separation between them decreases to zero .
the consequence is unfortunate first , because , to the extent that we can regard the approach of two particles ( rather than a large sample of particles ) as representing a compression , the fact that the pressure force decreases to zero as they approach coincidence means that our intuitive expectation that pressure forces continue to increase during a compression is violated .
secondly , and as noted above , it may lead to unphysical particle clumping . as described by @xcite for non - self gravitating simulations
, the reason is that during the natural course of a simulation , particles that find themselves closer than a critical separation distance ( @xmath91 for the kernel in equation [ eq : sphkern ] ) where the pressure force is at its maximum , experience relatively smaller mutual pressure forces pushing them apart , and relatively similar external forces perturbing their motion . because of these small forces , particles continue to travel together with an end state in which pairs of particles actually coincide , defining what herant calls a ` pairing instability '
. the astute reader will not fail to note that discussing a ` pairing instability ' in the context of a sph simulation would seem to be rather irrelevant .
sph particles are of course not assumed to represent actual physical particles at all , but rather some ( statistical ) realization of an underlying physical system . therefore any such discussion would appear to be meaningless . in our defense ,
we point out that our discussion is of a failure mode for the method where the statistical assumptions break down , and therefore will require a careful analysis . as in the case for smoothing ,
softened mutual gravitational forces first increase in magnitude as pairs of particles approach each other , then decrease to zero as they approach still further . in this case however , the decrease is not due to any constraint on the kernel , but rather on the combined assumptions that the softening represents a mass distribution and that , at some separation , the mass enclosed by a sphere whose radius equals that separation is less than the total mass .
then , by invoking gauss s law , only the fraction of mass inside the sphere contributes to the force . in 2d simulations , where the geometric arguments leading to the gauss s law result do not hold ( see e.g. figure 3 of
* ) , it is necessary to relax the strict identification of the kernel with a mass distribution . for the purpose of avoiding infinite force contributions at coincidence , softening according to this procedure is effective , though as we show below , problems remain due to the fact that the force still does not decrease to zero at coincidence . in both cases ,
the reason for the decrease in force magnitude at small distances is to avoid numerical pathologies : for gravity , the development of unphysical point - like force contributions , for hydrodynamics , unphysical discontinuities in the forces and other hydrodynamic quantities .
the important point to note regarding both softening and smoothing is that on such scales , the magnitudes of the forces are dependent on assumptions made outside the realm of the physical model . in other words ,
the forces computed there are under resolved and any developing phenomena sensitive to effects in that region can be due only to an external assumption rather than to any physical process . in order to ensure physically valid simulations , it will therefore be important to ensure that effects originating on unresolved scales do not drive the results .
moreover , because they act on similar scales but with effects of opposing sign , it will be important to consider the resolution limits of both softening and smoothing lengths together . setting a softening length much smaller than the smoothing length is equivalent to the statement that at some distances pressure forces are under resolved ( i.e. that an error is made in evaluating them ) , but gravitational forces are not under resolved ( that no error is made in evaluating them ) . the same statement is true in reverse when the smoothing length is smaller than the softening . regardless of whether or not one or the other force actually can be resolved better than the other , the numerical assumptions always limit the effective , physically correct force resolution to the larger of the two scales . as pointed out by bb97 ,
the consequences of differing values of softening and smoothing lengths are net force imbalances of up to a factor seven at small separations , even when the softening and smoothing are only different by a factor of two .
much larger force imbalances develop when the ratios become more extreme , and bb97 attribute the unphysical outcomes of several simulations presented in the literature to this source .
figure [ fig : force - ratios ] shows the ratio of the gravitational and pressure forces between two particles located in a disk where the toomre @xmath4 value set to unity , realized in 2d . as for the 3d case , forces are nearly balanced when softening and smoothing are set to identical scales , but large imbalances are present when they are not identical . unlike the 3d case however , the force ratios for the standard kernel softening / smoothing option do not remain near unity as the separation , @xmath74 , goes to zero , but instead approach infinity . due to the reduced dimensionality , the gravitational force approaches a finite , non - zero value rather than decreasing to zero as it does in the 3d case .
force imbalances such as these , whether caused by unequal softening and smoothing lengths , or by the form of the softening or smoothing themselves ( or both ) , may influence the susceptibility of the simulation to artificially induced fragmentation .
when the imbalance favors gravity , particles are not only passively drawn to each other , but are also actively attracted . given an imbalance of the opposite sense , particles may be artificially repelled from each other , preventing fragmentation of physical origin from occurring .
given the intrinsic force imbalances present in 2d simulations using the standard kernel for smoothing , even when used with identical length scales , simulations may therefore be susceptible to artificially induced fragmentation .
on the other hand , the modified kernel derivative of @xcite allows the force ratio to remain near unity all the way to zero separation because the pressure approaches a finite , non - zero value in step with softened gravity .
similarly , the modified gravitational softening defined by equation [ eq : mhat-2d ] also results in force ratios near unity for equal softening and smoothing lengths , but in this case , both forces approach zero as their limiting value rather than a finite quantity . in both cases , large force imbalances
can develop when the softening and smoothing are unequal . with the modified softening variant , force imbalances of up to a factor @xmath92 at @xmath93 develop when the lengths are not equal , nearly twice the imbalance present with the alternate kernel derivative .
because the actual magnitudes are smaller , the net influence of the imbalance might be smaller .
we will investigate this question , the influence consequences of using the alternate kernel derivative formulation and the alternate softening formulation and the importance of using fixed softening or variable softening for which force imbalances may or may not develop respectively , in our work below .
as computational facilities have become more and more powerful , simulations of greater and greater complexity have become possible to perform at acceptable cost .
an important step in increasing complexity is to increase the dimensionality , first from one to two dimensions , and then to three . in the context of circumstellar disks ,
the full transition to 3d remains incomplete .
some workers prefer to limit their simulations to two dimensions , while others have begun work in 3d .
a number of significant consequences derive from each choice . of primary importance
is the computational cost and its relation to the spatial resolution affordable to the simulation . for a grid based method , computational cost in 3d will be approximately proportional to the fourth power of the number of cells in any one dimension , while for 2d
, the proportionality drops to the third power .
a similar proportionality will hold for particle simulations as well .
therefore , simulations in 3d will be able to employ fewer total cells or particles per dimension : linear resolution is intrinsically coarser in 3d than in 2d for simulations of comparable cost .
in contrast , while 2d simulations allow dramatically higher linear resolution in the two dimensions they actually model , they do so at the cost of requiring assumptions be made about the character of the system and its behavior in the third dimension .
for a 3d simulation , if it is to be truly three dimensional , the physical extent the system in each direction must be resolved by some number of particles or grid cells greater than one .
the exact minimum number will be a function of both the problem and of the method employed to evolve the system . in the context of circumstellar disks , the fact that the disk is spatially thin means that the available resolution must be allocated inhomogeneously in space if the cost of calculation is not to become too exorbitant . in a grid simulation for example , many more grid cells must be allocated to the same physical length in the @xmath65 coordinate than either the @xmath94 or @xmath95 coordinates , to resolve the vertical extent of the disk .
problems may arise from asymmetric grid resolution , including incorrect gravitational forces @xcite , or incorrect hydrodynamic evolution .
the problem is especially acute for sph simulations because the kernel used to reconstruct the hydrodynamic quantities is spherically symmetric .
although experiments with non - spherical kernels have been published , they are not common , due in part to the difficulty of implementing them in a form that retains angular momentum conservation ( e.g. * ? ? ?
* ) . with a spherical kernel and in the low resolution limit , particle smoothing lengths may actually exceed that of the disk s vertical extent , so that only a single fluid element spans the entire system in that coordinate .
similar conditions may occur in grid based simulations , if extremely asymmetric zone dimensions are not to be encountered .
any 3d simulation carried out under such conditions will effectively model only two dimensions , but will not include any supporting assumptions present in a simulation explicitly limited to 2d .
one such assumption will be in the description of the fluid itself . in
most ` normal ' fluids , pressure is by definition a scalar quantity whose gradient causes a force to be exerted in all directions .
the procedure used in sph to construct the hydrodynamic quantities however assumes a roughly spherical distribution of particles .
otherwise , the interpolations at the heart of sph are no longer interpolations but instead extrapolations in directions where few particles exist . as a result , and as is well known to its practitioners , the method becomes very inaccurate at boundaries . in circumstellar disk simulations , performed with sph at resolution where only one or few particles span the entire vertical extent of the disk , essentially _ all _ particles will be located near boundaries
. whether or not a particular simulation with one dimensionality will be more physically meaningful than a comparable one of the other depends on the quality of the assumptions made about the third dimension in a 2d simulation and whether those assumptions offset the loss of linear resolution affordable in 3d . in section [ sec : disks-3d ]
, we will show that 3d simulations of disks require much higher resolution than one might naively expect in order to reproduce basic hydrodynamic features of the flow correctly .
such high cost means that a given study will be able to perform many fewer simulations , severely constraining the possibility of performing the large parameter studies often required to fully explore the implications of a given physical model .
we have therefore concentrated on the study of disks in 2d throughout the rest of this paper .
because disks are in fact truly three dimensional , in spite of our approximation that they are thin , it will be important for correctly interpreting the results of the simulations to understand the consequences of a 2d approximation and where it may break down . for 2d simulations , two fundamentally different assumptions about the third coordinate
are possible .
the modeler may either assume that the simulation is modeling an infinite cylinder or that the simulation is modeling a thin system in which the system s dynamics and morphology are either negligible in the third dimension or some approximation is made regarding their behavior .
each assumption leads to quite different treatments of the hydrodynamics and gravitation in the simulation . in the case of the infinite cylinder assumption ,
each point in the plane actually corresponds to a line extending to infinite distance in the positive and negative third coordinate , which we will assume to be the @xmath65 coordinate . for simple hydrodynamical problems
, the interpretation poses no particular conceptual difficulty since hydrodynamic quantities will carry over from 3d to 2d unaltered .
the consequence for gravitational or electrostatic forces however , is that they become inversely proportional to the separation in the @xmath96 plane , rather than to the inverse square of the separation . for a thin system ,
the opposite situation holds .
inverse square law forces retain their familiar 3d form , but hydrodynamic quantities must be altered , specifically into integrals of their true 3d forms .
for example , surface density may replace volume density .
the circumstellar disks in this study are spatially thin , and the most natural interpretation of a 2d simulation of such a disk is that of a thin , vertically integrated model . in keeping with this characteristic and with many previous analytic and numerical treatments of disks ,
our 2d disk simulations are performed in this context . in order to allow readers to evaluate the results of our study more thoroughly ,
we now discuss several factors important for their interpretation and similar work by others .
in addition to the requirement for 2d simulations that state variables be integrated over the @xmath65 coordinate , a more subtle modification must also be made to other hydrodynamic quantities . @xcite and @xcite each discuss the modifications in the effective value of its polytropic index , @xmath97 , of the gas , corresponding to a degree of freedom corresponding to the disk ` puffing up ' in the third coordinate . both conclude that a value of @xmath16 slightly reduced from that expected for the 3d case should be used , due to the additional freedom .
when an isothermal equation of state is employed , the @xmath16 value will retain its limiting value of unity , so no affect will be present in the work here
. the change will also not be required in a truly ` razor thin ' 2d model , where motion is restricted in the third dimension entirely . in this case , the effective @xmath16 would increase instead , due to a decrease in the number of internal degrees of freedom for the gas .
vertically integrated quantities also require special consideration in realization and interpretation of gravitational forces in the system , even though they retain their familiar inverse square form . as noted in section [ sec : our - softening ] , gravitational forces obtained from a straightforward carryover of the 3d method of kernel softening exhibit a non - zero force at zero separation . from a physical perspective , such a condition will simply be erroneous , since in reality the disk mass described by the particles is spread over some vertical extent : it is no longer ` thin ' compared to interparticle separations .
where in reality , the force of one vertical column on another falls to zero at zero separation , the force based on a vertically integrated mass located at the disk midplane does not . from a numerical perspective
, this condition challenges the assumption that the particles are collisionless , in turn the assumption that softening was introduced to ensure . while considerably weakened , consequences will be less severe then in the 3d case because the interaction force does not become infinite at any separation , as it would if two truly collisional particles were to interact .
we may therefore attempt to salvage the gravitational forces via some simple modification , as we suggest with the altered softening prescription defined in equation [ eq : mhat-2d ] , or the altered description of pressure forces derived from equation [ eq : mod - kerderiv ] .
care must still be taken using either modification in any 2d simulation .
interparticle forces will deviate from their vertically integrated forms over spatial scales comparable to the scale height of the disk . at extremely high resolution
, where particles correspond to very thin columns of finite extent in the vertical direction , the softening or kernel derivative modifications will affect the calculated forces only over a small fraction of that distance . at separations between the particle size , @xmath88 , and the scale height , @xmath98
, forces will therefore be overestimated .
numerical experimentation has shown that deviations will not be large at resolutions where @xmath99 , but become much more significant when @xmath100 .
simulations presented here do not fall into the latter category .
what resolution is required ( i.e. what values of @xmath37 and from equations [ eq : toom - cond ] and [ eq : max - surfdens ] ) to ensure that a simulation that produces collapsed objects in a disk is producing numerically valid results ? as was done to develop criteria for jeans collapse , we will define a specific problem on which to compare the results of a numerical code at different resolutions . in parallel
, we will investigate the influence of three different strategies for the treatment of small scale interactions between particles : a ` base ' version and two variants that modify the kernel derivative in one case or the kernel softening in the other .
we use a small variation ( see below ) of a simulation discussed in @xcite , who used an sph code to model the evolution of disks in two dimensions .
simulations using sph are especially sensitive to violation of a resolution criterion because resolution is dynamically allocated . particle smoothing lengths are ordinarily considered to be functions of the local flow variables so that in high density regions , they shrink in an attempt to follow the small scale motions of the fluid there . in most respects
, this feature can be extremely desirable because there is no _ a priori _ reason to expect fragmentation in one or another part of a given simulation .
on the other hand and as we show below , insufficient care in its use can lead to numerically induced fragmentation .
we use the vine code ( wetzstein , nelson , in preparation ) in its ` sph only ' mode and using its leapfrog integrator to perform our simulations . an earlier version of this code , with a second order runge - kutta integrator , was used in the original calculations in @xcite .
exploratory tests with vine using this same integrator in the present simulations showed that similar results were obtained from both .
most calculations were performed with a single , global time step for all particles in order to assure that our results were unaffected by as few systematic effects as possible .
some tests were made with individual time steps in order to explore effects of numerical stability due to this source .
these calculations required @xmath101 times less computer time to complete and , while some differences between global and individual time step versions were present in the results , none materially affect the conclusions made from them .
vine uses a binary tree to organize particle data , so that they may be accessed efficiently for use in both the hydrodynamic and the gravitational force calculations . in order to avoid calculation times for the gravitational forces of order @xmath102 ,
sufficiently distant particles are approximated as nodes in the tree , resolved to quadrupole order in the actual calculation .
the acceptability criterion for the nodes was set so that forces on @xmath103% of particles would be accurate to @xmath104% .
vine employs an artificial viscosity with both bulk and von neumann - richtmyer terms to stabilize the evolution and to convert kinetic energy into thermal energy in shocks .
the coefficient for each term were set to @xmath105 and @xmath106 , which are values standard in the literature . using these values , simulations of disks using sph
are afflicted with a quite large and unphysical level of shear viscosity . in order to minimize such effects
, the simulations here were run with the shear viscosity reduction switch of @xcite , which reduces the magnitude by a factor @xmath101 .
important parameters from the simulations presented here and in the following sections are listed in table [ tab : sims ] .
the columns of the table show the resolution of each simulation , the type of gravitational softening used ( for sph simulations ) and , finally , the time at which the time at which the first clump is formed and the duration of the simulation .
simulation names ending in ` tc ' and ` grv ' denote simulations run with the modified kernel derivative of equation [ eq : mod - kerderiv ] or the modified gravitational softening of equation [ eq : mhat-2d ] , respectively , while those suffixes represent identical initial conditions but run with the unaltered kernel derivative or softening .
all simulations are performed in 2d and include the effects of self gravity , except the _ sgoff2d4 _ simulation in 2d which does not , and the remaining _ sgoff _ simulations , discussed in section [ sec : disks-3d ] , which also do not include self gravity and are done in 3d . [
tab : sims ] .simulation parameters [ cols="<,>,^,<,<,<",options="header " , ] @xcite modeled the evolution of self gravitating disks in 2d with masses between 0.05 and 1.0 times the mass of the central star , using sph .
other simulations from that work , performed using ppm , are not considered here because they could not be carried out far enough into the high amplitude regime to make fragmentation likely .
the specific model we consider here ( labeled ` scv2 ' in that work ) , had an assumed disk mass of = 0.2@xmath107 and a minimum toomre @xmath4 value defining its stability of = 1.5 .
we believe that model will be a particularly challenging test of the criterion because collapse was observed after only about one orbit of the outer disk edge , corresponding to about 11 orbits in the region where clumps first started to form . by the conclusion of the run at 1.6 ,
more than 30 clumps had formed . as in the originals , we have simulated the evolution in two dimensions , so that the surface density is directly available from the calculation .
the temperature and surface density of the gas were defined in that model with softened power laws as : @xmath108^{-q/2}\ ] ] and @xmath109^{-p/2}\ ] ] where @xmath110 and @xmath111 , respectively , and the core radius for both power laws was set to @xmath112 .
the constants @xmath113 and @xmath114 were determined from the assumed toomre stability parameter and the radial dimensions of the disk , defined at its inner edge by @xmath115 and its outer edge by @xmath116 .
particles were laid out on concentric rings and an equilibrium state accounting for stellar and disk self gravity as well as pressure forces defined the velocities of each particle .
the gas was evolved under the forces of stellar gravity , self gravity and gas pressure , which was computed using an isothermal equation of state ( i.e. @xmath117 with fixed temperature as a function of radius ) .
because of the simple physical assumptions , the dimensions of the system were scalable . given a star of one solar mass and a disk radius of 50 au , one orbit at the outer edge of the disk requires approximately 353 yr . during the preparation of this manuscript , we determined that simulations evolved using initial conditions identical to those in @xcite were not possible because at high resolution the large pressure gradient at the disk s outer edge caused unphysical behavior in the system . in order to sidestep this problem , we modified the form of the surface density power law near the outer disk edge so that the discontinuity is spread over a larger radial range . in this work
, the surface density is distributed according to : @xmath118 where the factor , @xmath119 , is a linearly decreasing function near the disk boundary and is defined by : @xmath120 with this definition , the disk edge is smoothed in the region within a distance @xmath121 inward and outward of the nominal disk radius . in the simulations here , we define the smoothing parameter @xmath122 au . the other initial conditions and
physical assumptions remained the same . using the conditions defined above , we ran a set of four simulations with varying resolution using each of three treatments for the small scale interactions between particles , but otherwise identical .
the first treatment , with simulations labeled _ mod1-mod4 _ in table [ tab : sims ] , uses gravitational softening of the form defined by equation [ eq : mass - kernsoft ] and the standard kernel gradient defined derived directly from equation [ eq : sphkern ] . the second treatment , with simulations labeled _ mod1tc - mod4tc _ , uses the modified kernel derivative defined by equation [ eq : mod - kerderiv ] to determine mutual pressure forces between particles , again with the standard kernel softening . the third treatment , with simulations labeled _ mod1grv - mod4grv _ , uses the modified gravitational softening defined in equation [ eq : mhat-2d ] with the standard kernel gradient .
figures [ fig : dyn1-morph ] , [ fig : dyn1-morph - modk ] and [ fig : dyn1-morph - modsoft ] show these four realizations of the model , each at the termination of the simulation .
each model develops multiple armed spiral structures over most of its radial extent , as in the previous @xcite work .
the spiral structures are filamentary and change the details of their appearance as the simulation proceeds .
evolution after the formation of the spiral structure however , was strongly dependent on the resolution employed .
in evolution of the three lower resolution realizations of the _ mod _ series , spiral structures eventually fragmented into multiple clumps , as in the @xcite work .
the time at which clumps begin to form , measured from the beginning of the simulation , was later at higher resolution than at lower .
moreover , fewer clumps formed in the higher resolution realizations than in the lower , and in the highest resolution realization , no clumps were produced at all : the delay before clumps begin to form has become longer than the duration of the simulation ( 12 , or about 4200 yr ) .
similar statements are true of both the _ modtc _ and _ modgrv _ runs , though there are a number of important differences as well . of particular interest
is the fact that the structure seen in figures [ fig : dyn1-morph - modk ] and [ fig : dyn1-morph - modsoft ] is substantially smoother than in the corresponding panels of figure [ fig : dyn1-morph ] . the overall relative smoothness is also reflected in the number of clumps that form in each realization . where the _
simulation formed several clumps , simulations with the two modified treatments _
mod3tc / mod3grv _ did not .
although each of the corresponding lower resolution realizations did form clumps , the number that formed was smaller in each case and delayed in time relative to the standard kernel versions .
figures [ fig : dyn1-dens ] , [ fig : dyn1-dens - modk ] and [ fig : dyn1-dens - modsoft ] show azimuth averaged , radially binned surface density profiles of the same models as shown in figures [ fig : dyn1-morph ] , [ fig : dyn1-morph - modk ] and [ fig : dyn1-morph - modsoft ] , at the same times . also shown
are instantaneous maximum surface densities , defined as the maximum in each ring seen at the time shown .
the average in each bin is weighted by the number of particles in the bin , with the width of each bin set to 0.02 au . at small radii , accretion onto
the star has depleted the initial power law density distribution , in each case by a differing amount due to the differing magnitude of the dissipation derived from the artificial viscosity included in the simulation ( see @xcite for a discussion of this effect ) .
further out , a number of spikes are visible in the profiles , each corresponding to the radial location of a clump in the disk .
the lowest resolution realizations display a very large number of clumps , while successively higher resolution realizations produce fewer , or none .
both the number of clumps that formed and the radial extent over which they are found are smaller when we use the two modified kernel treatments , compared to the standard form .
though not possible to see in snap shots of a single time , an important difference between the evolution with the standard and either of the modified kernel treatments is that the cumulative maxima in the latter case remain far lower than the former and also do not appear to increase over time .
we believe that this near steady state behavior represents the correct evolution of the system , even if it were to be evolved further in time .
what is the origin of the differences in the morphology between each series and the others ?
we postulate that it lies in the treatment of the small scale interactions between particles , and specifically in force imbalances between pressure and gravity that may be present .
if so , we may expect distortions in the distribution of interparticle separations , relative to what may be expected and to what may be present with other realizations .
figure [ fig : neigh - hist ] shows histograms of the distances between all neighbor particles for each of the three realizations of the _ mod4 _ simulations , as well as a simulation with the same resolution and initial condition , but without self gravity .
each histogram bin is defined to be of width @xmath123 , and all neighbors for all particles in each simulation are accounted for in the histograms . on the scale of the smoothing length , we can approximate the surface density as constant . for a uniform density distribution in a 2d simulation ,
realized with a random distribution of particles , we expect that the number of neighbors to be found at a given separation to increase linearly with the separation itself . at large separations ,
we recover the linear behavior . at smaller separations ,
all four distributions depart from the linear proportionality because the assumption underlying the linear proportionality neglects the influence of interparticle forces . due to the different treatments of the small scale interactions
, some differences also appear in the neighbor distributions .
the most striking feature of the plot however is the behavior near @xmath93 . in the standard case ,
a large number of particles actually coincide in space , while in each of the other three variants , none do .
this coincidence represents an extreme example of herant s pairing instability discussed above , that develops due to the interparticle pressure vs. gravitational force imbalance at small separations .
the pairing instability seen here is clearly not identical to that discussed by herant however , since no pairing is present in the non self gravitating realization for which conditions are most similar to those discussed by herant .
the pairing is also not similar to the numerical instabilities seen by @xcite for the same reason .
figure [ fig : neig - pairs ] shows that the number of paired particles continues to grow as the simulation proceeds .
after 12 , more than 65000 pairs have formed from @xmath124 particles : of the @xmath125 original particles , about half have become paired .
the existence of paired particles means first that the effective resolution of the simulation decreases with time as more and more particles become paired , and is consistent with our qualitative observation above that the cumulative maxima in these simulations also increased .
their presence also means that even though the force contribution from a single neighbor particle is small relative to the individual contributions from the rest of the system , when the particles approach coincidence the pairwise contribution can have an important and large scale influence on the behavior of the simulation .
based on these results , and the impact they have on the large scale disk morphology and its tendency to fragment , we conclude that using the standard kernel based gravitational softening in combination with the standard kernel derivative is likely to produce results contaminated with numerical artifacts in 2d simulations , and should not be used .
for each of the three series of simulations , the propensity of simulations to fragment decreased as resolution increased , until at sufficiently high resolution no fragmentation occurred at all .
this behavior reflects that seen by @xcite , where fragmentation was enhanced when resolution was insufficient , rather than that seen by bb97 , where fragmentation was delayed .
this is fortunate because it allows us to use the change in behavior of realizations of the same initial condition but differing resolution to determine empirically the approximate resolution ( in number of particles ) necessary to obtain ` correct ' evolution .
specifically , we can apply the criterion to two simulations which straddle the boundary between those that produced fragments and those that did not .
we can then note the value of at which the criterion succeeds for the entire simulation at the higher resolution , but fails for the lower resolution version .
bb97 scaled the value of the jeans resolution criterion through the value of as a multiple of the average number of neighbors , .
it will be convenient to scale the two dimensional criterion similarly here although in both cases , the correct measure will be a quantity independent of the specific neighbor count .
we also note that in two dimensional sph simulations , it is usual to use a smaller number of neighbors for each particle due to the lower dimensionality . in this work ,
we have used a number ranging between 10 and 30 , depending on the local flow ( see * ? ? ? * for details ) .
we therefore scale by a factor @xmath126 for these simulations .
the cumulative maximum surface density reached at each point in the disk , binned as a function of radius in the _ mod3 _ and _ mod4 _ simulations . for reference , the azimuth averaged surface density shown in figure [ fig : dyn1-dens ] as well ( lower curve ) .
the three smooth curves show the condition from equation [ eq : max - surfdens ] with a value of of 20 , 120 and 240 particles ( top to bottom in each panel).,width=302 ] figure [ fig : dyn1-crit ] shows the cumulative maximum surface density binned as a function of radius , along with the maximum resolvable surface densities determined from equation [ eq : max - surfdens ] using three values of equal to 1 , 6 , and 12 times the average number of neighbors for sph particles evolved in 2d ( i.e. 20 ) .
the cumulative maximum for each given bin is defined as the maximum surface density achieved by any particle in that bin over the entire course of the simulation up until the time shown . for the _ mod3 _
realization , the cumulative maximum exceeds the critical value using the @xmath12712 criterion for all radii inside @xmath128 au . at the same time , the @xmath1276 criterion is not violated except at a few localized radii .
interestingly , a small density spike relatively early in the simulation near 7 au did not lead to clump formation , but later interactions near 10 - 15 au did . for the _ mod4 _ simulation , the @xmath129 condition is satisfied over the entire radial range for the life of the simulation , and no clumps formed .
we can make only a tentative assignment of the value of from this result however because of the importance particle pairing may have on the effective resolution .
figure [ fig : dyn1-crit - tcgrv ] shows the cumulative maximum surface densities for two of the simulations using the tc92 kernel derivative and the modified gravitational softening . in this case we plot the curves for _ mod2tc / mod3tc _ and _ mod2grv / mod3grv _ rather than for _ mod3 _ and _ mod4 _ in order to maintain the straddle of the fragmenting / non - fragmenting outcomes in these simulations . in both variants ,
the cumulative maximum densities in the _ mod2 _ realizations exceed the critical value using the both the @xmath1276 and @xmath12712 criteria , while the _ mod3 _ realizations obey the 6 criterion but violate the 12 criterion .
since no clumps formed during the evolution , the latter must be considered too conservative , and we conclude that the resolution required to avoid fragmentation due to unphysical growth of self gravitating structures in the disk is @xmath1276 . at first sight ,
the required resolution for the toomre condition appears much larger in comparison to that for the jeans condition analysis discussed by bb97 .
the apparent paradox is resolved if we note , as in section [ sec : consistent ] , that their assumed value of the jeans mass was quite low and resulted in a relatively lower neighbor requirement .
our definition of the toomre mass corresponds to an analogue of the larger definition of the jeans mass defined as in equation [ eq : sphere - wavel - mjeans ] .
the origin of the large required value of becomes clear when we observe that the values of the cumulative maxima ( or indeed , also the slightly lower instantaneous maxima not shown ) are typically a factor of several higher than the averages .
most of the difference can be accounted for by the existence of spiral structure in the disk , however fluctuations due to the effect of the exact , time varying positions of particles relative to each other on the calculation of the density make up a smaller , but still significant component of the difference .
such fluctuations are intrinsic to the sph method itself and , while a small decrease is observable between the low and high resolution simulations in figure [ fig : dyn1-crit ] , their existence will be a part of all sph simulations .
it is interesting to note that the densities can exceed the resolvable maximum in the lower resolution realizations for some time before clump formation begins .
further , the amount of time before the onset of clump formation is a resolution dependent quantity , with higher resolution leading to more time before clump formation .
it is therefore clear that resolution studies are particularly important for particle simulations showing evidence of clump formation . applying this statement to all three variants of our own _
series of simulations , it is clear that if clumps do indeed form in disks similar to those studied , the process requires a time scale than longer the 12 ( @xmath1304200 yr ) for which we have evolved the system . in the two lowest resolution realizations of all three variants ( i.e. _ mod1/tc / grv _ and _ mod2/tc / grv _ )
, the numerical stability criterion is violated even in the initial condition .
since these simulations are only slight modifications from those presented in @xcite and at the same resolution , the same statement applies to those simulations as well .
regarding the clump formation seen in the simulations from that work , we originally concluded that although the physical model ( in particular the isothermal equation of state ) was insufficient to correctly model clump formation , if it was to occur at all , it would be most likely in between about 10 and 40 au .
the conclusion that the physical model in @xcite may be insufficient may indeed remain valid ( see e.g. * ? ? ? * ) ) .
however , the location and existence of clump formation is definitely invalid : it was due purely to failure of the toomre criterion and not to any physical process , however modeled .
after investigating the effects of resolution and choice of kernel on disk simulations , in this section , we turn to an investigation of gravitational softening .
we first discuss factors that motivate the selection of the value of a fixed softening length , then perform a side by side comparison of otherwise identical simulations employing either fixed or variable softening . due to our conclusion in the last section that the standard softening / smoothing combination should not be used , we limit our investigations here to
the tc92 and modified gravitational softening variants .
unlike the case of variable softening , where the softening length is set locally and dynamically , with fixed softening the modeler must make a specific choice of the length that is appropriate for the entire simulation at all times .
choosing an appropriate value in disk models is substantially less trivial in the case of disk models than in the cloud collapse scenario discussed by bb97 because , simply as a consequence of the initial conditions ( i.e. the density gradient as a function of distance from the central object ) , the smoothing lengths of particles are relatively steep functions of position .
this is important because for any constant choice of softening , the relative magnitudes of softening to smoothing will also vary with position , perhaps artificially suppressing fragmentation in one region while enhancing it in another .
moreover , because the toomre wavelength will itself be a function of radius through both density and temperature , both lengths will vary relative to it as well .
naively , one might expect that variable softening could lead to results that are quite susceptible to small scale fragmentation .
for example , if a region begins to collapse and the particles representing it approach each other , their mutual gravitational attraction continues to increase as their softening and smoothing lengths decrease , perhaps instigating the very problem softening is meant to avoid .
it is not clear that such a condition exists in practice however .
variable softening was implemented for both of our _ mod _ series of simulations discussed in section [ sec:2dtest - evo ] , as well as our previous work in @xcite .
these models did not produce clumps , except as a consequence of insufficient resolution . on the other hand ,
fixed softening has been strongly advocated by @xcite and both their lower and higher resolution simulations do produce clumps .
there are a number of questions that we must answer in order to understand the implications of the softening choice and magnitude on the outcome of a simulation .
first , since we have no _ a priori _ knowledge of where in the disk we may expect fragmentation to occur , if it is to occur at all , we must decide how to choose a value for the gravitational softening that allows the best reproduction of the real system .
given such a choice , what is the difference between the influence that fixed softening has on the simulation compared to variable softening ?
for example , to what extent can one or the other choice suppress fragmentation in models where it should not occur , but which may be under resolved ?
to what extent are physically valid evolutionary signatures also suppressed ? can one or the other choice actually instigate fragmentation in models that are otherwise stable ? can some values of fixed softening length both artificially suppress and enhance fragmentation in different parts of the same disk ? in this section ,
we investigate the influence of particular fixed values of the softening with a series of simulations ( denoted the _ fix tc _ and _ fix grv _ series in table [ tab : sims ] ) that are identical to the _ mod2tc _ and _ mod2grv _ simulations which we have shown to be unstable to numerically induced fragmentation with variable softening .
these simulations implement a fixed softening using the same sph spline kernel as is used for the hydrodynamic evolution , but are run , in the first case , with the modified kernel derivative for the pressure force calculation or , in the second case , with the modified 2d softening .
each of the four pairs of simulations uses a different softening length corresponding to the magnitude of the initial smoothing length at different locations in the disk .
these are , respectively , the inner disk edge ( 0.5au ) , the orbit radius corresponding to the region most susceptible to clump formation ( 15au ) , and near the middle of the radial span of the disk ( 30au ) , and its outer edge ( 50au ) .
we show the mophologies of these simulations in figures [ fig : fix - morph - modk ] and [ fig : fix - morph - grv ] , and the azimuth averaged surface densities in figures [ fig : fix - crit - modk ] and [ fig : fix - crit - grv ] .
overall , the behavior and morphology observed for each variant is quite similar to that of the other
. with small softening ( _ fix1tc / fix1grv _ ) fragmentation occurs at essentially all orbit radii , on the same time scale as required for the disk itself to become active . due to this rapid onset , we terminated it very soon after it began , so that radially more distant parts of the disk simply had insufficient time to become active .
we believe that fragmentation would occur there too if the simulation were to be evolved further . only at the inner disk edge , where the softening and smoothing lengths become comparable ,
is fragmentation suppressed . at all other locations , it is actually enhanced relative to the case of variable length softening shown in the _ mod2tc / mod2grv _ simulations above , where clumps only occurred at much larger radii .
the _ fix2tc
/ fix2grv _ pair of simulations were evolved for longer in time than the corresponding _
mod2tc _ version , but ultimately they too produced clumps .
fragmentation occurred only in the region exterior to @xmath131 au , where softening was smaller than smoothing . in the _
fix3tc / grv _ pair , evolved with a still larger softening , strong spiral structures developed in the outer half of the disk , while comparatively weak spiral structure developed closer to the star .
significantly , the disk was evolved for 5 and did not produce any clumps over that time . for simulations _ fix4tc
/ fix4grv _ , with the largest softening , spiral structure is largely suppressed throughout the entire disk , relative to their smaller softening cousins and to _ mod2tc_. the gross structure of the spiral arms that form also appear somewhat weaker than those in the _ mod2tc / mod2grv _ simulations , however in the absence of a quantitative analysis of the pattern amplitudes ( beyond the scope of this work ) this statement remains somewhat subjective . from these simulations
, we can make the rather unsurprising conclusion that with a large enough softening value , clumping can be suppressed in simulations where it would otherwise occur due only to insufficient resolution . while not surprising
, it is still important to quantify the both the value of ` large enough ' and what changes in behavior occur as a function of gravitational softening , in order to be able to separate out real and artificial effects . in this case , ` large enough ' gravitational softening appears to be larger than the hydrodynamic smoothing lengths in each part of the disk . in order to completely suppress clumping , softening must therefore be set comparable to the smoothing values in the outer part of the disk .
while we have seen that large enough softening can alter the behavior of an already numerically suspect ( under resolved ) simulation , it is also important to determine its effect on what might otherwise be considered a well resolved simulation . to investigate this question ,
we have run four simulations , denoted _ m1fxtc _ , _ m4fxtc _ , _
m1fxgrv _ and _ m4fxgrv _ in table [ tab : sims ] , with identical initial conditions and resolution to our _ mod4tc / mod4grv _ simulations , but with fixed gravitational softening . as in the last section
, the value of the softening is set approximately to the value of the hydrodynamic smoothing length at a predetermined radius in the disk .
m1fxtc / grv _ pair of simulations , we set the softening equal to the smoothing at the inner edge of the disk ( 0.5 au ) , so that they correspond to higher resolution versions of the _ fix1tc / grv _ pair .
m4fxtc / grv _ pair , we set the softening equal to the smoothing at @xmath132 au .
these simulations also correspond to high resolution versions of the _
fix2tc / grv _ simulations .
figure [ fig : stable - fixcrit ] shows the azimuth averaged and cumulative maximum surface densities for the two tc92 simulations .
in contrast to the _ mod4tc / grv _ realizations , the two _ m1fx _ realizations produce fragments in the disk in much less than 1 . as for their lower resolution cousins , _
fix1tc / fix1grv _ , we terminated the simulations quite soon after they began due to the fragmentation .
we expect that had evolution proceeded further in time , fragmentation would have occurred in the outermost parts of the disk as well .
consistent with our finding in section [ sec:2dtest - evo ] , that higher resolution realizations produced fragments later than lower resolution realizations of the same initial condition , fragmentation occurred later than in the _ fix1 _ variants .
fragmentation was delayed longer in the tc92 variant and was suppressed over a slightly larger portion of the inner disk . in both cases ,
fragmentation only occurred in regions where the fixed gravitational softening was smaller than the hydrodynamic smoothing , consistent with the expectation from the discussion in section [ sec : softsmooth - limits ] and in bb97 for 3d simulations .
even though the relative size of the softening and smoothing lengths were the same in the low and high resolution realizations , clumps appeared only at somewhat larger orbit radii than at lower resolution . in all cases ,
the maximum surface densities in and near the clumps exceed the critical value determined from equation [ eq : max - surfdens ] and must therefore be considered to be numerically induced rather than due to physical processes .
results in the _
m4fxtc / m4fxgrv _ simulations , with a much larger softening length , were quite different from each other : fragmentation did not occur with the modified kernel , but did with the modified softening , again in the outer part of the disk where interparticle force imbalances favored gravity over pressure .
although we have not shown the earlier evolutionary history of the _ m4fxgrv _ simulation , we note in passing that its character was markedly different than any other discussed in this work .
high amplitude , filamentary spiral structures developed in the outer disk as early as @xmath133 , much earlier than fragmentation occurred as defined by a violation of the toomre criterion .
although the structures developed with some regularity , they were unable to remain distinct , instead becoming progressively more sheared out , as they disappeared into the background flow . based on this behavior , we conclude that the softening length chosen for this simulation was very near the boundary between enhancing and suppressing fragmentation , as we saw for the _ fix2grv
/ fix3grv _ simulations at lower resolution .
if , as we have suggested , small scale interparticle force imbalances enhance fragmentation when the imbalance favors gravity , then we should expect simulations with fragmentation to contain a significant number of both paired particles and particles at small mutual separations , but few or none in simulations that do not fragment .
figure [ fig : m1pairs ] demonstrates that indeed , this supposition is true for the two smallest softening length variants .
the number of paired particles rises immediately and dramatically from zero to over @xmath134 over the short span of each simulation .
further , while the number of pairs appears to level out or even decrease at late times , the feature is actually a consequences of the increasing prevalence of higher order multiple particle groupings , such as triples or quadruples , as demonstrated by the lower pair of curves in the figure , representing the population of triples .
histograms of the neighbor separations for all neighbors of all particles in the _ m4fxtc _ ( top ) and _ m4fxgrv _ ( bottom ) simulations .
, width=321 ] figure [ fig : m4neigh ] shows histograms of the full neighbor distributions of the two larger softening variants and figure [ fig : m4pairs ] shows the number of pairs as a function of time for the _ m4fxgrv _ simulation . as we suppose , a significant population of pairs and near pairs ( i.e. with @xmath135 ) develops in the modified softening variant , which fragmented , while few do in the tc92 variant .
over the course of 12 , only a tiny population of near pairs developed compared to its cousin , and we observed only @xmath136 particles to become paired , a small enough number to be unimportant overall . while force imbalances must have been present at small separations in both models , we conclude that the enhanced interparticle pressure forces in the tc92 model , coupled with the higher resolution , were of sufficient magnitude to eliminate the numerically induced fragmentation that occurred when the softening length was smaller . in an apparent contradiction of our supposition , the total number of paired particles generated in the three simulations here , which fragmented , is smaller than in the _
mod4 _ simulation , which did not fragment .
although we have made no quantitative comparison , a brief inspection indicates that the contradiction may be resolved by accounting for the spatial distribution of pairs and of near pairs .
because the small softening simulations were not evolved for a long enough time to become fully active , the distribution of pairs is quite concentrated in the regions did become active and did fragment , relative to the distribution in the _ mod4 _ realization where pairs are found spread much more evenly throughout the system .
the _ m4fxgrv _ simulation , which did become fully active , generated many fewer pairs but did generate a large population of near pairs , indicating again the consequences of force imbalances favoring gravity .
given that the effective resolution in the _ mod4 _ run is decreasing with time , we expect that if it were to be evolved further , and the concentration of paired particles increased further , fragmentation would occur .
disks realized with small fix softening fragmented in both the tc92 and the modified softening variants , and both in their low resolution realizations and at resolution high enough to satisfy the toomre condition defined in section [ sec : test - crit ] .
an immediate conclusion from the similarities might be that , since the fragmentation occurred each realization , the simulations are in fact converged . while this conclusion may in fact be correct , it would also be seriously misleading . due to flaws in their design ,
the simulations converge to an incorrect result .
interpretation of that result in terms of the physical behavior of the system and the importance of other correctly implemented physical processes becomes difficult or impossible .
while the result and its invalidity may be clear in cases like our test simulations , the same statement may not be true in other simulations where the results are more difficult to verify . as a trivial example , the high and low resolution variants with larger fixed softening for which both low resolution variants fragmented , but only one at higher resolution .
our conclusion in this case must be limited only to the fact that the tc92 kernel modification is more likely to suppress fragmentation than the softening modification .
two other , more technical aspects of simulations with fixed gravitational softening deserve mention .
first , the cpu time required to run simulations with large , fixed softening to some specified time is far longer than with either the variable softening case or the cases with small , fixed softening .
the reason is that with a large softening length , the interparticle spacing requires that a much larger fraction of the total number of gravitational interactions between particles be calculated as ` atoms ' rather than grouped together as nodes , as is ordinarily done in tree - based gravity solvers in general use for particle simulations .
secondly , in cases where fragmentation does occur , a simulation with large fixed softening can proceed for much longer in time , because the time step size does not decrease nearly as much as occurs in a variable softening simulation at the same resolution . continuing a simulation for a long time
after clumping occurred and the resolution criterion has been violated would be of limited utility however , since the clump may cause large perturbations to the rest of the system that would not occur otherwise .
we conclude from our models that a fixed softening may either enhance or suppress density inhomogeneities depending on the specific choice of softening value relative to the hydrodynamic smoothing .
no single value of softening may be considered ` optimal ' in a disk with a large radial extent . in the extreme case ,
an incorrect ( too small ) softening value may induce fragmentation to occur in a simulation that would not , given another fixed choice or a variable softening set to the hydrodynamic smoothing length .
given an incorrect but too large value , suppression of physically correct structure may occur .
in contrast , while a variable gravitational softening length may produce some violation of energy conservation , in the examples shown here the outcomes of the simulations are not affected in a manner as drastically as with fixed softening .
we therefore recommend that simulations involving self gravitating hydrodynamic systems incorporate a spatially and temporally variable gravitational softening , whose length scale is the same as that over which the hydrodynamic quantities are smoothed .
our recommendation is similar to , but stronger than that made by bb97 because we observe that fragmentation may develop in simulations whether or not the toomre criterion ( equation [ eq : toom - cond ] or [ eq : max - surfdens ] ) is satisfied . rather than signaling a temporal endpoint , beyond which bb97 conclude that evolution towards some final collapsed state is under resolved , we find that force imbalances at small particle separations can alter the evolutionary trajectory of a system which would otherwise never approach gravitational instability into one that does .
given that bb97 did not specifically study systems starting in a dynamically stable equilibrium condition for which fragmentation is known _ not _ to occur , as found in our disks , the behavior we observed in our simulations may not have been observable in their work .
we suggest that it is independent of the specific system being simulated , and in fact applies to the jeans collapse situation as well .
our conclusion is also consistent with that of @xcite , who made a detailed study of the effects of various choices of softening on the force errors produced by a static distribution of particles , and who recommends spatially variable softening in that case as well . since in nearly all particle based hydrodynamic simulations , however , the smoothing length is not fixed in time , our recommendation goes beyond dehnen s study of a static distribution to including temporal variation as well .
we base our recommendation on the results of our simulations and on the following argument .
completely independent of whether or not either the gravitational or pressure force actually can be resolved on better or worse length scales than the other , is the consequence of the assumption underlying both softening and smoothing .
namely , an error in the net force is present _ by assumption _ when the contribution from either source is not fully resolved .
we therefore have only the choice of how to handle this error most gracefully in the code .
we submit that avoidance of large force imbalances in regions that are by definition under resolved is a trait much to be desired in the properties of the code because pathological behavior that results from such imbalances artificial fragmentation or suppression of physically realistic fragmentation are avoided .
such a recommendation is not without cost however because it is equivalent to recommending that simulations using sph should not require that energy be conserved .
although we have not attempted to quantify the magnitude of the violation , in general we believe it to be a relatively small component of the total energy of a given particle since only neighbor particles will contribute .
also , the violation will be smaller in higher resolution simulations because those neighbors will represent a proportionally smaller component of the total mass in the system , from which the potential energy of a given particle is derived .
finally , the fact that the gravitational forces and potential are approximated by a summation of terms derived from a tree search already violates strict energy conservation .
given the large body of literature using such approximations , often checked for veracity against other methods or observations , it would appear that few , if any , unacceptable consequences originate from a low level of violation .
assertions such as these are no replacement for a measurement however , and we refer the reader to @xcite for discussion of a method to make an approximate quantification .
we are also aware of current work , developing on the techniques discussed in @xcite and by the same authors ( price and monaghan 2006 , submitted ) , to create an sph formulation that allows variable softening , while still conserving energy .
adapting sph codes in existence to utilize this technique would appear to be highly desirable , since it would remove the most serious drawback of variable softening .
we have considered and discarded a third , intermediate softening option , in which the gravitational softening of each particle is set , for example , to the _ initial _ value of the hydrodynamic smoothing length , and is thereafter fixed .
although no specific tests have been performed to study this question , we believe that using such a prescription will yield results similar to the fixed softening case with a small softening length
. for example , a particle initially located in a dense region ( small softening value ) , which later moves into a less dense region ( larger softening value ) could act as an attractor for other particles in its current neighborhood .
an exactly opposite effect might be seen for a particle initially in a low density region which moves to a high density region .
secondarily , we question the aesthetic utility of such a prescription because it implies knowledge of the internal mass distribution of each particle that is both independent of any hydrodynamic expansion or contraction of the gas ( the assumption underlying a temporally variable softening ) and needlessly complex .
the complexity is due to the fact that using constant , individual softening lengths suggests that information is known about not only the internal mass distribution of particles taken as an ensemble , but also the information that each particle has a different internal mass distribution from every other .
we can postulate no circumstances where such exquisitely detailed information is likely to be available .
the wave analysis underlying our resolution criterion makes the assumption that the flow is two dimensional : an integration over the @xmath65 coordinate is assumed .
the test problem defined in section [ sec : testing ] was chosen to be similar to the models originally employed by @xcite , and was therefore performed in the same two dimensional approximation as the originals .
this approximation has historically been common in other works as well . in this section ,
we examine methods for applying the criterion to fully three dimensional models , and to the question of whether 3d simulations require additional constraints on resolution in order to ensure their accuracy . for fully three dimensional simulations , where a surface density is not directly available
, the toomre criterion can not be applied without some additional computation . for grid based simulations ,
the computations are trivial and involve only an integral of the volume density in the @xmath65 direction . in cylindrical or cartesian coordinates , for example the integral would pass to a summation of @xmath137 over a set of grid zones with spacing @xmath138 , with @xmath139 being the volume density in the @xmath140th grid zone in the @xmath65 direction . for 3d sph simulations ,
the computation of a surface density is more complex because there is no grid .
instead , we recommend the creation of a temporary grid onto which the volume density can be mapped . then the surface density can be computed as a summation in the @xmath65 direction as in the case for grid based simulations . in test models
, we have found good results by using a mapping to the grid based on the same smoothing technique used to determine the volume densities at the particle locations , for the mapping . at the location of each grid zone
, we determine a list of ` neighbor ' particles which may contribute to the density of that zone , then we use the formula @xmath141 where @xmath19 is the mass of each neighbor particle , @xmath142 is the smoothing kernel , @xmath143 is the distance between the particle and the grid zone at location @xmath144 and @xmath88 is the mutual smoothing length of the particle and zone , defined as @xmath145 .
the grid smoothing length is defined as @xmath146 where the components are the grid spacing in each of @xmath147 , @xmath148 and @xmath65 .
we include a smoothing length for the grid itself in order to ensure that every particle contributes to at least one grid zone even in the case of vanishingly small smoothing lengths . for the test problem discussed in sections [ sec:3dsph - nosg - test ] , [ sec:3dsph - nosg - result ] and [ sec:3dsph - nosg - signif ] below ,
we map the volume density onto a cylindrical coordinates grid where the radial coordinate is logarithmically spaced and the azimuth coordinate is uniformly spaced so that its spacing is @xmath149 everywhere . in order to resolve the smallest scale features of the flow
, we have chosen our grid resolution so that its spacing is comparable to the smallest smoothing lengths in the calculation .
this grid spacing will in practice over resolve the features of most of the flow , and will not strictly be required in most circumstances .
our choice is motivated by the desire to ensure that the choice of grid ( i.e. @xmath150 ) will contribute negligibly to the derived densities and the quality of the fitted quantities in our analysis that depend on them . in any case , the computational expense of choosing of a finely spaced grid for the mapping is small in terms of the time to run a simulation because the mapping is not required for the simulation itself , but only for post - processing .
a conceptually simpler mapping would be to use a particle in cell technique to assign the mass of each particle to the grid cell that contains it , thus determining a volume density for that cell .
unfortunately , we can not recommend this method because it is susceptible to large errors due to the fact that sph particles may have smoothing lengths much different than the grid spacing .
for example ,
when a particle s smoothing length is larger than the grid spacing , it should contribute to the density over a larger region than assumed by the pic technique .
when two dimensional simulations are performed , it is common to obtain an approximation to the volume density by dividing from the disk s surface density by some factor multiplying the disk s scale height . for an non - self gravitating isothermal gas , the factor will be @xmath152 ( see section [ sec:3dsph - nosg - test ] below ) , but will vary slightly depending on the exact input physics . for purposes of this section and for simplicity , we shall assume a factor of two exactly .
as discussed at the end of section [ sec : applicability ] , the jeans or toomre stability criteria may be applied using this conversion to obtain one or the other of surface or volume density as required . in this section
, we will investigate the impact that this conversion may have on the indirectly obtained values used in the jeans criterion .
we will refer here to the criteria directly applied through the use of equations [ eq : jeans - wavel ] or [ eq : toom - wavel ] as the jeans or toomre criteria , while the criteria applied through the use of the disk scale height to convert surface to volume density ( or vice versa ) as the approximate jeans or toomre criteria . in order to proceed ,
it is necessary to examine the results of a simulation modeled fully in 3d , so that both volume density and surface density are known . in one of a series of papers studying fragmentation in disks
, @xcite discusses 3d disk models and @xcite continues this study with a close examination of the properties of a few models ( primarily the model named ` edh ' ) from the earlier work .
we will use these models as a test bed for our investigation .
since our study has not specifically defined a value for @xmath37 in equation [ eq : toom - cond ] , we will assume that its value is @xmath153 , which is identical to the value of @xmath66 required for jeans collapse simulations . while we believe this to be a reliable assumption , only by examining the results of a specifically chosen test problem
can we be fully confident that the value is sufficient .
the consequences of the uncertainty in specifying @xmath37 are that comparisons between the magnitudes of @xmath37 and @xmath66 and the critical lengths associated with them , may be more difficult to interpret . in this section
however , we seek primarily to test the veracity of the approximate forms of the criteria compared to the exact forms , and the absolute magnitudes are less important .
boss s model consists of a circumstellar disk modeled between 4 and 20 au , evolved on a grid of @xmath154 ( or @xmath155 in some simulations ) zones asymmetrically distributed primarily near the disk midplane .
he includes an ideal gas equation of state and radiative cooling in the diffusion approximation , but omits an artificial viscosity , thereby omitting the viscous and shock heating modeled by it . with this model
, he evolves a disk for 345 yr , at which time the simulation begins to violate the jeans criterion due to onset of clump formation , and is terminated .
we will apply our stability criteria to this model using each of the four relevant versions discussed in section [ sec : sizes ] .
the top panel of figure [ fig : boss - crit ] shows the value of the critical wavelengths at the time the simulation was terminated , using each of the four applicable implementations of the stability criteria .
the failure of the jeans criterion was used to define the end of the simulation , so its minimum value matches the grid spacing very closely .
the exact toomre criterion yields a critical wavelength that is about 30% smaller than that from the jeans analysis . using it and assuming that @xmath156 , the model has already passed the boundary beyond which it becomes numerically suspect . due to the speed at which clumping occurs ,
once underway , we do not believe that this small difference is very significant however , except to note that the simulation would have been terminated slightly earlier and at a lower density . at all other locations , the toomre wavelength exceeds the jeans wavelength and the evolution of the boss simulation leading up to its termination passes the resolution test .
the conclusions made from it stand on valid numerical grounds according to the toomre resolution criterion . on the other hand , the approximations of the jeans ( light solid ) and toomre ( light dashed ) criteria lead to critical wavelengths in the forming clump that differ by a factor of 1.56 larger than and 4.6 smaller than the grid spacing , for the approximate jeans and approximate toomre criteria , respectively .
moreover , using the approximate toomre criterion , the simulation is already well past the time for which its evolution is numerically valid .
each of the approximate forms uses the isothermal scale height to render the conversion .
does this quantity remain relevant in a collapsing region ?
the bottom panel of figure [ fig : boss - crit ] shows the ratios of the jeans and toomre wavelengths to the scale height .
except near the forming clump , the jeans wavelength is many times the disk scale height , making its application as a resolution criterion unclear according to the argument in section [ sec : sizes ] .
the toomre wavelength is also many times the disk scale height , but this only makes its application to a resolution criterion more appropriate , according to the same arguments .
we therefore recommend that it be used in favor of the jeans criterion in simulations of disks . in and near the collapsing region ,
both wavelengths become significantly smaller than the scale height . since those wavelengths loosely define the physical size of the collapsing body
, we can conclude that the structure of that object is no longer well represented as a slightly perturbed disk structure under which assumption the scale height was originally defined .
thus , the approximate criteria should not be used because their values become inaccurate near the clump , due to the fact that the disk vertical profile becomes distorted . in section [ sec:2d - vs-3d ] , we discussed the possibility that a disk simulation performed in 3d may not in fact be truly three dimensional because the disk thickness may be comparable to the smoothing lengths of the particles in and sph simulation , or the grid dimension in a grid based simulation . here
we explore the consequences of such simulations using a simple test problem .
specifically , the problem of a non - self gravitating disk evolved with an isothermal equation of state . for this system , an analytic expression for the vertical structure of the disk
can be derived exactly , and direct a comparison between the analytic result and theory can be made .
this test problem is equally applicable to grid based simulations and to particle simulations , and we propose it as a general test for interested numericists .
the vertical structure of the gas in a non - self gravitating disk will obey the equation @xmath157 where @xmath158 and @xmath11 are the pressure and volume density , @xmath65 is the altitude above the disk midplane and @xmath147 is the cylindrical radius . for an isothermal equation of state , the pressure
is related to the density and sound speed through the definition @xmath159 . using the equation of state to replace pressure with density , the solution to the differential equation is @xmath160 where @xmath62 defines the isothermal scale height and @xmath161 . for
any given location in the disk , the surface density can be obtained by integrating equation [ eq : vert - profile ] over all @xmath65 , giving the relation @xmath162 we use a modification of the 2d test problem described in section [ sec:2dtest - defn ] for this test .
particle layout , temperature and surface density are defined as in the 2d problem , such that a minimum toomre @xmath4 value of @xmath163 would be obtained if self gravity were included . to obtain an initial @xmath65
coordinate for each particle , we use a gaussian pseudo random number generator whose width is defined by the local value of the disk scale height , consistent with equation [ eq : vert - profile ] .
definition of the initial velocities takes account only of stellar gravity and pressure forces rather than disk self gravity as well .
velocities in the @xmath65 coordinate are set to zero . before proceeding further
, it is significant to note that the specific model we propose is a step removed from those of greatest relevance for actual systems because vertical density distributions depend on the details of the physical model employed .
also , as a practical matter , the exact definition of ` scale height ' itself can lose meaning in such models . to the extent that it does remain meaningful , any resolution requirement may be sensitive to the details of the model if , for example , a larger fraction of the mass were in an extended envelope or were concentrated closer to the midplane than with an isothermal structure , as is the case for the isentropic structure discussed in section [ sec:3dsph - rt ] below . in the context of the spatial distribution of particles or grid cells that represent the mass and at a coarse level of examination
, models that result in a higher midplane mass concentration will be somewhat similar in structure to colder isothermal disks in which the scale height is simply not as large .
we therefore believe that any modifications to the critical resolution appropriate for other models will be not be large and proceed to define a criterion for isothermal structure that we believe will be generally applicable . in the following two sections ,
we will however , proceed to define a resolution criterion in two complimentary ways : as a requirement per scale height at the disk midplane , and as a resolution per vertical column , in order to facilitate its use more generally .
we have run a series of simulations modeling the 3d test problem at different resolutions , from @xmath164 to 10@xmath165 particles , each labeled _ sgoff3 - 6 _ and defined in table [ tab : sims ] .
we allow each simulation to run for two orbits of the outer edge of the disk in order to ensure that the configuration represents the true system as realized by the hydrodynamic code rather than any peculiarity of the initial configuration .
we then map the volume density and velocities onto a three dimensional grid and calculate the surface density at each location .
we pose the following question for the configuration , which must be answered affirmatively before the results can be validated : does the configuration generated by the simulation correctly reproduce the correct midplane densities and scale heights everywhere in the disk ? to answer the question , we generate least squares fits to the densities as mapped onto a temporary grid to compare to the analytical values defined by equation [ eq : vert - profile ] . each fit returns three parameters corresponding to the values of @xmath166 and @xmath98 in equation [ eq : vert - profile ] .
we limit the fits to the radial region between 1 au and 45 au in order to eliminate the possibility that conditions at the inner or outer boundary of the disk distort the true picture of the structure .
inspection of the disk structure shows that these limits ensure that the fits are limited to the regions where the rotation curve is unaffected by either the gravitational softening of the star at the inner boundary or large density gradients near the outer boundary . in order to remove any potential systematic errors derived from the resolution of the grid or simulation itself , we average the volume densities at each altitude over patches of approximately one scale height on a side , so that the same number of fits are generated for each simulation .
finally , in order to retain self consistency in the face of any evolution of the disk away from the initial surface density profile , we derive the analytic values for @xmath166 and @xmath98 using the locally determined value of the surface density and rotation velocity , through the relation @xmath62 and equation [ eq:3dtest - surfdens ] .
patches are defined in successive radial rings of the disk by the condition that a ring s radial extent is the local , analytically determined value of the scale height .
their azimuthal extent is determined by the same condition , so that the total number of patches in a ring is defined by its circumference divided by the scale height . at the grid resolution chosen for the analysis , each patch covers a region of @xmath167 or more cells in the vertical coordinate , and @xmath168 radial and azimuthal grid cells near the inner disk edge where the scale heights are small , and @xmath169 near its outer edge where they are much larger . in total , approximately 14000 patches are required to cover entire disk , so that 14000 separate fits of the vertical structure in the disk are performed .
figure [ fig : bestfit - vstruc ] shows histograms of the ratios of the fitted quantities @xmath166 and @xmath98 to their analytic predictions , normalized to the total number of patches . at the lowest resolution shown , the vertical structure in the simulation is clearly more extended vertically than predicted , with the distribution of scale heights extending more than 50% larger than the analytic values and the midplane density distribution extending nearly 50% below the analytic values .
both distributions are spread relatively evenly over the range extending from the most extreme under or over estimates on one end , to nearly the correct analytic value on the other .
the distributions become narrower for each of the each of the higher resolution variants of the model , and appear in fact to be converging to the analytically expected values .
however , even at a resolution of one million particles , the highest resolution in our study , the peaks of the distributions deviate from the analytical values by about 5% and a significant population of patches are best fit with midplane densities and scale heights as much as 20% below or above the analytic values , respectively .
we can convert the correspondence between simulation and theory into a resolution requirement if we can quantify the minimum number of particles required in a vertical column for which the fit parameters are accurately reproduced .
the top panels of figure [ fig : hiterat - rhodist ] show the vertical resolution of the disks quantified as a ratio between the expected scale height and the smoothing length of each particle .
the bottom panels show the quality of the reproduction of the analytically determined disk midplane densities as in figure [ fig : bestfit - vstruc ] , broken out into separate histograms for each of four radial regions of the disk .
the vertical resolution varies between less than one smoothing length per scale height in the inner disk to as many as six near the outer edge of the high resolution simulation , _ sgoff6_. the ratio takes values over a range from less than unity up to a orbit radius dependent maximum value because particles located near the midplane at some radius , where densities are high , will have correspondingly smaller smoothing lengths .
particles located at higher altitudes where densities are lower will have larger smoothing lengths . dividing
each by the scale height ( constant at a given radius ) produces the range .
the maximum value of the ratio varies as a function of orbit radius primarily because the scale height itself is a function of radius , through the combination of the predefined temperature profile ( equation [ eq : dyn1-temp ] ) and the rotation curve . in the 13 au region of the disk , the entire vertical structure of the disk
is represented by the smoothing length of only a single particle .
the exact positions of neighboring particles relative to each other therefore cause a correspondingly larger influence on the densities , as demonstrated by the very wide distribution of midplane densities present in both the _ sgoff4 _ and the _ sgoff6 _ realizations , relative to their analytic values . in each successively more distant region , and at both resolutions , the distribution becomes narrower and of better quality , demonstrating the value of the increased vertical resolution .
for example , the vertical structure is resolved by @xmath170 smoothing lengths per scale height at @xmath171 au , and the fitted midplane densities fall between 65 and 95 percent of their analytic value for the 260000 particle simulation and 75 and 100 percent in the 1 million particle run . at the other end of the spectrum , density fits fall within @xmath136% of their predicted values only in the two outermost regions in the high resolution simulations , corresponding to orbit radii @xmath172 au ( i.e. the green and dark blue histograms in the figure ) , and only part of the distribution in the single outermost region of its lower resolution counterpart , corresponding to radii @xmath17327 au . if we specify that midplane densities within 5% of their analytic values are sufficiently accurate for the purpose of simulating the evolution of a circumstellar disk , then only these outer regions of the disk have sufficient accuracy . in the
two best resolved regions of simulation _ sgoff4 _ , ratios of fit to analytic midplane density values peak near @xmath174% and extend to as low as 90% , resolving one scale height with at most @xmath175 particles . in the same regions of _
sgoff6 _ , the peaks of the distributions increase to near @xmath176% , with tails extending to as low as @xmath177% , resolving one scale height with as few as 3.5 particles at 9 au and as many as 6 further out .
we conclude that midplane densities within @xmath136% of their analytic values are achieved only in regions where the vertical structure of the disk is resolved with at least @xmath5 smoothing lengths per scale height at the disk midplane .
it is important to note that the requirement for @xmath178 smoothing lengths per scale height near the disk midplane is sufficient only for obtaining reasonably accurate _
midplane _ densities , and only for a mass distribution described by equation [ eq : vert - profile ] .
quantities at higher altitudes may still be significantly under resolved because smoothing lengths are larger there due to the correspondingly lower densities .
errors due to insufficient resolution at high altitudes will be of lesser importance for simulations modeling gravitational instabilities because the magnitudes of the pressure and gravitational forces important for fragmentation will be largest where the mass is concentrated , close to the midplane .
nevertheless , some particles must be present at high altitudes in order to compress those lower down to the required densities .
how many are enough ? in this section we quantify the minimum number of particles required per vertical column needed to obtain good correspondence with theoretical expectations for density near the midplane in our models . because it is a metric that is effectively an integral over a full vertical column , this quantification is likely to be much less sensitive to the exact distribution of mass over the column , and so more generally applicable than a quantification per scale height .
it also eliminates the possibility of misperceptions based on the idea that the total mass outside the midplane is small enough to be neglected in estimates of the total resolution required per vertical column . for the isothermal disks in this study
, 98% of the disk mass will reside within three scale heights of the midplane , according to equation [ eq : vert - profile ] , meaning that the mass is effectively distributed over a total of six scale heights accounting for symmetry above and below the midplane . at the simplest level
, we might therefore approximate the required vertical resolution to be @xmath179 smoothing lengths per vertical column to avoid serious errors in the midplane densities .
a somewhat more accurate quantification might also account for the fact that mass is preferentially located near the midplane , with fewer particles present at high altitudes . in principle
, we can derive an approximate number of particles required per vertical column if we simply quantify the particle density over all altitudes in a single vertical column . at the minimal resolution required to obtain accurate midplane densities
however , we must presume that the high altitudes remain inaccurate due to the fact that densities are much lower there and particle separations larger , and they may therefore remain under resolved .
any direct quantification of the resolution required based on the actual distribution of particles in a simulation would therefore suffer from the inaccuracies in their high altitude distribution .
a fact that allows the analysis to proceed is that , for the purpose of ensuring accurate midplane densities , only the total weight of particles at high altitudes is needed rather than their distribution .
this is important because weight is a quantity integrated over the vertical column of material and will therefore take the same value whether or not the mass distribution responsible for it is well resolved . any high altitude mass distribution yielding the same weight will yield the same midplane density so that , given a correct midplane density , we may infer that the total weight is approximately correct as well .
we may therefore assume that the high altitude mass distribution derived from our simulations is correct ( even though it may not be ) , for the purpose of converting densities at a given altitude to a corresponding number of particles needed to supply the correct compressional forces to midplane material .
we can quantify the total particle count per vertical column in the simulations using a modification of the patch averaging strategy used above .
rather than creating full 3d density distributions for each patch , it is sufficient only to quantify the averaged number of particles located in each , and to create a coarse vertical distribution of the particles by assigning each particle to a histogram bin according to its altitude .
the result is the patch averaged volume density of particles in each histogram bin , from which a linear ( vertical ) particle density per scale height can easily be obtained by taking a cube root .
a final link is to note that interparticle separations are @xmath180 and that , by definition , the histogram bins are of width @xmath98 , so that the linear particle density is equivalent to the ratio of smoothing length to scale height .
the patch averaged vertical , linear particle density binned in histograms of width @xmath98 , for the _
sgoff4 _ and _ sgooff6 _ simulations . from top to bottom , the histograms in each panel correspond to the dark blue ( solid ) , green ( dotted ) , red ( short dashed ) and black ( long dashed ) regions as defined in figure [ fig : hiterat - rhodist ] .
the numerical values associated with each curve are sums over all histogram bins thus defining the averaged total particle counts per vertical column.,width=321 ] figure [ fig : num - vert ] shows histograms of the patch averaged ratios of @xmath181 defined using this procedure , separated into the same radial regions defined in section [ sec:3dsph - nosg - result ] .
consistent with expectations based on the physical model and on the fit results above , each histogram displays a maximum at the midplane and non - zero populations to as high as @xmath175 scale heights in each direction .
histograms with larger net populations exhibit a progressively more peaked structure , reflecting the better agreement with the analytical model provided by the higher resolution . for the present purposes
, it is interesting to note that in every region of the disk , the total resolution per vertical column is several times the resolution assuming that only the midplane contributes significantly to the total . in section [ sec:3dsph - nosg - result ]
, we determined that only the two outermost regions of the _ sgoff6 _ simulation were well resolved in our simulations . here
, we see that on average , the resolution of vertical columns in these regions is @xmath182 and @xmath183 particles per column for the 927 au and 2745 au regions , respectively . for the lower resolution _ sgoff4 _ simulation , the highest overall vertical resolution reaches @xmath184 particles per column , and all other regions in both simulations fall progressively farther below this value .
since fits in all of these regions deviate from the analytic expectation by more than 5% , we therefore conclude that a minimum of @xmath185 particles per vertical column are required to adequately reproduce the density structure at the disk midplane . in self
gravitating systems , the exact balance of pressure and gravitational forces will tip that system towards or away from two dramatically different outcomes : fragmentation or continued smooth evolution . in marginally stable systems of interest to modelers
, the exact balance between the two forces will be determined from terms of nearly equal magnitude , but of opposite sign .
if one of those quantities is incorrectly calculated , the outcome will be dramatically different .
the results of the test above demonstrate that simulations with too few particles tend to underestimate the midplane densities , in the case of 2.6@xmath186 particles , by as much as 30 - 35% .
this is important because the pressure will be underestimated by a similar factor through the equation of state .
fragmentation will therefore be enhanced in an under resolved simulation compared to that of either a well resolved simulation or , more importantly , a real , physical system . because we have not performed self gravitating simulations
, we can not specify the exact effects the errors will have in simulations or the level of enhanced fragmentation that may occur however .
such details are difficult to specify with precision because any enhancements may be mitigated in part by the fact that the cause of erroneously low densities in sph is simply that particles separations are greater than they should be , so that gravitational attraction between them is correspondingly less .
any possibility of mitigation of this sort provides little comfort relative to simply solving the hydrodynamic equations accurately in the first place however . for the disk morphology discussed here , resolution sufficient to accurately reproduce the disk s vertical structure and midplane density beyond @xmath187 au requires at least one million particles .
although we have not attempted to quantify the differences through simulations , colder disks closer to the fragmentation boundary will have correspondingly smaller scale heights , and will therefore require still more total particles to ensure adequate resolution . for example , the disk scale height and @xmath4 values both depend directly on the sound speed , so a change in one will be reflected proportionally in the other . a disk with minimum @xmath188 or @xmath189 will decrease the scale height below that in our simulations by a factor @xmath190 or @xmath191 respectively . in order to retain adequate resolution of the vertical structure
, smoothing lengths must be decreased by a similar factor by increasing resolution . in 3d
, the magnitude of the increase will be a factors of @xmath192 or @xmath193 corresponding to @xmath163 or @xmath194 million particles .
no simulations of circumstellar disks in the context of planet formation have yet been performed at such high resolution . even at a resolution of 1 million particles ,
the vertical structure in our simulations is not accurately reproduced inside 10 au , where midplane densities drop well below their analytic values .
this is important for simulations of gravitational fragmentation because the jovian planets in our own solar system formed at such radii .
the erroneously low densities may lead to enhanced fragmentation in those regions , even if the structure is accurately modeled further out . also , the resolved and unresolved parts of the disk are in no way isolated from each other . as waves or other structures
propagate through the disk , their subsequent evolution will be perturbed by the change in resolution , so that the evolution throughout the entire simulation become suspect .
both the small values of the scale height at small orbit radii and the insufficient resolution there are consequences of the @xmath195 temperature profile assumed for the disk , and commonly used throughout the literature . through the sound speed and rotation curve , the scale height will exhibit a proportionality @xmath196 and , for minimum @xmath41 , will take values of @xmath197 near 1 au increasing to @xmath198 near 45 au .
the temperature profiles in our simulations reflect the stage of active disk evolution during which gravitational instabilities are most likely to be present .
a temperature profile inversely proportional to radius will yield a flat profile and so avoid the issues of variable vertical resolution , but such steep profiles are not observed in real systems @xcite , and so we discount them here .
above , we showed that the vertical structure of an accretion disk must be resolved with some minimum number of particles if mass densities in the disk midplane are not to be substantially underestimated .
although this requirement may be sufficient to ensure accurate simulations in the case when comparatively simple physical models are employed , it will be less so when radiative transfer is included .
here we describe an exercise meant to illustrate requirements for resolution of the high altitude structure of the disk , if serious errors in the cooling rates derived from radiative emission from the disk photosphere are also to be avoided . at the orbit
radii most relevant for planet formation ( @xmath199 au ) , disks will be optically thick in the sense that the optical depth , @xmath200 , calculated from infinite distance to the disk midplane is large .
this is important because it means that the cooling rate of packet of disk material will be well modeled by a blackbody cooling law whose temperature is defined at the disk photosphere . if the altitude of the photosphere is known only approximately , due to low resolution or simple miscalculation , the cooling rate will be similarly approximate because the temperature structure near the photosphere is not known precisely .
comparatively small errors in temperature will lead to much larger errors in the local cooling rate because of the @xmath201 proportionality of the blackbody cooling function .
accurate knowledge of the cooling rate and its associated photosphere temperature is important because the dynamical balance of heating and cooling processes in disks is quite precarious : most of the heating and cooling sources are capable of removing or replacing essentially all of the disk s thermal energy over the course of only a few orbits @xcite . specifically for the question of disk fragmentation , @xcite shows that cooling rates faster than @xmath202 , where @xmath29 is the local orbit frequency , lead to disk fragmentation , but longer cooling rates may not in a local calculation .
later global calculations of @xcite confirmed this result for low mass disks , where they found that a cooling time of @xmath203 was long enough to prevent fragmentation , but a cooling time of @xmath204 was not .
simulations of higher mass disks with cooling rates of 10 and 5 @xmath205 showed similar changes in behavior . from these results
it is clear that an error in the cooling rate as large as a factor of two will be sufficient to suppress or enhance fragmentation in a simulation which would otherwise follow a very different evolutionary path .
a error of factor of two in the cooling rate will be equivalent to an error of @xmath128% in the photosphere temperature , due to the @xmath206 dependence of the cooling on the temperature .
since an error of such magnitude may seriously alter the balance between cooling and heating , a more restrictive limit must be set to ensure that any fragmentation that does occur is not subject to numerical error .
if we set an arbitrary standard that a cooling rate accurate to 25% will be sufficient to ensure results are not contaminated by numerical errors , then we will in turn require a photosphere temperature accurate to @xmath207% .
given the requirement as stated , it remains to convert the constraint on the photosphere temperature determination to a constraint on required spatial resolution of the vertical disk structure .
one approximate model of the vertical structure of a disk is available from the work of @xcite , in which an isentropic vertical structure is assumed , for a locally plane parallel , self gravitating disk .
figure [ fig : vert - struct ] shows temperature and optical depth profiles as a function of altitude for a region of a disk expected to be interesting for disk fragmentation , derived from the @xcite model .
several points are of interest in evaluating the figure .
first , although the density structure follows an approximately gaussian structure , at least in coarse outline , it is dramatically compressed relative to that expected for an isothermal disk .
no material is present above an altitude of @xmath208 , due to the different physical assumptions employed in deriving the mass distribution .
second , this configuration is optically thick , with an optical depth to the midplane of @xmath52 , meaning that a disk photosphere surface may in fact be defined . for the purposes of evaluating the radiative energy losses , a third point will be of most importance .
namely , that near the disk photosphere surface both the density and temperature are changing rapidly with altitude . a calculation that resolves these profiles too coarsely will therefore also only coarsely resolve the photosphere surface itself , possibly exposing the hot disk interior to space .
for example , if the particle density near the photosphere for an sph simulation is such that smoothing lengths are @xmath209 , then to a first approximation , the photosphere surface itself will be resolved on the same spatial scale because interparticle separations are themselves @xmath180 .
a particle actually located anywhere within @xmath88 of the actual photosphere surface may be determined to define the photosphere surface , simply because no other particle happened to be at a higher altitude there .
an identical argument holds for grid based simulations , with the size of a zone at the photosphere surface replacing smoothing length .
figure [ fig : vert - struct ] shows that uncertainties in the altitude of the photosphere surface of @xmath210 results in an increase or decrease of the derived photosphere temperature of @xmath132% .
doubling the linear resolution to @xmath211 decreases the uncertainties to @xmath212% , comparable to our requirement of errors no larger than @xmath213% .
we therefore conclude that simulations including radiative cooling must resolve the disk photosphere surface at a spatial scale no coarser than @xmath214 , when structure models similar to ours are employed . after completing the exercise and deriving this requirement
, we must also immediately point out that the quoted required resolution is not universally applicable without modification .
any constraint on spatial resolution in the context of radiation will be considerably complicated by the fact that the vertical temperature profile will be significantly influenced both by preferential dissipation of shocks at high altitudes @xcite , by radiative heating by external sources like the central star @xcite and , if any remains at the time of the simulation , the surrounding molecular cloud .
as @xcite also note , wavelength dependent opacities will also play a role , to the extent that long wavelength radiation is able to cool the disk interior effectively , even while shorter wavelength radiation remains trapped . in spite of these cautions ,
we remain convinced that the above exercise remains a valuable illustration .
it is based only on the condition that the transition between optically thick and optically thin material is well resolved and , although it is applied to a particular model of the vertical structure , it is not tied to it specifically . given the overall gross vertical structure of disks , in which densities are high at the midplane and drop steeply as altitudes increase , optical depth profiles similar to that in figure [ fig : vert - struct ] will occur in other models as well . because the optical depth is a very steeply falling function of increasing altitude , the transition region will be narrow , as in the model above .
whether that narrowness can be translated directly into a constraint on the spatial resolution will depend on the details of the temperature profile there , and its influence on the accuracy of radiative energy transfer .
the specific criterion as applied above presumes a temperature profile decreasing with altitude as the density decreases , @xmath215 .
if instead a much flatter profile exists across the transition , a less restrictive criterion may still permit sufficiently accurate evolution .
after exploring the conditions required for simulations to produce numerically valid results , we turn now to summarizing those conditions , to comparing the work done here with previous work in the literature , and to commenting on additional requirements beyond numerics that are required for simulations to accurately reproduce the evolution of real systems .
our first result is to extend the numerical criterion for the validity of simulations involving collapse in clouds and based on the analytic jeans collapse formalism ( bb97 , @xcite ) , to the case of disk systems where the geometry of the mass distribution is both flattened and rotating so that the jeans treatment is inapplicable . whereas a minimum spatial resolution per jeans wavelength or minimum number of particles per jeans mass is required in the cloud collapse case for grid or particle based simulations , respectively , in simulations of disks we require a minimum spatial resolution per ` toomre wavelength ' or minimum number of particles per ` toomre mass ' , for grid based or particle base simulations , respectively .
the two forms of the new criterion are given in equations [ eq : toom - cond ] and [ eq : max - surfdens ] .
we also note that simulations failing to resolve one or the other condition yield quite different outcomes . in the case of the jeans condition ,
bb97 conclude that insufficient resolution will delay or suppress numerically induced fragmentation with softening and smoothing equal , and only enhance it when smoothing is much larger than softening .
our results demonstrate that a simulation that fails to resolve the toomre condition is characterized by enhanced fragmentation compared to one that is well resolved , even when softening and smoothing are equal .
this fact allows an empirical determination of the minimum resolution required in a simulation in order to ensure fragmentation that develops is not of numerical origin . for 2d sph simulations ,
we have determined the minimum number of particles per toomre mass as six times the average number of neighbors ( @xmath128 ) used for the realization of the hydrodynamic quantities themselves .
we have not determined a value of for 3d models , but its seems unlikely to be dramatically different from 2d , since the same physical system is being modeled with the same numerical technique . to the extent that differences may be present , we expect that the criterion will be more conservative ( more particles required ) based on two arguments .
first , unstable waves in the toomre analysis are intrinsically two dimensional , so resolving their third dimension is meaningless : small particle separations or small grid spacing in that coordinate do not contribute to increasing the resolution of the waves themselves . using the same empirical fragmentation / non - fragmentation limiting condition used in section [ sec : test - crit ] to establish a required particle count , will therefore overestimate the required resolution because some neighbors will be counted only for purposes of establishing a value for , but will not actually serve to increase the resolution insofar as the toomre condition is concerned . secondly , although the toomre criterion implicitly assumes validity of the hydrodynamic equations , in practice the numerical method may fail to do so ( section [ sec : disks-3d ] ) , with the consequence that pressure forces are underestimated .
although a vertical resolution requirement is nominally an entirely different condition from resolving instability wavelengths , we see no practical manner in which they might be separated from each other in the empirical tests we describe . therefore , calibration of the requirements for the toomre condition in 3d will implicitly incorporate both requirements from the wave analysis and from the vertical resolution condition , yielding a resolution requirement defined by contributions from both sources .
a second important result of our work is to reconcile a variety of mutually inconsistent definitions of the jean collapse stability criteria made by different authors , due to the use of different definitions of the jeans mass itself .
we show that using the same definition of the jeans mass , a particle based simulation will require approximately 600 particles to resolve one jeans mass , while a grid based simulation will require about 64 zones ( i.e. @xmath216 where @xmath66 is defined in equation [ eq : jeans - cond ] , and @xmath217 ) .
the minimum resolution of 600 particles per jeans mass corresponds to approximately 12 times the average number of neighbor particles ( @xmath218 ) used by bb97 in their sph code .
this value is a factor of two higher than the factor of six times the average number of neighbors as derived from the 2d disk simulations explored in section [ sec:2dtest - evo ] .
third , we discuss the importance of the implementation of interparticle forces in particle based simulations of self gravitating disks .
we show that a naive translation of the 3d method for gravitational softening in sph into 2d leads to an unphysical , finite gravitational force at zero separations between particles . in consequence , significant particle ` pairing ' occurs as simulations proceed , resulting in an effectively time dependent resolution .
we consider two alternatives to remedy the behavior .
first , we artificially modify the derivative of the sph kernel according to the tc92 prescription , generating a finite , repulsive pressure force at zero separation which compensates for the attractive gravitational force .
second , we artificially modify the gravitational softening itself , to produce instead a zero force at zero separation , as is physically appropriate .
both alternatives are effective in reducing the likelihood of disk simulations to produce fragmentation and of particles in those simulations to become paired .
although the tc92 variant appears to be slightly more effective ( see e.g. section [ sec : enhancement ] ) , we view it as marginally the less preferable of the alternatives because it too produces a finite force at zero separation , albeit of opposite sign so that a sum of gravitational and pressure forces yields a small net force .
finite forces at zero separation , whether or not they are combined with others to produce a near zero sum , must be considered a less physical and therefore less desirable alternative to a solution which retains the collisionless nature of the model more directly .
fourth , we show that force imbalances on the spatial scale of the smoothing and softening develop when the two length scales are not equal . while acting only at very small spatial scales , these imbalances can induce large scale changes in the outcome of simulations where they exist , including artificially induced fragmentation in otherwise stable systems . to avoid such induced outcomes ,
we conclude that the gravitational softening length and the hydrodynamic smoothing length of each particle must be very nearly identical .
as noted above , our conclusion for disk simulations is similar to that made by bb97 in their study of collapsing clouds , but stronger in the sense that it holds even when the systems are , and are expected to remain , properly resolved according to the toomre criterion . in neither
the disk or cloud case however , does our conclusion come without price . specifically , in order to better resolve the flow in both high and lower density regions of the same simulation , most modern sph codes set the smoothing length dynamically as a function of the local flow .
setting the smoothing and softening lengths equal means that the softening length also must vary in time .
this is important because allowing temporally varying softening is equivalent to explicitly allowing a violation of conservation of energy in the simulation . while we believe the magnitude of violation to be small ,
its effect must be considered when evaluating self gravitating particle based simulations .
finally , we discuss the importance of adequately resolving the vertical structure of 3d disk simulations .
we find that insufficient resolution of the vertical structure leads to substantial deficits in the realized midplane densities compared to the predictions of analytical models .
we find that resolution comparable to at least @xmath5 smoothing lengths per disk scale height in the disk midplane are required to accurately fit the coefficients in analytic formulae for the vertical structure of non self gravitating isothermal disks in particle simulations .
restated in terms of the full thickness of the disk , the criterion is equivalent to the statement that at least @xmath219 particles are required per vertical column .
while we have not studied the effects of resolution in 3d grid based simulations , we expect similar criterion to hold for them as well
. the consequences of failures to resolve the vertical structure in self gravitating systems will be that pressure forces will be incorrectly determined .
any balance between them and self gravitation will therefore be falsely biased in favor of gravitational fragmentation .
presumably , the bias will distort the evolution of waves in the disk ( whether self gravitating or not ) even when fragmentation does not occur , because the density and pressure deviations imply a loss of fidelity of the solution to the equations of hydrodynamics provided by the numerical method .
we discuss the fact that while satisfying the criterion above is necessary for simulations to accurately evolve circumstellar disks , it may not be sufficient for them to do so alone , depending on the goals of the researcher .
simulations for which the criterion is satisfied everywhere may still under resolve high altitude structure of the disk , a fact which will be of great importance for models in which radiative transport is included .
such models must also resolve the temperature structure at the disk photosphere in order not to be contaminated by unacceptably large numerical errors .
we provide an example of the requirements when an isentropic vertical structure is present , for which we derive the computationally extremely demanding condition that the photosphere must be resolved at a spatial scale smaller than @xmath220 to avoid serious errors .
we point out however that the exact requirement will be highly sensitive to the details of the model and its effects on the temperature , density and opacity structures at high altitudes .
our test problem clearly shows that the fragmentation processes seen in @xcite were due to insufficient resolution , and we believe that many other works in the literature both before and since may be similarly affected . of the work discussing particle based simulations ,
the simulations done by @xcite are among the highest resolution simulations so far published , evolving either @xmath221 or @xmath222 particles ( in various runs ) in three dimensions
. it may therefore be particularly useful to contrast our test problems with the conditions employed by them , their results and conclusions .
they perform a series of sph simulations with several disk masses between = 0.075 and = 0.125 , each with minimum toomre @xmath4 values between 0.8 and 1.9 .
they show that the more massive disks fragment on a time scale of a few hundred years ( about 10 orbits in the region where the clumps form ) when the minimum @xmath4 values in the disk are below @xmath223 .
they also find that clump formation is more vigorous with higher resolution simulations of the same initial condition that with lower .
an isothermal gas equation of state was used initially for most of the models , and after a critical over density was exceeded , was switched over a small transition time to an adiabatic evolution that included heating due to compression , viscosity and shocks .
other models used an adiabatic equation of state for all times , and resulted in fragmentation that was much weaker or non - existent .
we first attempt to obtain some indications of the degree to which the flow is well resolved by the condition in equation [ eq : max - surfdens ] at different times during the evolution , using their discussion of the physical features of the flow and resolution .
for example , prior to clump formation the most unstable wavelength is quoted as @xmath224 au , including a correction factor for 3d effects .
the flow is clearly well resolved by the simulation at this early stage , but the value of the toomre wavelength , but that their later definition of the most unstable wavelength matches that of equation [ eq : toom - wavel ] .
] at this time is only representative of the mildly non - linear regime of the background flow .
it is not indicative of the later , more strongly non - linear regime near the forming clump , for which both the locally applied jeans based or toomre based resolution criteria are most useful . well after a clump has formed , when the clump has become a well defined , condensed and independent entity rather than a part of the disk , we expect that the toomre based criterion will become invalid . at this time , @xcite state that the resolution employed in their simulations was sufficient to resolve the jeans mass ( for which the underlying analysis is valid if the clump is much smaller than a disk scale height ) with a few hundred to a few thousand particles , even though the densities there reached as much as @xmath225 higher than the background .
an important detail of the implementation of @xcite is that the number of neighbors used for realizing the hydrodynamic quantities is fixed at 32 , a value common throughout the cosmological sph literature , but significantly smaller than the value of @xmath218 used ( but not specifically recommended ) by bb97 .
correcting for this systematic difference would not appear to alter the statement that the jeans mass is properly resolved however , because while the number of particles per jeans mass would be lower , it would still remain above the @xmath52 quoted by bb97 .
we can make relatively few direct inferences about intermediate times when the clump has begun to form but is still unbound or only weakly bound . at such times
the local value of @xmath4 may fall below one half , and we expect a much stronger resolution constraint from the toomre criterion than from the jeans criterion ( see equation [ eq : tj - relative ] ) , especially if the approximation that @xmath59 is not made .
therefore , while the clump may be well resolved by the jeans criterion , it may not be similarly well resolved by the toomre criterion .
based solely on the number of particles and the similarity of the initial conditions to our own test problem , we conclude that the high resolution ( 1 million particle ) simulations of @xcite appear to be well resolved by our toomre based resolution criterion .
we can make a similar , but much more tentative conclusion about the lower resolution ( @xmath226 particle ) versions in the same work .
a more detailed examination of the simulations at both resolutions in the context of the toomre resolution criterion would be desirable , particularly in the context of the sensitive dependence of the resolution requirement to the initial minimum value of @xmath4 and on the local value of @xmath4 in the fragmenting region .
@xcite advocate a fixed gravitational softening whose length scale is half of the initial hydrodynamic smoothing length of the particles . with this ratio of softening to smoothing , bb97
show that the gravitational force can exceed the pressure force by a factor of as much as seven on size scales of @xmath227 , and by a factor of 23 on scales of @xmath88 .
the actual length scale used is determined ( mayer 2004 private communication ) as an average of the smoothing lengths in the ` flat ' temperature region of the disk , corresponding approximately to the outer half of its radial extent , meaning that the outer half of their disks will have small scale force imbalances that originate in the numerical treatment of softening and smoothing .
the results of our simulations show the consequences of such imbalances in disk simulations . in regions where the hydrodynamic smoothing length exceeded the gravitational softening length , fragmentation
was strongly enhanced over that of the variable softening case , even when sufficient numerical resolution was used according to the toomre criterion .
since the force imbalance is not due to any physical origin , but only to a breakdown in the simulation method itself , we conclude that clumping observed in the simulations may not be an accurate depiction of an underlying physical system , and should be reexamined to ensure that they are not subject to this numerical flaw . of still greater concern
is the resolution of their simulations in the context of the discussion in section [ sec:3dsph - nosg - signif ] .
both the morphology of the system modeled by @xcite and the physics itself are different than those employed by our test models , however it seems unlikely that these modifications will alter the required resolution by the large factors needed to conclude that their vertical structure is well resolved , especially for their most unstable ( lowest @xmath4 ) models .
the extent to which any violation of the criterion actually affects fragmentation is unclear however .
it is likely that a definitive statement of the effects will require a complete study of the problem at extremely high resolution , making the question quite costly to address at present .
several other recent papers have discussed fragmentation in disks .
similar concerns regarding the vertical resolution will apply to many of them .
for example , in @xcite , as well as his earlier , similar work , the vertical coordinate ( the spherical coordinate @xmath228 in his case ) is resolved with a total of 23 zones . although they are asymmetrically arranged to concentrate resolution where they are needed most , only some @xmath13010 - 12 lie within ten degrees ( @xmath229 ) of the midplane , where most disk matter will be located .
although our vertical resolution requirement has been tested only with sph simulations , there is no reason to believe that an analogous requirement will be required for grid based simulations as well , since they must solve the same set of equations . naively applying the @xmath5 smoothing lengths per scale height as a @xmath5 grid cells per scale height criterion directly to the boss simulations , we observe that the vertical structure may be very near the required limits
. the exact number of cells will undoubtedly be different than a straightforward carryover however , so the extent to which boss s or any other any grid based simulations will be affected by the errors arising from deficient vertical resolution is unclear .
@xcite showed that for sufficiently fast cooling rates , clumping could occur in disks resolved with @xmath230 particles . due to
their cooling prescription , disks in their work dropped to globally averaged @xmath4 values near unity , well below those in our simulations , for which we determined erroneously low midplane densities developed . in later work discussing similar systems ( with overall morphologies identical to those in our study , but with additional physics ) , @xcite claim that their vertical structure is resolved with @xmath136 smoothing lengths per scale height .
it is difficult to reconcile this result with our own figure [ fig : hiterat - rhodist ] , especially given the much lower @xmath4 values ( see their figure 3 ) realized in their study and in the earlier work of @xcite .
it appears likely however , that low vertical resolution has likely influenced their results through the artificially low pressure forces induced by density deficits .
apart from the issue of vertical resolution , we note that their overall resolution ( in 3d ) is similar to that in our highest resolution 2d test problem .
the authors have not discussed resolution issues in the context of either jeans or toomre stability , so we can draw few inferences about the viability of their results under either the criteria of equation [ eq : max - voldens ] or [ eq : max - surfdens ] . in the context of the latter condition however , the @xmath4 values throughout their simulations rapidly approach values near unity .
this is important because equation [ eq : max - surfdens - q ] shows that our resolution requirement is quite sensitive to the value of @xmath4 , so that much higher resolution will be required for later evolution even if the initial condition is known to be well resolved by the toomre condition .
other simulations of @xcite perform a large parameter study of disks already containing one planet , in order to explore the possibility that its present may trigger the formation of further objects . for their parameter study , they use simulations with @xmath231 particles , but with varying minimum @xmath4 or initial planet mass .
the initial conditions of these disks were in most respects quite similar to those of mayer , however the resolution employed was far lower in order to facilitate the large parameter study . given the much lower resolution , it appears likely that these authors have under resolved the vertical structure of the disk , resulting in incorrectly low densities and pressures and the possibility of artificially enhanced fragmentation . while an attempt to duplicate their study is beyond the scope of this work , we have run selected simulations sampling their parameter space and
have observed that they resolve ( or fail to resolve ) the toomre condition at similar points in the parameter space for which they conclude that clumping was induced or not induced by the passage of the seed planet . due to the probable violation of both criteria
, we therefore believe that their conclusions have been contaminated by insufficient resolution and must be reexamined . in this work
, it has not been our purpose to discuss all of the requirements for simulations modeling gravitational fragmentation in disks to produce correct and physically relevant results .
such a description of a real physical system requires not only numerically valid simulations , but also relevant initial conditions and a correct and complete physical model .
numerically valid simulations will be interesting only to the extent that they model real systems with real physics . of particular interest for the physical understanding of disk evolution
is the fact that our highest resolution sph test model did not produce _ any _ clumps although it was evolved for much longer in time than its lower resolution cousins .
this is an important physical result because the test problem employed a locally isothermal equation of state , which is usually ( and erroneously - see the discussion in @xcite ) equated with the strong cooling limit where fragmentation is nearly inevitable .
the fact that clump formation did not occur in this limit , in a properly resolved simulation , may place the disk instability model for planet and brown dwarf formation on considerably less solid ground . at the same time , the boss simulation , implementing conditions that are much less biased toward clumping , _ did _ produce clumps while remaining numerically valid according to the toomre condition , though some doubt may remain regarding the resolution of the vertical structure .
his simulations include a description of radiative cooling processes in 3d , and he concludes that the clumping is due both to the efficient radiative cooling and to the fact that convection in the disk s vertical direction is efficient .
convection is able to transport thermal energy out of the optically thick disk midplane to its photosphere surface at higher altitudes , so that even optically thick regions like forming clumps cool efficiently . on the other hand ,
previous 2d work of @xcite , which assumed efficient vertical convection as a consequence of the radiative cooling prescription employed , did not produce clumps .
moreover , the total radiated energy emitted by the disks modeled in those simulations seriously underestimated the values observed in real systems because they were too cold . in a more complete physical model of the disk energy budget of heating and cooling processes
, the disks would have to be warmer in order to reproduce the observations , making clumping even less likely .
other fully 3d radiative transfer @xcite models confirm neither the boss conclusion about the efficiency of vertical convection nor the clumping that results from it . the authors continue to investigate the origin of the discrepancies , but no certainty yet exists .
possibilities include both differences in the numerical treatments , for example of the boundaries , the resolution of both the poisson solver and the hydrodynamic codes themselves , and of the physical models , for example the treatment of external radiation field , what opacities are used in the radiative transport and the equation of state for the gas .
the important point to take from these discrepant results is that not only do questions of numerical validity remain to be addressed , but also the physical models .
although progress has certainly been made , to date an insufficient understanding of what physical processes are important for fragmentation has been developed .
determination of the true viability of the disk instability model for jovian planet formation therefore still lies in the future .
except for the simulation discussed in figure [ fig : boss - crit ] , generously shared prior to its publication by alan boss , the computations reported here were performed using the uk astrophysical fluids facility ( ukaff ) , specifically the ukaff fellow guaranteed time allocation , which also provided financial support .
this work was partially carried out under the auspices of the national nuclear security administration of the u.s .
department of energy at los alamos national laboratory under contract no .
de - ac52 - 06na25396 and w-7405-eng-36 , for which this is publication la - ur-05 - 5347 .
we gratefully acknowledge useful conversations and correspondence with alan boss , richard durisen , annie mejia , megan pickett , lucio mayer and willy benz during the evolution and preparation of this work .
special note is given to richard durisen , with whom discussion of the toomre resolution criterion work shortly before iau symposium 200 , gave original impetus to its eventual publication .
we also appreciate the comments of the co - referees matthew bate and daniel price during the contentious review process experienced by this paper over two years and five revisions .
finally , we would like to acknowledge the support of the university of edinburgh development trust . | we define three requirements for accurate simulations that attempt to model circumstellar disks and the formation of collapsed objects ( e.g. planets ) within them .
first , we define a resolution requirement based on the wavelength for neutral stability of self gravitating waves in the disk , where a jeans analysis does not apply . for particle based or grid based simulations ,
this criterion takes the form , respectively , of a minimum number of particles per critical ( ` toomre ' ) mass or maximum value of a ` toomre number ' , @xmath0 , where the wavelength , @xmath1 , is the wavelength for neutral stability for waves in disks .
the requirements are analogues of the conditions for cloud collapse simulations as discussed in @xcite and @xcite , where the required minimum resolution was shown to be twice the number of neighbors per jeans mass or 4 - 5 times the local jeans wavelength , @xmath2 , for particle or grid simulations , respectively .
we apply our criterion to particle simulations of disk evolution and find that in order to prevent numerically induced fragmentation of the disk , the toomre mass must be resolved by a minimum of six times the average number of neighbor particles used .
we investigate the origin of the apparent discrepancy between the number of particles required by the cloud and disk fragmentation criteria and find that it is due largely to ambiguities in the definition of the jeans mass , as used by different authors .
we reconcile the various definitions , and when an identical definition of the jeans mass is used , the condition that @xmath3 in the truelove condition is equivalent to requiring about 10 - 12 times the average number of neighbor particles per jeans mass in an sph simulation , reducing the difference between simulations of disks and clouds to about two .
while the numbers of particles per critical mass are similar for both the jeans and toomre formalisms , the toomre requirement is more restrictive than the jeans requirement when the local value of the toomre stability parameter @xmath4 falls below about one half .
second , we require that particle based simulations with self gravity use a variable gravitational softening , in order to avoid inducing fragmentation by an inappropriate choice of softening length . we show that using a fixed gravitational softening length for all particles can lead either to artificial suppression or enhancement of structure ( including fragmentation ) in a given disk , or both in different locations of the same disk , depending on the value chosen for the softening length . unphysical behavior can occur whether or not the system is properly resolved by the new toomre criterion .
third , we require that three dimensional sph simulations resolve the vertical structure with at least @xmath5 particle smoothing lengths per scale height at the disk midplane , a value which implies a substantially larger number per vertical column because the disk itself extends over many scale heights .
we suggest that a similar criterion applies to grid based simulations .
we demonstrate that failure to meet this criterion leads to underestimates in the midplane density of up to 3050% at resolutions common in the literature . as a direct consequence ,
gas pressures will be dramatically underestimated and simulations of self gravitating systems may artificially and erroneously inflate the likelihood of fragmentation .
we outline an additional condition on the vertical resolution in simulations that include radiative transfer in order to ensure a correct description of the cooling , specifically that the temperature structure near the disk photosphere must be well resolved . as an example , we demonstrate that for an isentropic vertical structure , the criterion translates to resolution comparable to @xmath6 near the disk photosphere , to avoid serious errors in transfer rates of thermal energy in and out of the disk . finally , we discuss results in the literature that purport to form collapsed objects and conclude that many are likely to have violated one or more of our criteria , and have therefore made incorrect conclusions regarding the likelihood for fragmentation and planet formation . solar system : formation , stars : planetary systems : protoplanetary disks , hydrodynamics , methods : numerical |
cosmic string is one of the relic structures that are predicted to be produced in the course of symmetry breaking in the hot , early universe , whose discovery will probably be an important landmark for the high energy physics .
the quest for cosmic strings has been conducted in two ways : statistical method and direct search for individual cosmic string . on the statistical side
, many studies agreed that the contribution from cosmic strings to statistical observables such as the power spectrum is at most @xmath2 @xcite , reconfirming that cosmic strings played a minor role , if any , in making the universe .
other workers have set upper limits on the cosmic string parameter @xmath3 @xcite . in this letter
, we introduce a technique for estimating how strong a signal from cosmic strings has to be in order to be identified unequivocally .
the detecting power of this technique is calibrated by applying it to simulated anisotropy maps based on reasonable modeling . in the last part of this letter
, we apply this technique to the _ wmap _ 3 year w - band data set and analyze the implication of the results .
a cosmic string can leave discrete temperature steps in a cmb anisotropy map due to the kaiser - stebbins effect @xcite , with the height of a step @xmath4 given by @xmath5 where @xmath6 is the universal background temperature of cmb .
however , those step patterns are probably obscured by instrumental noises and other physical structures of anisotropies .
decomposition of a cmb anisotropy sky map patch .
a square patch is decomposed into the uniform background temperature ( @xmath7 ) + gaussian fluctuation ( @xmath8)+ discrete temperature step ( @xmath9 ) .
the number of pixels in a patch is @xmath10.,width=491 ] we consider a square patch of the cmb anisotropy sky map that extends to a size of the horizon at the time of recombination so that the conic spacetime formed by a cosmic string can be in effect within the region .
then another square patch of the cmb anisotropy sky map that contains a segment of moving cosmic string can be decomposed into three parts , as illustrated in figure [ fig1 ] : ( 1 ) the uniform background temperature ( @xmath7 ) , ( 2 ) the gaussian fluctuation ( variance @xmath11 ) , and ( 3 ) the discrete temperature step ( @xmath9 ) from a moving cosmic string .
the uniform background temperature @xmath7 comes from superhorizon - scale primordial fluctuations .
since we pick a horizon - sized region , the super - horizon fluctuations will appear to be a constant temperature shift for the whole patch .
smaller scale ( sub - horizon ) fluctuations and instrumental noise add up incoherently to form a gaussian fluctuation with the variance @xmath12 , where @xmath13 and @xmath14 are variances for fluctuations of primordial origin and instrumental noise , respectively . we introduce five parameters that characterize a square patch of a string - embedded sky map , @xmath7 , @xmath8 , @xmath15 , @xmath16 ( blueshifted pixels / total pixels ) , and @xmath17 ( orientation of step ) .
we concoct a simulated patch of the cmb anisotropy map by assigning arbitrary legitimate values to the parameters and adding the three components illustrated in figure [ fig1 ] .
to recover these parameters from the cmb anisotropy map where the step pattern is intermixed , we employ five observables @xmath18 that can be expressed in terms of @xmath19 where @xmath20 represents the signal at the @xmath21th pixel and @xmath22 is the two - dimensional gridded position vector for the @xmath21th pixel defined in a square patch .
evolution of errors of characteristic parameters as functions of @xmath15 with input values of parameters @xmath23 .
_ left panels _ : @xmath24 pixels in a square . _
middle panels _ : @xmath25 pixels .
_ right panels _ : @xmath26 pixels . for each size of patch ,
the number of simulations performed is 10,000 .
the shaded regions indicate the ranges of @xmath15 for which the step feature in a patch is not obvious compared to its background fluctuation ( @xmath27 ) .
the result on @xmath28 is biased upward because the simulations with erroneous values of @xmath28 ( @xmath29 ) were not included in the statistics.,width=491 ] in the absence of the background fluctuation ( @xmath30 ) , the following analytic expressions for the characteristic parameters return the exact input values : @xmath31 with the background fluctuation turned on , an observed characteristic parameter @xmath32 is expressed as an unbiased estimate @xmath33 plus a gaussian error @xmath34 @xmath35 , @xmath36 the errors ( @xmath37 ) present in the equations ( [ 11 ] ) are measures of how precisely the information on a temperature step screened by background fluctuation is recovered .
figure [ fig2 ] shows the behaviors of errors as functions of @xmath15 for three different pixel numbers for a patch .
the standard deviation of the background fluctuation chosen here is the mean standard deviation of patches with angular radius @xmath38 ( the angular size of the horizon at the recombination for the @xmath39cdm model ) for the wmap 3 year w - band data set with the kp2 mask applied .
the most pronounced feature of the graphs is that the error bars are wildly undulating for small @xmath15 compared to @xmath8 , with attenuating envelopes as @xmath15 increases .
the algorithm works poorly for low @xmath15 compared to @xmath8 , such as the @xmath40 error on the orientation estimate .
.[table1]simulation results at the critical values of @xmath15 above which @xmath41s are small enough and do not decrease any more . [ cols="^,>,<,>,<,>,<",options="header " , ] even the breakdowns of the algorithm do happen in the shaded region , resulting in erroneous parameter values such as negative @xmath15 or @xmath16 not in the range between 0 and 1 .
simulations with collapsed results are not included in the statistics shown in figure [ fig2 ] since those cases are evidently the ones with insufficiently strong signals against background fluctuation .
as @xmath15 increases , we begin obtaining the computed parameters that are very close to the true values with acceptable errors ( dubbed the `` good '' results ) , and it allows us to recover the temperature step parameters faithfully .
we also find from figure [ fig2 ] that a patch with more pixels works better with a faster attenuation of uncertainties .
table [ table1 ] displays the performances of the algorithm for the values of @xmath15 at the borders , above which errors for the temperature step ( @xmath41 ) do not get any better .
we use circular patches for _ wmap _ data analysis rather than square patches because of the computational advantage .
algebraic expressions for the observables given in equations ( [ 6])-([10 ] ) work very well for circular patches with negligible numerical differences for the case of square patches . a square patch with @xmath26 pixels at the normal resolution of _ wmap _ w - band data
covers nearly the same area as the circular region with angular radius @xmath38 .
we performed a further detecting power test with @xmath26 pixels patches and found empirically that the relation between the critical value of @xmath15 , @xmath42 , above which `` good '' results start to come out , and @xmath8 is @xmath43 the different choices of input values for the parameters @xmath16 , @xmath17 , or @xmath7 showed no noticeable difference in the evolutions of errors .
an observed cmb anisotropy map is an aggregate of various independent modes of perturbation ranging from tiny sub - horizon scales to super - horizon scales well beyond the correlation length .
as illustrated in figure [ fig1 ] , fluctuations with larger or smaller scales compared to the size of a test patch are neatly prescribed , but the intermediate scales whose wavelengths are comparable to the size of a test patch will appear to be continuous temperature tilts that also give plausible values of the step parameters .
one drawback of this algorithm is that it does not distinguish between a discrete temperature step and a smooth temperature slope .
however , this shortcoming can be easily fixed : if an apparent temperature step is detected at a spot on the map , we repeat the analysis with a half - sized patch at the same spot . if the structure is a continuous slope , then it would return half the value of @xmath28 than the previous result , while for the signal from a discrete step , the returned @xmath28 should stay the same within the error .
we conducted the step signal search through the _ wmap _ 3 year w - band data set and found 193,160 unqualified signals against background fluctuation ( the constraint in eq .
[ [ 12 ] ] is not applied ) , 129,049 qualified steps+tilts ( signals that meet the constraint eq . [
[ 12 ] ] ; i.e. , @xmath44 ) , and 12,330 qualified discriminated steps .
distributions of @xmath28 detected in _ wmap _ 3 year w - band data set .
the number of step signals found are 193,160 unqualified signals ( _ dotted curve _ ) , 129,049 qualified steps+tilts ( _ dashed curve _ ) , and 12,330 qualified steps ( _ solid curve_).,width=264 ] the @xmath28 value of qualified steps is in the range @xmath45 as the curves show in figure [ fig3 ] . patches with radius @xmath38 in the _ wmap _ 3 year w - band have a background fluctuation @xmath8 less than 176 @xmath46 in the kp2-mask - cleared region .
this means , at the worst case , we can identify a step signal as low as @xmath47 with moderate errors .
therefore , if there are cosmic string signals with @xmath48 and if they are located in the available region ( out of the galactic plane or the kp2 masked region ) , they would not be missed .
thus , we can set an upper limit on the cosmic string signal @xmath49 where @xmath50 is arbitrary .
thus , with @xmath51 @xcite , the upper limit of the cosmic string parameter @xmath3 can be estimated as @xmath52
we have presented and tested a new technique to directly discover cosmic strings via the patterns they would produce in a cmb anisotropy map .
we found that the minimum magnitude of the step signal that is required to be unequivocally identified is @xmath53 for the _ wmap _ 3 year w - band data set .
this algorithm can be used to crop the reliable step signals from the cmb anisotropy data , and it will serve as the valuable ground work for future pattern searches with more refined data , such as further _
wmap _ data releases or _ planck _ data . 99 pogosian et al . , 2003 , phys.rev.d , 68 , 023506 wu , j - h . , 2005 , astro - ph/0501239 fraisse , a. , 2005 , astro - ph/0503402 davis , a .- c . & kibble , t.w.b , 2005 , contemp.phys , 46 , 313 wyman , m. , pogosian , l . , & wasserman , i. , 2005 , phys.rev.d , 72 , 023513 hindmarsh , m. , 1994 , apj , 431 , 534 perivolaropoulos , l. & simatos , n. , 1998 , astro - ph/9803321 jeong , e. & smoot , g. f. , 2005 , apj , 624 , 21 lo , a. s. & wright , e. l. , 2005 , astro - ph/0503120 seljak , u. & slosar , a. , 2006 , phys.rev.d , 74 , 063523 pogosian , l. , wasserman , i. , & wyman , m. , 2006 , astro - ph/0604141 fraisse , a. , 2006 , astro - ph/0603589 kaiser , n. & stebbins , a. , 1984 , nature , 310 , 391 vilenkin , a. & shellard , e. p. s. , 1994 , cosmic strings and other topological defects , ( cambridge , uk , cambridge university press ) grski , k. m. et.al . , 2005 , apj , 622 , 759 | we introduce a new technique to detect the discrete temperature steps that cosmic strings might have left in the cosmic microwave background ( cmb ) anisotropy map .
the technique provides a validity test on the pattern search of cosmic strings that could serve as the groundwork for future pattern searches .
the detecting power of the technique is only constrained by two unavoidable features of cmb data : ( 1 ) the finite pixelization of the sky map and ( 2 ) the gaussian fluctuation from instrumental noise and primordial anisotropy .
we set the upper limit on the cosmic string parameter as @xmath0 at the 95% confidence level ( cl ) and find that the amplitude of the temperature step has to be greater than @xmath1 in order to be detected for the _ wilkinson microwave anisotropy probe ( wmap ) _ 3 year data . |
fermilab experiment 789 is a study of two - prong decays of beauty and charm which took data during the 1990/1 fixed - target run .
e789 has published the most precise measurement of the charm production cross section and its @xmath4 dependence in 800gev proton collisions,@xcite as well as novel measurements at very forward @xmath5 of the @xmath2 production cross section and its @xmath4 dependence using our beam dump ( or a beryllium insert just upstream of the beam dump ) as the target.@xcite more recent results include the only measurement to date of the cross section for beauty production in proton collisions at fixed - target energy,@xcite observed via the process @xmath6 , and high - statistics studies of @xmath2 and @xmath3 production.@xcite
the e789 apparatus ( shown schematically in fig . [ fig : e789 ] ) has been described in detail elsewhere@xcite ; we summarize briefly here .
it is based on the pre - existing e605/772 spectrometer,@xcite upgraded for this run by the addition of two silicon - microstrip vertex telescopes ( fig .
[ fig : ssd789 ] ) , one above and one below the beam , replacement of the station-1 mwpcs with drift chambers , and a tenfold increase in data - recording capacity ( to 50mb / spill ) .
wire - like targets were used to localize the primary - interaction vertex in @xmath7 and @xmath8 , so that only the decay vertex need be reconstructed .
= 2.87 in = 2.54 in the spectrometer features two large analysis magnets , sm12 and sm3 , which deflect charged particles in opposite directions in @xmath7 .
a water - cooled copper beam dump suspended within sm12 absorbs non - interacting beam protons as well as secondaries emitted within @xmath9@xmath10mr of the beam in @xmath7 . shielding within and around sm12 absorbs neutral secondaries .
this geometry limits the pair acceptance to @xmath91% but allows operation at high interaction rates .
the vertex telescopes and the 23 planes of scintillation - counter hodoscopes and drift chambers at stations 1 , 2 , and 3 measure the tracks of charged particles passing above or below the beam dump . the sm3 magnet serves to remeasure charged - particle momenta and thus to confirm the target origin of tracks , allowing the copious background of muons created within the beam dump to be rejected .
particles are identified by electromagnetic and hadronic calorimeters , scintillation - hodoscope and proportional - tube muon detectors , and a ring - imaging cherenkov counter . in most events
only two oppositely - charged particle tracks traversing the spectrometer are fully reconstructed , one passing through each vertex telescope .
data were taken separately in charm and beauty spectrometer settings . in the charm running we were able to operate the spectrometer at interaction rates up to 5mhz , using an on - line vertex trigger processor@xcite to reject @xmath1180% of hadron pairs from the target .
the higher sm12 current used in the beauty setting allowed @xmath2 and @xmath3 data to be taken at a 50mhz interaction rate , with no on - line vertex cut needed .
we have measured differential cross sections for charm@xcite and charmonium@xcite production .
charm data were taken using gold and beryllium targets at sm12 currents of 900 and 1000 amperes .
[ fig : d - peak ] shows the observed hadron - pair mass distributions ( under the @xmath12 and @xmath13 assumptions mass region and was not used in this analysis . ] ) for the various e789 charm data samples , using two different decay - vertex cuts .
clear @xmath14 signals stand out above the dihadron background .
the looser vertex cuts ( @xmath15 ) were found to optimize the statistical significance of the signal and were used in the cross - section and @xmath4-dependence analyses .
= 2.4 in = 2.65 in our narrow range of acceptance in the longitudinal momentum of the pair ( due to the beam - dump and shielding geometry ) precludes a direct measurement of total cross section .
we measure@xcite the @xmath16 differential cross section @xmath17b / nucleon at @xmath18 .
this can be extrapolated over all @xmath5 ( using the @xmath5 shape measured by previous experiments@xcite ) to give a total @xmath16 cross section @xmath19b / nucleon.@xcite averaging with previous measurements using 800gev proton beams@xcite gives @xmath20b / nucleon , consistent with next - to - leading - order ( nlo ) qcd predictions@xcite within the broad range of theoretical uncertainty .
[ fig : mumu - mass ] also shows the dimuon mass distribution from the beauty running , and figs .
[ fig : psi - pt ] and [ fig : psip - pt ] compare our measured @xmath2 and @xmath3 differential cross sections with qcd predictions.@xcite here there is little uncertainty in the extrapolation over @xmath5 , with the @xmath5 shape well determined both by these open - aperture " data alone and in combination with our beam - dump data,@xcite as indicated in fig .
[ fig : psi - xf ] . as at the tevatron
collider,@xcite @xmath2 and @xmath3 production are substantially underestimated in the qcd calculation , with phenomenological @xmath21 factors " of 7 and 25 ( respectively ) required to give agreement in magnitude between data and theory . note that the model calculation includes only contributions from color - singlet charmonium states and neglects possible contributions from color - octet charmonium components@xcite and from postulated states above @xmath22 threshold.@xcite ( these discrepancies have been the subject of much attention and are discussed further in sec . 5 .
) we find@xcite @xmath23nb / nucleon and @xmath24nb / nucleon .
comparison with previous results@xcite@xmath25@xcite ( fig .
[ fig : sdep ] ) shows that the @xmath2 total cross section and its excitation curve are by now well determined experimentally .
= 3.75 in = 3.75 in = 3.75 in = 3.75 in
nuclear dependences of production cross sections can shed light on production mechanisms and thus are of intrinsic interest .
in addition , suppression of charmonium production in nucleus - nucleus collisions has been proposed as a signature for quark - gluon - plasma formation,@xcite so it is important to study processes responsible for charmonium suppression in proton - nucleus collisions , which might present a background to a quark - gluon - plasma signal .
the production of heavy quarks is naively expected to depend linearly on the atomic weight ( @xmath4 ) of the target nucleus , since the dominant qcd mechanisms ( gluon - gluon fusion and @xmath26 annihilation ) involve hard partons.@xcite however , in fermilab e772 we showed@xcite that charmonium production in proton - nucleus collisions in fact has a complicated dependence on @xmath4 , parametrized as @xmath27 , suggesting that other processes are at work in addition to those of perturbative qcd . the concentration of our charm sample in a narrow range of feynman-@xmath28 results in the most precise determination to date of the charm - production nuclear dependence at a point in @xmath5 , allowing a precise comparison to be made between open - charm and charmonium production . to augment the forward-@xmath5 @xmath4-dependence measurements of e772
, we took additional @xmath2 data in e789 , using a rotating wheel of beryllium , carbon , and tungsten targets placed 1.27 m downstream of the usual target position in order increase the acceptance near @xmath29 .
[ fig : negjpsi ] shows the dimuon mass distributions thus obtained and the resulting @xmath4-dependence exponent @xmath30 _ vs. _
@xmath5 for @xmath2 compared with that for @xmath31 .
we see that ( at least at small @xmath5 ) @xmath30 is significantly lower for charmonium than for charm : @xmath31 production depends linearly on @xmath4 ( @xmath32 at @xmath33),@xcite while for the @xmath2 , @xmath34.@xcite = 3.3 in = 3.3 in the increased nuclear suppression at low @xmath5 of charmonium as compared to charm is consistent with models in which charmonium production is suppressed in nuclei due to dissociation by interaction with co - moving partons.@xcite models in which the nuclear suppression of charmonium production is an initial - state effect ( _ e.g . _ due to possible shadowing or nuclear modification of the gluon structure function ) are disfavored , since they would predict similar nuclear dependences for @xmath2 and @xmath35 production .
effects on charmonium production at large @xmath5 due to intrinsic charm@xcite ( the presence of virtual @xmath36 pairs in the nucleon sea ) and initial - state parton energy loss@xcite have also been postulated .
the predictions of the intrinsic - charm model have not been borne out by our data since they feature significantly larger ( and more strongly @xmath4-dependent ) cross - section contributions at the largest @xmath5 than are seen.@xcite the qualitative trend we observe at large @xmath5 ( nuclear suppression increasing with @xmath5 ) is successfully accomodated in models which take account of parton energy loss@xcite in traversing nuclear matter . since gluons should interact more strongly with matter than quarks , the parton energy - loss model makes the intriguing prediction that at the highest feynman-@xmath28 ( where @xmath37 annihilation dominates over gluon - gluon fusion ) , the nuclear suppression should decrease.@xcite the apparent increase of @xmath30 at @xmath38 suggests that this may be occurring , but better statistics are needed for confirmation ; these should be forthcoming from fermilab e866 , which is to take data during the 1996/7 fixed - target run .
production of charm@xcite and beauty@xcite quarks is in reasonable agreement with perturbative qcd calculations . on the other hand , production of charm and beauty _ quarkonia _
are observed at rates from one to two orders of magnitude higher than naively predicted .
the experimental facts of enhanced quarkonium production in fixed - target experiments are not new , dating back to the mid-1970s when @xmath2 hadroproduction was first observed . at that time
it was realized that lowest - order production of @xmath36 in a color - singlet state ( via one intermediate virtual photon or three virtual gluons ) had difficulty accounting for the large cross sections measured,@xcite and the color evaporation"@xcite ( or local - duality ) mechanism achieved currency : the @xmath36 pair could be produced in a colored state and later emit a soft gluon to neutralize its color at little or no cost in probability .
recent advances in perturbative qcd have made the predictions of the color - singlet model computable with no free parameters,@xcite allowing it to be definitively ruled out by both fixed - target and collider@xcite measurements . in the regime accessible to fixed - target experiments , questions may still remain as to the applicability of factorization and the role of intrinsic parton @xmath39 .
but the collider data in the previously - inaccessible regime of @xmath40 ( where perturbative calculations ought to be most trustworthy ) have forced theorists to consider seriously additional non - perturbative mechanisms . as a result , the leading candidate models which have emerged are the color - octet model@xcite and an updated color - evaporation model.@xcite these models , while less predictive than the color - singlet model , nevertheless make strong predictions ( for example , that @xmath2 s should be highly polarized at high @xmath41 ) , which can be tested in detail in upcoming experiments . in this connection
we mention fermilab e866 , which should record @xmath42 @xmath1 decays ( in closed aperture ) in the upcoming run , as well as the c0 charm project,@xcite in which a sample of @xmath43 @xmath2 decays could be accumulated , with ( due to the open geometry ) most final - state particles accompanying the @xmath2 also measured .
thus c0 charm holds the possibility of high - statistics measurements of @xmath44 ( as well as @xmath2 , @xmath3 , and open - charm ) production at fixed - target energy
. further data on @xmath4 dependences could also be useful .
energy loss in nuclear matter may be a means to distinguish color - singlet and color - octet charmonium states , since the color - octet state has gluonic quantum numbers and may be strongly absorbed.@xcite distinguishing initial- and final - state @xmath4-dependence mechanisms calls for more data on the @xmath4 dependence of open - charm production at large @xmath5 , another area in which c0 charm could contribute .
m. l. mangano , phenomenology of quarkonium production in hadronic collisions , " hep - ph/9507353 , in _ proc .
xth topical workshop on proton - antiproton collider physics _ ,
fermilab , batavia il , may 913 , 1995 , r. raja and j. yoh , _ eds .
_ , aip conference proceedings 357 , american inst .
( 1996 ) , p. 120 .
see for example m. w. bailey ( cdf collaboration ) , charmonium and bottomonium production in @xmath46 collisions at cdf , " fermilab - conf-96/235-e , to appear in _ proc . 1996 meeting of the division of particles and fields , american physical society _ , minneapolis , mn , august 1015 , 1996 .
see for example k. a. bazizi ( d0 collaboration ) , inclusive @xmath47 quark and heavy quarkonium production at d0 , " in _ proc .
xth topical workshop on proton - antiproton collider physics _ ,
fermilab , batavia il , may 913 , 1995 , r. raja and j. yoh , _ eds .
_ , aip conference proceedings 357 , american inst .
( 1996 ) , p. 105 . c. n. brown _ et al .
_ , expression of interest for a high - sensitivity charm experiment at c0 , " d. m. kaplan , spokesperson , july 9 , 1996,http://fnphyx - www.fnal.gov / conferences / ftp_home / charm2000/c0_charm.ps .
r. wittmann and u. heinz , z. phys .
c * 59 * , 77 ( 1993 ) ; + p. l. mcgaughey , to appear in _ proc .
quark matter 96 , 12th international conference on ultra - relativistic nucleus - nucleus collisions _ , heidelberg , germany , may 2024 , 1996 . | using a sample of @xmath0 @xmath1 decays , fermilab experiment 789 has studied production of @xmath2 and @xmath3 in 800gev proton - nucleon collisions . differential cross sections and nuclear dependences have been measured for charmonium as well as for charm and beauty production .
while charm and beauty production are consistent with perturbative qcd calculations , charmonium cross sections exceed the predictions of the color - singlet model by large factors , suggesting that additional mechanisms ( such as color - octet production ) may play important roles .
nuclear dependences of production cross sections may offer a new tool for the detailed understanding of charmonium production .
= 20000 |
much of our knowledge about _ qcd _ and the structure of the nucleon has been derived from deep inelastic scattering ( _ dis _ ) experiments , either @xmath1 ( @xmath2 ) via the exchange of a photon or @xmath3 , @xmath4 via @xmath5 or @xmath6 respectively .
the cross section is determined by the structure functions @xmath7 and additionally for neutrino _ dis _ , @xmath8 .
considering non - singlet ( _ ns _ or @xmath9 ) combinations only to avoid any extra gluon terms , the operator product expansion relates moments of these structure functions to nucleon matrix elements @xmath10 as @xmath11 where @xmath12 is the wilson coefficient and @xmath13 is proportional to the matrix element of @xmath14 .
@xmath15 is a quark bilinear form , involving a @xmath16-matrix and @xmath17 covariant derivatives ( see @xcite for our conventions ) .
this ` renormalisation group invariant ' ( _ rgi _ ) form gives a clean separation between a non - perturbative _ number _ @xmath13 ( which can be computed on the lattice ) and a _ function _
@xmath18 ( which is perturbatively known ) .
this is a possible _ direct _ comparison between the experimental result and the lattice result .
more practical at present , however , is to use _ parton density _ functions ( e.g. _ mrs _ , @xcite ) determined from _ global _ fits and related to the structure function by a _ convolution_. finally , in a scheme @xmath19 at scale @xmath20 then @xmath21 with @xmath22^{-1 } \equiv \left [ 2b_0 g^{\cal s}(m)^2 \right]^{- { d_{{\cal o};0}\over 2b_0 } } } & & \nonumber
\\ & & \times \exp{\left\ { \int_0^{g^{\cal s}(m ) } d\xi \left [ { \gamma^{\cal s}_{\cal o}(\xi ) \over \beta^{\cal s}(\xi ) } + { d_{{\cal o};0}\over b_0 \xi } \right ] \right\ } } , \nonumber\end{aligned}\ ] ] which in the @xmath23 scheme with @xmath24 from @xcite at @xmath25 gives for quenched @xmath26 , @xmath27 , @xmath28 ( @xmath29 , @xmath30 , @xmath31 respectively ) while for unquenched , @xmath32^{-1 } \sim 0.695(10)$ ] .
the euclideanisation and form of the @xmath15 has been described in @xcite : @xmath33 can be determined with nucleon momentum @xmath34 , while @xmath35 , @xmath36 and @xmath37 need a three - momentum with one non - zero component . by considering the _ ns _ term
the difficult to compute one - quark - line - disconnected terms cancel .
a given operator can mix with three different classes of operators ( with the same quantum numbers under the hypercubic group ) : [ a ] of higher dimension ; [ b ] the same dimension ; [ c ] lower dimension . at present
the @xmath0 improvement operators ( class [ a ] ) are only known for @xmath38 .
the associated improvement coefficients are not completely known , @xcite . for @xmath36 and @xmath37
there are additionally two mixing operators belonging to class [ b ] and for @xmath37 a further operator in class [ c ] .
for @xmath38 we have found that the numerical values of the improvement terms are much smaller than the operator itself ; thus dropping them will cause only an insignificant error .
renormalisation can be considered in the @xmath39 scheme both perturbatively and non - perturbatively , @xcite , by defining @xmath40 from @xmath41 = \langle q(p ) | { \cal o}^{{\mbox{\tiny $ born$ } } } | q(p ) \rangle , \nonumber\end{aligned}\ ] ] at @xmath42 .
perturbation theory gives @xmath43 + o(g_0 ^ 4 ) , \nonumber\end{aligned}\ ] ] where we have now computed @xmath44 for a general value of @xmath45 , @xcite .
the perturbative renormalisation constant for the first mixing term for @xmath36 has also been computed ; it turns out to be very small .
numerically this term is also smaller than the @xmath36 matrix element so we shall also drop it .
problems arise with the remaining mixing terms though : the born term ( between quark states ) vanishes , making their determination using only quark states impossible . at the present stage of development
we shall simply ignore these extra terms . practically for the operator renormalisation we shall use a variant of ` tadpole
improved , renormalisation group improved , boosted perturbation theory ' , ( _ ti - rgi - bpt _ ) , @xmath46 @xmath47^{d_{{\cal o};0}\over 2b_0 } \times } & & \nonumber \\ & & \left [ 1 + { b_1 \over b_0 } g^2_{{\box } } \right]^ { { b_0 d_{{\cal o};1}^{{\box } } - b_1 d_{{\cal o};0 } \over 2 b_0b_1 } + { { p_1 \over 4 } { b_0 \over b_1 } ( 1-n_d ) } } , \nonumber\end{aligned}\ ] ] ( where @xmath48 is the number of derivatives in the operator , @xmath49 , @xmath50 and @xmath51 is the first perturbative coefficient in @xmath52 ) .
@xmath53 in the ` @xmath54 ' scheme may be found from the known @xmath55 ( from @xmath56 ) .
results for unimproved quenched fermions are shown in fig . [ fig_v2b_v3+v4_momms_nf0_rgi_lat01 ] and compared with the non - perturbative method , @xcite .
reasonable agreement is seen .
= 7.00 cm [ fig_v2b_v3+v4_momms_nf0_rgi_lat01 ]
there have recently been suggestions @xcite for the behaviour of @xmath57 close to the chiral limit , @xmath58 \end{array } \nonumber\end{aligned}\ ] ] rather than a linear form : @xmath59 .
we check this using quenched unimproved results ( because this data extends down to lighter quark mass ) in fig .
[ fig_x1u - d_wilson_rgi_lat01 ] .
= 7.00 cm [ fig_x1u - d_wilson_rgi_lat01 ] although not conclusive , it would seem that presently linear fits are adequate and any possible non - linearities can only show up at rather small quark mass outside the present range of data .
we first consider the continuum limit using quenched @xmath0 improved fermions ( at @xmath60 @xmath61 , @xmath62 , @xmath63 ) after taking the chiral limit . in fig .
[ fig_x1b+x2+x3_aor02_lat01 ] we show the results .
the data for the higher moments is unfortunately rather noisy .
= 7.00 cm [ fig_x1b+x2+x3_aor02_lat01 ] finally we consider _ unquenched _ results . in fig .
[ fig_x1b_1u-1d.p0_mpi2+a2_nf2_lat01 ] we use a fit function = 7.00 cm [ fig_x1b_1u-1d.p0_mpi2+a2_nf2_lat01 ] ansatz @xmath64 for @xmath0 improved fermions with @xmath65 data sets . there is no obvious difference to the _ quenched result _ which is surprising , @xcite .
while for quenched _ qcd _ we expect that the sea term is suppressed giving too large a quark contribution , this is not so for unquenched _
qcd_. in conclusion the results presented here would seem to indicate that we have to be much closer to the chiral limit in order to be able to perceive the partonic properties of the nucleon .
further details will be given in @xcite .
the numerical calculations were performed on the hitachi _ sr8000 _ at lrz ( munich ) , the cray _
t3e_s at epcc ( edinburgh ) , nic ( jlich ) and zib ( berlin ) as well as the ape / quadrics at nic ( zeuthen ) .
we wish to thank the _ ukqcd _ collaboration for sharing their unquenched configurations with us . | using unimproved and non - perturbatively @xmath0 improved wilson fermions , results are given for the three lowest moments of unpolarised nucleon structure functions .
renormalisation , chiral extrapolation and the continuum limit of the matrix elements are briefly discussed .
the simulations are performed for both quenched and two flavours of unquenched fermions .
no obvious sign of deviation from linearity in the chiral extrapolations are found .
( this is most clearly seen in our quenched unimproved data , which extends to lighter quark mass . )
possible quenching effects also seem to be small .
the lowest moment thus remains too large , so it seems to be necessary to reach smaller quark masses in numerical simulations .
15.0 true cm |
with numerous high precision telescopes and satellite missions operating already or becoming operational within the next decade einstein s theory of general relativity will be tested with increasing accuracy , in particular , in the strong gravity regime ( see , e.g. , the reviews @xcite ) . from a theoretical point of view general relativity
is expected to be superseded by a new theory of gravity , which on the one hand will not be suffering from those pathological singularities present in many solutions of general relativity ( gr ) , and , on the other hand , will be compatible with quantum theory .
string theory , as a leading candidate for a unified theory of the fundamental interactions and a quantum theory of gravity , for instance , predicts the presence of higher curvature corrections to einstein gravity as well as the existence of further fields .
the low energy effective theory deriving from heterotic string theory contains as basic ingredients a gauss - bonnet term and a dilaton field @xcite .
the effects of the presence of such terms on the properties of compact astrophysical objects can be strong , as has been shown for black holes @xcite and neutron stars @xcite , which have been studied in a truncated version of this theory , named dilatonic einstein - gauss - bonnet ( egbd ) theory .
moreover , egbd theory allows for wormhole solutions without the necessity of introducing exotic fields @xcite . in this letter
we focus on rapidly rotating compact objects in egbd theory and compare their properties with those of their einstein gravity counterparts . in particular , we here determine the quadrupole moments of rapidly rotating black holes and neutron stars .
black holes in egbd theory can slightly exceed the kerr bound , @xmath2 , given by the scale invariant ratio of the angular momentum @xmath3 and the square of the mass @xmath4 in appropriate units @xcite .
the innermost - stable - circular - orbits ( iscos ) can differ from the respective kerr values by a few percent in the case of slow rotation @xcite and up to 10% for rapid rotation @xcite .
similarly , the orbital frequencies exhibit the largest deviations from the kerr frequencies for rapid rotation , which can be as large as 60%
. such large effects might be observable for astrophysical black holes . for neutron stars , on the other hand
, recent calculations showed , that for slowly rotating neutron stars universal relations hold between the scaled moment of inertia @xmath0 , the love number , and the scaled quadrupole moment @xmath1 in einstein gravity @xcite .
subsequently , the @xmath0-@xmath1 relation was generalized to the case of rapidly rotating neutron stars , where for fixed rotation parameter @xmath5 again only little dependence on the equation of state was seen @xcite , after appropriately scaled quantities were considered @xcite . in the following we briefly recall egbd theory and obtain the quadrupole moment for compact objects in this theory .
we then discuss the quadrupole moments of rapidly rotating black holes and show that they may deviate from their kerr counterparts by 20% and more .
subsequently , we present rapidly rotating neutron stars in egbd theory and discuss their @xmath0-@xmath1 relation .
we conclude , that just like in einstein gravity the properly scaled dimensionless quantities exhibit little dependence on the equation of state .
as motivated by the low - energy heterotic string theory @xcite we consider the following low - energy effective action @xcite @xmath6 , \label{act}\end{aligned}\ ] ] where @xmath7 is newton s constant , @xmath8 is the speed of light , @xmath9 denotes the dilaton field with coupling constant @xmath10 , @xmath11 is a numerical coefficient given in terms of the regge slope parameter , and @xmath12 is the gauss - bonnet ( gb ) term .
we have chosen the dilaton coupling constant according to its string theory value , @xmath13 . to construct rotating compact objects we employ the lewis - papapetrou line element @xcite for a stationary , axially symmetric spacetime with two killing vector fields @xmath14 , @xmath15 . in terms of the spherical
coordinates @xmath16 and @xmath17 , the isotropic metric reads @xcite @xmath18 + r^2 \sin^2\theta \left(d\varphi-\omega c dt\right)^2\right ) , \label{metric}\ ] ] where @xmath19 , @xmath20 , @xmath21 and @xmath22 are functions of @xmath16 and @xmath17 only . in the asymptotic region
the metric and dilaton functions possess the expansion @xmath23 where @xmath24 is a legendre polynomial , @xmath4 is the mass , @xmath3 is the angular momentum , @xmath25 is the dilaton charge , and @xmath26 and @xmath27 are further expansion constants . to extract the quadrupole moment of the compact solutions , we follow geroch and hansen @xcite ( see also @xcite ) and obtain @xmath28 m^3 . \label{q}\ ] ] we note , that there is no explicit contribution from the gb term , since this term decays sufficiently fast .
the dilaton field enters only via the coulomb - like term @xmath29 , which coincides with the analogous term in einstein - maxwell theory to lowest order in the expansion .
thus the expressions for the quadrupole moment are completely analogous . in the vacuum limit , when the dilaton charge vanishes , the quadrupole moment @xmath30 reduces to the expression in @xcite ( up to an overall sign ) , with @xmath31 . on the other hand , in the static limit
the solution is spherically symmetric . in this case @xmath32 and @xmath33 , and the quadrupole moment vanishes .
the domain of existence of the egbd black hole solutions i.e. , @xmath34 . ]
resides within the boundaries formed by ( i ) the kerr black holes , ( ii ) the critical black holes , where a radicant vanishes @xcite , and ( iii ) the set of extremal black holes with @xmath35 , which have temperature @xmath36 , a regular metric , but a dilaton field that diverges on the horizon at the poles @xcite . in fig .
[ fig1]a the shaded area represents the domain of existence , with ( i ) the lower boundary , ( ii ) the upper boundary , and ( iii ) the right lower boundary in the inset . in fig .
[ fig1]a we present the scaled quadrupole moment @xmath37 of the rotating black hole solutions versus the scaled angular momentum @xmath5 for several values of the dimensionless horizon angular velocity @xmath38 .
the curves are obtained , by fixing the value of @xmath39 and the value of @xmath40 and by varying the size of the black hole . for fixed @xmath40
the range of the scaled area @xmath41 for egbd black holes is largest in the static case and decreases to zero for the maximal value of @xmath42 @xcite .
this is reflected in the range of the scaled quadrupole moment @xmath43 , which is likewise largest in the limit of slow rotation , and decreases with increasing @xmath42 .
thus for slow rotation deviations from the kerr value of up to 20% and more are seen .
moreover , the solutions with @xmath35 , not present in general relativity ( gr ) , always have @xmath44 . in fig .
[ fig1]b we exhibit the scaled moment of inertia @xmath45 versus the scaled quadrupole moment @xmath43 for fixed values of @xmath42 . for @xmath46 the families of solutions with fixed @xmath42 start at the respective kerr values , @xmath47 and @xmath48 , indicated by the dots on the axis .
they end at the boundary ( ii ) given by the critical solutions and represented by the dotted curve .
the inset again shows the region @xmath35 .
the figure also contains the perturbative results derived in @xcite for small @xmath40 and small @xmath42 .
clearly , the @xmath0-@xmath1 relation for black holes depends on the value of the scaled angular momentum @xmath42 .
however , for black holes there is no dependence of this relation on the coupling parameter @xmath40 .
this is a consequence of the invariance of @xmath49 , @xmath43 , @xmath42 and the egbd equations under the scaling transformation @xmath50 , @xmath51 , and @xmath52 .
the absolute value of the scaled dilaton charge @xmath53 decreases monotonically with decreasing scaled quadrupole moment @xmath1 . in the limit @xmath54 the scaled dilaton charge @xmath53 vanishes .
therefore the scaled moment of inertia @xmath49 approaches @xmath55 in this limit .
its dependence on the value of the scaled angular momentum @xmath42 is only weak .
rapidly rotating neutron stars have been studied extensively in gr ( see , e.g. , @xcite ) . here
we address rapidly rotating neutron stars in egbd theory @xcite . considering rigid rotation
, the four - velocity has the form @xmath56 with angular velocity @xmath57 .
the normalization condition @xmath58 then yields @xmath59 the differential equations for the pressure @xmath60 and the energy density @xmath61 are obtained from the constraints @xmath62 , @xmath63 these equations have to be supplemented by an equation of state ( eos ) , @xmath64 ( or @xmath65 ) . for a polytropic eos , @xmath66 , with central pressure @xmath67 and central density @xmath68 .
substitution in eq .
( [ dpres ] ) yields @xmath69 , where @xmath70 is an integration constant , which we express as @xmath71 to obtain the more convenient expression @xmath72 the boundary of the star is defined by @xmath73 , implying @xmath74 and consequently @xmath75 . outside the star @xmath76 . in fig .
[ fig2]a we exhibit the quadrupole moment @xmath30 in units of @xmath77 versus the angular velocity @xmath57 in hz for a family of neutron stars for a fixed scaled angular momentum of @xmath78 , the gb couplings @xmath79 , 1 and 2 , and two well - known equations of state .
the first eos corresponds to a polytropic eos with @xmath80 , taken from @xcite and denoted by di - ii , which is widely used in neutron star physics and also employed for neutron stars in scalar - tensor theory @xcite , while the second eos corresponds to an approximation to the fps eos @xcite .
clearly , the two equations of state give rise to rather different quadrupole moments , whereas the dependence on the gb coupling @xmath40 is only moderate , while increasing with increasing @xmath40 . to address the @xmath0-@xmath1 relation for neutron stars we exhibit in fig .
[ fig2]b @xmath49 versus @xmath43 for a fixed value of the scaled angular momentum , @xmath78 .
again , we employ the polytropic eos di - ii @xcite and the analytical fit to the fps eos @xcite .
comparison shows , that for a given value of @xmath40 and a given value of @xmath42 the dependence on the eos is only weak , but increasing with increasing @xmath40 .
moreover , the @xmath40-dependence is the stronger the smaller @xmath42 .
we did not go beyond @xmath81 , since the strongest current observational upper bound on @xmath40 , obtained from low - mass x - ray binaries , is slightly less than this value ( in the units employed ) @xcite .
following the procedure of geroch and hansen @xcite we have derived the spin induced quadrupole moment for rapidly rotating compact objects in egbd theory .
the scaled quadrupole moment of black holes can deviate by 20% and more from the corresponding kerr value for a given value of the scaled angular momentum @xmath42 .
their @xmath0-@xmath1 relation has no deplicit dependence on the gauss - bonnet coupling @xmath40 . for neutron stars the presence of the matter breaks the scale invariance of the @xmath0-@xmath1 relation .
we now observe a dependence on @xmath40 , which is however weak in its allowed range . while the unscaled quantities depend quite significantly on the equation of state , the scaled quantities satisfy to good approximation universal relations , which depend only on the scaled angular momentum @xmath42 , in egbd theory like in gr .
we are currently extending the calculations to quark stars , to see whether the universal relations hold also in this case @xcite for egbd theory . unlike the case of chern - simons theory also studied in @xcite , however , the @xmath0-@xmath1 relations do not differ significantly for egbd theory and gr . here
other observables including iscos and orbital periods seem more promising handles to test the viability of egbd theory .
finally , another type of compact object can be obtained in egbd theory , namely wormholes , which may possibly correspond to black hole foils @xcite .
we expect that the static egbd wormholes @xcite can be generalized to rapidly rotating wormholes , while retaining their linear stability with respect to changes of the size of their throat
. it will be interesting to extract their quadrupole moments and compare them with the quadrupole moments of the egbd black holes .
j. r. gair , m. vallisneri , s. l. larson and j. g. baker , living rev .
* 16 * , 7 ( 2013 ) . n. yunes and x. siemens , living rev .
* 16 * , 9 ( 2013 ) . c. m. will , living rev . rel .
* 17 * , 4 ( 2014 ) .
d. j. gross , j. h. sloan , nucl .
* b291 * , 41 ( 1987 ) .
p. kanti , n. e. mavromatos , j. rizos , k. tamvakis and e. winstanley , phys .
d * 54 * , 5049 ( 1996 ) .
p. pani and v. cardoso , phys
. rev .
d * 79 * , 084031 ( 2009 ) .
b. kleihaus , j. kunz and e. radu , phys .
lett . *
106 * , 151104 ( 2011 ) .
pani , c. f. b. macedo , l. c. b. crispino and v. cardoso , phys .
d * 84 * , 087501 ( 2011 ) .
d. ayzenberg and n. yunes , arxiv:1405.2133 [ gr - qc ] .
p. pani , e. berti , v. cardoso and j. read , phys .
d * 84 * , 104035 ( 2011 ) .
p. kanti , b. kleihaus and j. kunz , phys .
* 107 * , 271101 ( 2011 ) .
p. kanti , b. kleihaus and j. kunz , phys .
d * 85 * , 044007 ( 2012 ) .
k. yagi and n. yunes , science * 341 * , 365 ( 2013 ) .
k. yagi , k. kyutoku , g. pappas , n. yunes and t. a. apostolatos , phys .
d * 89 * , 124013 ( 2014 ) .
k. yagi , l. c. stein , g. pappas , n. yunes and t. a. apostolatos , arxiv:1406.7587 [ gr - qc ] .
s. chakrabarti , t. delsate , n. grlebeck and j. steinhoff , phys .
lett . * 112 * , 201102 ( 2014 ) .
d. d. doneva , s. s. yazadjiev , n. stergioulas and k. d. kokkotas , astrophys . j. * 781 * , l6 ( 2013 ) .
s. mignemi , n. r. stewart , phys .
d * 47 * , 5259 ( 1993 ) . c. m. chen , d. v. galtsov , n. ohta , d. g. orlov , phys .
* d81 * , 024002 ( 2010 ) .
r. m. wald , `` general relativity , '' ( university of chicago press , chicago , 1984 ) t. p. sotiriou and t. a. apostolatos , class .
* 21 * , 5727 ( 2004 ) .
g. pappas and t. a. apostolatos , phys .
lett . * 108 * , 231104 ( 2012 ) .
n. stergioulas , living rev .
* 6 * , 3 ( 2003 ) .
friedman , n. stergioulas , _ rotating relativistic stars _ ( cambridge university press , 2013 ) d. d. doneva , s. s. yazadjiev , n. stergioulas and k. d. kokkotas , phys .
d * 88 * , no .
8 , 084060 ( 2013 ) .
p. haensel and a. y. potekhin , astron
. astrophys .
* 428 * , 191 ( 2004 ) .
k. yagi , phys .
d * 86 * , 081504 ( 2012 ) .
t. damour and s. n. solodukhin , phys .
d * 76 * , 024016 ( 2007 ) . | we consider rapidly rotating black holes and neutron stars in dilatonic einstein - gauss - bonnet ( egbd ) theory and determine their quadrupole moments , which receive a contribution from the dilaton .
the quadrupole moment of egbd black holes can be considerably larger than the kerr value . for neutron stars , the universality property of the @xmath0-@xmath1 relation between the scaled moment of inertia and
the scaled quadrupole moment appears to extend to egbd theory . |
we consider the classical lagrange function interpolation problem in the following `` discrete '' setting .
let @xmath0 be a given set of mutually distinct interpolation nodes , and let @xmath1 be a given set of `` basis '' functions .
for any given function @xmath2 we seek a linear combination of the basis functions that interpolates @xmath2 at all interpolation nodes , @xmath3 the coefficients @xmath4 can be computed by solving the linear system @xmath5 where @xmath6 is the so - called _ collocation matrix _ containing the values of the basis functions at the interpolation nodes @xmath7 if the basis functions are linearly independent on @xmath8 , the matrix @xmath6 is nonsingular , and the interpolation problem has a unique solution for all functions @xmath2 .
the sensitivity of the solution is then determined by the condition number @xmath9 of the collocation matrix @xmath10 where @xmath11 denotes the standard operator @xmath12-norm , with @xmath13 .
the polynomial b - splines of a fixed degree @xmath14 are frequently used as the basis functions in practice . such a basis is uniquely determined by a given multiset of knots that defines the local smoothness of the basis functions .
the corresponding collocation matrix is always totally nonnegative ( see @xcite ) , regardless of the choice of the interpolation nodes .
a particular choice of nodes is of special interest , both in theory and in practice , for shape preserving approximation .
when the nodes are located at the so - called greville sites , i.e. , at the _ knot averages _ , the interpolant has the variation diminishning property . moreover , for low order b - spline interpolation , it can be shown that @xmath15 is bounded _ independently _ of the knot sequence @xcite . on these grounds , in 1975 , carl de boor @xcite conjectured that the interpolation by b - splines of degree @xmath14 at knot averages is bounded by a function that depends only on @xmath14 , regardless of the knots themselves . in our terms , the conjecture says that @xmath15 or , equivalently , @xmath16 is bounded by a function of @xmath14 only .
the conjecture was disproved by rong
qing jia @xcite in 1988 .
he proved that , for geometric meshes , the condition number @xmath15 is not bounded independently of the knot sequence , for degrees @xmath17 .
therefore , it is a natural question whether exists any class of meshes for which de boor s conjecture is valid . since geometric meshes are highly nonuniform , the most likely candidates for the validity are uniform meshes . here
we discuss the problem of interpolation at knot averages by b - splines with _ equidistant simple _ knots .
the corresponding b - splines are symmetric on the support , and have the highest possible smoothness .
it is easy to see that the condition of this interpolation does not depend on the knot spacing @xmath18 , and we can take @xmath19 .
so , just for simplicity , we shall consider only the cardinal b - splines , i.e. , b - splines with simple knots placed at successive integers .
it should be stressed that the only free parameters in this problem are the _ degree _ @xmath14 of the b - splines and the _ size _
@xmath20 of the interpolation problem .
our aim is to prove that the condition of @xmath6 can be bounded independently of its order .
the corresponding collocation matrices @xmath6 are symmetric , positive definite , and most importantly , tplitz .
but , it is not easy to compute the elements of @xmath21 , or even reasonably sharp estimates of their magnitudes .
so , the natural choice of norm in ( [ 1.1 ] ) is the spectral norm @xmath22 where @xmath23 , and @xmath24 denote the largest and the smallest singular value of @xmath6 , respectively . for low spline degrees @xmath25 ,
the collocation matrices are also strictly diagonally dominant , and it is easy to bound @xmath26 by a constant .
consequently , de boor s conjecture is valid for @xmath25 . for higher degrees ,
the condition number can be estimated by embedding the tplitz matrix @xmath6 into circulant matrices of higher orders .
the main advantage of this technique , developed by davis @xcite and arbenz @xcite , lies in the fact that the eigenvalues of a circulant matrix are easily computable .
the final bounds for @xmath26 are obtained by using the cauchy interlace theorem for singular values ( see @xcite for details ) , to bound both singular values in ( [ 1.2 ] ) .
the paper is organized as follows . in section 2
we briefly review some basic properties of cardinal b - splines .
the proof of de boor s conjecture for low degree ( @xmath25 ) cardinal b - splines is given in section 3 . in section 4
we describe the embedding technique and derive the estimates for @xmath26 . despite all efforts , we are unable to prove de boor s conjecture in this , quite probably , the easiest case .
the final section contains the results of numerical testing that strongly support the validity of the conjecture , as well as some additional conjectures based on these test results .
let @xmath27 , for @xmath28 , be a sequence of simple uniformly spaced knots .
this sequence determines a unique sequence of normalized b - splines @xmath29 of degree @xmath14 , such that the spline @xmath30 is non - trivial only on the interval @xmath31 .
each of these b - splines can obtained , by translation and scaling , from the basic b - spline @xmath32 with knots @xmath33 , @xmath34 here , @xmath35 denotes the truncated powers @xmath36 , for @xmath37 , while @xmath38 the normalized version of the basic spline is defined as @xmath39 from ( [ 2.1 ] ) and ( [ 2.2 ] ) , we obtain the _ normalized _ b - spline basis @xmath40 : @xmath41 if the interpolation nodes @xmath42 are located at the knot averages , i.e. , @xmath43 then @xmath44 and the schnberg whitney theorem @xcite guarantees that the collocation matrix is nonsingular
. moreover , this matrix is totally nonnegative @xcite , i.e. , all of its minors are nonnegative . due to symmetry of b - splines on uniform meshes ,
the collocation matrices are also symmetric and tplitz .
so , we can conclude that cardinal b - spline collocation matrix is tplitz , symmetric and positive definite .
it is easy to show that the elements of the collocation matrix do not depend on the step - size @xmath18 of the uniform mesh , so we take the simplest one with @xmath19 and @xmath45 .
such b - splines are called cardinal . the interpolation nodes ( [ 2.3 ] )
are integers for odd degrees , while for even degrees , the interpolation nodes are in the middle of the two neighbouring knots of the cardinal b - spline . of degree @xmath46 and @xmath47 , respectively .
big black dots denote the spline values at the knot averages.,title="fig:",width=217 ] of degree @xmath46 and @xmath47 , respectively .
big black dots denote the spline values at the knot averages.,title="fig:",width=217 ] the normalized basic cardinal spline @xmath48 suffices to determine all basis function values at the interpolation nodes @xmath49 the general de boor
cox reccurence relation @xcite , written in terms of the degree of a spline is : @xmath50 note that the elements of a collocation matrix are rational , because the interpolation nodes are rational , and the de boor
cox recurrence formula ( [ 2.4 ] ) involves only basic arithmetic operations on rational coefficients .
these elements are therefore exactly computable in ( arbitrary precise ) rational arithmetic .
let @xmath51 denote the values of cardinal b - splines at knot averages ( see ( [ 2.3 ] ) ) .
then , the cardinal b - spline collocation matrix @xmath6 with interpolation nodes @xmath52 is a banded tplitz matrix of order @xmath53 , to be denoted by @xmath54 the matrix @xmath55 is represented by its first row , usually called the symbol , @xmath56 it is useful to note that each b - spline of degree @xmath37 is a unimodal function , i.e. , it has only one local maximum on the support . in the case of cardinal b - splines , we have already concluded that the splines are symmetric , and therefore the maximum values of @xmath48 is attained at the middle of the support , for @xmath57 .
the maximum value is @xmath58 furthermore , unimodality implies that the values of the spline @xmath48 are decreasing in the interval @xmath59 $ ] , so @xmath60 to esitmate the condition number of a cardinal b - spline we need to bound both the minimal and the maximal singular value of @xmath55 . for a symmetric and positive definite matrix ,
the singular values are eigenvalues .
therefore , the bounds for the eigenvalues of @xmath55 are sought for . from the gergorin bound for the eigenvalues , and the partition of unity of the b - spline basis
, we obtain an upper bound for @xmath61 @xmath62 similarly , we also obtain a lower bound for @xmath63 , @xmath64 which is sensible only if @xmath55 is strictly diagonally dominant .
strict diagonal dominance is achieved only for b - spline degrees @xmath65 ( easily verifiable by a computer ) .
the corresponding gergorin bounds are presented in table [ tbl3.1 ] .
this directly proves de boor s conjecture for low order b - splines .
ciccccc @xmath66 & 2 & 3 & 4 & 5 & 6 + 64 & 1.998136 & 2.994873 & 4.785918 & 7.466648 & 11.727897 + 128 & 1.999541 & 2.998757 & 4.796641 & 7.492176 & 11.785901 + 256 & 1.999886 & 2.999694 & 4.799180 & 7.498105 & 11.799106 + 512 & 1.999971 & 2.999924 & 4.799797 & 7.499534 & 11.802256 + 1024 & 1.999993 & 2.999981 & 4.799950 & 7.499884 &
11.803026 + 2048 & 1.999998 & 2.999995 & 4.799987 & 7.499971 & 11.803216 + @xmath67 & 2 & 3 & @xmath68 & 10 & @xmath69 + note that in the case of tridiagonal tplitz matrices , i.e. for @xmath70 , and , thus , @xmath71 in ( [ 3.1 ] ) , the exact eigenvalues are also known ( see bttcher
grudsky @xcite ) @xmath72 the largest and the smallest eigenvalue can then be uniformly bounded by @xmath73 these uniform bounds are somewhat better than those obtained by the gergorin circles .
when the degree of a cardinal b - spline is at least @xmath74 , the eigenvalue bounds for tplitz matrices can be computed by circulant embeddings .
first , we will introduce the smallest possible circulant embedding , and give its properties .
then we will present some other known embeddings , with positive semidefinite circulants . to obtain a bound for @xmath63 ,
the collocation matrix @xmath55 is to be embedded into a circulant @xmath75 , \quad m = n - d + r. \label{4.1}\ ] ] it is obviously a tplitz matrix with the following symbol @xmath76 this circulant @xmath77 is called a periodization of @xmath55 by bttcher and grudsky @xcite
. the bounds ( [ 3.3])([3.4 ] ) for the eigenvalues of @xmath55 are also valid for @xmath77 .
moreover , @xmath77 is doubly stochastic , always having @xmath78 as its largest eigenvalue . interestingly enough
, the upper bound ( [ 3.3 ] ) is attained here ( the gergorin bounds are rarely so sharp ) .
the symmetry of @xmath55 immediately implies the symmetry of @xmath77 , and we can conclude that the eigenvalues of @xmath77 are real , but not necessarily positive . for symmetric matrices ,
the singular values are , up to a sign , equal to the eigenvalues , so @xmath79 if the eigenvalues of the circulant @xmath77 are known , the spectrum of embedded @xmath55 can be bounded by the cauchy interlace theorem for singular values , applied to @xmath77 .
let @xmath80 be given , and let @xmath81 denote a submatrix of @xmath82 obtained by deleting a total of @xmath83 rows and/or @xmath83 columns of @xmath82 .
then @xmath84 where we set @xmath85 if @xmath86 .
the proof can be found , for example , in @xcite .
if we delete the last @xmath87 rows and columns of @xmath77 , we obtain @xmath55 .
the cauchy interlace theorem will then give useful bounds for @xmath88 , provided that @xmath77 is nonsingular .
moreover , if we delete more than @xmath87 last rows and columns of @xmath77 , we obtain bounds for tplitz matrices @xmath89 , of order @xmath90 , for @xmath91 , @xmath92 now we need to calculate the smallest singular value of @xmath77 , and show that it is non - zero .
the eigendecomposition of a circulant matrix is well - known ( see @xcite ) .
a circulant @xmath82 of order @xmath93 , defined by the symbol @xmath94 , can be written as @xmath95 where @xmath96 & \smash{\ddots } & \smash{\ddots } & \\ & & \smash{\ddots } & \;\ ; 1 \;\ ; \\ \ ; 1 \;\ ; & & & \;\ ; 0 \;\ ; \\ \end{bmatrix}.\ ] ] the spectral decomposition of @xmath97 is @xmath98 , where @xmath99 while @xmath100 hence , @xmath82 can be decomposed as @xmath101 the eigenvalues of a real symmetric circulant @xmath82 are real , and given by @xmath102 they can also be viewed as the discrete fourier transform ( dft ) of the symbol @xmath94 . for real and symmetric @xmath82 ,
i.e. , when @xmath103 , for @xmath104 , from ( [ 4.4 ] ) it also follows that @xmath105 so , all the eigenvalues , except @xmath106 , and possibly @xmath107 , for even @xmath93 , are multiple . therefore , the eigenvalues of the circulant @xmath77 from ( [ 4.1 ] ) are @xmath108 for prime orders
@xmath93 , the nonsingularity of @xmath77 is a consequence of the following theorem from @xcite .
let @xmath93 be a prime number .
assume that the circulant @xmath82 of order @xmath93 has entries in @xmath109 .
then @xmath110 if and only if @xmath111 or all the symbol entries @xmath112 are equal .
[ thmgkps ] if @xmath93 is prime , then we must have @xmath113 , since ( [ 3.2 ] ) implies that @xmath112 s are not equal , and from ( [ 4.6 ] ) we get @xmath114 theorem [ thmgkps ] suggests how to get the nonsingular embedding of @xmath55 .
first , @xmath55 should be embedded into the tplitz matrix @xmath115 , of order @xmath116 , where @xmath117 is chosen so that @xmath118 is a prime number .
then , @xmath115 is embedded into the circulant @xmath77 .
the other possibility is to embed @xmath55 into the smallest circulant matrix @xmath77 , as in ( [ 4.1 ] ) , and calculate its eigenvalues from ( [ 4.6 ] ) , in hope that @xmath77 is nonsingular . in this case ,
extensive numerical testing suggests that @xmath77 is always positive definite , but we have not been able to prove it .
there are also several other possible embeddings that guarantee the positive semidefiniteness of the circulant matrix @xmath82 .
the first one , constructed by dembo , mallows and shepp in @xcite , ensures that the positive definite tplitz matrix @xmath119 , of order @xmath120 , can be embedded in the positive semidefinite circulant @xmath82 , of order @xmath93 , where @xmath121 a few years later , newsam and dietrich @xcite reduced the size of the embedding to @xmath122 note that among all positive semidefinite matrices @xmath82 of order greater or equal @xmath93 , we can choose one of prime order .
this embedding will be positive definite according to theorem [ thmgkps ] .
it is obvious that embeddings ( [ 4.7])([4.8 ] ) are bounded by a function of the condition number of @xmath119 , i.e. , the quantity which we are trying to bound .
ferreira in @xcite embeds a tplitz matrix @xmath119 of order @xmath120 , defined by the symbol @xmath123 , into the circulant @xmath82 of order @xmath124 , @xmath125 where the symbol of the tplitz matrix @xmath126 is @xmath127 .
if we take @xmath128 from ( [ 3.1 ] ) , the only difference between embeddings ( [ 4.1 ] ) and ( [ 4.9 ] ) is in exactly @xmath129 zero diagonals , added as the first diagonals of @xmath126 . a sufficient condition for positive semidefiniteness of @xmath82
is given by the next result .
let @xmath82 be defined as in ( [ 4.9 ] ) , and let @xmath130 $ ] , @xmath131 $ ] .
if @xmath119 is positive definite , and @xmath132 , then @xmath82 is positive semidefinite .
once again , there is no obvious efficient way to verify whether the condition @xmath132 is fullfiled or not .
extensive numerical testing has been conducted , by using _ mathematica _ 7 from wolfram research , for the symbolic , arbitrary - precision rational , and machine - precision floating - point computations .
whenever feasible , the full accuracy was maintained .
owing mostly to the elegance and the accuracy of these results , insight into and the following conjecture about the spectral properties of the collocation matrices and the corresponding periodizations were obtained . for spline of degree @xmath137 with @xmath138 , respectively .
the associated circulants have order @xmath139 , i.e. , @xmath140 and @xmath141 .
note that for @xmath142 there is only one minimal eigenvalue , while for @xmath143 we have two minimal eigenvalues.,title="fig:",width=217 ] for spline of degree @xmath137 with @xmath138 , respectively .
the associated circulants have order @xmath139 , i.e. , @xmath140 and @xmath141 .
note that for @xmath142 there is only one minimal eigenvalue , while for @xmath143 we have two minimal eigenvalues.,title="fig:",width=217 ] for even @xmath93 , @xmath144 ( and , therefore , @xmath145 ) depends solely on @xmath14 , i.e. , the order @xmath93 of a circulant is irrelevant here .
moreover , for @xmath93 odd and even alike , the limiting value of @xmath144 is the same : @xmath146 hence , the notation @xmath147 is justified , since that value is determined uniquely by the degree @xmath14 of the chosen cardinal splines .
this is consistent with de boor s conjecture . the equations ( [ 5.1 ] ) and ( [ 5.2 ] )
provide us with efficiently and exactly computable estimates of the spectral condition numbers of large collocation matrices @xmath55 .
as demonstrated in figure [ fig5.2 ] and table [ tbl5.1 ] , the smallest eigenvalues of the collocation matrices converge rapidly and monotonically to the smallest eigenvalues of the corresponding circulant periodizations @xmath77 , as well as to the limiting value ( [ 5.2 ] ) .
it is worth noting that the spectral bounds obtained in such a way for lower degrees ( @xmath150 ) of cardinal b - splines are quite sharper than those established by the gergorin circle theorem ( cf .
table [ tbl3.1 ] and table [ tbl5.1 ] ) , at no additional cost .
cicciccicc @xmath66 & @xmath151 & @xmath152 & @xmath153 & @xmath154 & @xmath155 & @xmath156 + 64 & 1.998137 & 1.998758 & 7.466648 & 7.472749 & 11.72790 & 11.74214 + 128 & 1.999541 & 1.999694 & 7.492176 & 7.493492 & 11.78590 & 11.78866 + 256 & 1.999886 & 1.999924 & 7.498105 & 7.498410 & 11.79911 & 11.79971 + 512 & 1.999971 & 1.999981 & 7.499534 & 7.499607 & 11.80226 & 11.80240 + 1024 & 1.999993 & 1.999995 & 7.499884 & 7.499902 & 11.80303 & 11.80306 + 2048 & 1.999998 & 1.999999 & 7.499971 & 7.499975 & 11.80322 &
11.80322 + @xmath157 & & 2.000000 & & 7.500000 & & 11.80328 + cicciccicc @xmath66 & @xmath158 & @xmath148 & @xmath159 & @xmath160 & @xmath161 & @xmath162 + 64 & 45.04067 & 45.17179 & 9012.21 & 9543.49 & 371000.6 & 502472.1 + 128 & 45.57648 & 45.59721 & 10100.96 & 10150.47 & 569223.5 & 579852.3 + 256 & 45.69092 & 45.69486 & 10273.67 & 10279.58 & 594976.6 & 596037.0 + 512 & 45.71737 & 45.71822 & 10308.14 & 10309.00 & 599497.1 & 599628.0 + 1024 & 45.72373 & 45.72393 & 10315.86 & 10316.01 & 600450.4 & 600469.5 + 2048 & 45.72529 & 45.72534 & 10317.69 & 10317.72 & 600669.7 &
600673.0 + @xmath157 & & 45.72581 & & 10318.28 & & 600739.5 + since @xmath163 are rational numbers , ( [ 5.2 ] ) is useful for the exact computation of @xmath147 .
but , in floating - point arithmetic , the direct computation of @xmath147 from ( [ 5.2 ] ) is numerically unstable , as it certainly leads to severe cancellation .
it can be easily shown from ( [ 2.1 ] ) or ( [ 2.4 ] ) that the smallest non - zero value of the cardinal b - spline of degree @xmath14 at an interpolation node is : @xmath164 n^d\left(\frac{1}{2}\right ) = \frac{1}{2^d \cdot d ! } , & \hbox{for even $ d$. } \end{cases}\ ] ] moreover , all other values @xmath163 in ( [ 5.2 ] ) and , consequently , @xmath147 are integer multiples of @xmath165 . with that in mind , yet another , somewhat surprising conjecture emerged from the test results : @xmath166 t_r^d \cdot 2^d e_d = \frac{1}{d ! } e_d , & \hbox{$d$ even , } \end{cases } \label{5.3}\ ] ] where , as in @xcite , @xmath167 are the tangent numbers , and @xmath168 are the euler numbers , defined by the taylor expansions of @xmath169 and @xmath170 , respectively , @xmath171 these numbers are also related to the sequences a000182 ( the tangent or `` zag '' numbers ) , a000364 ( the euler or `` zig '' numbers ) and a002436 , from @xcite .
if true , ( [ 5.3 ] ) would be of significant practical merit , for there exist very stable and elegant algorithms for calculation of @xmath167 and @xmath168 by knuth and buckholtz @xcite .
so , it deserved an effort to find the proof .
a unifying framework for handling both cases is provided by the euler polynomials @xmath172 , defined by the following exponential generating function ( see ( * ? ? ?
* , p. 804 ) ) @xmath173 which is valid for @xmath174 .
first , note that @xmath175 , for all @xmath176 .
the remaining nontrivial values can be expressed in terms of special values of euler polynomials .
for the tangent numbers , we have @xmath177 this follows easily , by comparing the taylor expansion of @xmath178 @xmath179 and ( [ 5.4 ] ) , with @xmath180 and @xmath181 , instead of @xmath182 .
similarly , by comparing the taylor expansion of @xmath183 @xmath184 and ( [ 5.4 ] ) , with @xmath185 and @xmath181 , instead of @xmath182 , we get @xmath186 consider the function @xmath190 defined by @xmath191 from ( [ 5.4 ] ) with @xmath192 , the taylor expansion of @xmath190 can be written as @xmath193 \frac{t^n}{n!},\ ] ] so @xmath194 on the other hand , the leibniz rule gives @xmath195 \
, d^{n - m } \left [ \frac{2}{e^t + 1 } \right].\ ] ] if @xmath196 , then @xmath197 $ ] is always divisible by @xmath198 .
hence , @xmath199 which proves the first identity ( [ 5.7 ] ) .
the second one follows similarly , by considering @xmath200 the taylor expansion of @xmath201 is then given by @xmath202 \frac{t^n}{n!}.\ ] ] if @xmath196 , from the first part of the proof , it follows immediately that @xmath203 which proves ( [ 5.8 ] ) . to simplify the notation ,
let @xmath205 .
due to the symmetry of interpolation nodes , the sum in ( [ 5.2 ] ) can be written as @xmath206 where @xmath207 . from ( [ 2.1 ] ) and ( [ 2.2 ] ) , it follows that @xmath208 then @xmath209 let @xmath14 be odd , @xmath210 , with @xmath176 .
then @xmath211 and @xmath212 , so ( [ 5.9 ] ) becomes @xmath213 from the definition of truncated powers with positive exponents , the second sum contains only the terms with @xmath214 , i.e. , for @xmath215 . by changing the order of summation , we get @xmath216 then we shift @xmath217 by @xmath218 , so that @xmath217 starts at @xmath219 , to obtain @xmath220 the second sum can be simplified as ( see ( * ? ? ?
* , p. 804 ) ) @xmath221 hence @xmath222.\ ] ] by reversing the summation , from ( [ 5.7 ] ) with @xmath210 and @xmath223 , we conclude that @xmath224 since @xmath225 , by using ( [ 5.5 ] ) , we have @xmath226 this proves the claim for odd values of @xmath14 .
let @xmath14 be even , @xmath227 , with @xmath176 . for @xmath228
, it is obvious that @xmath229 , so we may assume that @xmath230 .
then @xmath211 and @xmath231 , so ( [ 5.9 ] ) becomes @xmath232 the second sum contains only the terms with @xmath233 , i.e. , for @xmath215 . by exactly the same transformation as
before , we arrive at @xmath234 now we expand the last factor in terms of powers of @xmath217 . then @xmath235 can be written as @xmath236 with @xmath237 like before , the second sum can be simplified as @xmath238 which gives @xmath239.\ ] ] by reversing the summation , from ( [ 5.8 ] ) with @xmath227 , for @xmath189 , we see that @xmath240 therefore , @xmath241 from ( [ 5.10 ] ) we obtain @xmath242 finally , by using ( * ? ? ? * , p. 804 ) ) @xmath243 together with ( [ 5.6 ] ) , this gives @xmath244 this completes the proof for even values of @xmath14 .
we would like to conclude with an observation that , to the best of our knowledge , scarcely any result could be found about sufficient conditions for the non - negativeness of the dft in terms of its coefficients , apart from the classical result of young and kolmogorov ( cited in zygmund @xcite ) : convexity is not fulfilled in the case of cardinal b - spline coefficients , since there is always one inflection point on each slope of the spline . and
yet , our numerical experiments strongly suggest that the class of series with a positive dft is worth investigating further , for the theoretical and practical reasons alike .
, _ localization of the eigenvalues of the toeplitz matrices using additive decomposition , embedding in circulants , and the fourier transform _ , in 10th ifac symposium on system identification , sysid94 , copenhagen , danmark , july 1994 , 1994 , pp . | the famous de boor conjecture states that the condition of the polynomial b - spline collocation matrix at the knot averages is bounded independently of the knot sequence , i.e. , depends only on the spline degree . for highly nonuniform knot meshes , like geometric meshes , the conjecture is known to be false . as an effort towards finding an answer for uniform meshes ,
we investigate the spectral condition number of cardinal b - spline collocation matrices .
numerical testing strongly suggests that the conjecture is true for cardinal b - splines .
* keywords * : cardinal splines , collocation matrices , condition , tplitz matrices , circulants * ams subject classifications * : 65d07 , 65d05 , 65f35 , 15a12 |
the @xmath11 reaction is one of the critical links in the @xmath12 and @xmath13 branches of the @xmath14chain of solar hydrogen burning [ 13 ] .
the total capture rate determined by processes of this chain is sensitive to the cross section @xmath15 ( or the astrophysical @xmath0 factor @xmath6 ) for the @xmath16 reaction and predicted neutrino rate varies as @xmath17^{0.8}$ ] @xcite . despite the impressive improvements in our understanding of the @xmath11 reaction made in the past decades ( see refs [ 411 ] and references therein ) , however
, some ambiguities connected with both the extrapolation of the measured cross sections for the aforesaid reaction to the solar energy region and the theoretical predictions for @xmath15 ( or @xmath6 ) still exist and they may influence the predictions of the standard solar model @xcite
. experimentally , there are two types of data for the @xmath16 reaction at extremely low energies : i ) six measurements based on detecting of @xmath18-rays capture ( see @xcite and references therein ) from which the astrophysical @xmath0 factor @xmath19 extracted by the authors of those works changes within the range [email protected] @xmath21 and ii ) six measurements based on detecting of @xmath12 ( see @xcite references therein as well as [ 6 - 11 ] ) from which @xmath19 extracted by the authors of these works changes within the range [email protected] @xmath23 . all of these measured data have a similar energy dependence for the astrophysical @xmath0 factors @xmath6 .
nevertheless , the adaptation of the available energy dependencies predicted in @xcite for the extrapolation of each of the measured data to low experimentally inaccessible energy regions , including @xmath8=0 , leads to a value of @xmath19 that differs from others and this difference exceeds the experimental uncertainty .
the theoretical calculations of @xmath19 performed within different methods also show considerable spread [ 12,1419 ] and the result depends on a specific model used .
for example , the resonating - group method calculations of @xmath19 performed in ref.@xcite show considerable sensitivity to the form of the effective nn interaction used and the estimates have been obtained within the range of 0.312@xmath20 0.841 @xmath23 .
the estimation of @xmath19=0.52@xmath24 0.03 @xmath25 @xcite also should be noted .
the latter has been obtained from the analysis of the experimental astrophysical @xmath0 factors @xcite , which were performed within the framework of the standard two - body potential model in the assumption that the dominant contribution to the peripheral reaction comes from the surface and external regions of the nucleus @xmath12 @xcite . at this , in @xcite the contribution from the nuclear interior ( @xmath26 , @xmath27=4 fm ) to the amplitude is ignored . in this case , the astrophysical @xmath0 factor is directly expressed in terms of the nuclear vertex constants ( nvc ) for the virtual decays @xmath28 ( or respective the asymptotic normalization coefficients ( anc ) for @xmath29 ) @xcite . as a result , in ref .
@xcite , the nvc - values for the virtual decays @xmath30 and @xmath31 were obtained , which were then used for calculations of the astrophysical @xmath0 factors at @xmath32180 kev , including @xmath8=0 .
however , the values of the ancs ( or nvcs ) @xmath33 and the @xmath19 obtained in @xcite may not be enough accurate associated both with the aforesaid assumption in respect to the contribution from the nuclear interior ( @xmath26 ) and with a presence of the spread in the experimental data @xcite used for the analysis . as far available values of these ancs obtained in @xcite
, they depend noticeably on a specific model used .
therefore , determination of precise experimental values of the ancs for @xmath33(g.s . ) and @xmath29(0.429 mev ) is highly desirable since it has direct effects in the correct extrapolation of the @xmath34 astrophysical @xmath0 factor at solar energies .
recently , a modified two - body potential approach ( mtbpa ) was proposed in @xcite for the peripheral direct capture @xmath35 reaction , which is based on the idea proposed in paper @xcite that low - energy direct radiative captures of particle @xmath36 by light nuclei @xmath37 proceed mainly in regions well outside the range of the internuclear @xmath38 interactions .
one notes that in mtbpa the direct astrophysical @xmath0 factor is expressed in terms of anc for @xmath39 rather than through the spectroscopic factor for the nucleus @xmath40 in the ( @xmath41 ) configuration as it is made within the standard two - body potential method @xcite . in refs.@xcite ,
mtbpa has been successfully applied to the radiative proton and @xmath9-particle capture by some light nuclei .
therefore , it is of great interest to apply the mtbpa for analysis of the @xmath42 reaction . in this work new analysis of the modern precise experimental astrophysical @xmath0 factors for the direct capture @xmath16 reaction at extremely low energies ( @xmath43 90 kev )
[ 6 - 11 ] is performed within the mtbpa @xcite to obtain @xmath3indirectly determined values both of the ancs ( the nvcs ) for @xmath4(g.s . ) and @xmath5(0.429 mev ) , and of @xmath6 at @xmath44 90 kev , including @xmath8=0 . in this work
we quantitatively show that the @xmath45 reaction within the aforesaid energy region is mainly peripheral and one can extract ancs for @xmath46 directly from the @xmath47 reaction where the ambiguities inherent for the standard two -body potential model calculation of the @xmath48 reaction , which is connected with the choice of the geometric parameters ( the radius @xmath49 and the diffuseness @xmath36 ) for the woods
saxon potential and the spectroscopic factors @xcite , can be reduced in the physically acceptable limit , being within the experimental errors for the @xmath6 .
the contents of this paper are as follows . in section 2 the results of the analysis of the precise measured astrophysical @xmath0 factors for the direct radiative
capture @xmath45 reaction is presented ( subsections 2.12.3 ) .
the conclusion is given in section 3 . in appendix
basic formulae of the modified two - body potential approach to the direct radiative capture @xmath16 reaction are given .
let us write @xmath50 ( @xmath51 ) for the relative orbital ( total ) angular moment of @xmath52 and @xmath9-particle in nucleus @xmath53 , @xmath54 ( @xmath55 ) for the orbital ( total ) angular moment of the relative motion of the colliding particles in the initial state . for the @xmath56 reaction populating to the ground and first excited ( @xmath57=0.429 mev ; @xmath58=1/2@xmath59 ) states of @xmath60 ,
the values of @xmath51 are taken to be equal to 3/2 and 1/2 , respectively , the value of @xmath50 is taken to be equal to 1 as well as @xmath61=0 , 2 for the @xmath62-transition and @xmath54=1 for the @xmath63-transition .
the basic formulae used for the analysis are presented in appendix . to determine the anc values for @xmath29(g.s ) and @xmath65(0.429 mev ) the recent experimental astrophysical @xmath0 factors , @xmath66 , for the @xmath34 reaction populating to the ground ( @xmath50=1 and @xmath51=3/2 ) and first excited ( @xmath57=0.429 @xmath67 ; @xmath68,@xmath50=1 and @xmath51=1/2 ) states [ 611 ] are reanalyzed based on the relations ( a1)(a7 ) .
the experimental data analyzed by us cover the energy ranges @xmath8=92.9168.9 kev [ 79 ] , 420951 kev @xcite , 3271235 kev @xcite and 7011203 kev @xcite for which only the external capture is substantially dominant @xcite .
also , in @xcite the experimental astrophysical @xmath0 factors for the the @xmath34 reaction populating to the first and excited states of the @xmath60 nucleus have been separated only for the energies of @xmath8=92.9 , 105.6 and 147.7 kev .
whereas , in @xcite the experimental astrophysical @xmath0 factors have been separated for all experimental points of @xmath8 from the aforesaid energy region by means of detecting the prompt @xmath18 ray ( the prompt method ) and by counting the @xmath60 activity ( the activation ) . the woods
saxon potential split with a parity ( @xmath69-dependence ) for the spin - orbital term proposed by the authors of refs .
[ 3032 ] is used here for the calculations of both bound state radial wave function @xmath70 and scattering wave function @xmath71 .
such the choice is based on the following considerations .
firstly , this potential form is justified from the microscopic point of view because it makes it possible to take into account the pauli principle between nucleons in @xmath52- and @xmath9-clusters in the ( @xmath72 ) bound state by means of inclusion of deeply bound states forbidden by the pauli exclusion principle .
the latter imitates the additional node ( @xmath73 ) arising in the wave functions of @xmath74 relative motion in @xmath12 similarly as the result of the microscopic resonating - group method @xcite .
secondly , this potential describes well the phase shifts for @xmath75-scattering in the wide energy range @xcite and reproduces the energies of low - lying states of the @xmath76 nucleus @xcite .
we vary the geometric parameters ( radius @xmath49 and diffuseness @xmath36 ) of the adopted woods
saxon potential in the physically acceptable ranges ( @xmath49 in 1.621.98 fm and @xmath36 in 0.630.77 fm @xcite ) in respect to the standard values ( @xmath49=1.80 fm and @xmath36=0.70 fm @xcite ) using the procedure of the depth adjusted to fit the binding energies . as
an illustration , fig.[fig1 ] shows plots of the @xmath77 dependence on the single - particle anc , @xmath78 for @xmath50= 1 and @xmath51=3/2 and 1/2 only for the two values of energy @xmath8 .
the width of the band for these curves is the result of the weak @xmath3residual@xmath79-dependence of the @xmath80 on the parameters @xmath49 and @xmath36 ( up to @xmath81 ) for the @xmath82 @xcite . the same dependence is also observed at other energies .
for example , for fig.1 plotted for @xmath8=0.1056 ( 0.1477 ) mev overall uncertainty ( @xmath83 ) of the function @xmath80 in respect to the central values @xmath77 corresponding to the central values of @xmath84 comes to @xmath83=@xmath24 4.5(@xmath244.5)% for the ground state of @xmath12 and @xmath83 = @xmath24 3.4(@xmath242.9)% for the excited state of @xmath12 . as it is seen from here that the @xmath11(0.429 mev ) reaction is slightly more peripheral than the @xmath16(g.s . )
reaction since the binging energy for @xmath12(0.429 mev ) is less than that for @xmath85(g.s ) .
the similar dependence of @xmath80 on the @xmath78 values is observed for the other aforesaid energies @xmath8 and the value of @xmath83 is no more than @xmath86 5.0% .
it follows from here that the condition ( a2 ) is satisfied for the considered reaction at the energies above mentioned within the uncertainties not exceeding the experimental errors of @xmath66 .
it should be noted that values of @xmath83 becomes larger as the energy @xmath8 increases ( for @xmath8 more than 1.3 mev ) .
thus , over the energy region 92.9@xmath7e@xmath7 1200 kev the @xmath87-dependence on @xmath88 is exactly sufficiently weak being within the experimental uncertainties for the @xmath66 .
such dependence is apparently associated also with the indirect taking into account the pauli principle mentioned above within the nuclear interior in the adopted nuclear @xmath89 potential leading as a whole to reduction of the contribution from the interior part of the radial matrix element into the @xmath87 function , which is typical for peripheral reactions .
we also calculated the @xmath90-elastic scattering phase shifts by variation of the parameters @xmath49 and @xmath36 in the same range for the adopted woods
saxon potential . as an illustration ,
the results of the calculations corresponding to the @xmath91 and @xmath92 waves are only presented in fig.[fig2 ] in which the width of the bands corresponds to a change of phase shifts values with respect to variation of values of the @xmath49 and @xmath36 parameters . as it is seen from fig.[fig2 ] ,
the experimental phase shifts @xcite are well reproduced within uncertainty of about @xmath24 5% .
the same results are also obtained for the @xmath93 and @xmath94 waves .
this circumstance allows us to test the condition ( a3 ) at the energies of @xmath8= 92.9 , 105.6 and 147.7 kev for which the @xmath16(g.s . ) and @xmath16(0.429 mev ) astrophysical @xmath0 factors were separately measured in @xcite . as an illustration , for the same energies @xmath8 as in fig.[fig1 ] we present in fig.[fig3 ] ( the upper panels ) the results of @xmath95-value calculation given by eq.(a3 ) ( @xmath96=(1@xmath97 3/2 ) and ( 1@xmath97 1/2 ) ) in which instead of the @xmath98 the experimental @xmath0 factors for the @xmath99 reaction populating to the ground and first excited states of @xmath12 were taken .
it should be noted that the same dependence occurs for other considered energies .
the calculation shows the obtained @xmath95 values also weakly ( up to 5.0 % ) depend on the @xmath78-value .
consequently , the @xmath11 reaction within the considered energy ranges is peripheral and a use of parametrization in terms of the ancs given by eq.(a1 ) is adequate to the physics of the reaction under consideration .
however , the values of the spectroscopic factors @xmath100 and @xmath101 corresponding to the @xmath102-configuration for @xmath12(g.s . ) and @xmath85(0.429 kev ) , respectively , change strongly about 1.7 times since calculated @xmath103 that vary by 1.75 times ( see , the lower panels in fig.[fig3 ] ) . for each energy
@xmath8 experimental point ( @xmath8=92.9 , 105.6 and 147.7 kev ) the values of the ancs are obtained for @xmath104 and @xmath105(0.429 mev ) by using the corresponding experimental astrophysical @xmath0 factor ( @xmath106 and @xmath107 , ( the activation ) ) @xcite in the ratio of the r.h.s . of the relation ( a1 ) instead of the @xmath98 and the central values of @xmath77 corresponding to the adopted values of the parameters @xmath49 and @xmath36 .
the results of the ancs , @xmath108 and @xmath109 , for these three energy @xmath8 experimental points are displayed in fig.[fig5 ] @xmath36 and @xmath110 ( filled circle symbols ) and the second and third columns of table [ table1 ] .
the uncertainties pointed in this figure correspond to those found from ( a1 ) ( averaged square errors ( a.s.e . ) ) , which include the total experimental errors ( a.s.e . from the statical and systematic uncertainties ) in the corresponding experimental astrophysical @xmath0 factor and the aforesaid uncertainty in the @xmath80 .
one should note that the same results for the ancs are obtained when @xmath111 ( @xmath112 and @xmath113 ) @xcite are used in eq.(a5 ) ( in eq.(a6 ) and ( a7 ) ) instead of @xmath6 ( @xmath114 and @xmath115 ) . then in eq.(a6 ) , inserting the averaged means of @xmath116 ( @xmath116=0.666 ) , obtained from the three data , and replacing of the @xmath117 in the l.h.s . of eq.(a5 ) with @xmath111 for the others , the two experimental points of energy @xmath8 ( @xmath8=126.5 and 168.9 kev ) from @xcite , four one @xmath8 ( @xmath8=420.0 , 506.0 , 615.0 and 951.0 kev ) from @xcite , the three one @xmath8 ( @xmath8=93.3 , 106.1 and 170.1 kev ) from @xcite and the ten one @xmath8 ( @xmath8=7011203 kev ) from @xcite can also determine values of ancs , @xmath118 and @xmath119 .
the results of the ancs @xmath120 for @xmath121(g.s . ) and @xmath122(0.429 mev ) are displayed in fig.[fig5 ] in which the open cycle and triangle symbols obtained from the analysis of the data of [ 69 ] as well as the filled triangles symbols obtained from the analysis of the data of @xcite(the @xmath60 recoils ) . besides
, the results obtained from the data of [ 79 ] are also presented in the second and third columns of table [ table1 ] . the same way the values of the ancs are obtained by using the separated experimental astrophysical @xmath0 factors ( @xmath123 and @xmath124 ) @xcite .
the results for these ancs are also presented in figs.[fig5]@xmath125 and @xmath126 both for the activation ( filled star symbols ) and for the prompt method ( filled square symbols ) . as it is seen from figs.[fig5 ] , for each of the independent measured experimental astrophysical @xmath0 factors the ratio in the r.h.s . of the relation ( a4 )
does not practically depend on the energy @xmath8 within the experimental uncertainties , although absolute values of the corresponding experimental astrophysical @xmath0 factors for the reactions under consideration depend noticeably on the energy and change by up to about 1.7 times when @xmath8 changes from 92.6 kev to 1200 kev .
this fact allows us to conclude that the energy dependence of the experimental astrophysical @xmath0 factors [ 611 ] is well determined by the calculated function @xmath80 and @xmath127 .
hence , the experimental astrophysical @xmath0 factors presented in [ 611 ] can be used as an independent source of reliable information about the ancs for @xmath121(g.s . ) and @xmath128(0.429 mev ) . also , in fig.[fig5 ] and table [ table2 ] the weighted means of the ancs - values and their uncertainties ( the solid line and the band width , respectively ) , derived both separately from each experimental data and from all of the experimental points , are presented . as it is seen also from the first and second ( fifth and sixth ) lines of table [ table2 ] the weighted means of the ancs - values for @xmath121(g.s . ) and @xmath128(0.429 mev ) obtained by the analysis performed separately for the activation , the prompt method and the @xmath60 recoils of the experimental data from the works [ 69 ] ( @xcite ) are in a good agreement with one another .
these results are the first ones of the present work .
nevertheless , the weighted means @xcite of the anc - values obtained by using separately the experimental data of the works refs.[69 ] and of the works refs.@xcite noticeably differ from one another ( up to 1.13 times for @xmath129(g.s . ) and up to 1.12 times for @xmath122(0.429 mev ) , see the parenthetical figures in table [ table2 ] ) .
the main reason of this difference is in the systematical discrepancy observed in absolute values of the experimental astrophysical @xmath0 factors measured by authors of works refs [ 69](the set i ) and of the works ref .
@xcite ( the set ii , see fig.[fig6 ] ) .
also , the central values of the weighted means for the anc - values for @xmath130(g.s . ) and @xmath131(0.429 mev ) obtained from all of the experimental data [ 611 ] , which is presented in the last line of table [ table2 ] , differ up to 10% more ( 3% less ) than those deduced from the data of [ 69](@xcite ) . as , at present , there is no a reasonable argument to prosecute to some of these experimental data measured by two groups ( [ 6 - 9](the set i ) and @xcite(the set ii ) ) , it seems , it is reasonable to obtain the weighted means of the ancs derived from all these real experimental ancs with upper and lower limits corresponding to the experimental data of the set ii and set i , respectively .
this leads to the asymmetric uncertainty for the weighted means of ancs and this is caused with the aforesaid systematical discrepancy observed in absolute values of the experimental data of the sets i and ii ( see the last line of table [ table2 ] ) . in this connection , from our point of view a new precisely measurement of @xmath132 is highly encouraged since it allows one to get an additional information about the ancs .
nevertheless , below we will use these ancs for extrapolation of the astrophysical @xmath0 factors at energies down , including @xmath8=0 .
the corresponding values of nvcs obtained by using eq.(a8 ) are given in table [ table2 ] .
a comparison of the present result and that obtained in paper @xcite shows that the underestimation of the contribution both of the nuclear interior and of the nuclear exterior indeed occurs in @xcite since the contribution of the nuclear interior ( @xmath133 4.0 fm ) to the calculated astrophysical @xmath0 factors and use the experimental data [ 6 - 11 ] more accurate than those in ref.@xcite can influence the extracted values of the ancs . besides
, one would also like to note that in reality the values of the ancs , @xmath134 and @xmath135 , should not be equal , as it was assumed in @xcite and the values of @xmath136=14.4 @xmath137 were obtained from the analysis of @xmath132 performed in within the @xmath49-matrix method .
the resulting anc ( nvc ) value for @xmath138(g.s . ) obtained by us is in good agreement with the value @xmath118=25.3 @xmath137 ( @xmath139=1.20 fm ) and that for @xmath140(0.429 mev ) differs noticeably from the value @xmath119=22.0 @xmath137 ( @xmath141=1.04 fm ) , which were obtained in @xcite within the ( @xmath142)-channel resonating - group method .
also , the results of the present work differ noticeably from the values @xmath118=12.6@xmath24 1.1 @xmath137 and @xmath143 @xmath137(@xmath134=3.55@xmath24 0.15 @xmath144 , @xmath135=2.90@xmath24 0.10 @xmath144 , @xmath139=0.596@xmath24 0.052 fm and @xmath141=0.397@xmath24 0.030 fm ) @xcite as well as those @xmath136=14.4 @xmath137 ( @xmath145=3.79 @xmath144 and @xmath146=0.680 fm ) @xcite .
in this connection one would like to note that in @xcite the bound state wave functions , which correspond to the binding energy for @xmath60(g.s . ) in the ( @xmath147)-channel differing noticeably from the experimental ones ( see table i in ref.@xcite ) , and the initial state wave functions were computed with different potentials and , so , these calculations were not self - consistent . since the ancs for @xmath148 are sensitive to the form of the nn potential , it is desirable , firstly , to calculate the wave functions of the bound state using other forms of the nn potential , and , secondly , in order to guarantee the self - consistency , the same forms of the nn potential should be used for such calculation of the initial wave functions .
the @xmath3indirectly determined values of the ancs for @xmath150 presented in the last line of table [ table2 ] ) and those for @xmath151 deduced in ref.@xcite can be used for obtaining information on the ratio @xmath152 for the virtual @xmath9 decays of the bound mirror ( @xmath153)-pair , where @xmath154 is the spectroscopic factor for @xmath60 ( @xmath155 ) in the ( @xmath147)((@xmath156))-configuration . for this aim , from @xmath157 ( @xmath158 and @xmath60 ) we form the relation @xmath159 where @xmath160(@xmath161 ) is the ratio of squares of the ancs ( single - particle ancs ) for the bound mirror ( @xmath153)-pair and @xmath51=3/2(1/2 ) for the ground ( first excited ) state of the mirror nuclei . besides , it should be noted that the relation ( [ 18acd ] ) allows one to verify a validity of the approximation ( @xmath162 ,
i.e. @xmath163 1 ) used in refs.@xcite for the mirror ( @xmath153 ) conjugated @xmath9 decays . for the bound and first excited state of the mirror ( @xmath164)-pair the values of @xmath165 and @xmath166 change by the factor of 1.3 under the variation of the geometric parameters ( @xmath49 and @xmath36 ) of the adopted woods saxon potential @xcite within the aforesaid ranges , while the ratios @xmath167 and @xmath168 for the bound and first excited states of the mirror ( @xmath153)-pair change by only about 1.5% and 6% , respectively .
it is seen that the ratios do not depend practically from variation of the free parameters @xmath49 and @xmath36 , which are equal to @xmath169=1.37@xmath24 0.02 and @xmath168=1.40@xmath24 0.09 .
they are in good agreement with those calculated in @xcite within the microscopic cluster and two - body potential models ( see table i there ) .
the ratios for the ancs are @xmath170=1.83@xmath171 and @xmath172=1.77@xmath173 . from ( [ 18acd ] )
the values of the ratio @xmath174 are equal to @xmath175=1.34@xmath176 and @xmath177=1.26@xmath178 for the ground and the first excited states , respectively . within their uncertainties , these values differ slightly from those of @[email protected] and @xmath177=0.990 calculated in ref.@xcite within the microscopic cluster model .
one notes that the values of @xmath174 calculated in @xcite are sensitive to the model assumptions ( the choice of the oscillation radius @xmath125 and the form of the effective @xmath179 potential ) and such model dependence may actually influence the mirror symmetry for the @xmath9-particle spectroscopic factors .
the mirror symmetry breakup for the @xmath9-particle spectroscopic factors can also be signalled by the results for the ratio of @xmath180 at zero energies for the mirror ( @xmath153)-pair obtained in @xcite within the resonating - group method by using the seven forms for the effective @xmath179 potential . as shown in @xcite ,
this ratio is sensitive to a form of the effective @xmath179 potential used and changes from 1.0 to 1.18 times in a dependence from the effective @xmath179 potential used .
one of the possible reasons of the sensitivity observed in @xcite can apparently be associated with a sensitivity of the ratio @xmath174 to a form of the effective @xmath179 potential used . in a contrast of such model dependence observed in @xcite , the problem of the ambiguity connected with the model @xmath79-dependence for the values of the ratios @xmath181 found by us from eq.([18acd ] )
is reduced to minimum within the experimental uncertainty .
it is seen from here that the empirical values of @xmath174 exceed unity both for the ground state and for the first excited state of the mirror ( @xmath153)-pair .
this result for @xmath174 is not accidental and can be explained qualitatively by the following consideration .
the fact is that the spectroscopic factor @xmath182 ( or @xmath183 ) is determined as a norm of the radial overlap function of the bound state wave functions of the @xmath184 , @xmath9 and @xmath155 ( or @xmath185 , @xmath9 and @xmath186 ) nuclei and is given by eqs.(100 ) and ( 101 ) from ref .
the interval of integration ( @xmath187 ) in eq .
( 101 ) can be divided in two parts . in the first integral denoted by @xmath188 for @xmath155 and @xmath189 for @xmath60 , the integration over @xmath190 covers the region 0@xmath191 ( the internal region ) , where nuclear ( @xmath192 or @xmath193 ) interactions are dominate over the coulomb interactions . in the second integral @xmath194 where in the asymptotic region the radial overlap function entering the integrand
is replaced by the appropriate whittaker function ( see , for example , eq.(108 ) of ref.@xcite ) , interaction between @xmath36 and @xmath9-particle ( @xmath195 for @xmath196 or @xmath197 for @xmath198 ) is governed by the coulomb forces only ( the external region ) . in ( [ 18ace ] ) , @xmath199 and @xmath200 is the whittaker function .
one notes that the magnitudes @xmath188 ( @xmath201 ) and @xmath202 ( @xmath203 ) define the probability of finding @xmath184 ( or @xmath185 ) in the ( @xmath156 ) configuration ( or the ( @xmath147 ) configuration ) at distances of @xmath204 and of @xmath205 , respectively .
obviously @xmath206 and @xmath207 .
an information about values of @xmath202 and @xmath203 can be obtained from ( [ 18ace ] ) by using the values of the ancs for @xmath208 and @xmath209 recommended in @xcite and in the present work , respectively .
for example , for @xmath2104.0 fm ( the surface regions for the mirror ( @xmath211)-pair ) the calculation shows that the ratio @xmath212 is equal to 1.43@xmath213 ( 1.31@xmath214 ) for the ground ( excited ) states of the @xmath155 and @xmath186 nuclei , i.e. the ratio @xmath215 .
owing to the principle of equivalency of nuclear interactions between nucleons of the ( @xmath192)-pair in the nucleus @xmath155 and ( @xmath216)-pair in the nucleus @xmath60 @xcite , the values of @xmath188 and @xmath189 should not differ appreciably . if one suggests that @xmath2171 , then the ratio @xmath2181 .
the equation ( a1 ) and the weighted means of the ancs obtained for the @xmath5(g.s ) and @xmath5(0.429 mev ) can be used for extrapolating the @xmath219 astrophysical @xmath0 factor for capture to the ground and first excited states as well as the total astrophysical @xmath0 factor at solar energies ( @xmath220 kev ) .
we tested again the fulfilment of the condition ( a2 ) in the same way as it is done above for e@xmath221 90 kev and similar results plotted in fig.[fig1 ] are also observed at energies of e@xmath222 kev . the experimental and calculated astrophysical @xmath0 factors for the @xmath11(g.s . ) , @xmath16 ( 0.429 mev ) and @xmath16 ( g.s.+0.429 mev ) reactions are presented in table [ table1 ] and displayed in fig.[fig6]_a _ , _ b _ and _ c _ , respectively . in figs.[fig6]_a _ and _ b _ , the open diamond and triangle symbols ( the filled triangle symbols ) show our results for the @xmath11(g.s . ) and @xmath16(0.429 mev ) reactions ( see table [ table1 ] also ) , which are obtained from the analysis of the total experimental astrophysical @xmath0 factors of [ 79 ] and @xcite(@xcite ) , respectively , by using the corresponding values of the ancs for each energy @xmath8 experimental point presented figs.[fig5]_a _ and _ b _ ( see table [ table1 ] too ) . there the open circle symbols lying along the smooth solid lines are the results of the extrapolation obtained by us in which each of the quoted uncertainties is the uncertainty associated with that for the ancs adopted .
all these results are the second ones of the present work . in figs.[fig6]_a _ and _ b _ the experimental data plotted by the filled circle symbols ( filled star and square symbols ) are taken from @xcite ( from @xcite ) .
the solid lines present our calculations performed also with the standard values of geometric parameters @xmath49=1.80 fm and @xmath36=0.70 fm . in fig.[fig6]_c _ , the symbols are data of all experiments [ 611 ] and the solid line presents our calculations performed with the standard values of geometric parameters @xmath49=1.80 fm and @xmath36=0.70 fm by using the weighted means of the ancs ( @xmath118 and @xmath119 ) presented in the last line of table [ table2 ] . there
the dashed ( dot - dashed ) lines are the results of calculation obtained by using the aforesaid lower ( upper ) limit values of the ancs pointed out in the last line of table [ table2 ] and the standard values of the geometric parameters ( @xmath49 and @xmath36 ) , and the dotted line is the result of ref.@xcite .
as it is seen from these figures , the equations ( a1 ) , ( a4 ) and ( a5 ) allow one to perform a correct extrapolation of the corresponding astrophysical @xmath0 factors at solar energies .
but , the noticeable systematical underestimation between the results of calculations performed in ref.@xcite in respect to the experimental data occurs .
the weighted means of the total astrophysical @xmath0 factor @xmath6 at solar energies ( @xmath8=0 and 23 @xmath223 ) obtained by us are presented in the the last line of table [ table2 ] .
as it is seen from table [ table2 ] the weighted means of @xmath19 , deduced by us separately from each the activation and the prompt method of the experimental data from works [ 69 ] and @xcite ( the first and second lines as well as the fifth and sixth lines ) , agree well within their uncertainties with each other and with those recommended in [ 911 ] .
but , these weighted means of @xmath19 obtained by us from the independent analysis of the different data ( the set i and the set ii ) differ also noticeably from one another ( about 11% ) and this distinction is mainly associated with the aforesaid difference observed in magnitudes of the corresponding ancs presented in the third and seventh lines of table [ table2 ] .
nevertheless , the weighted mean of @xmath224(0)=0.613@xmath225 @xmath226 , obtained by us by using the weighted means of the ancs presented in the last line of table [ table2 ] , within the asymmetric uncertainty , which is caused with the asymmetric uncertainty for the ancs presented in the last line of table [ table2 ] , agrees also with that recommended in [ 911,38 ] .
but , it is interesting to note that the central value of it is closer to that given in the third line of table [ table2 ] , than to the central value of the weighted mean given in the seventh line of table [ table2 ] . also , the astrophysical @xmath0 factors , calculated by using the values of the ancs obtained separately from the set i , the set ii and both them ( see table [ table2 ] ) , are fitted independently using a second - order polynomial within three energy intervals ( 0@xmath227 500 kev , 0@xmath227 1000 kev and 0@xmath227 1200 kev ) .
the results for the slop @xmath228(0)/@xmath224(0 ) are to be -0.711 mev@xmath229 , -0.734 mev@xmath229 and -0.726 mev@xmath229 in dependence on the aforesaid intervals , respectively , and they do not depend on the values of the ancs used .
one notes that they also are in agreement with -0.73 mev@xmath229 @xcite and [email protected] mev@xmath229 @xcite .
it is seen from here that the @xmath6 calculated by us ( the solid lines in fig.[fig6]_a _ , _ b _ and _ c _ ) and that obtained in @xcite have practically the same energy dependence within the aforesaid energy interval but they differ mainly with each other by a normalization .
our result for @xmath19 differs noticeably on that recommended in refs.@xcite @xcite and @xcite(@xmath230 40 kevb , [email protected] kevb and [email protected] kev b , respectively ) .
this circumstance is apparently connected with the underestimation of the contribution from the external part in the amplitude admitted in these works .
besides , the result of the present work is noticeably larger than the result of @xmath19=0.516 ( 0.53 ) kev b @xcite(@xcite ) obtained within the standard two - body ( @xmath147 ) potential by using @xmath89 potential deduced by a double - folding procedure .
one of the possible reason of this discrepancy can be apparently associated with the assumption admitted in @xcite that a value of the ratio @xmath174 for the bound mirror ( @xmath231)-pair is taken equal to unity ( @xmath232 = @xmath233=1.17 @xcite ) , which in turn results in the observed systematical underestimation of the calculated @xmath6 in respect to the experimental @xmath0 factors ( see the dashed line in fig.[fig6]_c _ ) .
but , as shown in subsection 2.3 the values of @xmath174 for the ground and first excited states of the mirror ( @xmath234)-pair are large unity . therefore , the underestimated values of @xmath232 and @xmath233 used in @xcite also result in the underestimated value of @xmath19 for the direct capture @xmath235 reaction .
perhaps , the assumption about equal values of the spectroscopic factor ( @xmath236 = @xmath237=1.17 @xcite ) is correct only for the spectroscopic factors @xmath238 since the value of @xmath19 obtained in @xcite and @xcite for the direct capture @xmath239 reaction agree excellently with each other .
one notes that in ref.@xcite the analysis of the @xmath239 experimental astrophysical @xmath0 factors @xcite has also been performed within mtbpa and the anc for @xmath208(g.s . ) , which has been deduced there from the results of ref.@xcite , also is in good agreement with that recommended by authors of ref.@xcite .
nevertheless , we observe that the value of @xmath19= 0.56 @xmath226 @xcite obtained within the microscopical ( @xmath147)-cluster approach is in agreement with our result within the uncertainty .
besides , our result is also in excellent agreement with that of @xmath19= 0.609 @xmath226 @xcite and @xmath19= 0.621 @xmath240 @xcite obtained within the ( @xmath241)-channel of version of the resonating - group method by using the modified wildermuth - tang ( mwt ) and the near - serber exchange mixture forms for the effective nn potential , respectively .
it follows from here that the mutual agreement between the results obtained in the present work and works of @xcite , which is based on the common approximation about the cluster @xmath242 structure of the @xmath60 , allows one to draw a conclusion about the dominant contribution of the @xmath242 clusterization to the low - energy @xmath56 cross section both in the absolute normalization and in the energy dependence [ 611 ] .
therefore , single - channel @xmath243 approximation for @xmath60 @xcite is quite appropriate for this reaction in the considered energy range . also , it is interesting to note that the ratios of the @xmath3indirectly determined astrophysical @xmath0 factors , @xmath244(0 ) and @xmath245 , for the @xmath246 reaction populating to the ground and first excited states obtained in the present work to those for the mirror @xmath247 reaction populating to the ground and first excited states deduced in ref.@xcite are equal to @xmath248=6.87@xmath249 and @xmath250=6.11@xmath251 , respectively .
these values are in a good agreement with those of @xmath248=6.6 and @xmath250=5.9 deduced in ref.@xcite within the microscopic cluster model .
this result also confirms directly our estimation for the ratio @xmath252 obtained above since the ancs for @xmath253(g.s ) and @xmath253(0.478 mev ) as well as the ancs for @xmath65(g.s ) and @xmath33(0.429 mev ) determine the astrophysical @xmath0 factors for the @xmath254 and @xmath219 reactions at zero energies and , consequently , the ratios @xmath248 and @xmath250 are proportional to @xmath255 and @xmath172 , respectively . fig .
[ fig6]_d _ shows a comparison between the branching ratio @xmath113 obtained in the present work ( the open triangle and square symbols ) and that recommended in refs.@xcite ( the filled square symbols ) , in @xcite ( the filled circle symbols ) and @xcite ( the filled triangle symbols ) .
the weighted mean @xmath256 of the @xmath113 recommended by us is equal to @xmath256=0.43@xmath24 0.01 . as it is seen from fig.[fig6]_d
_ , the branching ratio obtained in the present work and in @xcite is in a good agreement with that recommended in ref.@xcite although the underestimation occurs for the @xmath132 obtained in ref.@xcite .
such a good agreement between two of the experimental data for the @xmath113 can apparently be explained by the fact that there is a reduction factor in @xcite , being overall for the @xmath48(g.s . ) and @xmath48(0.429 mev ) astrophysical @xmath0 factors .
the present result for @xmath256 is in an excellent agreement with those of 0.43@xmath24 0.02 @xcite and 0.43 @xcite but is noticeably larger than 0.37 @xcite and 0.32@xmath24 0.01 @xcite .
the analysis of the modern experimental astrophysical @xmath0 factors , @xmath111 , for the @xmath16 reaction , which were precisely measured at energies @xmath8=92.9 - 1235 kev [ 611 ] , has been performed within the modified two - body potential approach @xcite .
the performed scrupulous analysis shows quantitatively that the @xmath11 reaction within the considered energy ranges is mainly peripheral and the parametrization of the direct astrophysical @xmath0 factors in terms of ancs for the @xmath150 is adequate to the physics of the peripheral reaction under consideration .
it is shown that the experimental astrophysical @xmath0 factors of the reaction under consideration [ 611 ] can be used as an independent source of getting the information about the ancs ( or nvcs ) for @xmath257 ( or for the virtual decay @xmath258 ) , and the found ancs can predict the experimental astrophysical @xmath0 factors separated for the @xmath235(g.s . ) and @xmath235(0.429 mev ) reactions at low experimentally acceptable energy regions ( 126.5@xmath7 e@xmath7 1203 kev ) obtained from the total experimental astrophysical @xmath0 factors [ 69,11 ] .
the new estimation for the weighted means of the ancs for @xmath259 and nvcs for the virtual decay @xmath258 are obtained .
also , the values of ancs were used for getting the information about the @xmath9-particle spectroscopic factors for the mirror ( @xmath260)-pair .
the obtained values of the ancs were also used for an extrapolation of astrophysical @xmath0 factors at energies less than 90 kev , including @xmath8=0 . in particular , the weighted mean of the branching ratio @xmath256 ( @xmath256=0.43@xmath24 0.01 ) and the total astrophysical @xmath0 factor @xmath19 ( @xmath224(0)=0.613@xmath261 @xmath226 ) obtained here are in agreement with that deduced in [ 711 ] from the analysis the same experimental asprophysical @xmath0 factors . besides , our result for @xmath19 is in an agreement with that @xmath19=0.56 kev @xcite obtained within the microscopical single - channel ( @xmath147 ) cluster model , @xmath19= 0.609 @xmath262 @xcite and @xmath19= 0.621 @xmath262 @xcite obtained within the ( @xmath263)-channel of version of the resonating - group method , but it is noticeably larger than the result of @xmath19=0.516 ( 0.53 ) kev @xcite(@xcite ) obtained within the standard two - body ( @xmath147 ) potential by using @xmath75 potential deduced by a double - folding procedure .
the authors are deeply grateful to s. v. artemov , l.d .
blokhintsev and a.m. mukhamedzhanov for discussions and general encouragement .
the authors thank also d.bemmerer for providing the experimental results of the updated data analysis .
the work has been supported by the academy of science of the republic of uzbekistan ( grant no.fa-f2-f077 ) . according to @xcite , for fixed @xmath50 and @xmath51 we can write the astrophysical @xmath0 factor , @xmath98 , for the peripheral direct capture @xmath11 reaction in the following form @xmath264 here
, @xmath265 is the anc for @xmath266 , which determines the amplitude of the tail of the @xmath186 nucleus bound state wave function in the ( @xmath267)-channel and is related to the spectroscopic factor @xmath268 for the ( @xmath147)-configuration with the quantum numbers @xmath50 and @xmath51 in the @xmath60 nucleus by the equation @xmath269 @xcite , and @xmath270 , where @xmath271 is the single - particle astrophysical @xmath0 factor @xcite and @xmath78 ( @xmath272 @xcite ) is the single - particle anc , which determines the amplitude of the tail of the single - particle wave function of the bound ( @xmath273 ) state @xmath274(@xmath275 ) @xcite ) and in turn is itself a function of the geometric parameters ( radius of @xmath49 and diffuseness @xmath36 ) of the woods - saxon potential , i.e. @xmath276@xcite . in order to make the dependence of the @xmath87 function on @xmath88 more explicit , in the radial matrix element @xcite entering in the @xmath277 function , we split the space of interaction in two parts separated by the channel radius @xmath278 : the interior part ( @xmath279 , where nuclear forces between the @xmath280-pair are important , and the exterior part ( @xmath281 ) , where the interaction between the @xmath9-particle and @xmath282 is governed by coulomb force only . the contribution from the exterior part of the radial matrix element into the @xmath87 function does not depend on @xmath88 since for @xmath205 the wave function @xmath283 can be approximated by its asymptotic behavior @xcite .
consequently , the parametrization of the astrophysical @xmath0 factor in the form ( a1 ) makes one it possible to fix a contribution from the exterior region ( @xmath281 ) , which is dominant for the peripheral reaction , by a model independent way if the ancs @xmath95 are known .
it follows from here that the contribution from the interior part of the radial matrix element into the @xmath87 function , which depends on @xmath88 through the fraction @xmath284 @xcite , must exactly determine the dependence of the @xmath277 function on @xmath88 . in eq .
( a1 ) the ancs @xmath95 and the free parameter @xmath88 are unknown .
but , for the peripheral @xmath48 reaction the equation ( a1 ) can be used for determination of the ancs . for this aim ,
obviously the following additional requirements @xcite @xmath285 and @xmath286 must be fulfilled as a function of the free parameter @xmath88 for each energy @xmath8 experimental point from the range @xmath287 and values of the function of @xmath87 from ( a2 ) .
the fulfillment of the relations ( a2 ) and ( a3 ) or their violation within the experimental uncertainty for @xmath288 enables one , firstly , to determine an interval for energies e where the dominance of extra - nuclear capture occurs and , secondly , to obtain the value @xmath289 for @xmath290 using the experimental astrophysical @xmath0 factors @xmath66 , precisely measured by authors of refs . [ 611 ] , instead of @xmath98 , i.e. @xmath291 then , the value @xmath289 can be used for extrapolation of the astrophysical @xmath0 factor @xmath98 to the region of experimental inaccessible energies 0@xmath292 by using the obtained value @xmath289 in ( a1 ) . the total astrophysical @xmath0 factor for the @xmath16(g.s.+0.429 mev )
reaction is given by @xmath293 @xmath294\eqno(a6)\ ] ] @xmath295 \eqno(a7)\ ] ] in which @xmath296 and @xmath115 is a branching ratio .
one notes that in the two - body potential model the anc @xmath297 is related to the nvc @xmath298 for the virtual decay @xmath299 by the equation @xcite @xmath300 where @xmath301 is the coulomb parameter for the @xmath302 bound state . in ( a8 )
the combinatorial factor taking into account the nucleon s identity is absorbed in @xmath297 and its numerical value depends on a specific model used to describe wave functions of the @xmath303 , @xmath9 and @xmath60 nuclei @xcite .
hence , the proportionality factor in ( a8 ) , which relates nvc s and ancs , depends on the choice of nuclear model @xcite .
but , as it is noted in @xcite , the nvc @xmath298 is a more fundamental quantity than the anc @xmath297 since the nvc is determined in a model - independent way as the residue of the partial @xmath0-matrix of the elastic @xmath304-scattering at the pole @xmath305 ( @xmath306 is the binding energy of the bound ( @xmath147 ) state of @xmath60 ) @xcite .
therefore , it is also of interest to get an information about values of the nvcs from eqs .
( a4 ) and ( a8 ) .
* j. n. bahcall , a. m. serenelli , and s. basu , astrophys .
j. * 621 * , l85 ( 2005 ) .
j. n. bahcall , w.f .
huebner , s. h. lubow , p. d. parker , and r. k. ulrich , rev.mod.phys . * 54 * , 767 ( 1982 ) . j. n. bahcall , and m. h. pinsonneault , rev.mod.phys . *
64 * , 781 ( 1992 ) .
e.g. adelberger , s.m.austin , j.n .
bahcall , a.b.balantekin , g.bogaert , l.s.brown , l.buchmann , f.e.cecil , a.e.champagne , l. de braeckeleer , ch.a .
duba , s.r .
elliott , s.j .
freedman , m. gai , g.goldring , ch .
r. gould , a. gruzinov , w.c .
haxton , k.m .
heeger , and e.henley , rev .
mod . phys . * 70 * , 1265 ( 1998 ) .
c.angulo , m.arnould , m.rayet , p.descouvemont , d.baye , c.leclercq-willain , a.coc , s.barhoumi , p.aguer , c.rolfs , r.kunz , j.w .
hammer , a.mayer , t. paradellis , s.kossionides , c.chronidou , k.spyrou , s.deglinnocenti , g. fiorentini , b.ricci , s.zavatarelli , c.providencia , h.wolters , j.soares , c.grama , j.rahighi , a.shotter , m.l.rachti , nucl .
a * 656 * , 3 ( 1999 ) .
b. s. nara singh , m. hass , y. nir - el , and g. haquin , phys.rev.lett . * 93 * , 262503 ( 2004 ) .
d. bemmerer , f. confortola , h. costantini , a. formicola , gy .
gyrky , r. bonetti , c. broggini , p. corvisiero , z. elekes , zs .
flp , g. gervino , a. guglielmetti , c. gustavino , g. imbriani , m. junker , m. laubenstein , a. lemut , b. limata , v.lozza , m. marta , r. menegazzo , p. prati , v. roca , c. rolfs , c. rossi alvarez , e. somorjai , o. straniero , f. strieder , f. terrasi , and h.p .
trautvetter ( the luna collaboration ) , phys.rev.lett . *
97 * , 122502 ( 2006 ) ; private communication .
gyrky , f. confortola , h. costantini , a. formicola , d. bemmerer , r. bonetti , c. broggini , p. corvisiero , z. elekes , zs .
flp , g. gervino , a. guglielmetti , c. gustavino , g. imbriani , m. junker , m. laubenstein , a. lemut , b. limata , v.lozza , m. marta , r. menegazzo , p. prati , v. roca , c. rolfs , c. rossi alvarez , e. somorjai , o. straniero , f. strieder , f. terrasi , and h.p .
trautvetter ( the luna collaboration ) , phys.rev .
c * 75 * , 035805 ( 2007 ) .
f. confortola , d. bemmerer , h. costantini , a. formicola , gy .
gyrky , p. bezzon , r. bonetti , c. broggini , p. corvisiero , z. elekes , zs .
flp , g. gervino , a. guglielmetti , c. gustavino , g. imbriani , m. junker , m. laubenstein , a. lemut , b. limata , v.lozza , m. marta , r. menegazzo , p. prati , v. roca , c. rolfs , c. rossi alvarez , e. somorjai , o. straniero , f. strieder , f. terrasi , and h.p .
trautvetter ( the luna collaboration ) , phys.rev .
c * 75 * , 065803 ( 2007 ) .
brown , c. bordeanu , r.f .
snover , d.w .
storm , d. melconian , a.l .
sallaska , s.k.l .
sjue , s. triambak , phys.rev .
c * 76 * , 055801 ( 2007 ) .
a di leva , l. gialanella , r. kunz , d. rogalla , d. schrmann , f. strieder , m. de cesare , n. de cesare , a. donofrio , z. flp , g. gyrky , g. imbariani , g. mangano , a. ordine , v. roca , c. rolfs , m. romano , e. somorjai , and f. terrasi , phys.rev.lett . * 102 * , 232502 ( 2009 ) . t. kajino , nucl.phys .
a * 460 * , 559 ( 1986 ) .
p.descouvement , a. adahchour , c. angulo , a. coc , e. vangioni - flam .
atomic data and nuclear data tables . *
88 * , 203 ( 2004 ) .
k. langanke , nucl .
a * 457 * , 351 ( 1986 ) .
a. cst and k. langanke , few - body systems .
* 29 * , 121 ( 2000 ) .
nollett , phys .
c * 63 * , 054002 ( 2001 ) .
dubovitchenko , a.v .
dzhazairov - kakhramanov , yad . fiz . * 58 * , 635 ( 1995 ) [ phys . at
* 58 * , 579 ( 1995 ) ] . *
48 * , 1664 ( 1982 ) ; nucl.phys .
a * 419 * , 115 ( 1984 ) .
p. mohr , h. abele , r. zwiebel , g. staudt , h. krauss , h. oberhummer , a. denker , j.w .
hammer and f. wolf , phys .
c * 48 * , 1420 ( 1993 ) .
p. mohr , phys .
c * 79 * , 065804 ( 2009 ) .
igamov , t.m .
tursunmuratov , and r. yarmukhamedov , yad.fiz .
* 60 * , 1252 ( 1997 ) [ phys . at
. nucl . * 60 * , 1126 ( 1997 ) ] .
osborne , c. a. barnes , r.w .
ravanagh , r.m .
kremer , g.j .
mathews , j.l .
zyskind , p.d .
parker , and a.j .
howard , nucl.phys.a * 419 * , 115(1984 ) .
r.f . christy and i. duck , nucl.phys .
* 24 * , 89 ( 1961 ) .
igamov , and r. yarmukhamedov , nucl.phys.a * 781 * , 247 ( 2007);nucl.phys.a * 832 * , 346 ( 2010 ) .
blokhintsev , i.borbely , e.i .
dolinskii , fiz.elem.chastits at . yadra .
* 8 * , 1189 ( 1977)[sov .
* 8 * , 485 ( 1977 ) ] .
blokhintsev and v.o .
yeromenko , yad.fiz .
* 71 * , 1219 ( 2008 ) [ phys . at
. nucl . * 71 * , 1126 ( 2008 ) ] .
robertson , p. dyer , r.a .
warner , r.c .
melin , t.j .
bowles , a.b .
mcdonald , g.c .
ball , w.g .
davies , e.d .
earle , phys.rev.lett . * 47 * , 1867 ( 1981 ) .
kim , m.h .
park , b.t .
kim , phys.rev .
c * 35 * 363 ( 1987 ) .
s.b . igamov and r. yarmukhamedov , phys . at
. nucl . * 71 * , 1740 ( 2008 ) ] .
artemov , s.b .
igamov , k.i .
tursunmakhatov , and r. yarmukhamedov , izv . ran : seriya fizicheskaya , * 73 * , 176 ( 2009)[bull.ras : physics,*73 * , 165 ( 2009 ) ] .
neudatchin , v.i .
kukulin , a.n .
boyarkina and v.d .
korennoy , lett.nuovo cim.*5 * , 834 ( 1972 ) .
kukulin , v.g .
neudatchin , and yu .
f. smirnov , nucl.phys .
a * 245 * , 429 ( 1975 ) .
kukulin , v.g .
neudatchin , i.t .
obukhovsky and yu.f .
`` clusters as subsystems in light nuclei '' . in clustering phenomena in nuclei , vol.3 , eds .
k. wildermuth and p. kramer ( vieweg , braunschweig , 1983)p.1 .
dubovichenko and m.a .
zhusupov , ivz .
nauk kaz.ssr , ser .
fiz .- mat . * 4 * , 44 ( 1983 ) ; yad .
* 39 * , 1378 ( 1984 ) [ sov .
phys.*39 * , 870 ( 1984 ) . ]
barnard , c. m. jones , and g.c .
phillips , nucl.phys . *
50 * , 629 ( 1964 ) .
r. j. spiger , and t. a. tombrello , phys.rev .
* 163 * ( 1967 ) 964 .
h. walliser , h. kanada , and y.c .
tang , nucl.phys . a * 419 * , 133 ( 1984 ) .
timofeyuk , p. descouvemont , r.c .
johnson , phys.rev .
c * 75 * , 034302 ( 2007 ) .
r. h. cyburt and b. davids , phys.rev.c * 78 * , 064614(2008 ) .
d. kurath and d. j. millener , nucl.phys.a * 238 * , 269 ( 1975 ) .
brune , w.h .
geist , r.w .
kavanagh , k.d .
veal , phys.rev.lett .
* 83 * , ( 1999 ) 4025 .
h.krwinkel , h.w .
becker , l. buchmann , j. grres , k.u .
kettner , w.e .
kieser , r. santo , p. schmalbrock , h.p .
trautvetter , a. vlieks , c. rolfs , j.w .
hammer , r.e .
azuma , w.s .
rodney , z.phys .
a * 304 * , 307 ( 1982 ) . t. altmeyer , e. kolbe , t. warmann , k. langanke , h.j .
assenbaum , z.phys .
a * 330 * , 277 ( 1988 ) .
u. schrder , a. redder , c. rolfs , r.e .
azuma , l. buchmann , c. campbell , j.d .
king , t.r .
donoghue , phys.lett .
b * 192 * , 55 ( 1987 )
. s. a. goncharov , j. dobesh , e. i. dolinskii , a. m. mukhamedzhanov and j. cejpek , yad .
fiz.*35 * , 662 ( 1982)[sov
. phys . * 35 * , 383 ( 1982 ) ] .
a. m. mukhamedzanov , f. m. nunes , phys.rev .
c * 72 * , 017602 ( 2005 ) . .[table1]the
@xmath3indirectly determined values of the asymptotic normalization constants ( @xmath307 and ( @xmath308 ) for @xmath309 , the experimental astrophysical @xmath0 factors ( @xmath310 and @xmath111 ) and branching ratio ( @xmath113 ) at different energies @xmath8 .
[ cols="^,^,^,^,^,^,^ " , ] | a new analysis of the modern precise measured astrophysical @xmath0 factors for the direct capture @xmath1 reaction [ b.s .
nara singh _ et al .
_ , phys.rev.lett . * 93 * , 262503 ( 2004 ) ; d. bemmerer _ et al .
_ , phys.rev.lett . * 97 * , 122502 ( 2006 ) ; f.confortola _ et al . _ ,
phys.rev.c * 75 * , 065803 ( 2007 ) , t.a.d.brown _ et al . _ ,
phys.rev .
c * 76 * , 055801 ( 2007 ) and a di leva , _
et al._,phys.rev.lett . * 102 * , 232502 ( 2009 ) ] populating to the ground and first excited states of @xmath2 is carried out based on the modified two - body potential approach .
new estimates are obtained for the @xmath3indirectly determined values of the asymptotic normalization constants ( the nuclear vertex constants ) for @xmath4(g.s . ) and @xmath5(0.429 mev ) as well as the astrophysical @xmath0 factors @xmath6 at e@xmath7 90 kev , including @xmath8=0 .
the values of asymptotic normalization constants have been used for getting information about the @xmath9-particle spectroscopic factors for the mirror ( @xmath10)-pair .
pacs : 25.55.-e;26.35.+c;26.65.+t + |
this work has been done in collaboration with a. bianconi ( univ .
brescia ) and r. jakob ( univ .
wuppertal ) . enlighting discussions with s. boffi ( univ .
pavia ) and d. boer ( riken - bnl ) are greatfully acknowledged . | we discuss the properties of fragmentation functions arising from the interference of two leading hadrons produced inside the same jet in the current fragmentation region of a hard process . for the case of semi - inclusive deep inelastic scattering ( dis ) , a proper folding of the cross section , integrated over the azimuthal position of the detected hadrons , produces a factorized form that allows for the extraction of the quark transversity distribution at leading twist .
using an extended spectator model , explicit calculations are shown for the hadron pair being two pions with invariant mass inside the @xmath0 resonance width . fragmentation functions ( ff ) , like distribution functions ( df ) , describe the nonperturbative properties of partons when confined inside hadronic bound states at a low scale .
factorization theorems for hard processes , whenever available , ensure their univsersality .
several df have been experimentally extracted and are parametrized with high accuracy . however , one of them , the quark transversity @xmath1 , is still unknown because it is connected to soft processes that flip chirality ; as such , it is , e.g. , unaccessible in inclusive dis . among alternative measurements ,
the single spin asymmetries in semi - inclusive dis seem more favourable .
in fact , a new generation of machines ( hermes , compass , erhic ) allows for a better resolution in the final state and precise semi - inclusive measurements are becoming feasible .
in this context , naive time - reversal odd ( for brevity , `` t - odd '' ) ff naturally arise because no constraints from time - reversal invariance can be imposed due to the existence of final state interactions ( fsi ) with or inside the residual jet @xcite .
some of these t - odd ff are also chiral odd , and they can be shown to represent the natural partner to isolate @xmath1 already at leading twist . in a field - theoretical description that assumes factorization , the soft processes linking the fragmenting quark to the observed hadron(s ) are defined as matrix elements of nonlocal operators involving quarks and gluons @xcite .
the simplest one at leading twist is the quark - quark correlator describing the decay of a quark @xmath2 into a detected hadron @xmath3 .
if the transverse momentum @xmath4 is measured with respect to the 3-momentum transfer @xmath5 and the quark is transversely polarized , a t - odd chiral - odd ff arises and allows for the extraction of @xmath1 via a single spin asymmetry ( the socalled collins effect ) @xcite .
however , due to the lack of collinear factorization the soft - gluon radiation needs to be taken into account and can lead to a severe suppression of the effect @xcite .
moreover , knowledge of the collins function implies the difficult task of modelling fsi between the observed hadron and the residual jet @xcite . from the theoretical point of view , it is more convenient to select processes where two leading hadrons @xmath6 and @xmath7 ( @xmath8 ) are detected in the same jet , that acts as a spectator @xcite . in this case ,
collinear factorization holds and the cross section would not be suppressed by sudakov form factors ; moreover , fsi are more easily modelled inside the pair .
if the two hadrons are unpolarized , four ff appear at leading twist : @xmath9 @xcite .
they depend on the light - cone quark momentum fractions @xmath10 delivered to the hadrons ( @xmath11 ) , on the transverse relative momentum @xmath12 ( with @xmath13 ) , on @xmath14 and @xmath15 @xcite .
each ff is also related to a specific spin state of the fragmenting quark : @xmath16 is the analogue of the collins effect ; on the contrary , @xmath17 represents a genuine new effect relating the transverse polarization of the fragmenting quark ( @xmath18 ) to the transverse relative dynamics of the detected pair ( @xmath19 ) , i.e. it is an `` analyzing power '' that transforms @xmath18 into the relative orbital angular momentum of the pair .
@xmath20 are t - odd and are nonvanishing only in the presence of residual fsi , at least between the two hadrons .
both @xmath21 are also chiral odd and can be identified as the chiral partner needed to access the transversity @xmath1 @xcite .
the cross section at leading twist for the process @xmath22 has been worked out in detail in ref . @xcite . here
, we will reconsider the case for an unpolarized beam and a transversely polarized target .
the lab frame can be defined by the plane containing the beam 3-momentum @xmath23 and the target polarization @xmath24 ( see fig . [ fig1 ] ) .
the scattering plane , which contains the scattered lepton 3-momentum @xmath25 and @xmath23 , is rotated by the azimuthal angle @xmath26 .
finally , the socalled hadronic plane , which contains @xmath27 and @xmath28 , is rotated by @xmath29 .
a further plane , which is just sketched in fig .
[ fig1 ] for sake of simplicity , contains @xmath30 and , consequently , @xmath19 ; it is rotated by @xmath31 with respect to the lab .
the nine - fold differential cross section depends on the energy fraction taken by the scattered lepton ( @xmath32 ) , on @xmath26 , on the quark light - cone momentum fraction @xmath33 of the target momentum @xmath34 , on @xmath35 and @xmath4 . since @xmath36 @xcite ,
the cross section can be more conveniently considered differential with respect to the pair invariant mass @xmath37 and @xmath38 . by integrating over the `` internal '' dynamics ( i.e. on @xmath39 ) and by properly folding the cross section over the experimental set of beam and hadron - pair azimuthal positions
@xmath26 and @xmath38 , it is possible to come to the factorized expression @xmath40 where @xmath41 is the electromagnetic fine structure constant , @xmath42 is the total energy in the center - of - mass frame , @xmath43 are the masses of the two observed hadrons , and a sum over the flavor @xmath44 of each quark and antiquark with charge @xmath45 is performed ( see also ref .
@xcite ) . in general
, the integrated function @xmath17 will depend on both @xmath46 and @xmath47 , contrary to the assumption proposed in ref .
@xcite .
it is worth noting also that eq .
( [ eq : finalx ] ) suggests a simpler measurement with respect to the case where only one hadron is detected .
in fact , here only the azimuthal position @xmath38 of the pair is required , while in order to measure the collins effect both the azimuthal angle and the @xmath48 hadron momentum need to be known ( see ref .
@xcite and references therein ) . +
a b c quantitative predictions for @xmath49 in eq .
( [ eq : finalx ] ) can be produced by extending the spectator model of ref .
@xcite to the case of the emission of a hadron pair . for the hadron pair being a proton and a pion ,
results have been published in ref .
@xcite , where fsi arise from the interference between the direct production and the roper decay . here
, results are shown for the case of two pions with invariant mass in the range @xmath50 $ ] , with @xmath51 mev and @xmath52 mev .
the spectator state has the quantum numbers of an onshell quark with constituent mass @xmath53 mev .
the quark decay at leading twist with the minimal number of vertices is represented by the set of diagrams shown in fig .
[ fig2 ] , where now t - odd ff arise from the contribution of diagram [ fig2]c through the interference between the direct production of the two @xmath54 and the @xmath0 decay .
explicit check has been made that the direct diagrams [ fig2]a and [ fig2]b qualitatively reproduce the experimental strength in the @xmath34- and @xmath55-channels , respectively , which represent most of the total @xmath56 production strength . in fig .
[ fig3 ] , @xmath57 \times h_1^ { { < \kern -0.3em \scriptscriptstyle { ) } } } ( z , m_h^2=m_\rho^2)$ ] of eq .
( [ eq : finalx ] ) is shown for the case @xmath58 .
it shows that a nonvanishing integrated interference ff survives allowing for the extraction of @xmath1 at leading twist , as suggested by eq .
( [ eq : finalx ] ) .
it is interesting to note that charge conservation for the diagrams of fig .
[ fig2 ] implies that the same results are obtained for the processes @xmath59 and @xmath60 , which differ just by the interchange @xmath61 .
therefore , the @xmath62-integrated amplitudes for the @xmath63 and @xmath64 quarks leading to the same @xmath65 final state cancel each other .
if only the valence quark content is considered for the proton ( @xmath66 ) and neutron ( @xmath67 ) targets , then the cross section at leading twist for the @xmath68 process turns out the same as for the @xmath69 one .
experimental tests of this conjecture could shed light both on the validiy of the spectator approximation and on the kinematical range where only the leading twist can be safely considered . |
generalized permutohedra are a very well - behaved class of convex polytopes studied in @xcite , as generalizations of the classical permutohedra , associahedra , cyclohedra , etc .
that work explored their wonderful properties from the point of view of valuations such as volumes , mixed volumes , and number of lattice points .
this paper focuses on their further good behavior with respect to face enumeration in the case when they are simple polytopes .
simple generalized permutohedra include as an important subclass ( the duals of ) the _ nested set complexes _ considered by deconcini and procesi in their work on _ wonderful compactifications _ of hyperplane arrangements ; see @xcite . in particular ,
when the arrangement comes from a coxeter system , one obtains interesting flag simple polytopes studied by davis , januszkiewicz , and scott @xcite .
these polytopes can be combinatorially described in terms of the corresponding coxeter graph .
carr and devadoss @xcite studied these polytopes for arbitrary graphs and called them _ graph - associahedra . _
we mention here two other recent papers in which generalized permutohedra have appeared .
morton , pachter , shiu , sturmfels , and wienand @xcite considered generalized permutohedra from the point of view of _ rank tests _ on ordinal data in statistics .
the normal fans of generalized permutohedra are what they called _ submodular rank tests_. agnarsson and morris @xcite investigated closely the @xmath3-skeleton ( vertices and edges ) in the special case where generalized permutohedra are minkowski sums of standard simplices .
let us formulate several results of the present paper .
a few definitions are required .
a _ connected building set _
@xmath4 on @xmath5:=\{1,\dots , n\}$ ] is a collection of nonempty subsets in @xmath5 $ ] such that 1 .
if @xmath6 and @xmath7 , then @xmath8 , 2 .
@xmath4 contains all singletons @xmath9 and the whole set @xmath5 $ ] , see definition [ def : building - set - definition ] .
an interesting subclass of _ graphical building sets _ @xmath10 comes from connected graphs @xmath11 on @xmath5 $ ] .
the building set @xmath10 contains all nonempty subsets of vertices @xmath12 $ ] such that the induced graph @xmath13 is connected .
the _ nestohedron _
@xmath14 is defined ( see definition [ def : nestohedron ] ) as the minkowski sum @xmath15 of the coordinate simplices @xmath16 , where the @xmath17 are the endpoints of the coordinate vectors in @xmath18 . according to (
* theorem 7.4 ) and ( * ? ? ?
* theorem 3.14 ) ( see theorem [ th : nestohedra_dual_to_nested_complexes ] below ) , the nestohedron @xmath14 is a simple polytope which is dual to a simplicial nested set complex . for a graphical building set @xmath10 , the nestohedron @xmath19 is called the _ graph - associahedron . _ in the case when @xmath11 is the @xmath20-path , @xmath19 is the usual associahedron ; and in the case when @xmath21 is the complete graph , @xmath19 is the usual permutohedron .
recall that the @xmath0-vector and the @xmath1-vector of a simple @xmath22-dimensional polytope @xmath23 are @xmath24 and @xmath25 , where @xmath26 is the number of @xmath27-dimensional faces of @xmath23 and @xmath28 .
it is known that the @xmath1-vector of a simple polytope is positive and symmetric . since the @xmath1-vector is symmetric
, one can define another vector called the @xmath2-vector @xmath29 by the relation @xmath30 a simplicial complex @xmath31 is called a _ flag complex _ ( or a _ clique complex _ ) if its simplices are cliques ( i.e. , subsets of vertices with complete induced subgraphs ) of some graph ( 1-skeleton of @xmath31 ) .
say that a simple polytope is _ flag _ if its dual simplicial complex is flag .
gal conjectured @xcite that the @xmath2-vector has nonnegative entries for any flag simple polytope .
let us that say a connected building set @xmath4 is _ chordal _ if , for any of the sets @xmath32 in @xmath4 , all subsets @xmath33 also belong to @xmath4 ; see definition [ def : chordal_building ] . by proposition [ prop : graphical_chordal ] ,
graphical chordal building sets @xmath10 are exactly building sets coming from chordal graphs . by proposition [ prop :
chordal_are_flag ] , all nestohedra @xmath14 for chordal building sets are flag simple polytopes .
so gal s conjecture applies to this class of _ chordal nestohedra , _ which include graph - associahedra for chordal graphs and , in particular , for trees . for a building set @xmath4 on @xmath5 $ ] ,
define ( see definition [ def : characterization ] ) the set @xmath34 of _ @xmath4-permutations _ as the set of permutations @xmath35 of size @xmath20 such that , for any @xmath36 , there exists @xmath37 such that @xmath38 , and @xmath39 contains both @xmath40 and @xmath41 .
it turns out that @xmath4-permutations are in bijection with vertices of the nestohedron @xmath14 ; see proposition [ prop : b - trees - b - permutations ] .
let @xmath42 denote the number of descents in a permutation @xmath35 .
let @xmath43 be the subset of permutations @xmath35 of size @xmath20 without two consecutive descents and without final descent ,
i.e. , there is no @xmath44 $ ] such that @xmath45 , assuming that @xmath46 .
( corollary [ cor : h_des_chordal ] and theorem [ th : gamma - chordal ] ) let @xmath4 be a connected chordal building set on @xmath5 $ ]
. then the @xmath1-vector of the nestohedron @xmath14 is given by @xmath47 and the @xmath2-vector of the nestohedron @xmath14 is given by @xmath48 this result shows that gal s conjecture is true for chordal nestohedra .
the paper is structured as follows .
section [ sec : fpolynomial ] reviews polytopes , cones , fans , and gives basic terminology of face enumeration for polytopes ( _ @xmath0-vectors _ ) , simple polytopes ( _ @xmath1-vectors _ ) , and _ flag _ simple polytopes ( _ @xmath2-vectors _ ) . section [ sec : cones - fans - preposets ] reviews the definition of generalized permutohedra , and recasts this definition equivalently in terms of their normal fans .
it then sets up the dictionary between preposets , and cones and fans coming from the braid arrangement .
in particular , one finds that each vertex in a generalized associahedron has associated to it a _ poset _ that describes its normal cone .
this is used to characterize when the polytope is simple , namely when the associated posets have hasse diagrams which are trees . in section [ sec : simples
] this leads to a combinatorial formula formula for the @xmath1-vector in terms of descent statistics on these tree - posets .
the remainder of the paper deals with subclasses of simple generalized permutohedra .
section [ sec : zonotopes ] dispenses quickly with the very restrictive class of simple zonotopal generalized permutohedra , namely the _ simple graphic zonotopes_. section [ sec : nested ] then moves on to the interesting class of _ nestohedra _ @xmath14 coming from a building set @xmath4 , where the posets associated to each vertex are rooted trees .
these include graph - associahedra .
section [ sec : flagnestohedra ] characterizes the flag nestohedra .
section [ sec : b - trees - permutations ] discusses @xmath4-trees and @xmath4-permutations .
these trees and permutations are in bijection with each other and with vertices of the nestohedron @xmath14 .
the @xmath1-polynomial of a nestohedron is the descent - generating function for @xmath4-trees .
then section [ sec : chordal ] introduces the class of chordal building sets and shows that @xmath1-polynomials of their nestohedra are descent - generating functions for @xmath4-permutations .
section [ sec : three - graph - ass - examples ] illustrates these formulas for @xmath1-polynomials by several examples : the classical permutohedron and associahedron , the cyclohedron , the stellohedron ( the graph - associahedron for the star graph ) , and the stanley - pitman polytope .
section [ sec : gamma_nestohedra ] gives a combinatorial formula for the @xmath2-vector of all chordal nestohedra as a descent - generating function ( or peak - generating function ) for a subset of @xmath4-permutations .
this result implies gal s nonnegativity conjecture for this class of polytopes .
the warm - up example here is the classical permutohedron , and the section concludes with the examples of the associahedron and cyclohedron .
section [ sec : one - branching ] calculates the generating functions for @xmath0-polynomials of the graph - associahedra for all trees with one branching point and discuss a relation with simon newcomb s problem .
section [ sec : almost - paths ] deals with graphs that are formed by a path with two small fixed graphs attached to the ends .
it turns out that the @xmath1-vectors of graph - associahedra for such path - like graphs can be expressed in terms of @xmath1-vectors of classical associahedra .
the section includes explicit formulas for graph - associahedra for the dynkin diagrams of all finite and affine coxeter groups .
section [ sec : flossing ] gives some bounds and monotonicity conjectures for face numbers of generalized permutohedra .
the paper ends with an appendix which clarifies the equivalence between various kinds of deformations of a simple polytope .
acknowledgments : the authors thank federico ardila , richard ehrenborg , ira gessel , sangwook kim , jason morton , margaret readdy , anne shiu , richard stanley , john stembridge , bernd sturmfels , oliver wienand , and andrei zelevinsky for helpful conversations .
this section recalls some standard definitions from the theory of convex polytopes and formulate gal s extension of the charney - davis conjecture . a _ convex polytope _ @xmath23 is the convex hull of a finite collection of points in @xmath18 .
the _ dimension _ of a polytope ( or any other subset in @xmath18 ) is the dimension of its affine span .
a _ polyhedral cone _ in @xmath18 is a subset defined by a conjunction of weak inequalities of the form @xmath49 for linear forms @xmath50 .
a _ face _ of a polyhedral cone is a subset of the cone given by replacing some of the inequalities @xmath51 by the equalities @xmath52 .
two polyhedral cones @xmath53 _ intersect properly _ if their intersection is a face of each . a _
complete fan of cones _ @xmath54 in @xmath18 is a collection of distinct nonempty polyhedral cones covering @xmath18 such that ( 1 ) every nonempty face of a cone in @xmath54 is also a cone in @xmath54 , and ( 2 ) any two cones in @xmath54 intersect properly .
cones in a fan @xmath54 are also called _
faces _ of @xmath54 .
note that fans can be alternatively defined only in terms of their top dimensional faces , as collections of distinct pairwise properly intersecting @xmath20-dimensional cones covering @xmath18 .
a _ face _ @xmath55 of a convex polytope @xmath23 is the set of points in @xmath23 where some linear functional @xmath56 achieves its maximum on @xmath23 , i.e. , @xmath57 faces that consist of a single point are called _ vertices _ and 1-dimensional faces are called _ edges _ of @xmath23 . given any convex polytope @xmath23 in @xmath18 and a face @xmath55 of @xmath23 , the _ normal cone to @xmath23 at @xmath55 _ , denoted @xmath58 ,
is the subset of linear functionals @xmath59
whose maximum on @xmath23 is achieved on all of the points in the face @xmath55 , i.e. , @xmath60 then @xmath58 is a polyhedral cone in @xmath61 , and the collection of all such cones @xmath58 as one ranges through all faces @xmath55 of @xmath23 gives a complete fan in @xmath61 called the _ normal fan _ @xmath62 . a fan of the form @xmath62 for some polytope @xmath23 is called a _ polytopal fan . _
the combinatorial structure of faces of @xmath23 can be encoded by the _ lattice of faces _ of @xmath23 ordered via inclusion .
this structure is also encoded by the normal fan @xmath62 .
indeed , the map @xmath63 is an inclusion - reversing bijection between the faces of @xmath23 and the faces of @xmath62 . a cone
is called _ pointed _ if it contains no lines ( 1-dimensional linear subspaces ) , or equivalently , if it can be defined by a conjunction of inequalities @xmath64 in which the @xmath65 span @xmath61 .
a fan is called _ pointed _ if all its faces are pointed .
if the polytope @xmath66 is full - dimensional , that is @xmath67 , then the normal fan @xmath62 is pointed . for
polytopes @xmath23 of lower dimension @xmath22 , define the @xmath68-dimensional subspace @xmath69 of linear functionals which are constant on @xmath23 .
then all cones in the normal fan @xmath62 contain the subspace @xmath70 .
thus @xmath62 can be reduced to a pointed fan in the space @xmath71 . a polytope @xmath23 is called _ simple _ if any vertex of @xmath23 is incident to exactly @xmath72 edges . a cone
is called _ simplicial _ if it can be given by a conjunctions of linear inequalities @xmath73 and linear equations @xmath74 where the covectors @xmath65 and @xmath75 form a basis in @xmath61 .
a fan is called _ simplicial _ if all its faces are simplicial .
clearly , simplicial cones and fans are pointed .
a convex polytope @xmath66 is simple if and only if its ( reduced ) normal fan @xmath76 is simplicial .
the _ dual simplicial complex _
@xmath77 of a simple polytope @xmath23 is the simplicial complex obtained by intersecting the ( reduced ) normal fan @xmath76 with the unit sphere .
note that @xmath27-simplices of @xmath77 correspond to faces of @xmath23 of codimension @xmath78 . for a @xmath22-dimensional polytope @xmath23 , the _ face number _
@xmath79 is the number of @xmath27-dimensional faces of @xmath23 .
the vector @xmath80 is called the _ @xmath0-vector _ , and the polynomial @xmath81
is called the _ @xmath0-polynomial _ of @xmath23 . similarly , for a @xmath22-dimensional fan @xmath54
, @xmath82 is the number of @xmath27-dimensional faces of @xmath54 , and @xmath83 .
note that face numbers of a polytope @xmath23 and its ( reduced ) normal cone @xmath84 are related as @xmath85 , or equivalently , @xmath86 .
we will most often deal with the case where @xmath23 is a simple polytope , or equivalently , when @xmath54 is a simplicial fan . in these situations , there is a more compact encoding of the face numbers @xmath79 or @xmath82 by smaller nonnegative integers .
one defines the _ @xmath1-vector _ @xmath87 and _ @xmath1-polynomial _ @xmath88 uniquely by the relation @xmath89 for a simplicial fan @xmath54 , the @xmath1-vector @xmath90 and the @xmath1-polynomial @xmath91 are defined by the relation @xmath92 , or equivalently , @xmath93 , for @xmath94 .
thus the @xmath1-vector of a simple polytope coincides with the @xmath1-vector of its normal fan .
the nonnegativity of @xmath95 for a simple polytope @xmath23 comes from its well - known combinatorial interpretation @xcite in terms of the @xmath3-skeleton of the simple polytope @xmath23 .
let us extend this interpretation to arbitrary complete simplicial fans .
for a simplicial fan @xmath54 in @xmath96 , construct the graph @xmath97 with vertices corresponding to @xmath22-dimensional cones and edges corresponding to @xmath98-dimensional cones of @xmath54 , where two vertices of @xmath97 are connected by an edge whenever the corresponding cones share a @xmath98-dimensional face . pick a vector @xmath99 that does not belong to any @xmath98-dimensional face of @xmath54 and orient edges of @xmath97 , as follows .
orient an edge @xmath100 corresponding to two cones @xmath101 and @xmath102 in @xmath54 as @xmath103 if the vector @xmath104 points from @xmath101 to @xmath102 ( in a small neighborhood of the common face of these cones ) .
[ prop : h_combin_int ] for a simplicial fan @xmath54 , the @xmath27th entry @xmath105 of its @xmath1-vector equals the number of vertices with outdegree @xmath27 in the oriented graph @xmath97 .
these numbers satisfy the dehn - sommerville symmetry : @xmath106 .
[ cor : h_combin_int ] ( cf .
@xcite ) let @xmath107 be a simple polytope .
pick a generic linear form @xmath59 .
let @xmath108 be the @xmath3-skeleton of @xmath23 with edges directed so that @xmath109 increases on each edge .
then @xmath95 is the number of vertices in @xmath108 of outdegree @xmath27 .
the graph @xmath97 has a unique vertex of outdegree 0 .
indeed , this is the vertex corresponding to the cone in @xmath54 containing the vector @xmath104 . for any face @xmath55 of @xmath54 ( of an arbitrary dimension ) , let @xmath110 be the induced subgraph on the set of @xmath22-dimensional cones of @xmath54 containing @xmath55 as a face . then @xmath111 , where @xmath112 is the link of the face @xmath55 in the fan @xmath54 , which is also a simplicial fan of smaller dimension .
thus the subgraph @xmath110 also contains a unique vertex of outdegree @xmath113 ( in this subgraph ) .
there is a surjective map @xmath114 from all faces of @xmath54 to vertices of @xmath97 ( i.e. , @xmath22-dimensional faces of @xmath54 ) that sends a face @xmath55 to the vertex @xmath115 of outdegree @xmath113 in the subgraph @xmath110 .
now , for a vertex @xmath115 of @xmath97 of outdegree @xmath27 , the preimage @xmath116 contains exactly @xmath117 faces of dimension @xmath118 .
indeed , @xmath116 is formed by taking all possible intersections of @xmath115 with some subset of its @xmath98-dimensional faces @xmath119 on which the vector @xmath104 is directed towards the interior of @xmath115 ; there are exactly @xmath120 such faces because @xmath115 has indegree @xmath27 in @xmath97 . thus a face of dimension @xmath118 in @xmath116
has the form @xmath121 for a @xmath122-element subset @xmath123 $ ] .
let @xmath124 be the number of vertices of @xmath97 of outdegree @xmath27 .
counting @xmath118-dimensional faces in preimages @xmath116 one obtains the relation @xmath125 . comparing this with the definition of @xmath105 , one deduces that @xmath126 .
note that the numbers @xmath105 do not depend upon the choice of the vector @xmath104 .
it follows that the numbers @xmath124 of vertices with given outdegrees also do not depend on @xmath104 . replacing the vector @xmath104 with @xmath127 reverses the orientation of all edges in the @xmath22-regular graph @xmath97 , implying the the symmetry @xmath128 .
the dehn - sommerville symmetry means that @xmath1-polynomials are palindromic polynomials : @xmath129 . in this sense
the @xmath1-vector encoding is more compact , since it is determined by roughly half of its entries , namely @xmath130 .
whenever possible , we will try to either give further explicit combinatorial interpretations or generating functions for the @xmath0- and @xmath1-polynomials of simple generalized permutohedra .
a simplicial complex @xmath31 is called a _ flag simplicial complex _ or _ clique complex _ if it has the following property : a collection @xmath131 of vertices of @xmath31 forms a simplex in @xmath31 if and only if there is an edge in the @xmath3-skeleton of @xmath31 between any two vertices in @xmath131 .
thus flag simplicial complexes can be uniquely recovered from their @xmath3-skeleta .
let us say that a simple polytope @xmath23 is a _ flag polytope _ if its dual simplicial complex @xmath77 is a flag simplicial complex .
we next discuss @xmath2-vectors of flag simple polytopes , as introduced by gal @xcite and independently in a slightly different context by brnden @xcite ; see also the discussion in @xcite . a conjecture of charney and davis @xcite led gal @xcite to define the following equivalent encoding of the @xmath0-vector or @xmath1-vector of a simple polytope @xmath23 , in terms of smaller integers , which are conjecturally nonnegative when @xmath23 is _ flag .
_ every palindromic polynomial @xmath132 of degree @xmath22 has a unique expansion in terms of centered binomials @xmath133 for @xmath134 , and so one can define the entries @xmath135 of the _
@xmath2-vector _ @xmath136 and the @xmath2-polynomial @xmath137 uniquely by @xmath138 [ conj : gal - conjecture ] gal @xcite the @xmath2-vector has nonnegative entries for any flag simple polytope . more generally , the nonnegativity of the @xmath2-vector holds for every flag simplicial homology sphere .
thus we will try to give explicit combinatorial interpretations , where possible , for the @xmath2-vectors of flag simple generalized permutohedra . as will be seen in section [ ssec : when - is - flag ] , any graph - associahedron is a flag simple polytope .
[ boolean - classes - remark ] section [ sec : gamma_nestohedra ] later will employ a a certain combinatorial approach to gal s conjecture and @xmath2-vector nonnegativity that goes back to work of shapiro , woan , and getu @xcite , also used by foata and strehl , and more recently by brnden ; see @xcite for a thorough discussion .
suppose @xmath23 is a simple polytope and one has a combinatorial formula for the @xmath1-polynomial @xmath139 , where @xmath140 is some statistic on the set @xmath141 .
suppose further that one has a partition of @xmath141 into _ @xmath0-symmetric boolean classes _
, i.e. such that the @xmath0-generating function for each class is @xmath142 for some @xmath143 .
let @xmath144 be the set of representatives of the classes where @xmath140 takes its minimal value .
then the @xmath2-polynomial equals @xmath145 .
call @xmath140 a `` generalized descent - statistic . '' additionally , define @xmath146 and call it a `` generalized peak statistic . ''
the reason for this terminology will become apparent in section [ sec : gamma_nestohedra ] .
this section reviews the definition of generalized permutohedra from @xcite .
it then records some observations about the relation between cones and fans coming from the braid arrangement and preposets .
( normal fans of generalized permutohedra are examples of such fans . )
this leads to a characterization for when generalized permutohedra are simple , an interpretation for their @xmath1-vector in this situation , and a corollary about when the associated toric variety is smooth .
the material in this section and in the appendix ( section [ sec : deformations ] ) has substantial overlap with recent work on _ rank tests of non - parametric statistics _ @xcite .
we have tried to indicate in places the corresponding terminology used in that paper .
recall that a _
usual permutohedron _ in @xmath18 is the convex hull of @xmath147 points obtained by permuting the coordinates of any vector @xmath148 with strictly increasing coordinates @xmath149 .
so the vertices of a usual permutohedron can be labelled @xmath150 by the permutations @xmath35 in the symmetric group @xmath151 .
the edges of this permutohedron are @xmath152 $ ] , where @xmath153 is an adjacent transposition .
then , for any @xmath154 and any @xmath155 , one has @xmath156 where the @xmath157 are some _ strictly positive _ real scalars , and @xmath158 are the standard basis vectors in @xmath18 .
note that a usual permutohedron in @xmath18 has dimension @xmath159 , because it is contained in an affine hyperplane where the sum of coordinates @xmath160 is constant .
[ def : generalized_permutohedron ] ( * ? ? ?
* definition 6.1 ) a _ generalized permutohedron _ @xmath23 is the convex hull of @xmath147 points @xmath161 in @xmath18 labelled by the permutations @xmath35 in the symmetric group @xmath151 , such that for any @xmath154 and any adjacent transposition @xmath155 , one still has equation , but with @xmath157 assumed only to be nonnegative , that is , @xmath157 can vanish .
the appendix shows that all @xmath147 points @xmath161 in a generalized permutohedron @xmath23 are actually vertices of @xmath23 ( possibly with repetitions ) ; see theorem [ th : deformation ] .
thus a generalized permutohedron @xmath23 comes naturally equipped with the surjective map @xmath162 given by @xmath163 , for @xmath164 .
definition [ def : generalized_permutohedron ] says that a generalized permutohedron is obtained by moving the vertices of the usual permutohedron in such a way that directions of edges are preserved , but some edges ( and higher dimensional faces ) may degenerate . in the appendix
such deformations of a simple polytope are shown to be equivalent to various other notions of deformation ; see proposition [ prop : permutohedron - summand - characterization ] below and the more general theorem [ th : deformation ] .
let @xmath165 be the usual coordinates in @xmath18 .
let @xmath166 denote the quotient space modulo the 1-dimensional subspace generated by the vector @xmath167 .
the _ braid arrangement _ is the arrangement of hyperplanes @xmath168 in the space @xmath169 .
these hyperplanes subdivide the space into the polyhedral cones @xmath170 labelled by permutations @xmath171 , called _
weyl chambers _
( of type @xmath141 ) .
the weyl chambers and their lower dimensional faces form a complete simplicial fan , sometimes called the _ braid arrangement fan . _ note that a usual permutohedron @xmath23 has dimension @xmath159 , so one can reduce its normal fan modulo the 1-dimensional subspace @xmath172 .
the braid arrangement fan is exactly the ( reduced ) normal fan @xmath76 for a usual permutohedron @xmath66 .
indeed , the ( reduced ) normal cone @xmath173 of @xmath23 at vertex @xmath161 is exactly the weyl chamber @xmath174 .
( here one identifies @xmath18 with @xmath61 via the standard inner product . )
recall that the _ minkowski sum _
@xmath175 of two polytopes @xmath176 is the polytope @xmath177 .
say that @xmath23 is a _
minkowski summand _ of @xmath178 , if there is a polytope @xmath179 such that @xmath180 .
say that a fan @xmath54 is _ refined _ by a fan @xmath112 if any cone in @xmath54 is a union of cones in @xmath112 .
the following proposition is a special case of theorem [ th : deformation ] .
[ prop : permutohedron - summand - characterization ] a polytope @xmath23 in @xmath18 is a generalized permutohedron if and only if its normal fan ( reduced by @xmath181 ) is refined by the braid arrangement fan . also , generalized permutohedra are exactly the polytopes arising as minkowski summands of usual permutohedra .
this proposition shows that generalized permutohedra lead to the study of cones given by some inequalities of the form @xmath182 and fans formed by such cones .
such cones are naturally related to posets and preposets .
recall that a _ binary relation _ @xmath178 on a set @xmath183 is a subset of @xmath184 .
a _ preposet _ is a reflexive and transitive binary relation @xmath178 , that is @xmath185 for all @xmath186 , and whenever @xmath187 one has @xmath188 . in this case
we will often use the notation @xmath189 instead of @xmath190 .
let us also write @xmath191 whenever @xmath192 and @xmath193 .
an _ equivalence relation _
@xmath194 is the special case of a preposet whose binary relation is also symmetric .
every preposet @xmath179 gives rise to an equivalence relation @xmath195 defined by @xmath196 if and only if both @xmath197 and @xmath198 .
poset _ is the special case of a preposet @xmath179 whose associated equivalence relation @xmath195 is the trivial partition , having only singleton equivalence classes .
every preposet @xmath179 gives rise to the poset @xmath199 on the equivalence classes @xmath200 .
a preposet @xmath179 is uniquely determined by @xmath195 and @xmath199 , that is , a preposet is just an equivalence relation together with a poset structure on the equivalence classes .
a _ preposet _
@xmath179 on @xmath183 is _ connected _ if the undirected graph having vertices @xmath183 and edges @xmath201 for all @xmath197 is connected . a _ covering relation _
@xmath202 in a poset @xmath179 is a pair of elements @xmath203 such that there is no @xmath204 such that @xmath205 .
hasse diagram _ of a poset @xmath179 on @xmath183 is the directed graph on @xmath183 with edges @xmath206 for covering relations @xmath202 .
let us say that a poset @xmath179 is a _ tree - poset _ if its hasse diagram is a spanning tree on @xmath183 .
thus tree - posets correspond to directed trees on the vertex set @xmath183 . a _ linear extension _ of a poset @xmath179 on @xmath183 is a linear ordering @xmath207 of all elements in @xmath183 such that @xmath208 .
let @xmath209 denote the set of all linear extensions of @xmath179 .
the _ union _ @xmath210 of two binary relations @xmath211 on @xmath183 is just their union as two subsets of @xmath212 . given any reflexive binary relation @xmath179 ,
denote by @xmath213 the preposet which is its transitive closure .
note that if @xmath214 and @xmath215 are two preposets on the same set @xmath183 , then the binary relation @xmath216 is not necessarily a preposet .
however , we can obtain a preposet by taking its transitive closure @xmath217 .
let @xmath218 denote containment of binary relations on the same set , meaning containment as subsets of @xmath212 .
also let @xmath219 be the _ opposite _ binary relation , that is @xmath220 if and only if @xmath221 . for two preposets @xmath23 and @xmath179 on the same set ,
let us say that @xmath179 is a _ contraction _ of @xmath23 if there is a binary relation @xmath222 such that @xmath223 .
in other words , the equivalence classes of @xmath195 are obtained by merging some equivalence classes of @xmath224 _ along relations _ in @xmath23 and the poset structure on equivalence classes of @xmath195 is induced from the poset structure on classes of @xmath224 .
for example , the preposet @xmath225 ( where @xmath226 is an equivalence class ) is a contraction of the poset @xmath227 .
however , the preposet @xmath228 is not a contraction of @xmath23 because @xmath3 and @xmath229 are incomparable in @xmath23 .
[ def : poset_fan ] let us say that two preposets @xmath214 and @xmath215 on the same set _ intersect properly _ if the preposet @xmath217 is both a contraction of @xmath214 and of @xmath215 .
a _ complete fan of posets on @xmath183 _ is a collection of distinct posets on @xmath183 which pairwise intersect properly , and whose linear extensions ( disjointly ) cover all linear orders on @xmath183 .
compare definition [ def : poset_fan ] to the definitions of properly intersecting cones and complete fan of cones ; see section [ ssec : cones_fans ] .
this connection will be elucidated in proposition [ prop : cone - preposet - dictionary ] .
[ ex : intersections ] the two posets @xmath230 and @xmath231 on the set @xmath232 intersect properly . here
@xmath233 is equal to @xmath234 .
these @xmath235 and @xmath236 form a complete fan of posets . however , the two posets @xmath237 and @xmath238 on the set @xmath239 do not intersect properly . in this case
@xmath240 , which is not a contraction of @xmath214 .
let us say that a _ braid cone _ is a polyhedral cone in the space @xmath166 given by a conjunction of inequalities of the form @xmath241 .
in other words , braid cones are polyhedral cones formed by unions of weyl chambers or their lower dimensional faces .
there is an obvious bijection between preposets and braid cones . for a preposet @xmath179 on the set
@xmath5 $ ] , let @xmath242 be the braid cone in the space @xmath169 defined by the conjunction of the inequalities @xmath243 for all @xmath244 .
conversely , one can always reconstruct the preposet @xmath179 from the cone @xmath242 by saying that @xmath245 whenever @xmath243 for all points in @xmath242 .
[ prop : cone - preposet - dictionary ] let the cones @xmath246 correspond to the preposets @xmath247 under the above bijection
. then 1 .
the preposet @xmath248 corresponds to the cone @xmath249 .
the preposet @xmath179 is a contraction of @xmath250 if and only if the cone @xmath115 is a face @xmath251 .
3 . the preposets @xmath247 intersect properly if and only if the cones @xmath246 do . 4 .
@xmath179 is a poset if and only if @xmath115 is a full - dimensional cone , i.e. , @xmath252 .
the equivalence relation @xmath195 corresponds to the linear span @xmath253 of @xmath115 .
the poset @xmath199 corresponds to a full - dimensional cone inside @xmath254 .
the preposet @xmath179 is connected if and only if the cone @xmath115 is pointed .
if @xmath179 is a poset , then the minimal set of inequalities describing the cone @xmath115 is @xmath255 .
( these inequalities associated with covering relations in @xmath179 are exactly the facet inequalities for @xmath115 . )
@xmath179 is a tree - poset if and only if @xmath115 is a full - dimensional simplicial cone .
10 . for @xmath256 , the cone
@xmath115 contains the weyl chamber @xmath174 if and only if @xmath179 is a poset and @xmath35 is its linear extension , that is @xmath257 .
\(1 ) the cone @xmath258 is given by conjunction of all inequalities for @xmath115 and @xmath251 .
the corresponding preposet is obtained by adding all inequalities that follow from these , i.e. , by taking the transitive closure of @xmath259 .
\(2 ) faces of @xmath251 are obtained by replacing some inequalities @xmath243 defining @xmath251 with equalities @xmath260 , or equivalently , by adding the opposite inequalities @xmath261 .
\(3 ) follows from ( 1 ) and ( 2 ) .
\(4 ) @xmath115 is full - dimensional if its defining relations do not include any equalities @xmath260 , that is @xmath195 has only singleton equivalence classes .
\(5 ) the cone associated with the equivalence relation @xmath195 is given by the equations @xmath260 for @xmath262 , which is exactly @xmath253 .
\(6 ) follows from ( 4 ) and ( 5 ) .
\(7 ) the maximal subspace contained in the half - space @xmath263 is given by @xmath264 .
thus the maximal subspace contained in the cone @xmath115 is given by the conjunction of equations @xmath260 for @xmath265 .
if @xmath179 is disconnected then this subspace has a positive dimension .
if @xmath179 is connected then this subspace is given by @xmath266 , which is just the origin in the space @xmath169 .
\(8 ) the inequalities for the covering relations @xmath267 imply all other inequalities for @xmath115 and they can not be reduced to a smaller set of inequalities .
\(9 ) by ( 4 ) and ( 7 ) full - dimensional pointed cones correspond to connected posets
. these cones will be simplicial if they are given by exactly @xmath268 inequalities . by ( 8)
this means that the corresponding poset should have exactly @xmath268 covering relations , i.e. , it is a tree - poset .
\(10 ) follows from ( 4 ) and definitions . according to proposition [ prop : cone - preposet - dictionary ] ,
a full - dimensional braid cone @xmath115 associated with a poset @xmath179 can be described in three different ways ( via all relations in @xmath179 , via covering relations in @xmath179 , and via linear extensions @xmath209 of @xmath179 ) as @xmath269 let @xmath54 be a family of @xmath22-cones in @xmath96 which intersect properly .
since they have disjoint interiors , they will correspond to a complete fan if and only if their closures cover @xmath96 , or equivalently , their spherical volumes sum to the volume of the full @xmath98-sphere .
a braid cone corresponding to a poset @xmath179 is the union of the weyl chambers @xmath174 for all linear extensions @xmath270 , and every weyl chamber has the same spherical volume ( @xmath271 of the sphere ) due to the transitive weyl group action .
therefore , a collection of properly intersecting posets @xmath272 on @xmath5 $ ] correspond to a complete fan on braid cones if and only if @xmath273 cf . definition [ def : poset_fan ] .
[ cor : fans = fans ] a complete fan of braid cones ( resp . , pointed braid cones , simplicial braid cones ) in @xmath169 corresponds to a complete fan of posets ( resp .
, connected posets , tree - posets ) on @xmath5 $ ] .
using proposition [ prop : permutohedron - summand - characterization ] , we can relate proposition [ prop : cone - preposet - dictionary ] and corollary [ cor : fans = fans ] to generalized permutohedra .
indeed , normal cones of a generalized permutohedron ( reduced modulo @xmath181 ) are braid cones . for a generalized permutohedron @xmath23 ,
define the _ vertex poset _
@xmath274 at a vertex @xmath275 as the poset on @xmath5 $ ] associated with the normal cone @xmath276 at the vertex @xmath277 , as above .
[ cor : fans_posets ] for a generalized permutohedron ( resp .
, simple generalized permutohedron ) @xmath23 , the collection of vertex posets @xmath278 is a complete fan of posets ( resp . , tree - posets ) .
thus normal fans of generalized permutohedra correspond to certain complete fans of posets , which we call _ polytopal .
_ in @xcite , the authors call such fans _ submodular rank tests , _ since they are in bijection with the faces of the cone of submodular functions .
that cone is precisely the deformation cone we discuss in the appendix .
[ ex : simplicial - non - polytopal - example ] in @xcite , the authors modify an example of studen @xcite to exhibit a non - polytopal complete fan of posets . they also kindly provided us with the following further nonpolytopal example , having @xmath279 posets @xmath274 , all of them tree - posets : @xmath280 ( which means that @xmath281 and @xmath282 ) , @xmath283 , @xmath284 , @xmath285 , @xmath286 , @xmath287 , @xmath288 , @xmath289 , @xmath290 , @xmath291 , @xmath292 , @xmath293 , @xmath294 , @xmath295 , @xmath296 , @xmath297 .
this gives a complete fan of simplicial cones , but does _ not _ correspond to a ( simple ) generalized permutohedron .
recall that @xmath298 is the surjective map @xmath163 ; see definition [ def : generalized_permutohedron ] .
the previous discussion immediately implies the following corollary .
[ cor : map ] let @xmath23 be a generalized permutohedron in @xmath18 , and @xmath275 be its vertex . for @xmath171 , one has @xmath299 whenever the normal cone @xmath300 contains the weyl chamber @xmath174 .
the preimage @xmath301 of a vertex @xmath275 is the set @xmath302 of all linear extensions of the vertex poset @xmath274 .
we remark on the significance of this cone - preposet dictionary for toric varieties associated to generalized permutohedra or their normal fans ; see fulton @xcite for further background . a complete fan @xmath54 of polyhedral cones in @xmath96 whose cones are rational with respect to @xmath303 gives rise to a toric variety @xmath304 , which is normal , complete and of complex dimension @xmath22 .
this toric variety is _ projective _ if and only if @xmath54 is the normal fan @xmath62 for some polytope @xmath23 , in which case one also denotes @xmath304 by @xmath305 . the toric variety @xmath304 is _ quasi - smooth _ or _ orbifold _ if and only if @xmath54 is a complete fan of simplicial cones ; in the projective case , where @xmath306 , this corresponds to @xmath23 being a simple polytope . in this situation ,
the @xmath1-numbers of @xmath54 ( or of @xmath23 ) have the auxiliary geometric meaning as the ( singular cohomology ) betti numbers @xmath307 .
the symmetry @xmath308 reflects poincar duality for this quasi - smooth variety .
the toric variety @xmath304 is _ smooth _ exactly when every top - dimensional cone of @xmath54 is not only simplicial but _ unimodular _ , that is , the primitive vectors on its extreme rays form a @xmath309-basis for @xmath303 .
equivalently , the facet inequalities @xmath310 can be chosen to form a @xmath309-basis for @xmath311 inside @xmath312 .
one has @xmath304 both smooth and projective if and only if @xmath306 for a _ delzant _ polytope @xmath23 , that is , one which is simple and has every vertex normal cone unimodular .
@xcite ) a complete fan @xmath54 of posets gives rise to a complete toric variety @xmath304 , which will be projective if and only if @xmath54 is associated with the normal fan @xmath62 for a generalized permutohedron .
a complete fan @xmath54 of tree - posets gives rise to a ( smooth , not just orbifold ) toric variety @xmath304 , which will be projective if and only @xmath54 is associated with the normal fan @xmath62 of a simple generalized permutohedron .
in other words , simple generalized permutohedra are always delzant .
all the assertions should be clear from the above discussion , except for the last one about simple generalized permutohedra being delzant .
however , a tree - poset @xmath179 corresponds to a set of functionals @xmath313 for the edges @xmath314 of a tree , which are well - known to give a @xmath309-basis for @xmath315 , cf .
* proposition 7.10 ) .
the goal of this section is to combinatorially interpret the @xmath1-vector of any simple generalized permutohedron .
[ def : des_tree_poset ] given a poset @xmath179 on @xmath5 $ ] , define the _ descent set _
@xmath316 to be the set of ordered pairs @xmath317 for which @xmath318 is a covering relation in @xmath179 with @xmath319 , and define the statistic _ number of descents _ @xmath320 . [
th : shelling - and - descents ] let @xmath23 be a simple generalized permutohedron , with vertex posets @xmath321 . then one has the following expression for its @xmath1-polynomial : @xmath322 more generally , for a complete fan @xmath323 of tree - posets ( see definition [ def : poset_fan ] ) , one also has @xmath324 .
proof of proposition 7.10 in @xcite ) let us prove the more general claim about fans of tree - posets , that is , simplicial fans coarsening the braid arrangement fan . pick a generic vector @xmath325 such that @xmath326 and construct the directed graph @xmath97 , as in proposition [ prop : h_combin_int ] .
let @xmath327 be the cone of @xmath54 associated with poset @xmath274 , see proposition [ prop : cone - preposet - dictionary](8 ) .
let @xmath251 be an adjacent cone separated from @xmath115 by the facet @xmath260 , @xmath328 .
the vector @xmath104 points from @xmath115 to @xmath251 if and only if @xmath329 , or equivalently , @xmath330 .
thus the outdegree of @xmath115 in the graph @xmath97 is exactly the descent number @xmath331 .
the claim now follows from proposition [ prop : h_combin_int ] . for a usual permutohedron @xmath23 in @xmath18 ,
the vertex posets @xmath274 are just all linear orders on @xmath5 $ ] .
so its @xmath1-polynomial @xmath332 is the classical _ eulerian polynomial _ the eulerian polynomial . ]
@xmath333 where @xmath334 is the usual _ descent number _ of a permutation @xmath35 .
any element @xmath35 in the weyl group @xmath151 sends a complete fan @xmath335 of tree - posets to another such complete fan @xmath336 , by relabelling all of the posets .
since @xmath337 is an isomorphic simplicial complex , with the same @xmath1-vector , this leads to a curious corollary . given a tree - poset @xmath179 on @xmath5 $ ] ,
define its _ generalized eulerian polynomial _
@xmath338 note that @xmath339 depends upon @xmath179 only as an _ unlabelled _ poset .
when @xmath179 is a linear order , @xmath340 is the usual eulerian polynomial @xmath341 .
[ cor : h - as - average ] the @xmath1-polynomial @xmath332 of a simple generalized permutohedron @xmath23 is the `` average '' of the generalized eulerian polynomials of its vertex tree - posets @xmath274 : @xmath342 see example [ ex : zonotopal - example ] below for an illustration of theorem [ th : shelling - and - descents ] and corollary [ cor : h - as - average ] .
it is natural to ask for upper and lower bounds on the @xmath1-vectors of simple generalized permutohedra .
some of these follow immediately from an @xmath1-vector monotonicity result of stanley that applies to complete simplicial fans . to state it
, we recall a definition from that paper .
say that a simplicial complex @xmath343 is a _ geometric subdivision _ of a simplicial complex @xmath31 if they have geometric realizations which are topological spaces on the same underlying set , and every face of @xmath343 is contained in a single face of @xmath31 .
[ th : subdivision - monotonicity ] ( see ( * ? ? ?
* theorem 4.1 ) ) if @xmath343 is a geometric subdivision of a cohen - macaulay simplicial complex @xmath31 , then the h - vector of @xmath343 is componentwise weakly larger than that of @xmath31 . in particular
this holds when @xmath344 come from two complete simplicial fans and @xmath343 refines @xmath31 , e.g. , the normal fans of two simple polytopes @xmath345 in which @xmath23 is a minkowski summand of @xmath23 .
[ cor : upperbound ] a simple generalized permutohedron @xmath23 in @xmath18 has @xmath1-polynomial coefficientwise smaller than that of the permutohedron , namely the eulerian polynomial @xmath341 .
proposition [ prop : permutohedron - summand - characterization ] tells us that the normal fan of @xmath23 is refined by that of the permutohedron , so the above theorem applies .
does the permutohedron also provide an upper bound for the @xmath0-vectors , flag @xmath0- and flag @xmath1-vectors , generalized @xmath1-vectors , and @xmath346-indices of generalized permutohedra also in the non - simple case ?
is there also a monotonicity result for these other forms of face and flag number data when one has two generalized permutohedra @xmath345 in which @xmath23 is a minkowski summand of @xmath347 ?
the answer is `` yes '' for @xmath0-vectors and flag @xmath0-vectors , which clearly increase under subdivision .
the answer is also `` yes '' for generalized @xmath1-vectors , which stanley also showed ( * ? ? ?
* corollary 7.11 ) can only increase under geometric subdivisions of rational convex polytopes .
but for flag @xmath1-vectors and @xmath346-indices , this is not so clear .
later on ( example [ ex : simplex - example ] , section [ sec : stanley - pitman - example ] , and section [ sec : flossing ] ) there will be more to say about lower bounds for @xmath1-vectors of simple generalized permutohedra within various classes .
this section illustrates some of the foregoing results in the case where the simple generalized permutohedron is a zonotope ; see also @xcite .
zonotopal generalized permutohedra are exactly the graphic zonotopes , and the simple zonotopes among them correspond to a very restrictive class of graphs that are easily dealt with .
a _ zonotope _ is a convex polytope @xmath348 which is the minkowski sum of one - dimensional polytopes ( line segments ) , or equivalently , a polytope whose normal fan @xmath349 coincides with chambers and cones of a hyperplane arrangement . under this equivalence , the line segments which are the minkowski summands of @xmath348 lie in the directions of the normal vectors to the hyperplanes in the arrangement .
given a graph @xmath350 without loops or multiple edges , on node set @xmath351 $ ] and with edge set @xmath352 , define the associated _ graphic zonotope _
@xmath353 to be the minkowski sum of line segments in the directions @xmath354 .
proposition [ prop : permutohedron - summand - characterization ] then immediately implies the following .
the zonotopal generalized permutohedra are exactly the graphic zonotopes @xmath353 .
simple zonotopes are very special among all zonotopes , and simple graphic zonotopes have been observed ( * ? ? ?
* remark 5.2 ) to correspond to a very restrictive class of graphic zonotopes , namely those whose biconnected components are all complete graphs .
recall that for a graph @xmath350 , there is an equivalence relation on @xmath352 defined by saying @xmath355 if there is some circuit ( i.e. , cycle which is minimal with respect to inclusion of edges ) of @xmath11 containing both @xmath356 .
the @xmath357-equivalence classes are then called _ biconnected components _ of @xmath11 .
[ prop : simple - graphic - zonotopes ] ( * ? ? ?
* remark 5.2 ) the graphic zonotope @xmath353 corresponding to a graph @xmath350 is a simple polytope if and only if every biconnected component of @xmath11 is the set of edges of a complete subgraph some subset of the vertices @xmath358 . in this case , if @xmath359 are the node sets for these complete subgraphs , then @xmath353 is isomorphic to the cartesian product of usual permutohedra of dimensions @xmath360 for @xmath361 .
let us give another description for this class of graphs .
for a graph @xmath55 with @xmath20 edges @xmath362 , the _ line graph _ @xmath363 of @xmath55 is the graph on the vertex set @xmath5 $ ] where @xmath314 is an edge in @xmath363 if and only if the edges @xmath17 and @xmath364 of @xmath55 have a common vertex .
the following claim is left an exercise for the reader . for a graph @xmath11 ,
all biconnected components of @xmath11 are edge sets of complete graphs if and only if @xmath11 is isomorphic to the line graph @xmath363 of some forest @xmath55 .
biconnected components of @xmath363 correspond to non - leaf vertices of @xmath55 . for the sake of completeness , included here
is a proof of proposition [ prop : simple - graphic - zonotopes ] .
if the biconnected components of @xmath11 induce subgraphs isomorphic to graphs @xmath365 then one can easily check that @xmath353 is the cartesian product of the zonotopes @xmath366 .
since a cartesian product of polytopes is simple if and only if each factor is simple , this reduces to the case where @xmath367 . also note that when @xmath367 and @xmath11 is a complete graph , then @xmath353 is the permutohedron , which is well - known to be simple . for the reverse implication ,
assume @xmath11 is biconnected but not a complete graph , and it will suffice , by proposition [ prop : cone - preposet - dictionary](9 ) , to construct a vertex @xmath277 of @xmath353 whose poset @xmath274 is not a tree - poset .
one uses the fact @xcite that a vertex @xmath277 in the graphic zonotope @xmath353 corresponds to an acyclic orientation of @xmath11 , and the associated poset @xmath274 on @xmath358 is simply the transitive closure of this orientation .
thus it suffices to produce an acyclic orientation of @xmath11 whose transitive closure has hasse diagram which is not a tree .
since @xmath11 is biconnected but not complete , there must be two vertices @xmath201 that do not span an edge in @xmath352 , but which lie in some circuit @xmath131 .
traverse this circuit @xmath131 in some cyclic order , starting at the node @xmath368 , passing through some nonempty set of vertices @xmath369 before passing through @xmath370 , and then through a nonempty set of vertices @xmath371 before returning to @xmath368 .
one can then choose arbitrarily a total order on the node set @xmath358 so that these sets appear as segments in this order : @xmath372 it is then easily checked that if one orients the edges of @xmath11 consistently with this total order , then the associated poset has a non - tree hasse diagram : for any @xmath373 and @xmath374 , one has @xmath375 with @xmath376 incomparable .
let @xmath353 be a simple graphic zonotope , with notation as in proposition [ prop : simple - graphic - zonotopes ] .
then @xmath353 is flag , and its @xmath0-polynomial , @xmath1-polynomials , @xmath2-polynomials are all equal to products for @xmath377 of the @xmath0- , @xmath1- , or @xmath2-polynomials of @xmath378-dimensional permutohedra .
use proposition [ prop : simple - graphic - zonotopes ] along with the fact that a cartesian product of simple polytopes is flag if and only if each factor is flag , and has @xmath0- , @xmath1- and @xmath2-polynomial equal to the product of the same polynomials for each factor .
note that the @xmath1-polynomial for an @xmath379-dimensional permutohedron is the eulerian polynomial @xmath341 described in above , and the @xmath2-polynomial is given explicitly in theorem [ th : cdg ] below .
[ ex : zonotopal - example ] consider the graph @xmath350 with @xmath380:=\{1,2,3,4\}$ ] and @xmath381 , whose biconnected components are the triangle @xmath382 and the edge @xmath383 , which are both complete subgraphs on node sets @xmath384 and @xmath385 . hence the associated graphic zonotope @xmath353 is simple and flag , equal to the cartesian product of a hexagon with a line segment , that is , @xmath353 is a hexagonal prism .
its @xmath0- , @xmath1- and @xmath2-polynomials are @xmath386 one can arrive at the same @xmath1-polynomial using theorem [ th : shelling - and - descents ] .
one lists the tree - posets @xmath274 for each of the @xmath387 vertices @xmath277 of the hexagonal prism @xmath353 , coming in @xmath388 isomorphism types , along with the number of descents for each : @xmath389 \text{chain : } & 2<3<1<4 & 1 \\ & 3<2<1<4 & 2 \\ & 4<1<2<3 & 1 \\ & 4<1<3<2 & 2 \\[.1 in ] \text{vee : } & 1<2<3 \text { and } 1<4 & 0 \\ & 1<3<2 \text { and } 1<4 & 1 \end{array } \qquad\quad \begin{array}{clc } \text{type } & \text { poset } q_v & \des \\[.1 in ] \text{wedge : } & 2<3<1 \text { and } 4<1 & 2 \\ & 3<2<1 \text { and } 4<1 & 3 \\[.1 in ] \text{wye : } & 2<1<3 \text { and } 1<4 & 1 \\ & 3<1<2 \text { and } 1<4 & 1 \\[.1 in ] \text{lambda : } & 3<1<2 \text { and } 4<1 & 2 \\ & 2<1<3 \text { and } 4<1 & 2 \end{array}\ ] ] and finds that @xmath390 .
lastly one can get this @xmath1-polynomial from corollary [ cor : h - as - average ] , by calculating directly that @xmath391 and then the @xmath1-polynomial is @xmath392 = 1 + 5t+5t^2+t^3.\ ] ]
this section reviews some results from @xcite , @xcite , and @xcite regarding the important special case of generalized permutohedra that arise from building sets .
these generalized permutohedra turn out always to be simple .
their dual simplicial complexes , the nested set complexes , are defined , and several tools are given for calculating their @xmath0- and @xmath1-vectors .
the notion of nested sets goes back to work of fulton and macpherson @xcite , and deconcini and procesi @xcite defined building sets and nested set complexes
. however , our exposition mostly follows @xcite and @xcite .
[ def : building - set - definition ] ( * ? ? ?
* definition 7.1 ) let us say that a collection @xmath4 of nonempty subsets of a finite set @xmath393 is a _ building set _ if it satisfies the conditions : * if @xmath6 and @xmath7 , then @xmath8 .
* @xmath4 contains all singletons @xmath9 , for @xmath394 . for a building set @xmath4 on @xmath393 and a subset @xmath395 , define the _ restriction _ of @xmath4 to @xmath39 as @xmath396 .
let @xmath397 denote the inclusion - maximal subsets of a building @xmath4 .
then elements of @xmath398 are pairwise disjoint subsets that partition the set @xmath393 .
call the restrictions @xmath399 , for @xmath400 , the _ connected components _ of @xmath4 .
say that a building set is _ connected _ if @xmath398 has only one element : @xmath401 .
[ ex : graphical_building ] let @xmath11 be a graph ( with no loops nor multiple edges ) on the node set @xmath393 .
the _ graphical building _
@xmath10 is the set of nonempty subsets @xmath402 such that the induced graph @xmath403 on node set @xmath404 is connected .
then @xmath10 is indeed a building set .
the graphical building set @xmath10 is connected if and only if the graph @xmath11 is connected .
the connected components of the graphical @xmath10 building set correspond to connected components of the graph @xmath11 .
also each restriction @xmath405 is the graphical building set @xmath406 for the induced subgraph @xmath13 .
[ def : nestohedron ] let @xmath4 be a building set on @xmath5:=\{1,\dots , n\}$ ] .
faces of the standard coordinate simplex in @xmath18 are the simplices @xmath16 , for @xmath407 $ ] , where the @xmath17 are the endpoints of the coordinate vectors in @xmath18 .
define the _ nestohedron _ @xmath14 as the minkowski sum of these simplices @xmath408 where @xmath409 are strictly positive real parameters ; see ( * ? ? ?
* section 6 ) . note
that since each of the normal fans @xmath410 is refined by the braid arrangement fan , the same holds for their minkowski sum ( * ? ? ?
7.12 ) , and hence the nestohedra @xmath14 are generalized permutohedra by proposition [ prop : permutohedron - summand - characterization ] .
it turns out that @xmath14 is always a simple polytope , whose combinatorial structure ( poset of faces ) does not depend upon the choice of the positive parameters @xmath409 . in describing this combinatorial structure , it is convenient to instead describe the dual simplicial complex of @xmath14 .
[ def : nested ] ( * ? ? ?
* definition 7.3 ) for a building set @xmath4 , let us say that a subset @xmath411 is a _ nested set _
if it satisfies the conditions : * for any @xmath412 one has either @xmath413 , @xmath414 , or @xmath39 and @xmath404 are disjoint .
* for any collection of @xmath415 disjoint subsets @xmath416 , their union @xmath417 is not in @xmath4 .
define the _ nested set complex _
@xmath418 as the collection of all nested sets for @xmath4 .
it is immediate from the definition that the nested set complex @xmath418 is an abstract simplicial complex on node set @xmath4 .
note that this slightly modifies the definition of a nested set from @xcite , following @xcite , in that one does not include elements of @xmath398 in nested sets .
[ th : nestohedra_dual_to_nested_complexes ] ( * ? ? ?
* theorem 7.4 ) , ( * ? ? ?
* theorem 3.14 ) let @xmath4 be a building set on @xmath5 $ ] .
the nestohedron @xmath14 is a simple polytope of dimension @xmath419 .
its dual simplicial complex is isomorphic to the nested set complex @xmath418 .
an explicit correspondence between faces of @xmath14 and nested sets in @xmath418 is described in ( * ? ? ?
* proposition 7.5 ) .
the dimension of the face of @xmath14 associated with a nested set @xmath420 equals @xmath421 .
thus vertices of @xmath14 correspond to inclusion - maximal nested sets in @xmath418 , and all maximal nested sets contain exactly @xmath419 elements .
[ rem : bistellar - construction ] for a building set @xmath4 on @xmath5 $ ] , it is known ( * ? ? ?
* theorem 4 ) that one can obtain the nested set complex @xmath418 ( resp .
, the nestohedron @xmath14 ) via the following stellar subdivision ( resp . , shaving )
construction , a common generalization of 1 .
the barycentric subdivision of a simplex as the dual of the permutohedron , 2 .
lee s construction of the associahedron @xcite .
start with an @xmath379-simplex whose vertices ( resp . , facets ) have been labelled by the singletons @xmath422 for @xmath423 $ ] , which are all in @xmath4 .
then proceed through each of the non - singleton sets @xmath39 in @xmath4 , in any order that reverses inclusion ( i.e. , where larger sets come before smaller sets ) , performing a _ stellar subdivision _ on the face with vertices ( resp .
, _ shave off _ the face which is the intersection of facets ) indexed by the singletons in @xmath39 . note that if @xmath424 are the connected components of a building set @xmath4 , then @xmath14 is isomorphic to the direct product of polytopes @xmath425 .
thus it is enough to investigate generalized permutohedra @xmath14 and nested set complexes @xmath418 only for connected buildings .
[ rem : building - set - monotonicity ] the definition of the nestohedron @xmath14 as a minkowski sum should make it clear that whenever one has two building sets @xmath426 , then @xmath14 is a minkowski summand of @xmath427 . hence theorem [ th : subdivision - monotonicity ] implies the @xmath1-vector of @xmath427 is componentwise weakly larger than that of @xmath14 .
[ rem : graphical - monotonicity ] nestohedra @xmath19 associated with graphical building sets @xmath10 are called _ graph - associahedra _ , and have been studied in @xcite . in @xcite , the sets in @xmath10 are called _ tubes _ , and the nested sets are called _ tubings _ of @xmath11 . in particular , the @xmath1-vector monotonicity discussed in remark [ rem : building - set - monotonicity ] applies to graph - associahedra @xmath428 associated to graphs @xmath429 where @xmath11 is an edge - subgraph of @xmath430 .
[ ex : complete_upper ] ( _ upper bound for nestohedra : the permutohedron _ ) see ( * ? ? ? * sect .
8.1 ) for the complete graph @xmath431 , the building set @xmath432}\setminus \{\emptyset\}$ ] consists of all nonempty subsets in @xmath5 $ ] .
let us call it the _ complete building set .
_ the corresponding nestohedron ( the graph - associahedron of the complete graph ) is the usual @xmath379-dimensional permutohedron in @xmath18 .
the @xmath433-th component @xmath434 of its @xmath1-vector is the eulerian number , that is the number of permutations in @xmath151 with @xmath433 descents ; and its @xmath1-polynomial is the eulerian polynomial @xmath341 ; see .
this @xmath1-vector gives the componentwise upper bound on @xmath1-vectors for all @xmath98-dimensional nestohedra .
this also implies that the @xmath0-vector of the permutohedron gives componentwise upper bound on @xmath0-vectors of nestohedra .
[ ex : simplex - example ] ( _ lower bound for nestohedra : the simplex _ ) the smallest possible connected building set @xmath435 \}$ ] gives rise to the nestohedron @xmath14 which is the @xmath379-simplex in @xmath18 . in this case @xmath436 give trivial componentwise lower bounds on the @xmath0- , @xmath1-vectors of nestohedra .
it turns out that there are two useful recurrences for @xmath0-polynomials of nestohedra and nested set complexes .
let @xmath437 be the @xmath0-polynomial of the nestohedron @xmath14 : @xmath438 where @xmath439 is the number of @xmath27-dimensional faces of @xmath14 . as usual
, it is related to the @xmath1-polynomial as @xmath440 .
[ th : f_b_recurrence ] ( * ? ? ?
* theorem 7.11 ) the @xmath0-polynomial @xmath437 is determined by the following recurrence relations : 1 . if @xmath4 consists of a single singleton , then @xmath441 .
2 . if @xmath4 has connected components @xmath442 , then @xmath443 3 .
if @xmath4 is a connected building set on @xmath393 , then @xmath444 another recurrence relation for @xmath0-polynomials was derived in @xcite , and will be used in section [ sec : pde ] below .
it will be more convenient to work with the @xmath0-polynomial of nested set complexes @xmath445 where @xmath4 is a building set on @xmath393 . for a building set @xmath4 on @xmath393 and a subset @xmath446 , recall that the restriction of @xmath4 to @xmath39 is defined as @xmath447 . also define the _ contraction _ of @xmath39 from @xmath4 as the building set on @xmath448 given by @xmath449 see ( * ? ? ?
* definition 3.1 ) .
a link decomposition of nested set complexes was constructed in @xcite .
it implies the following recurrence relation for the @xmath0-vector .
[ th : f_recurrence_zelevinsky ] ( * ? ? ?
* proposition 4.7 ) for a building set @xmath4 on a nonempty set @xmath393 , one has @xmath450
let @xmath11 be a simple graph on @xmath393 and let @xmath451 , i.e. , @xmath39 is a connected subset of nodes of @xmath11 .
it has already been mentioned that @xmath452 ; see example [ ex : graphical_building ] .
let @xmath453 be the graph on the node set @xmath448 such that two nodes @xmath454 are connected by an edge in @xmath453 if and only if 1 .
@xmath27 and @xmath118 are connected by an edge in @xmath11 , or 2 .
there are two edges @xmath455 and @xmath456 in @xmath11 with @xmath457 .
then the contraction of @xmath39 from the graphical building set @xmath10 is the graphical building set associated with the graph @xmath453 , that is @xmath458 .
this section characterizes the flag nested set complexes and nestohedra , and then identifies those which are `` smallest '' . for a graphical
building set @xmath10 it has been observed ( @xcite , ( * ? ? ?
* corollary 7.4 ) ) that one can replace condition ( n2 ) in definition [ def : nested ] with a weaker condition : * for a disjoint pair of subsets @xmath412 , one has @xmath459 .
this implies that nested set complexes associated to graphical buildings are flag complexes .
more generally , one has the following characterization of the nested set complexes which are flag .
[ prop : flag - characterization ] for a building set @xmath4 , the following are equivalent . 1
. the nested set complex @xmath418 ( or equivalently , the nestohedron @xmath14 ) is flag .
the nested sets for @xmath4 are the subsets @xmath460 which satisfy conditions ( n1 ) and ( n2 ) .
if @xmath461 with @xmath462 are pairwise disjoint and their union @xmath463 is in @xmath4 , then one can reindex so that for some @xmath433 with @xmath464 one has both @xmath465 and @xmath466 in @xmath4 .
the equivalence of ( i ) and ( ii ) essentially follows from the definitions .
we will show here the equivalence of ( i ) and ( iii ) .
assume that ( iii ) fails , and let @xmath467 provide such a failure with @xmath468 minimal .
note that this means @xmath469 , and minimality of @xmath468 forces @xmath470 for each @xmath471 ; otherwise one could replace the two sets @xmath472 on the list with the one set @xmath473 to obtain a counterexample of size @xmath474 .
this means that all of the pairs @xmath475 index edges of @xmath418 , although @xmath476 does not . hence @xmath418 is not flag
, i.e. , ( i ) fails
. now assume ( i ) fails , i.e. , @xmath418 is not flag .
let @xmath467 be subsets in @xmath4 , for which each pair @xmath475 with @xmath477 is a nested set , but the whole collection @xmath478 is not , and assume that this violation has @xmath468 minimal . because @xmath475 are nested for @xmath477 , it must be that @xmath479 _ does _ satisfy condition ( n1 ) , and so @xmath479 must fail condition ( n2 ) . by minimality of @xmath468 , it must be that the @xmath467 are pairwise disjoint and their union @xmath463 is in @xmath4 . bearing in mind
that @xmath470 for @xmath477 , it must be that @xmath469 .
but then @xmath479 must give a violation of property ( iii ) , else one could use property ( iii ) to produce a violation of ( i ) either of size @xmath433 or of size @xmath480 , which are both smaller than @xmath468 . for graphical buildings @xmath10 ,
the graph - associahedron @xmath19 and nested set complex @xmath481 are flag .
one can now use proposition [ prop : flag - characterization ] to characterize the inclusion - minimal connected building sets @xmath4 for which @xmath418 and @xmath14 are flag . for any building set @xmath4 on @xmath5 $ ] with @xmath418 flag
, one can apply proposition [ prop : flag - characterization](iii ) with @xmath476 equal to the collection of singletons @xmath482 since they are disjoint and their union @xmath5 $ ] is also in @xmath4 .
thus after reindexing , some initial segment @xmath483 $ ] and some final segment @xmath5 \setminus [ k]$ ] must also be in @xmath4 .
iterating this , one can assume after reindexing that there is a _
plane binary tree _
@xmath484 with these properties 1 .
the singletons @xmath485 label the leaves of @xmath484 , 2 .
each internal node of @xmath484 is labelled by the set @xmath39 which is the union of the singletons labelling the leaves of the subtree below it ( so @xmath5 $ ] labels the root node ) , and 3 .
the building set @xmath4 contains of all of the sets labelling nodes in this tree .
it is not hard to see that these sets labelling the nodes of @xmath484 already comprise a building set @xmath486 which satisfies proposition [ prop : flag - characterization](iii ) , and therefore give rise to a nested set complex @xmath487 and nestohedron @xmath488 which are flag .
see figure [ stanley - pitman - relative ] for an example . and
building set @xmath486 , along with its complex of nested sets @xmath487 , drawn first as in the construction of remark [ rem : bistellar - construction ] , and then redrawn as the boundary of an octahedron.,height=220 ] the previous discussion shows the following .
the building sets @xmath486 parametrized by plane binary trees @xmath484 are exactly the inclusion - minimal building sets among those which are connected and have the nested set complex and nestohedron flag . as a special case , when @xmath484 is the plane binary tree having leaves labelled by the singletons and internal nodes labelled by all initial segments @xmath483 $ ] ,
one obtains the building set @xmath486 whose nestohedron @xmath488 is _ stanley - pitman polytope _ from @xcite ; see @xcite .
the stanley - pitman polytope is shown there to be combinatorially ( but not affinely ) isomorphic to an @xmath379-cube ; the argument given there generalizes to prove the following . for any plane binary tree @xmath484 with @xmath20 leaves ,
the nested set complex @xmath487 is isomorphic to the boundary of a @xmath379-dimensional cross - polytope ( hyperoctahedron ) , and @xmath488 is combinatorially isomorphic to an @xmath379-cube .
note that the sets labelling the non - root nodes of @xmath484 can be grouped into @xmath268 pairs @xmath489 of siblings , meaning that @xmath490 are nodes with a common parent in @xmath484 .
one then checks that the nested sets for @xmath486 are exactly the collections @xmath491 containing _ at most one set from each pair _ @xmath492 . as a simplicial complex
, this is the boundary complex of an @xmath379-dimensional cross - polytope in which each pair @xmath492 indexes an antipodal pair of vertices .
note that in this case , @xmath493 which gives a lower bound for the @xmath0- and @xmath1-vectors of flag nestohedra by remark [ rem : building - set - monotonicity ] .
if one assumes conjecture [ conj : gal - conjecture ] , then it would also give a lower bound for @xmath2-vectors of flag nestohedra ( and for flag simplicial polytopes in general ) .
note that the permutohedron is a graph - associahedron ( and hence a flag nestohedra ) .
therefore , corollary [ cor : upperbound ] implies that the permutohedron provides the upper bound on the @xmath0- and @xmath1-vectors among the flag nestohedra .
this section discusses @xmath4-trees and @xmath4-permutations , which are two types of combinatorial objects associated with vertices of the nestohedron @xmath14 .
the @xmath1-polynomial of @xmath14 equals the descent - generating function for @xmath4-trees .
this section gives a combinatorial interpretation of the @xmath1-polynomials of nestohedra .
since nestohedra @xmath14 are always simple , one should expect some description of their vertex tree - posets @xmath274 ( see corollaries [ cor : fans_posets ] and [ cor : map ] ) in terms of the building set @xmath4 .
[ sssec : bh ] recall that a _ rooted tree _
is a tree with a distinguished node , called its _
root_. one can view a rooted tree @xmath494 as a partial order on its nodes in which @xmath495 if @xmath118 lies on the unique path from @xmath27 to the root .
one can also view it as a directed graph in which all edges are directed towards the root ; both viewpoints will be employed here . for a node @xmath27 in a rooted tree @xmath494 ,
let @xmath496 denote the set of all descendants of @xmath27 , that is @xmath497 if there is a directed path from the node @xmath118 to the node @xmath27 .
note that @xmath498 .
nodes @xmath27 and @xmath118 in a rooted tree are called _ incomparable _ if neither @xmath27 is a descendant of @xmath118 , nor @xmath118 is a descendant of @xmath27 .
[ def : b - trees ] ( * ? ? ?
* definition 7.7 ) , cf .
@xcite for a connected building set @xmath4 on @xmath5 $ ] , let us define a _
@xmath4-tree _ as a rooted tree @xmath494 on the node set @xmath5 $ ] such that * for any @xmath499 $ ] , one has @xmath500 . * for @xmath415 incomparable nodes @xmath501 $ ]
, one has @xmath502 .
note that , when the nested set complex @xmath418 is flag , that is when @xmath4 satisfies any of the conditions of proposition [ prop : flag - characterization ] , one can define a @xmath4-tree by requiring condition ( t2 ) only for @xmath503 .
[ prop : b - trees ] ( * ? ? ? * proposition 7.8 ) , ( * ? ? ? * proposition 3.17 ) for a connected building set @xmath4 , the map sending a rooted tree @xmath494 to the collection of sets @xmath504i@xmath505 gives a bijection between @xmath4-trees and maximal nested sets .
( recall that maximal nested sets correspond to the facets of the nested set complex @xmath418 and to the vertices of the nestohedron @xmath14 . ) furthermore , if the @xmath4-tree @xmath494 corresponds to the vertex @xmath277 of @xmath14 then @xmath506 , that is , @xmath494 is the vertex tree - poset for @xmath277 in the notation of corollary [ cor : fans_posets ] .
does a simple ( indecomposable ) generalized permutohedron @xmath23 come from a ( connected ) building set if and only if every poset @xmath274 is a rooted tree , i.e. has a unique maximal element ?
proposition [ prop : b - trees ] and theorem [ th : shelling - and - descents ] yield the following corollary .
[ cor : h_des_t ] for a connected building set @xmath4 on @xmath5 $ ] , the @xmath1-polynomial of the generalized permutohedron @xmath14 is given by @xmath507 where the sum is over @xmath4-trees @xmath494 .
the following recursive description of @xmath4-trees is straightforward from the definition .
[ prop : b_trees_construction ] ( * ? ? ?
* section 7 ) let @xmath4 be a connected building set on @xmath393 and let @xmath394 .
let @xmath508 be the connected components of the restriction @xmath509 .
then all @xmath4-trees with the root at @xmath27 are obtained by picking a @xmath510-tree @xmath511 , for each component @xmath510 , @xmath512 , and connecting the roots of @xmath513 with the node @xmath27 by edges . in other words ,
each @xmath4-tree is obtained by picking a root @xmath394 , splitting the restriction @xmath514 into connected components , then picking nodes in all connected components , splitting corresponding restrictions into components , etc .
recall definition [ def : generalized_permutohedron ] of the surjection @xmath515 @xmath516 here and below one identifies vertices of the nestohedron @xmath14 with @xmath4-trees via proposition [ prop : b - trees ] . by corollary [ cor :
map ] , for a @xmath4-tree @xmath494 , one has @xmath517 if and only if @xmath35 is a linear extension of @xmath494 .
proposition [ prop : b_trees_construction ] leads to an explicit recursive description of the surjection @xmath518 .
[ prop : b - map ] let @xmath4 be a connected building set on @xmath5 $ ] . given a permutation @xmath519 , one recursively constructs a @xmath4-tree @xmath520 , as follows .
the root of @xmath494 is the node @xmath521 .
let @xmath522 be the connected components of the restriction @xmath523 .
restricting @xmath35 to each of the sets @xmath524 gives a subword of @xmath35 , to which one can recursively apply the construction and obtain a @xmath524-tree @xmath525 .
then attach these @xmath526 as subtrees of the root node @xmath521 in @xmath494 .
this association @xmath527 is the map @xmath528 .
it is natural to ask for a nice section of the surjection @xmath518 ; these are the @xmath4-permutations defined next .
[ def : characterization ] let @xmath4 be a building set on @xmath5 $ ] . define the set @xmath529 of _ @xmath4-permutations _ as the set of permutations @xmath256 such that for any @xmath499 $ ] , the elements @xmath40 and @xmath530 lie in the same connected component of the restricted building set @xmath531 .
the following recursive construction of @xmath4-permutations is immediate from the definition .
[ lem : b - permutation - backward ] a permutation @xmath256 is a @xmath4-permutation if and only if it can be constructed via the following procedure . pick @xmath521 from the connected component of @xmath4 that contains @xmath20 ; then pick @xmath532 from the connected component of @xmath533\setminus\{w(n)\}}$ ] that contains the _ maximal _ element of @xmath5 \setminus \{w(n)\}$ ] ; then pick @xmath534 from the connected component of @xmath533 \setminus \{w(n ) , w(n-1)\}}$ ] that contains the maximal element of @xmath5\setminus \{w(n),w(n-1)\}$ ] , etc . continue in this manner until @xmath535 has been chosen .
let @xmath494 be a rooted tree on @xmath5 $ ] viewed as a tree - poset where the root is the unique maximal element .
the _ lexicographically minimal linear extension _ of @xmath494 is the permutation @xmath256 such that @xmath535 is the minimal leaf of @xmath494 ( in the usual order on @xmath309 ) , @xmath536 is the minimal leaf of @xmath537 ( the tree @xmath494 with the vertex @xmath535 removed ) , @xmath538 is the minimal leaf of @xmath539 , etc .
there is the following alternative `` backward '' construction for the lexicographically minimal linear extension of @xmath494 .
[ lem : lex_min_backward ] let @xmath35 be the lexicographically minimal linear extension of a rooted tree @xmath494 on @xmath5 $ ] .
then the permutation @xmath35 can be constructed from @xmath494 , as follows : @xmath521 is the root of @xmath494 ; @xmath532 is the root of the connected component of @xmath540 that contains the maximal vertex of this forest ( in the usual order on @xmath309 ) ; @xmath534 is the root of the connected component of @xmath541 that contains the maximal vertex of this forest , etc . in general , @xmath40 is the root of the connected component of the forest @xmath542 that contains the vertex @xmath543 .
the proof is by induction on the number of vertices in @xmath494 .
let @xmath544 be the rooted tree obtained from @xmath494 by removing the minimal leaf @xmath545 .
then the lexicographically minimal linear extension @xmath35 of @xmath494 is @xmath546 , where @xmath547 is the lexicographically minimal linear extension of @xmath544 , and both @xmath35 and @xmath547 are written in list notation . by induction , @xmath547 can be constructed from @xmath544 backwards .
when one performs the backward construction for @xmath494 , the vertex @xmath545 can never be the root of the connected component of @xmath548 containing the maximal vertex , for @xmath549 .
so the backward procedure for @xmath494 produces the same permutation @xmath550 .
the next claim gives a correspondence between @xmath4-trees and @xmath4-permutations .
[ prop : b - trees - b - permutations ] let @xmath4 be a connected building set on @xmath5 $ ] .
the set @xmath34 of @xmath4-permutations is exactly the set of lexicographically minimal linear extensions of the @xmath4-trees .
( equivalently , @xmath34 is the set of lexicographically minimal representatives of fibers of the map @xmath518 . ) in particular , the map @xmath518 induces a bijection between @xmath4-permutations and @xmath4-trees , and @xmath34 is a section of the map @xmath518 .
let @xmath256 be a permutation and let @xmath520 be the corresponding @xmath4-tree constructed as in proposition [ prop : b - map ] .
note that , for @xmath551 , the connected components of the forest @xmath552 correspond to the connected components of the building set @xmath553 , and corresponding components have the same vertex sets . according to lemma [ lem : lex_min_backward ] , the permutation @xmath35 is the lexicographically minimal linear extension of @xmath494 if and only if @xmath35 is a @xmath4-permutation as described in lemma [ lem : b - permutation - backward ] .
this section describes an important class of building sets @xmath4 , for which the descent numbers of @xmath4-trees are equal to the descent numbers of @xmath4-permutations . in this case
, the @xmath1-polynomial of the nestohedron @xmath14 equals the descent - generating function of the corresponding @xmath4-permutations .
let us say that a _ descent _ of a permutation @xmath256 is a pair such that @xmath554 . ]
@xmath555 such that @xmath556 .
let @xmath557 be the set of all descents in @xmath35 . also recall that the descent set @xmath316 of a poset @xmath179 is the set of pairs @xmath558 such that @xmath559 and @xmath560 ; see definition [ def : des_tree_poset ] . [
lem : aux_descent ] let @xmath179 be any poset on @xmath5 $ ] , and let @xmath561 be the lexicographically minimal linear extension of @xmath179 .
then one has @xmath562 .
one must show that any descent @xmath558 ( with @xmath563 ) in @xmath35 must come from a covering relation @xmath559 in the poset @xmath179 .
indeed , if @xmath564 and @xmath565 are incomparable in @xmath179 , then the permutation obtained from @xmath35 by transposing @xmath564 and @xmath565 would be a linear extension of @xmath23 which is lexicographically smaller than @xmath35 .
on the other hand , if @xmath564 and @xmath565 are comparable but not adjacent elements in @xmath179 , then they can never be adjacent elements in a linear extension of @xmath179 .
in particular , this lemma implies that , for a @xmath4-tree @xmath494 and the corresponding @xmath4-permutation @xmath35 ( i.e. , @xmath35 is the lexicographically minimal linear extension of @xmath494 ) , one has @xmath566 .
the rest of this section discusses a special class of building sets for which one always has @xmath567 .
[ def : chordal_building ] let us say that a building set @xmath4 on @xmath5 $ ] is _ chordal _ if it satisfies the following condition : for any @xmath568 and @xmath569 , the subset @xmath33 also belongs to @xmath4 .
recall that a graph is called _ chordal _ if it has no induced @xmath433-cycles for @xmath570 .
it is well known @xcite that chordal graphs are exactly the graphs that admit a _ perfect elimination ordering , _ which is an ordering of vertices such that , for each vertex @xmath277 , the neighbors of @xmath277 that occur later than @xmath277 in the order form a clique .
equivalently , a graph @xmath11 is chordal if its vertices can be labelled by numbers in @xmath5 $ ] so that @xmath11 has no induced subgraph @xmath571 with the edges @xmath317 , @xmath455 but without the edge @xmath572 .
let us call such graphs on @xmath5 $ ] _ perfectly labelled chordal graphs._-avoiding graphs _ because they are exactly the graphs that have no induced @xmath573-path @xmath564@xmath565@xmath574 with the relative order of the vertices @xmath575 as in the permutation 312 .
note that , unlike the pattern avoidance in permutations , a @xmath576-avoiding graph is the same thing as a @xmath577-avoiding graph . ]
[ ex : decreasing_trees ] let us say that a tree on @xmath5 $ ] is _ decreasing _ if the labels decrease in the shortest path from the vertex @xmath20 ( the root ) to another vertex .
it is easy to see that decreasing trees are exactly the trees which are perfectly labelled chordal graphs . clearly , any unlabelled tree has such a decreasing labelling of vertices .
the following claim justifies the name `` chordal building set . ''
[ prop : graphical_chordal ] a graphical building set @xmath10 is chordal if and only if @xmath11 is a perfectly labelled chordal graph .
suppose that @xmath11 contains an induced subgraph @xmath571 with exactly two edges @xmath317 , @xmath455 .
then @xmath578 but @xmath579 . thus @xmath10 is not a chordal building set . on the other hand ,
suppose that @xmath10 is not chordal
. then one can find a connected subset @xmath580 of vertices in @xmath11 such that @xmath581 , for some @xmath582 .
in other words , the induced graph @xmath583 is disconnected .
let us pick a shortest path @xmath23 in @xmath584 that connects two different components of @xmath430 .
let @xmath27 be the minimal vertex in @xmath23 and let @xmath118 and @xmath433 be the two vertices adjacent of @xmath27 in the path @xmath23 . clearly , @xmath585 and @xmath586 .
it is also clear that @xmath317 is not an edge of @xmath11 .
otherwise there is a shorter path obtained from @xmath23 by replacing the edges @xmath317 and @xmath455 with the edge @xmath572 .
so one has found a forbidden induced subgraph @xmath587 .
thus @xmath11 is not a perfectly labelled chordal graph .
[ prop : des = des - chordal ] let @xmath4 be a connected chordal building set . then , for any @xmath4-tree @xmath494 and the corresponding @xmath4-permutation @xmath35 , one has @xmath588 .
let @xmath494 be a @xmath4-tree and let @xmath35 be the corresponding @xmath4-permutation , which can be constructed backward from @xmath494 as described in lemma [ lem : lex_min_backward ] .
let us fix @xmath589 .
let @xmath590 be the connected components of the forest @xmath591 , where @xmath513 are the subtrees whose roots are the children of the vertex @xmath592 , and @xmath593 are the remaining subtrees .
let @xmath594 $ ] be the set of all descendants of @xmath592 in @xmath494 . by definition [ def : b - trees](t1 )
, one has @xmath37 .
suppose that the vertex @xmath595 appears in one of the subtrees @xmath513 , say , in the tree @xmath596 .
then , by lemma [ lem : lex_min_backward ] , @xmath40 should be the root of @xmath596 .
we claim that all vertices in @xmath597 are less than @xmath592 .
indeed , this is clear if @xmath592 is the maximal element in @xmath39 .
otherwise , the set @xmath598 is nonempty , @xmath599 because @xmath4 is chordal , and @xmath600 contains the maximal vertex @xmath601 . since the vertex set @xmath404 of @xmath596 should be an element of @xmath4 , it follows that @xmath602 .
so all vertices of @xmath597 are less than @xmath592
. thus none of the edges of @xmath494 joining the vertex @xmath592 with the roots of @xmath603 can be a descent edge .
the only potential descent edge is the edge @xmath604 that attaches the subtree @xmath596 to @xmath592 .
this edge will be a descent edge in @xmath494 if and only if @xmath554 , i.e. , exactly when @xmath604 is a descent in the permutation @xmath35 .
now suppose that the maximal vertex @xmath595 appears in one of the remaining subtrees @xmath593 , which are not attached to the vertex @xmath592 , say , in @xmath605 .
in this case @xmath592 should be greater than all @xmath606 .
( otherwise , if @xmath607 , then at the previous step of the backward construction for @xmath35 , @xmath605 is the connected component of @xmath608 that contains the vertex @xmath609 .
so @xmath592 should have been the root of @xmath605 . ) in this case , none of the edges joining the vertex @xmath592 with the components @xmath513 can be a descent edge and @xmath604 can not be a descent in @xmath35 .
this proves that descent edges of @xmath494 are in bijection with descents in @xmath35 .
corollary [ cor : h_des_t ] and proposition [ prop : des = des - chordal ] imply the following formula .
[ cor : h_des_chordal ] for a connected chordal building set @xmath4 , the @xmath1-polynomial of the nestohedron @xmath14 equals @xmath610 where @xmath611 is the usual descent number of a permutation @xmath612 . let us give an additional nice property of nestohedra for chordal building sets .
[ prop : chordal_are_flag ] for a chordal building set @xmath4 , the nestohedron @xmath14 is a flag simple polytope . let us check that a chordal building set @xmath4 satisfies the condition in proposition [ prop : flag - characterization](iii ) . using the notation of that proposition , let @xmath613 .
let @xmath614 be the union of those subsets @xmath615 that have a nonempty intersection with @xmath616 .
since @xmath617 is in @xmath4 ( because @xmath4 is chordal ) , the subset @xmath614 should also be in @xmath4 ( by definition [ def : building - set - definition](b1 ) ) . clearly , @xmath618 is the union of all @xmath619 s and @xmath620 consists of a single @xmath619 .
it is also clear that @xmath621 either equals @xmath622 or is obtained from @xmath622 by removing a single subset @xmath619 .
it follows that there exists an index @xmath582 such that @xmath623 .
this gives an index @xmath27 such that @xmath624 and @xmath619 are both in @xmath4 , as needed .
let us give several examples which illustrate corollary [ cor : h_des_t ] and corollary [ cor : h_des_chordal ] .
the @xmath0- and @xmath1-numbers for the permutohedron and associahedron are well - known . for the complete building set @xmath625
the nestohedron @xmath14 is the usual permutohedron ; see example [ ex : complete_upper ] and ( * ? ? ?
8.1 ) . in this case
@xmath4-trees are linear orders on @xmath5 $ ] and @xmath4-permutations are all permutations @xmath626 .
thus , as noted before in example [ ex : complete_upper ] , the @xmath1-polynomial is the usual eulerian polynomial @xmath341 , and the @xmath1-numbers are the eulerian numbers @xmath627 .
let @xmath628 denote the graph which is a path having @xmath20 nodes labelled consecutively @xmath629 .
the graphical building set @xmath630 consists of all intervals @xmath631 $ ] , for @xmath632 .
the corresponding nestohedron @xmath633 is the usual stasheff associahedron ; see @xcite . in this case
, the @xmath4-trees correspond to unlabelled plane binary trees on @xmath20 nodes , as follows ; see ( * ? ? ?
8.2 ) for more details . a _
plane binary tree _ is a rooted tree with two types of edges ( left and right ) such that every node has at most one left and at most one right edge descending from it . from proposition [ prop :
b_trees_construction ] , one can see that a @xmath4-tree is a binary tree with @xmath20 nodes labelled @xmath634 so that , for any node , all nodes in its left ( resp .
, right ) branch have smaller ( resp . , bigger ) labels .
conversely , given an unlabelled plane binary tree , there is a unique way to label its nodes @xmath635 to create a @xmath4-tree , namely in the order of traversal of a _ depth - first search_. furthermore , note that descent edges correspond to right edges .
it is well - known that the number of unlabelled binary trees on @xmath20 nodes is equal to the _ catalan number _
@xmath636 , and the number of binary trees on @xmath20 nodes with @xmath637 right edges is the _ narayana number _ @xmath638 ; see ( * ? ? ?
6.19c and exer . 6.36 ) .
therefore , the @xmath1-numbers of the associahedron @xmath633 are the narayana numbers : @xmath639 , for @xmath640 .
it is also well - known that the @xmath0-numbers of the associahedron are @xmath641 .
this follows from a classical kirkman - cayley formula @xcite for the number of ways to draw @xmath433 noncrossing diagonals in an @xmath20-gon . in this case
, the @xmath4-permutations are exactly @xmath576-avoiding permutations @xmath256 .
recall that a permutation @xmath35 is _ @xmath576-avoiding _ if there is no triple of indices @xmath642 such that @xmath643 .
thus corollary [ cor : h_des_chordal ] says that the @xmath1-polynomial of the associahedron @xmath633 is @xmath644 where the sum runs over all @xmath576-avoiding permutations in @xmath151 .
this is consistent with the known fact that the narayana numbers count @xmath576-avoiding permutations according to their number of descents ; see simion ( * ? ? ?
* theorem 5.4 ) for a stronger statement . if @xmath645 is the @xmath20-cycle , then the nestohedron @xmath646 is the _ cyclohedron _ also introduced by stasheff ; see @xcite .
the @xmath1-polynomial of the cyclohedron was computed by simion ( * ? ? ?
* corollary 1 ) : @xmath647 note that the @xmath20-cycle ( for @xmath648 ) is not a chordal graph , so corollary [ cor : h_des_chordal ] does not apply to this case .
let @xmath649 .
let @xmath650 be the @xmath601-star graph with the central node @xmath651 connected to the nodes @xmath652 .
let us call the associated polytope @xmath653 the _ stellohedron .
_ from proposition
[ prop : b_trees_construction ] one sees that @xmath654-trees are in bijection with _ partial permutations _ of @xmath655 $ ] , which are ordered sequences @xmath656 of distinct numbers in @xmath655 $ ] , where @xmath657 .
the tree @xmath494 associated to a partial permutation @xmath656 has the edges @xmath658 where @xmath659 are the elements of @xmath655\setminus \{u_1,\dots , u_r\}$ ] .
the root of @xmath494 is @xmath660 if @xmath661 , or @xmath651 if @xmath662 . for @xmath661 ,
one has @xmath663 , where the descent number of a partial permutation is @xmath664 also for the tree @xmath494 associated with the empty partial permutation ( for @xmath662 ) one has @xmath665 .
corollary [ cor : h_des_t ] then says that @xmath666 where the first sum is over nonempty partial permutations @xmath35 of @xmath655 $ ] .
in particular , the total number of vertices of the stellohedron @xmath653 equals @xmath667 this sequence appears in sloan s on - line encyclopedia of integer sequences as a000522 . in this case , @xmath654-permutations are permutations @xmath668 such that @xmath651 appears before the first descent
. such permutations @xmath35 are in bijection with partial permutations @xmath669 of @xmath655 $ ] .
indeed , @xmath669 is the part of @xmath35 after the entry @xmath651 .
since our labelling of @xmath670 ( with the central node labelled @xmath651 ) is decreasing ( see example [ ex : decreasing_trees ] ) , corollary [ cor : h_des_chordal ] implies that the @xmath1-polynomial of the stellohedron @xmath671 is @xmath672 , where the sum runs over all such permutations @xmath668 .
this agrees with the above expression in terms of partial permutations .
let @xmath673 , \{i\}\mid i=1,\dots , n\}$ ] ( the collection of all intervals @xmath674 $ ] and singletons @xmath9 ) . this ( non - graphical ) building set is chordal . according to @xcite ,
the corresponding nestohedron @xmath675 is the stanley - pitman polytope from @xcite . by proposition [ prop : b_trees_construction ]
, @xmath676-trees have the following form @xmath677 . for an increasing sequence @xmath39 of positive integers @xmath678 , construct the tree @xmath677 on @xmath5 $ ] with the root at @xmath679 and the chain of edges @xmath680 ; also , for each @xmath681\setminus i$ ] , one has the edge @xmath682 where @xmath683 is the minimal element of @xmath39 such that @xmath684 . in this case , @xmath676-permutations are permutations @xmath256 such that @xmath685 , for some @xmath686 .
using @xmath676-trees or @xmath676-permutations one can easily deduce that the @xmath1-polynomial of the stanley - pitman polytope is @xmath687 .
this is not surprising since @xmath675 is combinatorially isomorphic to the @xmath379-dimensional cube .
recall that the @xmath2-vector @xmath688 of a @xmath22-dimensional simple polytope is defined via its @xmath1-polynomial as @xmath689 ; and the @xmath2-polynomial is @xmath690 ; see section [ ssec : flag_simple_gamma ] .
the main result of this section is a formula for the @xmath2-polynomial of a chordal nestohedron as a descent - generating function ( or peak - generating function ) for some set of permutations .
this implies that gal s conjecture ( conjecture [ conj : gal - conjecture ] ) holds for this class of flag simple polytopes .
we review here the beautiful construction of shapiro , woan , and getu @xcite that leads to a nonnegative formula for the @xmath2-vector of the usual permutohedron .
this subsection also serves as a warm - up for a more general construction in the following subsection .
some notation is necessary .
recall that a _ descent _ in a permutation @xmath171 is a pair @xmath604 such that @xmath554 , where @xmath691 $ ] .
a _ final descent _ is when @xmath692 , and a _
double descent _ is a pair of consecutive descents , i.e. a triple @xmath693 .
additionally , define a _
peak _ of @xmath35 to be an entry @xmath40 for @xmath694 such that @xmath695 . here
( and below ) set @xmath696 and so a peak can occur in positions @xmath3 or @xmath20 . on the other hand ,
a _ valley _ of @xmath35 is an entry @xmath40 for @xmath697 such that @xmath698 .
the _ peak - valley sequence _ of @xmath35 is the subsequence in @xmath35 formed by all peaks and valleys .
let @xmath43 denote the set of permutations in @xmath151 which do not contain any final descents or double descents .
let @xmath699 denote the number of peaks in a permutation @xmath35 .
it is clear that @xmath700 , for permutations @xmath701 ( and only for these permutations ) .
[ th : cdg ] ( cf .
* proposition 4 ) ) the @xmath2-polynomial of the usual permutohedron @xmath702 is @xmath703 let us calculate the @xmath2-polynomial of the two dimensional permutohedron @xmath704 .
one has @xmath705 . of these , @xmath706 has one peak ( and no descents ) , and @xmath707 and @xmath708 have two peaks ( and one descent ) .
therefore , the @xmath2-polynomial is @xmath709 .
say that an entry @xmath40 of @xmath35 is an _ intermediary _
entry if @xmath40 is not a peak or a valley .
say that @xmath40 is an _ ascent - intermediary entry _ if @xmath710 and that it is a _ descent - intermediary entry _ if @xmath711 .
( here again one should assume that @xmath696 . )
note that the set @xmath43 is exactly the set of permutations in @xmath151 without descent - intermediary entries .
it is convenient to graphically represent a permutation @xmath171 by a piecewise linear `` mountain range '' @xmath712 obtained by connecting the points @xmath713 , @xmath714 , @xmath715 , , @xmath716 , @xmath717 on @xmath718 by straight line intervals , for some @xmath719 ; see figure [ fig : perm - graph ] .
then peaks in @xmath35 correspond to local maxima of @xmath712 , valleys correspond to local minima of @xmath712 , ascent - intermediary entries correspond to nodes on ascending slopes of @xmath712 , and descent - intermediary entries correspond to nodes on descending slopes of @xmath712 .
for example , the permutation @xmath720 shown in figure [ fig : perm - graph ] has three peaks @xmath721 , two valleys @xmath722 , one ascent - intermediary entry @xmath723 , and four descent - intermediary entries @xmath724 .
its peak - valley sequence is @xmath725 . for @xmath720,height=153 ] as noted in section [ ssec : descents_of_tree_posets ] , the @xmath1-polynomial of the permutohedron is the descent - generating function for permutations in @xmath151 ( the eulerian polynomial ) . in order to prove theorem [ th : cdg ]
, one constructs an appropriate partitioning of @xmath151 into equivalence classes ( cf .
remark [ boolean - classes - remark ] ) , where each equivalence classes has exactly one element from @xmath43 . to describe the equivalence classes of permutations
, one must introduce some operations on permutations .
[ def : leap ] let us define the _ leap _ operations @xmath726 and @xmath727 that act on permutations .
informally , the permutation @xmath728 is obtained from @xmath35 by moving an intermediary node @xmath564 on the mountain range @xmath712 directly to the right until it hits the next slope of @xmath712 .
the permutation @xmath729 is obtained from @xmath35 by moving @xmath564 directly to the left until it hits the next slope of @xmath712 .
more formally , for an intermediary entry @xmath730 in @xmath35 , the permutation @xmath728 is obtained from @xmath35 by removing @xmath564 from the @xmath27-th position and inserting @xmath564 in the position between @xmath731 and @xmath732 , where @xmath118 is the minimal index such that @xmath585 and @xmath564 is between @xmath731 and @xmath732 , i.e. , @xmath733 or @xmath734 .
the leap operation @xmath726 is not defined if all entries following @xmath564 in @xmath35 are less than @xmath564 .
similarly , the inverse operation @xmath729 is given by removing @xmath564 from the @xmath27-th position in @xmath35 and inserting @xmath564 between @xmath735 and @xmath736 , where @xmath433 is the maximum index such that @xmath737 and @xmath564 is between @xmath735 and @xmath736 .
the operation @xmath727 is is not defined if all entries preceding @xmath564 in @xmath35 are less than @xmath564 .
for example , for the permutation @xmath35 shown on figure [ fig : perm - graph ] , one has @xmath738 and @xmath739 .
clearly , if @xmath564 is an ascent - intermediary entry in @xmath35 then @xmath564 is a descent - intermediary entry in @xmath740 , and vise versa .
note that if @xmath564 is an ascent - intermediary entry in @xmath35 , then @xmath728 is always defined , and if @xmath564 is a decent - intermediary entry , then @xmath741 is always defined .
let us also define the _ hop _
operations @xmath742 on permutations . for an ascent - intermediary entry @xmath564 in @xmath35 , define @xmath743 ; and , for a descent - intermediary entry @xmath564 in @xmath35 , define @xmath744 .
for example , for the permutation @xmath35 shown on figure [ fig : perm - graph ] , the permutation @xmath745 is obtained by moving the descent - intermediary entry @xmath229 to the left to the first ascending slope , and @xmath746 is obtained by moving the ascent - intermediary entry @xmath723 to the right to the last descending slope .
note that leaps and hops never change the shape of the mountain range @xmath712 , that is , they never change the peak - valley sequence of @xmath35 .
they just move intermediary nodes from one slope of @xmath712 to another .
it is quite clear from the definition that all leap and hop operations pairwise commute with each other .
it is also clear that two hops @xmath742 get us back to the original permutation .
[ lem : hops_permutohedron ] for intermediary entries @xmath564 and @xmath565 in @xmath35 , one has @xmath747 and @xmath748 .
thus the hop operations @xmath742 generate the action of the group @xmath749 on the set of permutations with a given peak - valley sequence , where @xmath601 is the number of intermediary entries in such permutations .
let us say that two permutations are _ hop - equivalent _ if they can be obtained from each other by the hop operations @xmath742 for various @xmath564 s .
the partitioning of @xmath151 into hop - equivalence classes allows us to prove theorem [ th : cdg ] .
the number @xmath611 of descents in @xmath35 equals the number of peaks in @xmath35 plus the number of descent - intermediary entries in @xmath35 minus @xmath3 ( because the last entry is either a peak or a descent - intermediary entry , but it does not contribute a descent ) .
notice that if @xmath564 is an ascent - intermediary ( resp .
, descent - intermediary ) entry in @xmath35 then the number of descent - intermediary entries in @xmath750 increases ( resp . , decreases ) by 1 and the number of peaks does not change .
if @xmath171 has @xmath751 peaks then it has @xmath752 valleys and @xmath753 intermediary entries .
lemma [ lem : hops_permutohedron ] implies that the hop - equivalence class @xmath131 of @xmath35 involves @xmath754 permutations .
moreover , the descent - generating function for these permutations is @xmath755 .
each hop - equivalence class has exactly one representative @xmath669 without descent - intermediary entries , that is @xmath756 .
thus , summing the contributions of hop - equivalence classes , one can write the @xmath1-polynomial of the permutohedron as @xmath757 comparing this to the definition of the @xmath2-polynomial , one derives the theorem . according to proposition [ prop : chordal_are_flag ] ,
nestohedra for chordal building sets are flag simple polytopes .
thus gal s conjecture ( conjecture [ conj : gal - conjecture ] ) applies .
this section proves this conjecture and present a nonnegative combinatorial formula for @xmath2-polynomials of such nestohedra as peak - generating functions for some subsets of permutations .
let @xmath4 be a connected chordal building set on @xmath5 $ ] . recall that @xmath34 is the set of @xmath4-permutations
; see definition [ def : characterization ] .
let @xmath758 be the subset of @xmath4-permutations which have no final descent or double descent .
the following theorem is the main result of this section .
[ th : gamma - chordal ] for a connected chordal building @xmath4 on @xmath5 $ ] , the @xmath2-polynomial of the nestohedron @xmath14 is the peak - generating function for the permutations in @xmath759 : @xmath760 as noted earlier , @xmath700 for @xmath701 . the proof of theorem [ th : gamma - chordal ] will be an extension of the proof given for the @xmath2-vector of the permutohedron in section [ ssec : permutohedron - gamma ] .
recall that corollary [ cor : h_des_chordal ] interprets the @xmath1-polynomial of @xmath14 as the descent - generating function for @xmath4-permutations @xmath761 .
theorem [ th : gamma - chordal ] will be proven by constructing an appropriate partitioning of the set @xmath34 into equivalence classes , where each equivalence class has exactly one representative from @xmath759 .
as before , one uses ( suitably generalized ) hop operations to describe equivalence classes of elements of @xmath34 .
one needs powers of the leap operations @xmath762 , for @xmath763 , and @xmath764 , for @xmath765 ; see definition [ def : leap ] .
in other words , for @xmath766 , @xmath767 is obtained from @xmath35 by moving the intermediary entry @xmath564 to the right until it hits the @xmath143-th slope from its original location ; and , for @xmath768 , by moving @xmath564 to the left until it hits @xmath769-th slope from its original location .
clearly , @xmath767 is defined whenever @xmath143 is in a certain integer interval @xmath770 $ ] .
it is also clear that , if @xmath564 is an ascent - intermediary entry in @xmath35 , then @xmath564 is ascent - intermediary in @xmath771 for even @xmath143 and @xmath564 is descent - intermediary in @xmath771 for odd @xmath143 , and vice versa if @xmath564 is descent - intermediary in @xmath35 .
note that for a @xmath4-permutation @xmath612 , the permutations @xmath771 may no longer be @xmath4-permutations .
the next lemma ensures that at least some of them will be @xmath4-permutations .
[ lem : leaps ] let @xmath4 be a chordal building on @xmath5 $ ] .
suppose that @xmath612 is a @xmath4-permutation .
\(1 ) if @xmath564 is an ascent - intermediary letter in @xmath35 , then there exists an odd positive integer @xmath766 such that @xmath772 and @xmath773 , for all @xmath774 .
\(2 ) if @xmath564 is a descent - intermediary letter in @xmath35 , then there exists an odd negative integer @xmath768 such that @xmath772 and @xmath773 , for all @xmath775 .
the proof of lemma [ lem : leaps ] will require some preparatory notation and observations . for a permutation @xmath256 and @xmath776 $ ] such that @xmath777 , let @xmath778 be the set of all entries in @xmath35 which are located to the left of @xmath564 and are greater than or equal to @xmath564 ( including the entry @xmath564 itself ) .
the arrow in this notation refers to our graphical representation of a permutation as a mountain range @xmath712 : the set @xmath779 is the set of entries in @xmath35 located to the north - west of the entry @xmath564 . according to definition [ def : characterization ] ,
the set @xmath34 is the set of permutations @xmath35 such that , for @xmath36 , there exists @xmath780 such that both @xmath40 and @xmath543 are in @xmath39 and @xmath781 . if @xmath4 is chordal , then @xmath782 $ ] also belongs to @xmath4 ( see definition [ def : chordal_building ] ) and satisfies the same properties .
clearly @xmath783 .
thus , for a chordal building set , one can reformulate definition [ def : characterization ] of @xmath4-permutations as follows .
[ lem : b - permutations - chordal ] let @xmath4 be a chordal building set .
then @xmath34 is the set of permutations @xmath256 such that for any @xmath776 $ ] , the elements @xmath564 and @xmath784 are in the same connected component of @xmath785 .
equivalently , there exists @xmath37 such that @xmath786 , @xmath787 , and @xmath788 .
let us now return to the setup of lemma [ lem : leaps ] .
there are 2 possible reasons why the permutation @xmath789 may no longer be a @xmath4-permutation , that is , fail to satisfy the conditions in lemma [ lem : b - permutations - chordal ] : * it is possible that the entry @xmath564 and the entry @xmath790 are in different connected components of @xmath791 .
* it is also possible that another entry @xmath792 in @xmath669 and @xmath793 are in different connected components of @xmath794 .
let us call these two types of failure _ a - failure _ and _ b - failure . _
the following auxiliary result is needed .
[ lem : leaps_aux ] let us use the notation of lemma [ lem : leaps ] .
\(1 ) for left leaps @xmath789 , @xmath768 , one can never have a b - failure .
\(2 ) for the maximal left leap @xmath795 , where the entry @xmath564 goes all the way to the left , one can not have an a - failure .
\(3 ) for the maximal right leap @xmath796 , where the entry @xmath564 goes all the way to the right , one can not have an a - failure .
\(4 ) let @xmath797 and @xmath798 , for @xmath799 , be two adjacent leaps such that @xmath564 is descent - intermediary in @xmath669 ( and , thus , @xmath564 is ascent - intermediary in @xmath800 ) .
then there is an a - failure in @xmath669 if and and only if there is an a - failure in @xmath800 .
\(1 ) since @xmath761 , there is a subset @xmath37 that contains both @xmath565 and @xmath801 and such that @xmath802 . the same subset @xmath39 works for @xmath669 because @xmath803 or @xmath804 .
\(2 ) in this case , @xmath564 is greater than all preceding entries in @xmath669 , so @xmath805 .
( 3 ) in this case , @xmath564 is greater than all following entries in @xmath669 . the interval @xmath806 $ ] contains both @xmath564 and @xmath790 , @xmath807 , and @xmath37 because @xmath4 is chordal .
\(4 ) in this case , all entries between the position of @xmath564 in @xmath669 and the position of @xmath564 in @xmath800 are less than @xmath564 .
thus @xmath808 .
so @xmath669 has an a - failure if and only if @xmath800 has an a - failure .
it is easier to prove the second part of the lemma .
\(2 ) by parts ( 1 ) and ( 2 ) of lemma [ lem : leaps_aux ] , there exists a negative @xmath143 such that @xmath772 .
let us pick such an @xmath143 with minimal possible absolute value
. then @xmath143 should be odd , by part ( 4 ) of lemma [ lem : leaps_aux ] , which proves ( 2 ) .
\(1 ) suppose that there is an entry @xmath792 in the permutation @xmath35 such that @xmath565 and @xmath809 are in different connected components of @xmath810 .
in this case , @xmath811 , that is @xmath812 and @xmath565 is located to the right of @xmath564 in @xmath35 .
( otherwise , @xmath565 and @xmath601 are in different connected components of @xmath813 , which is impossible because @xmath35 is a @xmath4-permutation . )
let us pick the leftmost entry @xmath565 in @xmath35 that satisfies this condition .
then the permutation @xmath797 has a b - failure if the letter @xmath564 moves to the right of this entry @xmath565 ; and @xmath669 has no b - failure if @xmath564 stays to the left of @xmath565 . by our assumptions , @xmath564 stays to the left of @xmath565 in @xmath814 , so such a @xmath669 exists .
let @xmath797 be the maximal right leap ( i.e. , with maximal @xmath766 ) such that the entry @xmath564 stays to the left of @xmath565 .
then all entries in @xmath669 between the positions of @xmath564 and @xmath565 should be less than @xmath564 .
thus @xmath815 .
since @xmath612 , there is an @xmath37 such that @xmath816 and @xmath802 .
this subset @xmath39 should also contain the entry @xmath564 .
( otherwise , @xmath565 and @xmath601 would be in the same connected component @xmath817 , contrary to our choice of @xmath565 . )
thus @xmath818\in\b$ ] contains both @xmath564 and @xmath601 and @xmath819 .
this means that there is no a - failure in @xmath669 .
thus @xmath820 . if there is no entry @xmath565 in @xmath35 as above , then none of the permutations @xmath771 has a b - failure . in this case @xmath821 by part ( 3 ) of lemma [ lem : leaps_aux ] . in all cases
, there exists a positive @xmath143 such that @xmath822 and only a - failures are possible in @xmath823 , for @xmath774 .
let us pick the minimal such @xmath143
. then @xmath143 should be odd by part ( 4 ) of lemma [ lem : leaps_aux ] , as needed .
let us define the _
operations @xmath824 . for a @xmath4-permutation @xmath35 with an ascent - intermediary ( resp .
, descent - intermediary ) entry @xmath564 , the permutation @xmath825 is the right leap @xmath826 , @xmath766 ( resp .
, the left leap @xmath826 , @xmath768 ) with minimal possible @xmath827 such that @xmath669 is a @xmath4-permutation .
informally , @xmath825 is obtained from @xmath35 by moving the node @xmath564 on its mountain range @xmath712 directly to the right if @xmath564 is ascent - intermediary in @xmath35 , or directly left if @xmath564 is descent - intermediary in @xmath35 ( possibly passing through several slopes ) until one hits a slope and obtain a @xmath4-permutation .
lemma [ lem : leaps ] says that the @xmath4-hop @xmath825 is well - defined for any intermediary entry @xmath564 in @xmath35 .
it also says that if @xmath564 is ascent - intermediary in @xmath35 then @xmath564 is descent - intermediary in @xmath825 , and vice versa .
moreover , according to that lemma , @xmath828 .
let @xmath11 be the decreasing tree shown on figure [ fig : behop ] .
then the graphical building @xmath829 is chordal ; see example [ ex : decreasing_trees ] .
figure [ fig : behop ] shows several @xmath4-hops of the @xmath4-permutation @xmath830 : @xmath831 -permutation @xmath35 and some @xmath4-hops , height=153 ] let us now show that the @xmath4-hop operations pairwise commute with each other . let @xmath564 and @xmath565 be two intermediary entries in a @xmath4-permutation @xmath35
. then @xmath832 .
let us first assume that both @xmath564 and @xmath565 are descent - intermediary entries in @xmath35 . without loss of generality assume that @xmath833 . in this case @xmath834 and @xmath835 for some _ negative _ odd @xmath143 and @xmath582 , that is the entries @xmath564 and @xmath565 of @xmath35
are moved to the left . according to lemma [
lem : leaps_aux](1 ) , in this case one does not need to worry about b - failures .
in other words , @xmath825 is the first left leap @xmath771 ( i.e. , with minimal @xmath836 ) that has no a - failure .
similarly , @xmath837 is the first left leap @xmath838 without a - failures ( where a - failures concern the entry @xmath565 ) .
since a - failures for permutations @xmath839 , @xmath840 , are described in terms of the set @xmath841 $ ] , moving the entry @xmath812 in @xmath35 will have no effect on these a - failures .
thus , for the permutation @xmath842 , one has @xmath843 with exactly the same @xmath143 as in @xmath844 .
however , for permutations @xmath845 , @xmath840 , the sets @xmath846 might change if one first performs the operation @xmath824 to @xmath35 .
namely , let @xmath847 and @xmath848 then @xmath849 if @xmath564 is located to the left of @xmath565 in @xmath850 and @xmath564 is located to the right of @xmath565 in @xmath669 ( and @xmath851 otherwise ) .
notice that one always has @xmath852 , since this maximum is the maximal peak preceding @xmath565 in @xmath669 ( or in @xmath850 ) , and leaps and hops have no affect on the peaks .
if @xmath565 and @xmath601 are in the same connected component of @xmath853 then they are also in the same connected component of @xmath854 , that is if there is no a - failure for @xmath669 then there is no a - failure for @xmath850 .
suppose that there is no a - failure for @xmath850 but there is an a - failure for @xmath669 .
then the sets @xmath846 and @xmath855 have to be different .
that means that @xmath564 is located to the left of @xmath565 in @xmath850 and @xmath564 is located to the right of @xmath565 in @xmath669 .
let @xmath39 be the element @xmath37 such that @xmath816 and @xmath856
. then @xmath39 should contain the entry @xmath564 .
( otherwise , @xmath857 and there would be no a - failure for @xmath669 . )
let @xmath858 be the left leap with maximal possible @xmath859 such that the position of @xmath564 in @xmath860 is located to the right of the position of @xmath565 in @xmath850 .
since @xmath861 , it follows that @xmath862 .
in other words , if one starts moving to the right from the node @xmath565 along the mountain range @xmath863 , the ( ascending ) slope that first crosses the level @xmath564 is the place where the entry @xmath564 is located in @xmath860 . note that @xmath864 is odd because @xmath564 should be an ascent - intermediary entry in @xmath860 ; in particular @xmath865 .
since all entries in @xmath860 located between the position of @xmath565 in @xmath850 and the position of @xmath564 in @xmath860 are less than @xmath564 , one deduces that @xmath866 = \{\hat w \nw a\}$ ] .
thus the subset @xmath867 $ ] has three important properties : it lies in @xmath4 ( because @xmath4 is chordal ) ; it contains both @xmath564 and @xmath868 ; and it is a subset of @xmath869 .
it follows that there is no a - failure in @xmath860 .
this contradicts the fact that @xmath870 is the first left leap that has no a - failure .
thus @xmath669 has an a - failure if and only if @xmath850 has an a - failure .
it follows that @xmath871 with exactly the same @xmath582 as in @xmath872 .
this proves that @xmath873 , in the case when both @xmath564 and @xmath565 are descent - intermediary in @xmath35 .
let us now show that the general case easily follows .
suppose that , say , @xmath564 is ascent - intermediary and @xmath565 is descent - intermediary in @xmath35 .
then , for @xmath874 both @xmath564 and @xmath565 are descent - intermediary .
one has @xmath875 .
thus @xmath876 .
applying @xmath824 to both sides , one deduces @xmath877 .
the other cases are similar .
thus the @xmath4-hop operations @xmath824 generate the action of the group @xmath749 on the set of @xmath4-permutations with a given peak - valley sequence , where @xmath601 is the number of intermediary entries in such permutations .
let us say that two @xmath4-permutation are _ @xmath4-hop - equivalent _ if they can be obtained from each other by the @xmath4-hop operations @xmath824 for various @xmath564 s .
this gives the partitioning of the set of @xmath4-permutations into @xmath4-hop - equivalence classes .
one can now prove theorem [ th : gamma - chordal ] by literally repeating the argument in the proof of theorem [ th : cdg ] .
for a @xmath4-permutation @xmath612 with @xmath751 , the descent - generating function of the @xmath4-hop - equivalence class @xmath131 of @xmath35 is @xmath755 .
each @xmath4-hop - equivalence class has exactly one representative without descent - intermediary entries , that is , in the set @xmath759 .
thus the @xmath1-polynomial of the nestohedron @xmath878 ( see corollary [ cor : h_des_chordal ] ) is @xmath879 comparing this to the definition of the @xmath2-polynomial , one derives the theorem .
gal s conjecture holds for all graph - associahedra corresponding to chordal graphs .
[ prop : gamma_assoc ] the @xmath2-polynomial of the associahedron @xmath633 is @xmath880 where @xmath881 is the @xmath143-th catalan number .
the @xmath2-polynomial of the cyclohedron @xmath882 is @xmath883 these two formulas can be derived from the expressions for the corresponding @xmath1-polynomials ( see sections [ ssect : path - graph - example-1 ] and [ ssect : cyclo ] ) using standard quadratic transformations of hypergeometric series ; e.g. , see ( * ? ? ?
* lemma 4.1 ) .
on the other hand , let us mention the following three combinatorial interpretations of the @xmath2-vector for the associahedron @xmath633 .
it is known that the narayana polynomial which is the @xmath1-polynomial of @xmath633 is also the rank generating function for the well - studied _ lattice of noncrossing partitions _
@xmath884 . an explicit _ symmetric chain decomposition _ for @xmath884
was given by simion and ullman @xcite , who actually produced a much stronger decomposition of @xmath884 into disjoint boolean intervals placed symmetrically about the middle rank(s ) of @xmath884 .
their decomposition contains exactly @xmath885 such boolean intervals of rank @xmath886 for each @xmath887 , which immediately implies the formula for the @xmath2-polynomial ; see ( * ? ? ?
* corollary 3.2 ) . by section [ ssect : path - graph - example-1 ]
, the @xmath1-polynomial of @xmath633 counts plane binary trees on @xmath20 nodes according to their number of right edges .
there is a natural map from binary trees to _
full _ binary trees , i.e. , those in which each node has zero or two children : if a node has a unique child , contract this edge from the node to its child .
if the original binary tree @xmath494 has @xmath20 nodes , then the resulting full binary tree @xmath544 will have @xmath888 nodes , @xmath889 edges and @xmath143 right edges for some @xmath890 .
there are @xmath891 such full binary trees for each @xmath143 .
given such a full binary tree @xmath544 , one can produce all of the binary trees in its preimage by inserting @xmath886 more nodes and deciding if they create left or right edges .
one chooses the locations of these nodes from @xmath888 choices , either an edge of the full binary tree they will subdivide or located above the root , giving @xmath892 possible locations . thus the generating function with respect to the number of right edges for the preimage of @xmath544 is @xmath893 , where the term @xmath894 comes from choosing whether each of the new nodes creates a left or a right edge .
it follows that the generating function for all binary trees on @xmath20 nodes is @xmath895 , where @xmath891 counts full binary trees .
this implies the needed expression for the @xmath2-vector of the associahedron @xmath633 .
equivalently , one can describe the subdivision of all binary trees into classes where two binary trees are in the same class if they can be obtained from each other by switches of left and right edges coming from single child nodes .
then one gets exactly @xmath896 classes having @xmath894 as its generating function counting number of right edges , for each @xmath897 .
this proof is based on our general approach to @xmath2-vectors of chordal nestohedra . according to section [ ssect : path - graph - example-1 ] , @xmath4-permutations for the associahedron are @xmath576-avoiding permutations and @xmath1-polynomial is equal to the sum @xmath898 over all @xmath576-avoiding permutations @xmath256 . by theorem
[ th : gamma - chordal ] , @xmath899 equals the number of @xmath576-avoiding permutations with no descent - intermediary elements and @xmath900 peaks .
the ( flattenings of ) peak - valley sequences of such permutations are exactly @xmath576-avoiding _ alternating permutations _ in @xmath901 , that is 312-avoiding permutations @xmath547 such that @xmath902 .
it is known that the number of such permutations equals the catalan number @xmath891 ; see ( * ? ? ? * theorem 2.2 )
. then there are @xmath903 ways to insert the remaining @xmath886 descent - intermediary elements .
our goal in this section is to compute a generating function that computes the @xmath1-polynomials of all graph - associahedra in which the graph is a tree having at most one _ branched _ vertex ( i.e. , a vertex of valence @xmath573 or more ) . first recall ( see section [ ssect : path - graph - example-1 ] ) that the @xmath1-numbers of the associahedron @xmath633 are the narayana numbers @xmath904 , and the @xmath1-polynomial of the associahedron is the _ narayana polynomial _ : @xmath905 recall the well - known recurrence relation and the generating function for the narayana polynomials @xmath906 .
the recurrence for the @xmath0-polynomials @xmath907 given by theorem [ th : f_b_recurrence ] can be written as follows . when one removes @xmath433 vertices from the @xmath20-path , it splits into @xmath908 ( possibly empty ) paths .
thus one obtains @xmath909 where the sum is over @xmath910 such that @xmath911 . here
one assumes that @xmath912 .
let @xmath913 be the generating function for the narayana polynomials : @xmath914 the recurrence relation is equivalent to the following well - known functional equation : @xmath915 see ( * ? ? ?
6.36b ) .
trees with at most one branched vertex have the following form . for @xmath916 ,
let @xmath917 be the graph obtained by attaching @xmath433 chains of lengths @xmath918 to one central node .
for example , @xmath919 is the graph with a single node and @xmath920 is the @xmath433-star graph @xmath921 .
[ th : t_formula ] one has the following generating function for the @xmath1-polynomials of graph - associahedra @xmath922 for the graphs @xmath917 : @xmath923 where @xmath924 , and @xmath913 is the generating function for the narayana polynomials from .
this theorem immediately implies the following formula from @xcite .
* proposition 8.7 ) the generating function for the number of vertices in the graph - associahedron @xmath922 is @xmath925 where @xmath926 is the generating function for the catalan numbers .
the claim is obtained from theorem [ th : t_formula ] in the limit @xmath927 .
note however that one needs to use lhpital s rule before plugging in @xmath928 .
the first proof of theorem [ th : t_formula ] is fairly direct , using corollary [ cor : h_des_t ] and the solution to simon newcomb s problem .
the second uses theorem [ th : f_recurrence_zelevinsky ] to set up a system of pde s and solve them ; it has the advantage of producing a generating function for the @xmath1-polynomials of one further family of graph - associahedra .
let us first review simon newcomb s problem and its solution .
let @xmath929 be a permutation of the multiset @xmath930 , that is , each @xmath27 appears in @xmath35 exactly @xmath931 times , for @xmath932 .
a _ descent _ in @xmath35 is an index @xmath27 such that @xmath554 .
let @xmath611 denote the number of descents in @xmath35 .
_ simon newcomb s problem _ is the problem of counting permutations of a multiset with a given number of descents , see ( * ? ? ?
iv , ch .
iv ) and ( * ? ? ?
let us define the _ multiset eulerian polynomial _ as @xmath933 where the sum is over all permutations @xmath35 of the multiset @xmath930
. by convention , set @xmath934 .
in particular , the polynomial @xmath935 is the usual eulerian polynomial .
it is clear that @xmath936 , the total number of multiset permutations .
a solution to simon newcomb s problem can be expressed by the following generating function for the @xmath937 .
[ prop : simon_newcomb ] ( * ? ? ?
4.2.13 ) one has @xmath938 theorem [ th : t_formula ] then immediately follows from proposition [ prop : simon_newcomb ] and the following proposition .
[ prop : multi_eulerian ] the generating function for the @xmath1-polynomials of the polytopes @xmath922 equals @xmath939 let us label nodes of the graph @xmath917 by integers in @xmath5 $ ] , where @xmath940 , so that the first chain is labelled by @xmath941 , the second chain is labelled by @xmath942 , etc .
, with all labels increasing towards the central node , and finally the central node has the maximal label @xmath20 .
let @xmath494 be a @xmath917-tree .
suppose that the root @xmath143 of @xmath494 belongs to the @xmath535-st chain of the graph @xmath917 .
if one removes the node @xmath143 from the graph @xmath917 , then the graph decomposes into 2 connected components , one of which is a chain @xmath943 and the other is @xmath944 , where @xmath945 and all other indices are the same as before .
( the first component is empty if @xmath946 . ) according to proposition [ prop : b_trees_construction ] , the tree @xmath494 is obtained by attaching a @xmath943-tree @xmath596 and a @xmath944-tree @xmath544 to the root @xmath143 .
( here one assumes that there is one empty @xmath947-tree @xmath596 , for @xmath946 . )
let us repeat the same procedure with the tree @xmath544 .
assume that its root belongs to the @xmath536-nd chain and split it into a @xmath948-tree @xmath949 and a tree @xmath950 .
then repeat this procedure with @xmath950 , etc .
keep on doing this until one gets a tree @xmath951 with the root at the central node @xmath20 .
finally , if one removes the central node @xmath20 from @xmath951 , then it splits into @xmath433 trees @xmath952 such that @xmath953 is a @xmath954-tree , for @xmath955 .
so each @xmath917-tree @xmath494 gives us the following data : 1 .
a sequence @xmath956^m$ ] ; 2 . a @xmath957-tree @xmath525 , for @xmath958 ; 3 . a @xmath954-tree @xmath953 , for @xmath955 .
this data satisfies the following conditions : @xmath959 where @xmath960 are the standard basis vectors in @xmath961 .
conversely , data of this form gives us a unique @xmath917-tree @xmath494 .
the number of descents in the tree @xmath494 is @xmath962 where @xmath545 is the number of nonempty trees among @xmath596 , , @xmath963 , @xmath964 , , @xmath965 .
indeed , all descents in trees @xmath525 and @xmath953 correspond to descents in @xmath494 , each nonempty tree @xmath525 or @xmath953 gives an additional descent for the edge that attaches this tree , and descents in @xmath35 correspond to descent edges that attach trees @xmath966 .
let us fix a sequence @xmath967 .
for @xmath968 $ ] , let @xmath931 be the number of times the integer @xmath27 appears in @xmath35 .
in other words , @xmath35 is a permutation of the multiset @xmath930 . then the total contribution to the generating function @xmath969 of trees @xmath494 whose
data involve @xmath35 is equal to @xmath970 .
indeed , the term 1 in @xmath971 corresponds to an empty tree , and the term @xmath972 corresponds to nonempty trees , which contribute one additional descent .
the term @xmath973 comes from the @xmath931 trees @xmath974 , where @xmath975 are all occurrences of @xmath27 in @xmath35 .
finally , additional @xmath3 s in the exponents of @xmath976 s come from the trees @xmath952 . summing this expression over all permutations @xmath35 of the multiset @xmath930 and
then over all @xmath977 , one obtains the needed expression for the generating function @xmath969 .
one can dualize all definitions , statements , and arguments in this section , as follows .
an equivalent dual formulation to theorem [ th : t_formula ] says @xmath978 where @xmath979 .
the equivalence to theorem [ th : t_formula ] follows from the relation @xmath980 , which is a reformulation of the functional equation .
the _ dual multiset eulerian polynomial _ is @xmath981 , where the sum is over permutations @xmath35 of the multiset @xmath982 , @xmath983 , and @xmath984 is the number of _ weak descents _ in the multiset permutation @xmath35 , that is , the number of indices @xmath27 for which @xmath985 . the bijection which reverses the word @xmath35 shows that @xmath986 and consequently one has an equivalent formulation of the solution to simon newcomb s problem : @xmath987 then one can modify the proof of proposition [ prop : multi_eulerian ] , by switching the labels @xmath988 in the graph @xmath917 , and applying a similar argument to show @xmath989
this section rederives theorem [ th : t_formula ] using theorem [ th : f_recurrence_zelevinsky ] .
it also calculates the generating function for @xmath0-polynomials of graph - associahedra corresponding to another class of graphs , the _ hedgehog graphs _ defined below .
recall that @xmath990 is the path with @xmath20 nodes , and @xmath917 is the graph obtained by attaching the paths @xmath991 , , @xmath992 to a central node .
let us also define the _ hedgehog graph _ @xmath993 as the graph obtained from the disjoint union of the chains @xmath994 ,
, @xmath992 by adding edges of the complete graph between the first vertices of all chains . for example
, @xmath995 is the empty graph , @xmath996 , and @xmath997 is a graph with @xmath998 vertices obtained from the complete graph @xmath999 by adding a `` leaf '' edge hanging from each of the @xmath433 original nodes . by convention , for the empty graph , one has @xmath1000 .
theorem [ th : f_recurrence_zelevinsky ] gives the following recurrence relation for @xmath0-polynomials of path graphs : @xmath1001 indeed , there are @xmath1002 connected @xmath143-element subsets @xmath39 of nodes of @xmath990 , the deletion @xmath1003 is isomorphic to @xmath1004 , and the contraction @xmath1005 is isomorphic to @xmath1006 . for graphs @xmath917
, theorem [ th : f_recurrence_zelevinsky ] gives the following recurrence relation @xmath1007 \displaystyle \qquad \qquad \qquad \qquad \qquad \qquad \qquad \qquad + \sum \tilde f_{t_{b_1,\dots , b_k}}(t)\cdot \tilde f_{h_{a_1-b_1,\dots , a_k - b_k}}(t ) , \end{array}\ ] ] where the second sum is over @xmath1008 such that @xmath1009 , for @xmath932 .
indeed , a connected subset @xmath39 of vertices of @xmath1010 either belongs to one of the chains @xmath1011 , or contains the central node . in the first case ,
the restriction is @xmath1012 and the contraction is @xmath1013 , where @xmath1014 . in the second case , the restriction @xmath13 has the form @xmath1015 and the contraction is @xmath1016 .
similarly , for hedgehog graphs @xmath993 , one obtains the recurrence relation @xmath1017 \displaystyle \qquad \qquad \qquad \qquad \qquad \qquad \qquad \qquad + \sum \tilde f_{h_{b_1,\dots , b_k}}(t)\cdot \tilde f_{h_{a_1-b_1,\dots , a_k - b_k}}(t ) , \end{array}\ ] ] where the second sum is over @xmath1008 such that @xmath1009 , for @xmath932 . in all cases
one has the initial conditions @xmath1018 , except @xmath1019 .
the above recurrence relations can be written in a more compact form using these generating functions : @xmath1020 note that @xmath1021 and @xmath1022 are related to generating functions from section [ sec : one - branching ] : @xmath1023 the above recurrence relations can be expressed as the following partial differential equations with initial conditions at @xmath1024 : @xmath1025 one can actually solve these partial differential equations for arbitrary initial conditions , as follows .
[ eq : diffeq_fghr ] the solutions @xmath1026 , @xmath1027 , @xmath1028 , and @xmath1029 to the following system of partial differential equations with initial conditions @xmath1030 are given by @xmath1031 where @xmath1032 , for @xmath932 .
let us first solve ( [ eq : dfdq ] ) . for a constant @xmath131 , consider the function @xmath1033 given implicitly as @xmath1034 , i.e. , the graph of @xmath1033 is a level curve for @xmath1026 .
the tangent vector to the graph of @xmath1033 at some point @xmath1035 such that @xmath1036 is @xmath1037 .
the derivative of the function @xmath1026 at the point @xmath1035 in the direction of this vector should be 0 , i.e. , @xmath1038 .
this equation , together with the differential equation ( [ eq : dfdq ] ) for @xmath55 , implies that @xmath1039 . solving this trivial differential equation for @xmath1033 one
deduces that @xmath1040 , where @xmath1041 is a function that depends only on the constant @xmath131 .
since @xmath131 can be an arbitrary constant , one deduces that @xmath1042 plugging the initial condition @xmath1043 in the last expression , one gets @xmath1044 thus the solution @xmath1026 is given by @xmath1045 , as needed .
direct verification shows that the function @xmath1046 satisfies the differential equation ( [ eq : dgdq ] ) , for an arbitrary @xmath1047 .
the initial condition for @xmath1024 gives @xmath1048 .
thus @xmath1049 , where @xmath1050 , as above .
since @xmath1051 , one deduces that @xmath1052 , as needed . making the substitution @xmath1053 in differential equation ( [ eq : dhdq ] ) for @xmath1054 , one obtains equation ( [ eq : dhdq ] ) for @xmath11 with @xmath1055 . by the previous calculation , one has @xmath1056 .
thus the solution for ( [ eq : dhdq ] ) is @xmath1057 . making the substitution @xmath1058 in equation ( [ eq : dhdq ] ) for @xmath178 , where @xmath1054 is the solution to ( [ eq : dhdq ] )
, one obtains equation ( [ eq : dgdq ] ) for @xmath11 with @xmath1059 . by the above calculation
, one has @xmath1060 .
thus , @xmath1061 as needed .
applying proposition [ eq : diffeq_fghr ] to differential equation ( [ eq : f_a_diff_eqn ] ) for @xmath1062 , one obtains the implicit solution : @xmath1063 this is equivalent to the quadratic equation ( [ eq : c ] ) for @xmath913 .
explicitly , one gets @xmath1064 applying proposition [ eq : diffeq_fghr ] to differential equations ( [ eq : f_t_diff_eqn ] ) and ( [ eq : f_h_diff_eqn ] ) for the generating functions @xmath1022 and @xmath1065 , one obtains the following result .
[ th : fh_ft ] the generating functions @xmath1066 and @xmath1067 are given by the following expressions @xmath1068 where @xmath1069 and @xmath1021 is given by ( [ eq : expression_for_fa ] ) .
note that the above expression for @xmath1022 is equivalent to theorem [ th : t_formula ] , using .
the goal of this section is to use theorem [ th : f_b_recurrence ] to compute the @xmath1-polynomials of the graph - associahedra of a fairly general infinite family of graphs , including all dynkin diagrams of finite and affine coxeter groups .
let @xmath141 and @xmath1041 be two connected graphs with a marked vertex in each , and let @xmath1070 be the total number of unmarked vertices in @xmath141 and @xmath1041 . for @xmath1071 , let @xmath1072 be the graph obtained by connecting the marked vertices in @xmath141 and @xmath1041 by the path
@xmath1073 so that the total number of vertices in @xmath1074 is @xmath20 .
call graphs of the form @xmath1074 _ path - like graphs _ because , for large @xmath20 , they look like paths with some `` small '' graphs attached to their ends .
the following claim shows that the @xmath1-polynomials of the graph - associahedra @xmath1075 can be expressed as linear combinations ( with polynomial coefficients ) of the @xmath1-polynomials @xmath906 of usual associahedra ; see .
[ th : almost_path_combination ] there exist unique polynomials @xmath1076 $ ] of degrees @xmath1077 such that , for any @xmath1071 , one has @xmath1078 the polynomial @xmath1079 is a palindromic polynomial , that is @xmath1080 , for @xmath1081 .
similarly , one can express the @xmath0-polynomials of @xmath1075 as a linear combination of the @xmath0-polynomials of usual associahedra , because @xmath1082 .
one can rewrite this theorem in terms the generating function @xmath913 for the narayana numbers ; see .
[ cor : path_like ] there exists a unique polynomial @xmath1083 $ ] such that , for any @xmath1084 , one has @xmath1085 the polynomial @xmath1086 has degree @xmath1087 with respect to the variable @xmath368 .
it satisfies the equation @xmath1088 .
this follows from theorem [ th : almost_path_combination ] , by setting @xmath1089 .
let us first prove the existence of the linear expansion ( and later prove its uniqueness and the palindromic property of the coefficients @xmath1079 ) .
the recurrence from theorem [ th : f_b_recurrence ] will be used to prove this clam by induction on the total number of vertices in @xmath141 and @xmath1041 . for
this argument one should drop the assumption that @xmath141 and @xmath1041 are connected .
suppose that @xmath141 or @xmath1041 is disconnected , say , @xmath141 is a disjoint union of graphs @xmath1090 and @xmath1091 where @xmath1090 contains the marked vertex .
let @xmath1092 and let @xmath143 be the number of vertices in @xmath1091
. then @xmath1093 , where @xmath1094 .
by induction , @xmath1095 can be expressed as a linear combination of @xmath1096 , which produces the needed expression for @xmath1097 .
now assume that both @xmath141 and @xmath1041 are connected graphs .
theorem [ th : f_b_recurrence](3 ) gives the expression for the @xmath1-polynomial as the sum @xmath1098 over nonempty subsets @xmath1099 of vertices of @xmath1074 , where @xmath1100 denotes the graph @xmath1074 with removed vertices in @xmath1099 .
( here one has shifted @xmath1101 by @xmath1102 to transform @xmath0-polynomials into @xmath1-polynomials . )
let us write @xmath1099 as a disjoint union @xmath1103 , where @xmath39 is a subset of unmarked vertices of @xmath141 , @xmath404 is a subset of unmarked vertices of @xmath1041 , and @xmath1104 is a subset of vertices in the path connecting the marked vertices .
the contribution of the terms with @xmath1105 to the above sum is @xmath1106 .
note that @xmath1107 , where @xmath1108 . by induction
, one can express each term @xmath1109 as a combination of @xmath1110 .
the remaining terms involve a nonempty subset @xmath1104 of vertices in the path @xmath1073 .
when one removes these @xmath1111 vertices from the path , it splits into @xmath908 smaller paths @xmath1112 with @xmath1113 ; cf .
paragraph before .
thus the remaining contribution to @xmath1097 can be written as @xmath1114 where @xmath1115 is the graph with a single vertex , @xmath1116 by induction , one can express @xmath1117 and @xmath1118 as linear combinations of the @xmath1119 .
so the remaining contribution to @xmath1097 can be written as a sum of several expressions of the form @xmath1120 where the sum is over @xmath1121 such that @xmath1122 , @xmath1123 , @xmath1124 , @xmath1125 .
this expression depends on nonnegative integers @xmath575 such that @xmath1126 and a polynomial @xmath1127 of degree @xmath1128 . if one extends the summation to all @xmath1129 , one obtains the expansion for @xmath1130 times @xmath1127 .
applying the inclusion - exclusion principle and equation , one deduces that the previous sum equals @xmath1131 which is a combination of @xmath1132 as needed .
it remains to show the uniqueness of the linear expansion and show that the @xmath1079 are palindromic polynomials .
( here one assumes that the graphs @xmath141 and @xmath1041 are connected . ) according to corollary [ cor : path_like ] , the polynomial @xmath1133 can be written as @xmath1134 , where @xmath1135 $ ] .
if @xmath1136 with @xmath1137 , then this would imply that @xmath913 is a rational function , which contradicts the formula involving a square root .
this proves the uniqueness claim .
one has @xmath1138 and @xmath1139 because @xmath1-polynomials are palindromic .
thus @xmath1140 this implies that @xmath1088 . otherwise , @xmath913 would be a rational function
the equation @xmath1088 says that the coefficients @xmath1079 of @xmath1141 are palindromic .
let us illustrate theorem [ th : almost_path_combination ] by several examples . for a series of path - like graphs @xmath1074 , let @xmath1142 denote the polynomial @xmath1086 that appears in the generating functions @xmath1143 .
for instance , the expression @xmath1144 ( see the example below ) is equivalent to the expression @xmath1145 , for @xmath1146 .
define _ daisy graphs _ as @xmath1147 ; see section [ sec : one - branching ] .
( here @xmath1148 means a sequence of @xmath433 ones . )
they include _ type @xmath1149 dynkin diagrams _
@xmath1150 . for fixed @xmath433 , the @xmath1151 form the series of path - like graphs for @xmath1152 ( the @xmath433-star with marked central vertex ) and @xmath1153 ( the graph with a single vertex )
also define _ kite graphs _ as @xmath1154 ; see section [ sec : pde ] .
they are path - like graphs for @xmath1155 and @xmath1153 .
the _ affine dynkin diagram of type @xmath1156 _ is the @xmath20th path - like graph in the case when both @xmath141 and @xmath1041 are @xmath573-paths with marked central vertices .
here are the polynomials @xmath1142 for several series of such graphs : @xmath1157 the formulas for @xmath1158 , @xmath1159 , and @xmath1151 were derived from theorems [ th : t_formula ] and [ th : fh_ft ] ( using the maple package ) .
the formula for the affine dynkin diagram @xmath1156 was obtained using the inductive procedure given in the proof of theorem [ th : almost_path_combination ] .
the induction from the proof of theorem [ th : almost_path_combination ] is quite involved .
it is very difficult to calculate by hand other examples for bigger graphs @xmath141 and @xmath1041 .
it would be interesting to find a simpler procedure for finding the polynomials @xmath1142 .
also it would be interesting to find explicit formulas for the polynomials @xmath1142 for all daisy graphs , kite graphs , and other `` natural '' series of path - like graphs .
section [ sec : stanley - pitman - example ] showed that the @xmath0- and @xmath1-vectors of flag nestohedra coming from connected building sets are bounded below by those of hypercubes and bounded above by those of permutohedra .
it is natural to ask for the bounds within the subclass of graph - associahedra corresponding to connected graphs , and to ask for bounds on their @xmath2-vectors .
permutohedra are graph - associahedra corresponding to complete graphs , and so they still provide the upper bound for the @xmath0- and @xmath1-vectors . for lower bounds on @xmath0- and @xmath1-vectors , the monotonicity discussed in remark
[ rem : graphical - monotonicity ] implies that the @xmath0- and @xmath1-vector of any graph - associahedron @xmath19 for a connected graph @xmath11 is bounded below by the graph - associahedron for any spanning tree inside @xmath11 .
thus it is of interest to look at lower ( and upper ) bounds for @xmath0- , @xmath1- and @xmath2-vectors of graph - associahedra for trees . a glance at figure [ fig : trees - on - seven - gammas ] suggests that , roughly speaking , trees which are more branched and forked ( that is , farther from being a path ) tend to have higher entries componentwise in their @xmath2-vectors , and hence also in their @xmath0- and @xmath1-vectors .
in fact , in that figure , which shows all trees on @xmath723 vertices grouped by their degree sequences , one sees several ( perhaps misleading ) features : 1 .
the degree sequences are ordered linearly under the _ majorization _ ( or _ dominance _ ) partial ordering on partitions of @xmath1160 ( @xmath1161 here ) .
the @xmath2-vectors of these trees are linearly ordered under the componentwise partial order .
trees whose degree sequence are lower in the majorization order have componentwise smaller @xmath2-vectors .
the trees are distinguished up to isomorphism by their @xmath2-vectors .
-vectors @xmath1162 for graph - associahedra of all trees on @xmath723 vertices , grouped according to their degree sequences.,height=240 ] additionally , it seems that the _ wiener index _
@xcite for graphs has some correlation with the @xmath2-vector .
the _ wiener index _
@xmath1163 of a graph @xmath11 is defined as the sum of distances @xmath1164 over unordered pairs @xmath1165 of vertices in the graph @xmath11 , where @xmath1164 is the number of edges in the shortest path from @xmath27 to @xmath118 .
the wiener index @xmath1166 of a tree is equal to the number of forbidden @xmath576 patterns ( see the remarks following definition [ def : chordal_building ] ) provided by the tree @xmath494 ( plus the constant @xmath1167 ) . thus , for two trees on @xmath20 vertices , if one has @xmath1168 , then roughly speaking one might expect that the generalized permutohedron @xmath1169 has a larger gamma vector than @xmath1170 .
this is exactly the case for trees on @xmath723 vertices , as shown in figure [ fig : trees - on - seven - gammas ] .
it shows that as the @xmath2-vectors decrease , the wiener indices ( weakly ) increase . note that in this case , the wiener index together with the degree sequence completely distinguish all equivalence classes of trees . for trees @xmath494 on @xmath20 vertices ,
the maximum and minimum values of the wiener index are , respectively , @xmath1171 for @xmath990 , and @xmath1172 for @xmath1173 .
none of the properties ( i)-(iv ) persist for all trees .
for example , when looking at trees on @xmath1174 nodes , one finds that 1 .
the degree sequences are only partially ordered by the majorization order on partitions of @xmath1175 : @xmath1176 2 .
there are trees , such as the two shown in figure [ fig : eight - node - ceg - gammas](a ) and ( b ) , whose @xmath2-vectors are incomparable componentwise , 3 .
there are trees , such as the two shown in figure [ fig : eight - node - ceg - gammas](d ) and ( c ) , where the degree sequence of one strictly majorizes that of the other , but its @xmath2-vector is strictly smaller , and 4 .
there are nonisomorphic trees , such as the two shown in figure [ fig : eight - node - ceg - gammas](d ) and ( e ) , having the same @xmath2-vector .
nodes and the @xmath2-vectors of their graph - associahedra.,height=192 ] nevertheless , we do make some monotonicity conjectures about the face numbers for graph - associahedra .
[ conj : flossing - order - conjecture ] there exists a partial order @xmath1177 on the set of ( unlabelled , isomorphism classes of ) trees with @xmath20 nodes , having these properties : 1 .
@xmath990 is the unique @xmath1177-minimum element , 2 .
@xmath1173 is the unique @xmath1177-maximum element , and 3 .
@xmath1178 implies @xmath1179 componentwise .
we suspect that such a partial order @xmath1177 can be defined so that @xmath1180 will , in particular , be comparable whenever @xmath1181 are related by one of the _ flossing moves _ considered in @xcite .
assuming conjecture [ conj : flossing - order - conjecture ] , the @xmath2-vectors ( and hence also the @xmath0- , @xmath1-vectors ) of graph - associahedra for trees on @xmath20 nodes would have the associahedron @xmath633 and the stellohedron @xmath1182 giving their lower and upper bounds .
this would also imply that the @xmath0- , @xmath1-vectors of graph - associahedra for connected graphs on @xmath20 nodes would have associahedra and permutohedra giving their lower and upper bounds . to make a similar assertion for @xmath2-vectors it would be nice to have the following analogue of stanley s monotonicity result ( theorem [ th : subdivision - monotonicity ] ) .
[ conj : gamma - monotonicity ] when @xmath344 are two flag simplicial complexes and @xmath343 is a geometric subdivision of @xmath31 , the @xmath2-vector of @xmath343 is componentwise weakly larger than that of @xmath31 . in particular , when @xmath1183 are building sets giving rise to flag nestohedra and @xmath1184 , ( such as graphical buildings @xmath1185 for an edge - subgraph @xmath1186 ) then the @xmath2-vector of @xmath427 is componentwise weakly larger than that of @xmath878 .
we close with a question suggested by the sets of permutations @xmath34 and @xmath759 for a chordal building set @xmath4 which appeared in corollary [ cor : h_des_chordal ] and theorem [ th : gamma - chordal ] . given a ( non - chordal ) building set @xmath4 ,
is there a way to define two sets of permutations @xmath1187 and @xmath1188 such that : 1 .
the @xmath1-polynomial for the nestohedron @xmath14 is given by the descent generating function for @xmath1187 , and 2 .
the @xmath2-polynomial is given by the peak generating function for @xmath1188 ?
the goal of this section is to clarify the equivalence between various definitions of the deformations of a simple polytope , either by 1 . deforming vertex positions , keeping edges in the same parallelism class , or 2 . deforming edge lengths , keeping them nonnegative , or 3 . altering level sets of facet inequalities , or but not allowing facets to move past any vertices .
there will be defined below three cones of such deformations which are all linearly isomorphic .
this discussion is essentially implicit in ( * ? ? ?
* definition 6.1 and 19 ) , but we hope the explication here clarifies this relationship .
let @xmath23 be a simple @xmath22-dimensional polytope in @xmath96 .
let @xmath358 be its set of vertices .
let @xmath1189 be its set of edge pairs .
let @xmath55 be an indexing set for its facets , so that @xmath23 is defined by facet inequalities @xmath1190 for @xmath1191 , in which each @xmath1192 is a linear functional in @xmath312 , and @xmath1193 .
\(2 ) the _ edge length deformation cone _
@xmath1197 of @xmath23 is the set of points @xmath1198 such that all @xmath1199 , and , for any 2-dimensional face of @xmath23 with edges @xmath1200 , @xmath1201 , , @xmath1202 , @xmath1203 , one has @xmath1204
\(3 ) for @xmath1205 , let @xmath1206 be the polytope obtained from @xmath23 by parallel translations of the facets .
in particular , @xmath1207 .
the _ open facet deformation cone _ described by mcmullen in ( * ? ? ? * , p. 88 ) . ]
@xmath1208 for @xmath23 is the set of @xmath1209 for which the polytopes @xmath1210 and @xmath23 have the same normal fan @xmath1211 .
( equivalently , @xmath1210 and @xmath23 have the same combinatorial structure . ) the _ ( closed ) facet deformation cone _ is the closure @xmath1212 of @xmath1208 inside @xmath1213 .
it is clear that the definitions of @xmath1194 and @xmath1197 translate into linear equations and weak linear inequalities . thus @xmath1194 and @xmath1197 are ( closed ) polyhedral cones in the spaces @xmath1214 and @xmath1215 , correspondingly .
the following lemma shows that @xmath1212 is also a polyhedral cone .
[ lem : def_cone_f ] for a simple polytope @xmath23 , the facet deformation cone @xmath1208 is a full @xmath1216-dimensional open polyhedral cone inside @xmath1213 , that is a nonempty subset in @xmath1213 given by strict linear inequalities .
thus @xmath1217 is the closed polyhedral cone in @xmath1218 given by replacing the strict inequalities with the corresponding weak inequalities .
since every polytope @xmath1210 has facet normals in directions which are a subset of those for @xmath23 , the rays ( = 1-dimensional normal cones ) in @xmath1219 are a subset of those in @xmath62 . therefore , one will have @xmath1220 if and only if @xmath1221 have the same face lattices , or equivalently , the same collection of vertex - facet incidences @xmath1222 .
this means that one can define the set @xmath1208 inside @xmath1213 by a collection of strict linear inequalities on the coordinates @xmath1223 .
it is next explained how to produce one such inequality for each pair @xmath1224 of a vertex @xmath1225 and facet @xmath1226 of @xmath23 such that @xmath1227 .
if @xmath1225 lies on the @xmath22 facets @xmath1228 in @xmath23 , then @xmath1225 is the unique solution to the linear system @xmath1229 for @xmath1230 .
its corresponding vertex @xmath1231 in @xmath1210 is then the unique solution to @xmath1232 for @xmath1230 .
note that this implies @xmath1231 has coordinates given by linear expressions in the coefficients @xmath1233 .
then the inequality corresponding to the vertex - facet pair @xmath1224 asserts that @xmath1234 . 1 .
the normal fan @xmath62 refines the normal fan @xmath1236 .
2 . the vertices of @xmath347 can be ( possibly redundantly ) labelled @xmath1237 , @xmath1238 , so that @xmath1239 is a point in the vertex deformation cone @xmath1240 , i.e. , the @xmath1237 satisfy conditions .
the polytope @xmath347 is the convex hull of points @xmath1237 , @xmath1241 , such that @xmath1239 is in the vertex deformation cone @xmath1240 .
4 . @xmath1242 for some @xmath1233 in the closed facet deformation cone @xmath1212 .
5 . @xmath347 is a minkowski summand of a dilated polytope @xmath1243 , that is there exist a polytope @xmath1244 and a real number @xmath766 such that @xmath1245 .
\(i ) implies ( ii ) .
the refinement of normal fans gives rise to the redundant labelling of vertices @xmath1247 as follows : given a vertex @xmath368 of @xmath347 , label it by @xmath1237 for every vertex @xmath277 in @xmath23 having its normal cone @xmath300 contained in the normal cone @xmath1248 .
there are then two possibilities for an edge @xmath1249 of @xmath23 : either @xmath1250 , in which case is trivially satisfied , or else @xmath1251 so that @xmath1252 lie in different normal cones @xmath1253 .
but then since @xmath62 refines @xmath1236 , these latter two cones must share a codimension one subcone lying in the same hyperplane that separates @xmath1254 and @xmath300 .
as this hyperplane has normal vector @xmath1255 , this forces @xmath1256 to be a positive multiple of this vector , as desired .
\(iii ) implies ( i ) .
let @xmath347 be the convex hull of the points @xmath1237 .
fix a vertex @xmath1257 .
let @xmath1258 be a generic linear functional that belongs to the normal cone @xmath1254 of @xmath23 at the vertex @xmath669 .
then the maximum of @xmath109 on @xmath23 is achieved at the point @xmath669 and nowhere else .
let us orient the @xmath3-skeleton of @xmath23 so that @xmath109 increases on each directed edge .
this connected graph has a unique vertex of outdegree 0 , namely the vertex @xmath669 .
thus , for any other vertex @xmath1241 , there is a directed path @xmath1259 from @xmath1260 to @xmath1261 in this directed graph . according to the conditions of the lemma ,
one has have @xmath1262 , so @xmath1263 .
thus the maximum of @xmath109 on the polytope @xmath347 is achieved at the point @xmath1264 .
this implies that @xmath1264 is a vertex of @xmath347 and the normal cone @xmath1265 of @xmath347 at this point contains the normal cone @xmath1254 of @xmath23 at @xmath669 .
since the same statement is true for any vertex of @xmath23 , one deduces that @xmath62 refines @xmath1236 .
\(i ) implies ( iv ) .
first , note that if @xmath1266 then @xmath1242 for some @xmath1233 in the open facet deformation cone @xmath1208 .
indeed , the facets of @xmath347 are orthogonal to the 1-dimensional cones in @xmath1236 , thus they should be parallel to the corresponding facets of @xmath23 .
now assume that @xmath62 refines @xmath1236 .
recall the standard fact ( * ? ? ?
7.12 ) that the normal fan @xmath1267 of a minkowski sum @xmath1268 is the common refinement of the normal fans @xmath1269 and @xmath1270 . thus , for any @xmath1271 , the normal fan of the minkowski sum @xmath1272 coincides with @xmath62 . by the previous observation
, one should have @xmath1273 for some @xmath1274 .
since all coordinates of @xmath1275 linearly depend on @xmath1276 , one obtain @xmath1277 for @xmath1278 .
\(iv ) implies ( iii ) .
given @xmath1279 , it is the limit point for some family @xmath1280 .
one may assume that @xmath1275 linearly depends on @xmath1271 and @xmath1281 .
hence @xmath1282 is the limit of the polytopes @xmath1283 , which each have @xmath1284 .
in particular , the vertices of @xmath1283 can be labelled by @xmath1285 , @xmath1241 .
these vertices linearly depend on @xmath1276 and satisfy @xmath1286 for any edge @xmath1287 . taking the limit @xmath1288 ,
one obtains that @xmath347 is the convex hull of points @xmath1289 satisfying .
\(iv ) implies ( v ) . note that @xmath1290 , for @xmath1291 .
let @xmath1282 for @xmath1292 . the point @xmath1235 ( such that @xmath1293 ) belongs to the _ open _ cone @xmath1208 .
thus , for sufficiently large @xmath143 , the point @xmath1294 also belongs to the open cone @xmath1208 .
let @xmath1295 .
then one has @xmath1296 , as needed .
we are being somewhat careful here , since theorem [ th : deformation ] can fail when one allows a broader interpretation for a simple polytope @xmath23 to deform into a polytope @xmath347 by parallel translations of its facets , e.g. if one allows facets to translate _ past _ vertices .
for example , letting @xmath347 be a regular tetrahedron in @xmath1297 , and @xmath23 the result of `` shaving off an edge '' from @xmath347 with a generically tilted plane in @xmath1297 , one finds that @xmath62 does not refine @xmath1236 .
let us now describe the relationship between the three deformation cones @xmath1240 , @xmath1298 , and @xmath1299 .
let @xmath1054 be the linear subspace in @xmath1214 given by @xmath1300 clearly , the cone @xmath1240 belongs to the subspace @xmath1054 .
let us define two linear maps @xmath1301 the map @xmath1302 sends @xmath1303 to @xmath1198 , where @xmath1304 , for any edge @xmath1305 .
the map @xmath1306 sends @xmath1307 to @xmath1239 , where , for each vertex @xmath277 of @xmath23 given as the intersection of the facets of @xmath23 indexed @xmath1308 , the point @xmath1309 is the unique solution of the linear system @xmath1310 . for @xmath1311 , @xmath1312 , where the @xmath1237 are the vertices of the polytope @xmath1210 .
one can easily check that @xmath1313 .
indeed , this is clear for @xmath1314 and thus this extends to all @xmath1315 by linearity .
note that the kernel of the map @xmath1302 is the subspace @xmath1316 embedded diagonally into @xmath1214 .
this follows from the fact that the @xmath3-skeleton of @xmath23 is connected .
the vertex deformation cone @xmath1240 can be reduced modulo the subspace @xmath1317 of parallel translations of polytopes .
similarly , the facet deformation cone can be reduced modulo the subspace @xmath1318 , where @xmath1319 for @xmath1320 . the map @xmath1306 gives a linear isomorphism between the cones @xmath1299 and @xmath1240 .
the map @xmath1302 induces a linear isomorphism between the cones @xmath1321 and @xmath1298 .
thus one has @xmath1322 in particular , @xmath1323 .
it remains to show that for any @xmath1327 there exists a unique ( modulo diagonal translations ) element @xmath1328 such that @xmath1329 for any edge @xmath1305 .
let us construct the points @xmath1309 , as follows .
pick a vertex @xmath1330 and pick any point @xmath1331 . for any other @xmath1241 ,
find a path @xmath1332 from @xmath1225 to @xmath1333 in the @xmath3-skeleton of @xmath23 and define @xmath1334 .
this point @xmath1237 does not depend on a choice of path from @xmath1225 to @xmath277 , because any other path in the @xmath3-skeleton can be obtained by switches along 2-dimensional faces of @xmath23 .
these @xmath1335 satisfy the needed conditions . | the aim of the paper is to calculate face numbers of simple generalized permutohedra , and study their @xmath0- , @xmath1- and @xmath2-vectors .
these polytopes include permutohedra , associahedra , graph - associahedra , simple graphic zonotopes , nestohedra , and other interesting polytopes .
we give several explicit formulas for @xmath1-vectors and @xmath2-vectors involving descent statistics .
this includes a combinatorial interpretation for @xmath2-vectors of a large class of generalized permutohedra which are flag simple polytopes , and confirms for them gal s conjecture on nonnegativity of @xmath2-vectors .
we calculate explicit generating functions and formulae for @xmath1-polynomials of various families of graph - associahedra , including those corresponding to all dynkin diagrams of finite and affine types .
we also discuss relations with narayana numbers and with simon newcomb s problem .
we give ( and conjecture ) upper and lower bounds for @xmath0- , @xmath1- , and @xmath2-vectors within several classes of generalized permutohedra .
an appendix discusses the equivalence of various notions of deformations of simple polytopes . |
the first evidence of the observation of a pentaquark @xmath0 particle came in early 2003 from the leps collaboration @xcite in which a sharp resonance was reported at a mass of @xmath4 ( gev)/c@xmath5 . the diana collaboration @xcite reported evidence for a resonance enhancement at @xmath6 mev / c@xmath5 and @xmath7 mev / c@xmath5 in the charge - exchange reaction @xmath8 . the clas collaboration @xcite reported a statistical significance of @xmath9 at the @xmath10 invariant mass peak of @xmath11 mev / c@xmath5 with a measured width of 21 mev for the reaction @xmath12 .
the saphir @xcite collaboration reported a peak at @xmath13 mev ( 4.8 @xmath14 confidence level ) and an upper limit of 25 mev for the width with a @xmath15 confidence level in the @xmath16 invariant mass distribution from the @xmath17 process .
the saphir collaboration also reported the absence of a signal in the @xmath18 invariant mass distribution in the @xmath19 reaction .
more recently , the hermes collaboration @xcite reported evidence of the @xmath0 particle from a deuterium target with the decay mode @xmath20 at @xmath21 ( mev ) with a width of 7.5 mev dominated by detector resolution
. the clas collaboration @xcite also reported evidence on @xmath0 from a hydrogen target from the @xmath22 process .
they found a peak in the invariant mass of @xmath23 at 1555 mev with a width of about 26 mev .
the cern na49 collaboration @xcite reported an exotic @xmath24 baryon resonance in proton - proton collisions with a mass of @xmath25 mev / c@xmath5 and width below the detector resolution of 18 mev / c@xmath5 in the @xmath26 , and @xmath27 invariant mass spectra .
these two states are believed to be candidates for the @xmath28 ( @xmath29 ) , and the @xmath30 ( @xmath31 ) pentaquark state , respectively . however , there is probably more evidence against the existence of the @xmath28 and @xmath30 pentaquark states @xcite .
mostly recently , the h1 collaboration reported @xcite evidence for a narrow anti - charmed baryon state , interpreting such a state as the pentaquark @xmath32 state . in summary , the experimental situation concerning the @xmath0 and @xmath3 pentaquark states is the following .
there is evidence from different experiments supporting the existence of the pentaquark @xmath0 state , but there is no evidence for the @xmath3 particle .
the mass of the @xmath0 particle is from 1527 mev to 1555 mev , and the width is found to be from 9 mev to 26 mev limited by detector resolutions . also , the existing data seem to suggest that the @xmath0 particle is an isoscalar particle . however , the existence of @xmath33 pentaquark is far from being confirmed experimentally @xcite .
one also needs to address possible experimental issues such as kinematic reflection @xcite due to the decay of higher mass mesons , such as the @xmath34 , the @xmath35 , and the @xmath36 .
the original chiral soliton model @xcite predicted a mass of 1530 mev and a total width of less than 15 mev for the @xmath0 particle .
it also predicted that the @xmath0 particle is a spin @xmath37 isoscalar particle and the predicted mass for the @xmath38 state is 2070 mev . in the correlated diquark picture by jaffe and wilczek @xcite ,
the diquarks qq are correlated in an antisymmetric color , flavor and spin state . in the case of the @xmath0 pentaquark state
, it is a bound state of two highly correlated @xmath39 pairs and an antiquark .
the narrowness of the width of the @xmath0 particle can be explained in this model by the weak coupling between the @xmath10 continuum and the bound state of the two diquark pairs and the antiquark , @xmath40^{2}\bar{s}$ ] .
one of the interesting predictions from this model , which differs from the chiral soliton model prediction , is the mass of the cascade isospin @xmath41 multiple @xmath42 ^ 2\bar{d})$ ] .
it is predicted to be around 1750 mev with a width about @xmath43 greater than that of the @xmath0 particle .
the @xmath0 particle was also hypothesized as an isotensor resonance to explain the observed narrow width of the @xmath0 particle @xcite via isospin - violating strong decays .
the hypothesis that @xmath0 is an isotensor particle implies the existence of @xmath44 , @xmath45 , @xmath46 , and @xmath47 states .
gerasyuta and kochkin @xcite calculated the mass spectra of the isotensor theta - pentaquarks with @xmath48 in a relativistic quark model .
the predicted mass for @xmath45 is 1575 mev ( 1761 mev ) for @xmath49 ( @xmath50 ) .
here we following the convention in the literature that @xmath51 represents the i=0 state in anti - decuplet , @xmath52 represents the i=1 state in 27-plet and @xmath53 the i=2 state in 35-plet . in the remaining of the paper
, we will not explicitly differentiate the @xmath45 and the @xmath54 states because the proposed experiment will not be able to determine which multiplet the observed @xmath3 particle belongs to .
walliser and kopeliovich @xcite investigated the implications of the @xmath0 exotic pentaquark for the baryon spectrum in topological soliton models .
they estimate the positions of other pentaquark and septuquark states with exotic and with non - exotic quantum numbers , particularly within the 27-plet baryon and the 35-plet multiplets in su(3 ) soliton model . in the 27-plet
, the @xmath55 multiple ( @xmath56 , @xmath57 , @xmath54 ) are estimated to have a mass of 1.65 to 1.69 gev .
most recently , wu and ma @xcite studied the mass and width of the pentaquark @xmath52 states in the 27 baryon multiplet from chiral soliton model .
their calculations show that the mass of @xmath52 is about 1.61 gev and the width for the process @xmath58 is less than 44 mev .
et al . _
@xcite constructed a complete classification of @xmath59 states in terms of the flavor - spin su(6 ) representation and found that only the anti - decuplet , 27-plets and the 35-plets contain exotic states which can not be constructed by three quarks only . in this model ,
the ground state pentaquark is identified as the observed @xmath60 state , and is predicted to be an isosinglet anti - decuplet state .
the predicted masses for the excited exotic baryons are 1660 mev and 1775 mev . while it is important to determine the quantum number of the @xmath0 particle , particularly the spin and the parity and to search for other members in the exotic baryon family , it is more essential to confirm the existence of the @xmath0 particle with significantly improved statistics , and to determine its mass and the width more precisely .
karliner and lipkin @xcite derived an upper bound on the mass difference between the @xmath61 and @xmath0 based on simple assumptions about @xmath62 symmetry breaking and quantum mechanical variational method .
the resulting rather robust bound is more than 20 mev below the experimentally reported @xmath61 - @xmath0 mass difference .
it is also very important to verify definitely the non - existence or the existence of the @xmath3 particle and to search for the @xmath54 particle .
mass spectra for the proposed experiment . in this
simulation , physical events were generated from the @xmath63 process with a @xmath33 mass of 1540 mev ; the background is from @xmath64 pairs and an accidental coincidental rate .
, width=336 ] we proposed @xcite a new , high resolution search for pentaquark @xmath0 , @xmath3 and @xmath54 in hall c at jefferson lab using high resolution magnetic spectrometers in combination with a neutron counter by studying two different physical processes : the @xmath65 , where @xmath0 decays via @xmath66 , and the @xmath67 , where @xmath68 . due to the lack of a free neutron target
, a liquid deuterium target will be used for the @xmath69 process .
the @xmath70 particle will be detected in the short orbit spectrometer ( sos ) in coincidence with the @xmath71 and the neutron which will be detected in the high resolution kaon spectrometer ( hks ) , and the neutron counter , respectively . by detecting all three particles , the untagged bremsstrahlung photon energy ,
the mass of the @xmath0 particle , and the initial neutron momentum inside the deuteron can be reconstructed completely .
for the two - body process @xmath72 process one can not reconstruct the mass of the undetected @xmath3 ( @xmath54 ) particle from detecting @xmath70 particle only unless the energy of the incident real photon is known .
assuming the @xmath73 particle decays into @xmath71 and @xmath74 , the incident real photon energy can be reconstructed by detecting @xmath71 in coincidence with the @xmath70 particle .
such coincidence detection also suppresses backgrounds from other physical processes which have final states different from the @xmath75 final state .
the proposed experiment allows for the search of the @xmath3 ( @xmath54 ) particle in the mass range of 1500 mev to 1800 mev .
the primary sources of the physics backgrounds for the proposed @xmath67 process and the @xmath78 process are : ( i ) @xmath79 meson production , ( ii ) @xmath80 where , @xmath81 , and ( iii ) three - body final state @xmath82 production .
monte carlo simulations have been carried out to study these backgrounds , taking into account the momentum and angular resolutions of the two hadron spectrometers and their full momentum and angular acceptances . the spectrometer momentum and angle settings based on the two - body kinematics of the @xmath67 and the @xmath69 processes together with the reconstructed photon energy cut ( @xmath83 mev ) , where @xmath84 is the incident electron beam energy , effectively suppress the background contributions from the aforementioned physical processes .
cuts can be applied in the invariant mass of the @xmath85 ( @xmath16 ) system in order to suppress these backgrounds further .
simulations show that both the @xmath79 production and the @xmath86 channels are suppressed even without a 100 mev of @xmath87 cut and kinematic reflections are not issues for the proposed experiment . fig .
[ fig : thetaplus ] shows the projected result on the @xmath0 particle search .
the @xmath76 reaction will be used for the mass calibration .
the @xmath88 particle will be reconstructed by its decay into neutron and @xmath89 particle .
the neutron will be detected in the neutron counter in coincidence with the @xmath89 , and the mass of the @xmath88 particle will be reconstructed in a similar way as that of the @xmath0 particle . to suppress the background the @xmath71 particle will be detected in the sos spectrometer in coincidence with the neutron and the @xmath89 particle .
[ fig : calibration ] shows the simulated @xmath88 particle mass determination , which is better than 0.3 mev .
the upper panel shows the simulation result for the @xmath76 , @xmath77 reaction only , while the lower panel result includes accidental and other possible physical backgrounds .
this new experiment will determine the absolute mass of the @xmath0 particle to better than 0.5 mev and determine the @xmath0 particle width to better than 1.3 mev ( @xmath14 ) .
this work is supported by the u.s .
department of energy under contract number de - fg02 - 03er41231 .
the author thanks the yukawa institute for theoretical physics at kyoto university . discussions during the yitp workshop yitp - w-03 - 21 on `` multi - quark hadron : four , five and more ? '' were useful for the completion of this work . | recent evidence for the existence of pentaquark @xmath0 particle from several experiments at several different laboratories around the world has caused great excitement and raised many unanswered questions .
we discuss a new , high resolution experiment searching for pentaquark states in hall c at jefferson lab using an untagged bremsstrahlung photon beam employing both the hydrogen and the deuterium targets by studying the following processes : @xmath1 , and @xmath2 .
this new experiment will significantly improve our current knowledge of the mass , the width of the @xmath0 particle if it is confirmed and provide unambiguous evidence for the existence or non - existence of the @xmath3 particle from the @xmath1 process . |
we present results for the hadronic production of a vector boson pair , including all spin correlations in the decay of the final state bosons , @xmath7 where @xmath8 or @xmath9 .
the calculations are performed in next - to - leading order in @xmath10 .
we implement the helicity amplitudes of @xcite and thus extend previous treatments of vector boson pair production ( @xcite@xcite and @xcite@xcite ) to include spin correlations in all the partonic matrix elements . by including the decay products in this way it is possible to impose experimental cuts , necessary to compare theory with experiment .
some cuts are experimentally necessary and more stringent cuts are often useful in order to reduce backgrounds in the search for new physics . although phenomenological predictions including the _ complete _ one loop predictions are presented here for the first time , this may be a matter of theoretical correctness rather than practical importance .
the early predictions were performed at @xmath11 and to a limited extent at @xmath12 , so an update of the phenomenological results is in any case appropriate .
we have therefore provided predictions for @xmath3 collisions at @xmath13 ( tevatron run ii ) and for @xmath5 collisions at @xmath14 ( lhc ) .
since the early predictions were made , there have been changes in @xmath15 and the determination of the gluon distribution , especially at small @xmath16 .
we include this information by using modern parton distributions .
our results are obtained using a monte carlo program @xmath17 which allows the calculation of any infra - red finite quantity through order @xmath15 .
the monte carlo program is constructed using the method of ref .
@xcite based on the subtraction technique of ref .
we hope to provide further details in a subsequent publication .
as already noted , there is a substantial existing literature on vector boson pair production in hadronic collisions . as a cross - check of our results
, we will compare the values of the total cross section obtained using our monte carlo ( @xmath17 ) with those of frixione , nason , mele and ridolfi @xcite .
since the earlier predictions were made at centre - of - mass energies of @xmath18 tev and @xmath19 tev , we will also provide up - to - date values for @xmath20 collisions at @xmath21 tev and @xmath5 collisions at @xmath22 tev .
these run ii tevatron and lhc predictions contain the latest parton distribution sets @xcite as well as more recent electroweak input .
we will first present a comparison with older results on the total cross section .
we use the structure function set hmrsb @xcite , which is common amongst the calculations of @xcite .
this set corresponds to a four - flavour value of @xmath23 mev . for the purposes of comparison we use the standard two loop expression @xcite @xmath24,\ ] ] and match to five flavors at @xmath25 .
this yields a strong coupling at the @xmath26-mass of , @xmath27 the renormalization and factorization scales are chosen to be equal to the average mass of the produced boson pair ( namely , @xmath28 , @xmath29 and @xmath30 ) . finally , since the early calculations present results for the production of two on - shell vector bosons , we need to ensure that in our monte carlo ( in which we produce 4 final - state leptons ) we use the narrow - width approximation where only doubly - resonant diagrams are included .
the on - shell boson cross sections can be obtained by dividing out by the relevant branching ratios . a full discussion of the other diagrams included in our approach is given in section [ beyondzw ] .
the comparison of the results for a @xmath5 collider at centre of mass energies of @xmath18 tev and @xmath19 tev is shown in tables [ comp16 ] and [ comp40 ] respectively .
for the cases of @xmath31 and @xmath32 pair production , the values for comparison are taken directly from @xcite and @xcite . for @xmath33 ,
the results are slightly different from those published in @xcite since the contribution from processes of the type @xmath34 with the top quark taken massless ( which it is no longer appropriate to include ) have been removed .
the comparison between our results and the results of refs .
@xcite@xcite is satisfactory .
apart from their role as a check of our programs the results in tables [ comp16 ] and [ comp40 ] should be considered obsolete .
.total cross section for the various di - boson processes at @xmath18 tev . for the cases of @xmath35 and @xmath36 production the electroweak parameters are , @xmath37 , @xmath38 , @xmath39 and @xmath40@xcite . for @xmath26 pairs
the input is instead , @xmath41 , @xmath42 and @xmath39@xcite . [ cols="^,^,^,^,^,^,^,^,^ " , ]
we have presented a monte carlo program for vector - boson pair production at hadron colliders , including for the first time the complete @xmath43 corrections with leptonic decay correlations .
we have employed this program to calculate the total di - boson cross - sections , first as a cross - check with existing results in the literature and secondly in order to provide an update of predictions for run ii at the tevatron and for the lhc , including the latest structure functions and strong coupling .
the cross sections are larger than previous estimates in the literature .
the advantage of this monte carlo is only realized when cuts are applied to the final - state leptons .
this is primarily of importance when estimating standard model backgrounds to new physics , which is especially crucial at the tevatron . for this reason we have provided two examples of such uses in run ii , tri - lepton production as a susy signal and a di - lepton analysis for an intermediate - mass higgs search .
we first introduce a separation of the total @xmath44-particle tree - level amplitude into its doubly- and singly - resonant components , @xmath45 where @xmath46 is an overall coupling factor depending on the di - boson process under consideration .
the contributions @xmath47 were first calculated in @xcite although we will closely follow the more recent notation and approach of @xcite . in particular , the amplitudes are presented in terms of particle momenta @xmath48 that are all outgoing , so that momentum conservation implies @xmath49 . for this process
we label the particles as , @xmath50 where the leptons are not necessarily of the same flavour , and we write the process in this manner to remind the reader that @xmath51 represents an outgoing quark ( so that when we cross to obtain the desired result , it becomes an incoming anti - quark ) .
then from @xcite we see that the non - vanishing doubly - resonant helicity amplitudes for up - quark annihilation are ( with labels suppressed where possible ) , @xmath52 with the sub - amplitudes @xmath53 given by equations ( 2.8 ) and ( 2.9 ) of @xcite .
the @xmath54 are propagator factors given by , @xmath55 with @xmath56 and @xmath57 .
the couplings that appear in these amplitudes are , @xmath58 where @xmath59 is the electric charge in units of the positron charge . for the additional singly - resonant diagrams we find , @xmath60 where we have introduced a further set of scaled couplings @xcite , @xmath61 note that here the sub - amplitudes needed for the singly - resonant diagrams are exactly those introduced in @xcite to describe the doubly - resonant diagrams .
this is because the diagrams are topologically equivalent and ( modulo couplings ) we can obtain the singly resonant amplitudes by simply re - labelling the external legs .
the down - quark amplitudes may be obtained simply by symmetry , @xmath62 as described in @xcite , the doubly - resonant amplitudes for the process with an additional gluon radiated from the quark line are exactly analogous to ( [ wwdr ] ) and require the introduction of functions @xmath63 . however , unlike the 6-particle amplitudes , these functions are not sufficient to describe the singly - resonant diagrams . in this case
, initial state gluon radiation in the singly - resonant diagrams would correspond to final - state radiation in the equivalent doubly - resonant diagrams and thus we need to introduce a new set of sub - amplitudes . for a positive helicity gluon of momentum @xmath64
we find , @xmath65 where the new functions are defined by , @xmath66 and , @xmath67 \over{\spa1.7 \spa2.7 s_{34 } t_{127 } t_{345}}}. \end{aligned}\ ] ] our definition of the spinor products follows ref . @xcite .
the remaining amplitudes are obtained as above , with the additional ( negative helicity ) gluon corresponding , as in the doubly - resonant case , to the operation @xmath68 defined in @xcite .
as will be the case for all the amplitudes in this appendix , the loop contributions for the singly - resonant diagrams may be simply obtained by following the prescription ( 3.20 ) of @xcite . here
we label the two processes slightly differently , @xmath69 in order to simplify the form of the amplitudes and the overall coupling ( c.f .
[ overall ] ) is , @xmath70 the doubly - resonant amplitude for a left - handed decay of the @xmath26 is , @xmath71 where the masses in the propagators are now @xmath72 and @xmath73 . for @xmath74 production
we have @xmath75 , @xmath76 and the third line has a positive contribution , whilst @xmath77 corresponds to @xmath78 , @xmath79 and a negative sign .
this reduces to the form given in @xcite if we set @xmath80 to neglect the virtual photon diagrams .
the right - handed amplitude is obtained by a symmetry transformation , @xmath81 the singly resonant diagrams are somewhat more complicated . with a @xmath26 propagator
we can couple both electrons and neutrinos , while @xmath9 may only couple directly to the electrons .
in addition , if the final state electrons are both left - handed , there is a contribution from a diagram containing two @xmath82 propagators . in total
we obtain , @xmath83 where the newly introduced couplings @xmath84 and @xmath85 depend upon the process under consideration and are given by , @xmath86 the right - handed contribution is similar , but does not contain the corresponding final term , @xmath87 we now turn to the graphs including gluon radiation . for the doubly resonant contribution with a positive helicity gluon we find a similar structure , @xmath88 whilst the singly resonant pieces again require the new amplitudes , @xmath89 @xmath90 the remaining amplitudes , with the helicity of the gluon reversed , can be obtained by the transformation , @xmath91 , b_7^{\tree , a } \leftrightarrow b_7^{\tree , b } \right).\ ] ] f.e .
paige and s.d .
protopopescu , in proc .
ucla workshop , eds .
bengtsson et al . ,
world scientific , singapore ( 1986 ) .
bengtsson and t. sjstrand , comp .
* 46 * ( 1987 ) 43 t. sjstrand , comp .
* 82 * ( 1994 ) 74 .
e. w. n .. glover , j. ohnemus and s. s. d. willenbrock , phys . rev .
* d37 * ( 1988 ) 3193 ; + v. barger , g. bhattacharya , t. han and b. a. kniehl , phys .
d43 * ( 1991 ) 779 ; + j. gunion and t. han , phys . rev .
* d51 * ( 1995 ) 1051 ; + m. dittmar and h. dreiner , phys . rev .
* d55 * ( 1997 ) 167 ; + t. han and r .- j .
zhang , phys .
* 82 * ( 1999 ) 25 .
t. han , a. s. turcot and r .- j .
zhang , phys .
* d59 * ( 1999 ) 093001 .
s. dawson , nucl .
b359 * ( 1991 ) 283 ; a. djouadi , m. spira , p.m. zerwas , phys . lett .
* b264 * ( 1991 ) 440 .
see , for example , r.k .
ellis , w.j .
stirling and b.r .
webber , _ qcd and collider physics _ , cambridge university press , ( 1996 ) . | we present numerical results ( including full one - loop qcd corrections ) for the processes @xmath0 and @xmath1 and @xmath2 followed by the decay of the massive vector bosons into leptons .
in addition to their intrinsic importance as tests of the standard model , these processes are also backgrounds to conjectured non - standard model processes . because of the small cross sections at the tevatron , full experimental control of these backgrounds will be hard to achieve .
this accentuates the need for up - to - date theoretical information .
a comparison is made with earlier work and cross section results are presented for @xmath3 collisions at @xmath4 tev and @xmath5 collisions at @xmath6 tev .
practical examples of the use of our calculations are presented .
= -2 cm = -1.5 cm = 23.0truecm = 16.5truecm 2ex # 1.#2#1 .15 mm # 2 # 1.#2 # 1.#2.#3 - 1mu#1 | # 2 | # 3 - 1mu [ 1]0= 0=0 1= 1=1 0>1 # 1 / fermilab pub99/146t + hep - ph/9905386 + may 1999 0.5 cm j.m .
campbell and r.k .
ellis + 0.2 cm theory department , fermilab , po box 500 , batavia , il 60510 , usa .
_ submitted to physical review d _ |
flavour groups @xmath0 and their peculiar breaking to residual symmetries @xmath16 and @xmath17 in the charged lepton and in the neutrino sector , respectively , have been applied in the past in order to predict lepton mixing angles and the dirac phase in a model - independent way , for reviews see @xcite .
one of the most prominent examples is tri - bimaximal ( tb ) mixing @xcite which can be derived with the help of the flavour groups @xmath18 @xcite and @xmath19 @xcite .
recent measurements of the reactor mixing angle @xmath20 , @xmath21 , @xcite , however , clearly show that many patterns which have been discussed in the literature are strongly disfavoured , because they predict vanishing or very small @xmath20 .
for this reason , proposals have been made in which either the groups @xmath0 are chosen to be large , e.g. @xmath22 and @xmath23 @xcite , or the symmetry is reduced , e.g. @xmath24 is considered instead of @xmath25 in the neutrino sector @xcite .
we explore a different approach in this paper and consider a scenario with three majorana neutrinos in which a discrete , finite flavour group @xmath0 and a @xmath1 symmetry are combined and are broken in such a way that the residual symmetry in the neutrino sector is @xmath6 with @xmath26 being a subgroup of @xmath0 .
the residual symmetry @xmath5 in the charged lepton sector is - as in preceding approaches @xcite - chosen as cyclic symmetry ( or product thereof ) which allows to distinguish between the three generations of charged leptons .
we show for a general @xmath0 that such a breaking pattern allows to predict lepton mixing angles and dirac as well as majorana phases in terms of a single real parameter @xmath2 .
a non - vanishing reactor mixing angle can be easily accommodated and at the same time relations between mixing angles and cp phases are obtained .
in contrast to the scenario without a @xmath1 symmetry , we also predict majorana phases with our present approach .
we discuss in detail the conditions which have to be fulfilled in order to consistently formulate a setup with a flavour symmetry @xmath0 and a @xmath1 symmetry and in order to define the group @xmath17 as a direct product of @xmath26 and @xmath1 .
we also show that in general several independent @xmath1 transformations might be compatible with a flavour group @xmath0 which lead , in general , to physically different results .
we exemplify our formalism with a comprehensive study of @xmath7 and show several interesting mixing patterns whose predictions for the mixing angles are close to the best fit results @xcite for certain values of the parameter @xmath2 . the idea to combine a flavour symmetry @xmath0 with a @xmath1 symmetry is not new and has been already discussed in some cases in the literature @xcite .
an interesting example is the so - called @xmath27 reflection symmetry which is a combination of the canonical @xmath1 transformation and the @xmath27 exchange symmetry .
this generalized @xmath1 transformation permutes a muon neutrino ( antineutrino ) and a tau antineutrino ( neutrino ) .
if imposed on the neutrino mass matrix in the basis in which the charged lepton mass matrix is diagonal , it forces the elements of the second and third rows of the pontecorvo - maki - nakagawa - sakata ( pmns ) mixing matrix to have the same absolute values . as a consequence
, one finds @xmath28 and @xmath29 .
thus , maximal atmospheric mixing is predicted and , in view of the latest global fits @xcite , also the dirac phase @xmath30 has to be maximal .
recently , also the combination of the flavour group @xmath19 with a certain @xmath1 transformation has been discussed in two models @xcite .
other contexts in which flavour symmetries and cp violation appear together are : the idea of so - called geometrical cp violation in which the potential of certain scalars is constrained by a flavour group in such a way that their vacuum expectation values spontaneously break @xmath1 symmetry with phases independent of the parameters of the lagrangian @xcite ; the accidental presence of @xmath1 symmetries has been noticed in potentials invariant under ( single- and double - valued ) dihedral groups @xcite , while attempts to relate the prediction of cp violation to particular properties of certain representations of the group @xmath31 can be found in @xcite .
the paper is organized as follows : in section [ sec2 ] we first recall several well - known facts about @xmath1 transformations and discuss their combination with a flavour group @xmath0 and which conditions have to be fulfilled in order to consistently define such a setup . assuming three generations of majorana neutrinos , we then present the results for mixing angles and cp phases which are obtained for a general @xmath0 combined with a @xmath1 transformation and broken to @xmath16 and to @xmath6 .
we furthermore study the possibility of and conditions for the presence of an accidental @xmath1 symmetry . in section [ sec3 ]
we present the case @xmath7 and the different possible @xmath1 transformations compatible with all requirements .
we show that there are only a few independent - and phenomenologically interesting - cases for which we discuss the results for cp phases and mixing angles in detail . apart from that we analyze particular values of the parameter @xmath2 for which the symmetry in the neutrino sector is enhanced , @xmath32 , with @xmath33 being a subgroup of a finite flavour group containing @xmath19 .
we comment on results which can also be obtained for @xmath34 .
we conclude in section [ concl ] . in appendix [ appa ]
we fix our notations and conventions for the mixing parameters and show that the invariance of the charged lepton and neutrino mass matrices under a common generalized @xmath1 transformation @xmath35 leads to trivial dirac and majorana phases . in appendix [ appb ]
we show that the mathematical structure of the group arising from the combination of a flavour symmetry @xmath0 and a @xmath1 symmetry is in general a semi - direct product of the form @xmath36 with @xmath37 being the group associated with @xmath1 .
in this section we recall some basic properties of @xmath1 transformations and of @xmath1 invariance in the lepton sector .
we then discuss how to combine @xmath1 and internal symmetries and we explain how to use @xmath1 and flavour symmetries to constrain lepton mixing parameters . finally , we comment on the possibility of accidental @xmath1 symmetries .
we can define a @xmath1 transformation on a set of fields , collectively denoted by @xmath38 , as (x)= x ^*(x_cp ) [ x ] in matrix notation and with @xmath39 .
we choose the transformation @xmath35 to be a constant unitary symmetric matrix , [ xcond1 ] x x^= x x^ * = . in this way , @xmath40 , since eq.([x ] ) implies that @xmath41 . in theories with internal symmetries , such as a flavour symmetry @xmath0 ,
it is possible to generalize this requirement on @xmath35 to @xmath42 being a transformation belonging to the group of internal symmetries .
we refer to the transformation in eq .
( [ x ] ) as a generalized @xmath1 transformation @xcite .
for spinors we use a two - component notation and omit the obvious action of @xmath1 on the spinor indices ( which contains suitable phases such that @xmath43 ) . in the following
we also omit the dependence of the fields on the space - time point @xmath44 .
let us consider a generalized @xmath1 transformation acting on the three generations of lepton doublets @xmath45 l=x l^ * which fulfills eq.([xcond1 ] ) . in the interaction basis gauge interactions are @xmath1 conserving , while the requirement of @xmath1 invariance of the yukawa interactions , including terms giving rise to neutrino masses , constrains both the charged lepton mass matrix @xmath46 . ] and the neutrino mass matrix @xmath47 : on the spinor indices .
+ if @xmath35 is not symmetric , the conditions in eq.([invcond ] ) have to be changed into : @xmath48 and @xmath49 . ]
[ invcond ] x^ * m_l^m_l x = ( m_l^m_l ) ^ * , x m_x = m_^ * . as we show explicitly in appendix [ appa ] and
is discussed in the literature , see e.g. @xcite , the existence of a @xmath1 transformation fulfilling eq.([invcond ] ) implies that the cp invariants @xmath50 and @xmath51 ( see appendix [ appa ] for notation and conventions of the mixing parameters ) vanish and thus the dirac phase @xmath30 as well as the majorana phases @xmath52 and @xmath53 are trivial , i.e. [ phasestriv ] = 0 , = 0 , = 0 .
if neutrinos are dirac particles , the second equality in eq.([invcond ] ) has to be changed into [ diracnu ] x^ * m_^m_x = ( m_^m_)^ * .
again , as we show in appendix [ appa ] , the dirac phase is trivial .
this result can also be applied to the quark sector , i.e. the up quark and down quark mass matrices instead of dirac neutrino and charged lepton mass matrices .
we now consider a theory that is invariant under both a flavour symmetry @xmath0 and @xmath1 .
we assume @xmath0 to be a discrete and finite group .
however , most of the following statements can also be applied to a continuous symmetry @xmath0 , if it is global .
an extensive discussion of how to consistently define a @xmath1 symmetry in gauge theories can be found in @xcite .
we assume that scalar and spinor fields transform according to some representation of the flavour group @xmath0 and we denote a set of fields transforming in a generic irreducible representation @xmath54 of @xmath0 by @xmath38 : =a , [ rho ] with @xmath55 being a unitary matrix depending on the representation @xmath54 and on the chosen group element @xmath56 of @xmath0 . under @xmath1
the multiplet @xmath38 transforms as = x ^ * .
[ x2 ] the matrix @xmath35 is unitary and symmetric and acts on the representation @xmath54 . on a single irreducible representation of @xmath0 . in this case @xmath38 of eq .
( [ x2 ] ) denotes the smallest combination of irreducible representations of @xmath0 on which the action of @xmath1 is well defined . ]
if we perform a @xmath1 transformation , followed by a transformation of @xmath0 and another @xmath1 transformation we end up with @xmath57 . by consistency this should be a transformation of @xmath0 for the representation @xmath54 , given by the matrix @xmath58 and representing a group element @xmath59 ( note in general @xmath60 ) ( x^-1a x)^*=a .
[ xcond2 ] for a given group @xmath0 , this equation constrains the form of @xmath35 and we would like to determine its general solution that also satisfies eq .
( [ xcond1 ] ) . as one sees from eqs.([xcond1],[xcond2 ] ) , for a solution @xmath35 also @xmath61 with @xmath62 being an arbitrary phase is a solution .
notice that at least one solution for @xmath35 exists and it is the canonical @xmath1 transformation @xmath63 , if the representation @xmath54 and also its representation matrices are real , since eqs.([xcond1],[xcond2 ] ) are then trivially fulfilled ( @xmath64 ) . since the application of a similarity transformation and complex conjugation preserves the usual ( matrix ) multiplication rules , it is sufficient to check whether a transformation @xmath35 satisfies the constraint in eq .
( [ xcond2 ] ) for a set of generators @xmath65 of the group @xmath0 which give rise , via products , to all elements of @xmath0 . if @xmath66 , @xmath67 , also @xmath68 , as can be checked by explicit computation .
thus , @xmath55 and @xmath58 have the same order .
if @xmath0 is abelian , all irreducible representations are one - dimensional and the constraint in eq .
( [ xcond2 ] ) is satisfied by @xmath69 , with an arbitrary phase @xmath62 .
indeed , @xmath70 is in this case always @xmath71 .
if @xmath0 is non - abelian , there are special bases in which we can recognize in a simple way whether there exists ( at least ) one transformation @xmath35 which fulfills the constraint in eq .
( [ xcond2 ] ) ; for instance , in a basis in which all non - diagonal generators are real the canonical @xmath1 transformation @xmath63 fulfills eq.([xcond2 ] ) for all representation matrices @xmath55 .
notice that a generalization of our choice @xmath35 being symmetric , as mentioned below eq.([xcond1 ] ) , can lead to further solutions for @xmath35 .
however , in the present paper we always assume @xmath35 to be symmetric . if we perform a change of basis with a unitary matrix @xmath72 in the field space = ^ , [ change1 ] the unitary matrices @xmath35 and @xmath55 transform as = ^x^ * , = ^a , [ change2 ] as can be seen using eqs .
( [ rho ] , [ x2 ] ) .
the constraints in eqs .
( [ xcond1 ] , [ xcond2 ] ) are covariant under such a transformation @xmath72 , i.e. also @xmath73 and the set of matrices @xmath74 fulfill eqs .
( [ xcond1],[xcond2 ] ) : @xmath75 and @xmath76 .
a change of basis can be useful in order to reach a basis in which the action of some elements of @xmath0 and/or of @xmath1 is particularly simple .
for example , we can use the result that any unitary symmetric matrix @xmath35 can be written as the product @xmath77 of a unitary matrix @xmath72 and its transpose , in order to go to a basis in which the action of @xmath1 is canonical , @xmath78 , see eq .
( [ change2 ] ) . as a consequence , the constraint in eq .
( [ xcond2 ] ) reads @xmath79 .
the flavour group @xmath0 together with the @xmath1 transformation @xmath35 defines a group which we call @xmath80 in the following .
this group turns out to be a semi - direct product of @xmath0 and @xmath1 , as we show in appendix [ appb ] .
one reason for imposing a flavour symmetry @xmath0 is to constrain the form of the lepton mixing matrix in order to explain the observed pattern of mixing angles
. a particular approach @xcite is to assume the invariance under @xmath0 to be broken in such a way that the combination @xmath81 and the neutrino mass matrix @xmath47 possess residual discrete symmetries @xmath16 and @xmath17 , respectively .
if we consider three generations of majorana neutrinos , we are naturally led to the choice @xmath82 , the largest symmetry of @xmath47 leaving neutrino masses unconstrained @xcite . for the group @xmath16 we require the following properties
: it should be abelian in order to avoid degeneracies among the charged lepton masses and it should allow to assign different charges to the three different generations , for details see @xcite .
thus , @xmath16 is in general a ( direct ) product of cyclic symmetries , @xmath83 .
this choice of subgroups of @xmath0 and their relative embedding into the latter predict the form of the lepton mixing matrix , up to permutations of rows and columns and up to arbitrary majorana phases @xcite .
this mechanism can be implemented in concrete models in which the desired symmetry breaking pattern of the group @xmath0 can be achieved via spontaneous or explicit breaking and corrections to such a pattern are calculable ( and are usually small ) , see @xcite for reviews . in the present paper
we instead consider the case in which the residual symmetry @xmath17 is @xmath84 .
as we will see , this allows us to determine all physical phases and mixing angles in terms of a single real parameter @xmath2 . a small non - vanishing mixing angle @xmath20
can then be accommodated by a suitable choice of the parameter @xmath2 and furthermore testable relations among the mixing parameters are predicted .
once the flavour group @xmath0 has been chosen , several independent definitions of @xmath1 are in general possible , leading to physically distinct results . in this subsection
we explain the setup and show the general form of the lepton mixing matrix , while we illustrate several interesting features with an explicit example based on the group @xmath7 in section [ sec3 ] .
we recall that lepton doublets transform in a three - dimensional irreducible representation @xmath54 of @xmath0 and that neutrinos are of majorana type .
we further assume that @xmath0 contains a subgroup @xmath26 and we denote its generator in the representation @xmath54 by @xmath85 , @xmath86 . in order to define the direct product @xmath84 ,
@xmath1 should commute with @xmath26 .
this requirement translates into xz^*-zx=0 , [ xz ] as can be checked by computing the subsequent action of @xmath26 and @xmath1 and vice versa on the fields @xmath45 and by requesting the result to be independent of the ordering of these transformations .
also the condition in eq.([xz ] ) is covariant under the change of basis given in eqs .
( [ change1],[change2 ] ) , i.e. @xmath87 with @xmath88 holds .
thus , it is possible to go to a basis in which @xmath85 is diagonal and @xmath35 canonical ( again , this can be checked by explicit computation ) .
we indicate this particular basis by a hat : z=^ , x=^t .
[ hat ] barring the trivial case ( @xmath89 ) , we can assume , without loss of generality , that @xmath90 is of the form = 1&0&0 + 0&-1&0 + 0&0&1 .
in general the conditions of invariance of the neutrino mass matrix @xmath47 under @xmath84 are is not symmetric , the second equality in eq.([condmnu ] ) has to be changed into @xmath91 .
from this equation and the fact that neutrinos are majorana particles follows that @xmath42 has to fulfill @xmath92 .
furthermore , the condition in eq.([xz ] ) constrains the form of @xmath35 .
the admissible non - symmetric @xmath35 then have the property that @xmath42 is proportional to @xmath85 ( so that the first equality in eq.([condmnu ] ) becomes redundant ) .
most importantly , such an @xmath35 constrains the neutrino mass matrix @xmath47 in such a way that the neutrino mass spectrum turns out to be partly degenerate which is not compatible with experimental data @xcite .
this statement holds for any choice of @xmath0 and @xmath85 as long as @xmath35 has to fulfill eq.([xz ] ) .
] [ condmnu ] z^tm_z = m _ , x m_x= m_^ * . by making use of eq .
( [ hat ] ) we see that eq.([condmnu ] ) takes the form ( ^t m _ ) = ( ^t m _ ) , ( ^t m_)=(^t m_)^ * .
these conditions are satisfied by ^t m_= m_11 & 0 & m_13 + 0&m_22&0 + m_13&0&m_33 , m_ij = m_ij^ * .
the original matrix @xmath47 can be diagonalized by [ unuomrk ] u_^t m_u_= m_^diag , u_= r
( ) k , where @xmath93 is a rotation matrix r()= & 0 & + 0 & 1 & 0 + - & 0 & , 2= .
[ rtheta ] the unitary matrix @xmath94 is diagonal with entries @xmath95 and @xmath96 which encode the cp parity of the neutrino states and it renders @xmath97 positive ( semi-)definite . note that the fundamental interval of the parameter @xmath2 is @xmath98 , since @xmath99 and the diagonal matrix can be absorbed into the matrix @xmath94 .
this fixes the contribution from the neutrino sector to the lepton mixing up to permutations of the columns , since neutrino masses are unconstrained in the present framework .
the unitary matrix @xmath100 from the charged lepton sector is determined by requiring invariance of @xmath81 under the subgroup @xmath16 which is in general the product of cyclic groups @xmath101 , @xmath102 . denoting the generator of the cyclic group @xmath101 by @xmath103 , the invariance conditions read [ condml ] q_i^m_l^m_l q_i = m_l^m_l for all @xmath104 .
since the set of generators @xmath103 distinguishes between the three generations of charged leptons , the unitary matrix @xmath100 which simultaneously diagonalizes them is determined ( up to permutations of columns and phases of the column vectors ) [ qicond ] u_e^q_i u_e = _ i with @xmath105 diagonal . plugging eq.([qicond ] ) into eq.([condml ] ) shows that @xmath100 also diagonalizes @xmath106 , i.e. @xmath107 . finally , we have [ pmnsres ] u_pmns = u_e^(q_i ) ( z , x ) r ( ) k , up to permutations of rows and columns . we have spelled out the dependence on the choice of subgroups , which is specified by the generators of @xmath16 and @xmath108 , and on the @xmath1 transformation , i.e. the set @xmath109 . in our approach neutrino and charged lepton masses remain as undetermined parameters and thus the ordering of rows and columns of the pmns matrix is not fixed . for a given
set @xmath110 all three mixing angles , the dirac phase and the two majorana phases are determined in terms of the parameter @xmath2 whose size depends on the neutrino mass matrix elements @xmath111 , see eq.([rtheta ] ) .
the majorana phases are fixed up to the contribution from the matrix @xmath94 which can only shift the phases by @xmath4 . due to the covariance of all relevant equations , i.e. the independence of the choice of a particular basis , also the results for the mixing parameters
are basis - independent .
furthermore , they also do not depend on the above - mentioned freedom to multiply @xmath35 with an arbitrary phase @xmath112 .
we conclude this section with some remarks about possible accidental @xmath1 symmetries .
in particular , we show conditions which have to be necessarily fulfilled by the generators of the subgroups @xmath16 and @xmath17 , if such an accidental symmetry is present .
such conditions turn out to be useful for understanding the results of the example @xmath7 that we discuss in the next section .
we have imposed @xmath1 conservation in the neutrino , but not in the charged lepton sector .
thus , we might be led to the conclusion that non - trivial dirac and majorana phases are always generated in our approach .
actually this is not the case .
trivial phases are found when the mass matrices @xmath81 and @xmath47 , constrained by our choice of @xmath109 , satisfy the invariance conditions [ invcond2 ] y^ * m_l^m_l y = ( m_l^m_l ) ^ * , y m_y = m_^ * , for some unitary symmetric matrix @xmath113 , to which we refer as an accidental @xmath1 symmetry . using the results of appendix [ appa ] together with those of subsection [ setup ] we can rewrite these conditions in terms of @xmath109 instead of the mass matrices @xmath46 and @xmath47 .
as one can see , @xmath113 satisfies the first equality of eq .
( [ invcond2 ] ) if and only if [ xqcond ] q_i y - y q_i^t = 0 for all @xmath104 .
this condition ensures that @xmath113 is diagonal in the same basis as @xmath103 , that is @xmath113 is diagonal in the charged lepton mass basis .
notice that if @xmath113 satisfies eq .
( [ xqcond ] ) also @xmath114 , @xmath115 , does so .
similarly , one can check that the second equality of eq .
( [ invcond2 ] ) implies [ xxzcond ] z y - y z^ * = 0 , x y^ * - y x^ * = 0 .
the first equality is of the same form as the requirement for having a direct product of @xmath116 , see eq.([xz ] ) , while the second equality states that the two @xmath1 transformations @xmath35 and @xmath113 commute .
these conditions are , however , only necessary but not sufficient to ensure that the second equality of eq.([invcond2 ] ) holds .
we can distinguish the following cases : @xmath117 we can not find a @xmath1 symmetry @xmath113 which fulfills eqs.([invcond2]-[xxzcond ] ) .
then we have to expect non - trivial dirac and majorana phases ; @xmath118 a @xmath1 symmetry @xmath113 exists which fulfills these equations and it is furthermore real and diagonal in the neutrino mass basis .
then all cp phases are trivial , see eq.([phasestriv ] ) and appendix [ appa ] ; @xmath119 we find a @xmath1 transformation @xmath113 which fulfills eq.([xqcond ] ) and is diagonal in the neutrino mass basis ; however , it is not real in this basis , i.e. @xmath113 satisfies the first equality in eq.([xxzcond ] ) , but not the second one .
then @xmath113 is not a @xmath1 symmetry of the setup , but it still leaves @xmath81 and @xmath120 invariant , as it is the case for dirac neutrinos , compare eq.([diracnu ] ) .
then we know @xmath121 and @xmath122 .
furthermore , we can show that @xmath123 , see appendix [ appa ] , implies that [ majofixed ] ||= |(x^_1 -x^_2)| , diagonal entries of @xmath113 in the neutrino mass basis . or @xmath124 . ] in a setup characterized by @xmath109 a candidate for an accidental symmetry is @xmath125 which always satisfies eq .
( [ xxzcond ] ) and which is real and diagonal in the neutrino mass basis.,[condmnu ] ) follows that @xmath126 satisfies @xmath127 and that it fulfills the constraints in eqs.([xcond1],[xcond2 ] ) . ]
we analyze here the mixing patterns originating from the breaking of @xmath19 and @xmath1 to @xmath128 and to @xmath16 being an abelian subgroup of @xmath19 .
we choose @xmath19 , since it is among the smallest discrete groups with an irreducible three - dimensional representation and it is well - known to be the smallest symmetry group which leads to tb mixing , if broken in a non - trivial way @xcite .
we first present the group @xmath19 in a basis convenient for our purposes and then discuss in detail the findings of our comprehensive study of the group @xmath19 and @xmath1 .
the group @xmath19 can be defined in terms of three generators @xmath129 , @xmath130 and @xmath131 @xcite which fulfill the following relations @xmath132 & & ( s t)^3 = e \ ; , \;\ ; ( s u)^2 = e \ ; , \;\ ; ( t u)^2 = e \ ; , \;\ ; ( s t u)^4 = e \end{aligned}\ ] ] with @xmath133 being the neutral element of @xmath19 . note that the generators @xmath129 and @xmath130 alone give rise to the group @xmath18 . in the following we are only interested in the two irreducible ( faithful ) three - dimensional representations , called @xmath134 and @xmath135 in the notation of @xcite , to which we assign the three generations of left - handed leptons .
we choose real representation matrices @xmath129 , @xmath130 and @xmath131 ( called in the same way as the abstract elements of the group @xmath19 ) for the representation @xmath135 : @xmath136 the representation matrices of the other three - dimensional representation @xmath134 read @xmath129 , @xmath130 and @xmath137 .
notice that our choice of basis for @xmath129 , @xmath130 and @xmath131 is related to the one of @xcite with @xmath138 , @xmath139 and @xmath140 , through the unitary transformation @xmath141 @xmath142 so that @xmath143 ] the group @xmath19 has @xmath26 , @xmath144 , @xmath145 and @xmath146 as abelian subgroups .
the nine @xmath26 symmetries are generated by the elements @xmath147 while the generators of the four @xmath144 symmetries can be chosen as @xmath148 and those of the three @xmath145 symmetries as @xmath149 the @xmath26 generating elements are divided into two classes ( the first three of the list in eq.([z2gen ] ) and the last six ones ) , while the @xmath144 subgroups as well as the @xmath145 subgroups of @xmath19 are all conjugate to each other .
there are four klein subgroups .
one of them , called @xmath150 , is normal , while the three other ones @xmath151 , @xmath152 are conjugate to each other .
possible sets of generators of the different klein subgroups are @xmath153 we briefly repeat the mixing patterns which can be derived from the group @xmath19 as well as @xmath18 , if the group @xmath17 is a klein group and @xmath16 an abelian subgroup of @xmath0 , capable to distinguish the three generations of charged leptons .
all these results can be found in @xcite .
we have three different viable choices for the group @xmath16 if @xmath7 : @xmath154 , @xmath155 and @xmath156 .
as is well - known , the choice @xmath154 leads to tb mixing @xcite ( for example , we can choose @xmath130 as generator for @xmath16 and @xmath129 and @xmath131 for @xmath17 ) , while @xmath155 ( one possible choice of generator is @xmath157 ) and @xmath156 ( for example generated by the elements @xmath158 and @xmath159 ) give rise to bimaximal ( bm ) mixing @xcite . in the case of @xmath34
, we have a unique choice for @xmath25 ( to be generated by @xmath129 and @xmath158 ) and as choice for @xmath16 only @xmath154 ( for example , generated by @xmath130 ) . the mixing pattern is given by the familiar democratic mixing matrix , in which all elements have the same absolute value and the mixing parameters read @xmath160 , @xmath161 , @xmath162 , @xmath163 and @xmath164 .
this matrix has already been discussed many years ago as possible lepton mixing matrix @xcite .
we show the results of a comprehensive study in which we assume @xmath7 , @xmath6 and @xmath16 being @xmath165 , @xmath155 or @xmath166 . in order to facilitate understanding we present our results in terms of examples for the different cases .
we find that it is sufficient to consider only a small number of cases which lead to different results for mixing angles and cp phases , since other possible choices of @xmath103 , @xmath85 and @xmath35 are related by similarity transformations -belonging to the group @xmath19- to our representative solutions and thus can not lead to new results .
we concentrate in our discussion on the representation @xmath167 of @xmath19 . however
, if we assigned the three generations of left - handed leptons to the representation @xmath134 instead , the results would be the same and no additional results would be found , because the generators of the triplet @xmath134 just differ in the overall sign of the generator @xmath131 from those of the representation @xmath167 .
it follows a short subsection with technical details necessary for the derivation of the results .
all possible choices of @xmath85 and @xmath35 are related through similarity transformations ( contained in @xmath19 ) to the following three @xmath85 @xmath168 and their corresponding @xmath169 which fulfill the requirements stated in eqs.([xcond1],[xcond2],[xz ] ) of section [ sec2 ] .
for all @xmath85 @xmath170 are admissible and for @xmath171 in addition @xmath172 as mentioned in section [ sec2 ] , these transformations are defined up to an overall phase .
@xmath173 is the canonical @xmath1 transformation .
notice that in our particular basis the different @xmath169 turn out to be proportional to elements of the group @xmath19 : @xmath174 , @xmath175 , @xmath176 , @xmath177 and @xmath178 .
however , this is just a coincidence and in general the transformations @xmath169 do not belong to the flavour group @xmath0 .
we list the transformations @xmath179 which bring the different @xmath169 into the canonical form @xmath180 , see eqs.([change2],[hat ] ) : @xmath181 we note that , obviously , for different choices of @xmath85 , but using the same form of the matrices @xmath179 , the rotation matrix @xmath182 defined in eq .
( [ rtheta ] ) changes its form : it is a rotation in the ( 13)-plane for @xmath171 , a rotation in the ( 23)-plane for @xmath183 and for @xmath184 a rotation in the ( 12)-plane .
the different choices of @xmath16 , for which we discuss lepton mixing , can be represented by lll q = t & & g_e = z_3 , + q = s t u & & g_e = z_4 , + q_1=t s t^2 s , q_2=u t^2 & & g_e = z_2 z_2 . the matrix @xmath100 which diagonalizes the charged lepton mass matrix @xmath81 is
then of the form @xmath185 @xmath186 and @xmath187 we have performed a comprehensive study in which we consider all possible choices of @xmath103 , @xmath85 and @xmath35 and all possible permutations of columns and rows of the mixing matrix .
however , we show in the following only results which we consider phenomenologically interesting in the sense that we can find a value of the parameter @xmath2 such that the resulting mixing angles are reasonably close to their best fit values which we take from @xcite . as measure we use a @xmath188 function defined in the usual way and require its minimal value to be less than @xmath189 . in this way ,
all solutions leading to vanishing @xmath20 independent of the parameter @xmath2 are excluded , since @xmath190 is disfavoured at the @xmath191 level by global fits , @xmath192 , @xcite .
we define two different @xmath188 functions , because @xmath193 has two best fit values @xmath194 and @xmath195 .
errors are not completely gaussian for @xmath193 .
however , we use for the smaller best fit value a @xmath196 error of @xmath197 and for the larger one an error of @xmath198 . ] for this reason , we display in the tables below in several occasions different best fit values @xmath199 for the parameter @xmath2 for which the @xmath188 function has a global minimum . in the tables
we also display the results for the sines of the cp phases @xmath30 , @xmath52 and @xmath53 ( and the jarlskog invariant @xmath50 ) .
these quantities are presented in terms of absolute values , since the sign of the jarlskog invariant @xmath50 depends on the ordering of rows and columns , while the sign of @xmath200 and @xmath201 depends on the cp parity of the neutrino states which is encoded in the matrix @xmath94 , see eq.([unuomrk ] ) in section [ sec2 ] ( changing cp parity shifts the majorana phase by @xmath4 ) . requiring @xmath202 , we find five viable solutions ( case i , ii , iv , v and the case with @xmath155 or @xmath156 ) .
however , in all these cases the majorana phases are trivial , i.e. @xmath203 and @xmath204 .
thus , we have included another case , called case iii , which leads to non - trivial majorana phases depending on the parameter @xmath2 , although the minimum value of its @xmath188 functions is above 100 . in the following
we put special emphasis on the results for the cp phases @xmath30 , @xmath52 and @xmath53 , because frequently an accidental @xmath1 symmetry is present which leads to a trivial dirac phase and/or majorana phases .
we first take @xmath154 and choose @xmath205 .
five different cases can be distinguished and are represented by ccc & z = s , & x = x_1 + & z = s , & x = x_3 + & z = s , & x= x_5 + & z = s u , & x= x_1 + & z = s u , & x = x_2
. notice that in case i @xmath206 is also a @xmath1 symmetry of the neutrino sector , since @xmath207 .
similarly , it holds in case ii that @xmath208 is also a @xmath1 symmetry and in case iii it is @xmath209 .
analogously , we find that in case iv @xmath210 is also present and in case v @xmath211 .
thus , the six possible choices of @xmath169 for @xmath171 , mentioned in subsection [ technical ] , give rise to three independent cases and for @xmath183 the four possible @xmath169 lead to effectively two different cases . in cases
i , ii , iv and v we find a value of @xmath2 for which the computed mixing angles agree rather well with the ones obtained in global fits @xcite .
this is shown in table [ tab_c1_c5 ] together with the formulae for the mixing parameters in terms of generic @xmath2 .
.results for the mixing parameters in terms of the parameter @xmath2 for the cases i , ii , iv and v. we display the best fit value @xmath199 for @xmath2 for which the @xmath188 functions have a global minimum @xmath212 . since the global fit @xcite gives two best fit values for @xmath193 , one with @xmath213 and one with @xmath214 , we distinguish these two possibilities . since the cp phases are independent of @xmath2 , we only give the values of the mixing angles and the absolute value of @xmath50 for @xmath215 .
@xmath216 in the special case @xmath217 @xmath218 vanishes , since then @xmath50 vanishes . however , then the mixing angles are in considerable disagreement with the best fit values from @xcite .
[ tab_c1_c5 ] [ cols="^,^,^,^,^",options="header " , ] the best fit value @xmath199 is @xmath219 because for this value the result for @xmath220 is minimized : @xmath221 .
notice that for @xmath222 the dirac phase @xmath30 is trivial and also one of the majorana phases .
the former is trivial because of a common @xmath1 symmetry @xmath223 of the matrices @xmath81 and @xmath120 , as explained in subsection [ cpacc ] .
this @xmath1 transformation fulfills eq.([xqcond ] ) for @xmath205 and its form in the neutrino mass basis is @xmath224 this matrix is for generic values of the parameter @xmath2 neither diagonal nor real .
however , if @xmath225 ( as is true for @xmath226 ) , @xmath227 becomes diagonal . since
its entries are not real ( independently of @xmath94 ) , @xmath228 is not a symmetry of the neutrino mass matrix itself but only of @xmath120 and consequently majorana phases are not trivial in general . according to eq .
( [ majofixed ] ) , they can be directly read off from @xmath229 : @xmath230 and @xmath231 . for the choice @xmath155
we find only one case which passes our selection criteria .
we choose -as mentioned above- as representative @xmath145 generating element @xmath232 and take for @xmath85 and @xmath35 @xmath233 notice that the same results would be obtained with the @xmath1 symmetry @xmath234 , since @xmath235 .
the results for the mixing parameters and the best fit values @xmath199 of the parameter @xmath2 can be found in table [ tab_z4 ] and the mixing angles are displayed in figure 3 .
we notice the existence of two solutions , case a and case b , which arise from the freedom to exchange the second and the third rows of the pmns matrix .
the @xmath188 function of case a has a global minimum with @xmath236 , while that of case b has a global minimum with @xmath237 .
these best fit points @xmath199 are indicated as red dots in figure 3 .
in addition , we mark certain values of @xmath2 in order to guide the eye .
cp phases are trivial independently of the parameter @xmath2 which indicates the presence of an accidental @xmath1 symmetry @xmath113 in the charged lepton and neutrino sector . indeed , the @xmath1 symmetry @xmath238 of the neutrino sector is also a @xmath1 symmetry of the charged lepton mass matrix @xmath106 , because @xmath232 and @xmath206 fulfill eq .
( [ xqcond ] ) . as one can check , the pmns matrix has one column of the form @xmath239 or @xmath240 which is in common ( up to permutation ) with the bm mixing pattern originating from the charged lepton sector ( see eq.([ueqstu ] ) for @xmath155 and eq.([ueq1q2 ] ) for @xmath156 ) .
patterns with such a column have been recently mentioned in @xcite .
if a klein group is preserved in the charged lepton sector and @xmath6 , we find the same results as for @xmath16 being a @xmath145 symmetry .
one representative choice of the generators of the klein group in the charged lepton sector is @xmath241 and for @xmath85 and @xmath35 of the neutrino sector we take @xmath242 all cp phases are trivial , because the imposed @xmath1 symmetry of the neutrino sector @xmath243 , the canonical @xmath1 symmetry , is also present in the charged lepton sector .
as one can check , the two generators @xmath244 and @xmath245 fulfill : @xmath246 with @xmath247 as required for @xmath243 by the constraint in eq.([xqcond ] ) . the choice of @xmath228 as @xmath1 symmetry in the neutrino sector leads to the same results as the choice of @xmath173 , since @xmath248 . in appendix [ appb ]
we also discuss the mathematical structure of the groups arising from the combination of the group @xmath19 with one of the possible @xmath1 transformations @xmath169 . as has been mentioned at the beginning of subsection [ setup ]
, the maximal symmetry of a neutrino mass matrix @xmath47 for three generations of majorana neutrinos is a klein group , if neutrino masses are unconstrained . in the present approach we assume in general only one of these @xmath26 symmetries to be a subgroup of the flavour symmetry @xmath7 .
( this is similar to the approach in @xcite . )
however , we notice that for the various cases presented particular values of the parameter @xmath2 ( in general different for the different cases ) exist for which the second @xmath26 symmetry of the neutrino mass matrix @xmath47 can be promoted to a subgroup of a finite flavour group which contains @xmath19 .
thus , in these cases the result of the mixing angles ( and the cp phases ) can be achieved through the symmetry breaking of this finite flavour group to a klein group ( and the symmetry @xmath1 ) in the neutrino sector instead to @xmath116 .
in all cases displayed in table [ tab_c1_c5 ] the limit @xmath249 corresponds to tb mixing ( and trivial cp phases ) . in this limit
the second @xmath26 symmetry of the neutrino sector turns out to be generated by @xmath131 , belonging to @xmath19 , so that the residual symmetry in the neutrino sector is @xmath250 instead of only @xmath116 .
the results of the mixing angles are thus the same as in the case without a @xmath1 symmetry in the neutrino sector , see subsection [ a4s4gfpatterns ] . in case i also the limit
@xmath251 is noteworthy in which the mixing parameters take the values known from the democratic mixing matrix .
this can be understood , because the second @xmath26 symmetry of the neutrino mass matrix is now generated by @xmath158 so that again the neutrino sector is invariant under a residual symmetry @xmath250 .
as mentioned in subsection [ groups4 ] , @xmath129 and @xmath130 generate the group @xmath18 and according to subsection [ a4s4gfpatterns ] the resulting mixing pattern is the democratic one .
these special points are marked in figures 1 and 2 with a cyan dot and a green diamond , respectively .
interestingly , in case ii the pmns matrix is of the form @xmath252 this result should be compared with the findings of @xcite in which it has been shown that the pmns matrix takes this form ( with a matrix @xmath94 with arbitrary phases , because majorana phases are not determined in that approach ) for @xmath253 and @xmath254 , if the groups @xmath255 and @xmath23 are broken to a @xmath144 symmetry in the charged lepton sector and to a klein group in the neutrino sector , respectively . for case iii note that the results of mixing angles and the dirac phase coincide for the best fit point @xmath256 with those found in the case of a flavour group @xmath255 being broken to a klein group in the neutrino sector and to a @xmath144 symmetry in the charged lepton one @xcite .
indeed , we can check that for @xmath226 the second @xmath26 symmetry under which the neutrino mass matrix @xmath47 is invariant can be generated by @xmath257 obviously , this matrix does not represent an element of the group @xmath19 in our real basis for @xmath129 , @xmath130 and @xmath131 . if we consider the group generated by @xmath129 , @xmath130 , @xmath131 and @xmath258 , we find @xmath255 @xcite using the computer program gap @xcite . another particular value of the parameter @xmath2 is @xmath259 , because in this case the element @xmath158 is promoted to a @xmath26 generator under whose action the neutrino mass matrix is invariant ( independently of any @xmath1 transformation ) .
both the residual symmetry @xmath154 in the charged lepton and @xmath146 in the neutrino sector are then generated through elements written in terms of @xmath129 and @xmath130 only so that the relevant flavour group is @xmath18 rather than @xmath19 . as has been recapitulated in subsection [ a4s4gfpatterns ]
, the flavour group @xmath18 broken in this way leads to the democratic mixing pattern which coincides with the results of case iii for @xmath259 . for @xmath155 one particular limit
is given by @xmath249 for which the ( absolute values of the ) pmns matrix of case a and case b take(s ) the form ||u_pmns , a|| = ( ccc 1/2 & 1/2 & 1/ + 1/ & 1/ & 0 + 1/2 & 1/2 & 1/ ) , ||u_pmns , b|| = ( ccc 1/2 & 1/2 & 1/ + 1/2 & 1/2 & 1/ + 1/ & 1/ & 0 ) , respectively . as one recognizes , this is the bm mixing matrix up to permutations of rows and columns . from the viewpoint of group theory
the value @xmath260 allows to promote @xmath129 to the generator of one of the @xmath26 symmetries of the matrix @xmath47 and thus the klein group containing @xmath129 and @xmath131 is conserved in the neutrino sector . as repeated in subsection [ a4s4gfpatterns ] , the choice @xmath7 , @xmath155 and @xmath17 being a klein group leads to bm mixing .
if @xmath16 is also a klein group , very similar statements hold : namely , @xmath259 promotes @xmath129 to the generator of a @xmath26 symmetry of the matrix @xmath47 and the breaking of @xmath7 to @xmath261 being ( non - normal ) klein groups with @xmath262 leads to bm mixing .
we can easily deduce results for @xmath34 from those obtained for @xmath7 .
as mentioned in subsection [ groups4 ] the generators @xmath129 and @xmath130 alone give rise to @xmath18 .
thus , the only thing to do is to consider the above results which can be obtained with generators @xmath103 and @xmath85 made up of @xmath129 and @xmath130 .
this constrains us to take @xmath16 as @xmath144 symmetry .
furthermore , only case i , ii and iii are admissible , if @xmath85 is required to be a product of @xmath129 and @xmath130 . from the viewpoint of @xmath18 case
i and case iii are qualitatively different from case ii , since in the latter case @xmath35 is not proportional to an element of @xmath18 , while it is for case i and case iii .
the recent measurement of the reactor mixing angle @xmath20 has ruled out several of the prominent discrete flavour symmetries leading to vanishing or very small @xmath20 .
this fact has led to several new ideas .
for instance , we can still adopt a discrete flavour symmetry giving rise to tb mixing in first approximation .
however , in such a case , rather large corrections are needed in order to achieve the measured value of the reactor mixing angle . at the same time , predictability is in general reduced , since such corrections tend to affect all mixing angles . to retain predictability and to reproduce already in the lowest order approximation realistic mixing patterns
it is necessary to consider groups with larger order and several candidates have been found .
another possibility consists in reducing the residual symmetries in the neutrino and/or charged lepton sector . in this case
the symmetries do not completely determine all three mixing angles and it is possible to accommodate a non - vanishing value of @xmath20 .
however , the latter is then rather fitted than predicted . in this paper
we have analyzed a further possibility .
we have considered a framework that is as predictive as possible , without necessarily requiring large discrete groups as starting point . following earlier ideas found in the literature , we have included @xmath1 in the set of symmetries of the theory and have analyzed which role @xmath1 could play in the symmetry breaking pattern . to define our framework
we have first derived conditions that @xmath1 and the flavour symmetries have to obey in order to provide a consistent set of transformations . given a flavour group ,
these conditions constrain the possible choice of @xmath1 transformations .
we have then discussed , in a model - independent way , a possible symmetry breaking pattern and its consequences for the prediction of the mixing parameters .
assuming three generations of majorana neutrinos we have studied the case in which a flavour symmetry @xmath0 and a generalized @xmath1 transformation are combined and are broken to @xmath6 in the neutrino sector , while @xmath5 remains preserved in the charged lepton one .
our main result is that all mixing angles and all cp phases , dirac and majorana , are determined in terms of a single real parameter @xmath2 , which can take values between @xmath3 and @xmath4 .
this has to be compared with the approach in which @xmath1 is absent , and thus majorana phases can not be constrained . in order to show concrete examples and find new interesting mixing patterns ,
we have performed a comprehensive study for the group @xmath7 .
we have found several @xmath1 transformations compatible with all our requirements , which lead to different results for mixing angles and cp phases .
out of all possibilities only five give rise to phenomenologically interesting patterns , i.e. patterns for which the mixing angles are reasonably close to the best fit values for a certain value of the parameter @xmath2 . in particular ,
a non - vanishing value of the mixing angle @xmath20 can be easily accommodated and all mixing parameters are strongly correlated . among the interesting cases we find two cases , case i and case iv in our classification , in which the neutrino mass matrix @xmath47 is invariant under @xmath27 reflection symmetry in the charged lepton mass basis .
this particular generalized @xmath1 transformation has already been studied in the literature .
surprisingly , several cases lead to trivial cp phases . for such cases
we have figured out the presence of accidental @xmath1 symmetries , not required a priori by the assumed symmetry breaking pattern . in order to illustrate a case
in which all cp phases are non - trivial functions of the parameter @xmath2 , we have also discussed another case ( case iii ) , which however can not accommodate the best fit values of the mixing angles so well .
furthermore , we have briefly mentioned the cases which can also be obtained , if @xmath0 is @xmath18 instead of @xmath19 . throughout this paper
we have discussed the model - independent implications of our proposal .
physical results only depend on the assumed symmetry breaking pattern and are independent of the details of a specific implementation , such as the particle content of the flavour symmetry breaking sector or the possible additional symmetries of the theory .
nevertheless we think that it would be interesting to implement one of the presented cases in an actual model , for example in a supersymmetric context in which the breaking of the symmetry group to @xmath16 and @xmath17 is spontaneous due to non - vanishing vacuum expectation values of some flavons .
in such a model for case i or case iv the prediction of a maximal dirac phase @xmath30 for generic @xmath2 would not depend on the parameters of the theory and thus could represent a concrete realization of geometrical cp violation .
apart from the usual challenge that corrections can spoil some of the leading order predictions , models in which the presented idea is implemented have to face another challenge , namely the prediction of the parameter @xmath2 which depends on the entries of the neutrino mass matrix , and whose size is important for accommodating the best fit values of the mixing angles well .
ff and ch have been partly supported by the european programme `` unification in the lhc era '' , contract pitn - ga-2009 - 237920 ( unilhc ) .
ch is also supported by the erc advanced grant no .
267985 , `` electroweak symmetry breaking , flavour and dark matter : one solution for three mysteries '' ( damesyfla ) .
in this appendix we first fix our notation and conventions for the mixing parameters and then prove that the presence of a generalized @xmath1 transformation implies the vanishing of dirac as well as majorana phases .
as parametrization of the pmns matrix we use u_pmns = diag(1 , e^i /2 , e^i ( /2 + ) ) , with @xmath263 being of the form of the cabibbo - kobayashi - maskawa ( ckm ) matrix @xmath264 @xcite = c_12 c_13 & s_12 c_13 & s_13 e^-
i + -s_12 c_23 - c_12 s_23 s_13 e^i & c_12 c_23 - s_12 s_23 s_13 e^i & s_23 c_13 + s_12 s_23 - c_12 c_23 s_13 e^i & -c_12 s_23 - s_12 c_23 s_13 e^i & c_23 c_13 the presence of a generalized @xmath1 transformation @xmath35 under which charged lepton as well as neutrino mass matrices are invariant leads to the vanishing of all three cp invariants @xmath50 , @xmath278 and @xmath279 and thus to trivial dirac and majorana phases . in the following , we focus on the case of majorana neutrinos , but also briefly comment on dirac neutrinos .
all these results are known in the literature , see e.g. @xcite . in order to study the consequences of @xmath1 invariance on the lepton mixing parameters , we diagonalize the mass matrices @xmath46 and @xmath47 by unitary matrices @xmath100 and @xmath280 , respectively , m_l^m_l = u_e ( m_l^m_l)^diag u_e^ , [ ldiag ] m_= u_^ * m_^diag u_^ , so that the invariance conditions in eq . ( [ invcond ] ) in subsection [ cpgen ] become [ condk ] x_e^ * ( m_l^m_l)^diag x_e = ( m_l^m_l)^diag , x_m_^diag x_= m_^diag , with [ xcond ] x_e = u_e^x u_e^ * , x_= u_^x u_^ * . the matrices @xmath281 and @xmath282 represent the @xmath1 transformation @xmath35 in the lepton mass basis . for non - degenerate charged lepton and neutrino masses one can show that the general solutions of eq .
( [ condk ] ) for @xmath281 and @xmath282 are diagonal matrices [ conddiag ] x_e = e^i x_1 & 0 & 0 + 0 & e^i x_2 & 0 + 0 & 0 & e^i x_3 , x_= s_1 & 0 & 0 + 0 & s_2 & 0 + 0 & 0 & s_3 with [ condxphases ] x_i [ 0 , 2 ) , s_i = 1 .
this can be stated differently : the lepton sector is invariant under a generalized @xmath1 transformation described by @xmath35 , if one can find matrices @xmath281 and @xmath282 of the form as in eqs.([conddiag ] , [ condxphases ] ) , such that [ supercond ] u_x_u_^t = u_e x_e u_e^t .
provided that such @xmath281 and @xmath282 exist , we can study the consequences for the pmns matrix which is given by [ upmns ] u_pmns = u_e^u _ . from eqs.([supercond],[upmns ] ) we see that [ xupmnscond ] x_e^ * u_pmns x_= u_pmns^holds . plugging in eqs.([conddiag],[condxphases ] ) , we find @xcite [ pmnscpcon ] u_pmns , ij e^-i x_i s_j = u^*_pmns , ij .
the jarlskog invariant @xmath50 in eq.([jcp ] ) and @xmath283 then vanish .
also @xmath271 and @xmath272 vanish , if the pmns matrix fulfills eq.([pmnscpcon ] ) .
thus , also the majorana phases @xmath52 and @xmath53 are trivial in this case . these conclusions
hold unless the mixing angles take special values , as commented at the end of appendix [ appa1 ] . for dirac neutrinos ,
@xmath47 and @xmath282 are subject to the same constraints as the matrices @xmath46 and @xmath281 , see eq.([diracnu ] ) , and thus @xmath47 and @xmath282 have to satisfy ( only ) [ cpdirac1 ] x_^ * ( m_^m_)^diag x_= ( m_^m_)^diag instead of eqs.([invcond],[condk ] ) . as a consequence , the most general form of @xmath282 fulfilling eq.([cpdirac1 ] ) for a non - degenerate neutrino mass spectrum
is the same as for @xmath281 in eqs.([conddiag],[condxphases ] ) ( with @xmath284 called @xmath285 ) .
if this is realized , eq.([xupmnscond ] ) implies [ cpdirac2 ] u_pmns , ij e^-i ( x_i - x^_j ) = u^*_pmns , ij and thus a vanishing jarlskog invariant ( and trivial dirac phase ) .
in this appendix , we comment on the mathematical structure of the group comprising @xmath0 and @xmath1 and show that it is in general a semi - direct product .
we assume to be given a solution @xmath35 of the constraint in eq .
( [ xcond2 ] ) and that @xmath54 is a faithful representation is represented by a different representation matrix . ] of @xmath0 ( which does not need to be discrete and/or finite in the following ) . in order to analyze
which group arises from combining the transformations of @xmath0 and those of @xmath37 , the parity group generated by @xmath1 , we enlarge the field space from @xmath38 to @xmath286 . in this space
the actions of @xmath0 and @xmath1 are given by : =a =x , with [ calax ] a= , respectively .
the unitary matrices @xmath287 generate a group which is isomorphic to @xmath0 .
the matrix @xmath288 satisfies x^2= , [ xcond1 ] since @xmath35 is unitary and symmetric , and thus generates a group @xmath289 isomorphic to @xmath37 .
we do not distinguish between two isomorphic groups and we use the same notation for both .
the consistency condition in eq .
( [ xcond2 ] ) reads x^-1 ax = a .
[ xcond2 ] the group @xmath80 we are looking for is that generated by the unitary matrices @xmath287 and @xmath288 .
we observe that the closure of this set of transformations is guaranteed by eqs .
( [ xcond1],[xcond2 ] ) .
indeed by using these equations we can show that any element of @xmath80 can be cast into the form @xmath290 with @xmath287 belonging to @xmath0 and @xmath291 belonging to the group @xmath37 .
such a decomposition is unique , namely @xmath292 implies @xmath293 and @xmath294 .
this is equivalent to the statement that the intersection of the two groups @xmath0 and @xmath37 is only the neutral element .
the multiplication rule of two elements of @xmath80 follows from eqs .
( [ xcond1],[xcond2 ] ) and is given by ah ah=aa
a=hah .
therefore the group @xmath80 is isomorphic to the semi - direct product of @xmath0 and @xmath37 , @xmath295 and @xmath37 acts on @xmath0 .
@xmath0 is a normal subgroup of @xmath80 , while @xmath37 is in general only a subgroup .
if @xmath0 is a finite group , the group @xmath80 has twice as many elements as @xmath0 .
the semi - direct product reduces to a direct one if and only if @xmath288 commutes with all elements of @xmath0 , that is when the condition in eq .
( [ xcond2 ] ) is satisfied with @xmath293 for all elements @xmath287 of @xmath0 .
( this is equivalent to the case in which the condition in eq . ( [ xcond2 ] ) is satisfied with @xmath296 for all @xmath55 . )
then the subgroup @xmath37 is also normal .
in the particular case @xmath7 , for which we discuss in detail the possible choices of @xmath35 as well as the phenomenological results for lepton mixing angles and cp phases in section [ sec3 ] , we find the following : among the various @xmath169 , see eqs.([admissx_1 ] , [ admissx_2 ] ) , only the canonical @xmath1 transformation @xmath243 leads to a direct product @xmath298 , because only in this case @xmath299 holds for all generators @xmath55 ( and thus all elements ) of @xmath19 ( we have used the fact that all representation matrices are real ) . in all the other cases @xmath300 is only satisfied for one or two of the three generators @xmath129 , @xmath130 and @xmath131 . in particular , for @xmath301 , @xmath302 and @xmath303 the relation @xmath299 obviously holds for @xmath304 and @xmath305 , whereas for @xmath306 we find @xmath307 , @xmath308 and @xmath309 , respectively .
for the two @xmath1 transformations @xmath310 and @xmath311 @xmath299 is only fulfilled for the generator @xmath304 , as required in order to preserve the direct product of the @xmath26 symmetry generated by @xmath129 and of the @xmath1 transformation @xmath35 in the neutrino sector , see eq.([xz ] ) . for @xmath305 we find instead @xmath312 in both cases and @xmath306 gives rise to @xmath313 and @xmath314 , respectively .
00 g. altarelli , f. feruglio , rev .
* 82 * ( 2010 ) 2701 - 2729 .
[ arxiv:1002.0211 [ hep - ph ] ] ; h. ishimori , t. kobayashi , h. ohki , y. shimizu , h. okada , m. tanimoto , prog . theor . phys .
* 183 * ( 2010 ) 1 - 163 .
[ arxiv:1003.3552 [ hep - th ] ] .
p. f. harrison , d. h. perkins and w. g. scott , phys .
b * 530 * ( 2002 ) 167 [ hep - ph/0202074 ] ; z. -z .
xing , phys .
b * 533 * ( 2002 ) 85 [ hep - ph/0204049 ] .
g. altarelli and f. feruglio , nucl .
b * 741 * ( 2006 ) 215 [ hep - ph/0512103 ] ; i. de medeiros varzielas , s. f. king and g. g. ross , phys .
b * 644 * ( 2007 ) 153 [ hep - ph/0512313 ] ; x. -g .
he , y. -y .
keum and r. r. volkas , jhep * 0604 * ( 2006 ) 039 [ hep - ph/0601001 ] ; e. ma , phys .
d * 70 * ( 2004 ) 031901 [ hep - ph/0404199 ] . c. s. lam , phys .
b * 656 * ( 2007 ) 193 [ arxiv:0708.3665 [ hep - ph ] ] ; c. s. lam , phys .
* 101 * ( 2008 ) 121602 [ arxiv:0804.2622 [ hep - ph ] ] ; c. s. lam , phys .
d * 78 * ( 2008 ) 073015 [ arxiv:0809.1185 [ hep - ph ] ] ; e. ma , phys .
b * 632 * ( 2006 ) 352 [ hep - ph/0508231 ] .
k. abe _ et al . _
[ t2k collaboration ] , phys .
lett . * 107 * ( 2011 ) 041801 [ arxiv:1106.2822 [ hep - ex ] ] ; p. adamson _ et al . _
[ minos collaboration ] , phys .
lett . * 107 * ( 2011 ) 181802 [ arxiv:1108.0015 [ hep - ex ] ] ; y. abe _ et al . _
[ double - chooz collaboration ] , phys .
* 108 * ( 2012 ) 131801 [ arxiv:1112.6353 [ hep - ex ] ] ; f. p. an _ et al . _ [ daya - bay collaboration ] , phys .
* 108 * ( 2012 ) 171803 [ arxiv:1203.1669 [ hep - ex ] ] ; j. k. ahn _
[ reno collaboration ] , phys .
* 108 * ( 2012 ) 191802 [ arxiv:1204.0626 [ hep - ex ] ] ; f. p. an _ et al . _ [ daya - bay collaboration ] , chin .
c * 37 * ( 2013 ) 011001 [ arxiv:1210.6327 [ hep - ex ] ] .
m. c. gonzalez - garcia , m. maltoni , j. salvado and t. schwetz , jhep * 1212 * ( 2012 ) 123 [ arxiv:1209.3023 [ hep - ph ] ] .
r. de adelhart toorop , f. feruglio and c. hagedorn , phys .
b * 703 * ( 2011 ) 447 [ arxiv:1107.3486 [ hep - ph ] ] ; r. de adelhart toorop , f. feruglio and c. hagedorn , nucl . phys .
b * 858 * ( 2012 ) 437 [ arxiv:1112.1340 [ hep - ph ] ] .
s. -f .
ge , d. a. dicus and w. w. repko , phys .
b * 702 * ( 2011 ) 220 [ arxiv:1104.0602 [ hep - ph ] ] ; s. -f .
ge , d. a. dicus and w. w. repko , phys .
* 108 * ( 2012 ) 041801 [ arxiv:1108.0964 [ hep - ph ] ] ; d. hernandez and a. y. smirnov , phys .
d * 86 * ( 2012 ) 053014 [ arxiv:1204.0445 [ hep - ph ] ] .
p. f. harrison and w. g. scott , phys .
b * 535 * ( 2002 ) 163 [ hep - ph/0203209 ] ; p. f. harrison and w. g. scott , phys .
b * 547 * ( 2002 ) 219 [ hep - ph/0210197 ] ; p. f. harrison and w. g. scott , phys .
b * 594 * ( 2004 ) 324 [ hep - ph/0403278 ] .
w. grimus and l. lavoura , phys .
b * 579 * ( 2004 ) 113 [ hep - ph/0305309 ] .
y. farzan and a. y. smirnov , jhep * 0701 * ( 2007 ) 059 [ hep - ph/0610337 ] .
r. krishnan , p. f. harrison and w. g. scott , jhep * 1304 * ( 2013 ) 087 [ arxiv:1211.2000 [ hep - ph ] ] .
r. n. mohapatra and c. c. nishi , phys .
d * 86 * ( 2012 ) 073007 [ arxiv:1208.2875 [ hep - ph ] ] .
g. c. branco , j. m. gerard and w. grimus , phys .
b * 136 * ( 1984 ) 383 ; i. de medeiros varzielas and d. emmanuel - costa , phys .
d * 84 * ( 2011 ) 117901 [ arxiv:1106.5477 [ hep - ph ] ] ; i. de medeiros varzielas , d. emmanuel - costa and p. leser , phys .
b * 716 * ( 2012 ) 193 [ arxiv:1204.3633 [ hep - ph ] ] ; i. de medeiros varzielas , jhep * 1208 * ( 2012 ) 055 [ arxiv:1205.3780 [ hep - ph ] ] ; g. bhattacharyya , i. de medeiros varzielas and p. leser , phys . rev . lett . * 109 * ( 2012 ) 241603 [ arxiv:1210.0545 [ hep - ph ] ] . k. s. babu and j. kubo , phys .
d * 71 * ( 2005 ) 056006 [ hep - ph/0411226 ] ; k. s. babu , k. kawashima and j. kubo , phys .
d * 83 * ( 2011 ) 095008 [ arxiv:1103.1664 [ hep - ph ] ] .
m. -c .
chen and k. t. mahanthappa , phys .
b * 681 * ( 2009 ) 444 [ arxiv:0904.1721 [ hep - ph ] ] ; a. meroni , s. t. petcov and m. spinrath , phys .
d * 86 * ( 2012 ) 113003 [ arxiv:1205.5241 [ hep - ph ] ] .
g. ecker , w. grimus and h. neufeld , j. phys .
a * 20 * ( 1987 ) l807 ; h. neufeld , w. grimus and g. ecker , int .
j. mod .
phys . a * 3 * ( 1988 ) 603 .
g. c. branco , r. g. felipe and f. r. joaquim , rev .
* 84 * ( 2012 )
515 [ arxiv:1111.5332 [ hep - ph ] ] . w. grimus and m. n. rebelo , phys . rept .
* 281 * ( 1997 ) 239 [ hep - ph/9506272 ] . c. s. lam , phys
b * 656 * ( 2007 ) 193 [ arxiv:0708.3665 [ hep - ph ] ] ; a. blum , c. hagedorn and m. lindner , phys . rev .
d * 77 * ( 2008 ) 076004 [ arxiv:0709.3450 [ hep - ph ] ] . c. hagedorn , s. f. king and c. luhn , jhep * 1006 * ( 2010 ) 048 [ arxiv:1003.4249 [ hep - ph ] ] . c. s. lam , phys .
d * 83 * ( 2011 ) 113002 [ arxiv:1104.0055 [ hep - ph ] ] . v. d. barger , s. pakvasa , t. j. weiler and k. whisnant , phys .
b * 437 * ( 1998 ) 107 [ hep - ph/9806387 ] ; r. n. mohapatra and s. nussinov , phys .
d * 60 * ( 1999 ) 013002 [ hep - ph/9809415 ] ; g. altarelli and f. feruglio , jhep * 9811 * ( 1998 ) 021 [ hep - ph/9809596 ]
. n. cabibbo , phys .
b * 72 * ( 1978 ) 333 ; l. wolfenstein , phys .
d * 18 * ( 1978 ) 958 .
p. m. ferreira , w. grimus , l. lavoura and p. o. ludl , jhep * 1209 * ( 2012 ) 128 [ arxiv:1206.7072 [ hep - ph ] ] .
x. g. he and a. zee , phys .
b * 645 * ( 2007 ) 427 [ arxiv : hep - ph/0607163 ] ; w. grimus and l. lavoura , jhep * 0809 * , 106 ( 2008 ) [ arxiv:0809.0226 [ hep - ph ] ] ; c. h. albright , w. rodejohann , eur .
j. * c62 * ( 2009 ) 599 - 608 [ arxiv:0812.0436 [ hep - ph ] ] ; y. lin , nucl .
b * 824 * ( 2010 ) 95 [ arxiv:0905.3534 [ hep - ph ] ] ; w. grimus , l. lavoura and a. singraber , phys .
b * 686 * , 141 ( 2010 ) [ arxiv:0911.5120 [ hep - ph ] ] . c. s. lam , phys .
d * 74 * ( 2006 ) 113004 [ hep - ph/0611017 ] ; c. h. albright and w. rodejohann , eur .
j. c * 62 * ( 2009 ) 599 [ arxiv:0812.0436 [ hep - ph ] ] ; x. -g .
he and a. zee , phys .
d * 84 * ( 2011 ) 053004 [ arxiv:1106.4359 [ hep - ph ] ] .
d. hernandez , _ discrete symmetries and lepton mixing _
, talk given at bene 2012 , 18 september 2012 , ictp , trieste , italy ; webpage : ` http://users.ictp.it/~smr2366/ ` .
g. a. miller , h. f. blichfeldt , l. e. dickson , `` theory and applications of finite groups '' , john wiley & sons , new york , 1916 , and dover edition , 1961 ; w. m. fairbairn , t. fulton , and w. h. klink , j. math .
phys . * 5 * ( 1964 ) 1038 ; j. a. escobar and c. luhn , j. math .
* 50 * ( 2009 ) 013524 [ arxiv:0809.0639 [ hep - th ] ] .
gap , `` gap groups , algorithms , and programming , version 4.5.5 '' , ( 2012 ) , ` http://www.gap-system.org ` ; h. u. besche , b. eick , e. obrien , `` smallgroups- library of all small groups , gap package , version included in gap 4.5.5 '' , ( 2002 ) , ` http://www.gap-system.org/packages/sgl.html ` .
e. e. jenkins and a. v. manohar , nucl .
b * 792 * ( 2008 ) 187 [ arxiv:0706.4313 [ hep - ph ] ] .
g. c. branco , l. lavoura and m. n. rebelo , phys .
b * 180 * ( 1986 ) 264 .
e. e. jenkins and a. v. manohar , jhep * 0910 * ( 2009 ) 094 [ arxiv:0907.4763 [ hep - ph ] ] .
j. f. nieves and p. b. pal , phys .
d * 36 * ( 1987 ) 315 ; f. del aguila and m. zralek , nucl .
b * 447 * ( 1995 ) 211 [ hep - ph/9504228 ] ; f. del aguila , j. a. aguilar - saavedra and m. zralek , comput .
commun .
* 100 * ( 1997 ) 231 [ hep - ph/9607311 ] ; j. a. aguilar - saavedra and g. c. branco , phys .
d * 62 * ( 2000 ) 096009 [ hep - ph/0007025 ] ; j. f. nieves and p. b. pal , phys .
d * 64 * ( 2001 ) 076005 [ hep - ph/0105305 ] .
s. m. bilenky and s. t. petcov , rev .
* 59 * ( 1987 ) 671 [ erratum - ibid .
* 61 * ( 1989 ) 169 ] [ erratum - ibid .
* 60 * ( 1988 ) 575 ] ; l. wolfenstein , phys .
b * 107 * ( 1981 ) 77 ; s. m. bilenky , n. p. nedelcheva and s. t. petcov , nucl .
b * 247 * ( 1984 ) 61 ; b. kayser , phys .
d * 30 * ( 1984 ) 1023 . | we consider a scenario with three majorana neutrinos in which a discrete , finite flavour group @xmath0 is combined with a generalized @xmath1 transformation .
we derive conditions for consistently defining such a setup .
we show that in general lepton mixing angles and cp phases ( dirac as well as majorana ) only depend on one single parameter @xmath2 which can take values between @xmath3 and @xmath4 , if the residual symmetries are @xmath5 in the charged lepton and @xmath6 in the neutrino sector .
we perform a comprehensive study for @xmath7 and find five cases which are phenomenologically interesting .
they naturally lead to a non - zero reactor mixing angle and all mixing parameters are strongly correlated .
some of the patterns predict maximal atmospheric mixing and maximal dirac phase , while others predict trivial dirac and majorana phases .
+ 1.5 cm + .3 cm 0.5 cm ferruccio feruglio @xmath8 , claudia hagedorn @xmath9 + and robert ziegler @xmath10 + .7 cm @xmath11 dipartimento di fisica e astronomia ` g. galilei ' , universit di padova , + via marzolo 8 , i-35131 padua , italy .1 cm @xmath12 infn , sezione di padova , via marzolo 8 , i-35131 padua , italy .1 cm @xmath13 sissa , via bonomea 265 , i-34136 trieste , italy .1 cm @xmath14 physik - department , technische universitt mnchen , + james - franck - strasse , d-85748 garching , germany + .1 cm @xmath15 tum institute for advanced study , technische universitt mnchen , + lichtenbergstrasse 2a , d-85748 garching , germany .5 cm _ e - mail : _ , , |
in recent years m33 has been the target of many systematic studies . from searches for variable stars to pin down one of the fundamental distance estimates in the cepheid distance latter ( e.g. macri et al . 2001 ) , to kinematical studies of the stellar populations ( e.g. chandar et al .
2002 ) , to x - ray surveys ( e.g. haberl & pietsch 2001 ) .
however , in spite of m33 being the second closest spiral galaxy after m31 , close enough for the brighter members of its stellar population to be resolved , the properties of its field halo stars have received comparatively little attention .
the first attempt to study the field halo stars in m33 using ccds was that of mould & kristian ( 1986 , hereafter mk86 ) . in this classic paper ,
the authors observed an approximately 5x5 arcmin field ( see fig .
1 ) with the palomar 5 m telescope along with one of the first generation science - grade ccd detectors , a texas instruments 800 x 800 pixel array .
their color - magnitude diagram ( cmd ) , based on aperture photometry of 215 stars transformed onto the cousins system ( cousins 1976a , b ) , extends from above the first ascent red giant branch ( rgb ) tip down to @xmath11.5 mags below the tip ( [email protected] ) . from this diagram , they drew two primary conclusions .
first , by comparing the m33 rgb with those of the galactic globular clusters m92 and 47 tuc , they found the mean metal abundance of the field stars to be @xmath9[m / h]@xmath10 = 2.2 @xmath5 0.8 .
this result was rather surprising given the much higher metallicity ( @xmath9[m / h]@xmath11 ) they found for the halo of m31 .
second , from the bolometric magnitude of the rgb tip and an adopted reddening of @xmath12 , they calculated an absolute distance modulus of 24.8 @xmath5 0.2 for m33 .
cuillandre , lequeux , & lionard ( 1999 ) used the uh8k ccd camera at the prime focus of the cfht 3.6 m to image a 28 x 28 arcmin field in a region of m33 that included the mk86 field .
their @xmath13 cmd reaches as faint as v @xmath1 25.5 .
cuillandre et al .
( 1999 ) adopt an m33 distance modulus of @xmath14 and estimate the line - of - sight reddening ( @xmath15 ) by averaging values based on the 21-cm line ( hartmann 1994 ) and from foreground stars ( johnson & joner 1987 ) .
based on these values , a comparison with the rgb sequences of da costa & armandroff ( 1990 ) yields a mean metal abundance of [ fe / h ] = 1.0 with a metallicity spread from @xmath11.5 to @xmath10.6 dex for the field halo stars in m33 .
this mean abundance is clearly much higher than the value advocated by mk86 and a few tenths of a dex higher than the peak abundance of milky way field halo stars ( e.g. norris 1994 ; carney et al .
1996 ; ) .
most recently , davidge ( 2003 ) has used the gemini north 8 m telescope equipped with gmos to image a field located @xmath19 kpc in projected distance from m33 ( @xmath115.5 kpc deprojected distance ) , approximately twice as far out as the mk86 field .
the g , r , i , z cmds of davidge ( 2003 ) are based on a field of view of only 5x5 arcmin and are not extraordinarily deep , reaching some 2 magnitudes below the tip of the rgb , but they are notable as being the first such m33 data at these extreme galactocentric distances .
davidge s conclusions are twofold .
first , based on the color of the upper rgb , davidge ( 2003 ) determines a mean abundance of [ fe / h ] = @xmath16 ( random ) @xmath17 ( systematic ) for this region of m33 .
second , the presence of a significant number of bright asymptotic giant branch stars suggests that an intermediate - age population exists outside of the ` young star - forming ' disk of m33 .
this latter result is consistent with the work of several investigators ( e.g. sarajedini et al .
1998;2000 , chandar et al .
2002 ) who have suggested that the halo star clusters in m33 are several gyr younger than similar clusters in the milky way s halo . within this context
, the present series of papers attempts to shed further light on the field halo stellar population(s ) of m33 .
this first paper presents deep ground - based photometry of the mk86 field along with an analysis designed to probe the metallicity distribution function of this region of m33 .
we begin with a discussion of the observational material in the next section along with details of the photometric reduction procedure .
section 3 describes our artificial star experiments and how they are used to gauge the photometric completeness and errors .
we compare our photometry to that of mk86 in sec .
the analysis portion of the paper begins in sec . 5 in which we use the luminosity function of the rgb to estimate the distance to m33 .
6 , the radial variations of the stellar populations present in our field are analyzed .
section 7 presents the metallicity distribution function and what it implies for this region of m33 .
finally , our results are summarized in sec .
we obtained deep v and i images of a field centered about @xmath18 southeast of the center of m33 , or about 10 kpc in deprojected distance .
the pointings were centered at @xmath19:35:15 , @xmath20:30:00 j2000 , chosen to sample the halo field population of m33 and to overlap the photometry obtained by mk86 and so facilitate comparison with their work .
figure 1 shows the position of our field ( large square ) and the position of the field studied in mk86 ( small square ) both relative to the center of m33 ( upper right ) .
our data were obtained on the wiyn 3.5 m telescope at kitt peak national observatory on the night of 1999 march 10 ut as part of the ` 2-hour queue . '
the instrument was the s2 kb ccd imager which provided a @xmath21 pixel@xmath22 plate scale .
the ccd is a @xmath23 array and provided a @xmath24 field .
the seeing was generally excellent at @xmath25 allowing us to take full advantage of the fine plate scale .
a log of the observations is provided in table 1 .
figure 2 is a reproduction of one of our v - band ccd images .
the night of 1999 march 10 ut was not perfectly photometric , therefore in order to photometrically calibrate the data , we obtained a set of landolt standards ( landolt 1992 ) and short exposures of the same m33 fields on 2002 february 11 ut .
these calibration data were taken with the wiyn 0.9 m telescope on kitt peak using the same s2 kb ccd imager mounted at the cassegrain focus .
the instrument configuration was identical but on the 0.9 m telescope it has a plate scale of @xmath26 pixel@xmath22 , and a @xmath27 field of view .
seeing was @xmath28 but with the shallower exposures crowding was not a problem .
a log of these observations are also provided in table 1 .
we did all preliminary data reduction on both sets of images using the iraf routines contained in _
ccdproc_. the raw ccd frames were processed to fit and remove the overscan level , trim the overscan region , subtract a zero exposure time bias , and finally to flatten the individual pixel responses with appropriate dome flats .
the final , flattened images were all flat to @xmath29 with the exception of the deep , 3.5 m @xmath30-band images which had fringing at the @xmath31 level . since the peak - to - peak pixel extent of the fringing was much greater than the typical fwhm of point sources in the frame , we made no attempt to correct it .
we measured the aperture photometry of the sa98 standard stars using the iraf routine _ phot_. in each set of exposures of the sa98 field , 35 standards were measured so that the final calibration for the night was determined from 70 measurements .
the night was determined to show no evidence of a @xmath32 color term , to @xmath33 , so we simply fit the equations : @xmath34 where the capital letters denote standard magnitudes , the lower case letters denote instrumental magnitudes , the @xmath35 variables are the respective airmasses , and the numbered subscript coefficients are the fit coefficients .
the values determined for the calibration coefficients along with their uncertainties are given in table 2 .
from these uncertainties and the formal error aperture photometry measurements , the typical zero point uncertainty from these calibration equations is @xmath36 .
after photometrically calibrating the night of 2002 february 11 ut , we calculated a photometric transformation between the shallow m33 exposures taken on that night and the deep m33 exposures taken on the night of 1999 march 10 ut .
we measured the instrumental magnitudes for the deep exposures using the allstar routine from daophot ii ( stetson 1987 ) .
a transformation for each deep exposure was calculated independently due to changing airmass , seeing , and possible transparency changes .
there were typically @xmath37 stars in each transformation .
the coefficients of these transformations are given in table 3 .
note that we found no color term in the @xmath30-band transformations so they are simple zeropoint offsets .
the @xmath38-band did have a @xmath32 color term of @xmath39 .
the @xmath38-band transformations were calculated by solving a least squares fit to the equation : @xmath40 and the @xmath30-band transformations were calculate by finding the mean difference @xmath41 . to exclude saturated stars in the deep 3.5 m data , blends in either data set , mismatches between the sets , and any stars contaminated by cosmic rays ,
the least squares fits for the @xmath38-band data were @xmath42 clipped and iterated until no additional stars were eliminated and the @xmath30-band mean differences were @xmath43 clipped .
the number of stars listed in the last column of table [ trans ] are the number surviving the clipping and used to calculated the calibration .
the typical zero - point uncertainty in the calibration of the frames is @xmath44 mag . after culling the data to remove spurious detections , extragalactic sources , and unreliable detections , ( see discussion below ) the final calibrated data set contains 19,350 stars , measured in both @xmath38 and @xmath30 .
table 4 presents all of these data and fig .
3 presents an @xmath30 versus @xmath32 color magnitude diagram .
the cmd shows a well defined upper giant branch with a first ascent giant branch tip around @xmath45 , and a population of asymptotic giant branch ( agb ) stars above the tip .
the upper giant branch shows inherent scatter larger than that expected from just photometric error , and so there is likely to be an intrinsic dispersion in the age and/or metallicity of the sample .
this is also apparent in the significant number of blue main sequence stars ( i.e. @xmath46-@xmath47 and @xmath48 ) , the presence of stars brighter than the rgb tip , and the appearance of a red clump dominated by core - helium burning stars at [email protected] . in the sections below
we examine these features in detail . although we do not possess a cmd of a field away from m33
, we can simulate the appearance of such a diagram using the galaxy model of robin et al .
( 1996 , and references therein ) . using the latest version of their model available from their web page , ] we have generated colors and magnitudes for stars in the line of sight to our m33 field .
these are shown as the gray points in fig .
4 overplotted on our m33 cmd from fig .
3 . note that the simulated cmd includes the effects of photometric error ( fig .
7 ) but not incompleteness ( fig . 6 ) .
figure 4 shows that some of the scatter seen redward of v
i @xmath1 2 in the range 22@xmath49i@xmath4924 is due to field star contamination .
in addition , although some stars above the first ascent rgb tip likely do not belong to m33 , the majority of the agb population with v
[email protected] probably does .
lastly , we note that essentially all of the blue stars ( v [email protected] ) are likely to belong to m33 .
in order to understand details about the cmd for a field with such a complex mixture of stellar populations , we need to quantitatively determine the photometric accuracy and star detection efficiency of our photometric measurements . to do this we created a set of artificial stars of known magnitude and placed them on each data frame and then measured their photometry using a procedure identical to the procedure used with the real stars .
since image crowding may be a limiting factor , we placed the artificial stars on the images in a regular grid pattern , so that no where on the frames do they significantly increase the image crowding .
we generated the artificial stars to be introduced into the data frames via the following procedure .
first we fit analytic functions to three different regions of the observational cmd as displayed in the left panel of fig .
region one ( r1 ) is the lower asymptotic giant branch which we fit with a second order polynomial in @xmath30 .
region two is the upper giant branch , which we fit with a line in @xmath30 .
region three is dominated by young main sequence stars .
the scatter proved too large for any kind of formal fit ; as a result , we represented it using a line from @xmath51 . in both r1 and
r2 the fits were least squares fits that were @xmath42 clipped with no iterations .
we then used these analytic expressions to populate each section of the cmd .
we extrapolated the r1 fit up to @xmath52 and distributed 540 stars randomly along the fit between @xmath53 .
we only used this relatively small number of stars because completeness is not at question in this magnitude range .
we only needed to test the photometric accuracy .
we extrapolated the r2 fit down to @xmath54 to be sure to populate the artificial stars to magnitudes well below our detection limit .
since this set is to be used for the determination of completeness as well as photometric accuracy , we randomly distributed 9135 stars along the extrapolated fit throughout this entire range .
the line segment representing the stars in r3 was randomly populated with 325 stars . in total
, we generated 10,000 artificial stars .
the 10,000 artificial stars were distributed across each frame in a @xmath55 grid .
the origin of this grid was pixel @xmath56 in the first @xmath38 frame with stars distributed every 19.5 pixels along each axis .
the exact positions of the stars were their respective grid center @xmath57 pixels randomly determined for each star so that the positioning of each star relative to the pixel grid would not be regular .
magnitude offsets between each frame were determined and the coordinate transformation of each subsequent frame onto the system of the first v frame ( v1 ) was calculated ; the input coordinates and magnitudes of the artificial stars were then appropriately offset .
we used the daophot ii routine addstar to calculate the flux and place each artificial star in each of the 7 data frames .
we measured the photometry and processed the photometry lists from the 7 frames containing the artificial stars in a manner identical to the original data frames .
the resulting photometry of the artificial stars is shown by the black points in the right panel of fig .
the solid grey lines in the right panel indicated the input colors and magnitudes , so the scatter in the plot gives a sense of the ultimate photometric accuracy , with the caveat that the vast majority of the points are essentially coincident with the grey lines . to get a complete understanding of our measured photometry
, we need to understand our detection efficiency as a function of magnitude ( i.e. , the completeness ) .
the @xmath30 magnitudes of the artificial data were generated randomly with emphasis on the dimmer part of the magnitude range , and then corresponding @xmath38 magnitudes were calculated to place each star along one of the sequences described above . as a result
the input @xmath30 magnitudes span @xmath58 whereas the @xmath38 magnitudes do not go as bright and only span @xmath59 .
figure 6 is a plot of the completeness as both a function of @xmath38 and @xmath30 in 0.25 mag bins .
the @xmath38 detections are @xmath60 complete down to the 22.7523 mag bin and the @xmath30 detections are @xmath60 complete down to the 21.7522 mag bin .
both bands then have a number of magnitude bins where the completeness is very nearly @xmath60 , and then fall off quickly in the usual fashion .
all detections in each frame , real and artificial stars , were filtered through a set of statistical parameters to prevent non - stellar detections from appearing in our dataset .
in addition to reporting an instrumental magnitude and error for each measurement , allstar also calculates a goodness of fit to the empirical psf of the detection ( @xmath61 ) and a measure of the similarity of the detection to a point source ( @xmath62 ) . to avoid keeping measurements of bad pixels , cosmic ray hits , error peaks in the halos of saturated stars , and extended objects like background galaxies
, we filtered the allstar output lists to remove all objects with @xmath63 mag or @xmath64 at a given magnitude , @xmath65 , and @xmath66 .
finally , both real and artificial stars are kept in the dataset only if they appear on at least 2 of the 3 @xmath38 images and 3 of the 4 i images .
( in the case of the artificial stars , they were also considered found if they fell off the edge of a given frame but were found in all of the frames they did fall upon . ) as a result , the detection probability falls significantly below @xmath60 near @xmath67 and @xmath68 . a comparison of the absolute value of the difference between the input magnitudes and the measured magnitudes and the error estimates reported by allstar shows that the two are in good agreement .
figure 7 illustrates the comparison between the reported errors ( lower panels ) , and the mean of the absolute value of the difference between the input magnitudes and the measured value averaged in bins 0.25 mag in width .
the error bars in the top panels are the scatter in each bin .
the curves are natural logarithm least - squares fits to the reported errors in the lower panels , which are plotted over the binned measured errors in the top panels .
note that these curves are in good agreement with the measured errors .
the only significant deviation is at magnitudes dimmer than @xmath69 and @xmath70 both of which are beyond the point where our completeness of detection begins to fall off rapidly .
to make a direct comparison of our photometric measurements to those of mk86 , we used their finder chart and data tables to find the coordinate transformation between the two data sets .
we then calculated the difference between our measured @xmath38 and @xmath30 photometry and theirs . excluding differences @xmath71 mag which are likely mismatches or extreme variable stars , the @xmath30 band comparison had 197 stars and the @xmath38 band 188 .
we fit equations of the form : @xmath72 where tsb denotes values from our photometry and mk denotes values from mould and kristian .
the resulting fit coefficients are tabulated in table 5 .
the @xmath30 band transformation has a small but definite color term , but the @xmath38 band transformation has a much less significant color term . as an alternative , we also calculated a simple zeropoint offset between the two data sets .
these values are also presented in table 5 .
the @xmath30 band offset is essentially zero , but the scatter in the residuals is larger than the scatter around the fit which includes the color term . the @xmath38 band has a 0.121 magnitude zero point offset , but in spite of its color term being small in the linear fit , the scatter in the residuals of the simple zeropoint offset are also larger than about the linear fit .
whichever transformation is used , the scatter in the residuals is always @xmath73 mag for both bands .
this suggests that the uncertainty in mould and kristian s photometric measurements was a little larger than they estimated , and likely a result of unresolved star light contaminating their aperture photometry .
additionally , the color term in the i band transformation means that there is an offset between our ( v i ) colors and those of mk86 .
some of the conclusions we draw from our data are different than those of mk86 primarily because of these disagreements in the photometry .
we constructed the @xmath30-band luminosity function of our field by selecting stars from the @xmath30 versus @xmath32 cmd ( fig .
we selected all stars brighter than @xmath74 that fell within a band 0.75 mags in width centered on the first ascent giant branch locus .
we also included any stars brighter than @xmath75 which fell redward of the blue edge of this band extrapolated up to @xmath76 .
the resulting luminosity function is displayed in fig .
8 . based on the simulated field
cmd shown in fig .
4 , we estimate that the number of contaminating field stars in this lf is small compared to the number of m33 stars .
this is supported by the findings of mk86 , who reached the same conclusion .
a number of features are apparent from this luminosity function .
the first ascent giant branch is approximated well by a power law .
a power law fit to the bins in the range @xmath77 results in a line with slope , @xmath78 , as indicated by the heavy solid line in the figure .
we restricted the luminosity function to magnitudes brigher than @xmath74 because this is our @xmath7995% completeness limit ( see fig .
6 ) . inspecting the bins at brighter magnitudes ,
the first bin to significantly depart from the power law is the bin centered at 20.625 .
this suggests that the tip of the first ascent giant branch ( trgb ) must fall somewhere between this bin and the 20.875 mag bin so that @xmath80 .
stars brighter than this tip are mostly agb stars in m33 up until @xmath81 at which point the level reaches that expected for forground contamination .
the tip of the red giant branch is the last stage of a star s life before it begins he burning on the zero - age horizontal branch .
many studies have shown that @xmath82 , the @xmath30-band absolute magnitude of the trgb , is remarkably constant over a wide range of metallicities and ages ( da costa & armandroff 1990 ; lee et al .
1993 ; bellazzini et al .
hence , one can estimate @xmath83 , the @xmath30-band apparent magnitude of the trgb , to obtain a direct measurement of the apparent distance modulus for a stellar population . to locate @xmath83 in a systematic and reproducable way , we convolved the lf of stars in the color range @xmath84 with a gaussian - smoothed sobel filter weighted by the poisson noise at each magnitude ( sakai , madore , & freedman 1996 ; sakai et al .
the edge detector is very sensitive to noise in the lf so one must restrict attention to a small range of magnitudes which are likely to include the trgb .
we chose to examine the range @xmath85 mag because it encompasses previous estimates ( mould & kristian 1986 ; lee et al .
1993 ; kim et al .
2002 ) and based on our examination of the lf above .
figure 9 shows the logarithmic lf and edge detector response in this magnitude range .
the highest peak in the filter response occurs at @xmath86 . to estimate this random error we ran a series of bootstrap simulations in a manner similar to that of @xmath87 et al .
the lf was randomly resampled with replacement and each star was perturbed by randomly drawing from a gaussian distribution with standard deviation equal to the star s error .
the edge detector was then applied to the resulting new lf .
we repeated this process 50 times resulting in 50 estimates of the trgb .
we found that these estimates were well fit by a gaussian with a standard deviation of 0.02 mag which we adopt as the random error in our estimate of @xmath83 .
given this value of @xmath83 , we now proceed to determine @xmath82 as follows .
we use eq .
4 of bellazzini et al .
( 2001 ) , @xmath88}^2 + 0.48 { \rm[fe / h ] } + 3.66,\ ] ] to calculate @xmath82 using an initial guess for the peak metallicity of the population .
we then use the resultant distance to construct a metallicity distribution function as described in sec .
. the peak metal abundance of this distribution is input into the above equation as the next guess for the peak [ fe / h ] .
this iterative procedure converges quickly to a solution yielding @xmath89 .
combined with our value for @xmath83 , we find @xmath90 , where the quoted uncertainty includes errors in the measurement of @xmath83 along with random ( 0.02 mag ) and systematic ( 0.03 mag ) errors in the photometry . with our adopted reddening of @xmath91 and @xmath92 ( von hippel & sarajedini 1998 ) , we obtain an intrinsic distance modulus of @xmath4 for m33 .
this value is within the range of distances tabulated by kim et al .
( 2002 ) . however , it is significantly greater than the result of mcconnachie et al .
( 2004 ) , who find @xmath93 also based on the trgb .
the reasons for this difference are unclear at the moment .
adopting the position angle ( 23@xmath94 ) and inclination ( 56@xmath94 ) of m33 derived by regan & vogel ( 1994 ) and the distance modulus based on the trgb , we have computed deprojected radial positions relative to the center of m33 for each star in our sample . using these , fig .
10 shows cmds for our m33 data divided into four radial bins .
the dashed boxes represent the main sequence ( ms ) region ( 19.0@[email protected] , 0.5@xmath95(v
i)@xmath950.5 ) , the asymptotic giant branch ( agb ) region ( 19.7@[email protected] , 1.4@xmath95(v
i)@xmath953.5 ) , and the red giant branch ( rgb ) region ( 20.85@[email protected] , 1.0@xmath95(v
i)@xmath952.15 ) .
the ms is representative of stars with ages younger than @xmath10.5 gyr , the agb is mainly stars between @xmath12 gyr and @xmath18 gyr , and the rgb is dominated by stars older than @xmath18 gyr ( see fig . 12 ) . inspection of fig .
10 suggests that the mean age of the m33 stellar population is increasing as we proceed to larger galactocentric distances in our observed m33 field . in order to quantify this impression , fig .
11 shows the radial behavior of the various ` age tracer ' populations we have identified above .
the filled squares in fig .
11 show the variation in the stellar density of all stars that are above our 95% completeness limit ( v@xmath4925 , i@xmath4924 ) .
the open circles , open squares , and filled circles represent the rgb , agb , and ms regions , respectively . for each region
, we have subtracted off the foreground field - star contamination determined from fig .
4 and then divided by the total stellar density at each radius to account for the overall decrease in the density as r increases . the lines are the weighted least squares fits to each set of points .
we see that the inner regions of our m33 field are , in the mean , younger than the outer regions ; this is most dramatically illustrated by examining the ms region . to estimate the age gradient in our field ,
we proceed as follows .
first , we note that the age gradients implied by the three cmd regions ( ms , agb , rgb ) must be consistent with each other . as such
, we concentrate on the ms region because the gradient is most conspicuous there .
based on the locations of the isochrones in the ms region of the cmd in fig .
12 , we estimate an upper limit of 1 gyr for the total age gradient in our m33 field .
we will investigate this question further in paper ii where we present a synthetic cmd analysis we have performed using the starfish code of harris & zaritsky ( 2001 ) .
it is instructive to examine the distribution of rgb stars as compared with empirical rgb sequences for a set of standard galactic globular clusters ( gc ) .
this approach has historically been used to derive a metallicity distribution function ( mdf ) for stellar populations ( e.g. durrell , harris , & pritchet 2001 ; sarajedini & van duyne 2001 ; harris & harris 2002 ) .
it involves using the dereddened color of rgb stars as a surrogate for metal abundance under the assumption that all stars on the dominant rgb sequence are old ( i.e. @xmath5010 gyr ) . in this section
, we apply this technique to our cmd and investigate the properties of the resultant mdf . the upper panel of fig .
12 shows our m33 cmd along with z=0.004 ( [ fe / h]@xmath10.7 ) theoretical isochrones ( girardi et al .
2002 ) for ages of @xmath96 , @xmath97 , @xmath98 , @xmath99 , and @xmath100 years .
a distance modulus of @xmath101 and a reddening of @xmath12 have been applied to the photometry .
we see that stars with a range of ages are present in our field
. however , the rgb stars appear to be dominated by a population that is of order @xmath100 years old , which means that the majority of the resultant stellar metallicities will not be significantly affected by age . this assertion is supported by preliminary starfish synthetic cmd work we have performed . the full starfish analysis and a complete discussion of the results
will be presented in paper ii of this series .
the lower panel of fig .
12 illustrates our technique for constructing the mdf by showing empirical gc rgbs formulated by saviane et al .
( 2000 ) for @xmath6 $ ] values of 2.0 , 1.5 , 1.0 , and 0.7 dex . for each star with @xmath102@xmath103@xmath104 and 1.0@xmath95@[email protected]
, we can convert dereddened color to metal abundance using this grid of rgbs .
we have chosen this range of color and magnitude to minimize the influence of old agb stars that are bluer than the first ascent rgb .
the result of this exercise is shown in fig .
the filled circles in the upper panel represent the binned histogram of metallicities for the 1123 rgb stars that meet our selection criteria .
the solid line is the generalized histogram of these values constructed by adding up unit gaussians with widths given by the error in each metalliciity .
the generalized mdf has at least two advantages over the binned one .
first , it is not subject to the vagaries associated with binning - the choice of endpoints and the bin sizes .
second , genuine ( and artifactual ) features of the distribution are revealed when the errors are taken into account in its construction .
we see that the mdf exhibits a prominent peak at [ fe / h]@xmath106 , a possible secondary peak at [ fe / h]@xmath107 , and a gradual tail to lower metallicities . to demonstrate the effects of reddening uncertainties on the construction of the mdf , the lower panel of fig .
13 shows the generalized mdf from the upper panel ( solid line ) along with two additional distributions - one constructed using a reddening that is 0.02 mag smaller than our adopted value of @xmath91 ( dashed ) and the other using a value that is 0.02 mag larger ( dotted ) .
we note that the essential features of the distribution discussed above are unchanged .
as noted above , our preliminary starfish results suggest that the ages of the rgb stars are not likely to significantly affect the derived mdf .
we estimate that the metallicity of the peak _ could _ be biased downward by 0.1 to 0.2 dex due to the presence of stars younger than @xmath18 gyr in the rgb sample that produced the mdf .
thus , the true peak of the mdf could be as high as [ fe / h]@xmath108 .
this number is somewhat uncertain , but it does represent a likely upper limit on the effect of age on the mdf peak .
figure 14a plots the metal abundance of each rgb star as a function of its deprojected radial location in kiloparsecs .
a robust iterative 2-@xmath109-rejection least - squares fit ( sarajedini & norris 1994 ) to these data reveals a modest , yet statistically significant , gradient yielding @xmath110 = -0.06 ( \pm 0.01 ) r_{deproj } - 0.45 ( \pm 0.08),\ ] ] as shown by the solid line in fig .
figure 14b illustrates radial abundance gradients derived from other m33 populations compared with our data .
first , we divided our data from fig .
14a into two bins at r=10.5 kpc .
we then fit the dominant peak of the mdf in each bin with a gaussian function .
the resultant peak metallicities are plotted as open circles in fig .
14b . the filled circles in fig .
14b are the m33 disk regions studied by kim et al .
( 2002 , hereafter kklsg ) .
they constructed cmds based on hubble space telescope wide field planetary camera 2 observations of 10 m33 disk regions .
their cmds extend to i@xmath127 and sample the disk stars in m33 .
kklsg followed a procedure similar to our s ( thus allowing an intercomparison of the results ) and used the dereddened color of the rgb to estimate the mean metal abundance in each of their fields .
a least - squares fit to all of the filled circles yields @xmath110 = -0.05 ( \pm 0.01 ) r_{deproj } - 0.55 ( \pm 0.02)\ ] ] shown by the solid line .
if we exclude the two innermost points , where the image crowding may have been problematic , we find @xmath110 = -0.07 ( \pm 0.01 ) r_{deproj } - 0.48 ( \pm 0.04),\ ] ] which is illustrated by the long dashed line . in order to reliably compare these abundance gradient relations with our result
, we need to adjust our data for the difference in distance modulus between the present work [ @xmath101 ] and that of kklsg [ @xmath112 .
having done this , we find that @xmath110 = -0.07 ( \pm 0.01 ) r_{deproj } - 0.49 ( \pm 0.09),\ ] ] which is statistically identical to both of the lines derived by kklsg .
this suggests that the vast majority of stars in our m33 field belong to the disk population of that galaxy , _ not _ the halo as previously thought ( e.g. mk86 ) .
if this assertion is borne out , then the m33 stellar disk would extend out to @xmath112 kpc . at this point
, it is important to note that the slight radial age variation determined in sec .
6 would have a negligible effect on our metallicity gradient ( fig .
the girardi et al .
( 2002 ) isochrones predict a change in the implied metallicity of @xmath10.05 dex / gyr due to the effect of age on the colors of rgb stars at @xmath103=3.5 . given our upper limit of 1 gyr for the total age gradient in our field , this implies a differential correction of @xmath50.025 dex from the inner regions to the outer regions .
the resultant corrected " slope would be @xmath7[fe / h]/@xmath7@xmath8 = 0.07 , which is still consistent with the equations given above . for reference , the dotted line in fig .
14b shows the radial abundance variation of the hii regions as derived by kim et al .
( 2002 , see their fig . 5 ) .
lastly , the open squares in fig .
14b are the nine m33 halo globular clusters from sarajedini et al .
they exhibit no radial abundance gradient .
this is one signature of a chaotic fragmentation type formation scenario similar to galactic globular clusters outside 8 kpc from the galactic center , which also show no radial abundance gradient ( searle & zinn 1978 ) . as pointed out by sarajedini et al .
( 1998,2000 ) , the m33 halo clusters suffer from the second parameter effect again analogous to the outer galactic halo globulars .
if our m33 field was dominated by halo stars , then there would be no significant radial abundance gradient among the stars just like the halo clusters .
in this work , we present deep @xmath0 photometry of a field in m33 , that includes the region studied by mould & kristian(1986 ) , observed with the wiyn 3.5 m telescope .
based on the cmd , which reaches past the helium burning red clump , we draw the following conclusions .
\1 ) the i - band apparent magnitude of the trgb is measured to be @xmath113 , where the error includes the measurement error ( 0.02 mag ) along with random ( 0.02 mag ) and systematic ( 0.03 mag ) errors in the photometry .
\2 ) the i - band apparent magnitude of the trgb implies a distance modulus of @xmath2 , which combined with an adopted reddening of @xmath3 yields an absolute modulus of @xmath4 ( 867@xmath528 kpc ) for m33 .
\3 ) by examining the spatial properties of stars in various parts of the cmd , we find that , over the range of deprojected radii covered by our field ( @xmath18.5 to @xmath112.5 kpc ) , there is a significant age gradient with an upper limit of @xmath11 gyr ( i.e. , @xmath10.25 gyr / kpc ) .
\4 ) using the empirical rgb grid constructed by saviane et al .
( 2000 ) , we have converted the dereddened color of each rgb star to a metallicity .
the resultant metallicity distribution function ( mdf ) displays a primary peak at a metallicity of @xmath6$]@xmath11.0 with a tail to lower abundances .
the peak shows a radial variation with a slope of @xmath7[fe / h]/@xmath7@xmath8 = 0.06 @xmath5 0.01 dex / kpc .
this gradient is consistent with the variation seen in the inner _ disk _ regions of m33 .
therefore , we conclude that the vast majority of stars in this field belong to the disk of m33 , not the halo as previously thought .
llcrrr m33 & 1999 - 03 - 10 & v & 800 & 01:35:15 & 30:30:00 + m33 & & v & 800 & 01:35:15 & 30:30:12 + m33 & & v & 800 & 01:35:14 & 30:30:23 + m33 & & i & 1100 & 01:35:14 & 30:30:23 + m33 & & i & 1100 & 01:35:15 & 30:30:12 + m33 & & i & 1100 & 01:35:15 & 30:30:00 + m33 & & i & 1100 & 01:35:14 & 30:30:12 + sa98 & 2002 - 02 - 11 & v & 60 & 06:52:13 & -00:19:02 + sa98 & & v & 60 & 06:52:12 & -00:19:16 + sa98 & & i & 50 & 06:52:13 & -00:19:02 + sa98 & & i & 50 & 06:52:12 & -00:19:16 + m33 & 2002 - 02 - 11 & v & 600 & 01:36:00 & 30:27:51 + m33 & & v & 600 & 01:35:57 & 30:26:20 + m33 & & v & 600 & 01:36:07 & 30:28:20 + m33 & & i & 300 & 01:35:58 & 30:26:20 + m33 & & i & 300 & 01:36:02 & 30:27:20 + m33 & & i & 300 & 01:36:06 & 30:28:20 + lccc v1 & @xmath115 & @xmath116 & 26 + v2 & @xmath117 & @xmath118 & 22 + v3 & @xmath119 & @xmath120 & 27 + i1 & @xmath121 & & 24 + i2 & @xmath122 & & 26 + i3 & @xmath123 & & 29 + i4 & @xmath124 & & 24 + rllrrrrr 1 & 01:35:02.47 & 30:31:14.8 & 16.587 & 0.005 & 0.594 & 0.006 + 2 & 01:35:15.53 & 30:29:39.2 & 16.739 & 0.012 & 0.903 & 0.013 + 3 & 01:35:25.45 & 30:32:28.8 & 16.838 & 0.016 & 2.118 & 0.018 + 4 & 01:35:03.77 & 30:30:04.7 & 17.107 & 0.008 & 1.888 & 0.012 + 5 & 01:35:14.13 & 30:31:50.6 & 17.298 & 0.005 & 2.489 & 0.007 + | we present deep ccd photometry in the @xmath0 passbands using the wiyn 3.5 m telescope of a field located approximately 20 southeast of the center of m33 ; this field includes the region studied by mould & kristian in their 1986 paper .
the color - magnitude diagram ( cmd ) extends to i@xmath125 and shows a prominent red giant branch ( rgb ) , along with significant numbers of asymptotic giant branch and young main sequence stars .
the red clump of core helium burning stars is also discernable near the limit of our cmd .
the i - band apparent magnitude of the red giant branch tip implies a distance modulus of @xmath2 , which combined with an adopted reddening of @xmath3 yields an absolute modulus of @xmath4 ( 867@xmath528 kpc ) for m33 . over the range of deprojected radii covered by our field ( @xmath18.5 to @xmath112.5 kpc )
, we find a significant age gradient with an upper limit of @xmath11 gyr ( @xmath10.25 gyr / kpc ) .
comparison of the rgb photometry to empirical giant branch sequences for galactic globulars allows us to use the dereddened color of these stars to construct a metallicity distribution function ( mdf ) .
the primary peak in the mdf is at a metallicity of @xmath6$]@xmath11.0 with a tail to lower abundances .
the peak does show radial variation with a slope of @xmath7[fe / h]/@xmath7@xmath8 = 0.06 @xmath5 0.01 dex / kpc .
this gradient is consistent with the variation seen in the inner _ disk _ regions of m33 . as such
, we conclude that the vast majority of stars in this field belong to the disk of m33 , not the halo as previously thought . |
the search for higgs bosons is among the prime tasks of the cern large hadron collider ( lhc ) @xcite .
while the standard model ( sm ) of elementary - particle physics contains one complex higgs doublet , from which one neutral @xmath0-even higgs boson @xmath2 emerges in the physical particle spectrum after the spontaneous breakdown of the electroweak symmetry , the higgs sector of the minimal supersymmetric extension of the sm ( mssm ) consists of a two - higgs - doublet model ( 2hdm ) and accommodates a quintet of physical higgs bosons : the neutral @xmath0-even @xmath3 and @xmath4 bosons , the neutral @xmath0-odd @xmath5 boson , and the charged @xmath6-boson pair . at the tree level , the mssm higgs sector has two free parameters , which are usually taken to be the mass @xmath7 of the @xmath5 boson and the ratio @xmath8 of the vacuum expectation values of the two higgs doublets . in the following ,
we focus our attention on the @xmath3 , @xmath4 , and @xmath5 bosons , which we collectively denote by @xmath9 .
a recent discussion of @xmath6-boson production at the lhc may be found in refs .
@xcite and the references cited therein .
the dominant source of @xmath9 bosons is their single production by @xmath10 fusion , @xmath11 , which is mediated by heavy - quark @xcite and squark @xcite loops .
another , less important mechanism of single @xmath9-boson production is @xmath12 @xcite .
the @xmath9 bosons thus produced have essentially zero transverse momentum ( @xmath13 ) . in order for the @xmath9 bosons to obtain finite @xmath13
, they need to be produced in association with one or more other particles or jets ( @xmath14 ) . in leading order ( lo ) , @xmath15 associated production proceeds through the partonic subprocesses @xmath16 , @xmath17 , and @xmath18 , which again involve heavy - quark @xcite and squark @xcite loops .
alternatively , the @xmath9 bosons can be produced , with interesting rates , in association with ( i ) a dijet via intermediate - boson fusion , @xmath19 , where @xmath20 and @xmath21 stand for any light flavor of quark or antiquark , @xmath22 , and virtual particles are marked by an asterisk @xcite ; ( ii ) a quark - antiquark pair of heavy flavor @xmath23 via @xmath24 @xcite ; ( iii ) an intermediate boson via @xmath25 @xcite , @xmath26 @xcite , and @xmath27 @xcite ; or ( iv ) another , possibly different @xmath9 boson via @xmath28 @xcite and @xmath29 @xcite .
note that , due to the absence of @xmath30 couplings at the tree level , @xmath19 and the drell - yan processes @xmath31 and @xmath32 are not possible for @xmath33 .
the partonic suprocesses @xmath27 and @xmath29 are mediated by heavy - quark @xcite and squark loops @xcite .
comprehensive reviews of quantum corrections to higgs - boson production within the sm and mssm may be found in refs .
@xcite , respectively . in this paper
, we revisit @xmath34 associated hadroproduction via gluon fusion in the mssm . in the sm case , mutual agreement between three independent calculations @xcite has been established , and compact formulae for the partonic cross section are available @xcite . in the mssm , the status is much less advanced and , perhaps , somewhat unsatisfactory . as for @xmath27 with @xmath35
, there exists only one analysis so far @xcite , which has not yet been verified by other authors .
the analytic expressions presented in ref .
@xcite are rather complicated ; they involve 14 form factors . in order to translate the sm results @xcite to the case of @xmath27 with @xmath35 in the mssm ,
it is not sufficient to adjust the @xmath36 and @xmath37 couplings ; it is also necessary to include certain quark triangle diagrams with an @xmath5 boson in the @xmath38 channel [ see fig . [
fig : tree](a ) ] . on the other hand ,
the squark loop contributions to these two mssm processes vanish @xcite for reasons explained below .
as for @xmath39 , the quark loop contributions were first studied on the basis of a numerical evaluation @xcite , which was recently employed for phenomenological signal - versus - background analyses taking into account the subsequent @xmath40 and @xmath41 decays @xcite .
an independent analysis , including also the squark loops , was reported in ref .
@xcite , which does not contain an analytic expression for the partonic cross section either .
unfortunately , comparisons between refs .
@xcite are not discussed in these papers .
recently , the helicity amplitudes of @xmath27 with @xmath42 were analyzed in ref .
@xcite with regard to the asymptotic helicity conservation property of supersymmetry , and their real and imaginary parts were graphically presented at a specific scattering angle as functions of the center - of - mass ( c.m . )
energy . with the higgs hunt at the lhc being in full swing
, it is an urgent matter to consolidate our knowledge of @xmath34 associated hadroproduction via gluon fusion in the mssm , which is the motivation of this paper .
specifically , we present compact analytic expressions , also for the partonic cross sections of @xmath43 @xcite , and perform a detailed numerical analysis using up - to - date input .
the importance of @xmath44-initiated subprocesses for higgs - boson production has been variously emphasized in the literature , in particular , in connection with the final states @xmath9 @xcite , @xmath45 @xcite , @xmath46 @xcite , and @xmath47 @xcite .
these subprocesses receive contributions from feynman diagrams involving @xmath48-quark yukawa couplings , which are generally strong for large values of @xmath49 .
( the @xmath50 and @xmath51 couplings are also strong for small values of @xmath49 . )
if the two final - state particles couple to a @xmath1 boson ( or photon ) , as is the case for the final states @xmath52 , @xmath53 , @xmath54 , @xmath55 , and @xmath46 , then there are additional contributions from drell - yan - type diagrams , which are already present for the light flavors @xmath56 . however , diagrams of the latter type are absent for the final states @xmath57 , @xmath58 , @xmath59 , @xmath60 , @xmath61 , and @xmath47 , which can still be produced through @xmath44 annihilation . as for @xmath44 annihilation
, it should be noted that the treatment of bottom as an active flavor inside the colliding hadrons leads to an effective description , which comprises contributions from the higher - order subprocesses @xmath62 , @xmath63 , and @xmath64 .
if all these subprocesses are to be explicitly included along with @xmath65 , then it is necessary to employ a judiciously subtracted @xmath48-quark pdf in order to avoid double counting @xcite .
the evaluation of @xmath65 with an unsubtracted @xmath48-quark pdf is expected to slightly overestimate the true cross section @xcite . for simplicity
, we shall nevertheless adopt this effective approach in our analysis , keeping in mind that a qcd - correction factor below unity is to be applied .
in fact , such a behavior has recently been observed for @xmath66 @xcite . in order to reduce the number of unknown supersymmetric input parameters ,
we adopt a scenario where the mssm is embedded in a grand unified theory ( gut ) involving supergravity ( sugra ) @xcite .
the mssm thus constrained is characterized by the following parameters at the gut scale , which come in addition to @xmath49 and @xmath7 : the universal scalar mass @xmath67 , the universal gaugino mass @xmath68 , the trilinear higgs - sfermion coupling @xmath69 , the bilinear higgs coupling @xmath70 , and the higgs - higgsino mass parameter @xmath71 .
notice that @xmath7 is then not an independent parameter anymore , but it is fixed through the renormalization group equation .
the number of parameters can be further reduced by making additional assumptions .
unification of the @xmath72-lepton and @xmath48-quark yukawa couplings at the gut scale leads to a correlation between @xmath73 and @xmath49 . furthermore ,
if the electroweak symmetry is broken radiatively , then @xmath70 and @xmath71 are determined up to the sign of @xmath71 .
finally , it turns out that the mssm parameters are nearly independent of the value of @xmath69 , as long as @xmath74 gev at the gut scale .
this paper is organized as follows . in sec .
[ sec : two ] , we list the lo cross sections of @xmath26 , including the yukawa - enhanced contributions for @xmath75 , and those of @xmath27 , including both quark and squark loop contributions , in the mssm .
the relevant quark and squark loop form factors are relegated to appendix [ sec : b ] . in sec .
[ sec : three ] , we present quantitative predictions for the inclusive cross sections of @xmath76 at the lhc adopting a favorable sugra - inspired mssm scenario .
[ sec : four ] contains our conclusions . for the reader s
convenience , the relevant feynman rules are summarized in appendix [ sec : a ] .
in this section , we present the lo cross sections of the partonic subprocesses @xmath26 and @xmath27 , where @xmath42 , in the mssm . we work in the parton model of qcd with @xmath77 active quark flavors @xmath78 , which we take to be massless .
however , we retain the @xmath48-quark yukawa couplings at their finite values , in order not to suppress possibly sizable contributions .
we adopt the mssm feynman rules from ref .
the couplings of the @xmath1 and @xmath9 bosons to quarks , @xmath79 , @xmath80 , and @xmath81 , are given in eq .
( 5 ) of ref .
@xcite and eq .
( a3 ) of ref .
@xcite , respectively . as for the @xmath82 couplings , @xmath83 and @xmath84
are given by eq .
( [ eq : hzz ] ) in appendix [ sec : a ] , while the @xmath85 coupling vanishes at tree level .
the @xmath86 and @xmath87 couplings , @xmath88 and @xmath89 , may be found in eq .
( a2 ) of ref .
@xcite . for each quark flavor @xmath20
there is a corresponding squark flavor @xmath90 , which comes in two mass eigenstates @xmath91 .
the masses @xmath92 of the squarks and their trilinear couplings to the @xmath9 bosons , @xmath93 , are listed in eqs .
( a.5 ) , ( a.7 ) , and ( a.8 ) and in table 1 of ref .
@xcite and @xmath93 are called @xmath94 and @xmath95 , respectively . ] , eq .
( a.2 ) of ref .
@xcite , and eq . ( a4 ) of ref .
@xcite , respectively . considering the generic partonic subprocess @xmath96
, we denote the four - momenta of the incoming partons , @xmath97 and @xmath48 , and the outgoing @xmath1 and @xmath9 bosons by @xmath98 , @xmath99 , @xmath100 , and @xmath101 , respectively , and define the partonic mandelstam variables as @xmath102 , @xmath103 , and @xmath104 .
the on - shell conditions read @xmath105 , @xmath106 , and @xmath107 .
four - momentum conservation implies that @xmath108 .
furthermore , we have @xmath109 , where @xmath13 is the absolute value of transverse momentum common to the @xmath1 and @xmath9 bosons in the c.m . frame .
the tree - level diagrams for @xmath65 with @xmath35 and @xmath33 are depicted in fig .
[ fig : tree](a ) and ( b ) , respectively .
as already mentioned above , there is no drell - yan diagram in fig .
[ fig : tree](b ) because of the absence of a @xmath85 coupling at the tree level .
the differential cross sections for the first class of partonic subprocesses may be generically written as @xmath110 , \label{eq : bbzh}\end{aligned}\ ] ] where @xmath111 is fermi s constant , @xmath112 is the cosine of the weak mixing angle , @xmath113 , and @xmath114.\end{aligned}\ ] ] here , @xmath115 is the propagator function of particle @xmath116 , with mass @xmath117 and total decay width @xmath118 . for the second class of partonic subprocesses , we have @xmath119 , \label{eq : bbza}\end{aligned}\ ] ] where @xmath120 as for @xmath54 and @xmath55 production , there are also sizable contributions from @xmath121 annihilation via a virtual @xmath1 boson for the quarks of the first and second generations , @xmath56 , whose yukawa couplings are negligibly small .
the corresponding drell - yan cross sections are obtained from eq .
( [ eq : bbzh ] ) by putting @xmath122 and substituting @xmath123
. the resulting expression agrees with eq .
( 2.8 ) of ref .
@xcite , appropriate for @xmath124 in the sm , after adjusting the @xmath36 coupling .
the full tree - level cross sections are then obtained by complementing the @xmath44-initiated cross sections of eq .
( [ eq : bbzh ] ) with the drell - yan cross sections for @xmath56 .
the non - vanishing one - loop diagrams pertinent to @xmath27 , with @xmath35 and @xmath33 are depicted in figs .
[ fig : loop](a ) and ( b ) , respectively . as already mentioned in the introduction , the presence of the quark triangle diagrams involving an @xmath38-channel @xmath5-boson exchange in fig .
[ fig : loop](a ) represents a qualitatively new feature of the mssm as compared to the sm .
furthermore , similarly to fig .
[ fig : tree](b ) , quark triangle diagrams with an @xmath38-channel @xmath1-boson exchange do not appear in fig . [
fig : loop](b ) . in the following ,
we refer to a squark loop diagram involving an @xmath38-channel propagator as a triangle diagram .
the residual squark loop diagrams are regarded to be of box type .
the squark triangle and box diagrams for @xmath27 with @xmath35 vanish , and so do the squark box diagrams for @xmath39 .
this may be understood as follows .
( i ) the @xmath125 , @xmath126 , @xmath127 , and @xmath128 couplings are symmetric in @xmath129 and @xmath14 , while the @xmath130 coupling is antisymmetric @xcite .
thus , squark loops connecting gluons and @xmath1 bosons with an odd number of @xmath5 bosons vanish upon summation over @xmath129 and @xmath14 .
( ii ) the @xmath125 and @xmath128 couplings are linear in the squark four - momenta , while the @xmath126 , @xmath127 , and @xmath93 couplings are momentum independent @xcite .
thus , a squark loop connecting gluons , @xmath1 bosons , and @xmath9 bosons vanishes upon adding its counterpart with the loop - momentum flows reversed if the total number of gluons and @xmath1 bosons is odd . as in refs .
@xcite , we express the quark and squark loop contributions in terms of helicity amplitudes .
we label the helicity states of the two gluons and the @xmath1 boson in the partonic c.m .
frame by @xmath131 , @xmath132 , and @xmath133 .
we first consider @xmath27 with @xmath35 .
the helicity amplitudes of the quark triangle contribution read @xmath134 .
\label{eq : trih}\end{aligned}\ ] ] the quark triangle form factor , @xmath135 , is given in eq .
( [ eq : fht ] ) . as for the quark box contribution ,
all twelve helicity combinations contribute . due to bose symmetry ,
they are related by @xmath136 keeping @xmath137 generic , we thus only need to specify four expressions .
these read @xmath138 , \nonumber\\ { \mathcal m}_{+-0}^\box&= & \frac{8}{\sqrt{z\lambda}}\sum_qg_{\phi qq}a_{zqq}m_q \left[f_{+-}^0-(t\leftrightarrow u)\right ] , \nonumber\\ { \mathcal m}_{++\lambda_z}^\box&= & -4\sqrt{\frac{2n}{s}}\sum_qg_{\phi qq}a_{zqq}m_q \left[f_{++}^1-(t\leftrightarrow u)\right ] , \nonumber\\ { \mathcal m}_{+-\lambda_z}^\box&= & -4\sqrt{\frac{2n}{s}}\sum_qg_{\phi qq}a_{zqq}m_q \left[f_{+-}^1+(t\leftrightarrow u,\lambda_z\leftrightarrow -\lambda_z)\right ] .
\label{eq : boxh}\end{aligned}\ ] ] the quark box form factors , @xmath139 , are listed in eq .
( [ eq : fhb ] ) . for the reasons explained above
, we have @xmath140 for the squark - induced helicity amplitudes .
we now turn to @xmath39 .
the helicity amplitudes of the quark and squark triangle contributions read @xmath141 respectively .
the quark and squark triangle form factors , @xmath135 and @xmath142 , may be found in eq .
( [ eq : fat ] ) .
again , the helicity amplitudes of the quark box contribution satisfy the bose symmetry relations of eq .
( [ eq : bos ] ) .
we find @xmath143 , \nonumber\\ { \mathcal m}_{+-0}^\box&= & -\frac{8}{\sqrt{z\lambda}}\sum_qg_{a^0qq}a_{zqq}m_q \left[f_{+-}^0+(t\leftrightarrow u)\right ] , \nonumber\\ { \mathcal m}_{++\lambda_z}^\box&= & -4\sqrt{\frac{2n}{s}}\sum_qg_{a^0qq}a_{zqq}m_q \left[f_{++}^1-(t\leftrightarrow u)\right ] , \nonumber\\ { \mathcal m}_{+-\lambda_z}^\box&= & -4\sqrt{\frac{2n}{s}}\sum_qg_{a^0qq}a_{zqq}m_q \left[f_{+-}^1-(t\leftrightarrow u,\lambda_z\to-\lambda_z)\right].\end{aligned}\ ] ] the quark box form factors , @xmath139 , are presented in eq .
( [ eq : fab ] ) .
we recall that @xmath144 .
the differential cross section of @xmath27 is then given by @xmath145 where @xmath146 is the strong - coupling constant at renormalization scale @xmath147 .
due to bose symmetry , the right - hand side of eq .
( [ eq : xs ] ) is symmetric in @xmath148 and @xmath149 .
the differential cross section of @xmath150 in the sm is obtained from eqs .
( [ eq : trih])([eq : boxh ] ) and ( [ eq : xs ] ) , with @xmath151 , by replacing @xmath152 , adjusting the @xmath153 and @xmath154 couplings , and discarding the contribution due to @xmath5-boson exchange . in this way
, we recover the result of ref .
@xcite , which is expressed in terms of lorentz - invariant form factors rather than helicity amplitudes .
the kinematics of the inclusive reaction @xmath155 , where @xmath69 and @xmath70 are colliding hadrons , is described in sec .
ii of ref .
its double - differential cross section @xmath156 , where @xmath157 and @xmath13 are the rapidity and transverse momentum of the @xmath1 boson in the c.m.system of the hadronic collision , may be evaluated from eq .
( 2.1 ) of ref .
we are now in a position to explore the phenomenological implications of our results .
the sm input parameters for our numerical analysis are taken to be @xmath158 gev@xmath159 , @xmath160 gev , @xmath161 gev , @xmath162 gev , and @xmath163 gev @xcite .
we adopt the lo proton pdf set cteq6l1 @xcite .
we evaluate @xmath146 and @xmath164 from the lo formulas , which may be found , _
e.g. _ , in eqs .
( 23 ) and ( 24 ) of ref .
@xcite , respectively , with @xmath77 quark flavors and asymptotic scale parameter @xmath165 mev @xcite .
we identify the renormalization and factorization scales with the @xmath34 invariant mass @xmath166 .
we vary @xmath49 and @xmath7 in the ranges @xmath167 and 180 gev@xmath168 tev , respectively . as for the gut parameters , we choose @xmath169 gev , @xmath170 , and @xmath171 , and tune @xmath67
so as to be consistent with the desired value of @xmath7 .
all other mssm parameters are then determined according to the sugra - inspired scenario as implemented in the program package suspect @xcite .
we do not impose the unification of the @xmath72-lepton and @xmath48-quark yukawa couplings at the gut scale , which would just constrain the allowed @xmath49 range without any visible effect on the results for these values of @xmath49 .
we exclude solutions which do not comply with the present experimental lower mass bounds of the sfermions , charginos , neutralinos , and higgs bosons @xcite .
we now study the fully integrated cross sections of @xmath76 at the lhc , with c.m .
energy @xmath172 tev .
figures [ fig : zh][fig : za ] refer to the cases @xmath42 , respectively . in part ( a ) of each figure
, the @xmath173 dependence is studied for @xmath174 and 30 while , in part ( b ) , the @xmath49 dependence is studied for @xmath175 and 600 gev .
we note that the sugra - inspired mssm with our choice of input parameters does not permit @xmath49 and @xmath7 to be simultaneously small , due to the experimental lower bound on @xmath176 @xcite .
this explains why the curves for @xmath174 in figs .
[ fig : zh][fig : za](a ) only start at @xmath177 gev , while those for @xmath178 already start at @xmath179 gev . in figs .
[ fig : zh ] and [ fig : zh ] , which refer to @xmath35 , respectively , the total @xmath121-annihilation contributions ( dashed lines ) , corresponding to the coherent superposition of drell - yan and yukawa - enhanced amplitudes , and the @xmath10-fusion contributions ( solid lines ) , which arise only from quark loops , are presented separately . for a comparison with future experimental data , they should be added . for comparison ,
also the pure drell - yan contributions ( dotted lines ) are shown .
as for @xmath151 , we observe from fig . [
fig : zh ] that the contribution due to @xmath121 annihilation is almost exhausted by the drell - yan process and greatly exceeds the one due to @xmath10 fusion , by a factor of 35 .
the @xmath121-annihilation contribution falls off by a factor of two as @xmath176 runs from 82 gev to 115 gev and feebly depends on @xmath49 , except for the appreciable rise towards the lower edge of the considered @xmath49 range .
the @xmath10-fusion contribution feebly depends on @xmath176 , @xmath7 , and @xmath49 .
the situation is very different for @xmath180 , as is obvious from fig .
[ fig : zh ] . here , @xmath181 annihilation is generally far more important than the drell - yan process , except for @xmath175 gev and @xmath174 , where the latter gets close .
the contribution due to @xmath181 annihilation monotonically increases with @xmath49 , while the one due to the drell - yan process decreases .
furthermore , @xmath10 fusion competes with @xmath121 annihilation and even dominates for @xmath182 . as for @xmath33 , the @xmath44-annihilation contribution ( dashed lines ) and the total @xmath10-fusion contribution ( solid lines ) , corresponding to the coherent superposition of quark and squark loop amplitudes , are presented separately in fig .
[ fig : za ] . for comparison , also
the @xmath10-fusion contribution due to quark loops only ( dotted lines ) is shown .
as in the case of @xmath180 , @xmath10 fusion competes with @xmath44 annihilation and even dominates for @xmath182 .
again , the @xmath181-annihilation contribution monotonically increases with @xmath49 . the bulk of the @xmath10-fusion contribution is due to the quark loops , especially at low values of @xmath7 . finally , we compare our results with the literature . as already mentioned in sec .
[ sec : two ] , we recover the well - known sm result @xcite , for @xmath183 , by taking the sm limit of our results for @xmath151 in eqs .
( [ eq : trih ] ) and ( [ eq : boxh ] ) .
the contribution due to @xmath5-boson exchange in eq .
( [ eq : trih ] ) , which is not probed in the sm limit , agrees with the analogous contribution to @xmath184 given in eq .
( 1 ) of ref .
@xcite after appropriately adjusting the masses and couplings . on the other hand , the residual terms in the latter equation , which arise from the exchanges of @xmath3 and @xmath4 bosons ,
coincide with eq .
( [ eq : qtria ] ) after substituting the appropriate masses and couplings .
similarly , by adjusting masses and couplings in eq .
( [ eq : stria ] ) , we reproduce eq .
( 2.3 ) in ref .
@xcite , which gives the squark triangle contribution to @xmath184 . in ref .
@xcite , numerical results for the cross section of @xmath185 via quark - loop - mediated @xmath10 fusion were presented . adopting the input parameters and proton pdf set specified in that reference
, we nicely reproduce the separate contributions due triangle and box diagrams shown in fig . 4 therein , while we fail to agree with their superposition .
furthermore , we find reasonable agreement with the cross section of @xmath186 via @xmath10 fusion represented graphically for different scenarios in figs . 6 and 7 of ref .
@xcite adopting the respective inputs from there .
we analytically calculated the cross sections of the partonic subprocesses @xmath26 and @xmath27 , where @xmath42 , to lo in the mssm .
we included the drell - yan and yukawa - enhanced contributions to @xmath121 annihilation ( see fig .
[ fig : tree ] ) and the quark and squark loop contributions to @xmath10 fusion ( see fig . [
fig : loop ] ) .
we presented these results as helicity amplitudes expressed in terms of standard scalar one - loop integrals .
we then quantitatively investigated the inclusive cross sections of @xmath76 at the lhc with @xmath187 gev adopting a favorable sugra - inspired mssm scenario , varying the input parameters @xmath7 and @xmath49 .
our results are presented in figs .
[ fig : zh][fig : za ] .
the total cross section for @xmath151 is typically of order 1 pb , while those for @xmath188 are of order 100 fb ( 10 fb ) for @xmath175 gev ( 600 gev ) . assuming design luminosity , @xmath189 @xmath190s@xmath191
, a cross section of 1 pb corresponds to @xmath192 events per year and experiment at the lhc ( see table i of ref .
we thank a. a. barrientos bendezu and r. ziegler for their collaboration at the initial stage of this work . the work of b.a.k .
was supported in part by the german federal ministry for education and research bmbf through grant no .
05 ht6gua , by the german research foundation dfg through the collaborative research centre no .
676 _ particles , strings and the early universe the structure of matter and space time _ , and by the helmholtz association hgf through the helmholtz alliance ha 101 _ physics at the terascale_. the work of c.p.p . was supported in part by the german academic exchange service ( daad ) reinvitation programme under reference code a/07/02820 and by the office of the vice president for academic affairs of the university of the philippines .
in this appendix , we collect the feynman rules used in this paper .
the feynman rules for the @xmath193 vertices are @xmath194 , with @xmath195 , @xmath196 being the proton charge , @xmath197 , and @xmath198 where @xmath199 and @xmath200 are the weak hypercharge and electric charge of quark @xmath20 , respectively .
the feynman rules for the @xmath201 ( @xmath35 ) and @xmath202 vertices are @xmath203 and @xmath204 , respectively , with @xmath205 where @xmath206 is the mixing angle that rotates the weak @xmath0-even higgs eigenstates into the mass eigenstates @xmath3 and @xmath4 .
the feynman rules for the @xmath82 vertices are @xmath207 , with @xmath208 the feynman rules for the @xmath209 vertices are @xmath210 , where @xmath211 is the incoming four - momentum of the @xmath9 boson , @xmath212 is the outgoing four - momentum of the @xmath5 boson , and @xmath213 the feynman rules for the @xmath214 vertices are @xmath215 , with @xmath216 where @xmath217 are the squark mixing matrices , with @xmath218 being the squark mixing angles .
in this appendix , we express the quark and squark triangle and box form factors , @xmath135 , @xmath142 , and @xmath139 , for @xmath35 and @xmath33 , in terms of the standard scalar three- and four - point functions , which we abbreviate as @xmath219 and @xmath220 , respectively .
the definitions of the latter may be found in eq .
( 5 ) of ref .
@xcite .
the quark box form factors for @xmath35 read @xmath222 \nonumber\\ & & { } -\left[n\left(t+u+\frac{\lambda}{s}\right)+2m_q^2\lambda\right ] d^{h0z0}_{qqqq}(t , u)-4\left(szh+m_q^2\lambda\right)d^{hz00}_{qqqq}(s , t ) , \nonumber\\ f_{+-}^0&=&\frac{(h - z - s)}{n}(t - u)\left[s(t+u)c^{00}_{qqq}(s ) -\lambda c^{hz}_{qqq}(s)-2m_q^2nd^{h0z0}_{qqqq}(t , u)\right ] \nonumber\\ & & { } + 2(t+u)(t - z)\left[1+\frac{t(t - u)(h - z - s)}{n(t+u)}\right]c^{z0}_{qqq}(t ) \nonumber\\ & & { } + \frac{2(t - h)}{n}\left[z(u^2-t^2-\lambda)+(t+u)(t^2-zh)\right]c^{h0}_{qqq}(t ) \nonumber\\ & & { } -(h - z - s)\left[2st\frac{t^2-zh}{n}+4m_q^2(t - u)\right]d^{hz00}_{qqqq}(s , t ) , \nonumber\\ f_{++}^1&=&(t - u)\left[\frac{z - h - s}{\sqrt\lambda}-\lambda_z\right ] \left[\frac{s}{n}c^{00}_{qqq}(s)-\frac{1}{2}d^{h0z0}_{qqqq}(t , u)- \frac{s}{n}\left(t+\frac{2n}{t - u}\right)d^{hz00}_{qqqq}(s , t)\right ] \nonumber\\ & & { } + \frac{2(h - u)}{\sqrt\lambda n}\left(\lambda_z\sqrt\lambda+t - u+\frac{2n}{h - u}\right ) \left[(h - t)c^{h0}_{qqq}(t)+(z - u)c^{z0}_{qqq}(u)\right ] , \nonumber\\ f_{+-}^1&=&\frac{s}{n}\left(\frac{4s(t+u)}{\sqrt\lambda}+\sqrt\lambda -\lambda_z(t - u ) \right)c^{00}_{qqq}(s)-\frac{2s}{n}\left(\sqrt\lambda
+ \lambda_z(t - u)\right)c^{hz}_{qqq}(s ) \nonumber\\ & & { } -\frac{2(t - h)}{\sqrt\lambda n}\left(-s(u+3t)-2n+(u - t)(t - z ) + \lambda_z(t - s - z)\sqrt\lambda\right)c^{h0}_{qqq}(t ) \nonumber\\ & & { } + \frac{2(u - z)}{\sqrt\lambda n}\left(3u(s - z)+th-2z(h-2u)-\lambda_z(h - u ) \sqrt\lambda\right)c^{z0}_{qqq}(u ) \nonumber\\ & & { } + \frac{s}{\sqrt\lambda n}\left[t\left(\lambda+8zh-4ts -2(t+u)(z+h)\right.\right .
\nonumber\\ & & { } + \left.\left.\lambda_z(-2h+3t+u-2z)\sqrt\lambda \right ) -16m_q^2n\right]d^{hz00}_{qqqq}(s , t ) \nonumber\\ & & { } + \frac{1}{2}\left(-\sqrt\lambda-\frac{16m_q^2s}{\sqrt\lambda}+ \lambda_z(t - u)\right)d^{h0z0}_{qqqq}(t , u ) .
\label{eq : fhb}\end{aligned}\ ] ] the quark box form factors for @xmath33 read @xmath224 \nonumber\\ & & { } -\left[n\left(t+u+\frac{\lambda}{s}\right)+2m_q^2\lambda\right ] d^{h0z0}_{qqqq}(t , u)-4\left(szh+m_q^2\lambda\right)d^{hz00}_{qqqq}(s , t ) , \nonumber\\ f^0_{+-}&= & -\left[2+\frac{(t - u)^2}{n}\right]\left[s(t+u)c^{00}_{qqq}(s ) -\lambda c^{hz}_{qqq}(s)\right ] \nonumber\\ & & { } -2\left[3t - u+\frac{t}{n}(t - u)^2\right ] \left[(t - z)c^{z0}_{qqq}(t)+(t - h)c^{h0}_{qqq}(t)\right ] \nonumber \\ & & { } -2\left(zn - m_q^2\lambda\right)d^{h0z0}_{qqqq}(t , u ) \nonumber\\ & & { } + 2\left\{st\left[3t - u+\frac{t}{n}(t - u)^2\right]+2m_q^2\lambda\right\ } d^{hz00}_{qqqq}(s , t ) , \nonumber\\ f^1_{++}&=&\left(\frac{z - h - s}{\sqrt\lambda}-\lambda_z\right ) \left\{(t - u)\left(\frac{s}{n}c^{00}_{qqq}(s)-\frac{1}{2}d^{h0z0}_{qqqq}(t , u ) \right)\right .
\nonumber\\ & & { } -\left.s\left[2+\frac{t}{n}(t - u)\right]d^{hz00}_{qqqq}(s , t)\right\ } \nonumber\\ & & { } + 2(t - z)\left\{\lambda_z\frac{h - t}{n}+\frac{1}{\sqrt\lambda } \left[2+\frac{(t - u)(t - h)}{n}\right]\right\}c^{z0}_{qqq}(t ) \nonumber\\ & & { } -2(t - h)\left\{\lambda_z\frac{h - u}{n}+\frac{1}{\sqrt\lambda } \left[2-\frac{(t - u)(u - h)}{n}\right]\right\}c^{h0}_{qqq}(t ) , \nonumber\\ f^1_{+-}&=&\left(\lambda_z-\frac{t - u}{\sqrt\lambda}\right ) \left[\frac{s}{n}(s - z+h)\left(c^{00}_{qqq}(s)-td^{hz00}_{qqqq}(s , t)\right ) + \frac{s+z - h}{2}d^{h0z0}_{qqqq}(t , u)\right ] \nonumber\\ & & { } -2(t - z)\left\{\lambda_z\frac{t - h}{n}-\frac{1}{\sqrt\lambda } \left[2+\frac{(t - u)(t - h)}{n}\right]\right\}c^{z0}_{qqq}(t ) \nonumber\\ & & { } -2(t - h)\left\{\lambda_z\frac{u - h}{n}+\frac{1}{\sqrt\lambda } \left[2-\frac{(t - u)(u - h)}{n}\right]\right\}c^{h0}_{qqq}(t ) .
\label{eq : fab}\end{aligned}\ ] ] r. k. ellis , i. hinchliffe , m. soldate , and j. j. van der bij , nucl .
* b297 * , 221 ( 1988 ) ; u. baur and e. w. n. glover , _ ibid . _ * b339 * , 38 ( 1990 ) ; u. langenegger , m. spira , a. starodumov , and p. trb , jhep * 0606 * , 035 ( 2006 ) [ arxiv : hep - ph/0604156 ]
. s. l. glashow , d. v. nanopoulos , and a. yildiz , phys .
d * 18 * , 1724 ( 1978 ) ; e. eichten , i. hinchliffe , k. d. lane , and c. quigg , rev . mod .
phys . * 56 * , 579 ( 1984 ) ; * 58 * , 1065(e ) ( 1986 ) ; z. kunszt , z. trocsanyi , and w. j. stirling , phys . lett .
b * 271 * , 247 ( 1991 ) .
m. spira , fortsch .
phys . * 46 * , 203 ( 1998 ) [ arxiv : hep - ph/9705337 ] ; m.s .
carena and h.e .
haber , prog .
phys . * 50 * , 63 ( 2003 ) [ arxiv : hep - ph/0208209 ] ; a. djouadi , phys .
rept . * 459 * , 1 ( 2008 ) [ arxiv : hep - ph/0503173 ] .
d. a. dicus , j. l. hewett , c. kao , and t. g. rizzo , phys .
d * 40 * , 787 ( 1989 ) ; d. a. dicus and c. kao , _ ibid .
_ , 832 ( 1990 ) ; o. brein , w. hollik , and s. kanemura , _ ibid .
_ , 095001 ( 2001 ) [ arxiv : hep - ph/0008308 ] ; w. hollik and s .- h .
zhu , _ ibid .
_ , 075015 ( 2002 ) [ arxiv : hep - ph/0109103 ] .
j. f. gunion , h. e. haber , f. e. paige , w .- k .
tung , and s. s. d. willenbrock , nucl .
* b294 * , 621 ( 1987 ) ; r. m. barnett , h. e. haber , and d. e. soper , _ ibid .
_ , 697 ( 1988 ) ; f. i. olness and w .- k .
tung , _ ibid . _ , 813 ( 1988 ) ; v. barger , r. j. n. phillips , and d. p. roy , phys .
b * 324 * , 236 ( 1994 ) [ arxiv : hep - ph/9311372 ] ; m. a. g. aivazis , j. c. collins , f. i. olness , and w. k. tung , phys .
d * 50 * , 3102 ( 1994 ) [ arxiv : hep - ph/9312319 ] ; j. c. collins , _ ibid .
_ , 094002 ( 1998 ) [ arxiv : hep - ph/9806259 ] ; m. krmer , f. i. olness , and d. e. soper , _ ibid .
_ , 096007 ( 2000 ) [ arxiv : hep - ph/0003035 ] .
h. e. haber and g. l. kane , phys .
rept . * 117 * , 75 ( 1985 ) ; j. f. gunion and h. e. haber , nucl .
phys . * b272 * , 1 ( 1986 ) ; * b402 * , 567(e ) ( 1993 ) ; * b278 * , 449 ( 1986 ) ; * b402 * , 569(e ) ( 1993 ) ; * b307 * , 445 ( 1988 ) ; * b402 * , 569(e ) ( 1993 ) ; j. f. gunion , h. e. haber , g. kane , and s. dawson , _ the higgs hunter s guide _ ( addison - wesley , redwood city , 1990 ) . | we study the hadroproduction of a @xmath0-even or @xmath0-odd neutral higgs boson in association with a @xmath1 boson in the minimal supersymmetric extension of the standard model ( mssm ) we include the contributions from quark - antiquark annihilation at the tree level and those from gluon - gluon fusion , which proceeds via quark and squark loops , and list compact analytic results .
we quantitatively analyze the hadronic cross sections at the cern large hadron collider assuming a favorable supergravity - inspired mssm scenario .
pacs numbers : 11.30.pb , 12.60.jv , 13.85.qk , 14.80.da 16.cm 22.cm
@=11 tempcntc citex[#1]#2@fileswauxout tempcnta@tempcntb@neciteaciteforciteb:=#2citeo#1 citeotempcnta > tempcntbciteacitea , tempcnta = tempcntbtempcnta @=12 |
let us consider a regression model in the continuous time with the levy noise @xmath0 where @xmath1 is an unknown @xmath2 function , @xmath3 is some unobserved noise and @xmath4 is the noise intensity .
the problem is to estimate the function @xmath5 on the observations @xmath6 when @xmath7 . in this paper
we consider the estimation problem in the adaptive setting , i.e. when the regularity of @xmath5 is unknown .
note that if @xmath3 is the brownian motion , then we obtain the well known `` signal+white noise '' model ( see , for example , @xcite , @xcite , @xcite and etc ) .
it should be noted also that the model is very popular in the statistical radio - physics .
this is the estimation problem of the signal @xmath5 , observed under the white noise , when the signal noise ratio goes to infinity . in this paper
we assume that the noise @xmath3 is the levy process with unknown distribution @xmath8 on the skorokhod space @xmath9 $ ] . we know only that this distribution beings to some distribution family @xmath10 specified below . by making use of the robust estimation approach developed for nonparametric problems in @xcite we set the robust risks as @xmath11 where @xmath12 is an estimate ,
i.e. any function of @xmath6 and @xmath13 the goal of this paper is to develop the sharp model selection method for estimating the unknown signal @xmath5 .
the interest to such statistical procedures can be explained by the fact that they provide adaptive solutions for the nonparametric estimation through the sharp non - asymptotic oracle inequalities which give non asymptotic upper bound for the quadratic risk including the minimal risk over chosen estimate family with some coefficient closed to one ( see , for example , @xcite for discrete time and @xcite for the continuous time ) .
the origin of the model selection method goes back to early seventies with the pioneering papers by akaike @xcite and mallows @xcite who suggested to use penalizing in a log - likelihood type criterion .
barron , birg , massart @xcite , massart @xcite and kneip @xcite developed a non - asymptotic model selection method which enables one to derive non - asymptotic oracle inequalities for the nonparametric regression models with gaussian disturbances .
unfortunately , these methods can not be applied to the non gaussian regression models , since the estimators for the fourier coefficients in such cases are not independent random variables . by these reasons to estimate the function in non - gaussian regression models we use the model selection method developed by @xcite for nongaussian heteroscedastic regression models in discrete time . in constructing the sharp model selection procedures , in this paper
, we will use the approach close to that of the papers @xcite , @xcite , @xcite , @xcite developed for the estimation of a @xmath14 - periodic function in the continuous time on the large time interval , i.e. @xmath15 note , that for any @xmath16 setting @xmath17 , we can represent this model as a model with small parameter of form @xmath18 where @xmath19 and @xmath20 .
main difference this model from the original one is that the jumps are small , i.e. @xmath21 but we have not such property in the model .
therefore , unfortunately , we can not use directly the method developed for the estimation problem on the large time interval to the model .
so , the main goal of this paper is to develop a new sharp model selection method for the problem estimation function @xmath5 as @xmath7 .
as an application of the sharp model selection method in this paper we consider the signals number detection problem for the model . in many areas of science and technology the problems
arise how to select the number of freedom degrees for a statistical model that most adequately describes phenomenons under studies @xcite
. an important class of such problems is the detection problems of the signal number with unknown parameters in the noise .
for example , in the signal multi - path information transmission there is a detection problem for the number of rays in a multipath channel .
this problem is often reduced to the detection of the signals number . as a result ,
effective detection signals number algorithms can significantly improve the noise immunity in the data transmission over a multipath channel @xcite . in all these paper the signals number detection problems are considered only for observation with white nose . in this paper
we consider this problem for non gaussian noise with jumps given by .
the rest of the paper is organized as follows . in section [ sec : tr ] we transform the observation model to delete the large jumps . in section [ sec : mo ] we construct the sharp model selection procedure . in section
[ sec : mrs ] we give the main results on non - asymptotic estimation . in section [ sec : adpest ] we give the results on asymptotic robust efficiency . in sections [ sec :
lobn ] and [ sec : upbn ] we study the lower and upper bounds for the asymptotic risks respecttively . in section [ sec : numsgn ] we study the signals number detection problem through the model selection method . in section [ sec : siml ] we give the simulations results . section [ sec : pr ] contains the proofs of all main results . in appendix
we bring all proofs for auxiliary results .
in this paper the noise process @xmath3 is defined by the following levy process @xmath22 here , @xmath23 and @xmath24 are some constants , @xmath25 is a standard brownian motion , @xmath26 is the jump measure with the deterministic compensator @xmath27 , @xmath28 is some positive measure on @xmath29 , ( see , for example @xcite for details ) .
@xmath30 note that @xmath31 may be equal to @xmath32 . in the sequel
we will denote by @xmath8 the distribution of the process @xmath3 in the skorokhod space @xmath33 $ ] and by @xmath34 we denote all these distributions for which the parameters @xmath23 and @xmath24 satisfy the conditions @xmath35 where the bounds @xmath36 and @xmath37 are such that for any @xmath38 @xmath39 first of all , we need to eliminate the large jumps in the observations , i.e. we transform this model as @xmath40 the parameter @xmath41 will be chosen later .
so , we obtain that @xmath42 where @xmath43 the functions @xmath44 and @xmath45 with the truncated threshold @xmath46 .
let @xmath47 be an orthonormal basis in @xmath48 $ ] with @xmath49 .
we assume that this basis is uniformly bounded , i.e. for some constant @xmath50 , which may be depend on @xmath4 , @xmath51 where @xmath52 $ ] and @xmath53 $ ] denotes integer part of @xmath54 .
for example , we can take the trigonometric basis defined as @xmath55 and for @xmath56 @xmath57 \sin(\omega_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}x)\quad\mbox{for odd}\quad j\ , , \end{array } \right.\ ] ] where the frequency @xmath58 $ ] .
moreover , note that for any @xmath59 \to\bbr$ ] function @xmath60 from @xmath48 $ ] , the integrals @xmath61 are well defined with @xmath62 , @xmath63 , @xmath64 where @xmath65 , @xmath66 and @xmath67 . in the sequel
we denote by @xmath68 . to estimate the function @xmath5 we use the following fourier serie @xmath69 where @xmath70
these coefficients can be estimated by the following way .
the first we estimate as @xmath71 and for @xmath56 @xmath72 taking into account here that for such @xmath73 the integral @xmath74 we obtain from that these fourier coefficients can be represented as @xmath75 setting @xmath76 we obtain that for any @xmath77 @xmath78 now , according to the selection model approach developed in @xcite - @xcite we need to define for any @xmath79 the following functions @xmath80 where @xmath81 and @xmath82 .
[ pr.sec : mapr.0 ] the following upper bound holds .
@xmath83^{n}}{u\in[0,1]^{n}}{\lower.25ex\hbox{$\scriptstyleu\in[0,1]^{n}$ } } { \lower0.25ex\hbox{$\scriptscriptstyleu\in[0,1]^{n}$}}}}\left\vert b_{{\mathchoice{1,\varepsilon}{1,\varepsilon}{\lower.25ex\hbox{$\scriptstyle1,\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle1,\varepsilon$}}}}(u ) \right\vert \le \varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$ } } } } \,.\ ] ] taking into account that @xmath84 and @xmath85 for @xmath56 we immediately the upper bound . before the formulation we recall the novikov inequalities , @xcite , also referred to as the bichteler
jacod inequalities , see @xcite , providing bounds moments of supremum of purely discontinuous local martingales for @xmath86 @xmath87 where @xmath88 is some positive constant .
now , for any @xmath89 we set @xmath90 [ pr.sec : mapr.1 ] for any fixed truncated model parameter @xmath91 and for any vector @xmath92 with @xmath93 @xmath94 where @xmath95 .
we estimate the function @xmath96 for @xmath97 $ ] by the weighted least squares estimator @xmath98 where @xmath99 $ ] , the weights @xmath100 belong to some finite set @xmath101 from @xmath59^n$ ] , @xmath102 is defined in and @xmath103 in .
now we set @xmath104 where @xmath105 is the cardinal number of @xmath101 . in the sequel
we assume that @xmath106 .
now we chose the truncating parameter @xmath107 as @xmath108 to choose a weight sequence @xmath109 in the set @xmath101 we use the empirical quadratic risk , defined as @xmath110 which in our case is equal to @xmath111 since the fourier coefficients @xmath112 are unknown , we replace the terms @xmath113 by @xmath114 where @xmath115 is a some estimate for the variance parameter @xmath116 from .
if it is known we set @xmath117 if non this estimator will be prescribed later . finally , to choose the weights we will minimize the following cost function @xmath118 where @xmath119 is some threshold which will be specified later and the penalty term @xmath120 note that , if the @xmath121 is known then the penalty is defined as @xmath122 here @xmath115 is a some estimator for the variance from the condition @xmath123 .
we define the model selection procedure as @xmath124 where @xmath125 we recall that the set @xmath101 is finite so @xmath126 exists . in the case when @xmath126 is not unique we take one of them .
now , we specify the weight coefficients @xmath127 . consider a numerical grid of the form @xmath128 where @xmath129 \,.\ ] ] we assume that both the parameters @xmath130 and @xmath131 are functions of @xmath132 , i.e. @xmath133 and @xmath134 , such that @xmath135 & \lim_{{\mathchoice{{\varepsilon}\to 0}{{\varepsilon}\to 0}{\lower.25ex\hbox{$\scriptstyle{\varepsilon}\to 0 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle{\varepsilon}\to 0$}}}}\varsigma^{*}_{{\mathchoice{{\varepsilon}}{{\varepsilon}}{\lower.25ex\hbox{$\scriptstyle{\varepsilon}$ } } { \lower0.25ex\hbox{$\scriptscriptstyle{\varepsilon}$}}}}\varpi_{{\mathchoice{{\varepsilon}}{{\varepsilon}}{\lower.25ex\hbox{$\scriptstyle{\varepsilon}$ } } { \lower0.25ex\hbox{$\scriptscriptstyle{\varepsilon}$}}}}=0 \quad\mbox{and}\quad \lim_{{\mathchoice{{\varepsilon}\to 0}{{\varepsilon}\to 0}{\lower.25ex\hbox{$\scriptstyle{\varepsilon}\to 0 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle{\varepsilon}\to 0$}}}}\,{\varepsilon}^{-\delta}\varpi_{{\mathchoice{{\varepsilon}}{{\varepsilon}}{\lower.25ex\hbox{$\scriptstyle{\varepsilon}$ } } { \lower0.25ex\hbox{$\scriptscriptstyle{\varepsilon}$}}}}\,=+\infty \end{array } \right.\ ] ] for any @xmath119 .
one can take , for example , for @xmath136 @xmath137 where @xmath138 is some fixed constant and the threshold @xmath37 is introduced in .
for each @xmath139 , we introduce the weight sequence @xmath140 where @xmath141 , @xmath142 here @xmath143 $ ] , @xmath144 now we define the set @xmath101 as @xmath145 note , that these weight coefficients are used in @xcite for continuous time regression models to show the asymptotic efficiency . in the sequel
we need to estimate the variance parameter @xmath116 from . to this end we set for any @xmath146 @xmath147 + 1}{j=[1/\varepsilon]+1}{\lower.25ex\hbox{$\scriptstylej=[1/\varepsilon]+1 $ } } { \lower0.25ex\hbox{$\scriptscriptstylej=[1/\varepsilon]+1$}}}}\,{\widehat}{t}^2_{{\mathchoice{j,\varepsilon}{j,\varepsilon}{\lower.25ex\hbox{$\scriptstylej,\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstylej,\varepsilon$}}}}\ , , \quad n=1/\varepsilon^{2}\,,\ ] ] where @xmath148 are the estimators for the fourrier coefficients with respect to the trigonometric basis , i.e. @xmath149 [ re.sec:mo.1 ]
note that the similar sharp oracle inequalities was obtained before in the papers @xcite and @xcite for the nonparametric regression models in the discrete and continuous times respectively . in this paper
we obtain these inequalities for the model selection procedures based on any arbitrary orthogonal basic function .
we use the trigonometric function only to estimate the noise intensity @xmath121 .
first we set the following constant which will be used to describe the rest term in the oracle inequalities .
we set @xmath150 where @xmath151 we start with the sharp oracle inequalities .
[ th.sec:mrs.1 ] assume that for the model the condition holds .
then for any @xmath152 , the estimator of @xmath5 given in satisfies the following oracle inequality @xmath153 [ co.sec:oi.1 ] assume that for the model the condition holds .
if the variance parameter @xmath121 is known , then for any @xmath152 , the estimator of @xmath5 given in with the truncate parameter @xmath154 satisfies the following oracle inequality @xmath155 we need to study the estimate .
[ pr.sec : mapr.3 ] assume that in the model the unknown function @xmath1 is continuously differentiable .
then , for any @xmath146 @xmath156 where @xmath157 . the proof of this proposition is given in section [ sec : pr ] .
it is clear that in the case when @xmath158 were obtain that @xmath159 now using this proposition we can obtain the following inequality .
[ th.sec:mrs.20 ] assume that for the model the condition holds and the unknown function @xmath1 is continuously differentiable .
then the procedure with @xmath158 , for any @xmath152 , satisfies the following oracle inequality @xmath160 \label{sec : mrs.1 + } & + \varepsilon^2 \frac{\psi_{{\mathchoice{q,\varepsilon}{q,\varepsilon}{\lower.25ex\hbox{$\scriptstyleq,\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq,\varepsilon$}}}}+(\vert\dot{s}\vert+1)^{2 } g_{{\mathchoice{1,q}{1,q}{\lower.25ex\hbox{$\scriptstyle1,q$ } } { \lower0.25ex\hbox{$\scriptscriptstyle1,q$}}}}+g_{{\mathchoice{2,q}{2,q}{\lower.25ex\hbox{$\scriptstyle2,q$ } } { \lower0.25ex\hbox{$\scriptscriptstyle2,q$}}}}}{\delta}\,,\end{aligned}\ ] ] where @xmath161 now we study the robust risks defined in for the procedure . moreover , we assume also that the upper bound for the basis functions in may be dependent on @xmath162 , i.e. @xmath163 , such that for any @xmath164 @xmath165 [ th.sec:mrs.2 ] assume that for the model the condition holds and the unknown function @xmath1 is continuously differentiable
. then robust risks of the procedure with @xmath158 , for any @xmath152 , satisfy the following oracle inequality @xmath166 where the term @xmath167 is such that under the conditions and for any @xmath168 and @xmath38 @xmath169
now we study the asymptotically efficience properties for the procedure , with respect to the robust risks defined by the distribution family . to this end
we assume that the unknown function belongs to the following ellipsoid in @xmath170 @xmath171\,:\ , \sum_{{\mathchoice{j=1}{j=1}{\lower.25ex\hbox{$\scriptstylej=1 $ } } { \lower0.25ex\hbox{$\scriptscriptstylej=1$}}}}^{\infty}\,a_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}\,\theta^2_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}\,\le \r\}\ ] ] where @xmath172\right)^{2i}$ ] .
it is easy to see that in the case when the functions @xmath173 are trigonometric , then this set coincides with the sobolev ball @xmath174 \,:\,\sum_{{\mathchoice{j=0}{j=0}{\lower.25ex\hbox{$\scriptstylej=0 $ } } { \lower0.25ex\hbox{$\scriptscriptstylej=0$}}}}^k\,\|f^{(j)}\|^2\le \r\}\,,\ ] ] where @xmath175 and @xmath176 are some parameters , @xmath177 $ ] is the set of @xmath178 times continuously differentiable functions @xmath179\to\bbr$ ] such that @xmath180 for all @xmath181 . similarly to @xcite we will show here that the asymptotic sharp lower bound for the robust risk is given by @xmath182 note that this is the well - known pinsker constant obtained for the nonadaptive filtration problem in `` signal + small white noise '' model ( see , for example , @xcite ) . let @xmath183 be the set of all estimators @xmath184 measurable with respect to the sigma - algebra @xmath185 generated by the process .
[ th.sec:ef.1 ] for the distribution family .
the robust risks admit the following lower bound @xmath186 where @xmath187 .
we set the parameter @xmath188 in as function of @xmath189 , i.e. @xmath190 is such that @xmath191 for any @xmath38 .
for example , we can take @xmath192 .
[ th.sec:ef.2 ] assume that the conditions hold .
then model selection procedure admits the following asymptotic upper bound @xmath193 theorem [ th.sec:ef.1 ] and theorem [ th.sec:ef.2 ] imply the following result [ co.sec:mr.1 ] under the conditions theorem [ th.sec:ef.2 ] @xmath194 [ re.sec.mrs.1 ] it is well known that the optimal ( minimax ) risk convergence rate for the sobolev ball @xmath195 is @xmath196 ( see , for example , @xcite ) .
we see here that the efficient robust rate is @xmath197 , i.e. if the distribution upper bound @xmath198 as @xmath199 we obtain the more rapid rate with respect to @xmath196 , and if @xmath200 as @xmath7 we obtain the more slow rate . in the case
when @xmath37 is constant the robuste rate is the same as the classical non robuste convergence rate .
firstly , note , that for any fixed @xmath201 @xmath202 now for any fixed @xmath203 we set @xmath204 \quad\mbox{and}\quad \check{\r}=(1-\check{\gamma})\r\,.\ ] ] using this definition we introduce the parametric family @xmath205 as @xmath206 to define the bayesian risk we choose a prior distribution on @xmath207 as @xmath208 where @xmath209 are i.i.d .
gaussian @xmath210 random variables and the coefficients @xmath211 denoting by @xmath212 the distribution of the random variables @xmath213 on @xmath207 we introduce the bayes risk as @xmath214 furthermore , for any function @xmath60 , we denote by @xmath215 its projection in @xmath48 $ ] onto @xmath216 , i.e. @xmath217 since @xmath218 is a convex and closed set in @xmath48 $ ] , this project exists and unique for any function @xmath219 $ ] and , moreover , @xmath220 so , setting @xmath221 , we obtain that @xmath222 taking into account now that @xmath223 we obtain @xmath224 and @xmath225 therefore , in view of @xmath226 as to the last term in this inequality , in appendix we show that for any @xmath227 @xmath228 now it is easy to see that @xmath229 where @xmath230 so , in view of lemma [ le.sec:app.3 ] and reminding that @xmath231 we obtain @xmath232 & = \frac{1}{v_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}}}\ , \sum_{j=1}^{d}\,\frac{s^{*}_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$ } } } } } { s^{*}_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}+\,1 } = \frac{1}{v_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}}}\ , \sum_{j=1}^{d}\ , \left ( 1 - \frac{j^k}{d^k_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$ } } } } } \right ) \,.\end{aligned}\ ] ] therefore , using now the definition , the inequality and the limit , we obtain that @xmath233 taking here limit as @xmath234 implies theorem [ th.sec:ef.1 ] .
first we suppose that the parameters @xmath235 , @xmath175 in and @xmath37 in are known .
let the family of admissible weighted least squares estimates @xmath236 given by . consider the pair @xmath237\ ] ] where @xmath238 and @xmath239 satisfy the conditions in .
denote the corresponding setimate as @xmath240 note that for sufficiently small @xmath189 the pair @xmath241 belongs to the set .
[ th.sec:up.1 ] the estimator @xmath242 admets the following asymptotic upper bound @xmath243 * proof . * substituting and taking into account the definition one gets @xmath244 where @xmath245 .
note now that for any @xmath246 the expectation @xmath247 and , in view of the upper bound , @xmath248 therefore , @xmath249 where @xmath250 .
setting @xmath251 we obtain that for each @xmath252 @xmath253 one can check directly here that @xmath254 where the coefficient @xmath255 is given in .
moreover , due to the condition @xmath256 therefore , @xmath257 to estimate the last term in the right hand of , we set @xmath258 it is easy to check that @xmath259 therefore , taking into account that by the definition of the pinsker constant in @xmath260 , we arrive at the inequality @xmath261 hence theorem [ th.sec:up.1 ] . combining theorem [ th.sec:up.1 ] and theorem [ th.sec:mrs.2 ] yields theorem [ th.sec:ef.2 ] .
in this section we apply the model selection procedure the following problem from the statistical signals theory .
assume that the observed through some noise unknown signal in the model has the following form @xmath262 where @xmath173 are orthonormal known basis functions in @xmath263 , but the signals number @xmath264 and the coefficients @xmath265 are unknown .
the problem consists to estimation of @xmath264 on the observation . in the statistical radio - physics
this mens that we need to detect the number of received signals @xmath173 in multipath connection channels . in this case
the coefficients @xmath265 are the signal amplitudes . for this problem we use the lse family @xmath266 defined as @xmath267 this estimate can be obtained from with the weights @xmath268 .
the number of estimators @xmath269 is a some function of @xmath189 , i.e. @xmath270 , such that @xmath271 for any @xmath38 . as a risk for the signals number
we use @xmath272 where the risk @xmath273 is defined in and @xmath274 is some integer number ( maybe random ) from the set @xmath275 . in this case
the cost function has the following form .
@xmath276 so , for this problem the lse model selection procedure is defined as @xmath277 note that theorem [ th.sec:mrs.2 ] implies that the robust risks of the procedure with @xmath158 , for any @xmath152 , satisfy the following oracle inequality @xmath278 where the last term satisfies the property .
in this section we report the results of a monte carlo experiment to assess the performance of the proposed model selection procedure .
in we chose @xmath279 with @xmath280 , @xmath281j$ ] .
we simulate the model @xmath282 the frequency of observations per period equals @xmath283 .
we use the weight sequence as proposed in galtchouk and pergamenshchikov ( 2009 ) for a discrete time model : @xmath284 and @xmath285 $ ] .
we calculated the empirical quadratic risk defined as @xmath286 and the relative quadratic risk @xmath287 the expectations was taken as an average over @xmath288 replications , i.e. @xmath289 we used the cost function with @xmath290 : empirical risks [ cols="<,^,>",options="header " , ] [ re.sec:siml.1 ] it should be noted that the lse procedure is more appropriate than shrinkage method for such number detection problem .
first note that @xmath291 where @xmath292 . it should be noted that @xmath293 to study the last term in the right hand side of the inequality we set for any function @xmath60 from @xmath48 $ ] @xmath294 note that that for @xmath56 we random variables @xmath295 .
so , @xmath296 by the ito formula we can write that for any function @xmath60 from @xmath48 $ ] @xmath297 where @xmath298 .
so , taking into account that @xmath299 we obtain that @xmath300 so , setting @xmath301 we obtain that @xmath302 so , we obtain that @xmath303 where @xmath304 .
moreover , taking into account that for any @xmath60 , @xmath305 from @xmath48 $ ] @xmath306 we get @xmath307 & = \check{\varkappa}_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}\ , \sum^{n}_{{\mathchoice{i , j=2}{i , j=2}{\lower.25ex\hbox{$\scriptstylei , j=2 $ } } { \lower0.25ex\hbox{$\scriptscriptstylei , j=2$}}}}u_{{\mathchoice{i}{i}{\lower.25ex\hbox{$\scriptstylei$ } } { \lower0.25ex\hbox{$\scriptscriptstylei$}}}}\,u_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}\,\left(\int^{1}_{{\mathchoice{0}{0}{\lower.25ex\hbox{$\scriptstyle0 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle0$}}}}\,\phi_{{\mathchoice{i}{i}{\lower.25ex\hbox{$\scriptstylei$ } } { \lower0.25ex\hbox{$\scriptscriptstylei$}}}}(t)\phi_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}(t)\ , \,\d t\right)^{2 } = \check{\varkappa}_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}\,.\end{aligned}\ ] ] thus , @xmath308 now , to estimate the second term in the inequality note that in view of the inequality for any bounded function @xmath60 and any @xmath309 @xmath310 & \le 24\varrho^{4}_{{\mathchoice{1}{1}{\lower.25ex\hbox{$\scriptstyle1 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle1$}}}}\int^{1}_{{\mathchoice{0}{0}{\lower.25ex\hbox{$\scriptstyle0 $ } }
{ \lower0.25ex\hbox{$\scriptscriptstyle0$}}}}f^{2}(t)\d t + c^{*}_{{\mathchoice{4}{4}{\lower.25ex\hbox{$\scriptstyle4 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle4$}}}}\left ( \left(\pi(h^{2}_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}})\,\int^{1}_{{\mathchoice{0}{0}{\lower.25ex\hbox{$\scriptstyle0 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle0$}}}}f^{2}(t)\d t\right)^{2 } + \pi(h^{4}_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}})\,\int^{1}_{{\mathchoice{0}{0}{\lower.25ex\hbox{$\scriptstyle0 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle0$}}}}f^{4}(t)\d t \right)\,,\end{aligned}\ ] ] i.e. @xmath311 now it is easy to see that through the holder inequality the term @xmath312 can be estimated as @xmath313 from here and the inequality it follows that @xmath314 and , therefore , @xmath315 this implies that @xmath316 thus , the ito formula implies @xmath317 in the same way we calculate @xmath318 & = \varrho^{2}_{{\mathchoice{2}{2}{\lower.25ex\hbox{$\scriptstyle2 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle2$}}}}\pi(h^{4}_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}})\,\int^{1}_{{\mathchoice{0}{0}{\lower.25ex\hbox{$\scriptstyle0 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle0$}}}}\,\psi^{2}_{{\mathchoice{t}{t}{\lower.25ex\hbox{$\scriptstylet$ } } { \lower0.25ex\hbox{$\scriptscriptstylet$}}}}\d t \le \pi(x^{2})(a/\varepsilon)^{2}\,\left(\phi^{*}\right)^{4}\,\#(u)\,.\end{aligned}\ ] ] so , we obtain that @xmath319 similarly we obtain that @xmath320 this implies the upper bound . first note , that we can rewrite the empirical squared error in as follows @xmath321 where @xmath322 .
now using the definition of @xmath323 in we obtain that @xmath324 where @xmath325 and @xmath326 . setting @xmath327 we can rewrite as @xmath328 where @xmath329 , the exact penalization is defined in and the functions @xmath330 and @xmath331 are defined in .
it should be noted that for the truncates parameter the bound implies @xmath332 where @xmath333 .
let @xmath334 be a fixed sequence in @xmath101 and @xmath335 be as in .
substituting @xmath336 and @xmath335 in the equation , we obtain @xmath337 & + 2\varepsilon^2 b_{{\mathchoice{1,\varepsilon}{1,\varepsilon}{\lower.25ex\hbox{$\scriptstyle1,\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle1,\varepsilon$}}}}(\varpi)+2 m_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}}(\varpi)\nonumber\\[2 mm ] & + 2\varepsilon \sqrt{p_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}}({\widehat}{\lambda } ) } \frac{b_{{\mathchoice{2,\varepsilon}{2,\varepsilon}{\lower.25ex\hbox{$\scriptstyle2,\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle2,\varepsilon$}}}}({\widehat}{u})}{\sqrt{\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}}}- 2\varepsilon \sqrt{p_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}}(\lambda_0 ) } \frac{b_{{\mathchoice{2,\varepsilon}{2,\varepsilon}{\lower.25ex\hbox{$\scriptstyle2,\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle2,\varepsilon$}}}}(u_{{\mathchoice{0}{0}{\lower.25ex\hbox{$\scriptstyle0 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle0$}}}})}{\sqrt{\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}}}\nonumber \\[2mm]\label{sec : pr.4 } & - \delta { \widehat}{p}_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}}({\widehat}{\lambda})+ \delta { \widehat}{p}_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}}(\lambda_0)\end{aligned}\ ] ] where @xmath338 , @xmath339 and @xmath340 . note that by @xmath341 the inequality @xmath342 implies that for any @xmath343 @xmath344 from the bound it follows that for @xmath345 @xmath346 & + \varepsilon^2 |{\widehat}{\varkappa } -\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}| ( |{\widehat}{\lambda}|^2 + |\lambda_0|^2 + 4\vert\lambda\vert_{{\mathchoice{*}{*}{\lower.25ex\hbox{$\scriptstyle*$ } } { \lower0.25ex\hbox{$\scriptscriptstyle*$}}}})+ 2 \delta p_\varepsilon(\lambda_0)\,,\end{aligned}\ ] ] where @xmath347 .
it should de noted that through we can estimate this term as @xmath348 taking into account that @xmath349 , we can rewrite the previous bound as @xmath350 & + \frac{6\varepsilon^2 \vert\lambda\vert_{{\mathchoice{*}{*}{\lower.25ex\hbox{$\scriptstyle*$ } } { \lower0.25ex\hbox{$\scriptscriptstyle*$}}}}}{n } |{\widehat}{\varkappa } -\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}| + 2 \delta p_\varepsilon(\lambda_0).\end{aligned}\ ] ] to estimate the second term in the right hand side of this inequality we introduce @xmath351 moreover , note that @xmath352 where @xmath353 .
therefore , thanks to we obtain that for any nonrandom @xmath354 @xmath355 to estimate this function for a random vector we set @xmath356 so , through the inequality @xmath357 it is clear that the last term here can be estimated as @xmath358 where @xmath359 .
moreover , note that , for any @xmath360 , @xmath361 where @xmath362 .
taking into account now , that for any @xmath363 the components @xmath364 , we can estimate the last term as in , i.e. @xmath365 similarly , setting @xmath366 we obtain @xmath367 by the same way we find that @xmath368 and , for any @xmath369 , @xmath370 so , from we get @xmath371 therefore , taking into account that @xmath372 , the term @xmath373 can be estimated as @xmath374 using this bound in we obtain @xmath375 & + \frac{2\varepsilon^{2}\,\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}}{1 - 3\delta } + \frac{6\varepsilon^{2}\,\vert\lambda\vert_{{\mathchoice{*}{*}{\lower.25ex\hbox{$\scriptstyle*$ } } { \lower0.25ex\hbox{$\scriptscriptstyle*$}}}}}{(1 - 3\delta ) } |{\widehat}{\varkappa } -\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}| + \frac{2\delta}{(1 - 3\delta ) } p_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}}(\lambda_0).\end{aligned}\ ] ] moreover , for @xmath376 we can rewrite this inequality as @xmath377 & + 4\varepsilon^{2}\,\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$ } } } } + 12\varepsilon^{2}\,\vert\lambda\vert_{{\mathchoice{*}{*}{\lower.25ex\hbox{$\scriptstyle*$ } } { \lower0.25ex\hbox{$\scriptscriptstyle*$ } } } } |{\widehat}{\varkappa } -\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}| + 4\delta\ , p_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}}(\lambda_0)\,.\end{aligned}\ ] ] using here the bounds , , we obtain that @xmath378 & + 4\varepsilon^{2}\,\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$ } } } } + 12\varepsilon^{2}\,\vert\lambda\vert_{{\mathchoice{*}{*}{\lower.25ex\hbox{$\scriptstyle*$ } } { \lower0.25ex\hbox{$\scriptscriptstyle*$ } } } } \e_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}\,|{\widehat}{\varkappa } -\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}| + \frac{2\delta}{1 - 3\delta}\ , p_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$}}}}(\lambda_0)\,.\end{aligned}\ ] ] now , lemma [ le.sec:a.1-06-11-01 ] implies directly the inequality .
hence theorem [ th.sec:mrs.1 ] .
@xmath379 we use here the same method as in @xcite . first
, note that from the definitions and we obtain @xmath380 where @xmath381 so , we have @xmath382 + 1}{j=[1/\varepsilon]+1}{\lower.25ex\hbox{$\scriptstylej=[1/\varepsilon]+1 $ } } { \lower0.25ex\hbox{$\scriptscriptstylej=[1/\varepsilon]+1$}}}}\,t^2_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$ } } } } + 2 \check{m}_{{\mathchoice{\varepsilon}{\varepsilon}{\lower.25ex\hbox{$\scriptstyle\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\varepsilon$ } } } } + \varepsilon^{2 } \sum^n_{{\mathchoice{j=[1/\varepsilon]+1}{j=[1/\varepsilon]+1}{\lower.25ex\hbox{$\scriptstylej=[1/\varepsilon]+1 $ } } { \lower0.25ex\hbox{$\scriptscriptstylej=[1/\varepsilon]+1$}}}}\,(\eta^{a}_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}})^{2 } \,,\ ] ] where @xmath383 + 1}{j=[1/\varepsilon]+1}{\lower.25ex\hbox{$\scriptstylej=[1/\varepsilon]+1 $ } } { \lower0.25ex\hbox{$\scriptscriptstylej=[1/\varepsilon]+1$}}}}\,t_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}\,\eta^{a}_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}$ ] .
note that for the continiously differentiable functions ( see , for example , lemma a.6 in @xcite ) the fourrier coefficients @xmath384 for any @xmath162 satisfy the following inequality @xmath385 + 1}{j=[1/\varepsilon]+1}{\lower.25ex\hbox{$\scriptstylej=[1/\varepsilon]+1 $ } } { \lower0.25ex\hbox{$\scriptscriptstylej=[1/\varepsilon]+1$}}}}\,t^2_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$ } } } } \le 4\varepsilon\left(\int^{1}_{{\mathchoice{0}{0}{\lower.25ex\hbox{$\scriptstyle0 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle0$}}}}\vert\dot{s}(t)\vert \d t\right)^{2 } \le 4\varepsilon \vert\dot{s}\vert^{2 } \,.\ ] ] the term @xmath386 can be estimated by the same way as in , i.e. @xmath387 + 1}{j=[1/\varepsilon]+1}{\lower.25ex\hbox{$\scriptstylej=[1/\varepsilon]+1 $ } } { \lower0.25ex\hbox{$\scriptscriptstylej=[1/\varepsilon]+1$}}}}\,t^{2}_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$ } } } } \le 4\varepsilon^{3}\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}\vert\dot{s}\vert^{2}\,.\ ] ] moreover , taking into account that for @xmath56 the expectation @xmath388 we can represent the last term in as @xmath389 + 1}{j=[1/\varepsilon]+1}{\lower.25ex\hbox{$\scriptstylej=[1/\varepsilon]+1 $ } } { \lower0.25ex\hbox{$\scriptscriptstylej=[1/\varepsilon]+1$}}}}\,(\eta^{a}_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}})^{2 } = \varepsilon^{2}\,\check{\varkappa}_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}(n-[1/\varepsilon ] ) + \varepsilon \ , b_{{\mathchoice{2,\varepsilon}{2,\varepsilon}{\lower.25ex\hbox{$\scriptstyle2,\varepsilon$ } } { \lower0.25ex\hbox{$\scriptscriptstyle2,\varepsilon$}}}}(x ' ) \,,\ ] ] where the function @xmath390 is defined in and @xmath391 .
we remind that @xmath392 $ ] . therefore , in view of proposition we obtain @xmath393 + 1}{j=[\sqrt{1/\varepsilon}]+1}{\lower.25ex\hbox{$\scriptstylej=[\sqrt{1/\varepsilon}]+1 $ } } { \lower0.25ex\hbox{$\scriptscriptstylej=[\sqrt{1/\varepsilon}]+1$}}}}\,\eta^2_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$ } } } } -\sigma_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$ } } } } \right\vert \le 2\varepsilon\ , \varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$ } } } } + \varepsilon \sqrt{u_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}}+\frac{\sqrt{6\varkappa_{{\mathchoice{q}{q}{\lower.25ex\hbox{$\scriptstyleq$ } } { \lower0.25ex\hbox{$\scriptscriptstyleq$}}}}}}{\vert\lambda\vert_{{\mathchoice{*}{*}{\lower.25ex\hbox{$\scriptstyle*$ } } { \lower0.25ex\hbox{$\scriptscriptstyle*$ } } } } } \,.\ ] ] so , we obtain the bound .
hence proposition [ pr.sec : mapr.3 ]
. * acknowledgements . *
the research is funded by the academic d.i .
mendeleev fund program of the tomsk state university ( research project nu
8.1.55.2015 l ) .
first , setting @xmath401 , we obtain that @xmath402 moreover , note that one can check directly that @xmath403 so , for sufficiently small @xmath189 we obtain that @xmath404 where @xmath405 , @xmath406 and @xmath407 through the correlation inequality ( see , proposition a.1 in @xcite ) we can get that for any @xmath86 @xmath408 as @xmath7 . therefore , for any @xmath227 using the chebychev inequality for @xmath409 we obtain that @xmath410 hence the equality . where @xmath1 is any arbitrary nonrandom square integrated function , i.e. from @xmath412 $ ] and @xmath413 is the levy process of the form with nonzero constants @xmath23 and @xmath24 .
we denote by @xmath414 and @xmath415 the distributions of the processes @xmath6 and @xmath3 on the skorokhod space @xmath416 $ ] .
now for any @xmath417 and @xmath418 from @xmath416 $ ] we set @xmath419 where @xmath420 is the continuous part of the process @xmath418 in @xmath416 $ ] , i.e. @xmath421 and for any @xmath422 and any measurable @xmath423 from @xmath424 @xmath425,\gamma ) = \sum_{{\mathchoice{0\le s\le t}{0\le s\le t}{\lower.25ex\hbox{$\scriptstyle0\le s\le t$ } } { \lower0.25ex\hbox{$\scriptscriptstyle0\le s\le t$}}}}\,\chi_{{\mathchoice{\{\delta x_{{\mathchoice{s}{s}{\lower.25ex\hbox{$\scriptstyles$ } } { \lower0.25ex\hbox{$\scriptscriptstyles$}}}}\in\varrho_{{\mathchoice{2}{2}{\lower.25ex\hbox{$\scriptstyle2 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle2$}}}}\gamma\}}{\{\delta x_{{\mathchoice{s}{s}{\lower.25ex\hbox{$\scriptstyles$ } } { \lower0.25ex\hbox{$\scriptscriptstyles$}}}}\in\varrho_{{\mathchoice{2}{2}{\lower.25ex\hbox{$\scriptstyle2 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle2$}}}}\gamma\}}{\lower.25ex\hbox{$\scriptstyle\{\delta x_{{\mathchoice{s}{s}{\lower.25ex\hbox{$\scriptstyles$ } } { \lower0.25ex\hbox{$\scriptscriptstyles$}}}}\in\varrho_{{\mathchoice{2}{2}{\lower.25ex\hbox{$\scriptstyle2 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle2$}}}}\gamma\}$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\{\delta x_{{\mathchoice{s}{s}{\lower.25ex\hbox{$\scriptstyles$ } } { \lower0.25ex\hbox{$\scriptscriptstyles$}}}}\in\varrho_{{\mathchoice{2}{2}{\lower.25ex\hbox{$\scriptstyle2 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle2$}}}}\gamma\}$ } } } } \,.\ ] ] now we study the measures @xmath414 and @xmath415 in @xmath416 $ ] .
* proof . *
note that to show this proposition it suffices to check that for any @xmath429 any @xmath430 for @xmath431 @xmath432 taking into account that the processes @xmath433 and @xmath413 have the independent homogeneous increments , to this end one needs to check only that for any @xmath434 and @xmath435 @xmath436 to check this equality note that the process @xmath437 is the gaussian martingale .
from here we directly obtain the squation . hence proposition [ pr.sec : app.1++ ] . in this section
we consider the following continuous time parametric regression model @xmath438 where @xmath439 with the unknown parameters @xmath440 and the process @xmath3 is defined in .
note now that according to proposition [ pr.sec : app.1++ ] the distribution @xmath441 of the process is absolutely continuous with respect to the @xmath415 on @xmath33 $ ] and the corresponding radon - nikodym derivative is @xmath442 where @xmath443 is arbitrary function from @xmath33 $ ] .
let @xmath444 be a prior density on @xmath445 having the following form : @xmath446 where @xmath447 is some continuously differentiable density in @xmath29 .
moreover , let @xmath448 be a continuously differentiable @xmath449 function such that , for each @xmath450 , @xmath451 where @xmath452 for any @xmath453 measurable integrable function @xmath454 we denote @xmath455 & = \int_{\bbr^d}\,\int_{{\mathchoice{{{\cal x}}}{{{\cal x}}}{\lower.25ex\hbox{$\scriptstyle{{\cal x}}$ } } { \lower0.25ex\hbox{$\scriptscriptstyle{{\cal x}}$}}}}\ , h(x,\theta)\,f(x,\theta)\,\phi(\theta)\d \p_{{\mathchoice{\xi}{\xi}{\lower.25ex\hbox{$\scriptstyle\xi$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\xi$}}}}(x)\ , \d \theta\,,\end{aligned}\ ] ] where @xmath456 $ ] .
* proof . *
first of all note that , the density on the process @xmath461 is bounded with respect to @xmath462 and for any @xmath450 @xmath463 now , we set @xmath464 taking into account the condition and integrating by parts yield @xmath465 & = \int_{{{\cal x}}\times\bbr^{d-1}}\left(\int_{\bbr}\ , g^{\prime}_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}(\theta)\ , f(x,\theta)\phi(\theta)\d \theta_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}\right)\left(\prod_{i\neq j}\d \theta_i\right)\,\p_{{\mathchoice{\xi}{\xi}{\lower.25ex\hbox{$\scriptstyle\xi$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\xi$}}}}(\d x ) = \lambda_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}\,.\end{aligned}\ ] ] now by the bouniakovskii - cauchy - schwarz inequality we obtain the following lower bound for the quadratic risk @xmath466 to study the denominator in the left hand of this inequality note that in view of the reprentation @xmath467 therefore , for each @xmath468 , @xmath469 and @xmath470 taking into account that @xmath471 we get @xmath472 hence lemma [ le.sec:app.3 ] . bichteler k. , jacod j. _ calcul de malliavin pour les diffusions avec sauts : existence dune densit dans le cas unidimensionnel_. sminaire de probabilit , xvii , lecture notes in math . , * 986 * , springer , berlin , 1983 , 132157 . el - behery i.n . and macphie r.h .
( 1978 ) maximum likelihood estimation of the number , directions and strengths of point radio sources from variable baseline interferometer data .
_ ieee transactions on antennas and propagation _ , * 26 * ( 2 ) , 294 301 .
flaksman , a.g .
( 2002 ) adaptive signal processing in antenna arrays with allowance for the rank of the impule - response matrix of a multipath channel _ izvestiya vysshikh uchebnykh zavedenij .
radiofizika _ , * 45 * ( 12 ) , 1064 1077 .
galtchouk , l. and pergamenshchikov , s. ( 2013 ) uniform concentration inequality for ergodic diffusion processes observed at discrete times . _ stochastic processes and their applications _ , * 123 * , 91109 galtchouk , l. and pergamenshchikov , s. ( 2009 ) adaptive asymptotically efficient estimation in heteroscedastic nonparametric regression via model selection , _
/ hal-00326910/fr/_. konev v. v. and pergamenshchikov s. m. sequential estimation of the parameters in a trigonometric regression model with the gaussian coloured noise .
_ statistical inference for stochastic processes _ * 6 * ( 2003 ) 215235 .
konev v. v. and pergamenshchikov s. m. nonparametric estimation in a semimartingale regression model .
. _ journal of mathematics and mechanics of tomsk state university _ * 3 * ( 2009 ) 2341 .
konev v. v. and pergamenshchikov s. m. nonparametric estimation in a semimartingale regression model .
part 2 . robust asymptotic efficiency .
_ journal of mathematics and mechanics of tomsk state university _ * 4 * ( 2009 ) 3145 .
manelis , b.v .
( 2007 ) algorithms for tracking and demodulation of the mobile communication signal in the conditions of the indiscriminating multipath propagation , _ radiotekhnika _ , 4 , 1621 ( in russian ) .
trifonov a.p .
, kharin a.v .
, chernoyarov o.v .
, kalashnikov k. s. ( 2015 ) determining the number of radio signals with unknown phases . _ international journal on communications antenna and propagation _ , * 5 * ( 6 ) , 367374 . trifonov a.p . , kharin a.v .
, chernoyarov o.v .
( 2016 ) estimation of the number of orthogonal signals with unknown parameters .
_ journal of communications technology and electronics _ , * 61 * ( 9 ) , 1026 1033 . | we consider a nonparametric robust estimation problem in the continuous time for the functions observed on the fixed time interval with a noise defined by the levy processes with jumps .
an adaptive model selection procedure is proposed .
a sharp non - asymptotic oracle inequalities for the robust risks is obtained and the robust efficiency is shown .
we apply this procedure to signals number detection problem in the multipath connection channel .
_ msc : _
primary 62g08 , secondary 62g05 _ keywords _ : non - asymptotic estimation ; robust risk ; model selection ; sharp oracle inequality ; asymptotic efficiency . |
active galactic nuclei ( agns ) are characterized by strong energy output over a very broad frequency range , from the infrared to hard x - ray .
although their spectral energy distribution ( sed ) has been studied for several decades , one of the main difficulties is the unobservable region between 0.01 kev and 0.2 kev , which is due to galactic and intrinsic photoelectric absorption . as a result
most previous studies have focused mainly either on the uv / optical / infrared side or the soft / hard x - ray side , of the unobservable region . in the optical band ,
information about the agn can be derived from the balmer emission line series e.g. h@xmath0 , h@xmath1 and h@xmath9 , strong forbidden lines e.g. the [ oiii]@xmath84959/5007 doublets , and the optical underlying continuum .
timing analysis of variations in the balmer lines via reverberation mapping study is the most reliable method to estimate agn black hole masses ( e.g. kaspi et al .
2000 ; peterson et al . 2004 ) .
measurement of the [ oiii ] @xmath85007 emission line is used to provide an orientation independent estimate of the agn s central ionizing flux ( kauffmann et al .
2003 ; heckman et al .
2004 ; brinchmann et al . 2004 ; heckman et al .
it is also a principal diagnostic of the properties of the narrow line region ( nlr ) , such as the stellar velocity dispersion ( e.g. nelson & whittle 1995 ; nelson & whittle 1996 ; boroson 2003 ; greene & ho 2005 ) and outflow speed ( e.g. bian , yuan & zhao 2005 ; komossa et al .
the underlying continuum in the optical is believed to be dominated by emission from the accretion disc ( koratkar & blaes 1999 ; jin et al .
2011 , hereafter : paper - i ) and agn s host galaxy ( e.g. hao et al . 2010 ; landt et al .
2011 ) .
the soft to hard x - ray emission is emitted from a compact region .
the hard x - ray power law tail above 2 kev is produced by disc photons which are compton up - scattered by a hot , optically thin electron population ( e.g. haardt & maraschi 1991 ; zdziarski , poutanen & johnson 2000 ) .
however , the origin of the soft x - ray excess below 1 kev is still not clear ( e.g. gierliski & done 2004 ; done et al .
2011 ; also see paper - i ) .
significant absorption of the continuum by ionized gas located near to the central region ( i.e. the ` warm absorber ' ) can further complicate interpretation of the observed x - ray spectra ( reynolds 1997 ; crenshaw & kraemer 1999 ) .
the infrared band especially beyond 2 @xmath10 m is believed to be dominated by emission from the dusty torus , and so provides information about its physical condition ( e.g. rees et al .
1969 ; rieke 1978 ; lebofsky & reike 1980 ) . the observed 1 @xmath10 m minimum in the sed is also a common feature ( sanders et al . 1989 ) .
recently , landt et al . (
2011 ) showed that the continuum luminosity at 1 @xmath10 m can be used to estimate the agn black hole mass .
many previous studies have attempted to construct agn broadband sed by combining multi - waveband observational data to model the emission across the uv - soft x - ray energy gap ( see references in paper - i ) .
whichever method is used to construct the broadband sed , it is found that in almost all cases the sed peaks in the unobservable region ( rest - frame ) which encompasses the ` big blue bump ( bbb ) ' .
one important consequence of this is that the unobservable region contains the largest fraction of an agn s radiated energy ( walter & fink 1993 ; grupe et al .
1998 , 1999 ; paper - i ) . in paper - i
, we presented modelling results for a sample of 51 unobscured nearby type 1 agns .
from the spectral fitting we derived numerous spectral parameters that were not contaminated by x - ray or optical obscuration . * in this paper , we investigate the link between the central ionization flux characterized by the unobscured hard x - ray , and the properties of the various optical emission lines and the continuum .
we conduct this investigation by studying the profiles of various emission lines , and the luminosity correlations between the hard x - ray continuum and various optical emission features .
* narrow line seyfert 1s ( nls1s ) are often considered as a special type of agn whose permitted line width is comparable to other forbidden lines , and their [ oiii ] @xmath85007/h@xmath1 flux ratio is lower than is typical of normal seyfert 1s ( shuder & osterbrock 1981 ; osterbrock & pogge 1985 ) . the conventional definition for a nls1 is a seyfert 1 with h@xmath1 fwhm@xmath112000 km s@xmath12 and [ oiii ] @xmath85007/h@xmath1 @xmath113 ( goodrich 1989 ) .
previous studies focused on such kind of seyfert galaxies and showed many intriguing characteristics .
for example , nls1s were often found to have low black hole masses ( e.g. wang & lu 2001 ; bian & zhao 2004 ; zhou et al .
2006 ; komossa & xu 2007 ; zhu , zhang & tang 2009 , hereafter : zhu09 ) .
* these may be systematically lower than predicted by the m-@xmath13 relation which holds well for bls1s ( grupe & mathur 2004 ; mathur & grupe 2005 ) . *
the nls1s also have high eddington ratios ( boroson 2002 ; komossa 2008 ; paper - i ) .
it is thus proposed that nls1 s central black hole may still be growing ( e.g. mathur , kuraszkiewicz , & czerny 2001 ; komossa & mathur 2001 ; komossa 2008 ) .
in addition , nls1s have softer 2 - 10 kev spectra , lower 2 - 10 kev luminosities , higher @xmath14 values and more energetic bbb ( see paper - i and references therein ) . in this paper , we continue to pay special attention to the nls1 subset in our sample and show how they behave differently from other sources in the cross - correlation study .
this paper is organized as follows .
we first review some most important characteristics of the sample in section 2 , in order to emphasize that our study conducted in the following sections should be related to the most intrinsic properties of agn s bare core .
section 3 will present the ` optical to x - ray correlation spectrum ( oxcs ) ' based on our new ` correlation spectrum technique ( cst ) ' , from which various correlation features related to the hard x - ray luminosity are found in the optical spectrum .
section 4 will study the cross - correlation between different balmer line components and broadband sed components .
section 5 will focus on correlations related to the balmer line equivalent width ( ew ) .
section 6 will study the physical properties of different balmer emission line regions .
[ oiii ] @xmath85007 line s property and its correlation with different sed components are put in section 7 .
summary and conclusion will be made in section 8 . following paper - i
, flat universe model is adopted with the hubble constant h@xmath15 km s@xmath12 mpc@xmath12 , @xmath16 and @xmath17 . | in this second paper in a series of three , we study the properties of the various emission features and underlying continuum in the optical spectra of type 1 active galactic nuclei ( agns ) by using the unobscured hard x - ray emission as a diagnostic .
we introduce the use of the ` correlation spectrum technique ' ( cst ) for the first time .
we use this to show the strength of correlation between the hard x - ray luminosity and each wavelength of the optical spectrum .
this shows that for broad line seyfert 1s all the strong emission lines ( broad component of h@xmath0 and h@xmath1 , [ neiii ] @xmath23869/3967 , [ oi ] @xmath26300/6364 , [ oii ] @xmath23726/3729 , [ oiii ] @xmath24959/5007 ) and the optical underlying continuum all strongly correlate with the hard x - ray emission .
by contrast , the narrow line seyfert 1s show stronger correlation of the optical continuum but weaker correlation in the lines .
a cross - correlation with luminosity between the various balmer line components and the broadband sed components shows that the best correlation exists between the hard x - ray component and broad component ( bc ) of the balmer lines .
such a correlation is weaker for the intermediate ( ic ) and narrow components , which supports the view that broad line region ( blr ) has the closest link with the agn s compact x - ray emission .
the equivalent widths of balmer line ic and bc are found to correlate with l@xmath3 , @xmath4 , balmer line fwhm and black hole mass .
there is a non - linear dependence of the balmer line ic and bc luminosities with l@xmath3 and l@xmath5 , which suggests that a second - order factor such as the ilr and blr covering factors affect the balmer line component luminosities .
the balmer decrement is found to decrease from @xmath65 in the line core to @xmath62 in the extended wings , with mean decrements of 2.1 in blr and 4.8 in ilr .
this suggests different physical conditions in these regions , such as variations in electron density and dust abundance .
the [ oiii ] line is composed of a narrow core together with a blue - shifted component with average outflow velocity of @xmath7 .
the total luminosity of [ oiii ] @xmath85007 shows the best correlation with the luminosity of hard x - ray emission , and so can be used to estimate the intrinsic x - ray luminosity of obscured agns .
we use the cst to show the correlation of the [ oiii ] @xmath85007 luminosity with each wavelength of the full continuum sed .
this shows as before that the [ oiii ] @xmath85007 strongly correlates with power law tail , but also that it correlates almost as strongly with the optical continuum from the disc , but not with the soft excess .
[ firstpage ] accretion , optical / x - ray correlations , active - galaxies : nuclei |
the detrended fluctuation analysis@xcite ( dfa ) is a novel method that has been developed to address the problem of accurately quantifying long range correlations in non - stationary fluctuating signals .
it has been already applied to a multitude of cases including dna@xcite , human motor activity@xcite and gait@xcite , cardiac dynamics@xcite , meteorology@xcite , climate temperature fluctuations@xcite .
traditional methods such as power spectrum and autocorrelation analysis@xcite are not suitable for non - stationary signals@xcite .
dfa is , in short , a modified root - mean - square ( rms ) analysis of a random walk and consists of the following steps : starting with a signal @xmath2 , where @xmath3 , and @xmath4 is the length of the signal , the first step is to integrate @xmath2 and obtain @xmath5\ ] ] where @xmath6 stands for the mean @xmath7 we then divide the profile @xmath8 into boxes of equal length @xmath9 . in each box , we fit @xmath8 using a polynomial function @xmath10 which represents the local trend in that box . ( if a different order @xmath11 of polynomial fit is used , we have a different order dfa-@xmath11 , for example dfa-1 if @xmath11=1 , dfa-2 if @xmath11=2 , etc . )
next , the profile @xmath8 is detrended by substracting the local trend @xmath10 in each box of length @xmath9 : @xmath12 finally , the rms fluctuation for the integrated and detrended signal is calculated @xmath13 ^ 2}.\ ] ] the behavior of @xmath14 over a broad number of scales is obtained by repeating the aforementioned calculation of @xmath14 for varied box length @xmath9 . for scale
invariant signals , we find : @xmath15 where @xmath0 is the scaling exponent .
if @xmath16 , the signal is uncorrelated ( white noise ) , while if @xmath17 the signal is correlated . by employing the dfa method it was found@xcite that long - range correlations exist in the original time - series of the so called seismic electric signals ( ses ) activities .
these are low frequency ( @xmath18hz ) electric signals preceding earthquakes@xcite , the generation of which could be understood in the following context .
a change of pressure affects the thermodynamic parameters for the formation , migration or activation , in general , of defects in solids@xcite . in an ionic solid doped with aliovalent impurities
a number of extrinsic defects is produced,@xcite due to charge compensation , a portion of which is attracted by nearby aliovalent impurities , thus forming electric dipoles that can change their orientation in space through a defect motion@xcite .
hence , pressure variations may affect the thermodynamic parameters of this motion , thus resulting in a decrease or increase@xcite of the relaxation time of these dipoles , i.e. , their ( re)orientation is taking place faster or slower when an external electric field is applied . when the pressure , or the stress in general , reaches a
_ critical _
value@xcite a _ cooperative _ orientation of these electric dipoles occurs , which results in the emission of a transient electric signal .
this may happen in the focal region of a ( future ) earthquake since it is generally accepted that the stress gradually changes there _ before _ rupture .
it has been shown@xcite that ses activities are better distinguished from electric signals emitted from man - made sources , if dfa is applied to a signal after it has been analyzed in a newly introduced time domain , termed natural time @xmath19 . in a time series comprising @xmath4 events
, the natural time @xmath20 serves as an index@xcite for the occurrence of the @xmath21-th event .
the evolution of the pair ( @xmath22 ) is studied@xcite , where @xmath23 denotes a quantity proportional to the _ energy _ released in the @xmath21-th event . for dichotomous signals , which is frequently the case of ses activities , the quantity @xmath23 stands for the duration of the @xmath21-th pulse .
the normalized power spectrum @xmath24 was introduced , where @xmath25 and @xmath26 , @xmath27 ; @xmath28 stands for the _
natural frequency_. in natural time analysis , the properties of @xmath29 or @xmath30 are studied for natural frequencies @xmath28 less than 0.5 , since in this range of @xmath28 , @xmath29 or @xmath30 reduces to a _
characteristic function _ for the probability distribution @xmath31 in the context of probability theory .
when the system enters the _ critical _ stage , the following relation holds@xcite : @xmath32 for @xmath33 , eq.([fasma ] ) leads to@xcite @xmath34 which reflects@xcite that the variance of @xmath19 is given by @xmath35 where @xmath36 .
the entropy @xmath37 in the natural time - domain is defined as@xcite @xmath38 which depends on the sequential order of events@xcite .
it exhibits@xcite concavity , positivity , lesche@xcite stability , and for ses activities ( critical dynamics ) its value is smaller@xcite than the value @xmath39 ) of a `` uniform '' ( u ) distribution ( as defined in refs .
@xcite , e.g. when all @xmath31 are equal or @xmath23 are positive independent and identically distributed random variables of finite variance . in this case ,
@xmath40 and @xmath37 are designated @xmath41 and @xmath42 , respectively . ) .
thus , @xmath43 . the same holds for the value of the entropy obtained@xcite upon considering the time reversal @xmath44 , i.e. , @xmath45 , which is labelled by @xmath46 . in summary , the ses activities , in contrast to the signals produced by man - made electrical sources , when analyzed in natural time exhibit _ infinitely _ ranged temporal correlations@xcite and obey the conditions@xcite : @xmath47 and @xmath48 for major earthquakes , i.e. , with magnitude mw6.5 or larger , ses activities are accompanied@xcite by detectable variations of the magnetic field * b*@xcite .
these variations , which are usually measured by coil magnetometers ( see below ) , have the form of ` spikes ' of alternating sign .
it is therefore of interest to investigate whether these ` spikes ' exhibit long range temporal correlations .
this investigation , which is of major importance since only magnetic field data are usually available in most countries@xcite ( since it is easier to conduct magnetic field measurements than electric field ones ) , is made here in section ii . in the up to date applications of dfa
, long - range correlations have been revealed in ses activities of duration up to a few hours@xcite . during the last few years
, however , experimental results related to some ses activities of appreciably longer duration , i.e. , from several hours to a couple of days , have been collected .
these data now enable the investigation of scaling in a wider range of scales than hitherto known .
this provides an additional scope of the present paper and is carried out in section iii .
a discussion of the results concerning the magnetic and electric data follows in section iv .
finally , our conclusions are presented in section v.
the measurements have been carried out by three dansk coil magnetometers ( dmm ) oriented along the three axes : ew , ns and vertical .
details on the calibration of these magnetometers can be found in the supplementary information of ref.@xcite . in particular , this calibration showed that for periods larger than around half a second , the magnetometers measure@xcite the time derivative @xmath49 of the magnetic field and their output is ` neutralized ' approximately 200ms after the ` arrival ' of a heaviside unit step magnetic variation .
it alternatively means that a signal recorded by these magnetometers should correspond to a magnetic variation that has ` arrived ' at the sensor less than 200ms before the recording .
the data were collected by a campbell 21x datalogger with sampling frequency @xmath50sample / s .
figure [ fig1m](a ) provides an example of simultaneous recordings on april 18 , 1995 at a station located close to ioannina ( ioa ) city in northwestern greece .
variations of both the electric ( e ) and magnetic field are shown .
they were followed by a magnitude mw6.6 earthquake ( according to the centroid moment tensor solutions reported by the united states geological survey ) with an epicenter at [email protected]@xmath51e that occurred almost three weeks later , i.e. , on may 13 , 1995 .
the recordings of the two horizontal magnetometers oriented along the ew- and ns - directions labelled @xmath52 and @xmath53 , respectively , are shown in the lower two channels .
they consist of a series of ` spikes ' of alternating sign as it becomes evident in a 10min excerpt of these recordings depicted in fig.[fig1m](b ) .
these ` spikes ' are superimposed on a background which exhibits almost pseudo - sinusoidal variations of duration much larger than 1s that are induced by frequent small variations of the earth s magnetic field termed magnetotelluric variations ( mt ) .
in addition , the horizontal @xmath54variations were monitored at the same station by measuring the variation @xmath55 of the potential difference between ( pairs of ) electrodes -measuring dipoles- that are grounded at depths of @xmath56 m .
several such dipoles were deployed along the ew- and ns - directions with lengths ( l ) a few to several tens of meters ( short - dipoles ) or a couple of km ( long - dipoles ) ( thus the electric field is given by @xmath57 and is usually measured in mv / km ) .
for example , the following measuring dipoles are shown in the upper three channels of fig.[fig1m](a ) : two short electric dipoles at site ` c ' of ioa station ( see the supplementary information of ref.@xcite as well as ref.@xcite where the selection of site ` c ' has been discussed ) of length 50 m ( labelled e@xmath58-w@xmath58 and n@xmath58-s@xmath58 , respectively ) while the uppermost channel corresponds to a long dipole ( labelled l@xmath59-i ) with length @xmath60 km at an almost ns - direction . as it becomes obvious in fig.[fig1m](b ) , the @xmath61 variations consist of a series of almost rectangular pulses ( cf . the initiation and cessation of each rectangular pulse correspond to two ` spikes ' with opposite sign in the @xmath62 recordings ) .
we now apply dfa to the original time - series of the magnetic field variations and focus our attention on the @xmath52 component where the intensity of ` spikes ' is higher . dividing the time series of length @xmath4 into @xmath63 non - overlapping fragments ,
each of @xmath9 observations , and determining the local trend of the sub - series we find the corresponding @xmath64 versus @xmath65 plot where @xmath66 shown in fig.[fig1mdfa ] .
if we fit the data with two straight lines ( which are also depicted in fig.[fig1mdfa ] ) , the corresponding scaling exponents are @xmath67 and @xmath68 for the short- and long- time lags ( i.e. , smaller than around 12s and larger than @xmath69s ) , respectively .
the crossover occurs at a value ( @xmath70s ) which is roughly equal to the average duration @xmath71s of each electric pulse , corresponding to the interval between two consecutive alternating ` spikes ' .
thus , fig.[fig1mdfa ] shows that at time - lags @xmath72 larger than @xmath73 long - range power law correlations prevail ( since @xmath17 ) , while at shorter time - lags the @xmath0 value is very close to that ( @xmath16 ) of an uncorrelated signal ( white noise ) .
the above findings are strikingly reminiscent of the case of signals with superposed _
random _ spikes studied by chen et al.@xcite .
they reported that for a correlated signal with spikes , they found a crossover from uncorrelated behavior at small scales to correlated behavior at large scales with an exponent close to the exponent of the original stationary signal .
they also investigated the characteristics of the scaling of the spikes only and found that the scaling behavior of the signal with _
random _ spikes is a superposition of the scaling of the signal and the scaling of the spikes .
the case studied by chen et al.@xcite , however , is different from the present case , because the ` spikes ' studied here correspond to the pre - seismic magnetic field variations and hence are _ not _ random ( cf . recall that when applying dfa to the ` durations ' of the electric field rectangular pulses shown in fig.[fig1m](b ) , we found@xcite an exponent around unity ) .
in fig.[fig1 ] the following four long duration ses activities are depicted _ all _ of which have been recorded with sampling frequency @xmath50sample / s at a station close to pirgos ( pir ) city located in western greece ( in this station only electric field variations are continuously monitored with a multitude of measuring dipoles the deployment of which has been described in detail in the supplementary information of ref.@xcite ) .
first , the ses activity on september 17 , 2005 with duration of several hours that preceded the mw6.7 earthquake with an epicenter at [email protected]@xmath51e on january 8 , 2006 .
second , the ses activity that lasted from january 21 until january 26 , 2008 and preceded the mw6.9 earthquake at [email protected]@xmath51e on february 14 , 2008 .
third , the ses activity during the period from february 29 to march 2 , 2008 that was followed@xcite by a mw6.4 earthquake at [email protected]@xmath51e on june 8 , 2008 ( ses activity information is issued _ only _ when the expected magnitude is around 6 units or larger@xcite ) . finally , fig.[fig1](d ) depicts the most recent ses activity of duration of several hours that was recorded on december 12 , 2008 .
the latter ses activity was followed by a mw5.4 earthquake at [email protected]@xmath51e on february 16 , 2009 ( according to athens observatory the magnitude is ms(ath)=ml+0.5=6.0 ; we clarify that predictions are issued _ only _ when the expected magnitude is comparable to 6.0 units or larger ) . note that this earthquake occurred after the initial submission of the present paper ( and its time of occurrence was approached@xcite by means of the procedure developed in ref.@xcite ) . here ,
we analyze , as an example , the long duration ses activity that lasted from february 29 until march 2 , 2008(fig.[fig1](c ) ) .
the time - series of this electrical disturbance , which is not of obvious dichotomous nature , is depicted in the upper channel ( labelled `` a '' ) of fig.[figfeb ] .
the signal , comprising a number of pulses , is superimposed on a background which exhibits frequent small mt variations .
the latter are simultaneously recorded at _ all _ measuring sites , in contrast to the ses activities that solely appear at a restricted number of sites depending on the epicentral region of the future earthquake@xcite .
this difference provides a key for their distinction . in order to separate the background
, we proceed into the following steps : first , we make use of another measuring dipole of the same station , i.e. , the channel labelled `` b '' in fig.[figfeb ] , which has not recorded the signal but does show the mt pseudo - sinusoidal variations .
second , since the sampling rate of the measurements is 1sample / s , we now compute the increments every 1s of the two time - series of channels `` a '' and `` b '' . assuming the `` 1s increments '' of channel `` a '' lying along the x - axis and those of `` b '' along the y - axis , we plot in the middle panel of fig.[figfeb ] ( labeled `` c '' ) the angle of the resulting vector with dots .
when a non - mt variation ( e.g. , a dichotomous pulse ) appears ( disappears ) in channel `` a '' , this angle in `` c '' abruptly changes to 0@xmath51 ( @xmath74 ) .
thus , the dots in panel `` c '' mark such changes . in other words , an increased density of dots ( dark regions ) around 0@xmath51 or @xmath75 marks the occurrence of these pulses , on which we should focus . to this end , we plot in channel `` d '' of fig.[figfeb ] the residual of a linear least squares fit of channel `` a '' with respect to channel `` b '' . comparing channel `` d '' with channel `` a '' , we notice a significant reduction of the mt background but not of the signal .
the small variations of the mt background that still remain in `` d '' , are now marked by the light blue line , which when removed result in channel `` e '' of fig.[figfeb ] .
hence , channel `` e '' solely contains the electric field variations that precede rupture .
this channel provides the time - series which should now be analyzed .
the dfa analysis ( in conventional time ) of the time - series of channel `` e '' of fig.[figfeb ] is shown in fig.[fig3 ] .
it reveals an almost linear @xmath76 vs @xmath77 plot ( where @xmath78 ) with an exponent @xmath1 practically _ over all scales available _ ( approximately five orders of magnitude ) .
note , that this value of the exponent remains the same irrespective if we apply dfa-1 , dfa-2 or dfa-3 .
this result is in agreement with the results obtained@xcite for ses activities of _ shorter _ duration . for the classification of this signal ,
i.e. , to distinguish whether it is a true ses activity or a man - made electric signal , we now proceed to its analysis in natural time . in order to define ( @xmath22 )
the individual pulses of the signal depicted in channel `` e '' of fig.[figfeb ] have to be identified .
a pulse starts , of course , when the amplitude exceeds a given threshold and ends when the amplitude falls below it .
moreover , since the signal is not obviously dichotomous , instead of finding the duration of each pulse , one should sum the `` instantaneous power '' during the pulse duration in order to find @xmath23 , as already mentioned . to this end , we plot in fig.[febhis ] the histogram of the instantaneous power p of channel `` e '' of fig.[figfeb ] , computed by squaring its amplitude .
an inspection of this figure reveals a bimodal feature which signifies the periods of inactivity ( p@xmath79v@xmath80hz ) and activity ( p@xmath81v@xmath80hz ) in channel `` e '' of fig.[figfeb ] . in order to find @xmath23 , we focus on the periods of activity and select the power threshold p@xmath82 around the second peak of the histogram in fig.[febhis ] .
let us consider , for example , the case of p@xmath82=1400 @xmath83v@xmath80hz . in fig.[febexa](a ) , we depict the instantaneous power p of the signal in channel `` e '' of fig.[figfeb ] .
starting from the beginning of the signal , we compare p with p@xmath82 and when p exceeds p@xmath82 we start summing up the p values until p falls below p@xmath82 for the first time , @xmath84 .
the resulting sum corresponds to @xmath85 .
this procedure is repeated until p falls below p@xmath82 for the second time , @xmath86 , and the new sum represents @xmath87 , etc .
this leads to the natural time representation depicted in fig.[febexa](b ) .
the result depends , of course , on the proper selection of p@xmath82 .
the latter can be verified by checking whether a small change of p@xmath82 around the second peak of the histogram , leads to a natural time representation resulting in approximately the same values of the parameters @xmath88 and @xmath46 . by randomly selecting p@xmath82 in the range 500 to 2000
@xmath83v@xmath80hz , we obtain that the number of pulses in channel `` e '' of fig.[figfeb ] is @xmath89 with @xmath90 , @xmath91 and @xmath92 .
when p@xmath82 ranges between 1000 to 1500 @xmath83v@xmath80hz , the corresponding values are @xmath93 with @xmath94 , @xmath95 and @xmath96 .
thus , we observe that irrespective of the p@xmath82 value chosen , the parameters @xmath88 and @xmath46 obey the conditions ( [ eq1 ] ) and ( [ eq2 ] ) for the classification of this signal as ses activity .
the same holds for a non - dichotomous signal recorded on march 28 , 2009 at keratea station located close to athens ( fig.[fig8 ] ) .
to approach the occurrence time of the impending event , the procedure developed in ref.@xcite has been employed for the seismicity within the area n@xmath97e@xmath98 .
the results , when considering the seismicity until early in the morning on june 19 , 2009 , are shown in fig.[fig9 ] .
they show that a local maximum of the probability@xcite prob(@xmath40 ) of the seismicity is observed at @xmath99 upon the occurrence of the following events : ( a ) an m@xmath100=3.2 event at 12:42 ut on june 18 with epicenter at 38.7n 23.0e and an m@xmath100=3.2 event at 03:44 ut on june 19 at 37.9n 23.0e if we take a magnitude threshold m@xmath82=3.0 ( b ) when taking m@xmath82=3.1 , the maximum occurs again upon the occurrence of the aforementioned event at 12:42 on june 18 ( c ) when increasing the threshold at m@xmath82=3.2 , the maximum occurs somewhat earlier , i.e. , upon the occurrence of the m@xmath100=3.5 event at 08:26 ut on june 18 at 38.7n 23.0e .
let us now consider a larger area , i.e. , n@xmath101e@xmath102 , and repeat the same calculation .
the results are shown in fig.[fig10 ] .
they reveal again the occurrence of a local maximum of prob(@xmath40 ) at @xmath99 upon the occurrence of the above mentioned m@xmath100=3.2 event at 12:42 ut on june 18 when considering either m@xmath82=3.2 ( fig.[fig10](c ) ) or m@xmath82=3.1 ( fig.[fig10](b ) ) ( for the latter case a maximum also occurs at the last event shown in the figure with m@xmath100=3.2 at 03:44 ut on june 19 ) .
the maximum appears upon the occurrence of a m@xmath100=3.0 at 23:48 ut on june 18 when decreasing the threshold at m@xmath82=3.0 . in summary ,
the results exhibit both magnitude- and spatial- invariance , which is characteristic when approaching the critical point .
_ note added on july 21 , 2009 . _ repeating the aforementioned calculation for the area n@xmath101e@xmath102 , we find the results depicted in fig.[fig11](a ) for m@xmath82=2.8 .
an inspection of this figure reveals a local maximum of prob(@xmath40 ) at @xmath99 upon the occurrence of a m@xmath100=3.0 at 21:10 ut on july 19 , 2009 with an epicenter at 38.1n 22.7e ( the same result is found for m@xmath82=2.9 . )
interestingly , this local maximum turns to be the prominent one when studying the seismicity in the same area ( for the same threshold m@xmath82=2.8 ) since june 21 , 2009 , see fig.[fig11](b ) ; the same holds for m@xmath82=3.0 .
furthermore , the latter study was made for the smaller area n@xmath97e@xmath98 for m@xmath82= 2.6 , 2.7 and 3.0 and showed a local maximum of prob(@xmath40 ) at @xmath99 upon the occurrence of the same event ( i.e. , at 21:10 ut on july 19 , 2009 ) .
the investigation still continues in order to clarify whether the latter local maximum will eventually turn to become a prominent one .
_ note added on september 5 , 2009 . _ at 8:17 ut on september 2 , 2009 , a m@xmath100=4.3 earthquake occurred almost at the center of the predicted area n@xmath97e@xmath98 .
the continuation of the investigation of the previous paragraph in the smaller area n@xmath97e@xmath98 showed a prominent maximum of prob(@xmath40 ) at @xmath99 ( see fig.[fig12 ] ( a ) ) upon the occurrence of a m@xmath100=3.3 event at 09:35 ut on september 2 , 2009 , with an epicenter at 38.1n 23.3e .
this was found to hold for the aforementioned larger area ( see fig.[fig12 ] ( b ) ) as well , thus revealing that the _ criticality _ condition -as developed in ref . -
is obeyed . _ note added on september 19 , 2009 . _
the previous note was followed by three earthquakes of magnitude 4.4 , 4.7 and 4.8 that occurred at 21:15 ut september 9 , 06:04 ut september 10 and 07:12
ut september 16 , 2009 with epicenters at 38.7n 22.8e , 38.3n 24.1e and 39.0n 22.2e , respectively .
the continuation of the natural time investigation of the seismicity for @xmath103=3.0 in the smaller area n@xmath97e@xmath98 revealed a prominent maximum prob(@xmath40 ) at @xmath99 ( see fig.[fig13 ] ( a ) ) upon the occurrence of a m@xmath100=3.0 event at 01:55 ut on september 18 , 2009 , with epicenter at 38.7n22.8e . since the same was found to be true for the aforementioned larger area as well ( see fig.[fig13 ] ( b ) ) , the investigation is now extended to additional magnitude thresholds in order to clarify whether ( beyond spatial and ) magnitude threshold invariance holds , thus identifying that the system approaches the _ critical _ point . _ note added on october 13 , 2009 . _
the previous note was actually followed by an earthquake of magnitude 4.6 that occurred at 22:27 ut on october 2 , 2009 with an epicenter at 38.3n 22.9e .
the natural time investigation of the seismicity still continues with the following results .
first , in the smaller area n@xmath97e@xmath98 : for @xmath103=2.7 , the criticality condition started to be approached since october 7 and the _ most prominent _ maximum of prob(@xmath40 ) at @xmath99 was observed upon the occurrence of the m@xmath100=2.7 event at 01:56 ut on october 11 , 2009 .
the same behavior was found for @xmath103=2.8 ( but the last event for this case occurred at 08:15ut on october 10 , 2009 at 38.2n 23.1e ) .
second , in the larger area n@xmath101e@xmath102 : for both thresholds @xmath103=2.7 and @xmath103=2.8 , the criticality condition was also started to be approached since 20:18 ut on october 7 , 2009 .
the _ most prominent _ maximum of prob(@xmath40 ) at @xmath99 for @xmath103=2.7 have been observed upon the occurrence of the aforementioned m@xmath100=2.7 event at 01:56 ut on october 11 , 2009 , while for @xmath103=2.8 it was observed at 4:51 ut on october 9 , 2009 , upon the occurrence of the m@xmath100=2.9 event at 38.1e 23.2e .
interestingly , the results for the larger area are almost identical if the investigation starts at 15:00 ut on july 31 , 2009 .
_ note added on november 14 , 2009 .
_ meanwhile , two ses activities were recorded at pat on october 24 and november 11 , 2009 , see fig.[fig14 ] .
hence , the _ ongoing _ study ( fig.[fig15 ] ) of the previous note should be accompanied by a complementary study , focused on the determination of the occurrence time of the impending mainshock , that also considers the seismicity in the western part ( i.e. , n@xmath104e@xmath105 ) of the pat selectivity map ( see area a in fig.8 of ref .
such a complementary study , when taking into account the seismic data until early in the morning on november 13 , 2009 , leads to the results depicted in fig.[fig16 ] for @xmath103=3.1 . a local maximum of prob(@xmath40 ) at @xmath99 was observed ( see the rightmost arrow in fig.[fig16 ] ) upon the occurrence of m@xmath100=3.2 event at 18:55 ut on november 2 , 2009 ( that was followed by a m@xmath100=5.6 eq at 05:25 ut on november 3 , 2009 , with an epicenter at @xmath106n @xmath107e ) . in addition , the _ most prominent _ maximum in fig.[fig16 ] of prob(@xmath40 ) at @xmath99 is observed later , i.e. , upon the occurrence of a m@xmath100=3.1 event at 01:42 ut on november 13 , 2009 ( see the leftmost arrow in fig.[fig16 ] ) .
this , which has been also checked for @xmath103=3.0 , provides evidence that the critical point is approached .
calculations are now carried out to further investigate the magnitude threshold invariance of this behavior .
_ note added on november 27 , 2009 . _ actually , after the previous note the following two earthquakes occurred at 18:30 ut on november 16 , 2009 and at 04:29 ut on november 19 , 2009 with magnitudes ( m@xmath59(ath ) ) 4.3 and 4.4 and epicenters at @xmath108n @xmath109e and @xmath110n @xmath111e , respectively , that both lie within the pat selectivity map , i.e , n@xmath104e@xmath112 .
the continuation of the two studies mentioned in the previous note , i.e. , in the pat selectivity map as well as in the area n@xmath101e@xmath113 , both show that the _ most prominent _ maximum of prob(@xmath40 ) at @xmath99 was observed upon the occurrence of the m@xmath100=3.2 event at 19:31 ut on november 25 , 2009 , thus indicating that the system is near the critical point . _ note added on february 9 , 2010 . _ two earthquakes with magnitudes m@xmath59(ath)(@xmath114)=5.7 and 5.6 occurred at @xmath108n @xmath109e ( and hence inside the expected area ) at 15:56 ut on january 18 and at 00:46 ut on january 22 , 2010 ( after observing a maximum of prob(@xmath40 ) versus @xmath40 at @xmath40=0.070 for @xmath103=3.1 upon the occurrence of the @xmath115=3.2 event at 15:56 ut on january 15 , 2010 ) .
these two magnitudes are consistent with the condition m@xmath59(ath)@xmath1166.0 ( under which ses information are publicized ) if a reasonable@xcite experimental error 3@xmath117=0.7 is accepted
. however , since the amplitudes of all the ses activities recorded could warrant an impending mainshock with magnitude exceeding m@xmath59(ath)=6.0 , and in order to determine the occurrence time of such a mainshock , if any , the natural time analysis of seismicity still continues .
this analysis starting at 04:00 ut on december 27 , 2009 , revealed that in the area n@xmath118e@xmath119 ( almost 200 km @xmath120 100 km ) the criticality condition ( i.e. , prob(@xmath40 ) versus @xmath40 maximizes at @xmath40=0.070 ) for @xmath103=3.5 is obeyed during the last three events of fig.[fig17 ] , the latest of which ( that exhibits the strongest maximum ) occurred at 10:18 ut on february 7 , 2010 at @xmath121n @xmath122e with @xmath115=4.3 .
this has been checked to be valid for broader areas as well ( spatial invariance ) , but has not yet been found to be confirmed ( until early in the morning on february 8 , 2010 ) for lower magnitude thresholds .
the investigation of the latter is still in progress , because the following might happen : due to coarse - graining the criticality condition might have been obeyed at a larger time - window for large magnitude thresholds , while for smaller thresholds the time - window may become smaller thus achieving a better accuracy in the determination of the occurrence time of the impending earthquake .
in general , electric field variations are interconnected with the magnetic field ones through maxwell equations .
thus , it is intuitively expected that when the former exhibit long range correlations the same should hold for the latter .
this expectation is consistent with the present findings which show that , at long time - lags , the original time - series of both electric- and magnetic field variations preceding rupture exhibit dfa - exponents close to unity .
the above can be verified when data of both electric- and magnetic - field variations are simultaneously available .
this was the case of the data presented in fig.[fig1 m ] .
in several occasions , however , as mentioned in section i , only magnetic field data exist , in view of the fact that it is easier to conduct magnetic field measurements compared to the electric field ones . when using coil magnetometers , the magnetic field variations have the form of series of ` spikes ' . whenever the amplitude of these ` spikes ' significantly exceed the pseudo - sinusoidal variations of the mt background , as in the case of b@xmath123 in fig.[fig1 m ] , a direct application of dfa ( see fig.[fig1mdfa ] )
elucidates the long range correlations in the magnetic field variations preceding rupture . on the other hand , when considerable pseudo - sinusoidal mt variations are superimposed ,
a direct application of dfa is not advisable .
one must first subtract the mt variations ( following a procedure similar to that used in the electric field data in fig.[figfeb ] ) and then apply dfa .
the preceding paragraph refers to the analysis of the signal in conventional time . as already shown in ref.@xcite , natural time analysis ,
has the privilege that allows the distinction between true ses activities and man made signals .
this type of analysis , however , demands the knowledge of the energy released during each consecutive event .
the determination of this energy is easier to conduct in the case of electric field variations ( this is so , because coil magnetometers , as mentioned in section ii , act as db / dt detectors ) .
when these variations are of clear dichotomous nature , the energy release is proportional to the duration of each pulse@xcite . on the other hand , in absence of an obvious dichotomous nature an analysis of the ` instantaneous power ' similar to that presented in the last paragraph of section iii should be carried out to determine the parameters @xmath40 , s and s@xmath124 in natural time .
first , dfa was used as a scaling analysis method to investigate long - range correlations in the original time series of the magnetic field variations that precede rupture and have the form of ` spikes ' of alternating sign .
we find a scaling exponent @xmath0 close to 0.9 for time - lags larger than the average time interval @xmath73 between consecutive ` spikes ' , while at shorter time - lags the exponent is close to 0.5 thus corresponding to uncorrelated behavior .
second , using electric field data of long duration ses activities ( i.e. , from several hours to a couple of days ) recorded during the last few years , dfa reveals a scale invariant feature with an exponent @xmath1 , over _ all _ scales available ( approximately five orders of magnitude ) .
( computed by squaring its amplitude ) . the solid line parallel to the x - axis marks an example of a threshold p@xmath82(=1400 @xmath83v@xmath80hz ) chosen .
( b ) the resulting representation of the signal depicted in channel `` e '' of fig.[figfeb ] in natural time , when considering p@xmath82=1400 @xmath83v@xmath80hz . ] ) of the seismicity within the area n@xmath97e@xmath98 subsequent to the ses activity recorded at ker on march 28 , 2009 , when considering the following magnitude thresholds : m@xmath82=3.0 ( a ) , 3.1 ( b ) and 3.2 ( c ) . ] ) vs @xmath40 upon considering the seismicity ( m@xmath82=3.1 ) until 9:35 ut on september 2 , 2009 : panel ( a ) corresponds to the area n@xmath97e@xmath98 , whereas panel ( b ) to the larger area n@xmath101e@xmath102 . ] ) vs @xmath40 upon considering the seismicity ( m@xmath82=3.0 ) until 1:55 ut on september 18 , 2009 : panel ( a ) corresponds to the area n@xmath97e@xmath98 , whereas panel ( b ) to the larger area n@xmath101e@xmath102 . ] ) of the seismicity within the area n@xmath104e@xmath112 subsequent to the ses activity recorded at pat on october 24 , 2009 , when considering m@xmath82=3.1 .
the seismic data until early in the morning on november 13 , 2009 , have been used . ] | magnetic field variations are detected before rupture in the form of ` spikes ' of alternating sign .
the distinction of these ` spikes ' from random noise is of major practical importance , since it is easier to conduct magnetic field measurements than electric field ones .
applying detrended fluctuation analysis ( dfa ) , these ` spikes ' look to be random at short time - lags . on the other hand ,
long range correlations prevail at time - lags larger than the average time interval between consecutive ` spikes ' with a scaling exponent @xmath0 around 0.9 . in addition
, dfa is applied to recent preseismic electric field variations of long duration ( several hours to a couple of days ) and reveals a scale invariant feature with an exponent @xmath1 over all scales available ( around five orders of magnitude ) .
* keywords : * detrended fluctuation analysis , complex systems , scale invariance * many physical and biological complex systems exhibit scale - invariant features characterized by long - range power - law correlations , which are often difficult to quantify due to the presence of erratic fluctuations , heterogeneity and nonstationarity embedded in the emitted signals . here
, we focus on different types of nonstationarities such as random spikes and pseudo - sinusoidal trends , that may affect the long - range correlation properties of signals that precede rupture
. since these nonstationarities may either be epiphenomena of external conditions or may arise from the intrinsic dynamics of the system , it is crucial to distinguish their origin .
this is attempted in the present paper for both the magnetic and the electric field variations that appear before rupture by employing the detrended fluctuation analysis ( dfa ) as a scaling method to quantify long - range temporal correlations . in particular , for the magnetic field variations which have usually the form of ` spikes ' of alternating sign , we find that at short time scales they look to be random ( thus may then be confused with random noise ) , but at larger scales long - range correlations prevail .
as for the electric field variations , that are usually superimposed on a pseudo - sinusoidal background , upon using the longest time series available to date ( i.e. , with duration up to a couple of days ) , a scale - invariant feature over five orders of magnitude emerges with an exponent close to unity . * |
the delta - kicked rotor ( dkr ) is a nonlinear dynamical system which exhibits starkly contrasting behaviour in classical and quantum regimes . for example , the well - known chaotic diffusion exhibited by the classical dkr is completely suppressed by coherence effects ( dynamical localization @xcite ) in the quantum regime .
the field of quantum chaos ( see , e.g. , @xcite ) , which brings together the study of classically chaotic systems and their quantum mechanical analogues , is relatively new , especially given the maturity of the two parent fields . indeed , most of the progress in quantum chaos has been made only during the last quarter century . from an experimental point of view
, the field of quantum chaos received a major boost in the 1990 s with the use of ultra - cold atoms and pulsed standing - wave laser fields to realize a near - ideal quantum version of the delta - kicked rotor @xcite . at the low temperatures achievable using laser cooling , quantum behavior of the atomic particles becomes manifest , and optical manipulation of the atoms offers unprecedented control over the forces they experience .
furthermore , one can identify for this system an `` effective planck s constant '' , @xmath0 , which is directly proportional to the period of the laser pulsing and can therefore be adjusted to , in a sense , make the system `` more '' or `` less '' quantum mechanical . this so - called `` atom - optical kicked rotor '' has since been the subject of intense investigation by a number of experimental groups in a variety of different contexts ( see , e.g. , @xcite ) .
these investigations have in general focused on the long - time behavior of the system ; that is , on the properties of the system after a relatively large number ( at least several tens ) of kicks .
for these purposes , widths of the initial momentum distributions on the order of a few photon recoils ( i.e. , @xmath1 , where @xmath2 is the wave number of the laser light ) have sufficed , since the effects under investigation have typically not been dependent on starting from extremely precise initial states of the atomic motion . however , for detailed investigations of early - time behavior of the kicked rotor and of certain uniquely quantum mechanical phenomena , it is extremely desirable , or even essential , to have yet more control over the initial state .
while most work on the kicked rotor has concentrated on differences between classical and quantum behavior in the late - time regime early - time behavior was investigated by shepelyansky @xcite .
this work showed that significant differences also exist in _ initial _ diffusion rates , and the initial quantum diffusion rate exhibits a strong dependence on the effective planck s constant @xmath0 .
also this dependence can also lead to signatures in the late - time energies @xcite and diffusion rates @xcite , but for a direct study of initial rates a very narrow initial momentum distribution is vital from the point of view of being able to resolve small energy changes as a function of the system parameters after just a small number of kicks .
furthermore , more recent theoretical work @xcite has revealed that , with a very narrow initial momentum distribution , initial diffusion rates exhibit an even richer structure ( as a function of @xmath0 ) than that predicted by shepelyansky , whose calculations assumed broad ( uniform ) initial conditions . at certain specific values of @xmath0 in particular , where @xmath0 is a rational multiple of @xmath3 quite remarkable phenomena can occur in the form of so - called `` quantum resonances '' and `` anti - resonances '' @xcite .
these phenomena require particular initial momentum states which evolve in such a way that during the free evolution period in between kicks the different components of the state vector of the system experience either identical phase shifts , or a phase shift that alternates in sign from one momentum component to the next . where the phase factor is identical for all components , ballistic energy growth is observed ( quantum resonance ) .
where the phase factor alternates in sign , the system returns identically to its initial state after every second kick ( quantum anti - resonance ) . with a broad initial momentum distribution such resonance and anti - resonance behavior
can still be observed experimentally in the atom - optical kicked rotor @xcite , but it is far less pronounced than in the ideal case of initial momentum eigenstates . with a dilute atomic bose - einstein condensate however , it is possible to realize an initial state that is , to all intents and purposes , a momentum eigenstate and therefore allows a much clearer investigation of these phenomena . in this work ,
we follow the suggestion of daley and parkins in ref .
@xcite and investigate the early - time behavior of the atom - optical kicked rotor using a bose - einstein condensate to provide a very narrow and precise initial momentum state of the atoms .
we focus on investigating the energy as a function of kick number for specific values of effective planck s constant , and our results demonstrate the behaviors predicted theoretically . note that our work complements recent experimental studies of atom - optical versions of ( classically chaotic ) nonlinear dynamical systems which also make use of extremely narrow atomic momentum distributions to provide very precise initial conditions @xcite .
the basic model describing the atom - optical kicked rotor has been described by a number of authors , and here we briefly summarize this following the notation of ref .
a cold atomic sample interacts with a standing wave of laser light with frequency @xmath4 , far - detuned from resonance .
the laser is pulsed with period @xmath5 and pulse profile @xmath6 . due to the large detuning
, the internal atomic dynamics can be eliminated and the hamiltonian determining the motion of the atoms can be written as @xmath7 where @xmath8 and @xmath9 are the atomic position and momentum operators , respectively , and @xmath10 = @xmath11/@xmath12 is the effective potential strength , with @xmath13 the ( single - beam ) resonant rabi frequency and @xmath12 the detuning from atomic resonance .
we can rewrite eq .
( [ eqn : ham1 ] ) as a scaled dimensionless hamiltonian in the form @xmath14 which is standard for the kicked rotor system .
the position operator is defined by @xmath15@xmath8 , the momentum operator @xmath16@xmath5@xmath9/@xmath17 , the scaled time is @xmath18 , and @xmath19 = ( 4@xmath20@xmath5@xmath21/@xmath17 ) @xmath22 .
the classical stochasticity parameter ( or kick strength ) is given by @xmath23@xmath24@xmath5@xmath25 , where @xmath25 is the pulse length and @xmath24 = @xmath26@xmath20/2@xmath17 is the recoil frequency . in this work @xmath27
is taken to represent a square pulse , i.e. @xmath28 = 1 for 0 @xmath29 @xmath30 @xmath29 @xmath31 , where @xmath32 = @xmath25/@xmath5 , in which case @xmath33 . in these scaled units ,
we have [ @xmath34 = @xmath35 , with @xmath0 = 8@xmath36 .
thus the quantum nature of the system is reflected by an effective planck s constant , @xmath0 , which changes as we adjust the pulsing period @xmath5 . in the case of the @xmath12-kicked rotor ( i.e. , @xmath37 , @xmath38 )
the evolution of the system can be represented by the quantized standard map , @xmath39 where @xmath40 and @xmath41 , with the values recorded immediately after the kick at @xmath42 . in this version of the standard map
, the first kick occurs at @xmath43 . in our experiment ,
an image of the atomic cloud allows us to determine the momentum distribution and hence the kinetic energy after a set number of kicks . with the change in kinetic energy between consecutive kicks
we then determine the momentum diffusion rate @xmath44 an analytical investigation of early - time quantum diffusion rates in the dkr was made by shepelyansky @xcite , whose calculations assumed uniform ( broad ) initial position and momentum distributions and involved the evaluation of quantum correlation functions of the form @xmath45_+\rangle$ ] for @xmath46 . from a sum of such correlation functions an estimate of the initial quantum diffusion rate was obtained , which predicts ( broad ) peaks , or resonances , as a function of @xmath0 . in particular ,
prominent peaks appear in the diffusion rate where @xmath0 is an integer multiple of @xmath47 , together with other maxima whose number and positions ( with respect to @xmath0 ) vary with kick strength @xmath48 . more recently , daley and parkins @xcite re - examined the early - time diffusion rates for very narrow initial momentum distributions , as is appropriate to atom - optical experiments with bose - einstein condensates .
they find an even more complex and interesting structure in the diffusion rates as a function of @xmath0 , as exemplified by their result for @xmath49 , which takes the form @xmath50\nonumber \\ & & -\kappa j_1(k_q)[\sigma_\rho^2 { { \rm e}}^{-\sigma_\rho^2/2 } \cos(\bar{\rho}_0)+\bar{\rho}_0{{\rm e}}^{-\sigma_\rho^2/2 } \sin(\bar{\rho}_0)]\nonumber \\ & & + \frac{\kappa^2}{2 } [ j_0(k_q)-j_2(k_q)]\cos({\mathchar'26\mkern-9muk}/2 ) { { \rm e}}^{-\sigma_\rho^2/2 } \cos(\bar{\rho}_0 ) , \label{eqn : diffrate2kicks}\end{aligned}\ ] ] where @xmath51 , @xmath52 , and a gaussian initial momentum distribution of mean @xmath53 and variance @xmath54 is assumed .
the rates @xmath55 exhibit still more structure than for @xmath49 , but were computed numerically using wave function simulations @xcite ( which allow for finite pulse widths and atomic spontaneous emission ) .
note that for a broad initial momentum distribution @xmath49 is independent of @xmath0 and given simply by @xmath56 .
the phenomena of quantum resonances and anti - resonances occur for particular values of @xmath0 , and in their `` idealized '' forms require very specific initial conditions .
take , for example , an initial momentum eigenstate @xmath57 , where @xmath58 is an integer . through the kicking process , this state couples only to eigenstates @xmath59 , where @xmath60 is also an integer .
if @xmath0 is an _ even _ multiple of @xmath47 , then it is straightforward to show that @xmath61 , i.e. , during the free evolution period in between kicks the state vector of the system is unchanged .
this leads to ballistic energy growth ( i.e. , the energy depends quadratically on kick number ) and dynamical localization does not occur .
this is known as a quantum resonance and is related to the talbot effect in wave optics @xcite .
in contrast , if @xmath0 is an _ odd _ multiple of @xmath47 , then @xmath61 if @xmath62 is even , and @xmath63 if @xmath62 is odd . in this case , the system returns identically to its initial state , @xmath57 , after every second kick .
this effect is known as a quantum anti - resonance . other features ( i.e. , peaks or dips ) appearing in the diffusion rates as a function of @xmath0 ( see fig . [
fig : e1e2vskbar ] ) can also be related to behavior such as that described above . however , unlike quantum resonances and anti - resonances , the @xmath0 values for which these features occur depend on the kick strength @xmath48 in a nontrivial manner @xcite .
we focus on the behavior of the energy as a function of kick number and of the effective planck s constant , @xmath0 .
the kick strength is fixed by keeping the laser intensity constant and varying both the pulsing period @xmath5 ( which is proportional to the effective planck s constant ) and pulse length @xmath25 , such that the product @xmath64 is a constant .
the experiment is performed with a bose condensate of approximately @xmath65 @xmath66rb atoms in the @xmath67 = 2 , @xmath68 = 2 hyperfine state .
the condensate is formed ( as described in ref .
@xcite , but with minor modifications @xcite ) in a time - averaged orbiting potential trap with harmonic oscillation frequencies of @xmath69/2@xmath70 = 71 hz radially and @xmath71/2@xmath72 = 201 hz axially .
after radio frequency evaporation to form a bose condensate , the trap is relaxed to @xmath69/2@xmath72 = 32 hz and @xmath71/2@xmath70 = 91 hz over a period of 200 ms .
once formed , the condensate is released from the trap and exposed to pulsed optical standing waves after 1.7 ms of free expansion ( so that condensate mean - field effects can essentially be ignored @xcite ) . for our parameters ,
the momentum fwhm of the bose condensate is 0.03@xmath73 .
these standing waves are generated by two counterpropagating laser beams with parallel linear polarizations , derived from a single beam which is detuned 1.48 ghz from the 5s@xmath74 , @xmath67 = 2 @xmath75 @xmath76 , @xmath77@xmath78 = 3 transition .
for a chosen @xmath48 value , the pulse period is scanned from 21.12 @xmath79s ( @xmath80 ) to the quantum resonance at @xmath81 = 66.38 @xmath79s ( @xmath82 ) .
consequently the pulse length is varied from 1.25 @xmath83s to 400 ns .
there are limitations on the precise values of @xmath5 and @xmath84 caused by the incremental stepsize values of the pulse generator .
the laser detuning and intensity were chosen to give the desired kicking strength while maintaining a negligible spontaneous emission rate ( @xmath29 34 s@xmath85 ) .
the momentum distribution is determined using time - of - flight ( 4 ms of free expansion ) and absorption imaging of the atomic sample . when the condensate is released from the magnetic trap it receives an ( unwanted ) impulse corresponding approximately to [email protected] mm / s , as determined by bragg scattering @xcite . in order to apply a standing wave which is stationary with respect to the condensate the frequency difference between the two laser beams
is adjusted to remove the relative motion .
a double - pass acoustic - optic modulator is used in each beam for altering its frequency and switching the optical potential on and off .
the kicks given to the atomic sample populate momentum classes separated by @xmath87 ( as shown in fig .
[ fig : exptdataqr ] ) .
following the kicks , we determined the kinetic energy of the atomic sample , which entailed counting the number of atoms in each momentum state and multiplying by the energy of that state .
subsequently , to obtain the average kinetic energy per atom this energy value was divided by @xmath88 and the total number of atoms , and then multiplied by @xmath89 ( where @xmath17 is the mass of the rubidium atom ) .
this gave us an overall average energy in dimensionless units , @xmath90 .
[ -30cm][l ] [ -30cm][l ] [ -30cm][l ] [ -30cm][l ] [ -50cm][l]1 kick [ -50cm][l]2 kicks [ -50cm][l]3 kicks [ -50cm][l]4 kicks for 1(top ) to 4(bottom ) kicks ( b ) corresponding momentum distributions .
this particular case illustrates enhanced energy growth close to the quantum resonance at @xmath91 . ] for 1(top ) to 4(bottom ) kicks ( b ) corresponding momentum distributions .
this particular case illustrates enhanced energy growth close to the quantum resonance at @xmath91 . ]
( 0,0 ) ( -65,70)(a ) ( 60,70)(b )
we begin by highlighting the rich initial diffusion behavior of our kicked rotor system with experimental data for the energy after one and two kicks , @xmath92 and @xmath93 , as a function of the effective planck s constant , @xmath0 , for a fixed value of the kicking strength @xmath48 .
this data , plotted in fig .
[ fig : e1e2vskbar ] , confirms the prediction ref .
@xcite of a uniform value for @xmath94 , but a strong dependence of @xmath95 on @xmath0 , given a sufficiently narrow initial momentum distribution .
similar behavior is obtained for the higher kicking strength , as we shall see later .
the energy after one kick is given theoretically by @xmath96 , which allows us to infer a value @xmath97 ( @xmath98 ) from the experimental data for the lower ( higher ) kicking strength .
these values are consistent with those calculated from the laser intensity , detuning and pulse details .
note that for the one - kick data the duration of the kicking pulse was chosen for each @xmath0 value to match that used for sequences of two or more kicks , so the one - kick energies do in fact correspond to different experimental conditions .
however , in dimensionless units the energy after one kick ( @xmath99 ) is predicted to be constant as a function of @xmath0 . before continuing
, we note that a related experiment in which a released bose - einstein condensate of sodium atoms was subjected to a sequence of two standing - wave laser pulses separated by a varying time delay was recently reported by deng _
et al_. @xcite .
while effects related to those displayed above could be inferred from their results , their work was not set in the context of the quantum kicked rotor
if one does so , then the experiment they performed corresponds to varying @xmath0 and @xmath48 simultaneously ( since both parameters are proportional to the kicking period @xmath5 , and the laser intensity and @xmath84 were fixed for their measurements ) .
the energy as a function of kick number exhibits a complex variety of behaviors as the effective planck s constant is varied ( for fixed @xmath48 ) .
examples are plotted in figs .
[ fig : evsn15 ] and [ fig : evsn25 ] , where experimental energies for up to four kicks are shown .
close to @xmath100 ( figs .
[ fig : evsn15](a ) and [ fig : evsn25](a ) ) one observes the anti - resonance phenomenon described earlier , in which the system returns ( approximately ) to its initial state after every second kick this manifests itself as an oscillation in the energy .
in contrast , near @xmath82 ( figs . [ fig : evsn15](c ) and [ fig : evsn25](c ) ) one sees a continual growth in the energy as a consequence of the phenomenon of quantum resonance .
for the lower kicking strength , we also observe behavior suggesting a period-4 anti - resonance close to @xmath101 ( fig .
[ fig : evsn15](b ) ) . numerical simulations confirm this behavior , for @xmath102 , at a value of @xmath0 in this vicinity .
for the larger kicking strength similar behavior occurs to a certain extent , but is most pronounced at slightly larger values of @xmath0 .
unfortunately , falling signal - to - noise and the effects of a finite initial mean momentum means that it is very difficult to extract useful energy values beyond about 4 kicks and hence to rigorously confirm this higher - order behavior ( at least with the experimental setup that was used ) . also plotted in figs .
[ fig : evsn15 ] and [ fig : evsn25 ] are the results of numerical simulations for the appropriate values of @xmath0 and @xmath103 , and for @xmath104 and @xmath105 , respectively .
quantitative agreement between theory and experiment is reasonably good , but notable deviations appear after several kicks for the anti - resonance and resonance at @xmath106 and @xmath3 , respectively .
we believe that these deviations are due partly to fluctuations in the initial mean momentum of the atomic cloud relative to the laser standing wave .
inspection of the images of the kicked atomic clouds reveals that the position of the `` zero momentum '' peak can fluctuate from shot - to - shot by an amount of up to a few tenths of an atomic recoil .
this was confirmed by bragg scattering measurements @xcite and is associated with the precise details of the switching - off of the time - averaged orbiting potential trap used to confine the initial bose - einstein condensate . as shown in @xcite , the quantum anti - resonance and resonance phenomena are particularly sensitive to any non - zero initial mean momentum of the atomic ensemble . in fig .
[ fig : finitep0 ] we demonstrate this sensitivity with simulations for the parameters associated with figs .
[ fig : evsn15](a ) and [ fig : evsn15](c ) , but now for initial mean momenta @xmath107 and @xmath108 .
the width of the initial momentum distribution is the same as in the experiment .
as shown , with a finite value of @xmath109 , one finds behavior more consistent with the experimental data .
the sensitivity of the energy to the initial mean momentum of the atomic cloud adds an extra variable to the problem , over which we at present have little control . to further emphasize the dependence of the energy on the initial mean momentum , in fig .
[ fig : e2theory_p0 ] we plot the energy after two kicks , @xmath93 , using the theoretical result of eq .
( [ eqn : diffrate2kicks ] ) for several values of the initial mean momentum , @xmath109 .
this plot highlights the extreme sensitivity to initial motion of the quantum resonance at @xmath82 , which in fact changes to an _ anti - resonance _ when @xmath110 .
it follows that , in order to see resonance and antiresonance behavior controllably in their clearest forms , one requires very precise control over the initial mean motion of the atomic cloud .
in summary our results show a rich structure of resonances as a function of the effective planck s constant and kick number for the atom - optical kicked rotor with a narrow initial momentum distribution . for two kicks
we have shown that the resonant behavior is a more complex than those predicted by shepelyansky for a system with a broad initial momentum distribution .
moreover we observed quantum anti - resonance ( @xmath111 ) and quantum resonance ( @xmath91 ) in the energy as a function of @xmath0 for different kick strengths .
comparison between theory and experiment showed reasonable agreement , but by introducing an initial mean momentum to the numerical model better agreement could be obtained .
consequently , we have shown that the quantum features are very sensitive to precise initial conditions .
one possible method of eliminating the initial mean momentum of the condensate would be to perform the measurements with the condensate still confined by the magnetic trap .
one difficulty with this is that for our apparatus , condensate micro - motion will be introduced @xcite .
in addition , mean field effects will occur , although the nonlinearity associated with these provide an opportunity to investigate instability of the condensate wave function @xcite .
g. casati , b. v. chirikov , f. m. izrailev and j. ford , in _
stochastic behaviour in classical and quantum hamiltonian systems _ , vol .
93 of _ lecture notes in physics _ , edited by g. casati and j. ford ( springer - verlag , berlin , 1979 ) . | we report the experimental observation of resonances in the early - time momentum diffusion rates for the atom - optical delta - kicked rotor . in this work
a bose - einstein condensate provides a source of ultra - cold atoms with an ultra - narrow initial momentum distribution , which is then subjected to periodic pulses ( or `` kicks '' ) using an intense far - detuned optical standing wave .
a quantum resonance occurs when the momentum eigenstates accumulate the same phase between kicks leading to ballistic energy growth .
conversely , an anti - resonance is observed when the phase accumulated from successive kicks cancels and the system returns to its initial state .
our experimental results are compared with theoretical predictions . |
we are grateful for support from the sfb / trr 102 ( project b04 ) , the leipzig graduate school of excellence gsc185 `` buildmona '' and the deutsch - franzsische hochschule ( dfh - ufa ) under grant no .
cdfa-02 - 07 .
j.z . and m.m .
were funded by the european union and the free state of saxony .
10 b. a. berg , t. neuhaus , phys .
b 267 ( 1991 ) 249 ; + b. a. berg , t. neuhaus , phys .
68 ( 1992 ) 9 . w. janke , int . j. mod
c3 ( 1992 ) 1137 ; + w. janke , physica a 254 ( 1998 ) 164 .
f. wang , d. p. landau , phys .
86 ( 2001 ) 2050 ; + f. wang , d. p. landau , phys .
e 64 ( 2001 ) 056101 .
w. janke ( ed . ) , rugged free - energy landscapes , lect .
notes phys . 736 ( 2008 )
203 ; + a. mitsutake , y. sugita , y. okamoto , biopolymers ( peptide science ) 60 ( 2001 ) 96 .
l. zhan , comput .
( 2008 ) 339 .
j. yin , d. p. landau , comput .
( 2012 ) 1568 .
v. v. slavin , low temp .
36 ( 2010 ) 243 ; + a. ghazisaeidi , f. vacondio , l. a. rusch , j. lightwave technol . 28 ( 2010 ) 79 .
j. zierenberg , m. marenz , w. janke , comput .
( 2013 ) 1155 .
w. janke , b. a. berg , m. katoot , nucl .
b 382 ( 1992 ) 649 .
a. milchev , a. bhattacharaya , k. binder , macromolecules 34 ( 2001 ) 1881 ; + s. schnabel , m. bachmann , w. janke , j. chem .
( 2009 ) 124904 . | we implemented a parallel version of the multicanonical algorithm and applied it to a variety of systems with phase transitions of first and second order .
the parallelization relies on independent equilibrium simulations that only communicate when the multicanonical weight function is updated .
that way , the markov chains efficiently sample the temporary distributions allowing for good estimations of consecutive weight functions .
the systems investigated range from the well known ising and potts spin systems to bead - spring polymers .
we estimate the speedup with increasing number of parallel processes .
overall , the parallelization is shown to scale quite well . in the case of multicanonical simulations of the @xmath0-state potts model ( @xmath1 ) and multimagnetic simulations of the ising model , the optimal performance
is limited due to emerging barriers .
multicanonical simulations , parallel implementation , ising , potts , bead - spring polymer umbrella sampling algorithms like the multicanonical method @xcite and the wang - landau method @xcite have been applied to a variety of complex systems over the last two decades .
they are well suited for the investigation of phase transitions , especially of first order and may be applied to systems with rugged free - energy landscapes in physics , biology , and chemistry @xcite . for complex systems ,
a carefully chosen set of monte carlo update moves is usually the key to a successful simulation . but with computer performance increasing mainly in terms of parallel processing on multi - core architectures , it is of advantage when the algorithm can be parallelized .
this has been done recently for the wang - landau recursion @xcite and for the standard multicanonical recursion @xcite . while the former implementation relies on shared memory , introducing racing conditions and frequent communication , the latter benefits from independent markov chains with occasional communication .
after a short summary of the parallel multicanonical method , we will proceed with a demonstration of its performance for several spin systems and a flexible polymer .
the multicanonical ( muca ) method can be applied to a variety of ensembles .
still , it is probably easiest to understand using the example of the canonical ensemble . for a fixed temperature , all configurations that the system may assume are weighted with the boltzmann weight @xmath2 $ ] , resulting in a temperature dependent energy distribution
the idea is to replace this boltzmann weight by an arbitrary weight function @xmath3 , which may be modified iteratively such that the resulting energy distribution covers the full range of canonical distributions , hence it is called multicanonical . in terms of the partition function , this may be written as @xmath4 where @xmath5 is the density of states .
the canonical distributions and expectation values are recovered by reweighting : @xmath6 the most difficult task is the weight modification , which requires some effort .
the easiest way is to construct consecutive weights from the last weights and the current energy histogram : @xmath7 .
more sophisticated methods use the full statistics of previous iterations for a stable and efficient approximation of the density of states @xcite .
all our simulations use the latter version implemented with logarithmic weights in order to avoid numerical problems . the basic idea of the parallel implementation
is shown in fig .
[ fig : pmuca ] .
the system is initialized in @xmath8 independent realizations with the same weight function , which are distributed onto different cores . after each iteration ,
the individual histograms are merged and provide an estimate of the distribution @xmath9 belonging to the current weight function @xmath10 .
this is used to determine the consecutive weight function @xmath11 , which is again distributed to all @xmath8 cores .
the whole procedure is repeated until the weight function results in a flat energy distribution . since the weight modification is usually very fast compared to a single iteration , communication is kept to a bare minimum
. cores . after each iteration with independent markov chains
( but identical weights ) , the histograms are merged , the new weights are estimated and distributed to all processes again . , width=340 ] moreover , the parallelization may easily be applied to other ensembles , for example multimagnetic ( muma ) simulations . in this case , the coefficients of the partition function are modified by a correction weight function , which depends for example on the magnetization and is again modified iteratively in order to yield a flat histogram in the parameter : @xmath12 the parallelization is completely analogous to the standard case and the performance of the parallel multimagnetic simulation will be demonstrated below . for a fair comparison of the performance of this parallel implementation , we need to consider a few aspects .
first of all , the parallelization relies on independent markov processes , which leads to different simulations for different degrees of parallelization .
consequently , we may only compare the average performance per degree of parallelization .
furthermore , a number of parameters can influence the results and need to be fixed .
one example is the number of sweeps per iteration . in order to provide a proper estimate of the performance
, we determined the optimal number of sweeps per iteration @xmath13 for all degrees of parallelization in the cases of the ising model and the @xmath14-state potts model .
a detailed description can be found in @xcite .
the results of this analysis are @xmath15 we extrapolated this result for the remaining spin systems .
another factor to consider is the thermalization time . in order to remove additional parameters in our investigation
, we decided to thermalize only in the beginning and not in between iterations .
since the parallelization changes the outcome of the simulation , also the number of iterations until convergence is influenced .
this allows us to consider two different measures of the performance .
one is given by the speedup in convergence time , comparing the time @xmath16 a @xmath8-core simulation needs until convergence of the muca weights with the time @xmath17 a single - core simulation needs : @xmath18 obviously , this answers the question of required simulation time with increasing parallelization but depends strongly on the involved hardware .
thus , the result may differ if investigated on different compute clusters .
another possibility is to consider a time - independent statistical speedup by comparing the total number of sweeps per core until convergence .
as the optimal number of sweeps per iteration @xmath19 is fixed for all realizations , this results in measuring the average number of iterations until convergence @xmath20 .
@xmath21_{1 } } { [ \bar{n}_{\rm iter } m_{\rm opt}(p)]_{p}},\ ] ] in the following , we will use the ising model to demonstrate the differences .
afterwards we will restrict ourselves to the time - independent statistical speedup for simplicity .
we consider the two - dimensional ising system as a first test case .
this spin model with nearest - neighbor interaction exhibits a temperature driven second - order phase transition .
the hamiltonian is defined as @xmath22 .
figure [ fig : ising ] shows the performance of the method .
it can be seen that the statistical speedup scales nicely for all system sizes , in fact @xmath23 .
this means that the total statistics is efficiently distributed onto all cores .
the time speedup also scales well , except for the small system sizes where the duration of a single iteration was of the order of milliseconds for which our network communication is insufficient .
the two - dimensional @xmath0-state potts model is described by @xmath24 , where @xmath25 and interaction is restricted to nearest neighbors .
the system shows a temperature driven first - order phase transition for @xmath26 and a second - order phase transition otherwise .
applying the parallel multicanonical method to the potts model demonstrates its effect on systems with first - order phase transitions .
indeed , also in this case the parallelization works well , but with increasing degree of parallelization the speedup seems to saturate ( see fig .
[ fig : spinsystems ] ) .
a similar effect is observed when applying the parallelization to a multimagnetic simulation of the ising system , also shown in the figure . in this case , the ising system is simulated at fixed temperature @xmath27 , while it is attempted to achieve a flat distribution of the magnetization .
the occurring field - driven phase transition is of first order .
the saturation of the speedup may be explained by large integrated autocorrelation times accompanying concealed barriers .
thus , when reducing the number of sweeps per core with increasing degree of parallelization this might reach a point where the individual sweeps are too short in order to efficiently cross emerging concealed barriers . for a detailed description
we refer to @xcite .
the @xmath0-state potts model is furthermore well suited to take a look at the performance of the parallelization in the crossover regime from a second - order phase transition to a first - order phase transition . to this end ,
we considered extrapolated @xmath13 for different @xmath0 values .
the result is shown in fig .
[ fig : qpotts ] . for @xmath28
the temperature - driven phase transition is of second - order and the scaling of the performance is very well .
this still holds for @xmath29 where a so - called weak first - order phase transition occurs .
already for @xmath30 , we can see the drop in performance for large degrees of parallelization , as before .
-state potts models on a @xmath31 square lattice . for @xmath28
the potts model exhibits a second - order phase transition , while for @xmath32 the phase transition becomes first order . ] leaving the constraint of a lattice , we applied the parallel multicanonical method to a more complex system , a single flexible bead - spring polymer .
it consists of @xmath33 identical monomers , which are connected to their bonded neighbors by a fene spring potential and which interact with other monomers via a lennard - jones potential .
the hamiltonian is given by @xmath34 ^ 2\right),\ ] ] where @xmath35 is the average bond length , @xmath36 , @xmath37 , and @xmath38 is the spring constant . the parameters where chosen @xmath39 , @xmath40 , and @xmath41 according to @xcite .
as mentioned above , the choice of updates is crucial .
here , we used a combination of single monomer shift , spherical rotation , and double - bridging moves . using the example of a polymer system , we want to show the general applicability of the parallel multicanonical method
thus , we took an existing code of a multicanonical simulation with a fixed number of sweeps per iteration and some thermalization between iterations .
the total number of sweeps per iteration was distributed onto the cores . while we considered the full energy range for the spin systems , we restricted the multicanonical simulation of the homo - polymer to an energy range around the collapse transition . figure [ fig : polymer ] shows that the straightforward parallelization works also well for complex off - lattice systems , which involve computationally more expensive energy calculations .
moreover , this shows that the parallelization may be applied straightforwardly without taking too much care about the involved parameters . .
] in summary , the application of the parallel multicanonical method presented here is straightforward and very efficient for a range of systems .
we studied the performance on the example of the ising model , the @xmath0-state potts model and a coarse - grained polymer model . furthermore
, we showed that the parallelization may easily be adapted to flat histogram simulations in other ensembles .
overall , we could demonstrate a good performance yielding a close - to - perfect scaling @xmath23 for up to @xmath42 cores . |
in this paper , we study the problem of reconstructing a simple polygon @xmath0 from the visibility angles measured at the vertices of @xmath0 and from the cyclically ordered vertices of @xmath0 along its boundary .
precisely , for an unknown simple polygon @xmath0 of @xmath1 vertices , suppose we are given ( 1 ) the vertices ordered counterclockwise ( ccw ) along the boundary of @xmath0 , and ( 2 ) for each vertex @xmath2 of @xmath0 , the angles between any two adjacent rays emanating from @xmath2 to the vertices of @xmath0 that are visible to @xmath2 such that these angles are in the ccw order as seen around @xmath2 , beginning at the ccw neighboring vertex of @xmath2 on the boundary of @xmath0 ( e.g. , see fig .
[ fig : angles ] ) .
a vertex @xmath6 of @xmath0 is _ visible _ to a vertex @xmath2 of @xmath0 if the line segment connecting @xmath6 and @xmath2 lies entirely in @xmath0 .
the objective of the problem is to reconstruct the simple polygon @xmath0 ( up to similarity ) that fits all the given angles .
we call this problem the _ polygon reconstruction problem from angles _ , or _ pra _ for short . figure [ fig : example ] gives an example .
: the angle sequence @xmath7 is given . ]
the problem has been studied by disser , mihalk , and widmayer @xcite , who showed that the solution polygon for the input is unique ( up to similarity ) and gave an @xmath3 time algorithm for reconstructing such a polygon . using the input
, their algorithm first constructs the visibility graph @xmath8 of @xmath0 and subsequently reconstructs the polygon @xmath0 . as shown in @xcite , once @xmath8 is known , the polygon @xmath0 can be obtained efficiently ( e.g. , in @xmath4 time ) with the help of the angle data and the ccw ordered vertex sequence of @xmath0
. given a visibility graph @xmath8 , the problem of determining whether there is a polygon @xmath0 that has @xmath8 as its visibility graph is called the visibility graph _ recognition _ problem , and the problem of actually constructing such a polygon @xmath0 is called the visibility graph _ reconstruction _ problem .
note that the general visibility graph recognition and reconstruction problems are long - standing open problems with only partial results known ( e.g. , see @xcite for a short survey ) .
everett @xcite showed that the visibility graph reconstruction problem is in pspace , but no better upper bound on the complexity of either problem is known . in our problem setting , we have the angle data information and the ordered vertex list of @xmath0 ; thus @xmath0 can be constructed efficiently after knowing @xmath8 .
hence , the major part of the algorithm in @xcite is dedicated to constructing the visibility graph @xmath8 of @xmath0 . as indicated in @xcite ,
the key difficulty is that the vertices in this problem setting have no recognizable labels , e.g. , the angle measurement at a vertex @xmath2 gives angles between visible vertices to @xmath2 but does not identify these visible vertices globally .
the authors in @xcite also showed that some natural greedy approaches do not seem to work .
an @xmath3 time algorithm for constructing @xmath8 is given in @xcite .
the algorithm , called the _ triangle witness algorithm _
, is based on the following observation : suppose we wish to determine whether a vertex @xmath9 is visible to another vertex @xmath10 ; then @xmath9 is visible to @xmath10 if and only if there is a vertex @xmath11 on the portion of the boundary of @xmath0 from @xmath12 to @xmath13 in the ccw order such that @xmath11 is visible to both @xmath9 and @xmath10 and the triangle formed by the three vertices @xmath14 , and @xmath11 does not intersect the boundary of @xmath0 except at these three vertices ( such a vertex @xmath11 is called a _ triangle witness vertex _ ) . in this paper , based on the triangle witness algorithm @xcite , by exploiting some new geometric properties , we give an improved algorithm with a running time of @xmath4 .
the improvement is due to two key observations .
first , in the triangle witness algorithm @xcite , to determine whether a vertex @xmath9 is visible to another vertex @xmath10 , the algorithm needs to determine whether there exists a triangle witness vertex along the boundary of @xmath0 from @xmath12 to @xmath13 in the ccw order ; to this end , the algorithm checks every vertex in that boundary portion of @xmath0 .
we observe that it suffices to check only one particular vertex in that boundary portion .
this removes an @xmath15 factor from the running time of the triangle witness algorithm @xcite .
second , some basic operations in the triangle witness algorithm @xcite take @xmath16 time each ; by utilizing certain different data structures , our new algorithm can handle each of those basic operations in constant time .
this removes another @xmath16 factor from the running time .
note that since the input size is @xmath5 in the worst case ( e.g. , the total number of all visibility angles ) , our algorithm is worst - case optimal .
as shown in @xcite , if only the angle measurements are given , i.e. , the ordered vertices along the boundary of @xmath0 are unknown , then the information is not sufficient for reconstructing @xmath0 . in other words , it may be possible to compute several simple polygons that are not similar but all fit the given measured angles ( see @xcite for an example ) .
the problems of reconstructing geometric objects based on measurement data have been studied extensively ( e.g. , @xcite ) . as discussed above , the general visibility graph recognition and reconstruction problems are in pspace @xcite and no better complexity upper bound is known so far ( e.g. , see @xcite ) . yet , some results have been given for certain special polygons .
for example , everett and corneil @xcite characterized the visibility graphs of spiral polygons and gave a linear time reconstruction algorithm .
choi , shin , and chwa @xcite , and colley , lubiw , and spinrad @xcite characterized and recognized the visibility graphs of funnel - shaped polygons . by adding extra information , some versions of the problems become more tractable .
orourke and streinu @xcite considered the _ vertex - edge _ visibility graph that includes edge - to - edge visibility information .
wismath @xcite introduced the _ stab graphs _ which are also an extended visibility structure and showed how parallel line segments can be efficiently reconstructed from it .
snoeyink @xcite proved that a unique simple polygon ( up to similarity ) can be determined by the interior angles at its vertices and the cross - ratios of the diagonals of any given triangulation , where the cross - ratio of a diagonal is the product of the ratios of edge lengths for the two adjacent triangles .
jackson and wismath @xcite studied the reconstruction of orthogonal polygons from horizontal and vertical visibility information and gave an @xmath17 time reconstruction algorithm .
biedl , durocher , and snoeyink @xcite considered the problem of reconstructing the two - dimensional floor plan of a polygonal room using different types of scanned data , and proposed several problem models .
sidlesky , barequet , and gotsman @xcite studied the problem of reconstructing a planar polygon from its intersections with a collection of arbitrarily - oriented cutting " lines .
reconstructing a simple polygon from angle data was first considered by bil _
et al_. @xcite , who aimed to understand what kinds of sensorial capabilities are sufficient for a robot moving inside an unknown polygon to reconstruct the visibility graph of the polygon .
it was shown in @xcite that if the robot is equipped with a compass to measure the angle between any two vertices that are currently visible to the robot and also has the ability to know where it came from when moving from vertex to vertex , then the visibility graph of the polygon can be uniquely reconstructed .
reconstruction and exploration of environments by robots in other problem settings have also been studied ( e.g. , see @xcite ) .
the rest of this paper is organized as follows . in section [ sec : pre ] , we give the problem definitions in detail and introduce some notations and basic observations .
to be self - contained , in section [ sec : review ] , we briefly review the triangle witness algorithm given in @xcite .
we then present our improved algorithm in section [ sec : improve ] .
in this section , we define the problem in detail and introduce the needed notations and terminology . for ease of discussion and comparison , some of our notations follow those in @xcite .
let @xmath0 be a simple polygon of @xmath1 vertices @xmath18 in the ccw order along @xmath0 s boundary . denote by @xmath19 the visibility graph of @xmath0 , where @xmath20 consists of all vertices of @xmath0 and for any two distinct vertices @xmath9 and @xmath10 , @xmath21 contains an edge @xmath22 connecting @xmath9 and @xmath10 if and only if @xmath9 is visible to @xmath10 inside @xmath0 . in this paper ,
the indices of all @xmath9 s are taken as congruent modulo @xmath1 , i.e. , if @xmath23 , then @xmath24 is the same vertex as @xmath11 , where @xmath25 ( or @xmath26 ) ; similarly , if @xmath27 , then @xmath28 is the same vertex as @xmath11 , where @xmath29 . for each @xmath30 ,
denote by @xmath31 its degree in the visibility graph @xmath8 , and denote by @xmath32 the sequence of vertices in @xmath0 visible to @xmath9 from @xmath12 to @xmath33 ordered ccw around @xmath9 .
we refer to @xmath34 as @xmath9 s _ visibility angle sequence_. note that since both @xmath33 and @xmath12 are visible to @xmath9 , @xmath35 and @xmath36 .
for any two vertices @xmath14 in @xmath20 , let @xmath37 denote the sequence @xmath38 of the vertices ordered ccw along the boundary of @xmath0 from @xmath9 to @xmath10 .
we refer to @xmath37 as a _
chain_. let @xmath39 denote the number of vertices of @xmath0 in the chain @xmath37 .
for any two distinct vertices @xmath40 , let @xmath41 be the ray emanating from @xmath9 and going towards @xmath10 .
for any three vertices @xmath42 , denote by @xmath43 the ccw angle defined by rotating @xmath44 around @xmath2 to @xmath45 ( @xmath9 or @xmath10 need not be visible to @xmath2 ) .
note that the values of all angles we use in this paper are in @xmath46 . for any vertex @xmath47 and @xmath48 , let @xmath49 be @xmath50
. the problem can then be re - stated as follows : given a sequence of all vertices @xmath18 of an unknown simple polygon @xmath0 in the ccw order along @xmath0 s boundary , and the angles @xmath51 for each vertex @xmath2 of @xmath0 with @xmath52 , we seek to reconstruct @xmath0 ( up to similarity ) to fit all the given angles . without loss of generality , we assume that no three distinct vertices of @xmath0 are collinear . it is easy to see that after @xmath4 time preprocessing , for any @xmath53 and any @xmath48 , the angle @xmath49 can be obtained in constant time . in the following discussion , we assume that this preprocessing has already been done . sometimes we ( loosely ) say that these angles are given as input . the algorithm given in @xcite does not construct @xmath0 directly .
instead , the algorithm first computes its visibility graph @xmath19 .
as mentioned earlier , after knowing @xmath8 , @xmath0 can be reconstructed efficiently with the help of the angle data and the ccw ordered vertex sequence of @xmath0 .
the algorithm for constructing @xmath8 in @xcite is called the _ triangle witness algorithm _ , which will be briefly reviewed in section [ sec : review ] .
since @xmath20 consists of all vertices of @xmath0 , the problem of constructing @xmath8 is equivalent to constructing its edge set @xmath21 , i.e. , for any two distinct vertices @xmath54 , determine whether there is an edge @xmath22 in @xmath21 connecting @xmath9 and @xmath10 ( in other words , determining whether @xmath9 is visible to @xmath10 inside @xmath0 ) . to discuss the involved algorithms , we need one more definition . for any two vertices
@xmath55 with @xmath56 , suppose a vertex @xmath57 is visible to both @xmath9 and @xmath10 ; then we let @xmath58 be the _ first _ visible vertex to @xmath9 on the chain @xmath59 and let @xmath60 be the _
last _ visible vertex to @xmath10 on the chain @xmath59 ( e.g. , see fig . [
fig : sangle ] ) .
intuitively , imagine that we rotate a ray from @xmath61 around @xmath9 counterclockwise ; then the first vertex on the chain @xmath59 hit by the rotating ray is @xmath58 .
similarly , if we rotate a ray from @xmath62 around @xmath10 clockwise , then the first vertex on the chain @xmath59 hit by the rotating ray is @xmath60 .
note that if @xmath9 is visible to @xmath10 , then @xmath58 is @xmath10 and @xmath60 is @xmath9 .
we denote by @xmath63 the angle @xmath64 and denote by @xmath65 the angle @xmath66 .
it should be pointed out that for ease of understanding this paper , the above statement of defining @xmath67 is different from that in @xcite but they refer to the same angles in the algorithm .
the motivation for defining @xmath67 will be clear after discussing the following lemma , which has been proved in @xcite . and @xmath58 , and @xmath63 and @xmath65 which are the angles @xmath64 and @xmath66 , respectively .
] [ lem:10]@xcite for any two vertices @xmath55 with @xmath56 , @xmath9 is visible to @xmath10 if and only if there exists a vertex @xmath11 on @xmath68 such that @xmath11 is visible to both @xmath9 and @xmath10 and @xmath69 .
since the above lemma is also critical to our improved algorithm in section [ sec : improve ] , we sketch the proof of the lemma below . for any two vertices
@xmath55 with @xmath56 , if @xmath9 is visible to @xmath10 , then it is not difficult to see that there must exist a vertex @xmath57 that is visible to both @xmath9 and @xmath10 .
since the three vertices @xmath14 , and @xmath11 are mutually visible to each other , it is clear that the triangle formed by these three vertices does not intersect the boundary of @xmath0 except at the three vertices , implying that @xmath70 .
such a vertex @xmath11 is called a _
triangle witness _ of the edge @xmath22 in @xmath21 .
suppose @xmath9 is not visible to @xmath10 ; then it is possible that there does not exist a vertex @xmath57 which is visible to both @xmath9 and @xmath10 .
in fact , if there exists no vertex @xmath57 that is visible to both @xmath9 and @xmath10 , then @xmath9 can not be visible to @xmath10 .
hence in the following , we assume that such a vertex @xmath11 exists , i.e. , a vertex @xmath71 is visible to both @xmath9 and @xmath10 ( yet @xmath9 is not visible to @xmath10 ) . note that @xmath11 is not collinear with @xmath9 and @xmath10 .
if @xmath72 ( i.e. , the chain @xmath73 is part of the boundary of @xmath0 that blocks the visibility between @xmath9 and @xmath10 ) , then the lemma obviously holds .
otherwise , @xmath74 , and in this case , the visibility between @xmath9 and @xmath10 is blocked by the chain @xmath75 . since @xmath9 is not visible to @xmath10 , for any choice of such a vertex @xmath11 , the angle @xmath76 is not given by @xmath9 s visibility angle sequence @xmath34 .
the closest approximation " for @xmath76 in this case is determined by a vertex @xmath2 on the chain @xmath75 such that @xmath2 becomes @xmath10 if and only if @xmath10 is visible to @xmath9 . as in the definition of @xmath63 ,
the vertex @xmath58 is such a vertex @xmath2 , i.e. , the first visible vertex to @xmath9 on the chain @xmath59 .
similarly , the vertex @xmath9 in @xmath77 is replaced " in the definition of @xmath65 by @xmath60 , i.e. , the last visible vertex to @xmath10 on @xmath59 .
clearly , when @xmath10 is not visible to @xmath9 , it must hold that @xmath78 and @xmath79 .
therefore , @xmath80 .
lemma [ lem:10 ] thus follows .
in this section , we briefly review the triangle witness algorithm in @xcite that constructs the visibility graph @xmath19 of the unknown simple polygon @xmath0 .
the triangle witness algorithm is based on lemma [ lem:10 ] .
the algorithm has @xmath81 iterations . in the @xmath82-th iteration ( @xmath83 ) , the algorithm checks , for each @xmath84 , whether @xmath9 is visible to @xmath85 . after all iterations , the edge set @xmath21 can be obtained .
to this end , the algorithm maintains two maps @xmath86 and @xmath87 : @xmath88[v_j]=t$ ] if @xmath10 is identified as the @xmath89-th visible vertex to @xmath9 in the ccw order , i.e. , @xmath90 ; the definition of @xmath87 is the same as @xmath86 . during the algorithm ,
@xmath86 will be filled in the ccw order and @xmath87 will be filled in the clockwise ( cw ) order .
when the algorithm finishes , for each @xmath9 , @xmath88 $ ] will have all visible vertices to @xmath9 on the chain @xmath91 while @xmath92 $ ] will have all visible vertices to @xmath9 on the chain @xmath93 .
thus , @xmath88 $ ] and @xmath92 $ ] together contain all visible vertices of @xmath0 to @xmath9 .
for ease of description , we also treat @xmath88 $ ] and @xmath92 $ ] as sets , e.g. , @xmath94 $ ] means that there is an entry @xmath88[v_l]$ ] and @xmath95|$ ] means the number of entries in the current @xmath88 $ ] . initially , when @xmath96 , since every vertex is visible to its two neighbors along the boundary of @xmath0 , we have @xmath88[v_{i+1}]=1 $ ] and @xmath92[v_{i-1}]=deg(v_i)$ ] for each @xmath9 .
in the @xmath82-th iteration , we determine for each @xmath9 , whether @xmath9 is visible to @xmath85 .
below , we let @xmath97 .
note that @xmath95|+1 $ ] is the index of the first visible vertex to @xmath9 in the ccw order that is not yet identified ; similarly , @xmath98|)$ ] is the index of the first visible vertex to @xmath99 in the cw order that is not yet identified .
if @xmath9 is visible to @xmath99 , then we know that @xmath99 is the @xmath100|+1)$]-th visible vertex to @xmath9 and @xmath9 is the @xmath98|)$]-th visible vertex to @xmath99 , and thus we set @xmath88[v_{j}]=|f[v_i]|+1 $ ] and @xmath101[v_i]=deg(v_{j})-|b[v_{j}]|$ ] .
if @xmath9 is not visible to @xmath99 , then we do nothing .
it remains to discuss how to determine whether @xmath9 is visible to @xmath99 . according to lemma [ lem:10 ]
, we need to determine whether there exists a triangle witness vertex @xmath11 in the chain @xmath68 , i.e. , @xmath11 is visible to both @xmath9 and @xmath99 and @xmath69 . to this end ,
the algorithm checks every vertex @xmath11 in @xmath68 . for each @xmath11 ,
the algorithm first determines whether @xmath11 is visible to both @xmath9 and @xmath99 , by checking whether there is an entry for @xmath11 in @xmath88 $ ] and in @xmath101 $ ] .
the algorithm utilizes balanced binary search trees to represent @xmath86 and @xmath87 , and thus checking whether there is an entry for @xmath11 in @xmath88 $ ] and in @xmath102 $ ] can be done in @xmath16 time . if @xmath11 is visible to both @xmath9 and @xmath99 , then the next step is to determine whether @xmath69 .
it is easy to know that @xmath103 is @xmath104[v_{j}],b[v_l][v_i])$ ] , which can be found readily from the input . to obtain @xmath63
, we claim that it is the angle @xmath105[v_l],|f[v_i]|+1)$ ] .
indeed , observe that all visible vertices in the chain @xmath68 are in the current @xmath88 $ ] .
as explained above , @xmath95|+1 $ ] is the index of the first visible vertex to @xmath9 in the ccw order that has not yet been identified , which is the first visible vertex to @xmath9 in the chain @xmath59 , i.e , the vertex @xmath58 .
thus , @xmath105[v_l],|f[v_i]|+1)$ ] is @xmath64 , which is @xmath63 .
similarly , @xmath65 is the angle @xmath106|,b[v_j][v_l])$ ] .
note that both the angles @xmath105[v_l],|f[v_i]|+1)$ ] and @xmath106|,b[v_j][v_l])$ ] are known .
algorithm [ algo:10 ] in the appendix summarizes the whole algorithm @xcite . to analyze the running time of the above triangle witness algorithm ,
note that it has @xmath81 iterations . in each iteration
, the algorithm checks whether @xmath9 is visible to @xmath85 for each @xmath107 . for each @xmath9 ,
the algorithm checks every @xmath11 for @xmath108 , i.e , in the chain @xmath109 . for each such @xmath11 ,
the algorithm takes @xmath16 time as it uses balanced binary search trees to represent the two maps @xmath86 and @xmath87 . in summary ,
the overall running time of the triangle witness algorithm in @xcite is @xmath3 .
in this section , we present an improved solution over the triangle witness algorithm in @xcite sketched in section [ sec : review ] .
our improved algorithm runs in @xmath4 time . since the input size ( e.g. , the total number of visibility angles ) is @xmath5 in the worst case , our improved algorithm is worst - case optimal .
our new algorithm follows the high - level scheme of the triangle witness algorithm in @xcite , and thus we call it the _ improved triangle witness algorithm_. as in @xcite , the new algorithm also has @xmath81 iterations , and in each iteration , we determine whether @xmath9 is visible to @xmath85 for each @xmath107 .
for every pair of vertices @xmath9 and @xmath85 , let @xmath97 . to determine whether @xmath9 is visible to @xmath10 , the triangle witness algorithm @xcite checks each vertex @xmath11 in the chain @xmath68 to see whether there exists a triangle witness vertex . in our new algorithm , instead , we claim that we need to check only one particular vertex , @xmath110|}(v_i)$ ] , i.e. , the last visible vertex to @xmath9 in the chain @xmath68 in the ccw order , as stated in the following lemma .
: @xmath111 is @xmath110|}(v_i)$ ] , the last visible vertex to @xmath9 in @xmath68 ; @xmath9 is visible to @xmath10 if and only if @xmath111 is a triangle witness vertex . ] for the example in fig .
[ fig : visible ] : the gray ( pink ) area is @xmath112 , which is a star - shaped polygon with the vertex @xmath9 as a kernel point . ] for the example in fig .
[ fig : visible ] : the thick line segments form @xmath113 . ]
[ lem:20 ] the vertex @xmath9 is visible to @xmath10 if and only if the vertex @xmath110|}(v_i)$ ] is a triangle witness vertex for @xmath9 and @xmath10 .
let @xmath111 denote the vertex @xmath110|}(v_i)$ ] , which is the last visible vertex to @xmath9 in the chain @xmath68 in the ccw order . recall that @xmath111 being a triangle witness vertex for @xmath9 and @xmath10 is equivalent to saying that @xmath111 is visible to both @xmath9 and @xmath10 and @xmath114 .
if @xmath9 is visible to @xmath10 , then we prove below that @xmath111 is a triangle witness vertex for @xmath9 and @xmath10 , i.e. , we prove that @xmath111 is visible to both @xmath9 and @xmath10 and @xmath114 .
refer to fig .
[ fig : visible ] for an example .
let @xmath115 denote the subpolygon of @xmath0 that is visible to the vertex @xmath9 .
usually , @xmath115 is called the _
visibility polygon _ of @xmath9 and it is well - known that @xmath115 is a star - shaped polygon with @xmath9 as a kernel point ( e.g. , see @xcite ) . figure [ fig : vispolygon ] illustrates @xmath115 .
since @xmath111 is the last visible vertex to @xmath9 in @xmath68 and @xmath10 is visible to @xmath9 , we claim that @xmath10 is also visible to @xmath111 . indeed , since both @xmath111 and @xmath10 are visible to @xmath9 , @xmath111 and @xmath10 are both on the boundary of the visibility polygon @xmath115 .
let @xmath116 denote the portion of the boundary of @xmath115 from @xmath111 to @xmath10 counterclockwise ( see fig .
[ fig : boundary ] ) .
we prove below that @xmath116 does not contain any vertex of @xmath0 except @xmath111 and @xmath10 .
first , @xmath116 can not contain any vertex in the chain @xmath117 ( otherwise , @xmath10 would not be visible to @xmath9 ) .
let the index of the vertex @xmath111 be @xmath118 , i.e. , @xmath119 .
similarly , @xmath116 can not contain any vertex in the chain @xmath120 ( otherwise , @xmath111 would not be visible to @xmath9 ) .
finally , @xmath116 can not contain any vertex in the chain @xmath121 , since otherwise @xmath111 would not be the last visible vertex to @xmath9 in @xmath68 .
thus , @xmath116 does not contain any vertex of @xmath0 except @xmath111 and @xmath10 .
therefore , the region bounded by @xmath116 , the line segment connecting @xmath9 and @xmath10 , and the line segment connecting @xmath9 and @xmath111 must be convex ( in fact , it is always a triangle ) and this region is entirely contained in @xmath0 .
this implies that @xmath10 is visible to @xmath111 .
hence , the three vertices @xmath14 , and @xmath111 are mutually visible to each other , and we have @xmath70 . on the other hand , if @xmath111 is a triangle witness vertex for @xmath9 and @xmath10 , then by lemma [ lem:10 ] , the vertex @xmath9 is visible to @xmath10 .
the lemma thus follows . by lemma [ lem:20 ] , to determine whether @xmath9 is visible to @xmath10 , instead of checking every vertex in @xmath68 , we need to consider only the vertex @xmath110|}(v_i)$ ] in the current set @xmath88 $ ] .
hence , lemma [ lem:20 ] immediately reduces the running time of the triangle witness algorithm by an @xmath15 factor .
the other @xmath16 factor improvement is due to a new way of defining and representing the maps @xmath86 and @xmath87 , as elaborated below . in the following discussion .
let @xmath111 be the vertex @xmath110|}(v_i)$ ] . in our new algorithm ,
to determine whether @xmath9 is visible to @xmath10 , we check whether @xmath111 is a triangle witness vertex for @xmath9 and @xmath10 . to this end , we already know that @xmath111 is visible to @xmath9 , but we still need to check whether @xmath111 is visible to @xmath10 . in the previous triangle witness algorithm @xcite , this step is performed in @xmath16 time by representing @xmath86 and @xmath87 using balanced binary search trees . in our new algorithm , we handle this step in @xmath122 time , by redefining @xmath86 and @xmath87 and using a new way to represent them .
we redefine @xmath86 as follows : @xmath88[v_j]=t$ ] if @xmath10 is the @xmath89-th visible vertex to @xmath9 in the ccw order ; if @xmath10 is not visible to @xmath9 or @xmath10 has not yet been identified , then @xmath88[v_j]=0 $ ] . for convenience , we let @xmath88=f[v_i][v_{i+1},v_{i+2},\ldots , v_{i-1}]$ ] .
thus , in our new definition , the size of @xmath88 $ ] is fixed throughout the algorithm , i.e. , @xmath95|$ ] is always @xmath123 .
in addition , for each @xmath9 , the new algorithm maintains two variables @xmath124 and @xmath125 for @xmath88 $ ] , where @xmath124 is the number of non - zero entries in the current @xmath88 $ ] , which is also the number of visible vertices to @xmath9 that have been identified ( i.e. , the number of visible vertices to @xmath9 in the chain @xmath109 ) up to the @xmath82-th iteration , and @xmath125 is the index of the last non - zero entry in the current @xmath88 $ ] , i.e. , @xmath125 is the last visible vertex to @xmath9 in the chain @xmath109 in the ccw order .
similarly , we redefine @xmath87 in the same way as @xmath86 , i.e. , for each @xmath9 , @xmath92=b[v_i][v_{i+1},v_{i+2},\ldots , v_{i-1}]$ ] and @xmath92[v_j]=t$ ] if @xmath10 is the @xmath89-th visible vertex to @xmath9 in the ccw order .
further , for each @xmath9 , we also maintain two variables @xmath126 and @xmath127 for @xmath92 $ ] , where @xmath126 is the number of non - zero entries in the current @xmath92 $ ] , which is also the number of visible vertices to @xmath9 in the chain @xmath128 ( up to the @xmath82-th iteration ) , and @xmath127 is the index of the first non - zero entry in the current @xmath92 $ ] , i.e. , @xmath127 is the first visible vertex to @xmath9 in the chain @xmath128 in the ccw order . during the algorithm ,
the array @xmath88 $ ] will be filled in the ccw order , i.e. , from the first entry @xmath88[v_{i+1}]$ ] to the end while the array @xmath92 $ ] will be filled in the cw order , i.e , from the last entry @xmath92[v_{i-1}]$ ] to the beginning .
when the algorithm finishes , @xmath88 $ ] will contain all the visible vertices to @xmath9 in the chain @xmath129 , and thus only the entries of the first half of @xmath88 $ ] are possibly filled with non - zero values .
similarly , only the entries of the second half of @xmath92 $ ] are possibly filled with non - zero values .
below , we discuss the implementation details of our new algorithm , which is summarized in algorithm [ algo:20 ] .
@xmath130 initially , when @xmath96 , for each @xmath9 , we set @xmath88[v_{i+1}]=1 $ ] and @xmath92[v_{i-1}]=deg(v_i)$ ] , and set all other entries of @xmath88 $ ] and @xmath92 $ ] to zero .
in addition , we set @xmath131 , @xmath132 and @xmath133 , @xmath134 . in the @xmath82-th iteration , with @xmath135 , for each @xmath9 , we check whether @xmath9 is visible to @xmath10 , with @xmath97 .
if @xmath9 is not visible to @xmath10 , then we do nothing . otherwise , we set @xmath88[v_j]=l_i+1 $ ] and increase @xmath124 by one ; similarly , we set @xmath136[v_i]=deg(v_j)-l'_j$ ] and increase @xmath137 by one .
further , we set @xmath138 and @xmath139 .
it remains to show how to check whether @xmath9 is visible to @xmath10 .
by lemma [ lem:20 ] , we need to determine whether @xmath111 is a triangle witness vertex for @xmath9 and @xmath10 .
since @xmath111 is the last visible vertex to @xmath9 in the chain @xmath68 in the ccw order , based on our definition , @xmath111 is the vertex @xmath125 . after knowing @xmath111
, we then check whether @xmath111 is visible to @xmath10 , which can be done by checking whether @xmath136[v']$ ] is zero , in constant time . if @xmath136[v']$ ] is zero , then @xmath111 is not visible to @xmath10 and @xmath111 is not a triangle witness vertex ; otherwise , @xmath111 is visible to @xmath10 .
( note that we can also check whether @xmath140[v_j]$ ] is zero . ) in the following , we assume that @xmath111 is visible to @xmath10 . the next step is to determine whether @xmath114 . to this end , we must know the involved three angles .
similar to the discussion in section [ sec : review ] , we have @xmath141[v_j],b[v'][v_i])$ ] , @xmath142[v'],l_i+1)$ ] , and @xmath143[v'])$ ] .
thus , all these three angles can be obtained in constant time .
hence , the step of checking whether @xmath9 is visible to @xmath10 can be performed in constant time , which reduces another @xmath16 factor from the running time of the previous triangle witness algorithm in @xcite .
algorithm [ algo:20 ] summarizes the whole algorithm .
clearly , the running time of our new algorithm is bounded by @xmath4 . [ theo:10 ] given the visibility angles and an ordered vertex sequence of a simple polygon @xmath0 , the improved triangle witness algorithm can reconstruct @xmath0 ( up to similarity ) in @xmath4 time .
as discussed in @xcite , the above algorithm can also be used to determine whether the input angle data are consistent .
namely , if there is no polygon that can fit the input angle data , then the algorithm in theorem [ theo:10 ] can be used to report this case as well .
10 t. asano , s.k .
ghosh , and t. shermer .
chapter 19 : visibility in the plane , in _ handbook of computational geometry _ , j. sack and j. urrutia ( eds . ) , pages 829876 .
elsevier , amsterdam , the netherlands , 2000 .
t. biedl , s. durocher , and j. snoeyink . reconstructing polygons from scanner data . in _ proc .
of the 20th international symposium on algorithms and computation _ , volume 5878 of _ lecture notes in computer science _ , pages 862871 .
springer - verlag , 2009 .
d. bil , y. disser , m. mihalk , s. suri , e. vicari , and p. widmayer .
reconstructing visibility graphs with simple robots . in _ proc .
of the 16th international colloquium on structural information and communication complexity _ , volume 5869 of _ lecture notes in computer science _ , pages 8799 .
springer - verlag , 2009 .
choi , s.y .
shin , and k .- y . chwa . characterizing and recognizing the visibility graph of a funnel - shaped polygon . , 14(1):2751 , 1995 .
colley , a. lubiw , and j. spinrad .
visibility graphs of towers .
, 7(3):161172 , 1997 .
y. disser , m. mihalk , and p. widmayer .
reconstructing a simple polygon from its angles . in _ proc .
of the 12th scandinavian symposium and workshops on algorithm theory ( swat ) _ , volume 6139 of _ lecture notes in computer science _ , pages 1324 .
springer , 2010 .
g. dudek , m. jenkins , e. milios , and d. wilkes .
robotic exploration as graph construction .
, 7(6):859865 , 1991 .
h. everett . .
phd thesis , university of toronto , toronto , 1990 .
h. everett and d. corneil .
recognizing visibility graphs of spiral polygons .
, 11(1):126 , 1990 .
h. everett and d. corneil .
negative results on characterizing visibility graphs .
, 5(2):5163 , 1995 .
p. flocchini , g. prencipe , n. santoro , and p. widmayer .
hard tasks for weak robots : the role of common knowledge in pattern formation by autonomous mobile robots . in _ proc . of the 10th international symposium on algorithms and computation _ , pages 93102 , 1999 .
l. jackson and s. wismath .
orthogonal polygon reconstruction from stabbing information .
, 23(1):6983 , 2002 .
j. orourke and i. streinu .
the vertex - edge visibility graph of a polygon .
, 10(2):105120 , 1998 .
a. sidlesky , g. barequet , and c. gotsman .
polygon reconstruction from line cross - sections . in _ proc of the 18th canadian conference on computational geometry _ ,
pages 8184 , 2006 .
j. snoeyink .
cross - ratios and angles determine a polygon .
, 22:619631 , 1999 .
s. suri , e. vicari , and p. widmayer .
simple robots with minimal sensing : from local visibility to global geometry . , 27(9):10551067 , 2008 .
s. wismath .
point and line segment reconstruction from visibility information .
, 10(2):189200 , 2000 .
@xmath144 $ ] @xmath145 $ ] @xmath130 | in this paper , we study the following problem of reconstructing a simple polygon : given a cyclically ordered vertex sequence of an unknown simple polygon @xmath0 of @xmath1 vertices and , for each vertex @xmath2 of @xmath0 , the sequence of angles defined by all the visible vertices of @xmath2 in @xmath0 , reconstruct the polygon @xmath0 ( up to similarity ) .
an @xmath3 time algorithm has been proposed for this problem .
we present an improved algorithm with running time @xmath4 , based on new observations on the geometric structures of the problem . since the input size ( i.e. , the total number of input visibility angles ) is @xmath5 in the worst case , our algorithm is worst - case optimal .
= 14.0pt |
in the last decades some generalized entropies , different from the boltzmann - gibbs one , have been proposed in literature to study the statistical mechanics properties of complex systems with asymptotic free - scale behavior . among the various proposals , the tsallis entropy @xcite @xmath0={1\over1-q}\int\big(p(v)^q - p(v)\big)\,dv \ , \ ] ]
is a paradigm .
+ after its introduction in 1988 , it has been fruitfully employed in the investigation of statistical systems characterized by long - rang interactions and time - persistent memory effects . among the many ,
we quote astrophysical and cosmological systems , nuclear systems , low - dimensional chaotic systems , self - organized critical systems , turbulence .
it is also widely employed in geological , biological and social - economical sciences ( see @xcite for a currently update bibliography ) . + as observed in @xcite , the original proposal @xcite introduced the lagrange multiplier associated with the linear mean energy in an unsatisfied manner so that some important properties necessary for a well stated theory were lacked .
this insight induced in @xcite to reformulate the original formalism ( hereinafter we call it the modified 1st - formalism ) based on the definition of the linear mean value given by @xmath1 in @xcite alternative definitions of the mean value were introduced , and the authors proposed the 2nd - formalism based on the un - normalized @xmath2-mean value @xmath3 which unfortunately has the drawback that @xmath4 and the 3rd - formalism based on the escort probability mean value @xmath5 more recently , in @xcite the optimal lagrange multiplier formalism ( olm - formalism ) were introduced , where for the same physical input of the 3rd - formalism , the constraint in the maxent procedure is written as the _ centered_-form @xmath6 the 1st- and 3rd - formalisms have the disadvantage of dealing with self - referential expressions for the distribution functions which originate computational difficulties in some cases .
differently , the 2nd- and the olm - formalisms produce distribution functions that are not self - referential though the mean values are still self - referential ( see @xcite for a comparison between these two formalisms ) .
+ throughout all these alternative formulations , in @xcite , another formalism based on the entropy @xmath7 $ ] and the linear averages ( hereinafter @xmath8-formalism ) has been developed .
in particular , in @xcite it has been shown that the @xmath8-formalism produces a non self - referential expression both in the distribution and in the mean values .
( for a comparison of some physically relevant quantities in the different formalisms see appendix a ) .
+ in @xcite , the entropy @xmath7 $ ] and its equilibrium distribution have been employed to show the equivalence among the different formulations of the tsallis theory . the same @xmath8-formalism plays an important role in the discussion of the present work . for this reason , and also to fix the notations used in this paper , we shortly collect hereinafter the main results obtained in @xcite . + according to the following relations @xmath9 the equilibrium distribution obtained in the @xmath8-formalism @xmath10 with @xmath11 and @xmath12^{1/(1-q)}$ ] is the @xmath2-deformed exponential , can be rewritten in the form @xmath13 with @xmath14 , where , by replacing @xmath15 with @xmath2 , we obtain the probability distribution function of the modified 1st - formalism @xcite . in eq .
( [ pdf0 ] ) , @xmath16 and @xmath17 are the lagrange multipliers related , respectively , to the normalization @xmath18 and to the linear mean energy @xmath19 in the ( @xmath15)-formalism , as well as @xmath20 is related to the lagrange multiplier associated to the linear mean energy @xmath21 in the 1st - formalism .
+ equivalently , the same expression ( [ pdf0 ] ) can be transformed in @xmath22 corresponding to the distribution function derived in the 3rd - formalism @xcite , where @xmath23 is the lagrange multiplier of the escort mean energy @xmath24 and @xmath25 .
+ the expressions ( [ pdf1 ] ) and ( [ pdf3 ] ) are equivalents , in the sense that @xmath26 , if the conditions @xmath27 and @xmath28 \ , \label{u}\ ] ] are fulfilled .
( see also @xcite for an alternative discussion on the equivalence among these different formalisms ) .
+ in a previous tentative to establish a correspondence among the various formulations described above , it has been introduced the escort entropy @xcite @xmath29={1\over1-q}\left\{\frac{\int p^{\rm(e)}(v,\,t)\,dv}{\left[\int \big(p^{\rm(e)}(v,\,t)\big)^{1/q}\,dv\right]^q}-1\right\ } \ .\label{escort}\ ] ] this entropy is derived from @xmath30 $ ] by using the transformation @xmath31 where distributions @xmath32 and @xmath33 came , respectively , from the entropy @xmath30 $ ] constrained by the escort mean value @xmath34 and from the entropy @xmath35 , with @xmath36 , constrained by linear mean value @xmath37 .
+ in the following , we are concerned with self - similar solutions of a nonlinear fokker - planck equation ( nfpe ) derived from the @xmath7 $ ] entropy with a linear drift .
this kinetic equation , and some related versions , are widely considered in the literature @xcite to study anomalous diffusion processes like fractal media diffusion or diffusion in a non - newtonian fluid .
self - similar solutions corresponding to these nfpes were also investigated in @xcite because their relevance in the description of several phenomenologies in condensed matter physics . + the purpose of the present work is twofold .
+ firstly , we extend the discussion held in @xcite to the case of a system out of equilibrium .
this is performed in the next section 2 , where we show that the self - similar solutions derived in the @xmath8-formalism are equivalent to those of the same kinetic equations obtained by employing the modified 1st - formalism and 3rd - formalism .
+ secondly , in section 3 , we show that , by means of a time - scale conversion , there is a `` correspondence '' among the ( self - similar ) solutions of the nfpe derived in the @xmath8-formalism and the ( self - similar ) solutions of the nfpe derived in the escort - formalisms . as a by product
, it is shown that there exists a `` correspondence '' with the ( self - similar ) solutions of the nfpe derived in the 3rd - formalism with a time dependent diffusion coefficient .
to establish a relationship among different nfpes is useful not only in the framework of the statistical mechanics based on the tsallis entropy but in general with the purpose to classify the various kinetic equations in equivalence class such that any equation belonging to the same class describes substantially the same physical process .
the conclusions are given in section 4 whilst a comparison among the different formalisms discussed in this paper is reported in the appendix a.
in order to describe our problem we need to relate the production of entropy to a fokker - planck equation describing the kinetics of the system .
this issue has been discussed in literature by several authors , both in a very general nonlinear kinetic framework @xcite and , more specifically , in the tsallis entropy framework @xcite .
+ quite generally , the nfpe associated with a given entropy @xmath39 $ ] can be obtained from the relation @xmath40 and assuming for the expression of the nonlinear current @xmath41 the quantity @xmath42 is the thermodynamic force , where @xmath43 $ ] is the lyapunov functional for the given problem @xcite .
+ by posing @xmath44 = { \mathcal l}^{(1)}_q[p ] \equiv u^{(1)}_q - d^{(1)}\,s_{2-q}[p ] \ , \label{lyapunov1}\ ] ] where @xmath45=\int\frac{\big(p(v,\,t)\big)^{2-q}-p(v,\,t)}{q-1}\,dv \ , \ ] ] is the tsallis entropy in the @xmath8-formalism and @xmath46 is the constant diffusion coefficient , we obtain the following nfpe @xmath47 \ , \label{fp1}\ ] ] where @xmath48 is the drift coefficient . equation ( [ fp1 ] ) differs , in @xmath49 , from the kinetics equation derived in @xcite from entropy @xmath30 $ ] . the same eq .
( [ fp1 ] ) has been first proposed in @xcite and after reconsidered by different authors @xcite .
+ in the following , we assume @xmath50 with @xmath51 , which deals with a specific but very common expression for the drift coefficient , corresponding to the uhlenbeck - ornstein process in the @xmath52 limit . without loss of generality , we can pose @xmath53 . with this choice for @xmath54 ,
( [ fp1 ] ) admits self - similar solutions @xcite . in the @xmath8-formalism
such solutions are obtained by introducing the ansatz @xmath55 in this way , we derive the following system of differential equations for the unknown quantities @xmath56 and
@xmath57 : @xmath58\,\left(1-d^{(1)}\,\beta^{(1)}(t)\right ) = 0 , \label{s1}\\ & & \frac{d}{d\,t}\,\beta^{(1)}(t ) -(3-q)\,\beta^{(1)}(t)\,\left(1-d^{(1)}\,\beta^{(1)}(t)\right)=0 \ .\label{s2}\end{aligned}\ ] ] moreover , accounting for the constraints on the normalization and the linear mean energy , which give the relations to guarantee the convergence of all integrals on the full real axe @xmath59 .
the results can be straightforward generalized to @xmath60 , accounting for the cut - off condition on the integration interval . ]
@xmath61 with @xmath62 from eqs .
( [ s1 ] ) or ( [ s2 ] ) we obtain the further differential equation @xmath63=0 \ , \label{eve}\ ] ] describing the time evolution of the mean energy . + solutions of eqs .
( [ s1 ] ) , ( [ s2 ] ) and ( [ eve ] ) are given by @xmath64\,\ln_q\left({\beta^{(1)}(t)\over\beta_0}\right)^{1\over3-q } \
, \label{sol1a}\\ & & \beta^{(1)}(t)=\beta_0\bigg[\big(1-\beta_0\,d^{(1)}\big)\,e^{(q-3)\,t } + \beta_0\,d^{(1)}\bigg]^{-1 } \ , \label{sol1b}\\ & & u^{(1)}_q(t)=u_0\,\bigg[\big(1-\mu_0\,d^{(1)}\big)\,e^{(q-3)\,t } + \mu_0\,d^{(1)}\bigg]^{\frac{2}{3-q } } \ , \label{sol1c}\end{aligned}\ ] ] with @xmath65 , @xmath66 and @xmath67 . at equilibrium , @xmath68 , from eq . (
[ sol1b ] ) we obtain the relation @xmath69 , where @xmath70 , which means , as already known @xcite , that the quantity @xmath71 can be identified with the lagrange multiplier related to the mean energy @xmath72 . +
differently , by using in eq .
( [ fp1 ] ) the ansatz @xmath73 for the modified 1st - formalism , we obtain the following set of equations @xmath74 where @xmath75 means @xmath76 and @xmath77 , respectively , with @xmath78 and we posed @xmath79 , according to eq .
( [ b1 ] ) . + by means of the relations ( [ b1 ] ) and ( [ g1 ] ) the system ( [ eq0 ] ) is equivalent to the system ( [ s1 ] ) , ( [ s2 ] ) and ( [ eve ] ) .
correspondingly , its solutions @xmath80 with @xmath81 are equivalent to the solutions ( [ sol1a])-([sol1c ] ) , whilst the time evolution of @xmath77 is still given by eq .
( [ sol1c ] ) . + in a similar way , with the ansatz @xmath82 from eq .
( [ fp1 ] ) we obtain the system of the differential equations @xmath83 where @xmath84 means @xmath85 and @xmath86 , respectively , and @xmath87 with @xmath88 .
we note that at equilibrium ( @xmath89 ) , @xmath90 can be identified with the lagrange multiplier of the escort mean energy @xmath91 . + by using eqs .
( [ beta ] ) and ( [ u ] ) , it is easy to verify that the system ( [ eq1 ] ) can be transformed in the system ( [ eq0 ] ) and therefore , it is also equivalent to the system ( [ s1 ] ) , ( [ s2 ] ) and ( [ eve ] ) .
+ the solutions of eqs .
( [ eq1 ] ) are given by @xmath92 where @xmath93 whilst the solution for the mean energy @xmath94 has the same expression given in eq .
( [ sol1c ] ) , with @xmath95 and @xmath96 , being now @xmath97 and @xmath98 .
in this section , we study a different kind of relationship , which we call `` correspondence '' , between the modified 1st - formalism and 3rd - formalism + to start with , let us observe that by posing @xmath99 through the definitions ( [ zq1 ] ) and ( [ zq3 ] ) we verify that @xmath100 on the other hand , by assuming the same initial values for the both mean energy @xmath101 and by comparing eqs .
( [ sol22 ] ) and ( [ sol3 ] ) we get @xmath102 as a consequence , the following relation @xmath103 and its inverse @xmath104 hold , where we denoted ( @xmath105 respectively ) in the distribution . ] with @xmath106 the escort distribution obtained from the optimization problem with the entropy @xmath107 $ ] and the corresponding linear mean energy @xmath108 .
+ the consistency of eqs .
( [ p1-p3 ] ) and ( [ p3-p1 ] ) can be easily verified as follows .
first , by integrating eq .
( [ p1-p3 ] ) over @xmath109 and utilizing the normalization of @xmath110 and @xmath111 , it reduces to the definition of the partition functions @xmath112 and @xmath113 .
second , by multiplying eq .
( [ p1-p3 ] ) by @xmath54 and integrating over @xmath109 , we can verify that @xmath114 which is the initial condition ( [ cond ] ) .
+ all of these results are already known for the equilibrium system .
notwithstanding , let us now pose the following question : does the `` correspondence '' established through eqs .
( [ p1-p3 ] ) and ( [ p3-p1 ] ) still hold , in some sense , when the system evolves toward the equilibrium according to a suitable nfpes ?
+ for reasons which will be clarified through this section , it is more appropriate to start by studying the `` correspondence '' among @xmath115 and @xmath116 and only successively recover the `` correspondence '' with @xmath117 by means of eq .
( [ p1-p3 ] ) .
in fact , by employing the escort - formalism we get the advantage to deal with a linear expression for the mean energy @xmath118 . in this way , by posing @xmath119 \equiv u^{\rm(e)}_{q'}(t)-d^{\rm ( e)}\,s_{q'}^{\rm ( e)}[p ] \ , \label{lyapunov3}\ ] ] where @xmath120 $ ] is the escort entropy ( [ escort ] ) , we derive the following kinetic equation : @xmath121 \ , \\\label{fpne}\end{aligned}\ ] ] where @xmath122 and @xmath123 is the constant diffusion coefficient for the given problem .
this equation , for @xmath124 , becomes equivalent to the nfpe obtained previously in @xcite .
+ let us observe that eqs .
( [ fp ] ) and ( [ j ] ) provide a general mechanism to construct nfpes starting from a give lyapunov functional @xmath43 $ ] .
this method works also when the mean energy term in the lyapunov functional has a nonlinear dependence on the probability distribution functions like , for instance , in the case of @xmath125 .
notwithstanding , in these cases the resulting nonlinear current @xmath126 will contain , among to a nonlinear diffusion term , also a nonlinear drift term .
this fact makes problematic the comparison between the different formalisms which is the main purpose of the present work . a possible way to avoid such complication
is actually furnished by means of the escort formalism centered on the linear definition for the mean energy @xmath127 .
+ come back to the original question , self - similar solutions of eq .
( [ fpne ] ) can be obtained by posing @xmath128^{q'\over1-q ' } \ , \label{est}\ ] ] where @xmath129 and at equilibrium @xmath130 is the lagrange multiplier related to the escort mean energy @xmath131 .
+ we get the following system of equations @xmath132 where @xmath133 means @xmath134 and @xmath135 , respectively , and @xmath136 the solution of the system ( [ eq2 ] ) can be easily written in the form @xmath137 and @xmath138^{\frac{2}{1+q'}}\ ] ] with @xmath139 .
let us now introduce the two different time - scales ( @xmath140 and @xmath141 ) according to the relation @xmath142 accounting for the expressions ( [ f1 ] ) and ( [ fe ] ) , eq .
( [ tt ] ) can be integrated in @xmath143^{q'}-d^{\rm(e)}\,\mu_0^{\rm(e)}\right\ } \ , \\
\label{time}\end{aligned}\ ] ] where @xmath144 by converting the time - scale @xmath140 into @xmath141 , we can easily verify that the set of equations for the modified 1st formalism are transformed to those for the escort formalism @xmath145 which implies the following `` correspondence ''
@xmath146 since the quantities @xmath147 and @xmath148 determine completely the shape of the distribution @xmath115 and that of @xmath149 , respectively , we conclude that the `` correspondence '' stated in eq .
( [ p1-p3 ] ) at the initial time actually holds for any later time , i.e. @xmath150 finally , taking into account the expressions ( [ 1st ] ) and ( [ est ] ) , from the relation ( [ pp ] ) we obtain @xmath151\hspace{10mm}\rightarrow \\ \nonumber & & \hspace{-8mm}\frac{\partial}{\partial t ' } p_{q'}^{\rm(e)}(v,\,t')=\frac{\partial}{\partial v}\left[v\,p_{q'}^{\rm(e)}(v,\,t')+d^{\rm(e)}\,\left(\bar z_{q'}^{\rm(e)}(t')\right)^{{1\over q'}-q'}\frac{\partial}{\partial v}\big(p_{q'}^{\rm(e)}(v,\,t')\big)^{1\over q'}\right ] \ , \\\end{aligned}\ ] ] i.e. , concerning self - similar solutions , both the nfpes ( [ fp1 ] ) and ( [ fpne ] ) describe the same kinetic process when considered in the own time - scale @xmath140 and @xmath141 , respectively .
+ in figure 1 , we plot the relevant quantities @xmath152 [ resp . @xmath153 $ ] , @xmath154 [ resp . @xmath155 $ ] , @xmath156 [ resp . @xmath157 $ ] and @xmath158 [ resp .
@xmath159 $ ] , for the case @xmath160 ; @xmath161 and @xmath162 . in all panels
the full line depicts the evolution of these quantities in the modified 1st - formalism v.s .
time @xmath140 , whilst the dashed line depicts the evolution of the same quantities in the escort - formalism v.s .
the time @xmath140 . finally , the dotted line depicts the evolution of all these quantities in the escort - formalism v.s . the transformed time @xmath141 .
we can see that all the functions in the escort - formalism , referred to the time @xmath141 , coincide with the corresponding ones in the modified 1st - formalism , referred to the time @xmath140 for all time .
+ as known , some concepts like irreversibility of the process described by the fokker - planck equation or the @xmath163-theorem ensuring uniqueness and stability of equilibrium distribution , can be well described by means of the lyapunov function .
existence of a lyapunov function for eqs .
( [ fp1 ] ) and ( [ fpne ] ) in connection with the @xmath163-theorem has been investigated in @xcite and more in general , for an arbitrary nfpe , in @xcite . in the following ,
we make some considerations concerning the `` correspondence '' among the lyapunov functionals @xmath164 $ ] and @xmath165 $ ] arising in the two formalisms .
+ firstly , we observe that the both functionals @xmath166 $ ] given in eqs .
( [ lyapunov1 ] ) and ( [ lyapunov3 ] ) , where i=(1 ) or ( e ) , are actually lyapunov functionals for the kinetic equations ( [ fp1 ] ) and ( [ fpne ] ) , respectively , since they fulfill the relation @xmath167 where equality holds at equilibrium ( @xmath168 ) . according to eq .
( [ en ] ) we have @xmath169 which relates the variation in time of the lyapunov functions in both formalisms . + by recalling eqs .
( [ eq0 ] ) and ( [ eq2 ] ) , with @xmath170 , we can rewrite eq . ( [ 3 ] ) in the form @xmath171
so that @xmath172 i.e. , the functions @xmath173 , responsible for the different time - scale in the two formalisms , originate from the variation of the lyapunov function with respect to the mean energy .
+ finally , let us remind that the distributions @xmath174 and @xmath175 are closely related to each other according to eq .
( [ p1-p3 ] ) . by inserting this transformation into eq .
( [ fpne ] ) we obtain the following kinetic equation for the distribution @xmath117 @xmath176 \ , \\\label{fp33}\end{aligned}\ ] ] where we assumed @xmath177 . by means of the relation @xmath178 equation ( [ fp33 ] ) assumes the same expression of eq .
( [ fp1 ] ) , but with a time dependent diffusion coefficient .
+ it is worthy to observe that , contrarily to the nfpes ( [ fp1 ] ) and ( [ fpne ] ) , eq . ( [ fp33 ] )
does not follow from eqs .
( [ fp ] ) and ( [ j ] ) and it has been derived just on the basis of the `` correspondence '' between self - similar solutions in the different formalism . from a formal point of view , we can state that kinetic evolution of self - similar solutions , both in the @xmath8-formalism and in the 3rd - formalism , are described by the same nfpe with a constant diffusion coefficient @xmath46 in the former case and a time dependent diffusion coefficient @xmath179 in the later case ( or vice versa ) .
remarkably , since @xmath180 and at equilibrium @xmath181 , eq .
( [ dd ] ) reduces to eq . ( [ beta ] ) with @xmath2 replaced by @xmath105
. + by following the same steps described at the begin of this section , by replacing the relation @xmath182 in eq .
( [ fp33 ] ) , it can be rewritten in the form @xmath183 in this way , by introducing the ( time - evolution ) function @xmath184 we can establish a `` correspondence '' between self - similar solutions in the modified 1st - formalism and those of the 3rd - formalism , in the sense that , according to the transformation @xmath185 and @xmath186 the nfpe ( [ fp33 ] ) turns into eq .
( [ fp1 ] ) .
in the present letter , we have compared self - similar solutions of some nfpes obtained from different generalized entropies . for the sake of simplicity
we assumed a quadratic form of the energy density but extension to a more general expression @xmath187 , with @xmath188 , is straightforward .
+ our results can be summarized in the following two main points .
+ 1 ) according to eqs .
( [ beta ] ) and ( [ u ] ) we have shown that solutions @xmath189 of the nfpe ( [ fp1 ] ) , derived in the @xmath8-formalism , can be written in the form @xmath190 obtained in the modified 1st - formalism or in the form @xmath191 obtained in the 3rd - formalism .
+ 2 ) we have shown that the solution @xmath190 of the nfpe ( [ fp1 ] ) `` corresponds '' to the solution @xmath116 of the nfpe ( [ fpne ] ) derived in the escort - formalism , where @xmath99 .
remark that this relation embodies both dualities @xmath192 and @xmath193 .
these two transformations appear recurrently , alone or combined , in the framework of the generalized statistical mechanics based on the tsallis entropy @xcite .
+ the `` correspondence '' @xmath194 is established by means of different time - scale , stated through eq .
( [ tt ] ) . as a consequence
, it has been shown that the ( self - similar ) solutions , obtained in the modified 1st - formalism and the ones obtained in the 3rd - formalism , obey to similar kinetic equations where the diffusion coefficients are related according to eq .
( [ dd ] ) .
+ this paper makes some progress about the still discussed question of the equivalence between the various formalisms , introduced in the generalized statistical mechanics .
we have shown that the formulation of the theory centered around the escort mean values ( [ esc ] ) , whose physical meaning is still unclear at all , are in any way valid , being related to the same results obtained by employing the orthodox mean value definition ( [ lin ] ) .
this encourages the use of the escort probabilities which , in some cases , expedites the derivation of several results concerning this theory .
+ on a general basis , we have verified that concerning the class of self - similar solutions , there are some different nonlinear diffusive equations which describe the same kinetic process when it is observed in the appropriate time - scale .
the time - scale is derived by means of the lyapunov function which generates the corresponding kinetics evolution equation . in this way , by modifying the expression of the lyapunov function , we can produce a family of nonlinear fokker - planck equations belonging to the same equivalent class , in the sense discussed in this work
. it should be interesting to verify the existence of these equivalences also for more general kind of solutions .
we summarize , for the sake of comparison , the expressions of some physically relevant quantities like the entropy , the mean energy constraint and the distribution for the different formalisms discussed in this paper . + all these quantities
are related by means of the following variational principle @xmath195-\gamma\,\phi_0[p]-\beta\,\phi_1[p]\big)=0 \ , \ ] ] where @xmath196 and @xmath197 are the lagrange multipliers related to the constraints ( given through @xmath198 $ ] and @xmath199 $ ] ) of the normalization and that of the mean energy , respectively .
+ a ) ( @xmath15)-formalism .
@xmath200={1\over q-1}\int \left[p(v)^{2-q}-p(v)\right]\,dv \ , \\ & & \phi_1[p]=\int\epsilon(v)\,p(v)\,dv - u_q \ , \\ & & p(v)=\alpha\,\exp_q\big(-\gamma-\beta\,\epsilon(v)\big ) \ .\end{aligned}\ ] ] \b ) modified 1st - formalism .
@xmath201={1\over 1-q}\int\left[(p^{(1)}(v))^q - p^{(1)}(v)\right]\,dv \ , \\ & & \phi_1[p]=\int\epsilon(v)\,p^{(1)}(v)\,dv - u_q^{(1 ) } \ , \\ & & p^{(1)}(v)={1\over\bar z_q^{(1)}}\,\exp_{2-q}\big(-\beta_q^{(1)}\,\big(\epsilon(v)-u^{(1)}_q\big)\big ) \ , \\ & & \beta^{(1)}_q={\beta\over q}\,\left(\bar z_q^{(1)}\right)^{q-1 } \ , \\ & & \bar z_q^{(1)}=q^{1/(q-1)}\,\exp_q(\gamma+\beta\,u^{(1)}_q ) \ .\end{aligned}\ ] ] \c ) 2nd - formalism .
@xmath201={1\over1-q}\int \left[(p^{(2)}(v))^q - p^{(2)}(v)\right]\,dv \ , \\ & & \phi_1[p]=\int\epsilon(v)\,(p^{(2)}(v))^q\,dv - u_q^{(2 ) } \ , \\ & & p^{(2)}(v)={1\over z_q^{(2)}}\,\exp_q\big(-\beta\,\epsilon(v)\big ) \ , \\ & & z_q^{(2)}=q^{1/(q-1)}\,\exp_q(\gamma ) \ .\end{aligned}\ ] ] \d ) 3rd - formalism .
@xmath201={1\over1-q}\int \left[(p^{(3)}(v))^q - p^{(3)}(v)\right]\,dv \ , \\ & & \phi_1[p]={\int\epsilon(v)\,(p^{(3)}(v))^q\,dv\over\int ( p^{(3)}(v))^q\,dv}-u_q^{(3 ) } \ , \\ & & p^{(3)}(v)={1\over\bar z_q^{(3)}}\,\exp_q\big(-\beta_q^{(3)}\,\big(\epsilon(v)-u^{(3)}_q\big)\big ) \ , \\ & & \beta^{(3)}_q=\beta\,\left(\bar z_q^{(3)}\right)^{q-1 } \ , \\ & & \bar z_q^{(3)}=q^{1/(q-1)}\,\exp_q(\gamma ) \ .\end{aligned}\ ] ] \e ) olm - formalism .
@xmath201={1\over1-q}\int \left[(p^{(\rm olm)}(v))^q - p^{(\rm olm)}(v)\right]\,dv \ , \\ & & \phi_1[p]={\int\epsilon(v)\,(p^{\rm(olm)}(v))^q\,dv - u_q^{\rm(olm)}\int ( p^{\rm(olm)}(v))^q\,dv } \ , \\ & & p^{\rm(olm)}(v)={1\over\bar z_q^{\rm(olm)}}\,\exp_q\big(-\beta\,\big(\epsilon(v)-u^{\rm(olm)}_q\big)\big ) \ , \\ & & \bar z_q^{\rm(olm)}=q^{1/(q-1)}\,\exp_q(\gamma ) \ .\end{aligned}\ ] ] \f ) escort - formalism .
@xmath202={1\over1-q}\left\{{\int p^{\rm(e)}(v)\,dv\over\left[\int\left(p^{\rm ( e)}(v)\right)^{1/q}\,dv\right]^q}-1\right\ } \ , \\ & & \phi_1[p]=\int\epsilon(v)\,p^{\rm(e)}(v)\,dv - u_q^{\rm(e ) } \ , \\ & & p^{\rm(e)}(v)={1\over \bar z_q^{(e)}}\,\left[\exp_q\left(-\beta_q^{\rm(e ) } \big(\epsilon(v)-u^{\rm ( e)}_q\big)\right)\right]^q \ , \\ & & \beta_q^{\rm(e)}=\beta\,\left(\bar z_q^{\rm(e)}\right)^{q-1 } \ , \\ & & \gamma=-\beta\,u^{\rm(e)}_q \ .\end{aligned}\ ] ] | in a recent paper [ phys . lett .
a * 335 * , 351 ( 2005 ) ] the authors discussed the equivalence among the various probability distribution functions of a system in equilibrium in the tsallis entropy framework . in the present letter
we extend these results to a system which is out of equilibrium and evolves to a stationary state according to a nonlinear fokker - planck equation . by means of time - scale conversion
, it is shown that there exists a `` correspondence '' among the self - similar solutions of the nonlinear fokker - planck equations associated with the different tsallis formalisms .
the time - scale conversion is related to the corresponding lyapunov functions of the respective nonlinear fokker - planck equations .
tsallis entropy , nonlinear fokker - planck equation , escort probability , lyapunov function . |
let @xmath0 be a simple connected graph with vertex set @xmath1 , @xmath2 and @xmath3-adjacency matrix @xmath4 .
an eigenvalue @xmath5 of @xmath4 is said to be a main eigenvalue of @xmath0 if the eigenspace @xmath6 is not orthogonal to the all-1 vector @xmath7 , i.e. , it has an eigenvector the sum of whose entries is not equal to zero .
an eigenvector @xmath8 is a main eigenvector if @xmath9 . for details of the main eigenvector of adjacency matrix ,
the readers are suggested to refer to @xcite .
very recently , we introduced in @xcite the main signless laplacian eigenvalue of a graph .
let @xmath10 , where @xmath11 is the diagonal matrix of vertex degrees and @xmath12 is the degree of vertex @xmath13 .
then @xmath14 is called the signless laplacian matrix of @xmath0 .
some results on the signless laplacian matrix of a graph can be found in @xcite .
an eigenvalue @xmath15 of @xmath16 is said to be a main signless laplacian eigenvalue of @xmath0 ( or a main eigenvalue of @xmath14 ) if the eigenspace @xmath17 is not orthogonal to the all-1 vector @xmath7 , i.e. , it has an eigenvector the sum of whose entries is not equal to zero .
it was showed in @xcite that a graph @xmath0 with exactly one main signless laplacian eigenvalue if and only if @xmath0 is regular , trees and unicyclic graphs with exactly two main signless laplacian eigenvalues were characterized .
a graph @xmath0 is called a bicyclic graph , if it is a simple connected graph in which the number of edges equals the number of vertices plus one .
the aim of this work is to characterize all bicyclic graphs with exactly two main signless laplacian eigenvalues .
in @xcite , an alternative characterization of a graph with exactly two main signless laplacian eigenvalues is given by 2-walks .
a graph @xmath0 is 2-walk * @xmath18-parabolic * @xcite if there exist uniquely a positive integer @xmath19 and a non - negative integer @xmath20 such that such that @xmath21 and @xmath22 holds for all @xmath23 , where @xmath24 and @xmath25 is the set of all neighbors of @xmath26 in @xmath0 .
note that any 2-walk @xmath27-parabolic graph is irregular , since one has @xmath28 and @xmath29 for a @xmath30-regular graph , i.e. , @xmath31 or @xmath32 is not unique . *
lemma 1 * @xcite .
a graph @xmath0 has exactly two main signless laplacian eigenvalues if and only if there exist uniquely a positive integer @xmath19 and a non - negative integer @xmath20 such that @xmath21 and @xmath33 for all @xmath23 , i.e. , @xmath0 is a 2-walk @xmath18-parabolic graph . in @xcite
, we showed that a graph has exactly one main signless laplacian eigenvalue if and only if it is regular .
so , a 2-walk @xmath27-parabolic graph @xmath0 is non - regular , and there are @xmath34 such that @xmath35 . from @xmath36 and @xmath22 , we have @xmath37 * remark 1*. if @xmath0 is a 2-walk @xmath18-parabolic graph with @xmath38 , then @xmath39 since there is a pendent vertex @xmath40 with the only incident edge @xmath41 in @xmath0 and @xmath42 .
let @xmath43 @xmath44 .
for each @xmath45 , @xmath46 is the graph obtained from @xmath0 by deleting all pendant vertices . if @xmath47 , we use @xmath48 to denote the degree of the vertex @xmath26 in @xmath46 .
the following lemmas are true for unicyclic graphs , and it is easy to see that they are also true for bicyclic graphs from their proofs in @xcite . * lemma 2 * @xcite . if @xmath49 and @xmath50 , then @xmath51 or @xmath52
* lemma 3 * @xcite . if @xmath53 , then ( i ) @xmath54 ; ( ii ) @xmath55 and @xmath56 . *
lemma 4*. let @xmath57 and @xmath58 a cycle of @xmath0 .
if @xmath59 and @xmath60 , then there is @xmath61 such that @xmath62 .
* proof*. by the definition of @xmath46 , @xmath63 is also a cycle of @xmath46 .
if @xmath64 , then from ( [ eq:1 ] ) @xmath65 and @xmath66 so , @xmath67 , it contradicts ( [ eq:3 ] ) .
@xmath68 if @xmath69 is a cycle or a path of @xmath0 , then the length of @xmath69 , denoted by @xmath70 , is defined as the number of edges of @xmath69 .
* lemma 5*. let @xmath0 be a 2-walk @xmath18-parabolic graph .
@xmath71 is a path or cycle of length at least 2 in @xmath0 such that @xmath72 , @xmath73 and @xmath74 . then ( i ) if @xmath75 , then @xmath76 ; moreover , if @xmath77 , then there exists no path @xmath78 in @xmath0 such that @xmath79 and @xmath80 ; ( ii ) if @xmath81 , then @xmath82 .
* proof*. ( i ) if @xmath83 , then @xmath84 . by ( [ eq:1 ] ) , we have @xmath85 , and @xmath86 .
this contradicts @xmath72 .
so , @xmath76 .
if @xmath77 , and there is a path @xmath78 in @xmath0 such that @xmath79 and @xmath80 , then by ( [ eq:1 ] ) , @xmath87 , and @xmath88 .
this is impossible since @xmath89 .
\(ii ) if @xmath90 , then @xmath91 by ( [ eq:1 ] ) . on the other hand , @xmath92 and @xmath93 , a contradiction since @xmath81 . @xmath68
* lemma 6*. let @xmath49 , @xmath94 a path or cycle of @xmath46 with length at least 3 such that @xmath95 and @xmath96 for @xmath97 .
\(i ) if there is @xmath98 such that @xmath99 , then @xmath77 and @xmath100 .
moreover , if @xmath101 , then @xmath102 , @xmath103 , @xmath104 , @xmath105 ; or @xmath106 , @xmath107 .
if @xmath108 , then @xmath109 , @xmath110 , @xmath104 , @xmath111 ; or @xmath112 , @xmath107 .
\(ii ) if @xmath113 , then @xmath114 . moreover , if @xmath69 is a cycle , then @xmath77 and @xmath115 .
* proof*. ( i ) by lemma 2 , @xmath116 , @xmath117 . if there is @xmath118 such that @xmath119 , then @xmath120 and @xmath121 , or @xmath122 and @xmath123 . without loss of generality , assume that @xmath120 , @xmath121 . from ( [ eq:2 ] )
, we have @xmath124 where @xmath40 is a pendent vertex of @xmath0 , and @xmath125 @xmath126 since @xmath127 .
so , @xmath128 .
* case 1*. @xmath100 .
then @xmath129 , and @xmath130 from lemma 2 .
if @xmath131 , then @xmath132 .
so , @xmath133 , @xmath134 . because @xmath135 and @xmath136 , by lemma 2 , we have @xmath137 and @xmath138 . from ( [ eq:1 ] ) , @xmath139 . on the other hand , @xmath140 .
so , @xmath141 . this implies @xmath142 , a contradiction with @xmath59 .
so , @xmath143 , and @xmath144 . since @xmath120 , @xmath145 from ( [ eq:1 ] ) , and @xmath146 . by lemma 2 , @xmath147
so , @xmath148 , and @xmath149 , @xmath150 . if @xmath101 , then @xmath151 from above , and @xmath102 . @xmath152 and @xmath153
similarly , if @xmath108 , then @xmath154 , and @xmath109 .
@xmath104 and @xmath111 . *
case 2*. @xmath155 . if @xmath156 , then from ( [ eq:4 ] ) and lemma 3 , @xmath157 .
this contradicts lemma 2 whether @xmath94 a path or cycle . if @xmath158 , then from ( [ eq:4 ] ) and lemma 3 , @xmath159 . by lemma 2
, we have @xmath160 .
so , @xmath143 , and @xmath161 .
this implies @xmath162 and @xmath102 . from ( [ eq:1 ] ) , @xmath163 . on the other hand , @xmath164 , a contradiction . *
case 3*. @xmath165 . from ( [ eq:4 ] ) , @xmath166 .
this contradicts lemma 2 .
\(ii ) it is obvious that @xmath167 from lemma 2 , since @xmath168 and @xmath169 . moreover , if @xmath69 is a cycle and @xmath170 , then @xmath171 from lemma 4 , and @xmath137 from lemma 2 . by ( [ eq:1 ] ) , @xmath172
on the other hand , @xmath173 .
so , @xmath174 if @xmath175 , then @xmath176 .
this is impossible . if @xmath155 , then @xmath177 .
this contradicts @xmath171 . if @xmath165 , then @xmath178 . but
@xmath171 from lemma 4 , it contradicts lemma 2 .
so , @xmath179 . and
@xmath77 from lemma 5(i ) .
@xmath68 * lemma 7*. let @xmath49 , @xmath180 and @xmath181 .
if @xmath182 and @xmath183 , then @xmath184 if and only if @xmath185 . * proof*. from ( [ eq:1 ] ) and @xmath181 , we have @xmath186 . on the other hand , @xmath187 and @xmath188
so , @xmath184 if and only if @xmath185 .
in the following , we determine all bicyclic graphs with exactly two main signless laplacian eigenvalues . from lemma 6 , we know that @xmath197 and ( i ) if @xmath198 , @xmath199 , then @xmath200 ; ( ii ) if @xmath198 , @xmath201 ( or @xmath202 , @xmath199 ) , then @xmath203 ; ( iii ) if @xmath202 , @xmath201 , then @xmath204 . from lemma 6 , we know that @xmath197 and ( i ) if @xmath202 , @xmath199 ( or @xmath198 , @xmath201 ) , then @xmath214 ; ( ii ) if @xmath202 , @xmath201 , then @xmath215 ; ( iii ) if @xmath198 , @xmath199 , then @xmath216 .
\(i ) in the graph @xmath217 , we have @xmath102 , @xmath100 , @xmath104 , @xmath105 , @xmath218 , @xmath219 , @xmath220 by lemma 6 . @xmath221 by using eq.([eq:1 ] ) to @xmath222 , and it can be found that a pendant vertex @xmath40 of @xmath223 does not satisfy eq.([eq:1 ] ) .
so , @xmath217 is not a 2-walk @xmath18-parabolic graph .
\(ii ) in the graph @xmath224 , we have @xmath102 , @xmath100 , @xmath104 , @xmath105 , @xmath218 , @xmath219 , @xmath225 , @xmath226 by lemma 6 . first , @xmath227 by using eq.([eq:1 ] ) to @xmath228 , and then @xmath229 by using eq.([eq:1 ] ) to @xmath230 , and finally @xmath231 by using eq.([eq:1 ] ) to @xmath232 . if @xmath233 , then @xmath234 , i.e. , @xmath223 does not satisfy eq.([eq:1 ] ) ; otherwise , a pendant vertex @xmath40 of @xmath235 does not satisfy eq.([eq:1 ] ) .
so , @xmath224 is not a 2-walk @xmath18-parabolic graph .
first , we show that @xmath237 . if @xmath238 , then @xmath239 and @xmath177 from lemma 2 . by eq.([eq:1 ] ) , we have @xmath240 . on the other hand , @xmath241 .
so , @xmath242 and @xmath133 .
it implies that @xmath243 since @xmath20 is a nonnegative integer and @xmath244 , and @xmath245 . using eq.([eq:1 ] ) again , @xmath246 , a contradiction .
so , @xmath218 .
similarly ( or by symmetry ) , @xmath226 .
if there @xmath262 such that @xmath119 , then by lemma 6(i ) , @xmath263 ( i.e. , @xmath264 ) , @xmath100 , @xmath102 , @xmath103 , @xmath265 , @xmath104 , @xmath105 or @xmath106 , @xmath107 . by symmetry , we may assume that @xmath104 , @xmath105 .
using eq.([eq:1 ] ) , @xmath266 implies @xmath267 .
it can be obtained that @xmath268 by using lemma 7 for @xmath269 and @xmath270 , respectively . and @xmath271 ( i.e. , @xmath272 ) from lemma 6(i ) .
now , @xmath273 , it contradicts eq.([eq:1 ] ) .
so , @xmath259 . in the following , we show that @xmath300 .
otherwise , @xmath301 from claim 1 . by eq.([eq:1 ] ) , @xmath302 since @xmath281 . but
@xmath303 and @xmath304 , it implies @xmath274 , a contradiction .
so , @xmath300 , and @xmath305 , @xmath306 .
we only need to prove that @xmath218 by claim 2 .
if @xmath320 , then @xmath177 by lemma 2 , and @xmath321 from claim 3 , at least two of @xmath314 have degree 2 . without loss of generality , assume that @xmath322 . by eq.([eq:1 ] ) , @xmath289 . on the other hand , @xmath323 and @xmath324 .
we have @xmath325 . and
( i ) @xmath326 , implying that @xmath126 and @xmath327 ; ( ii ) @xmath328 , implying that @xmath329 .
so , @xmath100 .
if there is a 2-walk @xmath18-parabolic graph @xmath0 such that @xmath252 , let @xmath40 be a pendent vertex of @xmath0 and @xmath41 is the only edge incident with @xmath40
. then , @xmath336 by claim 4 and lemma 3 , @xmath337 and @xmath338 by lemma 2 .
let @xmath339 , @xmath340 by lemma 2 . by eq.([eq:1 ] ) , @xmath341 , i.e. , @xmath342 .
so , @xmath343
. the left inequality implies @xmath344 , and the right inequality implies that @xmath345 , i.e. , @xmath346 , since @xmath347 by lemma 3 .
so , @xmath100 or @xmath155 .
\(i ) if @xmath100 , then by eq.([eq:8 ] ) , @xmath352 , and @xmath353 , @xmath306 .
so , @xmath109 . by eq.([eq:1 ] ) , we have @xmath354 , and @xmath355 . since @xmath356 by lemma 2 and @xmath357 , @xmath358 and @xmath359
so , @xmath360 , @xmath361 , and @xmath362 , i.e. , @xmath363 , contradicts that @xmath0 is a simple graph .
\(ii ) if @xmath155 , then by eq.([eq:8 ] ) , @xmath364 . if @xmath305 , then @xmath306 , and @xmath365 which implies @xmath102
. otherwise , @xmath366 , and @xmath367 by lemma 2 .
so , @xmath368 and @xmath102 . in a word
, we have @xmath102 , @xmath155 , @xmath369 .
now , @xmath370 from claim 1 , @xmath371 and @xmath372 from claims 2 - 4 .
moreover , @xmath373 from lemma 5 , i.e. , @xmath189 .
also , it is easy to check that @xmath374 is a 2-walk @xmath375-parabolic graph . *
case 2*. @xmath383
. then @xmath379 from lemma 6(ii ) , and @xmath384 or @xmath385 from lemma 5(i ) .
so , @xmath386 or @xmath387 . it can be checked easily that @xmath388 and @xmath389 are 2-walk @xmath390-parabolic and @xmath391-parabolic graphs , respectively . * case 3*. @xmath392 . without loss of generality ,
we assume that @xmath393 .
then @xmath394 since @xmath0 is a simple graph .
by lemma 5(i ) , @xmath395 ( @xmath396 ) and if one of @xmath397 is 3 , then the other two are not 2 .
so , @xmath398 or @xmath399 if @xmath286 ; @xmath400 if @xmath401 ; @xmath402 if @xmath401
. also , @xmath403 are 2-walk @xmath375-parabolic , @xmath390-parabolic , @xmath404-parabolic and @xmath391-parabolic graphs .
therefore , a bicyclic graph @xmath0 with @xmath376 has exactly two main signless laplacian eigenvalues has exactly two main signless laplacian eigenvalues if and only if @xmath0 is isomorphic to one of @xmath377 . @xmath68 | a signless laplacian eigenvalue of a graph @xmath0 is called a main signless laplacian eigenvalue if it has an eigenvector the sum of whose entries is not equal to zero . in this paper , all connected bicyclic graphs with
exactly two main eigenvalues are determined . * keywords * : spectra of a graph , main signless laplacian eigenvalue , bicyclic graph . * ams classification * : 05c50 = 0.30 in |
the discovery of superconductivity in the doped-@xmath8fe@xmath9as@xmath9 ( @xmath8 = ba , ca , sr and eu ) spin density wave compounds has generated a great deal of activity in search of new compounds , particularly with ba , ca , sr and eu which exhibit interesting magnetic and superconducting properties @xcite . very recently ,
et al _ have reported superconductivity at 2.25 k in baptsi@xmath0 which crystallizes in the tetragonal banisn@xmath0-type non - centrosymmetric crystal structure with the space group @xmath10 @xcite .
compounds possessing a non - centrosymmetric crystal structure have been studied quite extensively after the discovery of superconductivity in cept@xmath0si ( @xmath11 = 0.75 k ) @xcite .
several other ce compounds with the general formula ce@xmath12si@xmath0 , where @xmath12 is a transition metal ( rh , ir and co ) crystallizing in the banisn@xmath0-type crystal structure were later found to exhibit superconductivity under pressure @xcite .
the anisotropic magnetic properties of ceptsi@xmath0 were studied by kawai _ et al _ and they observed that the ce moments order antiferromagnetically with two successive transitions at 4.8 k and 2.4 k respectively @xcite . in this work ,
we report on the crystal growth , electrical and magnetic properties of euptsi@xmath0 which also crystallizes in the banisn@xmath0 type crystal structure .
interest in the eu compounds stems from their unusual magnetic properties like valence fluctuation or magnetic ordering due to the strong coloumb interaction between the @xmath13 and conduction electrons .
eu compounds are also appealing as the charged state and the magnetic interactions of the eu ions can be probed by the technique of @xmath3eu mssbauer spectroscopy , supplementing the information derived from the usual bulk techniques like magnetization , resistivity and heat capacity .
we find that the magnetic properties of euptsi@xmath0 show an unexpected anisotropy ( eu@xmath14 is an @xmath15-state ion , with @xmath16 = 0 ) , whose origin is discussed .
the specific heat and mssbauer data suggest that two magnetic transitions take place at @xmath1 = 17 k and @xmath2 = 16 k. the transition at 17 k is a paramagnetic incommensurate magnetic transition followed by a lock - in transition to a single moment , commensurate phase at 16 k ; a phenomenon which is encountered in other eu@xmath14 or gd@xmath17 compounds .
the single crystals of euptsi@xmath0 were grown by the high temperature solution growth using the fourth element , sn as the solvent as reported by kawai _ et al _
@xcite for the crystal growth of ceptsi@xmath0 .
we took the charge of high purity individual metals eu , pt , si and sn in the ratio 1:1:3:19 . the metals were placed in a recrystallized alumina crucible and sealed in a quartz ampoule with a partial pressure of argon gas .
the sealed crucible was then slowly heated to 1050 @xmath18c and kept at this temperature for one day to achieve proper homogenization .
then the temperature of the furnace was cooled down to 500 @xmath18c over a period of 3 weeks .
the grown crystals were extracted out of the sn solvent by means of centrifuging .
several platelet like single crystals with typical dimension 4 @xmath193 @xmath19 0.6 mm@xmath20 were obtained .
the [ 001 ] axis is perpendicular to the plane of crystals .
our attempts to make captsi@xmath0 by this method were unsuccessful and resulted in the formation of good single crystals of ca@xmath9pt@xmath0si@xmath21 which has recently been reported by takeuchi _
et al _ @xcite . a polycrystalline sample of laptsi@xmath0 , as reference
, non - magnetic analogue of the eu compound , was prepared by the usual method of arc melting in an inert atmosphere of argon and annealed at 900 @xmath22c for two weeks .
the dc magnetic susceptibility and the in - field magnetization measurements were performed in the temperature range 1.8 - 300 k using a quantum design superconducting quantum interference device ( squid ) magnetometer and oxford vibrating sample magnetometer ( vsm ) .
the temperature dependence of electrical resistivity in the range 1.8 - 300 k was measured using a home made dc electrical resistivity set up . the heat capacity and magnetoresistance measurements
were performed using a quantum design physical properties measurement system .
@xmath3eu mssbauer spectra were recorded at various temperatures using a conventional acceleration spectrometer with a @xmath3smf@xmath0 source .
laue diffraction spots were recorded on a huber laue diffractometer , while powder diffraction pattern was recorded using phillips pan - analyitcal set - up .
electron probe micro - analysis ( epma ) measurements were performed on a cameca sx100 electron microprobe .
in order to study the magnetic properties along the principal crystallographic directions , the single crystal of euptsi@xmath0 was subjected to laue diffraction ; the good quality of the single crystals with four fold symmetry was confirmed by laue diffraction spots .
the flat plane of the crystal corresponded to the ( 001 ) plane .
the crystal was then cut along the principal crystallographic directions by means of spark erosion cutting machine .
since the single crystal was grown using sn as flux , a few pieces of the single crystals were crushed into powder form and subjected to powder x - ray diffraction to check the phase purity of the sample . from the rietveld analysis , it is confirmed that the sample possesses tetragonal crystal structure with the space group @xmath10 .
the estimated lattice constants are @xmath23 = 4.2660 and @xmath24 = 9.8768 with the unit cell volume 179.74 @xmath20 .
weak intensity peaks corresponding to free sn were also detected in the powder spectra even after the surface of the crystal was thoroughly scrapped to remove any superficial sn on the surface .
euptsi@xmath0 has a layered structure and a thin layer of free sn may get incorporated between the layers during the process of crystal growth .
the powder x - ray diffraction pattern of polycrystalline laptsi@xmath0 was similar to that of euptsi@xmath0 showing thereby that the two are iso - structural .
the lattice parameters obtained from the rietveld analysis are @xmath23 = 4.3467 and @xmath24 = 9.6335 with the unit cell volume 182.01 @xmath20 .
epma of the single crystals averaged over 10 different spots revealed the following composition : eu@xmath25pt@xmath26si@xmath27 : sn@xmath28 which is very close to 1:1:3 stoichiometry .
the trace amount of sn present in the single crystals is unavoidable .
the temperature dependence of the electrical resistivity of euptsi@xmath0 for current parallel to [ 100 ] direction is shown in fig .
the magnitude of electrical resistivity is typical of intermetallic compounds and its decrease with decreasing temperature indicates a metallic state . at 17 k the resistivity shows a change of slope ( shown on an enlarged scale in the inset of fig .
[ fig1 ] ) due to reduction in spin disorder scattering arising from magnetic ordering of the eu moments .
the resistivity shows only one transition ; however a second magnetic transition is clearly seen in the magnetic susceptibility and the heat capacity data as described in the following . near @xmath12 = 3.7
k which is close to the superconducting transition temperature of sn , the resistivity shows a drop which does not however attain a zero value .
this indicates the presence of some minor filamentary sn incorporated in the crystal during growth process .
( color online ) temperature dependence of electrical resistivity for @xmath29 [ 100 ] in the temperature range from 1.8 to 300 k. the inset shows the low temperature part and the arrow indicates the magnetic ordering.,scaledwidth=45.0% ] the temperature dependence of magnetic susceptibility from 1.8 to 300 k is shown in fig .
[ fig2 ] with the applied magnetic field @xmath30 parallel to [ 100 ] and [ 001 ] directions , respectively .
the low temperature magnetic susceptibility data clearly show two magnetic transitions at @xmath31 = 17 k and 16 k. the inverse magnetic susceptibility of euptsi@xmath0 is shown in fig .
[ fig2](b ) and the high temperature data were fitted to the curie - weiss law , @xmath32 ) . for @xmath4 [ 100 ] , the effective moment @xmath33 is 7.62 @xmath6/eu and the paramagnetic curie temperature @xmath34 = 4.7 k ; for @xmath4 [ 001 ] , @xmath33 = 7.56 @xmath6/eu and @xmath34 = 11.7 k. the effective moments are lower than the free ion value for eu@xmath14 ( @xmath33 = 7.94 @xmath6 ) , yielding a magnetic signal which is 8% lower than expected .
since we find the same signal deficit in the magnetization data , to be described below , we think it may be ascribed either to the presence of traces of an extra phase ( sn flux ) , resulting in a small over estimation of the actual amount of euptsi@xmath0 present in the sample , or to a slight off - stoichiometry of eu .
evidence for the presence of sn in the crystals , evoked above , favours the first possibility . since
the magnetization data described below show that the magnetic structure is antiferromagnetic , the positive values of the paramagnetic curie - weiss temperature along both directions are _ a priori _ surprising ; they are tentatively attributed to exchange anisotropy , as discussed below .
( color online)(a ) temperature dependence of magnetic susceptibility along the two principal directions in the temperature range from 1.8 to 300 k measured in an applied field of 0.1 t. the inset in ( a ) shows the low temperature part of the magnetic susceptibility clearly indicating the two antiferromagnetic ordering in euptsi@xmath0 .
( b ) inverse magnetic susceptibility of euptsi@xmath0 , the solid lines indicate the fit to curie - weiss law.,scaledwidth=45.0% ] to further elucidate the nature of magnetic ordering , we have performed isothermal magnetization versus field scans at t = 2 k in a vibrating sample magnetometer up to a field of 12 t as shown in fig .
[ fig3 ] . at low field ,
the magnetizations along the two principal directions increase linearly with the field , with similar slopes . for @xmath4 [ 001 ]
, there is an abrupt increase of the magnetization at 1 t followed by a faster linear increase up to 5.9 t ; above 5.9 t , the magnetization is saturated with a value 6.43 @xmath6/eu . for @xmath4 [ 100 ] , the magnetization shows a linear increase with field and saturates to the same value at a higher field of 9.2 t. the apparent saturation moment is lower than the eu@xmath14 free ion value ( 7 @xmath6/eu ) by 8% , which , as discussed above , leads us to think that there is a small amount of sn flux in the sample .
the overall behaviour of the magnetization versus field indicates an antiferromagnetic ( af ) structure . in a first analysis ,
the metamagnetic transition at @xmath5 = 1 t for @xmath7 [ 001 ] and the smaller value for the saturation field along [ 001 ] are characteristic of a two - sublattice af structure , with moments in zero field along [ 001 ] and with a small crystalline anisotropy along this axis @xcite . then , the field @xmath5 is the spin - flop field for which the moments along [ 001 ] reorient perpendicular to the field , and the critical fields of 5.9 and 9.2 t are the spin - flip fields ( @xmath35 ) for full alignment with the applied field in each direction .
the spin - flip field @xmath35 along [ 100 ] is larger because [ 100 ] is a hard magnetic axis .
however , the fact that the magnetization along [ 001 ] is not zero below @xmath5 and the presence of a slight metamagnetic transition along [ 100 ] at 1.5 t show that the magnetic structure is more complex than described above , and in particular that the eu@xmath14 moments ( in zero field ) are not exactly aligned along [ 001 ] .
anisotropy in the spin flip fields has been observed in some gd compounds @xcite , in euas@xmath0 @xcite and in eupdsb @xcite .
it arises in principle from the crystal field @xcite , although typically eu@xmath14 and gd@xmath17 ions , which have @xmath16 = 0 , show very small magneto - crystalline anisotropy .
another source of anisotropy could be two - ion exchange , although the simple model of an anisotropic molecular field tensor developed in ref .
@xcite for a two - sublattice af structure predicts that the spin - flip fields should be the same for both directions .
so we tentatively attribute the observed behaviour mainly to an anomalously large crystalline anisotropy of eu@xmath14 in euptsi@xmath0 .
( color online ) isothermal magnetization measurement of euptsi@xmath0 along the two principal crystallographic directions .
the metamagnetic transitions are indicated by the arrows.,scaledwidth=45.0% ] more quantitatively , the two - sublattice model of ref.@xcite involves the exchange field @xmath36 and the anistropy field @xmath37 , where @xmath38 is the crystalline anisotropy energy and @xmath39 the saturated eu@xmath14 moment . along the easy axis [ 001 ] , the spin - flop field is @xmath5 = @xmath40 and the spin - flip field @xmath41 , and along [ 100 ] the spin flip field is @xmath42 . applying the model to euptsi@xmath0
, the values of the critical fields along [ 001 ] can be reproduced with @xmath36 = 3.0 t and @xmath43 = 0.085 t , which are reasonable values , but then the critical field along [ 100 ] is 6.25 t , too small with respect to experiment ( 9.2 t ) .
so this simple model also fails to reproduce in detail the magnetization behaviour , which is probably due to a magnetic structure more complex than two af sublattices with moments aligned along [ 001 ] .
nevertheless , a hint to the presence of exchange anisotropy is given by the different values of the paramagnetic curie temperature along the two axes . indeed , in terms of the exchange integrals with nearest neighbours @xmath44 and with second neighbours @xmath45 along direction @xmath46 , one has ,
on the one hand @xcite : @xmath47 the anisotropy of @xmath34 then can stem from an anisotropy of the exchange constants . on the other hand
, one has : @xmath48 where @xmath49 labels the direction of the propagation vector of the magnetic structure . in case @xmath50
is negative and @xmath51 is positive with @xmath52 , then the magnetic structure is of af type . if these conditions are also fulfilled for the exchange integrals along directions [ 001 ] and [ 100 ] , then the paramagnetic curie temperatures are positive , and @xmath53 , as observed euptsi@xmath0 .
we have also recorded the magnetization data at various temperatures for @xmath7 [ 001 ] . as the temperature
is increased , the spin - flop transition shifts slightly to higher fields and then it decreases as the temperature is increased .
the spin - flip field monotonically decreases as temperature is increased . from the @xmath54 plots ( not shown here ) we have constructed the magnetic phase diagram as shown in fig .
[ fig4 ] .
( color online ) magnetic phase diagram of euptsi@xmath0 determined from the isothermal magnetization measured at various temperature . , scaledwidth=45.0% ] figure [ fig5 ] shows the temperature dependence of heat capacity of euptsi@xmath0 measured between 2 and 40 k , together with the heat capacity of non - magnetic , reference polycrystalline laptsi@xmath0 .
the heat capacity clearly shows two peaks at 17 k and 16 k confirming the bulk magnetic ordering in this compound .
incidentally , isostructural ceptsi@xmath0 also exhibits two magnetic transitions at 4.8 k and 2.4 k respectively @xcite .
the magnetic part of heat capacity @xmath55 was deduced by the usual method of subtracting the heat capacity of laptsi@xmath0 from that of euptsi@xmath0 after taking into account the renormalization due to different atomic masses of la and eu .
the @xmath56 versus temperature plot and the calculated entropy is shown in fig .
[ fig5](b ) .
the entropy is about @xmath57 ln8 at @xmath1 but above @xmath1 it exceeds the theoretical value of @xmath58 ln8 for @xmath59 .
while short range order above @xmath1 , which is typically present in the paramagnetic state , would explain the observed values of entropy up to @xmath1 , the increase of entropy beyond @xmath58 ln8 at higher temperatures suggests that laptsi@xmath0 may not be a good reference for the lattice heat capacity of euptsi@xmath0 .
we believe such an indication is reflected by the heat capacity plots of the two compounds which show an increasing divergence from each other at high temperatures .
the different valence states of eu@xmath14 and la@xmath17 may result in different inter - atomic potentials modifying the phonon spectra of these two compounds .
the heat capacity jump @xmath60 at @xmath31 is estimated to be 14.6 j / k mol . based on the mean field approximation , blanco _ et al _
@xcite have calculated the heat capacity jumps at the magnetic transition for two types of magnetic structures , namely , the equal moment ( em ) where the magnetic moments are the same at all sites and the amplitude modulated ( am ) , where the magnetic moment amplitude varies periodically from one site to another . for the case of em structure , the jump in the heat capacity at the ordering temperature
is given by , ( color online)(a ) temperature dependence of the heat capacity of euptsi@xmath0 .
the two magnetic orderings are indicated by the arrow .
( b ) @xmath61 versus @xmath12 of euptsi@xmath0 , the calculated entropy is shown as solid line.,scaledwidth=45.0% ] @xmath62 and for the amplitude modulated system , @xmath63 where @xmath64 is the total angular momentum .
since the effects of crystal electric field for @xmath65 , divalent state of eu are negligible , the total angular momentum @xmath64 ( = 7/2 ) can be used in the above equations .
@xmath66 amounts to 20.14 j / k@xmath67mol for the em case , while for the am case , @xmath68 equals 13.4 j / k@xmath67mol .
thus , theoretically , the jump in the heat capacity at the magnetic transition for am case is reduced to 2/3 of the value for em case @xcite .
our estimate of @xmath60 ( 14.6 j / k@xmath67mol ) at @xmath1 thus strongly suggests that euptsi@xmath0 possesses an amplitude modulated structure .
furthermore , blanco _ et al _ found that for the amplitude modulated system a hump in the heat capacity occurs below the ordering temperature , to compensate the loss of entropy just below @xmath31 , which incidentally is seen in euptsi@xmath0 as shown in fig .
[ fig5](b ) in the @xmath61 versus temperature plot .
( color online ) @xmath3eu mossbauer absorption spectra at selected temperatures in euptsi@xmath0 .
the lines are fits to a single magnetic hyperfine interaction at 4.2 and 16 k , to an inconmmensurate magnetic modulation at 17 k and to a sinlge eu@xmath14 line at 19 k.,scaledwidth=35.0% ] @xmath3eu mssbauer spectroscopy is an excellent tool to probe the valence and the magnetic state of eu ions in eu compounds . while information about the valence state of eu ions is obtained via the isomer shift , the magnitude of the hyperfine magnetic field can be inferred from the splitting of the spectrum due to hyperfine interaction . @xmath3eu
mssbauer absorption spectra were recorded in euptsi@xmath0 in the temperature range 4.2 to 19 k and at 300 k. spectra at selected temperatures are shown in fig .
all the spectra point to the presence of eu@xmath14 with no traces of eu@xmath17 impurity content .
the isomer shift of all the spectra was found to be -10.0(4 ) mm / s arising from the divalent state of eu ions .
a single line is observed from room temperature down to 19 k. at 300 k the line width is 2.65 mm / s which is a standard value for a eu@xmath14 compound and it increases to 3.2 mm / s at 19 k , due to the dynamic short range order close to the first magnetic transition at @xmath31 = 17 k. at 4.2 k , a magnetic hyperfine field of 33.0(5 ) t can be derived from the spectrum , which is a standard value for the fermi contact interaction in eu@xmath14 materials .
on increasing the temperature , the hyperfine field decreases steadily , reaching 16.8(5 ) t at 16 k and the lines are narrow , indicative of a commensurate magnetic order with a single moment value ( em structure ) . at 17 k ,
the spectral shape changes suddenly , as can be seen in the fig .
[ fig6 ] : a broad distribution of hyperfine fields is present at the nucleus sites .
the spectrum can be fitted to a gaussian - shaped distribution , but a physically more appealing interpretation is the following : below 17 k , where the specific heat data shows its first anomaly , an incommensurate am magnetic structure develops , which yields a distribution of eu@xmath14 moments , and hence of hyperfine fields .
the second peak at 16 k in the specific heat would correspond to a lock - in transition to a commensurate wave - vector , which is not uncommon in intermetallic eu or gd compounds @xcite .
the fit at 17 k shown in the figure corresponds to an incommensurate modulation , but it can not be distinguished from that to a gaussian shaped distribution .
however , it is difficult to understand why a broad gaussian distribution of hyperfine magnetic fields would appear at 17 k , if it were not because of a phase transition yielding a specific moment distribution .
indeed , it has been reported that such a `` cascade '' of transitions ( paramagnetic - incommensurate am - commensurate em ) occurs for instance in layered semi - metallic euas@xmath0 @xcite below @xmath1 = 11.3 k and @xmath2 = 10.26 k , as well as in eupdsb @xcite , below @xmath1 = 18 k and @xmath2 = 12 k. we therefore believe the mssbauer and the heat capacity data corroborate each other quite well in demonstrating the presence of a `` cascade '' of transitions in euptsi@xmath69 .
single crystals of euptsi@xmath0 were grown by flux method and the magnetic studies revealed antiferromagnetic transitions at @xmath1 = 17 and @xmath2 = 16 k. from the low temperature isothermal magnetization measurements , we found a spin - flop - like transition at @xmath5 = 1 t for @xmath4 [ 001 ] and spin - flip transitions at 5.9 t ( 9.2 t ) for @xmath4 [ 001 ] ( [ 100 ] ) .
this shows the presence of anisotropy , probably of both crystalline and exchange origin . a magnetic phase diagram has been constructed based on the isothermal magnetization data .
the heat capacity measurement also confirm the occurrence of two magnetic transitions .
further , the specific heat and mssbauer data suggest that euptsi@xmath0 undergoes a cascade of close transitions ( paramagnetic - incommensurate - commensurate ) as temperature decreases .
e. bauer , g. hilscher , h. michor , ch .
paul , e. w. scheidt , a. gribanov , yu .
seropegin , h. noel , m. sigrist and p. rogl , phys .
lett * 92 * , 027003 ( 2004 ) .
n. kimura , k. ito , k. saitoh , y. umeda , h. aoki and t. terashima , phys .
lett . * 95 * , 247004 ( 2005 ) .
t. kawai , h. muranaka , m. a. measson , t. shimoda , y. doi , t. d. matsuda , y. haga , g. knebel , g. lapertot , d. aoki , j. flouquet , t. takeuchi , r. settai and y.nuki , j. phys .
. jpn . * 77 * , 064716 ( 2008 ) . | single crystals of euptsi@xmath0 , which crystallize in the banisn@xmath0-type crystal structure , have been grown by high temperature solution growth method using molten sn as the solvent .
euptsi@xmath0 which lacks the inversion symmetry and has only one eu site in the unit cell is found to be an antiferromagnet with two successive magnetic transitions at @xmath1 = 17 k and @xmath2 = 16 k , as inferred from magnetic susceptibility , heat capacity and @xmath3eu mssbauer measurements .
the isothermal magnetization data for @xmath4 [ 001 ] reveal a metamagnetic transition at a critical field @xmath5 = 1 t. the magnetization saturates to a moment value of 6.43 @xmath6/eu above 5.9 t ( 9.2 t ) for @xmath7 [ 001 ] ( [ 100 ] ) , indicating that these fields are spin - flip fields for the divalent eu moments along the two axes .
the origin of this anisotropic behaviour is discussed .
a magnetic ( h , t ) phase diagram has been constructed from the temperature dependence of isothermal magnetization data .
the reduced jump in the heat capacity at @xmath1 indicates a transition to an incommensurate , amplitude modulated antiferromagnetic structure .
the shape of the hyperfine field split mssbauer spectrum at @xmath1 provides additional support for the proposed nature of this magnetic transition . |
in this paper we continue the study of cartan subalgebras of infinite - dimensional root - reductive lie algebras initiated in @xcite . we refine and extend the description of cartan subalgebras of @xmath3 given in @xcite to the case of a general root - reductive lie algebra .
we also solve the problem of describing the set of conjugacy classes of cartan subalgebras of the three simple root - reductive lie algebras @xmath0 , @xmath1 , @xmath2 , as well as of @xmath3 , under the group of automorphisms of the natural representation which preserve the lie algebra .
root - reductive lie algebras are a specific class of locally finite lie algebras , and as background literature on locally finite lie algebras we recommend @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
here is a description of the contents of the paper .
its first main part , sections [ sectiona ] and [ sectionb ] , is concerned with the definition and description of cartan subalgebras . according to @xcite ,
a cartan subalgebra of @xmath3 can be defined as a locally nilpotent subalgebra which is the centralizer of the set of all semisimple parts of its own elements .
we accept this definition for any locally reductive lie algebra @xmath4 and prove that similarly to the case of @xmath3 , cartan subalgebras are nothing but centralizers of arbitrary maximal toral subalgebras . in particular , a cartan subalgebra is always self - normalizing .
we then prove that a cartan subalgebra can be characterized equivalently as a subalgebra @xmath5 which coincides with the set of @xmath6-locally nilpotent vectors in @xmath4 , where @xmath6 runs over all finite - dimensional subalgebras of @xmath7 .
in addition , we show that if @xmath4 is a root - reductive lie algebra , any cartan subalgebra @xmath7 is nilpotent and coincides with the set of all @xmath7-locally nilpotent vectors in @xmath4 .
the main new phenomenon in the case of a general root - reductive lie algebra versus the case of @xmath3 is that cartan subalgebras are no longer necessarily commutative and that the adjoint action of a cartan subalgebra @xmath7 on itself no longer has to be locally finite .
we treat in detail the simple lie algebras @xmath0 , @xmath1 , and @xmath2 .
the case of @xmath0 is very similar to the case of @xmath3 considered in @xcite .
the main new result in the other two cases is that maximal toral subalgebras are in one - to - one correspondence with maximal orthogonal or symplectic self - dual systems of @xmath8-dimensional subspaces in the natural representation .
similarly to the case of @xmath3 , such maximal self - dual systems do not have to span the natural representation , a phenomenon that gives rise to nonsplitting maximal toral subalgebras . in the case of @xmath1 the centralizer of a maximal toral subalgebra
is in general nilpotent of depth no greater than @xmath9 but not necessarily commutative .
remarkably enough , the ideal of nilpotent elements of a cartan subalgebra of @xmath1 is itself a nilpotent orthogonal lie algebra of a vector space with a degenerate symmetric form , and all such nilpotent degenerate algebras occur inside cartan subalgebras of @xmath1 . in the case of @xmath0 ,
the analogous nilpotent degenerate subalgebras do not occur inside cartan subalgebras of @xmath0 , and for @xmath3 and @xmath2 , there are no analogous nilpotent nonabelian subalgebras , which is consistent with the fact that @xmath3 , @xmath0 , and @xmath2 admit only abelian cartan subalgebras . in section [ sectionc ]
we address the conjugacy problem for cartan subalgebras posed in @xcite .
more precisely , in the case of @xmath3 it was established that certain standard discrete invariants of cartan subalgebras are not sufficient to characterize the conjugacy classes of cartan subalgebras of @xmath3 with respect to the group of automorphisms of the natural representation which preserve @xmath3 . in this paper
we describe the missing continuous invariants in terms of a linear algebraic structure which we call a complement datum . our main theorem in the second part , theorem [ equivthm ] , solves the conjugacy problem in terms of the combinatorics of complement data .
it turns out that for a generic set of fixed standard invariants , there are uncountably many conjugacy classes of cartan subalgebras . in all remaining cases , there are only finitely many conjugacy classes , and in fact at most @xmath9 .
we give criteria to distinguish the uncountable and finite cases and provide representatives in each finite case .
all vector spaces and lie algebras are defined over the field of complex numbers @xmath10 .
a lie algebra @xmath4 is _ locally finite _
( respectively _ locally nilpotent _ ) if every finite subset of @xmath4 is contained in a finite - dimensional ( resp .
nilpotent ) subalgebra .
if @xmath4 is at most countable dimensional , then being locally finite is equivalent to being isomorphic to a union @xmath11 of nested finite - dimensional lie subalgebras @xmath12 .
a @xmath4-module @xmath13 is _ locally finite _ if each element @xmath14 is contained in a finite - dimensional submodule , and @xmath13 is _ locally nilpotent _ if for any @xmath14 there exists an @xmath15 with @xmath16 .
furthermore , we say that a @xmath4-module @xmath13 is a _ generalized weight module _ if @xmath17 , where @xmath18 we define a generalized weight module @xmath13 to be a _ weight module _ if in addition , @xmath19 for each @xmath20 .
the _ support _ in @xmath21 of a module @xmath13 is the set @xmath22 for an arbitrary lie subalgebra @xmath5 , we set @xmath23 where @xmath24 is any exhaustion of @xmath7 by finite - dimensional subalgebras @xmath25 . in other words , @xmath26 is the subalgebra of all @xmath6-locally nilpotent vectors in @xmath4 , where @xmath6 runs over all finite - dimensional subalgebras of @xmath4 .
we call a lie algebra _ locally reductive _ if it is the union @xmath11 of nested finite - dimensional reductive lie algebras @xmath27 such that @xmath27 is reductive in @xmath28 .
recall that for finite - dimensional lie algebras we have a notion of a jordan decomposition , and in particular of semisimple elements and nilpotent elements . since these notions are preserved under injections @xmath29 where @xmath27 is reductive in @xmath28 , for any locally reductive lie algebra we can talk about a _
jordan decomposition _ of an element .
in addition , note that any nilpotent element @xmath30 of a locally reductive lie algebra @xmath4 lies in @xmath31 $ ] .
indeed , @xmath32 for some @xmath33 , and thus @xmath34 \subseteq [ \g,\g]$ ] .
we call a subalgebra @xmath35 of a locally reductive lie algebra @xmath4 _ splittable _ if for every @xmath36 , both the semisimple and nilpotent jordan components @xmath37 and @xmath38 of @xmath39 belong to @xmath35
. a subalgebra @xmath40 is _ toral _ if every element is semisimple .
every toral subalgebra is abelian : the standard proof of this fact extends from the case of a reductive lie algebra to a locally reductive lie algebra , cf .
@xcite . by definition ,
a toral subalgebra @xmath40 is _ splitting _ if @xmath4 is a weight module with respect to the adjoint action of @xmath41 on @xmath4 . for any subset @xmath42 and any subalgebra @xmath43
, we define the centralizer of @xmath44 in @xmath35 , denoted @xmath45 , to be the set of elements of @xmath35 which commute in @xmath4 with all elements of @xmath44 .
the center of @xmath4 is denoted @xmath46 .
let @xmath47 and @xmath48 be vector spaces of countable dimension with a nondegenerate pairing @xmath49 .
we define the lie algebra @xmath50 to be the lie algebra associated to the associative algebra @xmath51 ( @xmath51 is an associative algebra whose multiplication satisfies @xmath52 for @xmath53 , @xmath54 ) .
the derived subalgebra of @xmath55 has codimension @xmath8 and is denoted @xmath56 . if @xmath47 is endowed with a nondegenerate symmetric ( resp .
antisymmetric ) bilinear form , then one can choose @xmath57 and denote by @xmath58 ( resp .
@xmath59 ) the lie subalgebra @xmath60 ( resp .
@xmath61 ) of the lie algebra associated to the associative algebra @xmath62 .
a result of mackey @xcite implies that all associative algebras @xmath51 are isomorphic , as long as @xmath63 and the pairing @xmath64 is nondegenerate .
hence , up to isomorphism , the lie algebras @xmath55 , @xmath56 , @xmath65 , and @xmath59 do not depend on @xmath47 , and the usual representatives of these isomorphism classes are denoted @xmath3 , @xmath0 , @xmath1 , and @xmath2 , respectively .
clearly the lie algebras @xmath3 , @xmath0 , @xmath1 , and @xmath2 are locally reductive , as one can also choose exhaustions @xmath66 and @xmath67 ( where @xmath68 for @xmath69 ) such that the pairings @xmath70 are nondegenerate .
let @xmath4 be a locally reductive lie algebra .
if @xmath7 is a subalgebra of @xmath4 , let @xmath71 denote the set of semisimple jordan components of the elements of @xmath7 .
following @xcite , we say that a subalgebra @xmath7 of @xmath4 is a _
cartan subalgebra _ if @xmath7 is locally nilpotent and @xmath72 .
we start with the following proposition which generalizes a result of @xcite .
[ centralizer ] let @xmath7 be a locally nilpotent subalgebra of a locally reductive lie algebra @xmath4 .
then the following assertions hold : 1 . @xmath73 ; 2 .
[ toral ] @xmath71 is a toral subalgebra of @xmath4 ; 3 .
[ self - normalizing ] @xmath74 is a self - normalizing subalgebra of @xmath4 .
let @xmath75 .
the local nilpotence of @xmath7 implies that @xmath76 for some @xmath77 .
since @xmath78 is a polynomial with no constant term in @xmath79 , it follows that @xmath80 . because an element commutes with its semisimple part , @xmath81 , and it follows by induction that @xmath82 .
hence @xmath83 .
thus @xmath73 . furthermore , by the same argument
, @xmath84 implies that @xmath85 . therefore any two elements of @xmath71 commute .
since the sum of any two commuting semisimple elements is semisimple , @xmath71 is a subalgebra .
finally , suppose @xmath30 is in the normalizer of @xmath86 .
for any @xmath87 , we have that @xmath88 \in \z_\g(\h_{ss})$ ] .
thus @xmath89,y]=0 $ ] , and as @xmath90 is semisimple it follows that @xmath88=0 $ ] .
hence @xmath91 , i.e. @xmath92 is self - normalizing . the following theorem is our main general result characterizing cartan subalgebras of locally reductive lie algebras .
it generalizes a well - known result for finite - dimensional reductive algebras ( cf .
vii , 4 ) ) , as well as a result from @xcite .
[ main ] let @xmath4 be a locally reductive lie algebra and @xmath7 a subalgebra of @xmath4 .
the following conditions on @xmath7 are equivalent : 1 .
[ c ] @xmath7 is a cartan subalgebra ; 2 .
[ d ] @xmath72 and @xmath71 is a subalgebra ; 3 .
[ e ] @xmath93 for some maximal toral subalgebra @xmath40 ; 4 .
[ f ] @xmath94 .
in addition , any cartan subalgebra @xmath7 is splittable and self - normalizing .
we first prove a short lemma .
[ lemma1 ] if @xmath7 is locally nilpotent and splittable , then @xmath95 . by (
vii , 5 , prop .
5 ) @xmath96 , with @xmath97 being the subalgebra of all nilpotent elements in @xmath7 .
it follows that @xmath98 . since elements of @xmath71 are semisimple , @xmath99 .
clearly @xmath100 . hence @xmath95 .
fix an exhaustion @xmath101 , where @xmath27 is a finite - dimensional reductive lie algebra with @xmath27 reductive in @xmath28 .
clearly ( [ c ] ) implies ( [ d ] ) , by proposition [ centralizer ] ( [ toral ] ) . to show that ( [ d ] ) implies ( [ c ] )
, we first prove that ( [ d ] ) implies that @xmath7 is splittable .
suppose that @xmath7 satisfies ( [ d ] ) .
for any @xmath15 note that @xmath102 since @xmath103 , we have @xmath104 for some sufficiently large @xmath105 . since @xmath106 is a subalgebra of @xmath107 , we know from ( * ? ? ? * ch .
vii , 5 , prop . 3 cor .
1 ) that @xmath108 is a splittable subalgebra of @xmath107 .
the reductive lie algebra @xmath27 is also splittable , by ( * ? ? ?
vii , 5 , prop .
the intersection @xmath109 is splittable , too .
being a union of splittable algebras , @xmath7 is splittable . to show the local nilpotence of @xmath7 , notice that the equality @xmath72 implies that every element of @xmath71 is in the center of @xmath7
now consider a general element @xmath110 .
choose @xmath39 such that @xmath111 .
for any @xmath112 , @xmath113 hence @xmath7 is locally nilpotent , i.e. ( [ d ] ) implies ( [ c ] ) .
next , it is clear that ( [ d ] ) implies ( [ e ] ) .
the equality @xmath114 shows that any semisimple element of @xmath4 which centralizes @xmath71 is already in @xmath71 . thus @xmath71 is a maximal toral subalgebra and ( [ e ] ) holds
let @xmath7 satisfy ( [ e ] ) .
the same argument as above shows that @xmath7 is splittable , whence @xmath115 .
then clearly @xmath116 . if @xmath117 , the existence of a semisimple element @xmath118 contradicts the maximality of @xmath41 .
therefore @xmath119 , and ( [ e ] ) implies ( [ d ] ) .
note that ( [ c ] ) implies ( [ f ] ) .
indeed , let @xmath7 be a cartan subalgebra .
we know @xmath7 is splittable and locally nilpotent , so by lemma [ lemma1 ] , @xmath120 . to show that ( [ f ] ) implies ( [ c ] ) , assume that @xmath94 .
then clearly @xmath7 is locally nilpotent .
an argument similar to that above shows that @xmath7 is also splittable .
indeed , for any @xmath15 we have @xmath121 the finite dimensionality of @xmath27 yields @xmath122 for some sufficiently large @xmath105 .
it is well known that @xmath123 is a splittable subalgebra of @xmath107 , see ( * ? ? ?
vii , 1 , prop .
11 ) . since @xmath27 is also splittable , the intersection @xmath124 is splittable , too .
hence @xmath125 is splittable .
being a union of splittable algebras , @xmath7 is splittable .
therefore lemma [ lemma1 ] implies that @xmath126 .
in addition , by proposition [ centralizer ] ( [ self - normalizing ] ) , a subalgebra @xmath7 satisfying ( [ c ] ) is self - normalizing . as we have already seen that a cartan subalgebra is splittable , the proof of theorem [ main ] is complete .
the most common characterization of a cartan subalgebra of a finite - dimensional lie algebra is as a nilpotent self - normalizing subalgebra .
according to theorem [ main ] , any cartan subalgebra of a locally reductive lie algebra is self - normalizing .
any cartan subalgebra of a root - reductive lie algebra is in addition nilpotent .
in fact , we show in theorem [ depth2 ] that it is nilpotent of depth at most @xmath9 .
in contrast with the case of @xmath3 or @xmath0 , where all cartan subalgebras are abelian and hence act locally finitely on themselves for a trivial reason ( cf .
@xcite ) , the cartan subalgebras of @xmath1 do not necessarily act locally finitely on themselves , as discussed at the end of subsection [ sosubsection ] below .
we do not know whether all nilpotent self - normalizing subalgebras of a root - reductive lie algebra are cartan subalgebras .
we also do not know whether a cartan subalgebra of a general locally reductive lie algebra is necessarily nilpotent . in what follows
we call a cartan subalgebra @xmath7 _ splitting _ if @xmath71 is a splitting toral subalgebra of @xmath4 .
this section begins with some brief introductory material on root - reductive lie algebras .
we then consider separately the cases of a root - reductive lie algebra @xmath4 containing a simple component of type @xmath0 , @xmath1 , and @xmath2 .
in particular , we describe all cartan subalgebras of the simple lie algebras @xmath0 , @xmath1 , and @xmath2 .
the section concludes with a general theorem about the cartan subalgebras of an arbitrary root - reductive lie algebra and with examples .
an inclusion of finite - dimensional reductive lie algebras @xmath127 is a _ root inclusion _ if , for some cartan subalgebra @xmath128 of @xmath129 , the subalgebra @xmath130 is a cartan subalgebra of @xmath131 and each @xmath130-root space @xmath132 is also a root space of @xmath129 .
informally , root spaces of @xmath131 are root spaces of @xmath129 .
a lie algebra @xmath4 is called _ root - reductive _ if it is isomorphic to a union @xmath11 of nested reductive lie algebras with respect to root inclusions for a fixed choice of nested cartan subalgebras @xmath133 with @xmath134 .
finite - dimensional reductive lie algebras are root - reductive , as are @xmath3 , @xmath0 , @xmath1 , and @xmath2 .
the following theorem is a slightly more general version of a result in @xcite .
[ structure ] let @xmath4 be a locally reductive lie algebra . 1 . [ split ]
there is a split exact sequence of lie algebras @xmath135 \rightarrow \g \rightarrow \g/[\g , \g ] = : \a \rightarrow 0,\ ] ] i.e. @xmath136 \subsetplus \a$ ] , where @xmath44 is abelian .
2 . if @xmath4 is root - reductive , the lie algebra @xmath137 $ ] [ simple ] is isomorphic to a direct sum of finite - dimensional simple lie algebras , as well as of copies of @xmath0 , @xmath1 , and @xmath2 , with at most countable multiplicities . 1 .
fix an exhaustion @xmath138 , where @xmath27 is finite dimensional and reductive .
let @xmath133 be a family of cartan subalgebras such that @xmath139 .
define @xmath140 .
then @xmath141 + \h$ ] . since @xmath7 is abelian
, there is an abelian lie algebra @xmath142 such that @xmath143 \oplus \a$ ] .
one may check that @xmath144 \subsetplus \a$ ] .
2 . for the proof see ( * ? ? ?
* theorem 1 ) .
we will apply the following proposition only in the case of root - reductive lie algebras , but we prove it in more generality for locally reductive lie algebras .
[ extendrep ] let @xmath138 be a locally reductive lie algebra , with @xmath27 finite dimensional and reductive , and @xmath27 reductive in @xmath28 .
let @xmath145 be a @xmath31$]-module which is a direct limit of injections of simple @xmath146$]-modules @xmath147 .
then the @xmath31$]-module structure on @xmath145 can be extended to a @xmath4-module structure . for each @xmath33
, choose an abelian lie algebra @xmath148 such that @xmath149 ) \oplus \a_i$ ] .
let @xmath150 $ ] and @xmath151 be the corresponding projections .
consider the inclusion @xmath152 , and note that @xmath153 is injective .
let @xmath154 be a vector space complement of @xmath155 in @xmath156 .
given an action of @xmath157 on @xmath147 by scalar endomorphisms , we define an action of @xmath156 on @xmath158 by scalars as follows .
let @xmath159 be arbitrary , and let @xmath160 denote the scalar by which @xmath161 acts on @xmath147 .
observe that the image of @xmath162 in @xmath163 must commute with the image of @xmath164 $ ] .
hence the image of @xmath165 in @xmath163 is the direct sum of a scalar endomorphism of @xmath147 and an endomorphism of a vector space complement of @xmath147 .
let @xmath166 denote the scalar by which @xmath165 acts on @xmath147 . define the action of @xmath167 on @xmath158 to be by the scalar @xmath168 .
since @xmath169 is injective , this procedure defines an action of @xmath155 on @xmath158 .
define the action of @xmath154 on @xmath158 to be trivial .
we are now ready to define an action of @xmath4 on @xmath145 . if @xmath170 and @xmath171 , then for some @xmath33 we have @xmath32 and @xmath172 . set @xmath173 , where the second action on the right hand side comes from the preceding paragraph .
one must check that this is well - defined , i.e. compatible with the inclusions @xmath29 and @xmath174 .
explicitly , the composition @xmath175 coincides with the map @xmath176 . since elements of @xmath157 act as scalars on @xmath147 ,
@xmath145 is a well - defined @xmath4-module . in what follows @xmath4
will always denote a root - reductive lie algebra . in this subsection @xmath177
is a direct summand of @xmath31 $ ] , and @xmath41 is a maximal toral subalgebra of @xmath4 .
we consider @xmath47 as a @xmath31$]-module in which all direct summands other than @xmath58 act trivially . by proposition [ extendrep ] ,
@xmath47 may be endowed with the structure of a @xmath4-module .
we write @xmath178 ( and more generally @xmath179 ) for the maximal locally finite @xmath41-submodule of @xmath47 ( resp . of a @xmath41-module @xmath180 ) .
the key fact about maximal locally finite submodules is that if @xmath180 and @xmath181 are two @xmath41-modules then @xmath182 .
this is a special case of a result of dimitrov and zuckerman appearing in ( * ? ? ?
* appendix prop .
a ) . thus @xmath183 .
recall that for any lie subalgebra @xmath43 , @xmath184 is a generalized weight @xmath35-submodule of @xmath47 .
if @xmath185 , it is easy to verify that @xmath186 coincides with the maximal locally finite @xmath41-submodule @xmath187 .
therefore @xmath188 and @xmath189 is the @xmath190-weight space of this weight decomposition of @xmath191 .
thus @xmath192 where @xmath193 .
[ so ] let @xmath4 be a root - reductive lie algebra for which @xmath58 is a direct summand of @xmath137 $ ] .
the following statements hold for any maximal toral subalgebra @xmath40 : 1 .
[ one ] if @xmath194 , then @xmath195 . if furthermore @xmath196 , then @xmath197 ; 2 .
[ two ] if @xmath198 then @xmath199 ; 3 .
[ three ] @xmath200 has rank 0 or 1 , and @xmath201 ; 4 .
[ four ] @xmath202 is nilpotent of depth at most @xmath9 . if @xmath203 , a subalgebra @xmath40 is maximal toral if and only if ( [ three ] ) , ( 5 ) and ( 6 ) hold , where 1 . [ five ] @xmath204 and @xmath205 for any @xmath206 ; 2 .
[ six ] @xmath207 , where @xmath208 . 1 .
let @xmath209 and @xmath210 be such that @xmath211 . if @xmath212 , then the @xmath4-invariance of @xmath213 yields @xmath214 .
thus @xmath215 .
since @xmath211 , it follows that @xmath195 .
in particular , if @xmath196 , then @xmath216 .
+ suppose @xmath196 and without loss of generality that @xmath217 .
each of @xmath218 and @xmath219 satisfies the equation @xmath220 and therefore is semisimple .
furthermore , @xmath221 = \langle w , w \rangle v \otimes v - \langle v , v \rangle w \otimes w = 0 $ ] .
hence @xmath222 is semisimple . since @xmath223 , @xmath224
is contained in @xmath41 by maximality
. then @xmath225 implies @xmath226 . for @xmath227
, we calculate @xmath228 .
therefore @xmath229 is @xmath8-dimensional .
2 . in view of part ( [ one ] ) ,
@xmath229 and @xmath230 are in the radical of the form @xmath231 and hence all elements of @xmath191 centralize @xmath232 .
3 . suppose , for the sake of a contradiction , that the form @xmath200 has rank greater than @xmath8 .
then there exist @xmath233 with pairings given by the matrix @xmath234 $ ] . as in part ( [ one ] )
, we see that @xmath235 is semisimple and hence lies in @xmath41 . since @xmath236 is in the @xmath190-weight space of @xmath41 , we calculate @xmath237 , which contradicts the choice of @xmath238
. therefore the form @xmath200 has rank @xmath190 or @xmath8 .
+ to prove that @xmath201 , it is enough to show that elements of @xmath239 are nilpotent as endomorphisms of @xmath47 .
let @xmath240 be arbitrary .
if the form @xmath200 has rank @xmath190 , one may compute that @xmath241 . if the form @xmath200 has rank @xmath8
, one may compute that @xmath242 .
we have @xmath243 , and since @xmath244 is central in @xmath189 , it is enough to show that @xmath245 is nilpotent of depth at most @xmath9 .
+ if @xmath246 , then @xmath239 is abelian .
now suppose that @xmath200 has rank @xmath8 .
let @xmath247 be the radical of the form @xmath248 .
let @xmath238 be any vector in @xmath249 , and we may as well assume that @xmath250 .
we have @xmath251 . to show that @xmath245 is nilpotent of depth at most @xmath9 , it is enough to show that @xmath252 \subseteq \bigwedge^2 k$ ] , since @xmath253 is central . for arbitrary @xmath254
, we compute @xmath255 & = & [ a \otimes v - v \otimes a , b \otimes v - v \otimes b ] \\ & = & ( a \otimes v - v \otimes a ) \cdot ( b \otimes v - v \otimes b ) - ( b \otimes v - v \otimes b ) \cdot ( a \otimes v - v \otimes a ) \\ & = & - a \otimes b + b \otimes a \\ & = & b \wedge a \in { \bigwedge } ^2 k.\end{aligned}\ ] ] 5 .
for the first equality , let us compute @xmath256 in two different ways .
on the one hand , by definition , @xmath257 . on the other hand , the assumption @xmath258 yields @xmath259 , and since @xmath260 for @xmath261 , we obtain @xmath262 .
this implies @xmath263 , and hence @xmath264 . +
now consider the fact that @xmath265 for @xmath196 .
it follows that @xmath266 .
part ( [ one ] ) gives @xmath205 for any @xmath206 .
6 . by the proof of part ( [ one ] )
, we have @xmath267 for @xmath268 .
this gives the inclusions @xmath269 in part ( [ three ] ) we showed that @xmath270 .
hence @xmath207 .
conversely , assume that conditions ( [ three ] ) , ( 5 ) , and ( 6 ) are satisfied .
since @xmath271 for @xmath272 , the subalgebra @xmath41 is a direct sum of @xmath8-dimensional lie subalgebras , hence abelian .
if @xmath196 , we have @xmath273 , hence @xmath274 .
thus @xmath41 is spanned by elements of the form @xmath235 where @xmath209 , @xmath275 , @xmath196 , and @xmath276 .
these elements are semisimple as in part ( [ one ] ) .
hence @xmath41 is a toral subalgebra of @xmath258 .
the centralizer of @xmath41 in @xmath4 is contained in @xmath277 and coincides with @xmath278 condition ( [ three ] ) implies that each element in @xmath239 is nilpotent , so @xmath41 is maximal toral . for any maximal toral subalgebra @xmath279
, we have @xmath280 which is a lie algebra nilpotent of depth at most @xmath9 . the result from theorem
[ so ] ( [ three ] ) that @xmath200 has rank 0 or 1 enables us to make the following definition .
a maximal toral subalgebra of @xmath1 is _ even _ if the rank of @xmath200 is 0 , and _
odd _ if the rank is 1 .
likewise , a cartan subalgebra @xmath7 is _ even _ if @xmath71 is an even maximal toral subalgebra , and @xmath7 is _ odd _ if @xmath71 is odd .
there are two types of exhaustions of @xmath1 via root inclusions : as @xmath281 and as @xmath282 .
a union of cartan subalgebras of @xmath283 yields an even splitting cartan subalgebra of @xmath1 , and a union of cartan subalgebras of @xmath284 yields an odd splitting cartan subalgebra of @xmath1 .
it is easy to see that any splitting cartan subalgebra of @xmath1 arises in one of these two ways and is necessarily abelian .
notice that @xmath239 is the degenerate orthogonal lie algebra @xmath285 . for a degenerate symmetric form on a vector space @xmath145 ,
the lie algebra @xmath286 is nilpotent precisely when the form has rank @xmath190 or @xmath8 .
explicitly , the lie algebra @xmath286 is nilpotent and nonabelian exactly when the form has rank @xmath8 and @xmath287 .
hence any nonabelian cartan subalgebra of @xmath58 must be odd and must have @xmath288 .
here is an example of a nonabelian cartan subalgebra of @xmath58 .
let @xmath289 be a basis for @xmath47 with @xmath290 .
let @xmath291 observe that @xmath279 is a maximal toral subalgebra , and we have @xmath292 we see that the cartan subalgebra @xmath189 is not abelian as @xmath293 = e_{-2 } \wedge e_1 $ ] .
it is easy to check that the centralizer of any even maximal toral subalgebra acts locally finitely on itself , whereas the centralizer of an odd maximal toral subalgebra acts locally finitely on itself if and only if @xmath294 is finite dimensional .
let @xmath295 be a direct summand of @xmath31 $ ] , and @xmath41 a maximal toral subalgebra of @xmath4 .
the following results are proved in the same way as the analogous statements for @xmath58 .
we will write @xmath296 for the symmetrizer of vector spaces @xmath180 and @xmath181 .
set also @xmath297 for @xmath298 , @xmath299 .
we consider @xmath47 as a @xmath31$]-module in which all direct summands other than @xmath59 act trivally , and by proposition [ extendrep ] , @xmath47 may be endowed with the structure of a @xmath4-module .
we have the equality @xmath300 and the weight decomposition @xmath301 observe that @xmath302 where @xmath303 .
[ sp ] let @xmath4 be a root - reductive lie algebra for which @xmath59 is a direct summand of @xmath137 $ ] .
the following statements hold for any maximal toral subalgebra @xmath40 : 1 . if @xmath194 , then @xmath195 . if furthermore @xmath196 , then @xmath197 ; 2 . if @xmath198 then @xmath304 ; 3 .
@xmath305 , and @xmath306 ; 4 .
[ spfour ] @xmath307 is abelian .
if @xmath308 , a subalgebra @xmath309 is maximal toral if and only if ( [ three ] ) , ( 5 ) , and ( 6 ) hold , where 1 .
@xmath204 and @xmath205 for any @xmath310 ; 2 .
@xmath311 , where @xmath208 . for any maximal toral subalgebra @xmath312
, we have @xmath313 which is an abelian lie algebra . in this subsection
we generalize a theorem from @xcite , which we first recall .
@xcite [ thmc ] a subalgebra @xmath314 is maximal toral if and only if the following conditions are satisfied : 1 .
@xmath315 and @xmath316 for any @xmath206 ; 2 .
@xmath317 , and @xmath318 ; 3 .
@xmath319 . thus , for any maximal toral subalgebra @xmath320 , we have @xmath321 which is an abelian lie algebra .
assume that @xmath56 is a direct summand of @xmath31 $ ] , and let @xmath41 be a maximal toral subalgebra of @xmath4 .
we consider @xmath47 and @xmath48 as @xmath31$]-modules in which all direct summands other than @xmath56 act trivially , and by proposition [ extendrep ] , @xmath47 and @xmath48 may be endowed with the structure of @xmath4-modules .
we have the equality @xmath322 and the weight decomposition @xmath323 observe that @xmath324 [ sl ] let @xmath4 be a root - reductive lie algebra for which @xmath56 is a direct summand of @xmath137 $ ] .
the following statements hold for any maximal toral subalgebra @xmath40 : 1 .
[ slone ] if @xmath325 , then @xmath195 . if furthermore the rank of @xmath326 is at least @xmath9 , then @xmath327 ; 2 . [ sltwo ] if @xmath328 then @xmath329 ; 3 .
[ slthree ] if the rank of @xmath326 is at least @xmath9 , then @xmath330 .
hence @xmath331 has rank @xmath190 or @xmath8 .
also , @xmath318 ; 4 .
[ slfour ] @xmath332 is nilpotent of depth at most @xmath9 . if @xmath333 , a subalgebra @xmath40 is maximal toral if and only if ( 3 ) , ( 5 ) , and ( 6 ) hold , where 1 .
@xmath317 , and @xmath318 ; 2 . @xmath315 and @xmath316 for any @xmath206 ; 3 .
@xmath334 . 1 .
suppose @xmath335 and that the rank of @xmath326 is at least @xmath9 .
then let @xmath209 and @xmath336 be such that @xmath337 . since the rank of @xmath326 is at least @xmath9 , there exist @xmath338 and @xmath339 such that @xmath340 and @xmath341 .
we have @xmath342 , by an argument similar to the @xmath1 case . then @xmath343 implies @xmath226 .
+ first assume @xmath344 .
let @xmath345 , and compute @xmath346 .
this implies that @xmath229 is @xmath8-dimensional .
a similar argument shows that @xmath347 is also @xmath8-dimensional .
+ second assume , for the sake of a contradiction , that @xmath348 . then compute @xmath349 , which implies @xmath350 , contradicting @xmath226 .
the argument from the @xmath1 case works ; see theorem [ so ] .
3 . if the rank of @xmath326 is at least @xmath9 , and if @xmath351 , then as in part ( [ slone ] ) there exists @xmath212 with @xmath352 . hence @xmath353 .
the proof of the last statement is similar to that in theorem [ so ] .
we have @xmath354 .
if the rank of @xmath326 is not @xmath8 , then one may check that @xmath355 is abelian .
suppose the rank of @xmath326 is @xmath8 .
let @xmath166 be the weight for which @xmath356 has rank @xmath8 .
then @xmath357 is central in @xmath355 .
therefore it suffices to show that @xmath358 is nilpotent of depth at most @xmath9 .
it is an easy computation to check that if either @xmath359 or @xmath360 has dimension @xmath8 , then @xmath358 is abelian , and that otherwise it is nilpotent of depth @xmath9 . if the rank of @xmath326 is @xmath190 or @xmath8 , then @xmath361 .
thus for any maximal toral subalgebra @xmath362 , we have that @xmath331 has rank @xmath190 , which yields ( 3 ) .
one can modify the proofs from theorem [ so ] to obtain ( 5 ) and ( 6 ) , as well as the reverse implication . for any maximal toral subalgebra @xmath362
, we have @xmath363 which is an abelian lie algebra . note
that the degenerate lie algebra @xmath364 is nilpotent if and only if the form @xmath365 has rank @xmath190 or @xmath8 . in the rank @xmath8 case , the algebra is again nilpotent of depth at most @xmath9 .
this phenomenon does not occur for @xmath3 or @xmath2 , since in these cases any nilpotent degenerate algebra must be fully degenerate . 1 .
a _ dual system for @xmath366 _ or _ for @xmath56 _ is a set of of @xmath8-dimensional vector subspaces @xmath367 and @xmath368 for @xmath369 such that @xmath370 . here
@xmath371 is a finite set or @xmath372 .
a _ self - dual system for @xmath58 _ or _ for @xmath59 _ is a set of @xmath8-dimensional vector subspaces @xmath367 for @xmath369 such that @xmath373 . here
@xmath371 is a finite subset of @xmath374 with @xmath375 , or @xmath376 .
the following is a corollary to theorems [ so ] , [ sp ] , [ thmc ] , and [ sl ] .
[ bijection ] cartan subalgebras of @xmath3 ( resp . of @xmath0 , @xmath1 , or @xmath2 ) are in one - to - one correspondence with maximal ( self-)dual systems for @xmath3 ( resp . for @xmath0 , @xmath1 , or @xmath2 ) .
a ( self-)dual system yields a toral subalgebra by the formulas in figure [ fig1 ] . in the case of @xmath3 , since @xmath377 for @xmath378 , the subalgebra @xmath41 is a direct sum of @xmath8-dimensional lie subalgebras , hence abelian . since @xmath379 , @xmath41
is spanned by elements of the form @xmath218 , where @xmath380 , @xmath381 , and @xmath276 .
these elements are semisimple since they satisfy the equation @xmath220 .
hence @xmath41 is a toral subalgebra .
a similar argument shows that self - dual systems also yield toral subalgebras .
consider the map which sends a ( self-)dual system to the toral subalgebra according to the formulas in figure [ fig1 ] .
this map is injective ( except in the case of @xmath0 , where one must not allow dual systems to have @xmath382 ) . to see injectivity in the case of @xmath3 ,
suppose @xmath383 and @xmath384 are dual systems with @xmath385
. then @xmath386 .
hence for some @xmath387 we have @xmath388 .
this argument can be adapted to work for self - dual systems . moreover
, this map preserves containment , and its image includes all maximal toral subalgebras by theorems [ so ] , [ sp ] , [ thmc ] , and [ sl ] .
therefore maximal toral subalgebras correspond to maximal ( self-)dual systems .
explicitly , the maximal toral subalgebra @xmath41 and the cartan subalgebra @xmath7 in @xmath4 associated to a maximal ( self-)dual system for @xmath4 are given in figure [ fig1 ] .
note that a maximal toral subalgebra @xmath40 is splitting if and only if @xmath389 and @xmath390 . in that case
necessarily @xmath391 and @xmath392 for @xmath3 or @xmath0 , @xmath391 for @xmath2 , and for @xmath1 @xmath393 is @xmath190 or @xmath8-dimensional .
in fact , these conditions are equivalent to being splitting .
@xmath394 the following theorem is our main result on cartan subalgebras of general root - reductive lie algebras and strengthens theorem [ main ] in this case .
[ depth2 ] let @xmath4 be a root - reductive lie algebra and @xmath5 a cartan subalgebra .
then @xmath7 is nilpotent of depth at most @xmath9 , and @xmath395 .
let @xmath396 , so @xmath93 .
we first show that @xmath397 } ( \t)$ ] is nilpotent of depth at most @xmath9 . recalling the decomposition of @xmath31 $ ] in theorem [ structure ] ( [ simple ] ) , let @xmath398 be the direct sum of all finite - dimensional simple direct summands of @xmath31 $ ] , and @xmath399 the direct sum of all infinite - dimensional simple direct summands of @xmath31 $ ] . as there are no nontrivial extensions of an abelian lie algebra by a finite - dimensional reductive lie algebra
, we get the decomposition @xmath400 where @xmath44 is abelian .
then @xmath401 where @xmath402 is a maximal toral subalgebra of @xmath398 and @xmath403 is a maximal toral subalgebra of @xmath404 .
so @xmath397}(\t ) = \z_{\g_{fd } } ( \t_1 ) \oplus \z_{\g_{id } } ( \t_2)$ ] .
as @xmath402 is self - centralizing in @xmath398 , it remains to show that @xmath405 is nilpotent of depth at most @xmath9 .
let @xmath406 be the decomposition of @xmath399 into simple direct summands .
since @xmath407 \subseteq \s_i$ ] , we have @xmath408 theorems [ so ] ( [ four ] ) , [ sp ] ( [ spfour ] ) , and [ sl ] ( [ slfour ] ) imply that @xmath405 is nilpotent of depth at most @xmath9 . thus @xmath409 } ( \t)$ ]
is a lie algebra nilpotent of depth at most @xmath9 .
recall from theorem [ main ] that @xmath410 is splittable .
if @xmath411 , then @xmath412 . by definition @xmath413 , and @xmath414 $ ] as @xmath415 is nilpotent .
thus @xmath416 } ( \t ) $ ] .
since @xmath41 is central in @xmath410 , it follows that @xmath410 is nilpotent of depth no greater than @xmath9 .
finally , the nilpotence of @xmath7 implies @xmath417 .
therefore the condition @xmath94 together with the inclusions @xmath418 yields @xmath395 . in the case of @xmath419 , any maximal toral subalgebra of @xmath4
surjects onto @xmath420 $ ] @xcite . in general , the map @xmath421 $ ] does not have to be surjective .
for example , consider the lie algebra @xmath4 defined as the direct limit of the inclusions @xmath422 .\end{aligned}\ ] ] clearly we have the exact sequence @xmath423 \rightarrow \g \rightarrow { \mathbb{c}}\rightarrow 0.\ ] ] let @xmath424 denote the matrix @xmath425,\ ] ] and let @xmath181 denote the element of @xmath4 defined by the sequence @xmath426 . then as a vector space @xmath4 is isomorphic to @xmath427 .
we will exhibit a cartan subalgebra of @xmath4 whose image in @xmath10 is trivial . for @xmath428 ,
let @xmath429 be the matrix whose only nonzero entry is a @xmath8 in the upper right corner , and let @xmath430 be the matrix whose only nonzero entry is a @xmath8 in the lower left corner .
then @xmath431 is a toral subalgebra of @xmath4 .
we will show that @xmath432 .
a general element of @xmath4 lies in @xmath433 for some @xmath434 , and hence it may be expressed as @xmath435 for some @xmath436 and @xmath437 .
the computation @xmath438 = a(c_{n+1 } - c_{-(n+1)})$ ] implies that if @xmath435 centralizes @xmath41 , then @xmath439 .
let @xmath7 be a cartan subalgebra which is the centralizer of any maximal toral subalgebra containing @xmath41 .
the containments @xmath440 imply that @xmath7 maps trivially in @xmath10 .
theorem [ depth2 ] also leaves open the question whether the intersection @xmath441 $ ] is in general a maximal toral subalgebra of @xmath31 $ ] .
clearly in the case of @xmath3 the intersection of a maximal toral subalgebra with @xmath0 is maximal . as the following example shows ,
the intersection is not in general maximal , and it can even be trivial .
let @xmath442 be a basis of @xmath47 , and let @xmath443 .
define constants @xmath444 for @xmath445 by @xmath446 let @xmath447 and notice that @xmath448 is a well - defined vector in @xmath449 as @xmath450 for @xmath451 .
observe also that @xmath452 is a basis of @xmath449 .
we construct a root - reductive lie algebra @xmath4 with @xmath31 = \sl(v , v_*)$ ] and @xmath453 $ ] countable dimensional .
we consider the maximal toral subalgebra in @xmath4 consisting of all elements of @xmath4 which have all the @xmath448 s as eigenvectors .
no nontrivial element of @xmath56 has this property . by construction , @xmath454 for some unique @xmath455 .
let @xmath456 be defined by @xmath457 .
in addition , define inductively positive integers @xmath458 for @xmath459 by @xmath460 let @xmath461 , and let @xmath4 be the root - reductive lie algebra defined as the direct limit of the inclusions @xmath462 , 0 \right)\\ \left ( 0 ,
\frac{\sum_{k=1}^i d_k^l}{i}x_l \right ) & \mapsto & \left ( \left [ \begin{matrix } -\frac{\gamma_i^l}{i}i_i & \\ & \gamma_i^l \end{matrix } \right ] , \frac{\sum_{k=1}^{i+1 } d_k^l}{i+1}x_l \right),\end{aligned}\ ] ] where @xmath463 is a basis of @xmath464 , @xmath465 is the identity matrix in @xmath466 , and @xmath467 .
clearly @xmath31 = \sl(v , v_*)$ ] , where @xmath468 . for @xmath469 ,
define the element @xmath470 by @xmath471 - \frac{l+1}{2}i_l \right ) ( c_{ij})_{1 \leq i , j \leq l } ^{-1 } , \frac{l+1}{2}x_l \right).\ ] ] since @xmath472 is the sum of a central element and the conjugate of a diagonal element of @xmath473 , @xmath472 is semisimple .
because @xmath472 and @xmath474 have the same set of eigenvectors , they commute .
explicitly , if @xmath475 , the image of @xmath472 in @xmath476 is @xmath477 - \frac{\sum_{k=1}^m d_k^l}{m } i_m \right ) ( c_{ij})_{1 \leq i , j \leq m } ^{-1 } , \frac{\sum_{k=1}^m d_k^l}{m } x_l \right),\ ] ] i.e. @xmath478 are eigenvectors of the image of @xmath472 in @xmath476 .
let @xmath41 be the toral subalgebra of @xmath4 generated by the elements @xmath472 .
we will show that no nontrivial element of @xmath56 centralizes @xmath41 .
suppose @xmath479 centralizes @xmath41 .
since @xmath480 for some @xmath387 , we may consider @xmath481 as @xmath482,0 \right ) \in \g_l$ ] for any @xmath483 .
in the same block notation , write @xmath484 , * \right)$ ]
. then @xmath485 , 0 \right ) , \left ( \left [ \begin{matrix } * & s_l \\ 0 & * \end{matrix } \right ] , * \right ) \right ] = \left ( \left [ \begin{matrix } * & cs_l \\ 0 & 0 \end{matrix } \right],0
\right ) .\ ] ] the @xmath39-th column of @xmath486 is @xmath487 .
let @xmath488 be sufficiently large that the vectors @xmath489 span @xmath449 .
since the columns of @xmath486 span @xmath449 , it follows that @xmath490 .
hence no maximal toral subalgebra containing @xmath41 intersects @xmath56 nontrivially .
we define a toral subalgebra @xmath41 of @xmath4 , where @xmath4 is one of @xmath3 , @xmath0 , @xmath1 , or @xmath2 , to be _ submaximal _ if @xmath41 is associated to a not necessarily maximal ( self-)dual system via the formulas in figure [ fig1 ] .
any nonzero submaximal toral subalgebra is associated to a unique ( self-)dual system .
clearly not every toral subalgebra of @xmath4 is submaximal , but any maximal toral subalgebra of @xmath4 is submaximal , as seen in corollary [ bijection ] .
let @xmath491 be the subgroup of @xmath492 preserving @xmath493 . in terms of dual bases for @xmath47 and @xmath48 , elements of @xmath491 are matrices with finitely many nonzero entries in each row and each column .
observe that @xmath494 is a representation of @xmath495 , as is @xmath56 .
similary , each of @xmath58 and @xmath59 is a representation of the corresponding subgroup of @xmath496 preserving the bilinear form .
these subgroups are denoted @xmath497 and @xmath498 . in what follows
we describe submaximal toral subalgebras of @xmath366 ( resp . of @xmath58 or @xmath59 ) up to conjugation by the group @xmath491 ( resp .
by @xmath497 or @xmath498 ) .
let @xmath4 be one of @xmath3 , @xmath0 , @xmath1 , or @xmath2 , and @xmath499 its corresponding group defined above .
two finite - dimensional submaximal toral subalgebras of @xmath4 are conjugate by an element of @xmath499 if and only if they have the same dimension .
let @xmath402 and @xmath403 be finite - dimensional submaximal toral subalgebras of @xmath4 of the same dimension .
then @xmath500 for some @xmath33 , where @xmath138 is an exhaustion by finite - dimensional reductive lie algebras under root inclusions @xmath12 .
it is clear that @xmath402 and @xmath403 , being submaximal , are conjugate in @xmath27 by an element of the classical algebraic group @xmath501 associated to @xmath27 .
hence they are conjugate in @xmath4 , since there is an obvious injective homomorphism of @xmath501 into @xmath499 . in what follows , we assume that all ( self-)dual systems we consider are infinite .
their corresponding submaximal toral subalgebras will be infinite dimensional .
clearly all maximal ( self-)dual systems are infinite . 1 . a dual system @xmath502 for @xmath55 ( resp . for @xmath56 ) with @xmath503 and a dual system @xmath504 for @xmath505 ( resp . for @xmath364 ) with @xmath506 are _ equivalent _ if there exist isomorphisms @xmath507 and @xmath508 and a bijection @xmath509 such that @xmath510 , @xmath511 , and @xmath512 .
2 . a self - dual system @xmath513 for @xmath58 ( resp . for @xmath59 ) with @xmath514 and a self - dual system @xmath515 for @xmath286 ( resp . for @xmath516 ) with @xmath517 are _ equivalent _ if there exist an isomorphism @xmath507 and a bijection @xmath518 such that @xmath510 and @xmath511 .
[ lemma3 ] any dual system @xmath504 for @xmath505 is equivalent to a dual system for @xmath55 .
similarly , any self - dual system @xmath515 for @xmath286 ( resp . for @xmath516 ) is equivalent to a self - dual system for @xmath58 ( resp . for @xmath59 ) .
the first statement follows directly from mackey s result @xcite , which implies the existence of isomorphisms @xmath519 and @xmath520 which send @xmath521 to @xmath213 . under these isomorphisms ,
the dual system @xmath504 maps to an equivalent dual system .
similarly , any nondegenerate ( symmetric or antisymmetric ) form on a countable dimensional vector space can be diagonalized , hence there exists an isomorphism @xmath519 which sends @xmath521 to @xmath213 .
this isomorphism sends @xmath515 to an equivalent self - dual system .
given a dual system @xmath383 for @xmath55 , consider the following construction .
let @xmath522 be a vector space complement in @xmath47 of @xmath523 , and @xmath524 a vector space complement in @xmath48 of @xmath525 .
choose for each @xmath15 a generator @xmath526 , and let @xmath527 be the vector such that @xmath528 .
forgetting the dual system , we can consider the bilinear form @xmath529 restricted to @xmath530 , together with the linear functionals @xmath531 and @xmath532 .
this information encodes the entire structure of the dual system , together with @xmath47 , @xmath48 , and @xmath529 , up to isomorphism .
hence we suggest the following as a useful concept . 1 .
a _ complement datum for @xmath3 or @xmath0 _ is a pair of at most countable dimensional vector spaces @xmath522 and @xmath524 , a bilinear map @xmath533 , a sequence of linear functionals @xmath534 on @xmath522 , and a sequence of linear functionals @xmath535 on @xmath524 , for @xmath15 .
2 . a _ complement datum for @xmath1 _ ( resp . for @xmath2 )
is an at most countable dimensional vector space @xmath522 together with a symmetric ( resp .
antisymmetric ) bilinear form @xmath536 on @xmath522 and a sequence of linear functionals @xmath534 on @xmath522 , for @xmath537 . from a dual system for @xmath55 and choices of @xmath538 , @xmath539 , @xmath522 and @xmath524 as described above , we can produce a complement datum for @xmath3 .
such a dual system and complement datum are said to be _
compatible_. similarly , a dual system for @xmath56 and a complement datum for @xmath0 obtained from it by the same procedure
are said to be _
compatible_. we will think of compatibility as a relation . the notion of compatibility between a self - dual system and a complement datum for @xmath1 or @xmath2 is similar .
fix a self - dual system @xmath513 for @xmath58 or @xmath59 , and we construct a complement datum as follows .
let @xmath522 be a vector space complement of @xmath523 in @xmath47 , and choose a nonzero @xmath526 for @xmath540 , and let @xmath541 be determined by @xmath542 . consider the restriction of @xmath529 to @xmath543 together with the linear functionals on @xmath522 given by @xmath544 .
again , the self - dual system and any complement datum obtained from it in this way are said to be _
we will say that two sequences @xmath545 and @xmath546 are _ almost equal _ , and write @xmath547 , if @xmath548 for all but finitely many @xmath33 .
let @xmath549 be the set of @xmath550 such that @xmath551 .
in the case of @xmath3 or @xmath0 , let @xmath552 be the set of @xmath553 such that @xmath554 . for a fixed ( self-)dual system , one might choose a compatible complement datum in such a way that @xmath522 would actually contain the set of @xmath555 such that @xmath556 for all @xmath33 . in this case
, @xmath549 would be this orthogonal complement .
the examples we give will all have this property .
we do not require this property in general because it is not necessary and complicates the statements and proofs .
given a complement datum @xmath557 for @xmath3 or @xmath0 , consider the formal sum @xmath558 where in what follows we consider @xmath536 as a linear map @xmath559 .
note that @xmath560 is a well - defined linear map when restricted to @xmath561 or @xmath562 . given a complement datum @xmath563 for @xmath1 ,
we define analogously @xmath564 in the case of @xmath2 , we define @xmath565 observe that @xmath560 is a linear map when restricted to @xmath566 or @xmath567 .
moreover , we see that @xmath560 retains from @xmath536 the property of symmetry or antisymmetry . 1 . a complement datum @xmath557 for @xmath3 or @xmath0 is _ nondegenerate _ if for any nonzero @xmath568 there exists @xmath553 such that @xmath569 , and for any nonzero @xmath570 there exists @xmath550 such that @xmath571
. moreover @xmath557 is _ maximal _ if @xmath560 restricted to @xmath572 is trivial .
a complement datum @xmath563 for @xmath1 is _ nondegenerate _ if for any nonzero @xmath568 there exists @xmath550 such that @xmath573 .
moreover @xmath563 is _ maximal _ if @xmath560 restricted to @xmath574 has rank @xmath190 or @xmath8 .
a complement datum @xmath563 for @xmath2 is _ nondegenerate _ if for any nonzero @xmath568 there exists @xmath550 such that @xmath573 .
moreover @xmath563 is _ maximal _ if @xmath560 restricted to @xmath574 is trivial .
[ compatibility ] any complement datum which is compatible with a ( self-)dual system is nondegenerate . for any compatible pair , the complement datum is maximal if and only if the ( self-)dual system is maximal .
suppose @xmath557 is a complement datum compatible with a dual system for @xmath55 .
take a nonzero vector @xmath568 . since @xmath575
, we see that @xmath576 is a well - defined nonzero vector in @xmath47 .
hence , by the nondegeneracy of @xmath213 , there exists @xmath577 such that @xmath578 .
let @xmath579 , with @xmath553 .
we compute : @xmath580 similarly , if @xmath570 the analogous calculation shows that there exists @xmath550 such that @xmath571 .
we turn to maximality .
suppose that the dual system @xmath581 for @xmath55 is not maximal .
then there exist vectors @xmath582 and @xmath583 such that @xmath584 .
let @xmath585 and @xmath586 for @xmath550 and @xmath553 .
note @xmath587 and @xmath588 .
we calculate @xmath589 , and @xmath590 .
hence @xmath551 and @xmath554 , so @xmath591 and @xmath592 .
now , @xmath593 in the other direction , let @xmath557 be a nonmaximal complement datum compatible with a dual system for @xmath55 .
fix @xmath591 and @xmath592 such that @xmath594 .
let @xmath595 and @xmath596 .
we calculate @xmath597 for @xmath598 .
similarly , @xmath599 for @xmath598 .
finally , @xmath600 let @xmath557 be a nondegenerate complement datum for @xmath3 or @xmath0 .
define @xmath601 and @xmath602 .
we define a bilinear map @xmath213 on @xmath603 by extending @xmath536 via @xmath604 for @xmath550 , and @xmath605 for @xmath553 , and @xmath606 .
consider a nonzero element @xmath607 .
we see that @xmath608 is nonzero for some @xmath387 , unless @xmath591 and a certain condition on @xmath545 is satisfied , namely that @xmath609 .
but if @xmath610 is nonzero , then @xmath30 must be nonzero , and there exists @xmath553 such that @xmath611 . similarly ,
any nonzero element in @xmath48 pairs nontrivially with an element of @xmath47 .
hence @xmath213 is a nondegenerate pairing between @xmath47 and @xmath48 .
so @xmath581 is a dual system for @xmath55 or for @xmath56 which is compatible with the given complement datum .
now suppose we are given a nondegenerate complement datum @xmath563 for @xmath1 or @xmath2 .
define @xmath601 .
we define a bilinear form @xmath213 on @xmath47 by extending @xmath536 appropriately ( i.e. symmetrically or antisymmetrically ) via @xmath612 for @xmath550 , and @xmath613 for @xmath614 , and @xmath615 for @xmath540 .
consider a nonzero element @xmath607 .
we see that @xmath616 is nonzero for some @xmath387 , unless @xmath591 and the coefficients @xmath545 satisfy some condition . in the @xmath1 case , the condition is @xmath617 .
but if @xmath618 , then @xmath30 is nonzero , and there exists @xmath619 such that @xmath620 . 1
. [ def1 ] two complement data @xmath557 and @xmath625 for @xmath3 or @xmath0 are _ equivalent _ if there exist a bijection @xmath509 , isomorphisms @xmath626 and @xmath627 , and nonzero constants @xmath628 , such that @xmath629 for all @xmath550 , @xmath630 for all @xmath553 , and @xmath631 on @xmath632 .
[ def2 ] two complement data @xmath563 and @xmath633 for @xmath1 are _ equivalent _ if there exist a bijection @xmath634 such that @xmath635 , an isomorphism @xmath636 , and nonzero constants @xmath628 with @xmath637 , such that @xmath629 for all @xmath550 and @xmath631 on @xmath638 .
[ def3 ] two complement data @xmath563 and @xmath633 for @xmath2 are _ equivalent _ if there exist a bijection @xmath634 such that @xmath635 , an isomorphism @xmath636 , and nonzero constants @xmath628 with @xmath639 , such that @xmath629 for all @xmath550 and @xmath631 on @xmath638 .
note that the left hand side of the equation @xmath631 in ( [ def1 ] ) above is in fact a finite sum when applied to any @xmath640 .
likewise , @xmath641 in ( [ def2 ] ) and ( [ def3 ] ) is a finite sum when applied to any @xmath642 .
[ equivthm ] let @xmath4 be one of @xmath55 , @xmath56 , @xmath58 , and @xmath59 .
the compatibility relation between ( self-)dual systems and complement data induces a bijection between conjugacy classes of submaximal toral subalgebras of @xmath4 and equivalence classes of nondegenerate complement data . under this bijection
, conjugacy classes of maximal toral subalgebras of @xmath4 correspond to equivalence classes of nondegenerate maximal complement data .
there is a bijection between conjugacy classes of submaximal toral subalgebras of @xmath4 and conjugacy classes of dual systems for @xmath4 , which comes from the bijection of submaximal toral subalgebras of @xmath4 and dual systems for @xmath4 .
there is also a bijection between conjugacy classes of dual systems for @xmath4 and equivalence classes of dual systems , by lemma [ lemma3 ] .
we will show that compatibility induces a bijection between equivalence classes of dual systems and equivalence classes of nondegenerate complement data .
the second statement then follows immediately from proposition [ compatibility ] , which says that compatibility preserves maximality .
we first prove that compatibility induces a well - defined map from equivalence classes of dual systems to equivalence classes of complement data .
let @xmath502 be a dual system for @xmath55 , and @xmath644 an equivalent dual system for @xmath645 .
choose complement data @xmath557 and @xmath625 compatible with these dual systems .
we have implicitly chosen generators @xmath526 , @xmath527 , @xmath646 , and @xmath647 .
we have isomorphisms @xmath507 and @xmath508 and a bijection @xmath509 such that @xmath648 , @xmath511 , and @xmath512 .
let @xmath649 be the nonzero constant such that @xmath650 ( and hence @xmath651 ) .
the map @xmath652 defines a map @xmath626 and linear functionals @xmath653 .
notice that @xmath654 for all @xmath550 .
clearly @xmath655 is an isomorphism .
similarly , @xmath656 defines @xmath657 and @xmath658 .
again , @xmath659 for all @xmath553 .
compute @xmath660 hence for any @xmath550 we have @xmath661 .
the analogous calculation shows @xmath662 for all @xmath553 .
then compute @xmath663 i.e. @xmath664 on @xmath632 .
this shows that the two complement data @xmath557 and @xmath625 are equivalent .
hence the compatibility relation induces a map from equivalence classes of dual systems to equivalence classes of nondegenerate complement data , where nondegeneracy follows from proposition [ compatibility ] .
the surjectivity of this map follows from lemma [ lemma2 ] . for injectivity we will show that if two dual systems are compatible with equivalent complement data , then the two dual systems are equivalent .
suppose that the dual systems @xmath665 for @xmath55 and @xmath666 for @xmath505 are compatible with complement data @xmath557 and @xmath625 , respectively , and that these complement data are equivalent .
there exist isomorphisms @xmath626 and @xmath627 , nonzero constants @xmath628 together with a bijection @xmath667 such that @xmath629 for all @xmath550 and @xmath630 for all @xmath553 , and @xmath668 on @xmath632 .
define linear functionals @xmath669 on @xmath522 and @xmath670 on @xmath524 by @xmath671 and @xmath672 then @xmath654 for all @xmath550 and @xmath673 for all @xmath553 .
we have @xmath674 and @xmath675 . similarly , @xmath676 and @xmath677 .
we define @xmath678 and @xmath679 these isomorphisms establish the equivalence of dual systems . in the spirit of @xcite
, we introduce the five _ standard invariants _ of an infinite dual system ( and hence of an infinite - dimensional submaximal toral subalgebra or a cartan subalgebra of @xmath3 or @xmath0 ) : @xmath680 where @xmath681 and @xmath682 .
the invariants take values in @xmath683 .
if a dual system for @xmath55 or @xmath56 is compatible with the complement datum @xmath557 , then one may check that the five standard invariants are @xmath684 similarly , there are three _ standard invariants _ of an infinite self - dual system ( and hence of an infinite - dimensional submaximal toral subalgebra or a cartan subalgebra of @xmath1 or @xmath2 ) : @xmath685 where @xmath681 .
if a self - dual system for @xmath58 or @xmath59 is compatible with the complement datum @xmath563 , then the three standard invariants are @xmath686 1 .
there exists a submaximal toral subalgebra of @xmath3 or @xmath0 with standard invariants @xmath687 precisely when @xmath688 and @xmath689 .
2 . there exists a submaximal toral subalgebra of @xmath1 with standard invariants @xmath690 precisely when @xmath691 .
there exists a submaximal toral subalgebra of @xmath2 with standard invariants @xmath690 precisely when @xmath691 and @xmath692 is even when it is finite . 1 . from the definition of a nondegenerate complement datum , @xmath560 yields surjections @xmath694 and @xmath695 .
thus @xmath696 and @xmath697 .
+ given @xmath688 and @xmath689 , we construct a nondegenerate complement datum for @xmath3 or @xmath0 with standard invariants @xmath687 as follows .
let @xmath522 be a vector space with basis @xmath698 , and let @xmath524 be a vector space with basis @xmath699 .
define @xmath700 by setting @xmath701 for @xmath702 , and setting @xmath703 for @xmath704 , and setting @xmath705 for @xmath706 and letting all other pairings in this basis be trivial . for @xmath707 define @xmath708 where @xmath709 and @xmath710 , and @xmath711 where @xmath712 and @xmath713 .
then @xmath549 has basis @xmath714 , and @xmath552 has basis @xmath715 , and @xmath716 has rank @xmath692 .
this complement datum is nondegenerate , hence it gives rise to a submaximal toral subalgebra of @xmath3 or @xmath0 with standard invariants @xmath687 .
2 . from the definition of a nondegenerate complement datum
, @xmath560 yields a surjection @xmath717 .
thus @xmath718 .
+ given any @xmath691 , we construct a nondegenerate complement datum for @xmath1 with standard invariants @xmath690 as follows .
let @xmath522 be a vector space with basis @xmath698 .
we define @xmath719 by setting @xmath720 for @xmath721 and setting @xmath722 for @xmath704 . for @xmath723
define @xmath724 where @xmath709 and @xmath710 .
then @xmath549 has basis @xmath714 , and the rank of @xmath725 is @xmath692 .
this complement datum is nondegenerate , hence it gives rise to a submaximal toral subalgebra with standard invariants @xmath690 .
the proof of the inequality is the same as for @xmath1 .
we see that @xmath692 , being the rank of an antisymmetric bilinear form , is even if it is finite .
+ given any @xmath691 and @xmath692 even , we construct a nondegenerate complement datum for @xmath2 with standard invariants @xmath690 as follows .
let @xmath522 be a vector space with basis @xmath698 .
we define @xmath719 by setting @xmath726 for @xmath727 and setting @xmath722 for @xmath704 . for @xmath723 define @xmath708 where @xmath709 and @xmath710 .
then @xmath549 has basis @xmath714 , and the rank of @xmath725 is @xmath692 .
this complement datum is nondegenerate , hence it gives rise to a submaximal toral subalgebra with standard invariants @xmath690 .
a submaximal toral subalgebra of @xmath3 , @xmath0 , or @xmath2 is maximal if and only if its first standard invariant is @xmath190 .
a submaximal toral subalgebra of @xmath1 is maximal if and only if its first standard invariant is @xmath190 or @xmath8 .
hence we immediately conclude the following corollary . 1 .
[ cora]there exists a cartan subalgebra of @xmath3 or @xmath0 with standard invariants
@xmath687 precisely when @xmath728 and @xmath729 and @xmath730 .
2 . there exists a cartan subalgebra of @xmath1 with standard invariants @xmath690 precisely when @xmath731 and @xmath732 .
3 . there exists a cartan subalgebra of @xmath2 with standard invariants @xmath690 precisely when @xmath728 and @xmath733 . 1 . [ corb ]
the cardinality of the set of conjugacy classes of cartan subalgebras of @xmath3 or @xmath0 with given standard invariants is @xmath190 , @xmath8 , @xmath9 , or continuum .
one conjugacy class occurs precisely for standard invariants @xmath734 .
two conjugacy classes occur precisely for standard invariants @xmath735 , @xmath736 , @xmath737 , and @xmath738 .
the cardinality of the set of conjugacy classes of cartan subalgebras of @xmath1 with given standard invariants is @xmath190 , @xmath8 , or continuum .
one conjugacy class occurs precisely for standard invariants @xmath739 and @xmath740 .
the cardinality of the set of conjugacy classes of cartan subalgebras of @xmath2 with given standard invariants is @xmath190 , @xmath8 , or continuum .
one conjugacy class occurs precisely for standard invariants @xmath739 . 1 . let @xmath741 be a complement datum for @xmath3 or @xmath0 with standard invariants @xmath742 .
after possibly interchanging the role of @xmath522 and @xmath524 , we may assume that @xmath743 .
first we consider each of the special cases for @xmath3 , and then prove that in all other nontrivial cases the set of conjugacy classes has cardinality continuum .
let @xmath744 and @xmath745 be dual bases of @xmath47 and @xmath48 , indexed by @xmath746 . 1 . here
both @xmath522 and @xmath524 have dimension @xmath190 , so there is exactly one linear functional on each space and exactly one pairing between them .
thus there is exactly one complement datum , and it is nondegenerate and maximal . up to equivalence there is still exactly one complement datum . a dual system in this class is @xmath747 .
we see that @xmath522 is @xmath8-dimensional and @xmath524 is @xmath190-dimensional . by setting @xmath748 where @xmath749 and @xmath750 , we define two complement data .
these are inequivalent because whether @xmath751 is zero infinitely often or finitely often is an invariant .
+ notice that in the case under consideration @xmath752 and @xmath753 for an arbitrary nondegenerate maximal complement datum @xmath754 .
up to equivalence , we may assume that @xmath755 , where @xmath756 must occur infinitely often since @xmath757 .
therefore , up to equivalence , there are just two possibilities : @xmath758 occurs infinitely often or finitely often .
these equivalence clases are represented by the above complement data .
a dual system in the first class is @xmath759 , and a dual system in the second class is @xmath760 , where @xmath761 and @xmath762 .
we see that @xmath522 is @xmath8-dimensional and @xmath763 is @xmath8-dimensional .
we define two complement data by setting @xmath748 where @xmath749 and @xmath764 .
these are inequivalent because whether or not @xmath765 occurs infinitely often is an invariant .
+ since @xmath763 , @xmath766 , so every complement datum is equivalent to one with @xmath767 for all @xmath33 .
fix @xmath550 , and we may assume @xmath768 for all @xmath33 .
since @xmath757 , we must have @xmath769 for infinitely many @xmath33 .
now @xmath770 is nonzero by nondegeneracy , so we may fix @xmath553 such that @xmath771 . up to equivalence , there are just two possibilities : @xmath772 occurs infinitely often or finitely often .
a dual system in the first class is @xmath773 , and a representative of the second class is @xmath774 , where @xmath761 and @xmath762 + now suppose that @xmath729 and @xmath730 and that we are not in one of the above cases .
thus either @xmath775 or @xmath776 .
since the cardinality of the set of all complement data for @xmath522 and @xmath524 is at most continuum , it is enough to show that the cardinality of equivalence classes in each of these cases is at least continuum .
1 . we will construct a family of complement data which represents continuum many equivalence classes .
the space @xmath777 is a vector space of dimension @xmath8 which we identify with @xmath10 . for any @xmath778
we define a complement datum @xmath779 as follows .
choose @xmath751 and @xmath780 to vanish on @xmath549 and @xmath552 respectively , such that @xmath781 is @xmath8 or @xmath782 and each of these two values occurs infinitely often .
choose @xmath536 to give surjections @xmath783 and @xmath784 .
since @xmath560 coincides with @xmath536 when restricted to @xmath562 or @xmath561 , such a complement datum @xmath779 is nondegenerate and maximal for any @xmath782 . + for @xmath779 and @xmath785 to be equivalent , there must be an element @xmath786 and a permutation @xmath787 such that @xmath788 . thus there must be an element of @xmath789 sending @xmath790 to @xmath791 .
hence the only other element from this family in the equivalence class of @xmath779 is @xmath792
. therefore there are continuum many equivalence classes . 2 .
consider the projective space @xmath793 , and let @xmath794 be an at most countable subset .
we define a complement datum @xmath795 as follows .
choose @xmath796 such that for each @xmath797 , @xmath751 lies on the line @xmath798 for infinitely many @xmath33 .
choose @xmath799 such that that @xmath800 implies @xmath592 . finally , choose @xmath536 to give surjections @xmath783 and @xmath784 . since @xmath560 coincides with @xmath536 when restricted to @xmath562 or @xmath561 ,
such a complement datum @xmath795 is nondegenerate and maximal for any @xmath801 .
clearly if @xmath795 and @xmath802 are equivalent , then there must be an element @xmath803 , a bijection @xmath787 , and nonzero constants @xmath628 such that @xmath804 .
thus there must be an element of @xmath805 mapping @xmath801 to @xmath806 .
hence it is enough to show that there are uncountably many equivalence classes under @xmath805 of at most countable subsets of @xmath807 .
+ suppose that @xmath808 is finite .
fix a set of @xmath809 points in general position in @xmath807 . for each @xmath810 in general position , let @xmath811 be the union of the fixed set and @xmath812 . since @xmath813 acts with finite stabilizers on sets of @xmath809 points in @xmath814 , each @xmath811 is equivalent to at most a finite number of other such @xmath815 .
thus there are at least continuum many equivalence classes .
+ suppose that @xmath816 , where we allow @xmath817 .
the following argument was suggested by scott carnahan and anton geraschenko .
choose countably many distinct lines @xmath818 in @xmath807 , which is possible since @xmath816 .
fix a subset @xmath819 of @xmath820 .
let @xmath821 be the @xmath33th smallest element of @xmath819 .
on @xmath822 choose @xmath823 distinct points .
on @xmath824 choose @xmath825 distinct points .
on @xmath826 choose @xmath827 distinct points each of which is not colinear with any pair of the chosen points on @xmath822 and @xmath824 .
this is possible because there are only finitely many lines going through two of the previously chosen points .
continue choosing points to get a set @xmath828 in @xmath807 satisfying the property that exactly one subset of @xmath821 points lie on the same line in @xmath807 , and such that @xmath821 points lie on @xmath829 .
since linear transformations preserve colinearity , we see that @xmath828 is equivalent to @xmath830 if and only if @xmath831 .
since there are continuum many subsets of @xmath820 , we have constructed representatives of continuum many equivalence classes .
let @xmath832 be a complement datum for @xmath1 with standard invariants @xmath690 .
there are two special cases for @xmath1 , and then we prove that in all other nontrivial cases , there are continuum many conjugacy classes . 1 .
in this case @xmath833 , hence there is exactly one equivalence class of complement data . a cartan subalgebra with these invariants arises naturally from the realization of @xmath1 as @xmath834 .
2 . in this case
@xmath835 is @xmath8-dimensional and @xmath560 is nontrivial . since @xmath836 , there is exactly one equivalence class of complement data . a cartan subalgebra with these invariants arises naturally from the realization of @xmath1 as @xmath837 .
3 . consider the vector space @xmath838 .
consider an at most countable set @xmath839 .
we define a complement datum @xmath795 as follows .
choose @xmath751 to vanish on @xmath549 such that @xmath840 and each element of @xmath801 occurs for infinitely many @xmath33 .
choose @xmath536 to have rank @xmath692 on @xmath549 and to give a surjection @xmath841 .
since @xmath560 coincides with @xmath536 when restricted to @xmath566 , such a complement datum @xmath795 is nondegenerate and maximal for any @xmath801 .
clearly if @xmath795 and @xmath802 are equivalent , then there must be an element @xmath803 and a bijection @xmath787 such that @xmath842 .
thus there must be an element of @xmath843 mapping @xmath801 to @xmath806 .
hence it is enough to show that there are uncountably many equivalence classes under @xmath843 of at most countable subsets of @xmath844 .
+ now we can use the same two arguments from the analogous case for @xmath3 .
if @xmath808 is finite , we can choose @xmath845 points in general position and proceed as before .
if @xmath846 , including @xmath817 , then we can choose points on lines in @xmath844 satisfying certain collinearity properties as before .
3 . let @xmath832 be a complement datum for @xmath2 with standard invariants @xmath847 .
there is one special case for @xmath2 .
1 . in this case
@xmath833 , thus there is exactly one conjugacy class of cartan subalgebras , and they arise naturally from the realization of @xmath2 as @xmath848 .
2 . repeat the proof of the @xmath849 case for @xmath1 , choosing @xmath751 such that @xmath850 . in particular , for @xmath1 and @xmath2
, if there are finitely many conjugacy classes of cartan subalgebras with given standard invariants , then the cartan subalgebras with these standard invariants are splitting .
we conclude the paper with a description of a case with uncountably many equivalence classes .
let @xmath744 and @xmath745 be dual bases of @xmath47 and @xmath48 , indexed by @xmath746 . consider for any pair of sequences @xmath851 and @xmath852 the dual system for @xmath55 given by @xmath853 .
this dual system has invariants @xmath854 if and only if the sequences @xmath851 and @xmath852 are not almost equal to @xmath190 .
the following are three binary invariants of these dual systems : whether for infinitely many @xmath33 , @xmath545 is zero and @xmath546 is nonzero ; whether for infinitely many @xmath33 , @xmath545 is nonzero and @xmath546 is zero ; and whether for infinitely many @xmath33 , both @xmath545 and @xmath546 are zero .
we define a _ multiset _ on a set @xmath801 as a map @xmath855 such that @xmath808 vanishes outside a countable subset of @xmath801 .
we say @xmath856 is the _ multiplicity _ of @xmath30 .
suppose the group @xmath789 acts on @xmath801 .
two multisets @xmath808 , @xmath857 on @xmath801 are _ almost proportional _ if there exists @xmath858 such that for all @xmath859 , @xmath860 , and @xmath861 for all but finitely many @xmath859 .
a sequence @xmath862 gives rise to a multiset @xmath855 by setting @xmath863 .
two sequences are _ almost proportional _ if the multisets arising from them are almost proportional .
consequently , two sequences @xmath864 and @xmath865 almost proportional if and only if there exists @xmath866 and a permutation @xmath787 such that @xmath867 . every dual system with standard invariants
@xmath854 is equivalent to one of the form @xmath853 .
two such dual systems are conjugate if and only if their three binary invariants agree and the sequence @xmath868 is almost proportional to the sequence @xmath869 . fix a dual system with standard invariants @xmath854 , and choose a compatible complement datum @xmath625 which satisfies the further condition that the @xmath870 vanish on @xmath871 and the @xmath872 vanish on @xmath873 .
choose nonzero vectors @xmath874 and @xmath875 . since the complement datum is nondegenerate
, we can choose vectors @xmath876 and @xmath877 such that @xmath878 .
as the complement datum is maximal , we have @xmath879 .
after possibly replacing @xmath30 with @xmath880 , we may assume without loss of generality that @xmath881 .
let @xmath882 and @xmath883 .
consider the dual system @xmath884 , where @xmath885 and @xmath886 .
let @xmath887 and @xmath888 . define a map @xmath889 by @xmath890 , @xmath891 , and define @xmath892 by @xmath893 , @xmath894 .
it is clear that @xmath655 is an equivalence of dual systems .
thus every equivalence class of dual systems has a representative of the required form .
it remains to show that two such dual systems are equivalent precisely when they have the same binary invariants and the sequences @xmath868 and @xmath869 are almost proportional .
suppose two dual systems , given by sequences @xmath851 , @xmath852 , @xmath895 , and @xmath896 , are equivalent .
there exist a bijection @xmath509 , isomorphisms @xmath897 and @xmath898 , and nonzero constants @xmath628 , such that @xmath899 and @xmath900 .
we adopt the notation @xmath901 , @xmath902 , @xmath903 , and @xmath904 .
suppose @xmath905 is given by @xmath906 , @xmath907 .
then @xmath908 implies @xmath909 , hence @xmath910 .
since @xmath655 is an isomorphism , it follows that @xmath161 and @xmath692 are nonzero .
we also have @xmath911 , thus @xmath912 .
similarly , there is a nonzero @xmath913 such that @xmath914 . therefore the binary invariants agree .
moreover @xmath915 , so the sequences @xmath868 and @xmath869 are almost proportional .
conversely , suppose that @xmath916 with @xmath917 nonzero and that the binary invariants agree .
choose nonzero constants @xmath628 such that @xmath918 and @xmath919 .
this will be possible precisely because the sequences binary invariants agree .
let @xmath920 be the constant @xmath921 .
define @xmath905 by @xmath922 and @xmath923 , and define @xmath924 by @xmath925 and @xmath926 .
one calculates that @xmath899 , @xmath900 , and @xmath631 .
we can visualize the preceding proposition as follows .
cartan subalgebras of @xmath3 with standard invariants @xmath854 are in bijection with certain `` admissible '' multisets on the parameter space depicted in figure [ fig2 ] , modulo almost proportionality .
more precisely , the parameter space is @xmath927 , where the action of @xmath789 on @xmath928 is @xmath929 .
a multiset is admissible if infinitely many points lie on each of the two lines .
the action of @xmath789 on the parameter space is @xmath930 = [ ( cx , cy ) ] = [ ( c^2x , y)]$ ] , and hence elements of @xmath789 simply rescale the line .
a similar analysis holds more generally .
let @xmath801 be the quotient @xmath931 , where @xmath789 acts on @xmath932 by @xmath929 .
there is a residual diagonal action of @xmath789 on @xmath801 , which enables us to define almost proportional multisets on @xmath801 .
if two complement data are equivalent , then the images of @xmath933 and @xmath934 in @xmath801 must be almost proportional
. however , the converse does not hold , as different choices of @xmath536 often yield inequivalent dual systems .
it is difficult to clarify this dependence in general ; nonetheless , in practice one should be able to distinguish nonconjugate dual systems .
for example , consider complement data with standard invariants @xmath935 and a fixed sequence @xmath933 of @xmath801 with no unusual symmetries .
it is easy to see that any two choices of @xmath536 yield inequivalent dual systems .
y. bahturin , g. benkart , _ highest weight modules for locally finite lie algebras _ , ams / ip stud . in adv
. math . 4 ( 1997 ) , 131 . y. bahturin , h. strade , _ some examples of locally finite simple lie algebras _ , arch
( basel ) 65 ( 1995 ) , 2326 .
a. baranov , _ diagonal locally finite lie algebras and a version of ado s theorem _
, j. algebra 199 ( 1998 ) , 139 .
a. baranov , _ simple diagonal locally finite lie algebras _
london math .
77 ( 1998 ) , 362386 .
a. baranov , a. zhilinski , _ diagonal direct limits of simple lie algebras _ , comm .
algebra 27 ( 1998 ) , 22492766 .
n. bourbaki , _ groupes et algbres de lie _ , hermann , paris , 1975
. i. dimitrov , i. penkov , _ weight modules of direct limit lie algebras _ , intern .
notices 1999 , no . 5 , 223 - 249 , math.rt/9808045 . i. dimitrov , i. penkov , _ borel subalgebras of @xmath936 _ , resenhas ime - usp 6 ( 2004 ) , no . 2/3 , 153 - 163 .
g. mackey , _ on infinite dimensional linear spaces _ , trans .
ams 1945 , vol .
57 , no . 2 , 155 - 207 .
neeb , i. penkov , _ cartan subalgebras of @xmath3 _ , canad .
46(2003 ) , 597 - 616 . | root - reductive lie algebras are direct limits of finite - dimensional reductive lie algebras under injections which preserve the root spaces .
it is known that a root - reductive lie algebra is a split extension of an abelian lie algebra by a direct sum of copies of finite - dimensional simple lie algebras as well as copies of the three simple infinite - dimensional root - reductive lie algebras @xmath0 , @xmath1 , and @xmath2 . as part of a structure theory program for root - reductive lie algebras ,
cartan subalgebras of the lie algebra @xmath3 were introduced and studied in @xcite . in the present paper
we refine and extend the results of @xcite to the case of a general root - reductive lie algebra @xmath4 .
we prove that the cartan subalgebras of @xmath4 are the centralizers of maximal toral subalgebras and that they are nilpotent and self - normalizing .
we also give an explicit description of all cartan subalgebras of the simple lie algebras @xmath0 , @xmath1 , and @xmath2 .
we conclude the paper with a characterization of the set of conjugacy classes of cartan subalgebras of the lie algebras @xmath3 , @xmath0 , @xmath1 , and @xmath2 with respect to the group of automorphisms of the natural representation which preserve the lie algebra . |
the last ten years have seen a crescendo of discoveries that have revolutionized our understanding of the process by which the tiny graininess seen in the cosmic microwave background ( cmb ) resulted in the formation of galaxies .
up to a decade ago , galaxy evolution models followed the approach pioneered by tinsley ( 1972 ) , larson & tinsley ( 1974 ) , tinsley & gunn ( 1976 ) , the evolutionary population synthesis approach , wherein galaxies form at some redshift @xmath0 , perhaps varying by type , with some star formation history @xmath1 and some stellar initial mass function ( imf ) .
the general framework was that postulated by eggen et al .
( 1962 ) and sandage ( 1986 ) : elliptical galaxies had essentially a simple passive evolution after formation at high @xmath2 in a single rapid collapse and starburst , while the timescale for star formation of late type galaxies is much longer and in fact comparable with the hubble time .
models were required to reproduce the photometric properties of local galaxies .
the more refined models dealt simultaneously with both photometric and chemical evolution .
guiderdoni & rocca - volmerange ( 1987 ) were the first to introduce _
dust absorption_. however , the general attitude was that dust extinction is a mere correction that does not change the main evolutionary trends seen in the optical band .
this view appeared to be grounded by the fact that a comparison of the optical and far - ir luminosity densities in the local universe shows that only @xmath3 of starlight is reprocessed by dust .
mazzei et al .
( 1992 , 1994 ) produced the first self - consistent chemical and photometric evolution models including both dust absorption and emission , to get complete spectral energy distributions ( seds ) from uv to far - ir both for late- and early - type galaxies .
these models showed that , while the seds of late - type galaxies are not expected to change drastically throughout their evolutionary history , under plausible assumptions spheroidal galaxies might have been very optically thick during their early phases , if indeed most of their stars were formed in a giant starburst .
their seds would then have been similar to that of hyper - luminous iras galaxies .
mazzei & de zotti ( 1996 ) showed that ( sub)-mm observations of a number of high redshift radio galaxies supported this scenario .
these models were exploited by franceschini et al .
( 1994 ) to produce the first set of galaxy counts at optical _ and _ far - ir to sub - mm wavelengths and to predict the sed of the cosmic ir background ( cirb ) ; a remarkable result was that the cirb intensity might be as large as , or larger than , that of the optical background .
up to that time , the optical quest for `` normal '' high-@xmath2 ( `` primeval '' ) galaxies yielded a very meager return : only a handful of objects were discovered , mostly serendipitously or through absorption features in the spectra of distant quasars or gravitational lensing .
the only effective method for finding high-@xmath2 galaxies was optical identifications of radio sources , which are generally associated to early type galaxies .
a major breakthrough occurred roughly at the same time , thanks to the color selection technique targeting the lyman discontinuity at @xmath4 , developed by steidel & hamilton ( 1992 , 1993 ) .
follow - up spectroscopy with the keck telescope confirmed the efficiency of this technique ( steidel et al .
1996 ) which led to the discovery of a widespread population of high-@xmath2 galaxies .
madau et al .
( 1996 ) applied the same technique to the multicolor images of the hubble deep field to identify galaxies at @xmath5 ( @xmath6 drop - outs ) and at @xmath7 ( @xmath8 drop - outs ) .
they compared the luminosity density in hdf high-@xmath2 galaxy samples to that found in low-@xmath2 survey and interpreted these data in terms of evolution with redshift of the star - formation / metal - production rate . although the results at high-@xmath2 were presented as lower limits ( because nearly all the corrections to be applied drive the derived star formation rates up ) it was argued that they may be not far from the true values and that they are consistent with a peak in the metal production rate at @xmath9 .
there was a remarkable , at least qualitative , agreement between the metal production rate , @xmath10 , estimated in this way and the prediction of the cosmic chemical evolution models by pei & fall ( 1995 ) , and the predictions of hierarchical cdm models ( baugh et al .
1998 ) , all of which show a peak in the metal production rate at @xmath1112 . on the other hand , their conclusions were affected by several large uncertainties due to ( ferguson et al .
2000 ; somerville et al .
2001 ) : 1 .
selection criteria : the adopted color selection was extremely conservative and spectroscopic surveys ( steidel et al .
1996 ; lowental et al .
1997 ) have identified at least a dozen @xmath12 galaxies that would have been missed 2 .
sampling of the luminosity function : madau et al .
( 1996 ) simply summed up the observed fluxes of galaxies , although only the bright tail of the lf was sampled ( thompson et al .
surface brightness corrections : for fixed luminosity and physical size , the surface brightness drops by about 1 mag between @xmath13 and @xmath14 .
this may produce a substantial underestimate of @xmath10 at high @xmath2 ( ferguson 1998 ; lanzetta et al .
effects of dust : dust is ubiquitous in star - forming regions and relatively modest amounts of dust suffice to strongly attenuate the uv emission .
analysis of uv spectral slopes yielded evidences of substantial dust extinction ( pettini et al .
1998 ; meurer et al . 1997 , 1999 ; sawicki & yee 1998 ; steidel et al . 1999 ; flores et al .
1999 ; mobasher & mazzei 2000 ) .
the derived upward corrections of the luminosity densities by madau et al .
( 1996 ) range from factors of @xmath15 to more than 10 , indicating that the correction based on optical / uv data is both large and very uncertain . even more : very dusty galaxies , akin to ultra - luminous iras galaxies , that were present at high @xmath2 , would be largely unaccounted by these studies because they would be unlikely to be identified as lyman - break galaxies . all that boils down to the conclusion that the optical view of `` primeval '' galaxies is incomplete and biased
. it must be complemented with far - ir / sub - mm observations .
almost simultaneous with the discovery of lbgs was a second major breakthrough : the discovery of the cosmic infrared background ( puget et al .
1996 ; guiderdoni et al . 1997 ; schlegel et al . 1998 ; fixsen et al . 1998
; hauser et al . 1998 ; lagache et al . 1999 , 2000 ; finkbeiner et al . 2000 ) at a level 10 times higher than the no - evolution predictions based on the iras local luminosity function of galaxies and twice as high as the cosmic optical background obtained from the optical counts , implying that most of starlight emitted at early phases of galaxy evolution is actually reprocessed by dust ( remember that in the local universe dust emission accounts for only @xmath3 of the bolometric luminosity of galaxies ) .
the cirb sed turned out to be in remarkably good agreement with the models by franceschini et al .
( 1994 ) whereby most of the star formation in spheroidal galaxies was enshrouded by dust .
thus , _ far - ir to sub - mm observations carry the most direct , minimally biased information on the cosmic star - formation / metal - production history .
_ sub - mm observations have a particularly privileged role in the game : their uniquely strong , negative k - correction , due to the steep increase with frequency of dust emission in galaxies , coupled with the strong cosmological evolution demonstrated by iso and scuba data ( elbaz et al . 1999 ; smail et al .
1997 ; hughes et al . 1998
; barger et al .
1999a , b ; eales et al . 1999 ) as well as by the cirb intensity , greatly emphasizes high - redshift galaxies .
such observations therefore hold the key to understanding the formation and evolution of massive elliptical galaxies , a still controversial issue in cosmology .
two quite different scenarios are still confronting each other .
one scenario , motivated by hierarchical clustering cold dark matter ( cdm ) cosmologies , predicts that most large galaxies arise from a series of merging events taking place over a major fraction of the cosmological time ( baugh et al .
1998 ; kauffmann 1996 ) .
the other scenario assumes instead that the bulk of the baryonic mass of the galaxy was assembled in gaseous form .
most stars formed in a single gigantic starburst , followed by essentially passive evolution , due to the ageing of stellar populations .
this scheme is sometimes qualified as _
it is worth stressing that the _ monolithic _ scenario too can be consistent with the hierarchical gravitational instability picture for the structure formation .
the difference with the _ merging _ scenario is that large ellipticals formed most of their stars _ in situ _ as soon as the corresponding dark matter halos condensed out . in the _ merging _
scenario most stars are formed in smaller galaxies that subsequently merged and the merging events eventually triggered further bursts of star formation . in the _ merging _ hypothesis , ellipticals would exhibit a broad range of colors and spectral properties , reflecting the variety of star formation histories , the bulk of stars having an intermediate age , and their number density should decrease with increasing redshift . on the contrary , in the _ monolithic _ scenario the number density should remain constant and the stars should be old and essentially coeval .
several lines of observational evidence ( the tightness of the fundamental plane for ellipticals in local clusters , the tight color - magnitude relation for ellipticals in clusters up to @xmath16 , the modest shift with increasing redshift in the zero - point of the fundamental plane , mg@xmath17 , and color - magnitude relations for cluster ellipticals , the small zero - point offset between cluster and field ellipticals ) converge in indicating that most stars in galactic spheroids are old ( formed at @xmath183 ) and essentially coeval ( see renzini & cimatti 1999 , shade et al . 1999 , and kodama et al .
1999 for recent discussions ) .
it is , however , still controversial whether ellipticals were assembled to their present mass at a later epoch , compared to that of formation of their stars ( see renzini & cimatti 1999 , ferguson et al . 2000 , and references therein , and the recent work by martini 2001 and rodighiero et al .
good agreement was recently found by daddi et al .
( 2000 ) between the observed surface density of extremely red objects ( eros ) in the widest field survey for such objects available , with that predicted by pure luminosity evolution ( ple ) models after applying the appropriate color and luminosity thresholds . since there are indications that the bulk of eros are passively evolving ellipticals , this result supports the view that most field ellipticals were fully assembled by @xmath19 .
sub - mm data provide information on an earlier stage of formation of ellipticals than optical data do , i.e. on the stage when these galaxies were forming most of their stars .
scuba surveys did in fact provide the most critical test for the _ merging _ scenario . a very comprehensive attempt to make detailed predictions for the observable properties of galaxies throughout the electromagnetic spectrum , at all redshifts , starting from an assumed initial spectrum of density fluctuations ,
was carried out by granato et al .
these authors combined the semi - analytical galaxy formation model of cole et al .
( 2000 ; galform ) with the evolutionary spectro - photometric models including the effect of dust by silva et al .
( 1998 ; grasil ) , to compute epoch - dependent galaxy luminosity functions from uv to sub - mm wavelengths .
galform evolves the primordial density fluctuations and applies simplified analytical descriptions of the main physical processes of gas cooling and collapse , star formation , feedback effect from supernovae , galaxy merging , chemical enrichment of stars and gas , and more , on top of a monte carlo description of the process of formation and merging of dark matter halos through hierarchical clustering .
grasil follows the evolution of stellar populations and absorption and emission by dust with a realistic 3d geometry , with a disk and a bulge , two - phase dust ( in clouds and in the diffuse ism ) , star formation in the clouds , radiative transfer of starlight through the dust distribution , a realistic dust grain model including pahs and quantum heating of small grains , and a direct evaluation of the dust temperature distribution at each point in the galaxy .
the models include both galaxies forming stars quiescently in disks , and starbursts triggered by galaxy mergers .
they reproduce quite well a large number of observables , including the seds of normal spirals and starbusts from uv to sub - mm ; their internal extinction properties and , in particular , the observed relationship between far - ir / uv luminosity ratio and the slope of the uv continuum and the observed starbust attenuation law ( calzetti et al . 2000 ) with the same dust mixture which reproduces the milky way extinction law ; the local galaxy luminosity functions from uv to far - ir ; the counts of galaxies in wavebands from uv to far - ir .
_ however , they fail by a factor @xmath20 to reproduce the sub - mm counts _ , which turn out to be the most critical test for semi - analytic approach .
in fact , analogous conclusions are reached by devriendt & guiderdoni ( 2000 ) who implemented a independent _ ab - initio _ models , building on earlier work by guiderdoni et al .
( 1997 , 1998 ) and exploiting the spectro - photometric evolution models , including dust , developed by devriendt et al .
( 1999 ; stardust ) . to reproduce the @xmath21 m counts they are forced to introduce an _ ad - hoc _ population of heavily extinguished , massive galaxies with large star - formation rates at intermediate and high redshifts .
these difficulties , affecting even the best current recipes , may indicate that new ingredients need to be taken into account
. a key new ingredient may be the mutual feedback between formation and evolution of spheroidal galaxies and of active nuclei residing at their centers .
in the past , evolutionary histories of quasars and of galaxies have been dealt with as essentially independent subjects .
this was due to two main reasons .
on one side , it was essentially impossible to have a synoptic view of the evolutionary history of the two populations , since the high redshift universe was almost exclusively the realm of quasars while the `` local '' universe ( @xmath22 ) was the realm of galaxies .
on the other side , and perhaps more important , physical processes involved in the formation and evolution of both quasar and galaxies are of great complexity and therefore difficult to handle without the guidance of direct observational data .
the situation has drastically improved in the last years with the detection of ordinary and starburst galaxies up to @xmath23 which has opened to direct investigation the early phases of galaxy evolution .
there is know convincing evidence of substantial star - formation activity at least up to @xmath24 ( lagache et al .
1999 ; steidel et al . 1999 ) .
on the other hand , extensive data on the demography of super - massive black holes ( bh ) in nearby galaxies has been accumulating , allowing to piece together an increasingly quantitative description of their mass function ( salucci et al .
1999 ) and of where ( and possibly when ) they formed .
there is no evidence of quasar activity outside galaxies and deep imaging of bright quasar samples have revealed that their hosts are most likely elliptical galaxies ( hall & green 1998 ; mclure et al .
1999 ; schade et al .
2000 ) . as recognized by haehnelt & rees ( 1993 ) , the epoch of quasar activity
coincides with the time when the first potential wells form in the standard cdm scenario . in the same framework ,
haehnelt et al . (
1998 ) demonstrated that the luminosity function of actively star - forming galaxies at @xmath13 and the b - band luminosity function of quasars at the same redshift can both be matched with the mass function of dark matter halos .
suggestive similarities have been found between the history of the blue luminosity density produced by stars and quasars ( cavaliere & vittorini 1998 ) .
the similarity persists and even strengthens when the history of star formation is compared with that of the global volume emissivity of active galactic nuclei ( franceschini et al .
. the evolution of the luminosity density of extragalactic radio sources also appears to remarkably parallel that of the star formation rate ( dunlop 1998 ) .
if quasars were in place roughly at the same time as galaxies , it is natural to expect that , given their high bolometric luminosities , they play an important role in the early evolution of galaxies ( silk & rees 1998 ; blandford 1999 ; monaco et al .
2000 ) : they may provide crucial feedback for limiting star formation , for arresting its growth , for inducing or inhibiting collapse of further density perturbations in the neighborhood and for ionizing the inter - galactic medium ( blandford 1999 ) . in turn ,
the fuelling of quasars is likely to be tightly related to galaxy evolution .
quasars are widely believed to be powered by accretion into a massive bhs . as discussed by cavaliere & vittorini ( 1998 , 2000 )
the data on quasar evolution can be interpreted in terms of fuelling of bhs during the formation of the host galaxy while , at later times , accretion is rekindled by interactions of the host galaxy with companions .
it follows that a consistent evolutionary scenario must deal with the combined evolution of bhs and of their host galaxies .
the link between the two classes of objects is given by the observed relationship between the bh mass and the mass ( or , better , the velocity dispersion ) of the host galaxy ( magorrian et al .
1998 ; ferrarese & merritt 2000 ; gebhardt et al .
2000 ) . if this relationship was imprinted during the early phases of the evolution ( silk & rees 1998 ; fabian 1999 )
, the formation rate of spheroids can be derived directly from the formation rate of quasars , which , in turn , can be inferred from the epoch - dependent luminosity function , now observationally determined ( or at least estimated ) up to @xmath25 ( boyle et al .
2000 ; fan et al . 2001 ) , given the quasar lifetime , @xmath26 .
the latter quantity can be determined from the condition of consistency with the present day mass function of spheroids ( salucci et al .
1999 ) , giving @xmath27yr , well within the range of current estimates yielding @xmath28@xmath29yr ( salucci et al . 1999
; martini & weinberg 2001 ; monaco et al . 2000 ) .
* once the effects of cooling and heating processes ( the latter being mostly due to stellar feedback ) are properly taken into account , the timescale for star formation within virialized dark matter ( dm ) halos turns out to be relatively short for massive spheroidal galaxies ( @xmath30gyr for @xmath31 ) , while in the case of less massive halos the feedback from supernovae ( and/or from the active nucleus ) slows down the star formation and can also expel a significant fraction of the gas . _ the canonical scheme implied by the hierarchical cdm scenario
small clumps collapse first is therefore reversed when we consider baryons .
_ during this phase , large spheroidal galaxies show up as luminous sub - mm sources , accounting for the @xmath21 m counts . * when the quasar luminosity reaches a high enough value , its action ( ionization and heating of the gas ) , together with that of sn explosions , stops the star formation and eventually expels the residual gas .
this naturally explains the observed correlation between bh and host spheroidal masses ( silk & rees 1998 ; fabian 1999 ) .
the same mechanism distributes in the igm a substantial fraction of metals . *
a `` quasar phase '' follows , lasting @xmath32@xmath29yr .
* intermediate- and low - mass spheroids have lower sfr s and less extreme optical depths .
they show up as lbgs . * a long phase of passive evolution follows , galaxy spectra becoming rapidly red . determining the redshift distribution of scuba source is obviously of utmost importance ,
since it allows us to obtain a unbiased view of the star - formation history of the universe and potentially holds the key to understanding the formation and evolution of the most massive ellipticals .
unfortunately this is a very difficult job , due to the rather poor angular resolution of scuba and to the faintness of the optical counterparts , making the identification ambiguous in most cases . of greatest importance
are deep vla maps of scuba fields which allow us to identify radio counterparts or place stringent limits on the radio flux of the sub - mm galaxies .
the recent work on sub - mm to radio spectral indices ( carilli & yun 1999 , 2000 ) has provided us with a useful tool for estimating , or constraining , the redshifts . the redshift information derived in this manner can be compared with the spectroscopic redshifts for individual candidate optical counterparts to determine the reliability of proposed identifications . in spite of the fact that only a handful of redshifts of sources detected in blanck - field scuba surveys have been reliably measured ,
the results summarized in dunlop ( 2001 ) indicate that most scuba sources are probably at @xmath33 , consistent with the scenario in which the peak agn emission corresponds to , or even causes the termination of , major star - formation activity in giant spheroidal galaxies ( granato et al .
given that the optical emission of powerful quasars peaks at @xmath34 , dust emission is expected to peak @xmath35gyr earlier , i.e. at @xmath36 ( dunlop 2001 ) .
at early times , collapsed objects are likely to be in the highest peaks of the density field , since one needs a dense enough clump of baryons in order to start forming stars .
such high-@xmath17 peaks are highly biased tracers of the underlying density field ( kaiser 1984 ; bardeen et al .
1986 ; mo & white 1996 ) . the bias factor is strongly dependent on the mass of dark halos .
therefore , clustering measurements allow estimates of involved masses .
although the samples of sub - mm galaxies are still too small to allow determinations of their clustering properties , large cell - to - cell fluctuations are seen , suggestive of strong clustering .
gravitational lensing of extragalactic light by line of sight mass concentrations can strongly amplify fluxes of distant sources ( peacock 1982 ) .
although the probability of strong lensing is very small , its distribution has a power - law tail ( @xmath37 ) extending up to very large values ( @xmath3830 for extended sources ; perrotta et al .
if counts decrease with increasing flux fast enough , the fraction of lensed sources at bright fluxes may be large .
the special properties of the sub - mm counts make this spectral region ideally suited for detecting lensed sources .
in fact , as noted above , this spectral region greatly emphasizes high - redshift sources ( which have the highest probability of being gravitationally lensed ) and yields very steep counts , thus maximizing the amplification bias ( peacock 1982 ; turner et al .
thus , we expect that the fraction of gravitationally lensed galaxies is boosted in large area , relatively shallow surveys at sub - mm wavelengths ( blain 1996 , 1998 ) . in the scenario by granato et al .
( 2001 ) the bulk of star - formation in massive early - type galaxies occurred at @xmath39 and was essentially completed at @xmath40 .
this translates into an essentially exponential turnover of the counts , reflecting the high - mass turnover of the mass ( and of the associated luminosity ) function of galactic halos in place at the relevant @xmath2 .
while this extremely fast decline of the counts at bright fluxes may imply that unlensed forming spheroidal galaxies will not be detectable by large area , relatively shallow surveys such as those to be carried out by esa s planck satellite , on the other hand , the amplification bias due to lensing is maximized . as a result ,
the counts of this population may be actually dominated , at the brightest flux levels , by highly magnified sources .
this characteristic feature of the model by granato et al .
( 2001 ) may help to discriminate between different scenarios , such as the one proposed by rowan - robinson ( 2001 ) , predicting less steep counts , implying a much lower fraction of highly magnified , bright sources ( perrotta et al . 2001b ) . | the intensity of the cosmic far - ir background and the strong evolution of galaxies in the far - ir to mm wavelength range demonstrate that the bulk of starlight emitted during the early phases of galaxy evolution was reprocessed by dust .
therefore , the optical view of the galaxy formation process is highly incomplete and biased , and must be complemented with far - ir / sub - mm observations .
we review the impact of sub - mm surveys on our understanding of the evolutionary history of spheroidal galaxies . a recent model , bringing into play also the inter - relationships between formation and early evolution of spheroidal galaxies and quasars , is described and some implications are outlined .
# 1 # 2 _ mem .
soc .
astron . it . _ * # 1 * , # 2 # 1 # 2 _ the messenger _ * # 1 * , # 2 # 1 # 2 _ astron .
nach . _ * # 1 * , # 2 # 1 # 2 _ astron .
astrophys . _
* # 1 * , # 2 # 1 # 2 _ astron .
astrophys .
lett . _ * # 1 * , l#2 # 1 # 2 _ astron .
astrophys .
rev . _ * # 1 * , # 2 # 1 # 2 _ astron .
astrophys .
suppl .
ser . _ * # 1 * , # 2 # 1 # 2 _ astron .
j. _ * # 1 * , # 2 # 1 # 2 _ ann .
rev .
astron .
astrophys . _ * # 1 * , # 2 # 1 # 2 _ astrophys .
j. _ * # 1 * , # 2 # 1 # 2 _ astrophys .. j. lett .
_ * # 1 * , l#2 # 1 # 2 _ astrophys .
j. suppl .
_ * # 1 * , # 2 # 1 # 2 _ astrophys .
space sci .
_ * # 1 * , # 2 # 1 # 2 _ adv .
space res .
_ * # 1 * , # 2 # 1 # 2 _ bull .
astron .
inst .
czechosl . _ * # 1 * , # 2 # 1 # 2 _ j. quant .
spectrosc .
. transfer _ * # 1 * , # 2 # 1 # 2 _ mon . not .
r. astr .
soc .
_ * # 1 * , # 2 # 1 # 2 _ mem .
r. astr .
soc . _ * # 1 * , # 2 # 1 # 2 _ phys .
lett .
rev . _ * # 1 * , # 2 # 1 # 2 _ publ .
astron .
. japan _ * # 1 * , # 2 # 1 # 2 _ publ .
astr .
soc .
pacific _ * # 1 * , # 2 # 1 # 2 _ nature _ * # 1 * , # 2 epsf.sty |
dilepton invariant - mass spectra provide the unique opportunity to directly probe the electromagnetic ( e.m . )
spectral function of the hot and dense medium created in energetic collisions of heavy nuclei @xcite . in the low - mass region ( lmr , @xmath8 gev )
, this enables the study of modifications of the light vector mesons ( v=@xmath1 , @xmath2 , @xmath3 ) caused by their interactions with the surrounding matter particles and/or changes in the underlying condensate structure . in the intermediate - mass region ( imr , @xmath9 )
, electromagnetic emission is expected to become continuum - like with rather well - defined strength rendering it a suitable tool to infer temperatures of the excited system ( well ) before interactions cease ( `` thermal freezeout '' ) .
dilepton measurements in heavy - ion collisions at the cern - sps ( at center - of - mass energies of @xmath10=17.3 and 8.8 agev ) have proven the presence of substantial excess radiation beyond e.m .
final - state decays of produced hadrons in both the lmr @xcite and imr @xcite .
this , in particular , corroborated the presence of interacting _
matter _ over a duration of @xmath1110 - 15 fm/@xmath12 in central pb - au collisions .
moreover , the spectral shape of the excess in the lmr could only be accounted for if major medium modifications of the vacuum @xmath1-meson line shape were incorporated @xcite .
both the implementation of a dropping mass or a strong broadening of its width as following from hadronic many - body approaches , were compatible with the large enhancement observed at masses below the free @xmath1 mass .
in addition , with the same underlying fireball model ( lifetime and temperature evolution ) , reasonable agreement with the enhancement measured in the imr @xcite has been established @xcite ( as well as with direct photon spectra @xcite ) .
the excess for @xmath13 gev was largely attributed to four - pion type annihilations in the hadronic phase , with a strength determined by the vacuum e.m .
spectral function , while the calculated quark - gluon plasma ( qgp ) contribution to the thermal yield amounted to @xmath11@xmath14 ( @xmath11@xmath15-@xmath16 in the lmr @xcite ) .
a recent , substantial , improvement in statistics and mass resolution in low - mass dimuon spectra in in - in collisions @xcite shows good agreement with predictions for thermal radiation with in - medium @xmath1-meson broadening as following from hadronic many - body calculations @xcite .
the shape of the excess radiation is well described from threshold ( @xmath17 ) to @xmath18 gev , while the absolute yield is overpredicted by about @xmath14 .
the latter can be accommodated by a minor adjustment in the thermal fireball evolution ( amounting to a @xmath14 lower fireball lifetime ) , and after inclusion of in - medium @xmath2 and @xmath3 decays ( whose contribution is mostly localized around their free mass ) , as well as four - pion type annihilation ( which sets in at masses @xmath19 gev ) , a quantitative description of the na60 invariant - mass spectra from threshold to @xmath11@xmath20 gev in central in - in collisions emerges @xcite ( see also refs .
while the inclusive mass spectra are well described , the comparison of the calculations with newly released transverse pair - momentum ( @xmath5 ) spectra @xcite reveals some discrepancies at @xmath21 gev in semicentral in - in . in the present article
we reiterate the main points of our previous study of dilepton invariant - mass spectra at the sps , and extend the analysis to @xmath5 spectra .
in particular , we conduct a detailed analysis of sources at high @xmath5 , in terms of ( i ) an improved treatment of @xmath1 decays at thermal freezeout ( which are subject to maximal blue shift due to transverse flow ) , ( ii ) a component of primordial ( hard - produced ) @xmath1 mesons subject to energy loss when traversing the medium ( using high-@xmath22 pion spectra as a guideline ) , ( iii ) drell - yan ( dy ) annihilation in primordial @xmath23-@xmath23 collisions which we extrapolate to small mass by imposing constraints from real photon production , and ( iv ) meson @xmath7-channel exchange contributions to the thermal production rate which are not included in the many - body vector - meson spectral functions ( but which have been found to be a significant source at high @xmath5 in real photon production @xcite ) gev .
therefore , the inclusion of these contributions does not upset our earlier description of inclusive @xmath24 spectra in terms of thermal radiation and freezeout @xmath1 s .
initial results of these studies have been reported in refs .
@xcite . ] .
another interesting issue which has received little attention thus far is how uncertainties in the critical temperature and hadro - chemical evolution of the fireball affect dilepton spectra . for hadro - chemical freezeout
we investigate the sensitivity to temperatures in the range @xmath25-@xmath26 mev , representative for top sps energy according to recent thermal model analyses @xcite . in connection with updates of lattice - qcd results indicating a critical temperature up to @xmath27-@xmath28 mev @xcite ,
this , in particular , opens the possibility of a chemically equilibrated hot and dense hadronic phase for , say , @xmath29-@xmath30 mev , which we also consider .
we furthermore conduct a quantitative study of effective slope parameters of the na60 @xmath5 spectra , where the investigations of the fireball chemistry are supplemented with variations of the transverse flow velocity .
finally , we revisit the consequences of our fireball refinements on our previous evaluations @xcite of dielectron spectra as measured by ceres / na45 in semicentral pb - au collisions @xcite , as well as recent data in central pb - au @xcite .
our article is organized as follows : in sec .
[ sec_therm ] we recall the main ingredients of our approach to calculate thermal emission rates , based on the e.m .
spectral function in the vacuum ( sec .
[ ssec_rate ] ) , followed by discussing medium effects on hadronic emission at low mass ( due to in - medium @xmath1 , @xmath2 and @xmath3 spectral functions ; sec .
[ ssec_lowmass ] ) and at intermediate mass ( due to finite-@xmath31 chiral mixing ; sec . [ ssec_intmass ] ) ; a new element not included in previous spectral - function calculations are @xmath7-channel meson - exchange reactions which therefore are elaborated in more detail in sec . [ ssec_omt ] , while we will be brief on partonic emission from a quark - gluon plasma ( sec .
[ ssec_qgp ] ) . in sec .
[ sec_nontherm ] we evaluate nonthermal dilepton sources : @xmath1-meson decays at thermal freezeout ( whose decay kinematics differ from thermal radiation ; sec .
[ ssec_fo ] ) , an estimate of primordial @xmath1 s at large @xmath5 ( which do not thermalize ; sec . [ ssec_hard ] ) , and drell - yan annihilation ( with an extrapolation to low mass ; sec .
[ ssec_dy ] ) . in sec .
[ sec_fireball ] we recollect the ingredients to our thermal fireball model for the space - time evolution of the medium in heavy - ion collisions , including variations in hadrochemical freezeout and critical temperature . in sec .
[ sec_spectra ] we implement all dilepton sources into the fireball to compute dimuon invariant - mass ( sec . [ ssec_mspec ] ) and transverse - momentum spectra ( sec .
[ ssec_qtspec ] ) in comparison to na60 data ; the effect of hadrochemistry and @xmath32 on na60 spectra is worked out in sec .
[ ssec_hadrochem ] , followed by a slope analysis of @xmath5 spectra ( sec .
[ ssec_slope ] ) in the context of which we also investigate different radial flow scenarios ; the improvements in the fireball and dilepton source description as deduced from the na60 data are confronted with previous and new ceres / na45 data in sec .
[ ssec_ceres ] . sec .
[ sec_concl ] contains a summary and conclusions .
in thermal equilibrium the rate of dilepton emission per four - volume and four - momentum can be related to the hadronic e.m .
spectral function as @xcite @xmath33 which in this article we will refer to as thermal dileptons ( or thermal radiation ) .
the retarded e.m .
current - current correlator is given by @xmath34 } \right \rangle } } \ , \ ] ] where @xmath35=@xmath36=1/137 denotes the fine structure constant , @xmath37 the dilepton invariant mass squared with energy @xmath38 and three - momentum @xmath39 , and @xmath40 the thermal bose distribution function ( @xmath31 : temperature , @xmath41 : baryon chemical potential ) .
the final - state lepton phase space factor , @xmath42 depends on the lepton mass , @xmath43=0.511(105.6 ) mev for electrons ( muons ; @xmath44 ) , but quickly approaches one above threshold , @xmath45 ( _ e.g. _ , @xmath46 for dimuons ) . in the vacuum , the e.m .
spectral function , @xmath47 , is well known from @xmath48 annihilation into hadrons .
it is characterized by the light vector resonances @xmath1(770 ) , @xmath2(782 ) and @xmath3(1020 ) at low mass ( vector dominance model ( vdm ) ) and a perturbative quark - antiquark continuum at higher mass , @xmath49 ( @xmath50 : number of colors , @xmath51 : strong coupling constant , @xmath52 : the electric quark charge in units of the electron charge , and @xmath53 is running over up , down and strange quark flavors for @xmath54 gev ) .
@xmath55 gev signifies a `` duality '' scale above which the total cross section for @xmath56 ( and thus the strength of @xmath57 ) essentially behaves perturbatively with little impact from subsequent hadronization .
since the hadronic final state in @xmath48 annihilation approximately resembles a thermal medium ( except for strangeness ) , time - reversal invariance of strong and electromagnetic interactions implies that , to leading order in temperature , the equilibrium dilepton emission rate of hadronic matter is determined by the free e.m .
correlator , @xmath58 , cf . also refs .
@xcite . in the hadronic basis , the e.m .
spectral function is dominated by the isovector ( @xmath1 ) channel while the isoscalar channels are suppressed .
the su(3)-flavor quark model , _ e.g. _ , predicts a weighting of 9:1:2 for the e.m .
couplings of @xmath1 , @xmath2 and @xmath3 , respectively .
these values are roughly in line with the electromagnetic decay widths @xmath59 , @xmath60 and @xmath61 kev for @xmath1 , @xmath2 and @xmath3 , respectively ( the empirical values of the @xmath3 width should be corrected for phase space by the ratio @xmath62=0.77 ) .
the vector - isovector channel furthermore provides the most direct link to chiral symmetry : under chiral rotations , the @xmath1 channel transforms into the axialvector - isovector ( @xmath63 ) one .
thus , the @xmath63 is commonly identified as the chiral partner of the @xmath1 .
is identified with the pion . ]
the pertinent spectral functions in free space have been accurately measured in hadronic @xmath64 decays @xcite , cf .
[ fig_vavac ] , and evaluated in terms of chiral order parameters ( @xmath65 , four - quark condensates ) using weinberg sum rules @xcite . to leading order in temperature ( _ i.e. _ , for a hot pion gas ) ,
the mutual in - medium modifications of vector and axialvector channels can be inferred from chiral symmetry alone , as will be discussed in sec .
[ ssec_intmass ] below .
the main objective for the remainder of this section is a realistic evaluation of the in - medium e.m .
spectral function . in the lmr
, we assume that vector dominance remains valid in the medium .
our calculations of the @xmath66-meson spectral functions , @xmath67 , utilize effective hadronic lagrangians , where the interaction vertices and coupling constants are constrained by gauge and chiral symmetry , as well as empirical decay and scattering data .
medium effects are calculated within hadronic many - body theory _ without _ introducing explicit medium dependencies of the bare parameters in the lagrangian ( which require information beyond the effective hadronic theory ) .
a careful comparison of our predictions for dilepton spectra in heavy - ion collisions with experiment can then serve as a basis for identifying medium effects that go beyond the hadronic many - body framework . since the prevalent role in the lmr of the e.m .
correlator is played by the @xmath1 meson , the latter has been the main focus for studying in - medium effects in thermal dilepton production .
however , with the current precision of the na60 dilepton data on the @xmath15-@xmath16 level , thermal emission from @xmath2 and @xmath3 decays becomes relevant .
their contributions are an inevitable consequence of the formation of a thermal medium , and we therefore incorporate them along with pertinent medium modifications . within the vdm the in - medium e.m .
correlator in the lmr remains directly proportional to the vector - meson spectral functions as in eq .
( [ piem ] ) .
the key objective is then to calculate the vector - meson selfenergies at finite temperature and baryon density ( @xmath68 ) , @xmath69 , figuring into the propagator as @xmath70 here , @xmath71 denotes the bare mass , and the selfenergy contributions are classified into three types : ( i ) medium modifications of the pseudoscalar meson cloud , @xmath72 ( @xmath73 , @xmath74 or @xmath75 for @xmath76 , @xmath2 or @xmath3 , respectively ) , and direct interactions of @xmath66 with ( ii ) mesons and ( iii ) baryons from the surrounding heat bath , @xmath77 and @xmath78 .
one should also note that without the implementation of a phase transition into a realistic hadronic model ( which has not been achieved thus far ) , many - body calculations are to be considered as an extrapolation beyond a certain temperature and/or density .
we estimate that , under sps conditions , our approach should be reliable up to temperatures of about @xmath79 mev , where the total hadron density amounts to about @xmath80 ( with a net - baryon density of about @xmath81 ) , while it has increased to about @xmath82 at the expected phase boundary at @xmath83 mev . in the following we briefly recall the construction of , and constraints on , the underlying effective lagrangians . the first step in the calculation of a realistic @xmath1 propagator is a proper description of its coupling to the pion cloud in vacuum ( giving rise to its two - pion decay width , @xmath84 mev ) .
a suitable chiral effective lagrangian can be obtained by , _
e.g. _ , introducing the vector fields as local gauge fields into the chiral pion lagrangians and implementing the photon field via vector dominance @xcite , @xmath85 \\ & + \frac{e m_\rho^2}{g_\rho } a_\mu \rho_3^\mu \ , . \end{split}\ ] ] the pertinent two - pion and one - pion - tadpole diagrams specify the @xmath1 s pion cloud in the vacuum with a transverse selfenergy ( using a proper regularization procedure ) .
the coupling constant , @xmath86 , a cutoff parameter , @xmath87 ( to render the pion loops finite ) , and the bare @xmath1-mass , @xmath88 , can be readily adjusted to reproduce @xmath89-wave @xmath90-@xmath90 scattering phase shifts and the pion e.m .
formfactor in free space @xcite .
this model also predicts the two - pion part of the @xmath64-decay data well , cf .
[ fig_vavac ] . based on the above model in the vacuum ,
we adopt the in - medium @xmath1 propagator from the hadronic many - body approach developed in refs .
@xcite , to which we refer the reader for more details .
modifications of the pion cloud are induced by an in - medium @xmath90 propagator which is dressed with standard @xmath91 and @xmath92 excitations in nuclear matter , supplemented with appropriate vertex corrections to maintain a conserved vector current .
while the calculations of nuclear pion - cloud modifications are often restricted to the case of vanishing @xmath1 three - momentum ( `` back - to - back '' kinematics ) , we here employ the extension to finite three - momentum worked out in ref .
this is particularly important in that it enables ( i ) to constrain the corresponding medium effects by nuclear photo - absorption data @xcite and @xmath93 scattering , and ( ii ) a meaningful application to dilepton transverse momentum spectra . at finite temperature , where a ( large ) fraction of nucleons is thermally excited into baryons ,
only the nucleon density should be employed . to account for ( i ) @xmath94 excitations into higher resonances and ( ii ) @xmath95 excitations ( @xmath96 ) ,
an `` effective '' nucleon density @xmath97 has been used in the pion propagators @xcite ( we estimate the uncertainty implied by this prescription to be about @xmath9810% ) .
in addition , effects of antibaryons are added by multiplying @xmath99 with a factor @xmath100 where @xmath101 denotes the measured antiproton - to - proton ratio ( this is a small correction at sps energies but becomes essential at rhic @xcite ) .
the modifications of the pion cloud induced by a pion gas have been found to be rather small in various analyses @xcite .
a large part of the pion - gas effect is already provided by the bose enhancement factor , @xmath102 $ ] @xcite , while low - energy @xmath90-@xmath90 interactions are suppressed by chiral symmetry .
the former is included in the present calculations , but the latter are neglected ( being much weaker than baryon - induced effects ) .
direct interactions of the ( bare ) @xmath1 with baryons and mesons have been approximated by resonance interactions ( pion - exchange interactions , including spacelike pions , are implicit in the nuclear modifications of the pion - cloud part ) . for @xmath103 (
_ i.e. _ , @xmath104 excitations ) , this includes both @xmath105-wave scattering into negative - parity baryons and @xmath89-wave scattering into positive - parity baryons , with @xmath106 and @xmath107 , respectively .
an estimate of the involved couplings and cutoff parameters can be obtained by reproducing the empirical @xmath108 decay widths , but more reliable constraints follow from comprehensive scattering data . for the present model , in addition to hadronic decay branchings
, cold nuclear matter effects induced in both the pion cloud and via resonant @xmath1-@xmath23 interactions have been checked against total photoabsorption cross sections on the nucleon and nuclei @xcite , as well as @xmath93 scattering data @xcite . whereas @xmath109 and @xmath110 scattering relates to the low - density limit of the @xmath1-spectral function ,
@xmath111 data constrain @xmath112 close to nuclear saturation density , @xmath113 @xmath114 .
for the finite - temperature case , excitations on hyperons of the type @xmath115 are incorporated for @xmath116 and @xmath117 resonances with quantum numbers equivalent to the nonstrange sector and appreciable empirical decay widths into @xmath1 s and/or photons .
@xmath118 excitations on thermally excited nonstrange baryons ( _ e.g. _ , @xmath119 ) are included but turned out to be small . in the meson gas ,
direct @xmath1-@xmath24 interactions on thermal mesons ( @xmath120 , @xmath121 , @xmath1 ) have been scrutinized in ref .
@xcite , again based on @xmath122-channel resonance dominance , @xmath123 , and with pertinent lagrangians satisfying basic requirements of vector - current conservation and chiral symmetry . as in the case of excited baryons , the couplings and cutoffs have been estimated from combined hadronic and radiative decay widths .
the most important contributions to the @xmath1 propagator arise from axialvector mesons , @xmath124 ( chiral partner of the @xmath1 ) , @xmath125 and @xmath126 , as well as the @xmath127 .
another valuable consistency check for the in - medium properties of hadrons are qcd sum rules .
the latter relate a dispersion integral over a spectral function to an operator product expansion ( ope ) for the correlation function at spacelike four - momenta , @xmath128 .
the nonperturbative coefficients of the ope involve quark and gluon condensates whose in - medium modifications reflect upon ( an energy integral over ) the spectral function .
generically , a reduction of the condensates leads to a softening of the spectral function , _
i.e. _ , a shift of strength to lower mass . however , as has been elaborated in ref .
@xcite for the case of cold nuclear matter , the additional low - mass strength in the in - medium @xmath1 spectral function can be due to both a reduced mass or a broadening of the width , or a combination thereof .
the pertinent `` allowed band '' for width and mass values at normal nuclear matter density has been quantified in ref .
@xcite . for the @xmath1 spectral
function employed here @xcite , one finds a mass and width of @xmath129 mev at nuclear saturation density and vanishing three - momentum .
these values turn out to be consistent with the band derived in ref .
@xcite ( cf . middle panel of fig . 2 in there ) , but more extensive ( _ i.e. _ , different densities and temperatures ) and quantitative comparisons are certainly desirable ( _ e.g. _ , the low - mass strength ( shoulder ) in the @xmath1 spectral function which is induced by `` pisobar '' ( @xmath130 ) and `` rhosobar '' ( @xmath104 ) excitations , can not be represented in breit - wigner form as underlying ref .
however , at this point , there is no indication for changes in the bare parameters in the lagrangian , _
i.e. _ , at nuclear saturation density hadronic many - body effects saturate the reduction of quark and gluon condensates as required by qcd sum rules .
( upper panel ) , @xmath2 ( middle panel ) and @xmath3 ( lower panel ) under conditions representative for the time evolution of heavy - ion collisions at full sps energy : @xmath131 } = ( 232,0,0,175)$ ] , @xmath132 , @xmath133 and vacuum for the dotted , short - dashed , long - dashed and solid lines , respectively.,scaledwidth=45.0% ] ( upper panel ) , @xmath2 ( middle panel ) and @xmath3 ( lower panel ) under conditions representative for the time evolution of heavy - ion collisions at full sps energy : @xmath131 } = ( 232,0,0,175)$ ] , @xmath132 , @xmath133 and vacuum for the dotted , short - dashed , long - dashed and solid lines , respectively.,scaledwidth=45.0% ] ( upper panel ) , @xmath2 ( middle panel ) and @xmath3 ( lower panel ) under conditions representative for the time evolution of heavy - ion collisions at full sps energy : @xmath131 } = ( 232,0,0,175)$ ] , @xmath132 , @xmath133 and vacuum for the dotted , short - dashed , long - dashed and solid lines , respectively.,scaledwidth=45.0% ] the @xmath1-spectral functions under fireball conditions representing the evolution of the hadronic phase at full sps energy ( @xmath134 agev ) are displayed in the upper panel of fig .
[ fig_avec ] .
the resonance strongly broadens with little , if any , mass shift , essentially leading to its `` melting '' close to the expected phase boundary .
these features are a rather generic consequence of a multilevel excitation spectrum in which real parts occur with altering signs while imaginary parts are negative definite and strictly add up .
also note the development of a pronounced low - energy strength around @xmath135 gev which leads to a very large enhancement over the vacuum spectral function ( and will be augmented by thermal bose factors in the dilepton rate ) . at lower temperatures ,
the presence of meson - chemical potentials ( for @xmath90 , @xmath121 , @xmath136 ) implies higher meson and baryon densities than in chemical equilibrium at a given temperature , entailing stronger medium effects .
the build - up of the chemical potentials ( necessary to preserve the experimentally observed hadron ratios throughout the hadronic cooling ) therefore facilitates the observation of in - medium effects in the dilepton spectra .
a strong broadening of the @xmath1 in hot and dense matter has also been reported in other calculations at finite @xmath31 and @xmath68 @xcite .
the in - medium @xmath1-spectral function has been rather extensively employed to calculate dilepton spectra in heavy - ion collisions . at full sps energy
it has been implemented into fireball models @xcite , transport simulations @xcite and hydrodynamic evolutions @xcite , with fair success ( and consistency ) in describing ceres / na45 data in central and semicentral pb - au collisions @xcite . when averaged over a typical space - time evolution , the in - medium @xmath1 width at sps amounts to @xmath137-@xmath138 mev , almost three times its vacuum value , implying @xmath139 when extrapolated to the expected qcd - phase boundary . despite an experimental pion - to - baryon ratio of @xmath140 ,
medium effects induced by baryons have been identified as essential at sps energies early on @xcite .
this has led to the prediction of an even larger excess at lower beam energies , which was tentatively confirmed in a @xmath141 agev pb+au run of ceres / na45 @xcite ( albeit with rather large errors ) .
the lightlike limit of the @xmath1-spectral function has been applied to calculate thermal photon spectra @xcite , supplemented with meson - exchange reactions which are not included in @xmath142 ( these become significant at momenta above @xmath143 gev and will be discussed in sec .
[ ssec_omt ] below ) .
after convolution over the same fireball model as used for dileptons , the direct - photon excess observed by wa98 in central pb - pb reactions at sps @xcite , as well as by phenix in central au - au collisions at rhic @xcite , can be reasonably well accounted for ( with additional primordial and qgp photon sources , the latter being subdominant at sps ) .
the in - medium @xmath1-spectral function following from hadronic many - body theory @xcite has thus passed a rather wide range of both theoretical and phenomenological tests , in both cold nuclear as well as hot hadronic matter .
medium effects on the @xmath2 have received less attention so far , especially in the context of ultrarelativistic heavy - ion collisions ( urhics ) where its dilepton decays are dominated by free decays after thermal freezeout ( commonly referred to as a `` cocktail '' contribution ) .
we here adopt the same approach as for the @xmath1 , following ref . @xcite . in the vacuum ,
the hadronic @xmath2 width , @xmath144 , is composed to approximately equal parts of a direct @xmath145 and a @xmath146 coupling , represented by anomalous terms in the interaction lagrangian , @xmath147 with a hadronic formfactor cutoff - parameter , @xmath148 gev , and using the vdm @xcite , one finds for the radiative decay width @xmath149 mev which approximately agrees with the updated experimental value of @xmath150 mev @xcite . in cold nuclear matter
, the renormalization of the @xmath90-@xmath1 cloud has been found to be rather sensitive to the medium effects on the @xmath1 @xcite , which is due to the fact that the @xmath151 decay proceeds via the low - mass tail of the @xmath1-spectral function below its nominal mass of @xmath152 mev . in ref .
@xcite @xmath153% of the estimated in - medium broadening at nuclear saturation density is attributed to the in - medium @xmath1 spectral function ( taken from ref .
@xcite ) , resulting in @xmath154 mev . likewise
, a large contribution to second order in the nuclear density , @xmath155 , to @xmath156 arises due to the simultaneous dressing of both @xmath90 and @xmath1 @xcite .
following ref .
@xcite , we restrict ourselves to the impact of the in - medium @xmath1 by folding the phase space in the @xmath157-decay width over the spectral function evaluated in the previous section , @xmath158 \ , f_{\omega\rho\pi}(q_{\mathrm{cm}})^2 \ , , \end{split}\ ] ] where @xmath159@xmath160@xmath161 , @xmath162 , @xmath163 , and @xmath164 is the three - momentum of @xmath90 and @xmath1 in the rest frame of the @xmath2 .
in addition we account for cold - nuclear matter effects due to the two most important direct @xmath2-@xmath23 excitations , @xmath165 and @xmath166 , as identified in ref .
@xcite in a coupled - channel analysis of empirical vector - meson nucleon scattering amplitudes .
recent measurements of @xmath167-invariant mass spectra in photon - induced production off hydrogen and nb targets have revealed a low - mass shoulder @xcite that has been interpreted as reduced @xmath2 mass in nuclear matter .
however , this interpretation is not free of controversy and may also be explained by a large increase of the @xmath2 width to about @xmath168 mev @xcite .
the latter is compatible with the @xmath2 propagator employed in the present paper .
finite - temperature effects on the @xmath2 propagator include bose - enhancement factors in the @xmath90-@xmath1 cloud , the in - medium @xmath1 propagator in eq .
( [ gamorp ] ) , a schematic estimate of the inelastic @xmath169 width based on ref .
@xcite , and a full calculation of the thermal @xmath170 loop @xcite .
the predicted average @xmath2 width in the hadronic phase of urhics is @xmath171@xmath172@xmath173 mev @xcite , cf .
middle panel of fig .
[ fig_avec ] , quite consistent with the results of ref .
this has the interesting consequence @xcite that if dilepton spectra in hics show a structure of @xmath11@xmath173 mev width in the @xmath1-/@xmath2 mass region , it would be an unambiguous signature of the in - medium @xmath2 spectral function since it is narrower than the lower limit given by a free @xmath1 contribution , @xmath174 mev . for the @xmath3 , collision rates in a meson gas
have been estimated to amount to a broadening by @xmath175 mev at @xmath176 mev @xcite .
the dressing of the kaon cloud is presumably the main effect for @xmath3 modifications in nuclear matter , increasing its width by @xmath177 mev at saturation density @xcite .
recent data on @xmath3 absorption in nuclear photoproduction suggest substantially larger values @xcite .
since a quantitative , empirically constrained treatment of the @xmath3 in hot and dense matter is not available at present , we will consider the @xmath3 width as a parameter .
following ref .
@xcite we augment the microscopic calculations of ref .
@xcite by a factor of @xmath178 to roughly agree with the phenomenological values extracted from nuclear photoproduction @xcite .
the corresponding @xmath3 spectral functions under sps conditions are displayed in the lower panel of fig .
[ fig_avec ] .
the average width over the fireball evolution amounts to @xmath179 . above the @xmath3 mass , the hadronic structure of the ( vacuum ) e.m .
correlator becomes more involved being characterized by overlapping broad resonances which combine into continuum - like multi - meson states ( most notably isovector four - pion states ) , recall fig .
[ fig_vavac ] . to estimate medium effects in this regime , we take recourse to model - independent predictions following from a low - temperature expansion evaluated using chiral - reduction techniques for the corresponding matrix elements involving thermal pions .
this method has been first worked out in ref .
@xcite in the chiral limit ( @xmath180 ) and leads to the chiral mixing formula for the isovector part of vector and axialvector correlators , @xmath181 the mixing coefficient @xmath182 is determined by pion tadpole diagrams via a loop integral @xmath183 ( @xmath184 : on - shell pion energy , @xmath185 mev : pion decay constant ) . for @xmath186 , @xmath66 and @xmath187 correlators
degenerate signaling chiral - symmetry restoration ( at @xmath188@xmath172160 mev ) .
note that the admixture of the axialvector part in the @xmath189-@xmath190 gev region ( which corresponds to the @xmath63 resonance ) fills in the `` dip '' structure of the vector correlator ( recall fig .
[ fig_vavac ] ) ; for @xmath186 both spectral functions merge into the perturbative - qcd continuum for @xmath191 gev which has been interpreted @xcite as a lowering of the quark - hadron duality scale , @xmath192 , cf .
( [ piem ] ) .
the simple form of eq .
( [ mix ] ) only holds in the soft - pion limit , _
i.e. _ , when neglecting the pion four - momenta in @xmath193 . more elaborate treatments @xcite will broaden and somewhat reduce the enhancement in the @xmath194-@xmath190 gev region @xcite .
we here implement the mixing effect on the ( isovector ) four - pion part utilizing eq .
( [ mix ] ) with a mixing parameter @xmath195 with finite pion mass ( @xmath196 mev ) , critical temperature , @xmath32 , and pion chemical potentials ( @xmath197 ) implemented as an overall fugacity factor , @xmath198 , in boltzmann approximation ( @xmath199 varies according to the thermal fireball evolution discussed in sec .
[ sec_fireball ] below ) . the two - pion piece , as well as the three - pion piece corresponding to @xmath63 decay , @xmath200 , have been removed as they are included via the @xmath1-spectral function as discussed above . a detailed analysis of the derivation of eq .
( [ mix ] ) in the presence of a finite pion - chemical potential leads to the following expression for the mixing effect on the vector - isovector current correlation function : @xmath201 as discussed in the introduction , a current discrepancy in the theoretical description of the na60 data concerns the dilepton yield at high transverse pair - momentum , @xmath21 gev , for masses @xmath202 gev .
it is therefore important to scrutinize the vector - meson spectral function calculations for high momentum sources .
the latter are typically associated with @xmath7-channel meson exchanges , rather than @xmath122-channel resonances whose decay products are suppressed at momenta far off the resonance . in the spectral - function approach ,
@xmath7-channel exchange processes are encoded in medium modifications of the pseudoscalar meson cloud . as elaborated in sec .
[ sssec_rho ] , the pion - cloud modifications of the @xmath1 explicitly include interactions with baryons , but not with thermal mesons .
consequently , @xmath7-channel meson exchanges are present for @xmath203 interactions , but not for @xmath204 .
e.g. , the experimental @xmath205 cross section is properly reproduced at large @xmath10 where @xmath7-channel @xmath90 exchange dominates .
indeed , it has been found in ref .
@xcite in the context of thermal photon production that @xmath7-channel exchanges in @xmath206 outshine the contributions from the @xmath1-spectral function at momenta above @xmath11@xmath143 gev for @xmath207 mev ( at smaller temperatures the importance of @xmath7-channel exchanges is reduced ) .
most notably , @xmath2 exchange emerged as the single most important hadronic high-@xmath5 thermal photon source .
meson via @xmath2-meson exchange in the @xmath7 channel.,scaledwidth=30.0% ] to estimate the relevance of @xmath7-channel exchanges in @xmath208 scattering for thermal dileptons , we adopt the same simplified treatment as in ref .
@xcite , properly extended to a finite invariant mass of the virtual photon ( lepton pair ) : rather than implementing the pertinent diagram into the @xmath1 selfenergy , we employ the standard kinetic theory expression for the thermal production rate @xmath209 \frac{z_{\pi}^3 f^b(\omega^\rho_{p_1};t ) f^b(\omega^\pi_{p_2};t ) } { 4 ( 2 \pi)^6 } \ , \end{split}\ ] ] which allows to calculate the contribution for @xmath210 in terms of the corresponding matrix element , @xmath211 , in born approximation .
the latter substantially facilitates the task of maintaining gauge invariance . for the @xmath212 vertex , the anomalous - coupling lagrangian , eq .
( [ rho - om - pi ] ) , is used to calculate the matrix element , @xmath213 \\ & \quad + t^2 [ ( m^2 + 2 m_{\pi}^2+m_{\rho}^2)^2 - 2(m^2 + 2 m_{\pi}^2+m_{\rho}^2)s+2s^2 ] -2 t^3 ( m^2 + 2 m_{\pi}^2+m_{\rho}^2-s ) + t^4 \ } \ , \end{split}\ ] ] corresponding to the feynman diagram depicted in fig .
[ fig - om - t ] ; @xmath24 denotes the invariant mass of the lepton pair ; @xmath214 and @xmath215 are the usual mandelstam variables with @xmath216 the exchanged four - momentum .
the @xmath1-@xmath217 vertex has been written in the form @xcite @xmath218 where @xmath219 .
medium effects on the @xmath1 meson are implemented by using the full in - medium propagator , @xmath220 , for the intermediate @xmath1 meson which couples to the photon ( upper @xmath1 line in fig .
[ fig - om - t ] , with @xmath221 ) and by folding eq .
( [ msqr ] ) over the in - medium spectral function for the incoming @xmath1 ( lower @xmath1 line in fig .
[ fig - om - t ] , with @xmath222 and a weight @xmath223 ) . to avoid double counting with the @xmath2 @xmath122-channel graph already included in the in - medium @xmath1-selfenergy , the integral in eq .
( [ rate - om - t ] ) is restricted to spacelike @xmath2-exchange momenta , _
i.e. _ , @xmath224 .
the coupling constant , @xmath225 , has been fixed in @xcite by a simultaneous fit to the hadronic and radiative @xmath2 decays including a hadronic dipole - form factor , @xmath226 with @xmath227 gev . to maintain gauge invariance in a simplified way as in ref .
@xcite , we pull the formfactor outside the integral and introduce an average momentum transfer , @xmath228 ( adapted to the finite mass of the virtual photon ) , according to @xmath229}{[m_{\omega}^2-t(x)]^2 } \ , , \\ t(x)&= m^2+m_{\pi}^2 - 2 ( m^2+q^2-q p x ) \ , .\end{aligned}\ ] ] the amplitude , eq .
( [ msqr ] ) , is then multiplied by @xmath230 .
( [ rate - om - t ] ) furthermore accounts for pion fugacity factors , @xmath231 due to finite pion chemical potentials , implemented in boltzmann approximation ( _ e.g. _ , for the incoming @xmath90 and @xmath1 this amounts to a total fugacity of @xmath232 ) .
@xmath7-channel exchange in the @xmath233 reaction .
upper panel : comparison of the rates when using either the free or the in - medium @xmath1 propagators in both the matrix element and the incoming @xmath1-mass distribution .
middle panel : comparison of the in - medium @xmath2 @xmath7-channel rates with thermal rates from the full in - medium @xmath1 spectral function .
lower panel : same as middle panel , but restricting the momentum integration to @xmath234 gev .
all curves for @xmath31=150 mev and 120 mev include pertinent pion fugacities at @xmath235 mev and 79 mev , respectively.,scaledwidth=45.0% ] @xmath7-channel exchange in the @xmath233 reaction .
upper panel : comparison of the rates when using either the free or the in - medium @xmath1 propagators in both the matrix element and the incoming @xmath1-mass distribution .
middle panel : comparison of the in - medium @xmath2 @xmath7-channel rates with thermal rates from the full in - medium @xmath1 spectral function .
lower panel : same as middle panel , but restricting the momentum integration to @xmath234 gev .
all curves for @xmath31=150 mev and 120 mev include pertinent pion fugacities at @xmath235 mev and 79 mev , respectively.,scaledwidth=45.0% ] @xmath7-channel exchange in the @xmath233 reaction .
upper panel : comparison of the rates when using either the free or the in - medium @xmath1 propagators in both the matrix element and the incoming @xmath1-mass distribution .
middle panel : comparison of the in - medium @xmath2 @xmath7-channel rates with thermal rates from the full in - medium @xmath1 spectral function .
lower panel : same as middle panel , but restricting the momentum integration to @xmath234 gev .
all curves for @xmath31=150 mev and 120 mev include pertinent pion fugacities at @xmath235 mev and 79 mev , respectively.,scaledwidth=45.0% ] in fig .
[ fig_drdm - om - t ] we summarize our results for the @xmath2 @xmath7-channel emission rate , integrated over three - momentum , for conditions along our default trajectory for in(158 agev)+in collisions at sps ( including finite @xmath199 ) .
the upper panel illustrates that the implementation of the in - medium @xmath1 propagator ( which was not done in ref .
@xcite ) leads to a notable reduction in the emission rate , which is mostly due to the reduction in the intermediate @xmath1 propagator , @xmath236 .
we have verified that medium effects on the @xmath2 propagator are negligible ( in the spacelike regime finite - width effects are insignificant as long as @xmath237 ) . in the middle panel , the comparison of the in - medium @xmath2 @xmath7-channel rates to the ones from the full in - medium @xmath1 spectral function @xcite , based on eq .
( [ rate ] ) , confirms that the former contribution is indeed small in magnitude .
however , if one applies a three - momentum cut of @xmath234 gev ( lower panel ) , the relative magnitude of the @xmath7-channel contribution increases , as anticipated at the beginning of this section .
but even in this case the relative strength of the @xmath2 @xmath7-channel exchange is rather moderate . the increasing contribution at higher momentum can be more directly seen when integrating the rate , eq .
( [ rate - om - t ] ) , over invariant mass bins and plotting it versus @xmath39 , cf .
[ fig_drdq - om - t ] .
@xmath7-channel exchange in @xmath233 ( computed with in - medium @xmath1 propagators ) and the leading contribution from the full in - medium-@xmath1 spectral function . as in fig .
[ fig_drdm - om - t ] , the results for @xmath31=150 mev include pion fugacity factors.,scaledwidth=45.0% ] @xmath7-channel exchange in @xmath233 ( computed with in - medium @xmath1 propagators ) and the leading contribution from the full in - medium-@xmath1 spectral function . as in fig .
[ fig_drdm - om - t ] , the results for @xmath31=150 mev include pion fugacity factors.,scaledwidth=45.0% ] as indicated above , in our calculations of dilepton spectra in sec .
[ sec_spectra ] , we will implement the @xmath2 @xmath7-channel emission rate , eq .
( [ rate - om - t ] ) , by an incoherent addition to the main contribution , eq .
( [ rate ] ) .
this neglects interference terms in the pertinent selfenergy contribution in the ( denominator of the ) @xmath1 spectral function , which is justified due to the relative smallness of the @xmath2 @xmath7-channel part . to simulate the presence of other @xmath7-channel processes ( _
e.g. _ , @xmath90 and @xmath63 exchange ) , guided by the photon rate calculations of ref .
@xcite , we will multiply the @xmath2 @xmath7-channel contribution by a factor of 2 in our calculation of dilepton spectra below .
emission from the qgp is calculated using the hard - thermal - loop improved rate for @xmath238 annihilation @xcite , including an extrapolation to finite three - momentum .
it turns out @xcite that this rate has the conceptually attractive feature that it closely coincides with the rate in hadronic matter when both are extrapolated to the expected phase - transition region .
this is suggestive for a kind of quark - hadron duality @xcite and has the additional benefit that the emission from the expanding fireball can be anticipated to become rather insensitive to details of how the phase transition is implemented ( _ e.g. _ , to the values for the critical temperature or `` latent heat '' ) .
this point will be studied explicitly in sec .
[ ssec_hadrochem ] below .
recent ( experimental and theoretical ) developments suggest that the qgp features significant nonperturbative effects for temperatures up to @xmath11@xmath190-@xmath239 , _
e.g. _ , the survival of ( hadronic ) resonances and/or bound states , even in the light - quark sector .
possible consequences for the dilepton emission rate have been estimated in ref .
@xcite . depending on the width of these states ,
a maximal enhancement of up to a factor of @xmath240 over the perturbative qgp emission rate at intermediate masses ( _ i.e. _ , in the mass range of @xmath190-@xmath241 gev ) is conceivable .
we will not further pursue this possibility in the present paper .
in this section we address dilepton sources other than radiation from a thermal source as given by eq .
( [ rate ] ) , _ i.e. _ , @xmath1 decays after thermal freezeout , decays of primordially produced @xmath1 mesons which escape the fireball , and primordial drell - yan annihilation . for simplicity , we generically refer to these sources as `` non - thermal '' , even though the freezeout decays , _
e.g. _ , are represented by thermal `` blast - wave '' spectra commonly used to characterize light hadron spectra in heavy - ion collisions .
dilepton decays of long - lived hadrons ( most notably dalitz decays of @xmath136 and @xmath2 , as well as exclusive @xmath242 decays of @xmath2 and @xmath3 ) can be rather well separated from the freezeout of the interacting fireball in heavy - ion collisions , which has given rise to the notion of the `` hadronic decay cocktail '' contribution to dilepton spectra , computed using the free spectral shape of the decaying mesons . for the @xmath1 meson , due to its short lifetime , such a separation is not well defined .
therefore , when adding thermal radiation to the cocktail , the @xmath1 is commonly removed from the latter and implemented into the thermal yield . in previous works
@xcite this was done by running the fireball an extra @xmath11@xmath143 fm/@xmath12 using the in - medium @xmath1 spectral function .
it turns out @xcite , however , that this description of @xmath1 decays at thermal freezeout carries an extra factor of @xmath243 relative to a standard blast - wave spectrum of hadrons at thermal freezeout , where @xmath244 is the usual lorentz factor . roughly speaking ,
the in - medium radiation given by eq .
( [ rate ] ) is proportional to the electromagnetic decay width times the fireball lifetime , @xmath245 , while decays after freezeout are proportional to the branching ratio of electromagnetic to total lifetime , @xmath246 , where the @xmath217 factor cancels ( as usual , we define @xmath247 as the partial decay width in the rest system of the particle ) .
the remainder of this section will give a more detailed discussion of this . to account for the correct time dilation effects in the calculation of dilepton decays of @xmath1-mesons after thermal freezeout
, we use the standard cooper - frye description @xcite , @xmath248 for the phase - space distribution of an on - shell particle ( @xmath249@xmath160@xmath250 ) at thermal freezeout ; @xmath31 and @xmath251 denote the local temperature and four velocity ( flow ) of the fluid cells of the medium , and @xmath252 is the hypersurface normal vector defined by an appropriate freezeout condition .
in accordance with the homogeneous fireball model described in sec .
[ sec_fireball ] , thermal freezeout at a constant time @xmath253 in the laboratory frame is assumed , _
i.e. _ , @xmath254 .
the in - medium @xmath1 spectral function at freezeout is introduced via the substitution @xmath255 where @xmath256 ( as before ) includes the average over the three polarizations via the lorentz trace which runs over the physical ( _ i.e. _ , four - momentum transverse ) components of the propagator . to properly treat the low - mass tails of the spectral function
, one would have to resolve the individual resonance decays figuring into the @xmath1 selfenergy . to simplify our task
, we circumvent this problem by employing the vacuum form of the @xmath1 selfenergy augmented with a width corresponding to the full - width - half - maximum of the in - medium spectral function at thermal freezeout ( @xmath257 mev ) .
the distribution of @xmath1 mesons at thermal freezeout then reads @xmath258 in accordance with the averaging procedure over the freezeout duration , we evaluate the transverse four - velocity for the freezeout @xmath1 at a time @xmath259 fm / c after the thermal emission , eq.([rate ] ) , has shut off .
the dilepton rate follows by folding with the appropriate partial decay width . within our vector - meson dominance model
this is given by the matrix element for the process @xmath260 , and after integration over @xmath261 this results in @xmath262 where @xmath263 is the dilepton - phase space factor ( [ lept - ps ] ) .
the second line of eq .
( [ fo - dilep ] ) shows that the momentum dependence of the dilepton distribution from @xmath1 decays after thermal freeze - out deviates from the rate from a thermal source ( [ rate ] ) by a lorentz factor @xmath264 .
the physical origin of this difference is the time dilation of the total lifetime of a freeze - out @xmath1 meson with three - momentum @xmath39 which is absent in the formula for radiation from a thermal source , because its @xmath1-meson abundance at each instant of time is fixed by the temperature and pion - chemical potential of the medium ( as required by detailed balance of @xmath1 formation and decay ) , and thus the total number of thermal dileptons is determined by the lifetime of the fireball .
note , however , that the thermal rate is proportional to @xmath265 with an associated time dilation factor @xmath243 .
the freezeout-@xmath1 dilepton spectra are thus equivalent to standard blast - wave descriptions of stable hadrons @xcite .
our default assumption for the radial profile of the flow field is a linear dependence on the radius according to @xmath266 where @xmath267 is the radius of the fire - cylinder , @xmath268 its surface speed , and @xmath269 the radial coordinate of the fluid cell related to the volume element @xmath270 ( since we neither address azimuthal asymmetries nor peripheral collisions in this work , we approximate the cross sectional area of the fire cylinder as a circle ) .
another source of non - thermal dileptons is the decay of @xmath1 mesons which originate from primordial hard - scattering processes and traverse the interaction zone without equilibrating .
we evaluate this contribution within a schematic jet - quenching model @xcite , as follows .
first , we construct the @xmath5 spectrum of primordial @xmath1 mesons assuming a power law , @xmath271 with parameters estimated from @xmath272-@xmath272 scattering data @xcite , @xmath273 and @xmath274 . the total number of primordial @xmath1 s in @xmath187-@xmath187 collision is determined based on the empirical freezeout systematics of light hadron production , _ i.e. _ , we calculate the expected total number of @xmath1 mesons in in - in at @xmath275 mev and match the norm of the primordial spectrum , eq .
( [ primord ] ) , to it . at the same time
, we have to take care of the correct scaling properties of the spectrum at high momentum : while the ( total ) yield at low @xmath5 is proportional to the number of participant nucleons , the high-@xmath5 yield should scale with the number of primordial @xmath23-@xmath23 collisions .
we implement this transition by a continuous linear switching between the two regimes over the range @xmath276 .
second , we implement a cronin effect for a - a collisions by a `` gaussian smearing '' of eq .
( [ primord ] ) , @xmath277,\ ] ] with a conservative estimate of @xmath278 gev@xmath279 as extracted in ref .
@xcite based on direct photon spectra in @xmath272-@xmath187 reactions .
( upper panel ) and @xmath90 mesons ( lower panel ) from primordial ( hard ) production plus thermal freezeout , compared to a compilation of data for @xmath280 of pions at full sps energy @xcite for central collisions with the standard equation of state ( eos - a , full line ) and an eos with @xmath281 ( eos - b+c as discussed in sec . [ sec_fireball ] , dashed line ) .
the primordial @xmath1 spectra include effects of initial cronin smearing and jet - quenching in the expanding fireball , while the thermal freezeout @xmath1 s are taken from sec .
[ ssec_fo ] ( same for the pions , but without finite - width effects).,scaledwidth=45.0% ] ( upper panel ) and @xmath90 mesons ( lower panel ) from primordial ( hard ) production plus thermal freezeout , compared to a compilation of data for @xmath280 of pions at full sps energy @xcite for central collisions with the standard equation of state ( eos - a , full line ) and an eos with @xmath281 ( eos - b+c as discussed in sec . [ sec_fireball ] , dashed line ) .
the primordial @xmath1 spectra include effects of initial cronin smearing and jet - quenching in the expanding fireball , while the thermal freezeout @xmath1 s are taken from sec .
[ ssec_fo ] ( same for the pions , but without finite - width effects).,scaledwidth=45.0% ] third , we calculate a suppression factor representing the probability for primordial @xmath1 mesons to escape the medium without rescattering , using monte carlo techniques . in line with our fireball model , we start from a spatially homogeneous distribution in the transverse plane at the qgp formation time .
the escape probability for a @xmath1 with momentum @xmath39 is then calculated as @xmath282 where the absorption cross section @xmath283 depends on the @xmath1-meson formation time in its rest frame , which is assumed to be @xmath284@xmath160@xmath143 fm , augmented by lorentz time dilation in the fireball frame ; @xmath285 and @xmath286 denote the pre - hadronic and hadronic absorption cross sections of the @xmath1 , respectively , and @xmath287 the corresponding particle density of the fireball medium ( partonic or hadronic ; if the @xmath1 meson has not formed by the time the system has hadronized , we use constituent quark scaling to infer the partonic density as @xmath288 , and vice versa ) .
a rough check of our construction may be obtained by comparing the @xmath1 spectra to measured pion spectra in @xmath187-@xmath187 collisions at the sps .
since pion spectra contain both a hard ( primordial ) and a soft ( thermal + flow ) component , a comparison with our @xmath1 spectra should contain both the primordial part constructed in this section and the thermal freezeout contribution described in the previous section , [ ssec_fo ] .
we do this comparison in terms of the nuclear modification factor , @xmath280 , defined as the ratio of the spectrum in @xmath187-@xmath187 collisions over the collision - number scaled spectrum in @xmath272-@xmath272 .
the upper panel of fig .
[ fig_raa ] shows that the jet - quenching + freezeout-@xmath1 model results in fair agreement with a recent compilation of @xmath280 for pions in s - au and pb - pb systems from various experiments at full sps energy @xcite . as a check of this procedure , as well as of the fireball model , we plot in the lower panel of fig .
[ fig_raa ] the calculated pion @xmath280 for central in - in at sps ( as a sum of the jet - quenching plus thermal - freezeout components ) for central in - in at sps ; the agreement with data is reasonable ( note that resonance decays are not included , which are mostly concentrated at low @xmath5 ) . in the high - mass region ( hmr , @xmath289 gev ) , dilepton spectra in nuclear collisions
are expected to be dominated by the drell - yan ( dy ) process , _ i.e. _ , primordial annihilation of quarks and antiquarks within the incoming nucleons . to leading order ,
this process does not depend on the strong coupling constant ( it is purely electromagnetic , @xmath290 ) and can therefore be rather reliably evaluated , provided one has a good knowledge of the parton - distribution functions ( pdfs ) within the nucleon . in a central collision ( @xmath291 ) of two equal nuclei with mass number @xmath187 , the invariant - mass spectrum of drell - yan pairs per unit rapidity
is given by @xmath292 in terms of the standard dy cross section in an elementary nucleon - nucleon ( @xmath23-@xmath23 ) collision , @xmath293 \ .
\label{dynn}\ ] ] here , @xmath294 and @xmath295 denote the ( collinear ) quark and anti - quark distribution functions , respectively ( neglecting nuclear effects ) .
their arguments are related to the center - of - mass ( cms ) rapidity , @xmath296 , and the invariant mass of the lepton pair as @xmath297 with @xmath298 , where @xmath10 is the cms energy of the @xmath23-@xmath23 collision .
the root - mean - squared radius parameter in eq .
( [ dyaa ] ) , @xmath299 fm , arises from a folding over a gaussian thickness function ; for simplicity , we will adopt eq .
( [ dyaa ] ) also for non - central collisions with an accordingly reduced number of participants , @xmath187 . for the parton distribution functions we employ the leading - order parameterization from ref .
@xcite ( grv94lo ) , which incorporate isospin asymmetries in the sea - quark distributions which significantly reduce the need for additional isospin corrections for nuclei with @xmath300 ( _ e.g. _ , less than @xmath301 for pb - pb collisions ) , which will be neglected here @xcite . in eq .
( [ dynn ] ) higher - order corrections in @xmath51 are encoded in an empirical @xmath121 factor , which turns out to be @xmath302 to reproduce dy production in @xmath272-@xmath187 collisions @xcite .
in addition to generating a @xmath121 factor , higher - order effects manifest themselves in a nonzero pair-@xmath5 of the dy dileptons . to obtain a realistic @xmath5 spectrum ,
we follow the procedure adopted by the na50 collaboration @xcite : based on a comprehensive analysis of @xmath272-@xmath187 and @xmath187-@xmath187 collisions at the sps it has been found that the imr and hmr dilepton spectra can be fairly well described by assuming a gaussian shape for the dy spectrum , @xmath303 with @xmath304 gev . in ref .
@xcite , a value of @xmath305 gev has resulted in reasonable agreement with the na50 @xmath5 spectra for imr dimuons in central pb - pb ; for @xmath306 gev , which increases the spectrum at @xmath307gev by 50% , the agreement with na50 would still be acceptable .
here , we adopt @xmath305 gev providing a conservative estimate of the dy contribution .
the extrapolation of the dy spectrum to the regions where both mass and momentum are small ( say , below @xmath308 gev ) is problematic , but fortunately its contribution in this regime is small compared to , _
e.g. _ , thermal emission in @xmath187-@xmath187 collisions .
there is , however , an additional constraint provided by the photon point ( @xmath309 ) which allows to extrapolate the dy spectra to low mass , at least for reasonably large momenta , @xmath21 gev ( which is the regime were the dy yield is noticeable ) .
formally , the photon - production rate follows from the dilepton one by taking the limit @xmath310 .
more specifically , this is encoded in the relation @xmath311 with the same e.m
. current - current correlation function , @xmath312 of eq .
( [ ret - se ] ) , as in the dilepton rate , eq .
( [ rate ] ) . for @xmath313 ,
the @xmath24-dependence of @xmath314 is weak , and the difference between @xmath315 and @xmath316 amounts to a factor of @xmath317 .
thus we can evaluate the dy @xmath5 spectrum at a mass @xmath318 and extrapolate it down in mass to the photon point using the factor @xmath319 . for @xmath318=0.8 - 1
gev , in connection with @xmath305 gev , reasonable agreement with the primordial photon spectrum of ref .
@xcite is found .
in addition , there could be `` pre - equilibrium '' contributions from secondary drell - yan processes @xcite ( _ e.g. _ , @xmath320 involving primordially produced pions ) , which turn out to be rather sensitive to the pion formation time .
_ e.g. _ , for @xmath321 fm/@xmath12 , the enhancement over primordial drell - yan annihilation in central s - u was found to be @xmath11@xmath322 at @xmath323 gev . in a thermal emission description , which we employ here after a rather early thermalization time of @xmath324 fm/@xmath12 , it is difficult to separate pre - equilibrium radiation from thermal emission ( e.g. , secondary drell - yan in @xmath90-@xmath23 interactions might overlap with thermal @xmath39-@xmath325 annihilation ) .
a rough estimate of pre - equilibrium effects may be obtained by varying the thermalization time ; decreasing , _
e.g. _ , @xmath326 to 0.8 fm/@xmath12 ( which is close to the overlap time of the 2 colliding nuclei at sps energies ) increases the qgp contribution at @xmath327 gev by @xmath1150% , which is less than 10% of the drell - yan contibution at all masses considered . for the dilepton contribution from correlated decays of @xmath328 and @xmath329 mesons
we use the experimental result from @xmath272-@xmath272 collisions extrapolated to in - in , as provided by the na60 collaboration @xcite . as a note of caution ,
we remark that recent measurements at the relativistic heavy - ion collider ( rhic ) report substantial modifications of heavy - quark spectra in au - au collisions , relative to @xmath272-@xmath272 ( as inferred from a suppression and elliptic flow of `` non - photonic '' single - electron spectra associated with semileptonic decays of open - charm ( and -bottom ) hadrons ) @xcite .
such ( possibly nonperturbative @xcite ) medium modifications of the charm momentum spectra presumably translate into a softening of the invariant - mass spectra of @xmath330 pairs as well . at the sps ,
the shorter qgp lifetime is likely to lead to smaller effects , but an explicit measurement of the ( delayed ) charm decays has been presented recently @xcite .
our description of the space - time evolution of central and semicentral @xmath187-@xmath187 collisions is approximated by an expanding thermal fireball characterized by a time dependent cylindrical volume as @xcite , @xmath331 where we neglect effects due to a finite ellipticity .
the initial transverse radius @xmath332 is determined by the centrality of the collision ( _ e.g. _ , @xmath333 for central ( semicentral ) in - in collisions ) .
the initial longitudinal size , @xmath334 , is equivalent to the formation time , @xmath326 , of the thermal medium , which we fix at the standard value of @xmath324 fm/@xmath12 ( translating into @xmath335 fm , where @xmath336 represents the rapidity width of a thermal fireball ) . for the longitudinal expansion
we employ a moderate acceleration , @xmath337/fm , together with an initial velocity of @xmath338 ( reminiscent to ref .
@xcite ) , but the dilepton invariant - mass spectra are essentially unaffected if we use @xmath339 as in ref . @xcite .
the most important parameter is the transverse acceleration .
more recent applications of the fireball model , both in the dilepton @xcite and charm diffusion @xcite context , have used larger values than in previous work @xcite , in the range @xmath340/fm . here , we employ @xmath341 as in ref .
@xcite .
the time evolution of the temperature is determined assuming entropy conservation . at a given collision energy , the hadro - chemistry of the fireball
is characterized by thermal - model fits to the observed hadron ratios . in our default scenario
we assume the chemical freezeout temperature to coincide with the critical temperature for qgp formation , at @xmath342 mev . for central and semicentral in(158a gev)+in collisions we fix the entropy per ( net ) baryon at @xmath343 , which , using a hadronic resonance gas ( hg ) equation of state ( eos ) , translates into an associated baryon chemical potential of @xmath344 mev , well within the uncertainties of recent thermal model fits at sps @xcite ( all other chemical potentials being zero ) .
the subsequent hadronic trajectory in the phase diagram is then constructed at fixed @xmath345 ( isentropic expansion ) with the additional constraints of pion , kaon , @xmath136 and antibaryon number conservation , requiring the build - up of corresponding chemical potentials . for a given number of nucleon participants ( _ i.e. _ , collision centrality ) ,
the fireball entropy amounts to @xmath346 for central ( semicentral ) in - in collisions , translating into a charged particle multiplicity of @xmath347 ( for a chemical freezeout temperature of @xmath348 mev , as considered below , the multiplicities increase by less than 5% ) . using the relation between the total entropy and volume , @xmath349 , the entropy density , @xmath350 , can be used to determine the time evolution of temperature and baryon density ( along the hadronic trajectory ) using the hg eos , and similarly for the qgp phase using a quasiparticle eos .
the volume partition in the hg - qgp mixed phase is calculated from the standard mixed - phase construction , where the fraction of matter in the hadronic phase is given by @xmath351 where @xmath352 denote the critical entropy densities at @xmath32 . with a formation time of @xmath324 fm/@xmath12
, the evolution for central ( semicentral ) in - in starts in the qgp at @xmath353 mev , passes through a mixed phase at @xmath354 , and terminates at thermal freezeout at around @xmath355-@xmath356 mev .
the main uncertainties associated with the fireball evolution are the transverse acceleration as well as the overall lifetime ( which is somewhat correlated to the longitudinal expansion ) .
it turns out that the latter is rather sensitive to the absolute magnitude of the experimentally observed dilepton excess radiation , resulting in @xmath357 for central in - in collisions .
the remaining uncertainty consists of the interplay between longitudinal and transverse expansion .
as mentioned above , the variations considered in our previous works @xcite , @xmath358-@xmath359 and @xmath360-@xmath361 , have negligible impact on the invariant mass spectra , while the longitudinal - acceleration scenario , which we will focus on here , allows for a slightly larger transverse expansion which appears to be favored by the rather hard @xmath5 spectra . in principle
, more accurate information on the final state of the expansion can be obtained once hadronic spectra for in - in are available ( our pion spectra are actually in reasonable agreement with s - au data , recall lower panel of fig .
[ fig_raa ] ) . for central pb - pb collisions at sps energies ,
the typical results for transverse surface velocity and thermal freezeout temperature are @xmath362)=(0.65\pm0.1,120\pm10)$ ] @xcite , which is quite comparable to the values used here for in - in .
we emphasize that all contributions to the dilepton spectrum ( qgp , @xmath1 , @xmath2 , @xmath3 , and four - pion ) are tied to the _ same _ evolution , thus fixing their relative weights .
finally , to illustrate the uncertainties associated with the underlying equation of state we will investigate three combinations of critical and chemical freezeout temperatures roughly covering the current theoretical and experimental ranges : * our default scenario , employed in most of our calculations thus far , consisting of identical @xmath32 and @xmath363 at an `` intermediate '' value of @xmath26 mev ; thermal freezeout is fixed at @xmath364 ( semicentral in - in ) . *
a scenario with a relatively small and identical critical and chemical freezeout temperature at @xmath365 mev , compatible with recent thermal model fits in refs .
@xcite . for dilepton spectra the most important consequences of this scenario
are a significantly extended qgp phase which will increase its thermal emission contribution and reduce the hadronic one , in particular at intermediate masses .
in addition , due to smaller pion chemical potentials in the subsequent hadronic phase , the freezeout temperature ( at fixed fireball lifetime ) will be larger than with eos - a , at @xmath366 . * a scenario with a large critical temperature @xmath367 mev ( as suggested by recent lattice qcd computations @xcite , which maximizes ( minimizes ) the space - time volume occupied by the hadronic ( qgp ) phase . since chemical freezeout at such a temperature is questionable , we allow for a chemically equilibrated hadronic phase until chemical - freeze out sets in under the same conditions as in eos - b , at @xmath368 mev .
we now turn to a systematic analysis of experimental dilepton spectra as measured at the sps by the na60 and ceres / na45 collaborations in in - in and pb - au collisions , respectively . based on the various ingredients developed in the previous sections , we first address the na60 invariant - mass and transverse - momentum spectra , followed by a consistency check with earlier and updated ceres data .
-mesons @xcite ( dash - dotted red line ) , 4@xmath90 annihilation with chiral @xmath66-@xmath187 mixing ( dashed blue line ) , qgp plus correlated open charm decays ( dotted orange line ) and drell - yan annihilation ( solid turquoise line ) ; the upper dashed brown line is the sum of the above , while the solid purple line additionally includes in - medium @xmath2 and @xmath3 decays as well as freezeout plus primordial @xmath1s ( solid black line ) . in the semi - central data , the uncertainty due to the @xmath136 cocktail subtraction
is indicated by the open and filled data points ( the former are based on an estimated @xmath136 yield at high @xmath5 while the latter represent an upper limit on the @xmath136 by subtracting the dimuon spectrum to zero at threshold).,scaledwidth=46.0% ] -mesons @xcite ( dash - dotted red line ) , 4@xmath90 annihilation with chiral @xmath66-@xmath187 mixing ( dashed blue line ) , qgp plus correlated open charm decays ( dotted orange line ) and drell - yan annihilation ( solid turquoise line ) ; the upper dashed brown line is the sum of the above , while the solid purple line additionally includes in - medium @xmath2 and @xmath3 decays as well as freezeout plus primordial @xmath1s ( solid black line ) . in the semi - central data , the uncertainty due to the @xmath136 cocktail subtraction
is indicated by the open and filled data points ( the former are based on an estimated @xmath136 yield at high @xmath5 while the latter represent an upper limit on the @xmath136 by subtracting the dimuon spectrum to zero at threshold).,scaledwidth=46.0% ] - , @xmath2- and @xmath3-mesons without baryon effects ( + qgp+charm+in - medium four - pion ; solid purple line ) , and ( ii ) free @xmath1 ( + qgp+charm+free four - pion ; dashed green line ) .
lower panel : na60 data @xcite in the imr compared to thermal dimuon spectra with different implementations of the four - pion contribution , using either its vacuum form ( lower dashed line ) or including chiral mixing ( upper dashed line ) , and corresponding total spectra ( lower and upper solid line , respectively).,scaledwidth=46.0% ] - , @xmath2- and @xmath3-mesons without baryon effects ( + qgp+charm+in - medium four - pion ; solid purple line ) , and ( ii ) free @xmath1 ( + qgp+charm+free four - pion ; dashed green line ) .
lower panel : na60 data @xcite in the imr compared to thermal dimuon spectra with different implementations of the four - pion contribution , using either its vacuum form ( lower dashed line ) or including chiral mixing ( upper dashed line ) , and corresponding total spectra ( lower and upper solid line , respectively).,scaledwidth=46.0% ] thermal @xmath369 invariant - mass spectra for @xmath187-@xmath187 collisions are computed by integrating the emission rate , eq .
( [ rate ] ) , over the fireball evolution ( as well as three - momentum ) , @xmath370 where @xmath371 denotes the detector acceptance which has been carefully tuned to na60 simulations @xcite .
the fugacity factor , @xmath372=@xmath373 , arises due to chemical off - equilibrium in the hadronic phase for @xmath374 ; it depends on the thermal emission source under consideration , cf .
( [ piem ] ) : for the @xmath1 , @xmath2 and four - pion contributions one has @xmath375 with @xmath376 , respectively , while for the @xmath3 one has @xmath377 , where @xmath378 accounts for strangeness undersaturation for medium - size nuclear collision systems at the sps @xcite .
in addition , appropriate fugacities figure into the various in - medium selfenergy contributions .
upper panel : @xmath379 gev ; lower panel : @xmath380 gev.,scaledwidth=45.0% ] @xcite .
upper panel : @xmath379 gev ; lower panel : @xmath380 gev.,scaledwidth=45.0% ] but for central collisions.,scaledwidth=45.0% ] but for central collisions.,scaledwidth=45.0% ] initial comparisons @xcite of na60 data to theoretical predictions @xcite have focused on the contribution from the @xmath1 meson which dominates in the lmr .
the shape of the in - medium @xmath1-spectral function describes the experimental spectra well , but the absolute yields have been overestimated by @xmath381 .
this discrepancy has been resolved @xcite by increasing the transverse fireball expansion ( @xmath382 ) , reducing the fireball lifetime to about @xmath383-@xmath384 fm/@xmath12 , cf .
[ sec_fireball ] above .
in addition , the larger transverse expansion leads to harder emission spectra in @xmath5 , which will be helpful in understanding the @xmath5 spectra , as discussed in the following sections .
[ fig_med ] summarizes our results for the mass spectra in semicentral and central in(158 agev)-in collisions , computed for the eos - a scenario ( @xmath385 mev ) .
the modifications relative to our previous work @xcite are : ( i ) the freezeout @xmath1 has been separated from the in - medium @xmath1 contribution , ( ii ) primordial @xmath1 and ( iii ) drell - yan contributions have been added . as a result of the separation ( i ) ,
the thermal emission lifetime is now slightly smaller , 6.5(6.2 ) fm/@xmath12 for central ( semicentral ) collisions .
the inclusive mass spectra ( and pertinent conclusions ) are essentially identical to those in ref .
@xcite , _ i.e. _ , the predicted in - medium effects on @xmath1 spectral function lead to good agreement with the data in the lmr , while the same fireball evolution also reproduces the observed excess in the imr well .
the largest source here is four - pion annihilation , together with smaller contributions from open - charm decays and qgp emission . in - medium
@xmath3 decays are noticeable but not very significant relative to the uncertainty in the data . the sensitivity to the medium effects in the e.m .
correlator is further illustrated in fig .
[ fig_vac ] .
the upper panel demonstrates that a free @xmath1 spectral function is ruled out , but also one which only includes modifications due to a meson gas clearly does not reproduce the data , due to both a too narrow peak and a lack of enhancement below the free @xmath1 mass , especially when approaching the dimuon threshold . in the lower panel
one sees that the effect of chiral mixing on the 4@xmath90 contribution amounts to up to a factor of @xmath112 enhancement in the @xmath63 resonance region , but the effect on the total is rather moderate .
thus , no strong case on the chiral mixing can be made at present .
the prevalence of the four - pion contribution in the imr is reminiscent to the hydrodynamic calculations in ref .
@xcite where hadronic rates calculated in the chiral reduction formalism have been employed . on the contrary , in the fireball calculations of ref .
@xcite qgp emission dominates in the imr .
we elucidate on this discrepancy in sec .
[ ssec_hadrochem ] below . in figs .
[ fig_mspecpt ] and [ fig_mspecpt_central ] we compare our theoretical calculations to na60 @xmath24-spectra binned into regions of low ( @xmath386 gev ) and high ( @xmath387 gev ) transverse pair momentum , for both semicentral and central collisions .
also in this representation the agreement is fair .
there is possibly an indication of a slight over- ( under- ) estimate in the high-@xmath5 bin for ( semi- ) central collisions . to further scrutinize this issue we now turn to @xmath5 spectra , binned in invariant mass . in analogy to eq .
( [ spect ] ) for @xmath24 spectra , @xmath5 spectra are computed by integrating the dilepton rate , eq .
( [ rate ] ) , over the space - time evolution of the fireball and a suitable interval in invariant mass , @xmath388 $ ] .
however , since the available experimental spectra have been corrected for the detector acceptance , the monte - carlo simulation of the acceptance function ( with radial and angular dependencies due to directed ( radial flow ) and random ( thermal ) motion of the virtual photon ) is now replaced by explicit integrations over the radial coordinate , @xmath389 , of the fireball , the azimuthal angle , @xmath390 , of the virtual photon momentum and its rapidity , @xmath296 , @xmath391 as before , @xmath392 denotes the radius of the expanding fire - cylinder ( corresponding to the first parenthesis in eq .
( [ vfb ] ) ) , @xmath393 its longitudinal length ( second parenthesis in eq .
( [ vfb ] ) ) .
note that the rate , eq .
( [ rate ] ) , is calculated in the thermal rest frame while the integrations in eq .
( [ qt - spect ] ) are in the laboratory frame .
the relation between the four - momentum , @xmath394 , in the laboratory frame and the one in the local rest frame of the fluid cells , @xmath395 , is determined by a boost with the radial flow velocity , eq .
( [ flow - profile ] ) . due to rotational invariance the thermal rate , eq .
( [ rate ] ) , depends only on the magnitude of the three - momentum in the thermal rest frame and therefore one of the integrations over the spatial and momentum azimuthal angles in eq .
( [ qt - spect ] ) becomes trivial yielding a factor of @xmath396 . to gain qualitative insights into the behavior of dilepton @xmath5 spectra from thermal sources ,
let us assume that the three - momentum dependence of the hadronic e.m .
current correlator is weak ( it would be absent in the absence of medium effects ) , so that for not too large invariant - mass intervals , the integrand in eq .
( [ qt - spect ] ) can be considered constant except for the bose distribution . for @xmath397 , it is possible to approximate the qualitative behavior of the @xmath5 spectra in analytic form @xcite . for a linear flow profile , eq .
( [ flow - profile ] ) , and in boltzmann approximation one has @xmath398 with @xmath399 weakly @xmath5 dependent . substituting @xmath400 ( @xmath401 ) the integrals over @xmath296 and @xmath3 are given by modified bessel functions .
for the @xmath296-integral we can use their asymptotic form for @xmath402 : @xmath403 in the non - relativistic limit , @xmath404 , we use the asymptotic form @xmath405 to obtain @xmath406 we have numerically verified that this expression can be further simplified by the following approximate treatment of the @xmath389 integral : in the pre - factor of the exponential substitute @xmath407 by its average over @xmath389 , @xmath408 use the @xmath389 average of the argument in the exponential , which , together with the non - relativistic approximation @xmath409 , results in @xmath410 where @xmath411 and @xmath412 .
the form for @xmath413 is the known `` pocket formula '' for the parametric dependence of the effective slopes of hadronic @xmath5 spectra on the fireball temperature , the particle mass and the surface - flow velocity @xcite .
we recall that in the ultrarelativistic limit , @xmath414 , the dependence of the effective temperature on the radial velocity is given by the doppler blue - shift expression for a thermalized gas of massless particles , @xmath415 which follows from the large-@xmath5 limit of @xmath416 in eq .
( [ spect - besseli0 ] ) @xcite which also provides an additional factor @xmath417 . for our @xmath389-dependent flow profile
the effective temperature is given by the blue - shift value for an average value @xmath418 where @xmath419 is not easily determined by simple approximations of the radial integral .
numerical studies show that in the case of a linear radial - flow profile , eq .
( [ flow - profile ] ) , and typical parameters in the region of the @xmath1 peak ( @xmath420 , @xmath421 ) , @xmath422-@xmath423 leads to a good estimate for @xmath424 at high @xmath5 .
it is important to note that for vector - meson decays after thermal freezeout , cf .
[ ssec_fo ] ) , the @xmath5 spectra are harder by an additional factor @xmath425 , _
i.e. _ , applying the same approximations to eq .
( [ fo - dilep ] ) as used to derive eq .
( [ fancy - pocket ] ) , one obtains @xmath426 with @xmath424 again given by eq .
( [ fancy - teff ] ) .
the additional factor @xmath425 in ( [ fancy - pocket - fo ] ) compared to ( [ fancy - pocket ] ) originates from the additional factor @xmath427 in the cooper - frye ( cf ) formula ( [ fo - dilep ] ) compared to the mclerran - toimela ( mt ) emission formula ( [ rate ] ) , as discussed in detail in sec .
[ ssec_fo ] .
gev ; middle panel : @xmath428 gev ; lower panel : @xmath429 gev . , scaledwidth=47.0% ] gev ;
middle panel : @xmath428 gev ; lower panel : @xmath429 gev .
, scaledwidth=47.0% ] gev ; middle panel : @xmath428 gev ; lower panel : @xmath429 gev . ,
scaledwidth=47.0% ] , but for central in(158 agev)-in.,scaledwidth=47.0% ] , but for central in(158 agev)-in.,scaledwidth=47.0% ] , but for central in(158 agev)-in.,scaledwidth=47.0% ] the above discussion of effective slopes in thermal emission spectra obviously does not apply to non - thermal dilepton sources , _
i.e. _ , decays of `` primordial @xmath1 mesons '' ( sec .
[ ssec_hard ] ) and drell - yan annihilation ( sec [ ssec_dy ] ) .
both are characterized by a power - law behavior at high @xmath5 , where their contribution becomes potentially important . toward lower @xmath5 ,
the primordial @xmath1 contribution is much suppressed due to @xmath1 absorption ( `` jet quenching '' ) , while the drell - yan process is no longer well - defined .
however , the drell - yan dileptons carry the hardest slope of all sources considered , which renders even a naive extrapolation of their spectra to low @xmath5 very small ( no more than a few percent for @xmath430 gev in all mass bins below @xmath431 gev ) . at @xmath432 gev , in turn , the drell - yan contribution is quite appreciable .
[ fig_ptspec ] and [ fig_ptspec_central ] summarize our calculations for semicentral and central in - in collisions , respectively , in three different mass bins in comparison to the na60 data @xcite .
open - charm decays have been removed from the experimental spectra and are consequently not included in the theory curves either .
the parameters of our fireball evolution are as described in sec .
[ sec_fireball ] within our default eos - a scenario ( @xmath433 mev ) .
since the overall normalization of the experimental spectra is not known at present , we have fixed it in each mass bin using the invariant mass spectra in the low-@xmath5 bin ( @xmath434 ) .
this easily translates into a @xmath15-@xmath16 uncertainty in the absolute normalization . for the central collisions
the agreement between theory and data is quite satisfactory in all mass bins . in the high - mass bin @xmath143 gev@xmath435@xmath24@xmath435@xmath20 gev , lower panel in fig .
[ fig_ptspec_central ] ) , this procedure appears to overestimate somewhat the @xmath5 spectra at @xmath436 gev , which , however , seems not to be reflected in the corresponding @xmath5 bin in the mass spectra ( lower panel in fig .
[ fig_mspecpt_central ] ) .
this could very well be due to the uncertainty in the underlying normalization procedure .
note again that the compared to ref .
@xcite additionally implemented hard components ( primordial @xmath1 s and drell - yan ) are insignificant at @xmath430 gev , _
i.e. _ , for the total yield and the understanding of the spectral shape of the inclusive @xmath24-spectra . on the other hand , at higher @xmath5 , these contributions are essential ( even dominant ) for a proper description of the spectra . in fact , in semicentral collisions , the experimental transverse momentum spectra in the two mass bins below ( @xmath437 gev@xmath438 gev ) and around ( @xmath439 gev@xmath440 gev ) the free @xmath1-meson mass turn out to be underestimated ( and too soft ) for @xmath441 gev . especially in the low - mass bin , the discrepancy again appears to be larger than one could anticipate from the mass spectrum in the @xmath437-@xmath439 gev regime in the @xmath442 gev bin ( lower panel in fig .
[ fig_mspecpt ] ) .
the fact remains , however , that the theoretical @xmath5 spectra in the low and intermediate mass bin are somewhat too soft .
even though the final transverse flow velocity of the fireball model is about @xmath443 smaller for semicentral relative to central in - in collisions , this difference would not be able to account for the discrepancy ( it amounts to a change in slope by about @xmath15 mev for the freeze - out @xmath1 , and even less for thermal radiation ) .
the nuclear suppression factor , @xmath444 , in fig .
[ fig_raa ] ( solid line ) suggests that there might be room for a 30% increase of the primordial plus freezeout @xmath1 contribution ( possibly more if one could compare to the @xmath280 of @xmath1 mesons ; due to their larger mass a more pronounced flow effect could enhance their @xmath280 ) . whether this would suffice in the low - mass bin
, is questionable . , but additionally including ( factor 2 augmented ) contributions from @xmath2 @xmath7-channel exchange in @xmath445.,scaledwidth=47.0% ] , but additionally including ( factor 2 augmented ) contributions from @xmath2 @xmath7-channel exchange in @xmath445.,scaledwidth=47.0% ] , but additionally including ( factor 2 augmented ) contributions from @xmath2 @xmath7-channel exchange in @xmath445.,scaledwidth=47.0% ] another possibility consists of meson @xmath7-channel exchange in elastic scattering of @xmath1 mesons off pions in the heat bath ( with subsequent conversion into a lepton pair ) .
these processes have been calculated previously for real photon production @xcite ( which are given by the same diagrams except that the final - state photon is on - shell ) , where @xmath2 exchange has been found to be the most important process at high @xmath5 ; however , contributions from other exchange processes ( _ e.g. _ , @xmath90 and @xmath63 exchange in @xmath446 , or reactions involving strange mesons @xcite ) are also appreciable . to roughly account for the latter , we augment the rates computed in sec .
[ ssec_omt ] by a factor of two ( in addition , we recall that @xmath208 initial states in the hadronic fireball emission carry a pion fugacity factor to the third power , @xmath447 ) .
the convolution of these rates over the fireball evolution leads to a contribution to dilepton-@xmath5 spectra as shown by the lower dash - dotted line in each of the panels in fig .
[ fig_ptspec_om_t ] ; it indeed provides the hardest spectrum among all thermal sources ( qgp , in - medium vector mesons and four - pion annihilation ) , in line with the rates displayed in fig .
[ fig_drdq - om - t ] .
consequently , the relative importance of the @xmath7-channel processes grows with @xmath5 but remains rather moderate even at @xmath448 , up to @xmath449 and @xmath301 of the total theoretical yield in the @xmath135-@xmath450 and @xmath439-@xmath451 mass bins , respectively ( negligible for @xmath452 gev ) .
note that the slopes of the @xmath7-channel emission spectra resemble the data quite well , but our present estimate of their strength is insufficient to resolve the discrepancies at high @xmath5 . however , their impact on the slope of the total spectra is not insignificant , as we will see in sec .
[ ssec_slope ] below . in all our calculations of thermal dilepton spectra
thus far , the medium evolution was based on the notion that the critical ( @xmath32 ) and chemical freezeout ( @xmath363 ) temperature coincide , at @xmath453 mev .
more recent theoretical ( lattice qcd ) and phenomenological ( thermal model fits to hadron ratios ) studies , however , allow for the possibilities that @xmath32 could be significantly larger ( @xmath30-@xmath28 mev ) and @xmath363 significantly smaller ( @xmath454-@xmath455 mev ) .
while a smaller @xmath456 mev is still viable , a larger @xmath363 appears unlikely since in a high - density hadronic phase number - changing reactions affecting the chemistry are to be expected ( in addition , thermal model fits start to become unstable at temperatures above @xmath11@xmath457 mev due to uncertainties in the high - lying , high degeneracy hadron resonance spectrum ) .
therefore , in this section we study the sensitivity of the invariant - mass and @xmath5 spectra to the hadro - chemistry of the fireball , keeping its geometry and flow parameterization as well as the total lifetime the same as before ( including all normalizations , where applicable ) .
as quoted in sec [ sec_fireball ] , in addition to our standard equation of state ( eos - a : @xmath458 ) , we investigate two alternative scenarios : eos - b with @xmath459 and eos - c with @xmath460 , @xmath281
. in general , a lower chemical freezeout temperature ( where all meson chemical potentials are by definition zero ) entails a larger volume at chemical freezeout , and thus also smaller baryon chemical potentials ( _ i.e. _ , smaller baryon densities ) , as well as smaller meson chemical potentials in the subsequent evolution .
this furthermore implies that , to obtain roughly the same overall dilepton yield , thermal freezeout occurs at a _ larger _ temperature ( and smaller @xmath461 ) , which , in principle , leads to ( somewhat ) harder @xmath5 spectra , both at thermal freezeout and for thermal emission throughout the hadronic evolution . in the remainder of this section , we focus on semicentral collisions . , but for a fireball with a critical temperature @xmath462 and a chemical - freezeout temperature @xmath463 ( eos - b).,scaledwidth=45.0% ] , but for a fireball with a critical temperature @xmath462 and a chemical - freezeout temperature @xmath463 ( eos - b).,scaledwidth=45.0% ] , but for a fireball with a critical temperature @xmath462 and a chemical - freezeout temperature @xmath463 ( eos - b).,scaledwidth=45.0% ] , but for a fireball with a critical temperature @xmath462 and a chemical - freezeout temperature @xmath463 ( eos - b).,scaledwidth=47.0% ] , but for a fireball with a critical temperature @xmath462 and a chemical - freezeout temperature @xmath463 ( eos - b).,scaledwidth=47.0% ] , but for a fireball with a critical temperature @xmath462 and a chemical - freezeout temperature @xmath463 ( eos - b).,scaledwidth=47.0% ] , but for a fireball with a critical temperature @xmath464 and a chemical - freezeout temperature @xmath463 ( eos - c).,scaledwidth=45.0% ] , but for a fireball with a critical temperature @xmath464 and a chemical - freezeout temperature @xmath463 ( eos - c).,scaledwidth=45.0% ] , but for a fireball with a critical temperature @xmath464 and a chemical - freezeout temperature @xmath463 ( eos - c).,scaledwidth=45.0% ] , but for a fireball with a critical temperature @xmath464 and a chemical - freezeout temperature @xmath463 ( eos - c).,scaledwidth=47.0% ] , but for a fireball with a critical temperature @xmath464 and a chemical - freezeout temperature @xmath463 ( eos - c).,scaledwidth=47.0% ] , but for a fireball with a critical temperature @xmath464 and a chemical - freezeout temperature @xmath463 ( eos - c).,scaledwidth=47.0% ] we start our investigation within the `` eos - b '' scenario , _
i.e. _ , @xmath465mev . the resulting @xmath24- and @xmath5-spectra
are compiled in figs .
[ fig_mspecpt_semi_fb6_tc160 ] and [ fig_ptspec_semi_fb6_tc160 ] , respectively .
the smaller @xmath32 implies a longer duration of both the qgp and mixed phase ( the latent heat for eos - b is larger than for eos - a since entropy density for the qgp eos drops slower with @xmath31 than for the hg eos ) , while the duration of the hadronic phase is accordingly reduced . as noted above
, the thermal - freezeout temperature has increased to @xmath466 mev ( compared to @xmath467 mev for eos - a ) . as an immediate consequence ,
qgp radiation is significantly enhanced , while the hadronic yield is slightly reduced ( both for in - medium @xmath1 in the lmr and four - pion in the imr ) .
the overall quality in the description of the invariant - mass spectra is quite comparable to the eos - a scenario .
the somewhat harder spectra implied by the larger hadronic temperatures for eos - b ( most notably for the freezeout @xmath1 ) lead to a slight increase of the dilepton yield in the @xmath1 mass region in the @xmath442 gev momentum bin ( compare lower panels in figs .
[ fig_mspecpt ] and [ fig_mspecpt_semi_fb6_tc160 ] ) .
however , in the low- ( @xmath437 gev@xmath468 gev ) and intermediate - mass ( @xmath439 gev@xmath469 gev ) bins of the @xmath5 spectra , the improvement at high @xmath5 is rather marginal , while in the higher mass ( @xmath470 gev@xmath471 gev ) there is essentially no change compared to the eos - a calculation ( recall fig .
[ fig_ptspec ] ) , despite the fact that the qgp contribution is now larger than the four - pion one . turning to the eos - c scenario ( summarized in figs .
[ fig_mspecpt_semi_fb6 ] and [ fig_ptspec_semi_fb6 ] ) , the large value of @xmath367 mev leads to a qgp and mixed phase which is shorter than for eos - a , but there is now a `` high - density '' ( chemically equilibrated ) hadronic phase down to @xmath348 mev , followed by a subsequent chemical off - equilibrium evolution which is identical to eos - b .
obviously , the qgp emission yield is the smallest of all three scenarios , while the hadronic yield ( both in - medium @xmath1 s and four - pion annihilation ) is the largest , possibly implying a slight overestimate in the @xmath1-mass region of the inclusive mass spectra ( @xmath472 larger than for eos - a , which could be readjusted by a slightly reduced lifetime ) .
the high - density hadronic phase helps in the @xmath5 spectra in the low- and intermediate - mass bin , for which , at high @xmath5 , the discrepancy is the smallest for the three scenarios . in the highest mass bin ,
even the rather large ratio of four - pion annihilation over qgp emission does not upset the agreement with the @xmath5 spectra , indicating that a large part of the four - pion emission emanates from the high - density hadronic phase .
gev , is not shown.,scaledwidth=45.0% ] since the question of four - pion vs. qgp emission in the imr has drawn considerable attention in the recent literature @xcite , we take a closer look at their interplay in our three scenarios in fig .
[ fig_qgp_4pi_eos_comp ] .
the plot illustrates again that with eos - b ( @xmath473 mev ) the qgp contribution exceeds the in - medium four - pion annihilation over the entire imr considered ( note that the tails of the @xmath1 spectral function are not included ) .
the opposite trend applies to our default eos - a scenario , and even more so for eos - c .
especially the large difference between eos - b and eos - c clearly demonstrates that the major portion of the four - pion emission emanates from the high - density hadronic phase ( eos - b and eos - c have an identical hadronic evolution below @xmath348 mev ) ; a similar argument applies to eos - a , _
i.e. _ , four - pion emission does not primarily arise in the late stages ( with large flow ) despite an enhancement through large pion - fugacity factors .
the basic reason for this can be understood by inspecting the temperature - mass differential emission yield which is roughly given by @xcite @xmath474 characterized by the standard thermal exponential factor and a power in temperature resulting from the three - momentum integral over the boltzmann factor and , most importantly , the volume expansion . assuming a weak temperature dependence of the e.m .
spectral function ( medium effects bias the emission toward higher @xmath31 ) , a differentiation over @xmath31 identifies the temperature of maximum emission as @xmath475 . in the imr of interest here , this means a @xmath476 right around @xmath32 ( note that , at @xmath32 , the hadronic phase has an extra `` advantage '' of a roughly twice larger volume compared to the qgp due to the latent heat ) , consistent with our fireball emission pattern .
we conclude this section by noting that `` reasonable '' variations in hadro - chemical freezeout and critical temperature have a very moderate impact on the invariant - mass spectra as seen by na60 . on the one hand , this implies little sensitivity to the concrete values of @xmath32 and @xmath363 . on the other hand , it means that our results are very robust with respect to uncertainties in these quantities .
the main reason for this robustness is the `` quark - hadron duality '' of the underlying ( medium - modified ) emission rates from the hadronic and qgp phase in the relevant temperature regime , @xmath29-@xmath30 mev , in both lmr ( @xmath1 melting ) and imr ( including chiral mixing ) . as such
, it provides additional support to the medium modifications in the employed rates . in the imr , this `` duality '' does not allow for a ( maybe even academic ) distinction between a high - density hadronic or partonic source .
partitions with either component dominant are viable in both mass and transverse momentum spectra . concerning the high-@xmath5 region of the lower two mass bins ,
our calculations indicate a slight preference for eos - b and eos - c , due to their smaller @xmath363 , implying larger temperatures in the hadronic evolution . to complete our analysis of the @xmath5 spectra we perform in this section a quantitative analysis of effective slope parameters following the procedure adopted by the na60 collaboration for the experimental data @xcite .
the @xmath5 spectra have been divided into several invariant - mass bins , in each of which the data have been fitted to the function @xmath477 where the fit range is typically taken as @xmath478 gev , but the extracted slopes are rather insensitive to variations as long as the lower limit is above @xmath437 gev .
in addition , the experimental @xmath5 spectra exhibit little dependence on centrality ( except for peripheral collisions ) , and therefore the slope analysis has been performed for inclusive @xmath5 spectra with @xmath479 . in our theoretical analyses we focus on semicentral collisions . we have tried to follow the same procedure ( see also ref .
@xcite , but as illustrated by the dash - dotted line in fig .
[ fig_teff_fit_range_dep ] , a fit to our theoretical spectra ( represented by the dots ) in the above range ( @xmath480 gev ) slightly underestimates the effective slopes toward higher @xmath5 , especially in the lower mass bins .
part of the problem is the bias of the fit toward low @xmath5 where the yield is the largest and thus dominates the total @xmath481 . to better reproduce the theoretical spectral shape with the above function we have therefore chosen to ( i ) restrict the fit range to @xmath482 , and ( ii ) minimize the @xmath481 for the logarithm of the spectra .
the resulting fit ( solid line in fig .
[ fig_teff_fit_range_dep ] ) indeed agrees better with the theoretical spectra .
this effect becomes more pronounced when the drell - yan component is included ( which , in fact , might indicate that our extrapolation overestimates the dy yield at low mass ) . based on the slightly revised procedure ,
our slope analysis is carried out for the total sum of all calculated @xmath5 spectra components in twelve @xmath483 mass bins in the range @xmath484 ( corresponding to the experimental ones ) for semicentral in - in ; by default , the drell - yan and meson @xmath7-channel contributions are not included , but studied in separate fits by adding them to the totals . as in the experimental analysis ,
correlated open - charm decays are removed altogether .
our systematic study addresses variations in the eos as well as the possibility of another @xmath11@xmath449 increase in the transverse acceleration of the fireball expansion , _
i.e. _ , @xmath485/fm ( which is , in fact , the value we have used in our calculations for heavy - flavor observables at rhic @xcite , as extracted from the hydrodynamic model of ref .
@xcite ) . specifically , we evaluate the following scenarios , as summarized in the six panels of fig .
[ fig_slopes ] : 1 . the default fireball expansion with eos - a ( @xmath433 mev ) , characterized by a thermal freezeout with temperature and radial flow surface velocity of @xmath486 ; 2 . the default fireball expansion with eos - b ( @xmath487 mev ) and @xmath488 ; 3 . the default fireball expansion with eos - c ( @xmath367 mev , @xmath489 mev ) and @xmath488 ; 4 . the same as in ( b ) , but with a transverse fireball acceleration @xmath490 yielding @xmath491 ; 5 . the same as in ( c ) , but with a transverse fireball acceleration @xmath490 yielding @xmath491 ; 6 .
the same as in ( c@xmath492 ) , but additionally including the contribution from meson-@xmath7-channel exchange with in - medium @xmath1 propagator ( cf . sec .
[ ssec_omt ] ) ; 7 . the same as in ( c@xmath493 ) , but using the vacuum @xmath1 propagator in the meson-@xmath7-channel exchange contribution .
the upper two panels of fig . [ fig_slopes ] compare the impact of the three different eos using our baseline acceleration of @xmath494/fm .
the default scenario eos - a ( a ) falls short of the empirical slopes by a significant margin of up to @xmath495-@xmath496 mev in the lmr , while reaching the lower end of the data in the imr .
the higher hadronic temperatures of scenarios eos - b ( b ) and eos - c ( c ) improve the situation by about @xmath497-@xmath498 mev in the lmr , approximately reflecting the increase in the hadronic fireball temperatures ( e.g. , @xmath466 mev compared to @xmath467 mev at thermal freezeout ) .
scenario eos - b ( b ) additionally improves around the @xmath1 peak due to a larger weight of the relatively hard components from decays of freezeout and primordial @xmath1 , since the overall thermal hadronic emission is smaller than for eos - a and eos - c ( compare fig .
[ fig_ptspec_semi_fb6_tc160 ] to figs .
[ fig_med ] and [ fig_mspecpt_semi_fb6 ] ) .
however , for both eos - b and eos - c the slopes in the lmr are still below the data , even though the shape is not too bad . in an attempt to improve on the slopes ,
we investigate a further increase of the transverse acceleration to values representative for rhic energies @xcite , from @xmath499 to @xmath500 for eos - b and eos - c ( scenarios ( b@xmath492 ) and ( c@xmath492 ) , shown in the middle panels of fig . [ fig_slopes ] ) . while this increases the effective slopes at all emission stages , it is more efficient for the later hadronic sources ( in accordance with the blue - shift pocket formulae given in sec .
[ ssec_qtspec ] ) .
consequently , the lmr exhibits further improvement , and , importantly , the consistency with the data in the imr is not spoiled due to the fact that the main contributions there arise from earlier stages with higher temperatures and less flow ( dominated either by the contributions from the qgp phase for eos - b or four - pion annihilation in the high - density hadronic phase for eos - c , cf .
[ fig_qgp_4pi_eos_comp ] ) . at this point ,
eos - b provides somewhat better agreement and one might be tempted to take this as evidence for a prevalently partonic emission source , as suggested in refs .
however , remaining uncertainties preclude an unambiguous interpretation , as we will now show . in the two bottom panels of fig .
[ fig_slopes ] we focus on the eos - c scenario with @xmath485/fm .
the left panel illustrates the sensitivity of the slopes to the additional inclusion of @xmath7-channel meson exchange processes . as discussed in the previous section ,
their spectra alone carry a slope very similar to the data ; not surprisingly , their inclusion increases the slopes in the lmr with no impact on the imr slopes .
when implementing the @xmath7-channel meson exchanges with the vacuum @xmath1 propagator , the effect on the slopes is somewhat larger compared to using the in - medium @xmath501 , especially in the free @xmath1 mass region .
a single @xmath208 scattering with unmodified @xmath1 mesons would most closely represent interactions of @xmath1 mesons in a kinetic regime ( rescattering without thermalization ) , and thus provide a missing link between the two extremes of thermal radiation and primordial @xmath1 s surviving jet quenching ( no reinteraction ) in our baseline calculations ( note that there is no contribution from `` quenched '' @xmath1 s ) .
the overall level of agreement with the data is now very similar to eos - b .
also recall that @xmath367 mev maximizes the ratio of hadronic to qgp emission ; for @xmath275 mev four - pion emission still outshines the qgp in the imr , while the slope parameters are closer to the eos - b scenario .
the effect of the drell - yan contribution on the slopes is shown in the bottom - right panel of fig .
[ fig_slopes ] .
in the imr the change in slope is consistent with the experimental results ( full vs. open squares ) , while in the lowest mass bins the uncertainties in the extrapolation of the dy are large .
we emphasize again that more stringent constraints can be enforced once an accurate knowledge of the absolute normalization in the experimental mass and momentum spectra becomes available .
there are uncertainties in the experimental extraction of the slope parameters beyond the statistical error bars shown in the figures . in ref .
@xcite a systematic error of @xmath117 mev is estimated due to the subtraction of background and decay ( cocktail ) sources . for the lmr , it has been estimated that the subtraction of a potential drell - yan contribution could lower the effective slopes by 5 - 10 mev .
one should also be aware of that the applicability of hydrodynamic ( or thermal fireball ) descriptions is limited in transverse momentum , typically to @xmath502 gev for pion spectra at the sps ( consistent with where the spectra reach the collision - scaling regime in fig .
[ fig_raa ] ) , and to significantly smaller @xmath22 for elliptic flow .
fireball models might compensate for this to some extent by implementing a larger acceleration , which may be ( part of ) the reason that the dilepton slopes prefer @xmath485/fm . to further scrutinize the @xmath5 spectra
, it would be very illuminating to reiterate earlier implementations of the dilepton rates from hadronic many - body theory into hydrodynamic @xcite ( with a more realistic hadro - chemsitry ) or transport @xcite simulations . in this context
, a measurement of dilepton elliptic flow could provide further insights in the detailed emission history of the medium at sps energies @xcite . after the refinements in our theoretical approach for sps dilepton spectra in the context of the na60 data , we revisit in this section the consequences for the comparison with ceres / na45 in semicentral and central pb - au collisions , which showed good agreement in our earlier works @xcite . in the following ,
we constrain ourselves to the above default scenario for the fireball evolution with a transverse acceleration of @xmath503 . as in the in - in case
, the increased expansion rate entails shorter fireball lifetimes ; we choose the latter as to obtain the same thermal freezeout temperature for central ( @xmath504 mev ) and semicentral ( @xmath505 mev ) pb - au collisions as in our previous work @xcite , implying a reduction by @xmath241 fm/@xmath12 for both centralities ( @xmath506 fm/@xmath12 and @xmath507 fm/@xmath12 , respectively ) .
also , since for the larger system sizes ( _ i.e. _ , lifetimes ) the freezeout- and cocktail-@xmath1 contributions @xcite are significantly suppressed relative to thermal radiation , and since we are not interested in ( high- ) @xmath508 spectra here , we adopt the ( old ) simplified treatment for the freezeout @xmath1 by running the fireball an additional duration of @xmath143 fm/@xmath12 @xcite .
the larger lifetimes relative to in - in collisions furthermore imply that the nonrelativistic formula for the surface expansion velocity , @xmath509 , approaches values uncomfortably close to @xmath12 in the late stages of the collision . to implement a relativistically covariant acceleration in a simple way , we model the acceleration of the fireball in analogy to the relativistic motion for a charged particle in a homogeneous electric field .
the radial surface velocity and fireball radius then take the form @xmath510 which tames the acceleration at late times and matches the non - relativistic case for early times .
( short - dashed lines ) and qgp emission ( dash - dotted lines ) using our previous ( labeled `` old '' ) and updated fireball evolution .
the solid line is the sum of in - medium @xmath1 , @xmath2 and @xmath3 contributions , while the long - dashed line results from four - pion annihilation including chiral mixing as described in sec . [ ssec_intmass ] .
the drell - yan contribution is below the displayed range ; correlated charm decays are neglected .
lower panel : the solid line is the total in - medium thermal emission + hadronic cocktail ( dash - dotted line , adopted from the experimental evaluation @xcite ) , compared to combined 1995/1996 ceres / na45 data @xcite . the short - dashed line is the sum of thermal emission with a vacuum @xmath1 spectral function and the cocktail . , scaledwidth=42.0% ] ( short - dashed lines ) and qgp emission ( dash - dotted lines ) using our previous ( labeled `` old '' ) and updated fireball evolution . the solid line is the sum of in - medium @xmath1 , @xmath2 and @xmath3 contributions , while the long - dashed line results from four - pion annihilation including chiral mixing as described in sec . [ ssec_intmass ] .
the drell - yan contribution is below the displayed range ; correlated charm decays are neglected .
lower panel : the solid line is the total in - medium thermal emission + hadronic cocktail ( dash - dotted line , adopted from the experimental evaluation @xcite ) , compared to combined 1995/1996 ceres / na45 data @xcite . the short - dashed line is the sum of thermal emission with a vacuum @xmath1 spectral function and the cocktail . , scaledwidth=45.0% ] in the upper panel of fig .
[ fig_ceres - semi ] the updated calculations for dielectron spectra in semicentral pb - au collisions are compared to our previous results ( using the same hadro - chemistry and @xmath1 spectral function ) , including experimental acceptance cuts as defined by ceres / na45 . for masses
@xmath511 gev , the @xmath1 spectral function yield is reduced by @xmath497-@xmath16 , directly reflecting the reduction in the lifetime of the hadronic phase in the fireball . at lower masses , however , the difference is smaller and eventually disappears for @xmath512 gev .
the reason for this is the interplay of the single - electron cuts and the increase in transverse flow , where the latter gives larger momentum to both electrons which in turn have a larger probability to make it into the acceptance . for the qgp contribution ,
the difference is smaller ( less than @xmath322 ) and weakly dependent on mass , since the increased transverse flow plays little role in the early phases ( we recall that , for the canonical choice of the formation time , @xmath324 fm/@xmath12 , the initial temperature for @xmath14 central pb - au amounts to @xmath513 mev , and @xmath514 mev for @xmath443 central ) .
the addition of in - medium @xmath2 and @xmath3 decays affects the total yield mostly around the respective free masses , more significantly in the @xmath3 region .
the in - medium four - pion contribution becomes relevant for masses @xmath515 gev , with a relative strength compared to the other sources that closely resembles the in - in case .
the drell - yan contribution does not show up on the scale of fig .
[ fig_ceres - semi ] ( and has been restricted to masses above @xmath111 gev ) .
correlated charm decays are not accounted for .
the lower panel of fig .
[ fig_ceres - semi ] summarizes the comparison of the theoretical calculations with the combined 1996/1996 ceres / na45 data @xcite , including the hadron decay cocktail .
the level of agreement is very similar to the one with our earlier results @xcite . ) '' the full in - medium @xmath1 spectral function has been replaced by the one which only includes medium effects in a meson gas.,scaledwidth=45.0% ] finally , we turn to the most recent ceres data for central pb(158 agev)-au collisions @xcite , which are shown in fig .
[ fig_ceres - centr ] in a form similar to the na60 data in the previous sections , _
i.e. _ , on a linear scale and with the cocktail subtracted . again , our calculations employing the updated fireball describe these data fairly well .
variations of @xmath516 in the fireball lifetime ( affecting the total dilepton yields at the same level with little change in the spectral shape ) , would still lead to reasonable agreement .
the importance of baryon - induced medium modifications at sps energies is reiterated by the dash - double - dotted curve where only meson - gas effects are included in the @xmath1 spectral function ; this scenario does not properly reproduce the experimental spectra . when comparing to the na60 dimuon mass spectra , the four - pion and qgp contributions in the imr appear smaller relative to the signal in the mass region below the free @xmath1 mass .
this is essentially due to the experimental acceptance , which for ceres / na45 is significantly larger in the low - mass region than for na60 , especially for low - momentum lepton pairs .
a particularly striking feature of the theoretical predictions is the large enhancement below the two - pion threshold , which is closely connected with the approach to ( and constraints from ) the photon point .
the very - low - mass excess is rather sensitive to the baryon - induced medium effects as well , as illustrated by the curve where the latter are switched off .
experimentally , this regime is only accessible with dielectrons ( and after subtraction of the @xmath517 dalitz decay ) , and the first hint from data is consistent with the theoretical yield .
clearly , it would be very worthwhile to further explore this region .
in the present paper , we have conducted a quantitative study of dilepton spectra in semi-/central heavy - ion collisions at full sps energy ( @xmath134 gev ) .
we have supplemented our earlier calculations of thermal radiation ( which reproduce available mass spectra ) with sources of non - thermal origin ( expected to become relevant at high transverse momentum and for small systems ) , and we have scrutinized the results with respect to hadro - chemical and flow properties of the underlying thermal fireball expansion . the key quantity to describe thermal dilepton rates is the ( imaginary part of the ) electromagnetic current correlator . in hadronic matter , we have evaluated its medium modifications in terms of many - body spectral functions for the light vector mesons at low mass , and chiral vector - axialvector mixing for the continuum part at intermediate mass .
partonic emission above @xmath32 has been approximated by perturbative quark - antiquark annihilation .
an appealing feature of this description is a smooth merging of the hadronic rates with the partonic ones at temperatures around the expected phase transition , @xmath518-@xmath30 mev . as a new ingredient
we have computed thermal dilepton rates induced by @xmath7-channel meson exchange in @xmath519 reactions , which we found to become relevant at momenta above @xmath11@xmath190 gev .
we have augmented our approach by non - thermal sources , _
i.e. _ , drell - yan annihilation , as well as decays of @xmath1 mesons from primordial production ( subject to jet quenching ) and from decoupling at thermal freezeout .
we have emphasized that these @xmath1 decays carry an extra kinematic lorentz-@xmath217 factor relative to thermal radiation , resulting in harder dilepton-@xmath5 spectra .
all sources have been implemented into a space - time evolution for @xmath187-@xmath187 collisions at sps , for which we employed an isentropic fireball model with qgp , mixed and hadron - gas phases and updated expansion parameters to better match the absolute dilepton yields measured by na60 .
hadrochemical freezeout and subsequent evolution are constructed in line with measured hadron multiplicities .
the calculated dimuon invariant - mass spectra agree well with na60 data in central and semicentral in - in collisions , including @xmath5 bins below @xmath520 gev and above @xmath470 gev , thus confirming our earlier results for thermal emission and validating the predicted in - medium vector spectral functions ( dominated , but not exhausted , by the @xmath1 ) .
when comparing to @xmath5 spectra , the new sources contribute to provide good agreement for central collisions , while discrepancies persist with semicentral data at @xmath521 for masses around and below the free @xmath1 mass .
in particular , effective slope parameters for @xmath522 amount to @xmath523-@xmath514 mev , @xmath11@xmath524 short of the empirically extracted values in the lmr .
we have investigated the sensitivity of our results to variations in the critical and chemical - freezeout temperature of the fireball .
the spectral shape of the invariant - mass spectra turned out to be insensitive for the range @xmath525 mev .
the reason is the `` duality '' of the thermal emission rates around @xmath32 , rendering the `` melting '' of the @xmath1 meson a robust signal .
a low value of @xmath518 mev entails that qgp radiation dominates over hadronic emission in the imr . a low value of @xmath348 mev implies smaller pion chemical potentials ( and thus higher temperatures ) in the hadronic evolution , increasing the effective slope parameters by about 15 - 20 mev , still @xmath1130 mev short of the data in the lmr .
in an attempt to resolve this discrepancy , we have implemented a @xmath11@xmath449 increase in the transverse fireball acceleration as previously employed for central au - au collisions at rhic energy .
this elevates the slope to @xmath11@xmath526 mev around the free @xmath1 mass ( and properly reproduces the increase below ) , while still being consistent with the ( upper ) experimental value of @xmath527 mev in the imr where the blue shift effect is less pronounced due to the prevalently early emission .
in particular , we have demonstrated that , within the current theoretical and experimental uncertainties , _ both _
scenarios for a predominant emission source in the imr
hadronic ( four - pion ) or partonic are viable . in either case , the radiation mostly emanates from matter at temperatures around @xmath528 mev .
an unambiguous distinction between a partonic and hadronic source in the imr therefore appears difficult at sps energies . the larger transverse flow developed in the partonic stage at rhic energies could facilitate this task @xcite , although additional complications arise due to the much larger open - charm contribution and its in - medium modifications .
finally we have checked that our refined assessment of dilepton sources , together with the improvements in the fireball expansion , preserves our previously found agreement with the ceres / na45 dielectron spectra in pb - au collisions .
the larger system size leads to an appreciable increase in thermal radiation , thus reducing the uncertainties associated with the modeling of the break - up stage .
in addition , dielectrons allow access to the mass region below the two - pion threshold , where the medium effects on the @xmath1-spectral function , augmented by the thermal bose factor ( and photon propagator ) , predict a large thermal dilepton signal .
first data in this region support this signal , and further corroborate the importance of baryonic interactions for the in - medium @xmath1 spectral function .
future efforts should be pursued along several directions : first and foremost , the theoretically calculated vector spectral functions must be extended to their chiral partners ( most notably the axialvector ( @xmath63 ) channel as the partner of the @xmath1 ) , to establish direct connections to chiral symmetry restoration .
the evaluation of in - medium weinberg sum rules @xcite , in connection with constraints from lattice qcd , will play an important role in this enterprise @xcite .
second , from the phenomenological side , the implementation of emission rates into both hydrodynamic and transport simulations should be reiterated , to compute @xmath5 spectra and check whether the rather large acceleration of the fireball model can be justified for the in - in system at sps .
the study of dilepton elliptic flow @xcite could shed further light on the time profile of the radiation . at rhic energies ,
partonic collectivity increases significantly ; an unexpectedly large dilepton excess in the lmr recently observed by phenix needs to be understood @xcite .
work in several of these directions is in progress .
we are grateful to s. damjanovic and h. j. specht for discussion and information on the na60 acceptance , and to a. marin for providing us with the new ceres data .
this work was supported in part by a u.s .
national science foundation career award under grant no
. phy-0449489 . | a quantitative evaluation of dilepton sources in heavy - ion reactions is performed taking into account both thermal and non - thermal production mechanisms .
the hadronic thermal emission rate is based on an electromagnetic current - correlation function with a low - mass region ( lmr , @xmath0 gev ) dominated by vector mesons ( @xmath1 , @xmath2 , @xmath3 ) and an intermediate - mass region ( imr , @xmath4 ) characterized by ( the onset of ) a multi - meson continuum .
a convolution of the emission rates over a thermal fireball expansion results in good agreement with experiment in the low - mass spectra , confirming the predicted broadening of the @xmath1 meson in hadronic matter in connection with the prevalence of baryon - induced medium effects .
the absolute magnitude of the lmr excess is mostly controlled by the fireball lifetime , which in turn leads to a consistent explanation of the dilepton excess in the imr in terms of thermal radiation .
the analysis of experimental transverse - momentum ( @xmath5 ) spectra reveals discrepancies with thermal emission for @xmath6 gev in noncentral in - in collisions , which we address by extending our calculations by : ( i ) a refined treatment of @xmath1 decays at thermal freezeout , ( ii ) primordially produced @xmath1 s subject to energy - loss , ( iii ) drell - yan annihilation , and ( iv ) thermal radiation from @xmath7-channel meson exchange processes .
we investigate the sensitivity of dilepton spectra to the critical temperature and hadro - chemical freezeout of the fireball .
the @xmath1 broadening in the lmr turns out to be robust , while in the imr quark - gluon plasma radiation is moderate unless the critical temperature is rather low . |
linchenko and montgomery @xcite introduced the @xmath5-th frobenius - schur ( fs ) indicator @xmath6 of a finite - dimensional representation @xmath7 of a semisimple hopf algebra by generalizing that for finite groups . this notion has been studied extensively and plays an important role in the study of semisimple hopf algebras ; see , _
e.g. _ , @xcite and references therein
. it also has a categorical generalization as well as some other notions of hopf algebra theory . recall that a monoidal category @xmath1 is said to be _ left rigid _
if every object @xmath8 has a left dual object @xmath9 , and a left rigid monoidal category @xmath1 is said to be _
pivotal _ if it is endowed with an isomorphism @xmath10 of monoidal functors .
based on the `` first formula '' @xcite of the fs indicator for semisimple hopf algebras , ng and schauenburg defined the @xmath3-th fs indicator @xmath11 of an object @xmath7 of a @xmath12-linear pivotal monoidal category @xmath13 to be the trace of the @xmath14-th power of a certain linear map @xmath15 if @xmath1 is the category of representations of a semisimple hopf algebra , then @xmath1 has a canonical pivotal structure and @xmath16 agrees with linchenko and montgomery s .
this generalization has some interesting applications to semisimple hopf algebras , fusion categories and conformal field theories ; see @xcite .
it is interesting to develop a theory of fs indicators for non - semisimple hopf algebras ( or , more generally , non - semisimple tensor categories ) .
if a given monoidal category would be pivotal , then fs indicators could be defined in the way of ng and schauenburg
. however , some interesting results have been obtained in `` non - pivotal '' settings : for example , jedwab s trace invariant @xcite can be considered as a generalization of the second fs indicator to a non - pivotal setting .
kashina , montgomery and ng s @xmath5-th indicator @xcite can be considered as a generalization of the @xmath5-th fs indicator of the regular representation to arbitrary finite - dimensional hopf algebras . for the above reasons , in this paper ,
we propose a generalization of fs indicators in `` non - pivotal '' settings .
our idea is very simple : going back to its definition , we find that we only need the @xmath7-th component of the pivotal structure @xmath10 to define .
now suppose that @xmath1 is a linear left rigid monoidal category .
then @xmath17 does not hold in general . if we are given a _ pair _
@xmath18 consisting of an object @xmath8 and an isomorphism @xmath19 , then we can define @xmath20 by replacing @xmath21 with @xmath22 in the definition of . for @xmath23 and @xmath24
, we define the @xmath3-th fs indicator of the pair @xmath25 by @xmath26 in the first half of this paper , we develop a category - theoretical framework to treat such a generalization of the fs indicators and explain how results of @xcite and @xcite relate to our framework . in the second half , based on our framework , we study the indicators of the `` adjoint object '' in a finite tensor category , which can be considered as a generalization of the adjoint representation of a finite - dimensional hopf algebra . now we describe the organization of this paper . in section [ sec : pre ] , for reader s convenience , we remember some definitions from the theory of monoidal categories and hopf monads . in section [ sec : duality - functor ] , we first recall from @xcite the definition and the basic properties of the _ left duality transformation _
@xmath27 for a strong monoidal functor @xmath28 between left rigid monoidal categories @xmath1 and @xmath29 .
we study its relation to adjunctions .
in particular , we observe that , if @xmath1 and @xmath29 are rigid , and @xmath30 is either left or right adjoint to @xmath31 , then there is a natural isomorphism @xmath32 given by using ( lemma [ lem : monoidal - adj - bidual ] ) . by a _ pivotal object _ in a left rigid monoidal category @xmath1 , we mean a pair @xmath33 consisting of an object @xmath7 and an isomorphism @xmath19 . in section
[ sec : piv - cov ] , we define the category @xmath0 of pivotal objects in @xmath1 . this category has a natural structure of a pivotal monoidal category and a certain universal property . thus we call it the _ pivotal cover _ of @xmath1 . by the universal property , a strong monoidal functor @xmath28 between left rigid monoidal categories induces a strong monoidal functor @xmath34 preserving the pivotal structure .
moreover , if @xmath1 and @xmath29 are rigid and @xmath31 has a left ( right ) adjoint functor @xmath35 , then @xmath36 has a left ( right ) adjoint functor given by @xmath37{\cong } i(v)^ { * * } \big).\ ] ] in section [ sec : fs - ind ] , we define the @xmath3-th fs indicator @xmath38 of a pivotal object @xmath25 in a rigid monoidal category over a field @xmath12 in the way as described above . after proving its basic properties ,
we explain how jedwab s trace invariant @xcite relate to our fs indicator ( [ sec : example - trace - inv ] ) .
we also show that kashina , montgomery and ng s @xmath5-th indicator @xmath39 of a finite - dimensional hopf algebra over @xcite can be expressed by using the above construction of adjunctions : we have @xmath40 where @xmath35 is a right adjoint of the fiber functor and @xmath41 is the unit object of the pivotal cover of the category of finite - dimensional vector spaces over @xmath12 ( [ subsec : fs - ind - examples ] ) .
as these examples illustrate , interesting pivotal objects are obtained by adjunctions .
now let @xmath1 be a finite tensor category @xcite .
the main topic of section [ sec : ind - adj - obj ] is the pivotal object @xmath42 obtained as the image of the unit object under @xmath43 where @xmath44 is the monoidal center of @xmath1 , @xmath45 is the forgetful functor , and @xmath35 is a right adjoint functor of @xmath46 .
as it generalizes the adjoint representation of a hopf algebra ( [ sec : adj - rep - hopf ] ) , we call @xmath47 ( or its underlying object @xmath48 ) the _ adjoint object_. one of its important properties is that there exists an isomorphism @xmath49 where @xmath50 is the functor @xmath51 and @xmath52 means the set of natural transformations from a functor @xmath31 to itself .
hence with @xmath53 induces an operator @xmath54 .
we analyze this operator deeply by using the techniques of the graphical calculus and hopf monads .
in particular , we show that the @xmath5-th power of this operator is given by @xmath55 this implies that the map @xmath56 is of finite order ( theorem [ thm : en - order ] ) and thus @xmath57 is a cyclotomic integer .
note that our @xmath38 is not a cyclotomic integer in general , while the fs indicator in the sense of ng and schauenburg @xcite always is . in section [ sec : adj - rep - hopf ]
, we apply our results to finite - dimensional hopf algebras .
we give a formula of @xmath58 for a finite - dimensional semisimple hopf algebra @xmath59 by using the normalized integral in @xmath59 .
this formula can not be generalized for non - semisimple @xmath59 .
thus it is an interesting problem to know the value of @xmath60 for non - semisimple @xmath59 .
we determine @xmath61 for certain non - semisimple and non - pivotal hopf algebras @xmath59 in a direct way and give some applications .
the study of the `` adjoint object '' was greatly motivated by a talk given by michihisa wakui at a workshop held at kyoto university in september 2012 .
the author thank him for valuable comments .
the author also thank the referee for careful reading of the manuscript and pointing out a number of errors in the previous version .
the author is supported by grant - in - aid for jsps fellows ( 24@xmath623606 ) .
a _ monoidal category _ is a data @xmath63 consisting of a category @xmath1 , a bifunctor @xmath64 ( called the tensor product ) , an object @xmath65 ( called the unit object ) , and natural isomorphisms @xmath66 satisfying the pentagon and the triangle axiom .
we refer the reader to @xcite for the basic theory of monoidal categories . the natural isomorphisms @xmath67 , @xmath68 and @xmath69 are called the associativity , the left unit , and the right unit isomorphism , respectively . note that if @xmath70 is a monoidal category , then @xmath71 are also monoidal categories , where @xmath72 for @xmath73 .
now let @xmath1 and @xmath29 be monoidal categories .
monoidal functor _ from @xmath1 to @xmath29 is a triple @xmath74 consisting of a functor @xmath28 , a natural transformation @xmath75 and a morphism @xmath76 in @xmath29 satisfying certain axioms ( * ? ? ?
if @xmath77 and @xmath78 are invertible , then @xmath31 is said to be _
strong_. suppose that @xmath31 and @xmath79 are monoidal functors from @xmath1 to @xmath29 .
monoidal natural transformation _ from @xmath31 to @xmath79 is a natural transformation @xmath80 such that @xmath81 for all @xmath73 . as is well - known , monoidal categories , monoidal functors and monoidal natural transformations form a 2-category .
an _ equivalence of monoidal categories _ ( or _ monoidal equivalence _ ) is defined as an equivalence in this 2-category .
formally , a _ comonoidal functor _ from @xmath1 to @xmath29 is defined to be a monoidal functor from @xmath82 to @xmath83 . in other words , it is a triple @xmath84 consisting of a functor @xmath28 , a natural transformation @xmath85 and a morphism @xmath86 in @xmath29 satisfying the axioms similar to those for monoidal functors . a _
comonoidal natural transformation _
is defined in a similar way to the monoidal case .
let @xmath1 be a monoidal category .
a _ left dual object _ of @xmath8 is a triple @xmath87 consisting of an object @xmath88 and morphisms @xmath89 ( called the _ evaluation _ ) and @xmath90 ( called the _ coevaluation _ ) satisfying @xmath91 suppose that a left dual object @xmath87 of @xmath7 is given . by using the evaluation and the coevaluation , one can establish isomorphisms @xmath92 natural in the variables @xmath73 ( * ? ? ?
* chapter xiv ) .
this means that the functors @xmath93 and @xmath94 are left adjoint functors of @xmath95 and @xmath96 , respectively .
the following lemma is essentially a consequence of the uniqueness of adjoint functors ( see , _
e.g. _ , ( * ? ? ?
* lemma 2.1.5 ) ) .
[ lem : left - dual - uniq ] suppose that @xmath97 and @xmath98 are left dual objects of @xmath7 .
then there exists an isomorphism @xmath99 characterized either of @xmath100 we say that @xmath1 is _ left rigid _ if every object of @xmath1 has a left dual object . in this paper ,
_ we always assume that a left dual object has been chosen for every object of a left rigid monoidal category_. the fixed left dual object of @xmath7 will be denoted by @xmath101 a _ right dual object _ of @xmath8 is a left dual object of @xmath102 .
we say that @xmath1 is _ right rigid _ if every object of @xmath1 has a right dual object , or , equivalently , @xmath103 is left rigid .
similarly to the above , we assume that a right dual object @xmath104 has been chosen for each object @xmath7 of a right rigid monoidal category .
we say that @xmath1 is _ rigid _ if it is both left and right rigid .
suppose that @xmath1 is a left rigid monoidal category .
then @xmath105 extends to a strong monoidal functor @xmath106 , called the _ left duality functor _ ( * ? ? ?
* xiv.2 ) . for a right rigid monoidal category ,
right duality functor _
@xmath107 is defined in a similar way .
let @xmath12 be a field . by a _ @xmath12-linear monoidal category _
, we mean a monoidal category @xmath1 such that * @xmath108 is a finite - dimensional vector space over @xmath12 for all @xmath73 , * the composition of morphisms is @xmath12-bilinear , and * the tensor product of morphisms is @xmath12-bilinear .
since the term `` tensor category '' has several different uses , we shall fix its meaning . in this paper , we follow the usage of etingof and ostrik @xcite .
thus a _ tensor category _ over @xmath12 is a @xmath12-linear abelian rigid monoidal category @xmath1 such that * every object of @xmath1 has finite length , and * the unit object @xmath65 is simple .
one of our particular interests is a _ finite tensor category _
@xcite , _ i.e. _ , a tensor category which is , as a @xmath12-linear category , equivalent to the category of finite - dimensional modules over a finite - dimensional algebra over @xmath12 .
let @xmath1 be a category .
a _ monad _ on @xmath1 is a triple @xmath109 consisting of an endofunctor @xmath110 on @xmath1 and natural transformations @xmath111 and @xmath112 satisfying @xmath113 for all @xmath8 . given a monad @xmath110 on @xmath1 , a _
@xmath110-module _ is a pair @xmath114 consisting of an object @xmath115 and a morphism @xmath116 , called the _ action _ , satisfying @xmath117 let @xmath114 and @xmath118 be two @xmath110-modules .
a morphism @xmath119 in @xmath1 is said to be _
@xmath110-linear _ if @xmath120 .
we denote by @xmath121 the category of @xmath110-modules and @xmath110-linear morphisms between them . in literature
( _ e.g. _ @xcite ) , a @xmath110-module is called a _
@xmath110-algebra _ and the category @xmath121 is referred to as the _ eilenberg - moore category _ of @xmath110 . now suppose that @xmath1 is a monoidal category .
a _ bimonad _ on @xmath1 is a monad @xmath109 on @xmath1 such that the functor @xmath122 is comonoidal and the natural transformations @xmath123 and @xmath124 are comonoidal . if @xmath110 is a bimonad on @xmath1 , then @xmath121 is a monoidal category : for @xmath110-modules @xmath114 and @xmath118 , their tensor product is defined by @xmath125 , where @xmath126 the pair @xmath127 is the unit object of @xmath121 .
the associativity and unit constraints of @xmath121 are defined so that the forgetful functor @xmath128 is a strict monoidal functor . a _
hopf monad _ on @xmath1 is a bimonad such that certain natural transformations , called the _ fusion operators _ , are invertible @xcite . if @xmath1 is a rigid monoidal category , then the notions of a left antipode @xmath129 and a right antipode @xmath130 for a bimonad @xmath110 on @xmath1 are defined as natural transformations @xmath131 satisfying certain axioms . a hopf monad on @xmath1 is characterized as a bimonad @xmath110 on @xmath1 such that a left and a right antipode for @xmath110 exist @xcite .
if @xmath110 is a hopf monad on a rigid monoidal category @xmath1 , then the monoidal category @xmath121 is rigid ( * ? ? ?
* theorem 3.8 ) .
more precisely , given a @xmath110-module @xmath114 , we define @xmath132 , where @xmath133 the axioms for a left antipode ensure that the triple @xmath134 is a left dual object of @xmath114 .
a right dual object of @xmath135 is given in a similar way but by using the right antipode @xmath130 .
let @xmath28 be a strong monoidal functor between left rigid monoidal categories @xmath1 and @xmath29 . for each @xmath8
, we set @xmath136 then the triple @xmath137 is a left dual object of @xmath138 .
hence , by lemma [ lem : left - dual - uniq ] , there exists a unique isomorphism @xmath139 characterized either of @xmath140 the family @xmath141 is in fact an isomorphism of strong monoidal functors @xcite .
following @xcite , we call @xmath142 the _ duality transformation _ for @xmath31
. the aim of this section is to collect basic properties of the duality transformation and the natural isomorphism @xmath143 defined by @xmath144 [ lem : dual - trans - composition ] if @xmath145 is a sequence of strong monoidal functors between left rigid monoidal categories , then , for all @xmath8 , we have @xmath146 the second equation follows immediately from the first .
hence we prove the first one . for @xmath8 ,
we compute : @xmath147 where @xmath148 and @xmath149 are the monoidal structure of @xmath150 .
now the first equation follows from the definition of the duality transformation .
the following lemma is just a rephrasing of ( * ? ? ?
* proposition 7.1 ) by using the duality transformation ; see also ( * ? ? ?
* appendix a ) .
[ lem : dual - trans - mon - nat - tr ] let @xmath151 be strong monoidal functors between left rigid monoidal categories @xmath1 and @xmath29 , and let @xmath80 be a monoidal natural transformation
. then @xmath152 is invertible and we have @xmath153 for all @xmath8 .
hence , @xmath154 for a strong monoidal functor @xmath28 between right rigid monoidal categories , the _
right duality transformation _
@xmath155 is defined in a similar way .
since it is the left duality transformation of the strong monoidal functor @xmath156 induced by @xmath31 , the right duality transformations can be treated in a similar manner to the left duality transformations . a _ pivotal monoidal category _
is a left rigid monoidal category @xmath1 endowed with a monoidal natural transformation @xmath157 . since the duality transformation for @xmath158 is the identity
, @xmath159 is invertible and @xmath160 ( * ? ? ? * appendix a ) .
a pivotal monoidal category is rigid since the triple @xmath161 is a right dual object of @xmath7 .
let @xmath28 be a strong monoidal functor between pivotal monoidal categories @xmath1 and @xmath29 .
we say that _ @xmath31 preserves the pivotal structure _
@xcite if @xmath162 for all @xmath8 , where @xmath159 s are the pivotal structures of @xmath1 and @xmath29 .
the following result follows from lemma [ lem : dual - trans - composition ] .
[ lem : piv - preserve - func - compose ] let @xmath145 be a sequence of strong monoidal functors between pivotal monoidal categories . if @xmath31 and @xmath79 preserve the pivotal structure , then so does their composition @xmath150 .
following mac lane ( * ? ? ?
* iv.1 ) , we write @xmath163 if @xmath28 is a functor , @xmath164 is right adjoint to @xmath31 , and @xmath165 and @xmath166 are the unit and the counit of the adjunction .
the following lemma is well - known ( see , _
e.g. _ , ( * ? ? ?
* proposition 3.84 ) ) : [ lem : monoidal - adj-1 ] let @xmath167 be an adjunction between monoidal categories @xmath1 and @xmath29 . if @xmath79 is monoidal , then @xmath31 is comonoidal by @xmath168 and @xmath169 .
similarly , if @xmath31 is comonoidal , then @xmath79 is monoidal . by using the duality transformation ,
we prove : [ lem : monoidal - adj-2 ] let @xmath170 be a strong monoidal functor between rigid monoidal categories .
suppose that @xmath31 has a left or a right adjoint functor and let @xmath35 be one of them
. then there are isomorphisms @xmath171 natural in the variables @xmath172 and @xmath8 .
we consider the case where @xmath35 is left adjoint to @xmath31 . for all @xmath173 and @xmath8 , there are the following natural isomorphisms : @xmath174 hence @xmath175 by yoneda s lemma .
the other cases are proved in a similar way .
the same proof applies in the case where @xmath1 and @xmath29 are closed ( _ i.e. _ , admits an internal hom ) and @xmath31 preserves internal hom s .
let @xmath28 be as in lemma [ lem : monoidal - adj-2 ] . by emphasizing the use of hopf monadic techniques , natural isomorphisms in that lemma
are obtained as follows : we first consider the case where @xmath35 is left adjoint to @xmath31 .
let @xmath176 be the counit of the adjunction .
by lemma [ lem : monoidal - adj-1 ] , @xmath35 is comonoidal .
moreover , the pair @xmath177 is a hopf adjunction by ( * ? ? ?
* proposition 3.5 ) .
hence the left fusion operator @xmath178 for this hopf adjunction , defined by @xmath179 for @xmath8 and @xmath172 , is a natural isomorphism .
similarly , the right fusion operator gives a natural isomorphism @xmath180 .
if @xmath35 is right adjoint to @xmath31 , then the functor @xmath181 induced by @xmath35 is left adjoint to @xmath182 . applying the above arguments to the pair @xmath183 , we obtain desired natural isomorphisms . for a functor @xmath184 between rigid monoidal categories , we set @xmath185 .
[ lem : monoidal - adj - dual ] let @xmath28 be a strong monoidal functor between rigid monoidal categories . if @xmath31 has a left adjoint @xmath186 , then @xmath187 is right adjoint to @xmath31 .
similarly , if @xmath31 has a right adjoint @xmath188 , then @xmath189 is left adjoint to @xmath31 . in other words , if one of a left adjoint @xmath190 or a right adjoint @xmath188 of @xmath31 exists , then the other one exists and there are natural isomorphisms @xmath191 hence we obtain natural isomorphisms @xmath192 one can obtain explicit descriptions of the above isomorphisms by going back to the proof of ( * ? ? ?
* lemma 3.5 ) . however , instead of doing so , we prove the following assertion directly : [ lem : monoidal - adj - bidual ] let @xmath28 be a strong monoidal functor between rigid monoidal categories .
if @xmath193 is an adjunction , then the composition @xmath194 is an isomorphism .
similarly , if @xmath195 is an adjunction , then @xmath196 is an isomorphism .
there is the following sequence of natural isomorphisms : @xmath197 let @xmath198 be the composition of the above isomorphisms with @xmath199 . then @xmath200 .
note that the functor @xmath158 is an equivalence , since we have assumed @xmath1 to be rigid .
thus , by yoneda s lemma , @xmath201 is invertible .
the invertibility of @xmath202 follows from the same argument applied to @xmath203 .
suppose moreover that @xmath1 and @xmath29 are pivotal monoidal categories and @xmath31 preserves the pivotal structure in lemma [ lem : monoidal - adj - bidual ] .
the following lemma says that adjoint functors of such an @xmath31 , in a sense , preserves the pivotal structure . by abuse of notation
, we denote the pivotal structures of @xmath1 and @xmath29 by the same symbol @xmath159 .
then : [ lem : piv - preserve - func - adj ] for all @xmath8 , we have @xmath204 consider the following diagram : @xmath205^{i_{\ell}(\eta^ { * * } ) } & i_{\ell}(f i_{\ell}(v)^ { * * } ) \ar[r]^{i_{\ell}(\xi^{-1 } ) } & i_{\ell } f ( i_{\ell}(v)^ { * * } ) \ar[r]^{\varepsilon } & i_{\ell}(v)^ { * * } \\
i_{\ell}(v ) \ar[r]^{i_{\ell}(\eta ) } \ar[u]^{i_{\ell}({\mathsf{p } } ) } & i_{\ell } f
i_{\ell}(v ) \ar[r]^{{\mathrm{id } } } \ar[u]^{i_{\ell}({\mathsf{p } } ) } & i_{\ell } f i_{\ell}(v ) \ar[r]^{\varepsilon } \ar[u]_{i_{\ell } f({\mathsf{p } } ) } & i_{\ell}(v ) \ar[u]_{{\mathsf{p } } } } \ ] ] the left and the right square commute by the naturality of @xmath159 and @xmath206 , respectively , and so does the middle square since @xmath31 preserves the pivotal structure .
the composition along the top and the bottom row are the morphism @xmath201 and the identity , respectively .
therefore we get the first equation .
the second one is obtained in a similar way .
for a strong monoidal endofunctor @xmath207 on a monoidal category @xmath1 , we define the category @xmath208 of `` objects fixed by @xmath31 '' as follows : an object of @xmath208 is a pair @xmath209 consisting of an object @xmath8 and an isomorphism @xmath210 in @xmath1 .
the set of morphisms is defined by @xmath211 for @xmath212 , and the composition of morphisms is defined in an obvious way .
note that @xmath208 is a monoidal category with tensor product @xmath213 and unit @xmath214 .
if , moreover , @xmath1 is left rigid , then so is @xmath208 .
indeed , @xmath215 is a left dual object of @xmath216 ( the fact that @xmath217 and @xmath218 are morphisms in @xmath208 follows from the definition of the duality transformation @xmath142 ) .
now we apply the above construction to the double - dual functor @xmath158 : for a left rigid monoidal category @xmath1 , we denote by @xmath0 the left rigid monoidal category @xmath208 with @xmath219 and call it the _ pivotal cover _ of @xmath1 . an object of @xmath0 will be referred to as a _ pivotal object_. in this paper , a pivotal object
is usually denoted in bold letters . unless otherwise noted , given a pivotal object ,
say @xmath25 , its underlying object is denoted by the corresponding letter @xmath7 and the isomorphism @xmath220 is denoted by @xmath221 .
to justify our terminology , we first prove the following theorem : @xmath0 is a pivotal monoidal category with pivotal structure @xmath222 let @xmath2 . since the duality transformation for @xmath158 is the identity ( * ? ? ?
* appendix a ) , the left dual object of @xmath25 is given by @xmath223 in particular , @xmath224 . from this
, one easily sees that @xmath225 is indeed an isomorphism in @xmath0 . for a morphism @xmath226 in @xmath0
, we have @xmath227 _ i.e. _ , @xmath228 is a natural isomorphism . that @xmath159 is a monoidal natural transformation follows directly from the definition of the tensor product of @xmath0 .
we define the forgetful functor @xmath229 by @xmath230 for @xmath2 .
this functor is faithful and strict monoidal .
the pivotal cover has the following universal property : [ thm : piv - cov - univ ] let @xmath1 be a pivotal monoidal category , and let @xmath29 be a left rigid monoidal category . for every strong monoidal functor @xmath28 , there uniquely exists a strong monoidal functor @xmath231 preserving the pivotal structure such that @xmath232 as monoidal functors define a functor @xmath231 by @xmath233 for all objects @xmath8 and all morphisms @xmath234 in @xmath1 , and endow this functor with the monoidal structure given by @xmath235 one can check that @xmath236 so defined satisfies the required conditions . hence the existence is proved .
now we show the uniqueness .
let @xmath236 be a functor satisfying the required conditions . by , we have @xmath237 for all morphisms @xmath234 in @xmath1 , and @xmath238 for some @xmath239 if @xmath8 is an object .
since the duality transformation of @xmath240 is the identity , we have @xmath241 for all @xmath8 . since @xmath236 preserves the pivotal structure , @xmath242 where @xmath159 s are the pivotal structures of @xmath1 and @xmath243 .
hence @xmath236 is given by .
it is obvious from that the monoidal structure of @xmath236 is given by .
[ rem : piv - cov - pdim ] for an endomorphism @xmath244 in a pivotal monoidal category with pivotal structure @xmath159 , the left and the right pivotal trace of @xmath234 are the endomorphisms of the unit object defined and denoted by @xmath245 respectively .
we call @xmath246 and @xmath247 the left and the right pivotal dimension of @xmath7 , respectively . by the definition of the pivotal structure of @xmath0
, we have @xmath248 for all @xmath2 . given a hopf algebra @xmath59 with bijective antipode @xmath129 , we define the hopf algebra @xmath249 to be the semidirect product @xmath250 , where @xmath251 is ( the group algebra of ) the infinite cyclic group with generator @xmath252 acting on @xmath59 as @xmath253 .
if @xmath7 is a finite - dimensional left @xmath249-module , then @xmath254 defines an @xmath59-linear isomorphism @xmath255 .
the assignment @xmath256 gives an isomorphism @xmath257 of monoidal categories .
observe that @xmath249 is infinite - dimensional even if @xmath59 is finite - dimensional .
thus one might think that the pivotal cover is often `` too big '' and some `` smaller '' category should be considered .
for example , a _
@xcite of a fusion category @xmath1 , which is effectively used in their study of fusion categories , is a full subcategory of the pivotal cover @xmath0 .
as such an example illustrates , some smaller subcategories of @xmath0 are often more convenient than @xmath0 itself . however , in this paper , we consider the whole @xmath0 since its universal property will play an important role in what follows .
now we investigate functorial properties of our construction .
first , suppose that @xmath28 is a strong monoidal functor between left rigid monoidal categories . applying theorem [ thm : piv - cov - univ ] to @xmath258 , we get a unique strong monoidal functor @xmath34 such that @xmath36 preserves the pivotal structure and @xmath259 as monoidal functors . by the proof of theorem [ thm : piv - cov - univ ] , @xmath36 is given explicitly by @xmath260 for @xmath2 .
if , in addition , @xmath261 is another strong monoidal functor between left rigid monoidal categories , then we have @xmath262 by theorem [ thm : piv - cov - univ ] ( or , more directly , by lemma [ lem : dual - trans - composition ] ) .
fix left rigid monoidal categories @xmath1 and @xmath29 . for a monoidal natural transformation @xmath80 between strong monoidal functors @xmath151 , we set @xmath263 by lemma [ lem : dual - trans - mon - nat - tr ] , we see that @xmath264 is in fact a monoidal natural transformation .
it is obvious that the assignment @xmath265 preserves the vertical composition and the horizontal composition of monoidal natural transformations .
now we introduce two 2-categories @xmath266 and @xmath267 as follows : * the 0-cells , 1-cells and 2-cells of @xmath266 are left rigid monoidal categories , strong monoidal functors , and monoidal natural transformations , respectively . *
the 0-cells , 1-cells and 2-cells of @xmath267 are pivotal monoidal categories , pivotal - structure - preserving strong monoidal functors , and monoidal natural transformations , respectively ( see lemma [ lem : piv - preserve - func - compose ] ) . in both cases ,
the composition of cells is defined in the same way as the 2-category of monoidal categories .
summarizing our results so far , we obtain : @xmath268 is a 2-functor . the following corollary , which is conceptually obvious , is rigorously proved by using the functoriality of the pivotal cover . if @xmath28 is a monoidal equivalence between left rigid monoidal categories , then @xmath269 is a monoidal equivalence preserving the pivotal structure .
[ rem : piv - cov-2-adj ] let @xmath270 be the 2-functor forgetting the pivotal structure .
the pivotal cover @xmath271 is a left adjoint of @xmath272 in the following sense : for a pivotal monoidal category @xmath1 , we define @xmath273 where @xmath159 is the pivotal structure of @xmath1 .
@xmath274 is a strong monoidal functor and preserves the pivotal structure .
then , for all 0-cells @xmath275 and @xmath276 , we have @xmath277 where the forgetful functor @xmath278 for @xmath279 is regarded as a strong monoidal functor @xmath280 .
define the 2-category @xmath281 of right rigid monoidal categories in the same way as @xmath266 .
the assignment @xmath282 extends to isomorphisms @xmath283 of 2-categories .
one can define the pivotal cover @xmath284 for right rigid monoidal categories by @xmath285 here we show that an adjoint of a strong monoidal functor @xmath31 induces an adjoint of @xmath36 . in the following theorem , we use the same notations as in lemma [ lem : monoidal - adj - bidual ] .
[ thm : piv - cov - induced - adj ] let @xmath28 be a strong monoidal functor between rigid monoidal categories .
if @xmath286 is an adjunction , then the functor @xmath287 is left adjoint to @xmath36 .
similarly , if @xmath195 is an adjunction , then @xmath288 is right adjoint to @xmath36 .
we only show that @xmath289 is left adjoint to @xmath36 .
define @xmath290 for @xmath291 and @xmath2 , respectively .
then @xmath292 is a morphism in @xmath243 . indeed , writing @xmath293 , we compute @xmath294 it is obvious from the definition that @xmath295 is a natural transformation .
one can show that @xmath296 is a natural transformation in a similar way .
hence we obtain an adjunction @xmath297 .
[ rem : f - piv - frobenius ] we use the same notations as in the above theorem
. one can apply results in [ subsec : dual - adj ] to @xmath289 and @xmath298 .
for example , we have natural isomorphisms @xmath299 by lemma [ lem : monoidal - adj - dual ] .
in particular , @xmath300 , where @xmath41 is the unit object .
hence , if @xmath289 and @xmath301 would be isomorphic as functors , then @xmath302 would be self - dual .
a functor @xmath31 is said to be _ frobenius _ @xcite if it has a left adjoint functor which is right adjoint to @xmath31 . based on the above argument
, we will give an example of a strong monoidal functor @xmath31 such that @xmath31 is frobenius but @xmath36 is not ( [ subsec : fs - ind - examples ] ) .
in a series of papers @xcite , ng and schauenburg introduced and studied the @xmath3-th frobenius - schur ( fs ) indicator for an object @xmath7 of a @xmath12-linear pivotal monoidal category , which we will denote by @xmath303 to avoid conflict of notation . in this section , we introduce the @xmath304-th fs indicator @xmath38 for a pivotal object in a @xmath12-linear rigid monoidal category @xmath1 and investigate their basic properties .
let @xmath1 be a left rigid monoidal category .
given @xmath305 , we write @xmath306 and set @xmath307 with @xmath308 . for an object @xmath2 and a positive integer @xmath5 , we define a bijection @xmath309 as follows : first , we recall that there is a natural isomorphism @xmath310 for @xmath311 ; see ( * ? ? ?
* definition 3.3 ) .
the map @xmath312 is defined by @xmath313 for @xmath314 , where @xmath315 is the isomorphism obtained by using the associativity constraint ( which is uniquely determined by mac lane s coherence theorem ) . strictly speaking
, this definition does not make sense in the case where @xmath316 .
however , modifying the definition in an obvious way , we obtain @xmath317 where @xmath318 is the monoidal structure of @xmath158 .
now let @xmath1 be a @xmath12-linear monoidal category . then , since the map @xmath312 is @xmath12-linear , we can make the following definition : let @xmath2 be a pivotal object , and let @xmath5 and @xmath14 be integers . if @xmath319 , then we define the _
@xmath3-th fs indicator _ of @xmath25 by @xmath320 if @xmath18 , then @xmath38 is also written as @xmath321 . it is convenient to extend the @xmath3-th fs indicator @xmath322 for all @xmath323 . motivated by the definition of the generalized fs indicator @xcite , we define @xmath324 for @xmath2 and set @xmath325 for @xmath326
. now one can uniquely extend @xmath322 for all integers @xmath5 and @xmath14 so that @xmath327 holds for all @xmath323 . by definition
, we have @xmath328 for all @xmath329 , @xmath323 and @xmath330 . if @xmath1 is pivotal with pivotal structure @xmath159 , then the map @xmath312 with @xmath331 is nothing but the map used in @xcite to define @xmath332 .
hence we get @xmath333 for all @xmath8 and integers @xmath5 and @xmath14 . in our setting
, an isomorphism @xmath255 does not need to be a component of a pivotal structure .
this is a small but crucial difference between our theory and @xcite .
now let @xmath28 be a @xmath12-linear strong monoidal functor between @xmath12-linear left rigid monoidal categories . for each @xmath334
, there is a canonical isomorphism @xmath335 obtained by using the monoidal structure of @xmath31 in an obvious way . in a similar way as @xcite , we can prove the following theorem : let @xmath2 .
then the map @xmath336 makes the following diagram commute : @xmath337 as an immediate consequence , we obtain : [ thm : fs - ind - invariance ] let @xmath28 be a @xmath12-linear strong monoidal functor between @xmath12-linear left rigid monoidal categories . if @xmath31 is fully faithful , then we have @xmath338 for all integers @xmath5 and @xmath14 and all objects @xmath2 . by theorem [ thm : fs - ind - invariance ] and mac lane s coherence theorem , we may assume that @xmath1 is strict when we study general properties of the fs indicators .
moreover , by the following lemma , we may also assume that the duality functor is strict : [ lem : strict - rigid ] for every rigid monoidal category @xmath1 , there exists a rigid monoidal category @xmath339 satisfying the following conditions : * @xmath1 is equivalent to @xmath339 as a monoidal category . * @xmath339 is a strict monoidal category . *
the left duality functor @xmath340 for @xmath339 is a strict monoidal functor . *
the right duality functor @xmath107 for @xmath339 is the inverse of @xmath340 .
our proof is based on the argument of ng and schauenburg @xcite , where a strictification of a pivotal monoidal category is considered .
let @xmath1 be a rigid monoidal category . without loss of generality
, we may assume that @xmath1 is a strict monoidal category .
we define a strict monoidal category @xmath339 as follows : objects of @xmath339 are sequences of length @xmath341 , like @xmath342 where @xmath343 s are objects of @xmath1 and @xmath344 s are integers . for such an @xmath345 , we set @xmath346 = x_1^{*n_1 } \otimes \dotsb \otimes x_r^{*n_r},\ ] ] where @xmath347 , @xmath348 if @xmath319 , and @xmath349 if @xmath350 . for the sequence @xmath351 of length zero , we understand @xmath352 = { \mathbbm{1}}$ ] . for objects @xmath353 , the set of morphisms between them is defined by @xmath354 , [ \mathbb{y}]),\ ] ] and the composition of morphisms in @xmath339 is defined by the composition in @xmath1 .
it is easy to see that the category @xmath339 is a strict monoidal category with tensor product given by the concatenation of sequences and unit object @xmath355 .
the assignment @xmath356 $ ] naturally extends to a strong monoidal functor from @xmath339 to @xmath1 .
this functor is in fact a monoidal equivalence , since it is fully faithful and surjective on objects .
hence @xmath339 fulfills the conditions ( s1 ) and ( s2 ) .
we now choose a left and a right dual object for each object of @xmath339 so that the conditions ( s3 ) and ( s4 ) are satisfied . for @xmath357 as in , we set @xmath358 and define @xmath359 and @xmath360 inductively as follows : if @xmath351 is the object of length zero , then set @xmath361 .
if @xmath362 is an object of @xmath339 of length one , then set @xmath363 if @xmath364 for some @xmath365 such that the morphisms @xmath366 , @xmath367 , @xmath368 and @xmath369 have been defined , then set @xmath370 one can check that the above rule defines @xmath371 and @xmath372 consistently and the triple @xmath373 is a left dual object of @xmath345 . by ,
the left duality functor for @xmath339 is strict , _
i.e. _ , the condition ( s3 ) holds . finally , we set @xmath374 for @xmath357 as in .
the condition ( s4 ) is satisfied if we choose the triple @xmath375 as a right dual object for @xmath357 .
now let @xmath1 be a @xmath12-linear rigid monoidal category satisfying the conditions ( s1)(s4 ) of the above lemma .
then we can freely use the graphical technique to express morphisms in @xmath1 .
our convention is that the source of a morphism is placed on the top of the diagram , and the target is placed on the bottom .
for example , @xmath376 represents the tensor product @xmath377 of @xmath378 and @xmath379 . the evaluation and the coevaluation
are expressed by @xmath380 respectively . and
@xmath381 for pivotal objects @xmath382 , we define @xmath383 by figure [ fig : def - e - map ] . by using their graphical expressions
, one can check that the diagram @xmath384 commutes .
note that the map @xmath385 used to define the fs indicator coincides with the map @xmath386 with @xmath387 . as a consequence , for all @xmath323 such that @xmath319 , we have @xmath388 where @xmath389 with @xmath387 .
let @xmath1 be a @xmath12-linear left rigid monoidal category .
for an integer @xmath5 and an object @xmath8 such that @xmath390 we choose an isomorphism @xmath19 and set @xmath391 @xmath392 does not depend on the choice of @xmath22 by .
now we consider the case where @xmath393 for some finite - dimensional hopf algebra @xmath59 over @xmath12 . for a self - dual absolutely simple
left @xmath59-module @xmath7 , _ jedwab s trace invariant _
@xmath394 @xcite is defined to be the trace of @xmath395 where @xmath396 is the antipode of @xmath59 and @xmath397 is the annihilator of @xmath7 .
now we claim : @xmath398 .
let @xmath399 be a basis of @xmath7 .
we denote by @xmath400 and @xmath401 the basis of @xmath9 and @xmath402 dual to @xmath403 and @xmath400 , respectively .
fix an isomorphism @xmath404 of @xmath59-modules and let @xmath405 be the matrix representing @xmath406 with respect to the basis @xmath403 and @xmath400 .
then @xmath394 is equal to the trace of @xmath407 , where @xmath408 means the matrix transpose ( see @xcite ) .
now we set @xmath18 , where @xmath409 .
note that the matrix @xmath410 represents the map @xmath22 with respect to the basis @xmath403 and @xmath401 . by remark [ rem : piv - cov - pdim ]
, we have @xmath411 to compute @xmath412 , we consider the following map : @xmath413 by the graphical expression of @xmath414 , we see that @xmath110 is conjugate to @xmath414 via the canonical isomorphism @xmath415 . since @xmath416 , and since @xmath417 is one - dimensional , we have @xmath418 we now conclude @xmath419 .
let @xmath28 be a @xmath12-linear fully faithful strong monoidal functor between left rigid monoidal categories @xmath1 and @xmath29 . if @xmath8 is an object satisfying , then so does @xmath138 . by the invariance of the fs indicators , we have : let @xmath28 be as above .
then @xmath420 for all @xmath421 and all @xmath8 satisfying .
the case where @xmath422 recovers ( * ? ? ? * theorem 4.1 ) .
let @xmath59 be a finite - dimensional hopf algebra over a field @xmath12 with comultiplication @xmath423 and antipode @xmath129 .
recall that the _ convolution product _ @xmath424 in @xmath425 is defined by @xmath426 where @xmath427 is the comultiplication of @xmath428 in the sweedler notation .
let @xmath5 be a positive integer .
the _ @xmath5-th indicator _ of @xmath59 , introduced by kashina , montgomery and ng in @xcite , is defined by @xmath429 where @xmath430 means the @xmath431-th power with respect to @xmath424 .
note that the identity map is invertible with respect to @xmath424 ( with inverse @xmath129 ) .
hence the @xmath5-th indicator can be defined for all integers @xmath5 by the same formula as above .
for example , @xmath432 the aim of this subsection is to explain that the @xmath5-th indicator can be expressed as the @xmath433-th fs indicator of a certain pivotal object .
we consider the map @xmath434 then @xmath435 is an object of @xmath436 ( if we regard @xmath59 as a left @xmath59-module by the multiplication from the left ) .
our claim is : [ thm : fs - ind - and - kmn - ind ] @xmath437 for all integers @xmath5 .
write @xmath438 . for @xmath439 ,
the claim is proved by remark [ rem : piv - cov - pdim ] , as follows : @xmath440 for @xmath319 , we define @xmath441 by @xmath442 for @xmath443 and @xmath444 .
we have showed that @xmath445 is equal to the trace of this map in ( * ? ? ?
* lemma 3.2 ) . on the other hand , this map coincides with the map @xmath446 introduced in
[ subsec : graph - pres ] .
hence , by , we conclude @xmath447 finally , we suppose that @xmath350 . set @xmath448 , and let @xmath449 be a non - zero left integral .
we recall from ( * ? ? ?
* proposition 3.13 ) that @xmath450 note that @xmath451 . by (
* proposition 3.16 ) and the above argument , we have @xmath452 by , @xmath453 is invertible .
hence , if we could prove @xmath454 then , by , and , we would have @xmath455 let us prove to complete the proof . following radford @xcite ,
the map @xmath456 is an isomorphism of left @xmath59-modules .
we will show that @xmath457 is in fact an isomorphism from @xmath458 to @xmath459 in @xmath436 , that is , @xmath460 for @xmath461 , we have @xmath462 this implies @xmath463 . by
, we also have @xmath464 by the above results , is proved as follows : @xmath465 the following assertions are equivalent : 1 .
@xmath448 is equal to @xmath466 .
@xmath467 in @xmath436 .
indeed , by , it is obvious that ( 1 ) implies ( 2 ) . to show that the converse holds
, we note that every @xmath59-linear map @xmath468 is of the form @xmath469 for some @xmath470 . for @xmath470
, one can check that : @xmath471 thus if @xmath472 and @xmath473 are isomorphic in @xmath436 , then there exists an invertible element @xmath470 satisfying the above equivalent conditions . applying the counit and substituting @xmath474 , we conclude that such an element @xmath475 does not exist unless @xmath476 .
therefore ( 2 ) implies ( 1 ) .
let @xmath477 denote the category of finite - dimensional vector spaces over @xmath12 , and let @xmath478 be the forgetful functor .
then @xmath479 is a left adjoint functor of @xmath46 with unit and counit given by @xmath480 the left adjoint functor of @xmath481 constructed in theorem [ thm : piv - cov - induced - adj ] is given by @xmath482 for @xmath483 .
hence , in particular , @xmath484 , where @xmath41 is the unit object of @xmath485 .
equation in introduction is obtained by theorem [ thm : fs - ind - and - kmn - ind ] and .
note that @xmath46 is frobenius , since a finite - dimensional hopf algebra is a frobenius algebra .
on the other hand , as we have seen , @xmath472 is not self - dual in general
. thus the functor @xmath481 is not frobenius in general ( _ cf_. remark [ rem : f - piv - frobenius ] ) .
let @xmath59 and @xmath486 be two finite - dimensional hopf algebras over @xmath12 .
suppose that there exists an equivalence @xmath487 of @xmath12-linear monoidal categories .
we are interested in the following question : @xmath488 a positive answer to this question , together with theorem [ thm : fs - ind - invariance ] , would give a conceptual explanation why the @xmath5-th indicator is a gauge invariant @xcite .
we conclude this section by noting the following result : the answer to is yes , provided that @xmath59 and @xmath486 are semisimple and @xmath12 is of characteristic zero . for an object @xmath7 of a pivotal monoidal category , set @xmath489 , where @xmath490 is the pivotal structure .
let @xmath31 be as above .
since @xmath491 as @xmath486-modules , and since @xmath31 preserves the canonical pivotal structure ( * ? ? ?
* corollary 6.2 ) , we have @xmath492
given a ( strict ) monoidal category @xmath1 , we denote by @xmath44 the center of @xmath1 . thus , an object of @xmath44 is a pair @xmath493 consisting of an object @xmath115 and a _ half - braiding _ for @xmath135 , _
i.e. _ , a natural isomorphism @xmath494 satisfying @xmath495 for all @xmath311 , and a morphism @xmath496 in @xmath44 is a morphism @xmath497 in @xmath1 compatible with the half - braiding .
we note that if @xmath1 is rigid , then so is @xmath44 .
more precisely , one can choose a left dual object for each @xmath498 so that the duality transformation of the forgetful functor @xmath499 is the identity .
let @xmath1 be a finite tensor category over an algebraically closed field @xmath12 .
then the functor has a right adjoint functor @xcite .
hence , by theorem [ thm : piv - cov - induced - adj ] , @xmath481 has a right adjoint functor , say @xmath298 .
we now set @xmath500 and write it as @xmath42 or @xmath501 if @xmath1 is clear . the isomorphism class of @xmath47 and that of @xmath48 do not depend on the choice of @xmath188 .
we call @xmath47 ( or its underlying object @xmath48 ) the _ adjoint object _ of @xmath1 , since , as we will see in [ subsec : adj - obj - hopf - case ] , @xmath48 is the adjoint representation of @xmath59 if @xmath393 for some finite - dimensional hopf algebra @xmath59 .
the aim of this section is to investigate the properties of the fs indicators of the adjoint object .
we first remark the following relation between @xmath502 and @xmath503 , the @xmath3-th fs indicator in the sense of ng and schauenburg @xcite .
[ thm : adj - obj - ind - ns ] suppose that @xmath1 has a pivotal structure
. then we have @xmath504 for all integers @xmath5 and @xmath14 .
thus , @xmath503 does not depend on the choice of the pivotal structures of @xmath1 .
let @xmath159 denote the pivotal structure of @xmath1
. then @xmath44 has a unique pivotal structure , which we denote by the same symbol @xmath159 , such that the forgetful functor @xmath46 preserves the pivotal structure . by lemma [ lem : piv - preserve - func - adj ]
, we have @xmath505 where @xmath506 is the natural isomorphism given in lemma [ lem : monoidal - adj - bidual ] and @xmath318 is the monoidal structure of @xmath158 .
hence the result follows from .
the indicators of the adjoint object have the following gauge invariance : [ thm : adj - obj - invariance ] let @xmath1 and @xmath29 be finite tensor categories over @xmath12 .
if there exists an equivalence @xmath28 of @xmath12-linear monoidal categories , then we have @xmath507 for all integers @xmath5 and @xmath14 .
let @xmath508 and @xmath509 denote the forgetful functors .
then there exists an equivalence @xmath510 of braided monoidal categories such that @xmath511 . applying the 2-functor @xmath512 ,
we have @xmath513 fix a right adjoint @xmath514 of @xmath515 and a quasi - inverse @xmath516 of @xmath31 . by
, @xmath517 is right adjoint to @xmath518 .
hence we have isomorphisms @xmath519 in @xmath243 .
now the result follows from theorem [ thm : fs - ind - invariance ] . for further investigation on the adjoint object ,
we realize @xmath44 as the category of modules over a certain hopf monad @xmath520 on @xmath1 .
we first recall the notion of _ coends _ to define @xmath520 .
let , in general , @xmath521 and @xmath522 be categories , and let @xmath523 be a functor .
a _ dinatural transformation _
@xmath524 , where @xmath525 , is a family @xmath526 of morphisms in @xmath522 such that @xmath527 for all morphisms @xmath528 in @xmath1 .
a _ coend _ of @xmath529 is a pair @xmath530 of an object @xmath525 and a dinatural transformation @xmath524 satisfying the following universal property : if @xmath531 is another dinatural transformation , then there uniquely exists a morphism @xmath532 such that @xmath533 for all @xmath534 .
if this is the case , we write @xmath535 and call @xmath536 the _ universal dinatural transformation_. now let @xmath1 be a finite tensor category over @xmath12 . for simplicity
we assume that @xmath1 fulfills the conditions ( s1)(s4 ) of lemma [ lem : strict - rigid ] . by ( * ? ? ?
* corollary 5.1.8 ) , the coend @xmath537 exists for all @xmath8 . by the parameter theorem for coends (
* ix.7 ) , we can extend @xmath538 as an endofunctor @xmath520 on @xmath1 .
day and street @xcite showed that the functor @xmath520 has a structure of a monad and the category @xmath44 is isomorphic to @xmath539 .
since @xmath44 is a braided rigid monoidal category in such a way that the forgetful functor to @xmath1 is strict monoidal , @xmath520 has a structure of a quasitriangular hopf monad .
following bruguires and virelizier @xcite , the hopf monad structure of @xmath520 is given as follows : first , let @xmath540 denote the universal dinatural transformation for the coend @xmath541 , which will be depicted as in figure [ fig : universal - dinat - trans-1 ] .
for @xmath8 and @xmath305 , define @xmath542 by figure [ fig : universal - dinat - trans - n ] . by repeated use of fubini s theorem for coends ( * ? ? ? *
ix.8 ) , we see that @xmath543 has a structure of the coend @xmath544 with universal dinatural transformation @xmath545 . the hopf monad structure on the functor @xmath520
is now defined by figure [ fig : hopf - monad - z ] .
we omit the description of the quasitriangular structure of @xmath520 since we will not use it ; see @xcite for details , where , more generally , the quantum double of a hopf monad is considered . ] ] ] following @xcite and @xcite , an isomorphism @xmath546 is given as follows : first , given an object @xmath115 , we define a natural transformation @xmath547 by @xmath548 for @xmath549 . then the map @xmath550 is an isomorphism of vector spaces .
a morphism @xmath551 makes @xmath135 into a @xmath520-module if and only if @xmath552 is a half - braiding for @xmath135 .
hence we get a bijection @xmath553 on objects .
it turns out that this bijection is in fact an isomorphism of monoidal categories ( and , moreover , an isomorphism of braided monoidal categories if we endow @xmath520 with an appropriate r - matrix ) .
let @xmath554 denote the forgetful functor .
to identify @xmath539 with @xmath44 via , we shall check that the following two assertions hold : under the identification , ( i ) a right adjoint functor of @xmath46 corresponds to that of @xmath110 , and ( ii ) the left dual object in @xmath44 corresponds to the left dual object in @xmath539 .
the first claim follows since the diagram @xmath555^{\cong}_{\eqref{eq : hopf - monad - z - cat - iso } } \ar[rd]_{t } & & \mathcal{z}(\mathcal{c } ) \ar[ld]^{u } \\ & \mathcal{c } } \ ] ] commutes . to verify ( ii ) , first check that the duality transformations of @xmath46 and @xmath110 are identities .
then , by lemma [ lem : dual - trans - composition ] and the commutativity of , the duality transformation for is shown to be the identity .
we now identify @xmath539 with @xmath44 via . in
what follows , a @xmath520-module will be denoted by the same symbol as its underlying object .
given @xmath556 , we denote by @xmath557 and @xmath558 the action of @xmath520 on @xmath135 and the corresponding half - braiding for @xmath135 , respectively . by definition
, we have @xmath559 for all @xmath549 , where the half - braiding @xmath560 is expressed by a crossing in the right - hand side .
now we describe the adjoint object @xmath47 by using the hopf monad @xmath520 . from the viewpoint of the hopf monadic description of the center ,
a left adjoint functor of the forgetful functor @xmath45 is easier to deal with than a right one .
indeed , a left adjoint of @xmath46 is just a free @xmath520-module functor @xmath561 by , the adjoint object is given by @xmath562 , where @xmath289 is the left adjoint of @xmath481 constructed in theorem [ thm : piv - cov - induced - adj ] .
thus we first give a description of @xmath563 which is the object such that @xmath564 .
[ lem : adj - obj - dual ] write @xmath565 as @xmath566 .
then @xmath567 and @xmath457 is the isomorphism determined by the following property : @xmath568 @xmath567 is obvious from . to prove
, we compute the isomorphism @xmath569 given in lemma [ lem : monoidal - adj - bidual ] . since the duality transformation of @xmath46 is the identity , it is given by the composition @xmath570 where @xmath571 is the action of @xmath520 on @xmath572 .
let , in general , @xmath135 be a @xmath520-module with action @xmath557 . by and the definition of the left antipode of @xmath520
, we have @xmath573 for all @xmath549 . replacing @xmath135 with @xmath574 , we obtain the following description of the action of @xmath520 on @xmath575 : @xmath576 applying this formula to @xmath577 , we compute : @xmath578 for all @xmath579 .
since @xmath580 , we obtain the result . for an object @xmath581 , the _ drinfeld isomorphism _
is defined by @xmath582 where @xmath583 is the half - braiding of @xmath135 . here
we show the following relation between the adjoint object and the drinfeld isomorphism : [ prop : adj - obj - z ] @xmath584 , where @xmath585 .
let @xmath566 be as in lemma [ lem : adj - obj - dual ] .
then the claim is equivalent to @xmath586 where @xmath587 . to show this
, we first note that @xmath588 is given by @xmath589 let @xmath549 be an object , and write @xmath590 .
then , by lemma [ lem : adj - obj - dual ] , @xmath591 & = \!\!\ !
\begin{array}{c } \includegraphics{prop-6 - 4 - 4.eps } \end{array } = \!\!\ !
\begin{array}{c } \includegraphics{prop-6 - 4 - 5.eps } \end{array } = \!\!\ !
\begin{array}{c } \includegraphics{prop-6 - 4 - 6.eps } \end{array } \!\!\ ! = \!\!\ !
\begin{array}{c } \includegraphics{prop-6 - 4 - 7.eps } \end{array } \!\!\ ! = \!\!\ !
\begin{array}{c } \includegraphics{prop-6 - 4 - 8.eps } \end{array } \\[1.5ex ] & = i_{{\mathbbm{1}}}(x ) .
\end{aligned}\ ] ] by the universality , @xmath592 .
hence @xmath593 .
we show that the order of the map @xmath595 is finite and give a bound of the order in terms of the distinguished invertible object of @xmath1 .
a key observation for our proof is that there is an isomorphism @xmath596 where @xmath50 is the functor defined by @xmath597 as in the proof of proposition [ prop : adj - obj - z ] , it is easier to deal with @xmath598 of that proposition than the adjoint object @xmath599 itself .
thus , instead of establishing an isomorphism as in , we prove the following proposition : [ prop : end - otimes - n - iso ] define @xmath600 by @xmath601 for a morphism @xmath602 in @xmath1 and objects @xmath305 .
then @xmath603 is an isomorphism of vector spaces .
an isomorphism @xmath604 is given in @xcite under the assumption that @xmath1 is a ribbon category .
replacing the braiding in the arguments in @xcite with the half - braiding for @xmath31 , one can see that is an isomorphism without the assumption that @xmath1 is a ribbon category . for the sake of completeness
, we provide a detailed proof for the bijectivity of .
define @xmath605 by @xmath606 for @xmath607 and @xmath305 , where @xmath608 is the morphism given in .
the dinaturality of @xmath536 in the variables @xmath305 implies the naturality of @xmath609 in @xmath305 , and the universality implies that the map @xmath610 is in fact an isomorphism of vector spaces ; see @xcite for details and more general result .
next , define an isomorphism @xmath611 by @xmath612 and @xmath613 for @xmath614 , where @xmath615 ( @xmath8 , @xmath556 ) is the fusion operator for the hopf monad @xmath520 given by .
we note that @xmath616 is graphically characterized as follows : @xmath617 by using this , we compute @xmath618 as in figure [ fig : computation - kappa - n ] .
one now sees that the map @xmath619 is in fact the map @xmath620 in consideration .
hence @xmath620 is an isomorphism . ]
let @xmath54 be the map induced by @xmath621 via the isomorphism @xmath620 of proposition [ prop : end - otimes - n - iso ] . since @xmath621 is conjugate to @xmath595 , so is @xmath622 .
it seems to be difficult to express this map in a familiar way .
however , the @xmath5-th power of @xmath622 is an exception ; it can be expressed in terms of the double duality functor @xmath158 as follows : [ prop : end - otimes - n - e - map ] for all @xmath623 and @xmath305 , we have @xmath624 write @xmath625 and @xmath598 . by proposition [ prop : end - otimes - n - iso ] , we have @xmath626 for all objects @xmath549 and morphisms @xmath602 . by using this formula repeatedly
, we have @xmath627 for all @xmath305 ( see figure [ fig : end - otimes - n - e - map - proof ] for the detailed computation for with @xmath628 ) . hence follows . ] the distinguished invertible object @xmath629 is an invertible object introduced by etingof , nikshych and ostrik in @xcite .
generalizing radford s @xmath630-formula for finite - dimensional hopf algebras , they showed that there exists an isomorphism @xmath631 of monoidal functors .
in particular , some positive powers of @xmath158 are isomorphic to @xmath632 as monoidal functors . following , let @xmath633 denote the smallest positive integer @xmath431 such that the @xmath431-th power of @xmath158 is isomorphic to @xmath632 as a monoidal functor .
for example , @xmath634 if and only if @xmath1 admits a pivotal structure
. by , we have @xmath635 where @xmath636 is the smallest positive integer @xmath14 such that @xmath637 .
[ thm : en - order ] the order of @xmath622 divides @xmath638 . since @xmath595 and @xmath621 are conjugate to @xmath622 , the orders of @xmath595 and @xmath621 are equal to the order of @xmath622 and , in particular , divide @xmath638
. by , the same order of these three maps divides @xmath639 .
let @xmath79 be the @xmath431-th power of @xmath158 , where @xmath640 .
we note that @xmath79 is a strict monoidal functor since so is @xmath158 by our assumption . by definition
, there exists an isomorphism @xmath641 of monoidal functors .
now let @xmath623 be a natural transformation and consider the following diagram : @xmath642 where @xmath305 . since @xmath152 and @xmath643 are natural , this diagram commutes . by the definition of monoidal natural transformations , the compositions along the top and the bottom row are the identities . by proposition [ prop : end - otimes - n - e - map ] , we have @xmath644 for all @xmath305 .
hence the order of @xmath622 divides @xmath645 .
let @xmath646 be the smallest positive integer @xmath431 such that the @xmath431-th power of @xmath158 is isomorphic to @xmath632 as a functor . in the case where @xmath316 , the proof of the above theorem works even if @xmath643 is only an isomorphism of functors .
thus we obtain the following theorem : [ thm : en - order-1 ] the order of @xmath647 divides @xmath646 .
the following two corollaries are immediate consequences of the above results : for every positive integer @xmath5 , the sequence @xmath648 is periodic .
the period of the sequence divides @xmath638 . for each integer @xmath649 , we define @xmath650 to be the subring of @xmath12 generated by the roots of the polynomial @xmath651 .
since the trace of a linear operator is the sum of the eigenvalues , we have : @xmath652 for all @xmath653 and @xmath24 .
let @xmath59 be a finite - dimensional hopf algebra over an algebraically closed field @xmath12 with comultiplication @xmath423 , counit @xmath655 and antipode @xmath129 . in this section ,
we describe the adjoint object in @xmath393 and study its indicators in detail . for this purpose , it is useful to realize the center of @xmath654 as the category of yetter - drinfeld modules . recall that a yetter - drinfeld module over @xmath59 is a pair @xmath656 consisting of a left @xmath59-module @xmath135 and a left @xmath59-coaction @xmath657 such that the yetter - drinfeld condition @xmath658 holds for all @xmath428 and @xmath659 , where @xmath660 .
we denote by @xmath661 the category of finite - dimensional yetter - drinfeld modules over @xmath59 .
this category is a braided monoidal category @xcite and is isomorphic to @xmath662 as a braided monoidal category via @xmath663 where @xmath583 is given by @xmath664 for @xmath665 and @xmath659 .
note that the isomorphism commutes with the forgetful functors to @xmath654 .
in what follows , we identify @xmath662 with @xmath666 via .
then a right adjoint functor of the forgetful functor @xmath667 is given as follows : given @xmath668 , we make @xmath669 into a yetter - drinfeld module over @xmath59 by defining the action and the coaction of @xmath59 by @xmath670 for @xmath671 and @xmath672 , respectively .
the assignment @xmath673 naturally extends to a functor @xmath674 , which can be considered as a variant of radford s induction functor @xcite .
now we define natural transformations @xmath124 and @xmath206 by @xmath675 it is easy to check that @xmath676 is an adjunction . by using this adjunction , we now give the following description of the adjoint object : [ prop : adj - obj - h - mod ] let @xmath677 be the adjoint representation of @xmath59 , _
i.e. _ , the vector space @xmath59 with the left @xmath59-module structure @xmath678 given by @xmath679 and let @xmath680 be the isomorphism of @xmath59-modules defined by @xmath681 then @xmath682 is the adjoint object for @xmath654 .
define @xmath683 by @xmath684 .
the canonical isomorphism @xmath685 of vector spaces induces an isomorphism @xmath686 of yetter - drinfeld modules over @xmath59 . in view of proposition
[ prop : adj - obj - z ] , it is sufficient to show that the map @xmath22 is equal to the drinfeld isomorphism @xmath687 . to compute @xmath687 , we fix a basis @xmath688 of @xmath689 and denote by @xmath690 and @xmath691 the bases of @xmath692 and @xmath693 dual to @xmath688 and @xmath690 , respectively .
then , for all @xmath694 and @xmath695 , we compute : @xmath696 let @xmath697 .
writing @xmath698 , we compute : @xmath699 where @xmath690 and @xmath691 have the same meaning as in the proof of the above proposition .
given @xmath668 , we denote by @xmath700 the set of @xmath59-invariants in @xmath7 . under the canonical isomorphism @xmath701 with @xmath702 ,
the map @xmath703 is conjugate to @xmath595 .
hence we obtain : therefore @xmath58 is computable in the following sense : suppose that one knows the structural coefficients of @xmath59 with respect to a basis @xmath688 of @xmath59 .
then one can obtain a basis of @xmath705 by solving the system of linear equations @xmath706 with variables @xmath707 , where @xmath678 is the action of @xmath59 on @xmath708 . since @xmath709 is defined only by using the antipode and the transposition of tensorands
, one can represent it as a matrix with respect to the basis of @xmath705 .
it is obvious that the trace of a matrix is computable .
hence , if @xmath712 is inner for some @xmath713 , then @xmath714 is a periodic sequence whose period divides @xmath431 .
this observation is a hopf - algebraic counterpart of theorem [ thm : en - order-1 ] . indeed , with the notation in that theorem , we have @xmath715 we now consider the case where @xmath59 is a finite - dimensional semisimple hopf algebra over an algebraically closed field @xmath12 of characteristic zero
. then there uniquely exists an integral @xmath449 such that @xmath716 ( the _ normalized integral _ ) . by using this element
, we define a linear form @xmath717 on @xmath59 by @xmath718 for @xmath428 , where @xmath719 is the integral on @xmath59 such that @xmath720 .
let @xmath721 and @xmath24 . with the above notations ,
the @xmath3-th fs indicator of the adjoint object in @xmath654 is given by @xmath722 where @xmath723 and @xmath724 , @xmath725 is a unique integer such that @xmath726 .
let @xmath332 denote the @xmath3-th fs indicator of ng and schauenburg with respect to the canonical pivotal structure of @xmath654 . by theorem [ thm : adj - obj - ind - ns ]
, we have @xmath728 by using the integrals , @xmath729 for @xmath668 is expressed as @xmath730 where @xmath731 is the character of @xmath7 ( see @xcite for the case of @xmath732 ; the general case is proved in a similar way ) . using radford s trace formula ( * ? ? ?
* theorem 2 ) and some identities on integrals ( * ? ? ?
* proposition 4 ) , we have @xmath733 for @xmath428 .
now the result follows from , and . for non - semisimple hopf algebras , we do not know any general formula to compute @xmath58 .
thus it is an interesting problem to know its value .
here we determine @xmath61 for certain non - semisimple and non - pivotal hopf algebras in a direct way .
suppose that @xmath12 is an algebraically closed field of characteristic zero .
let @xmath739 and @xmath431 be integers such that @xmath740 and @xmath741 , and let @xmath742 be a primitive @xmath739-th root of unity .
the algebra @xmath743 is generated by @xmath744 , @xmath745 and @xmath643 subject to the following relations : @xmath746 note that the following set is a basis of this algebra : @xmath747 the algebra @xmath743 has the hopf algebra structure determined by @xmath748 this hopf algebra is an example of liftings of quantum planes and isomorphic to the hopf algebra @xmath749 of @xcite .
our result is : note that an element @xmath759 is central if and only if @xmath760 one easily calculates that the adjoint action is given by @xmath761 for @xmath762 .
now suppose that @xmath759 is central . by and the first equation of , @xmath763 must be of the form @xmath764 for some @xmath765 . by , and
, we have @xmath766 this implies that @xmath767 is zero if @xmath768 and @xmath769 .
hence , @xmath770 where @xmath771 and @xmath772 .
\(a ) if @xmath773 , then @xmath774 for all @xmath775
. therefore an element @xmath763 of the form belongs to the space spanned by .
it is easy to see that the space spanned by is contained by @xmath754 .
hence the result follows .
\(b ) now suppose that @xmath751 .
the difference from the case where @xmath773 is that an element @xmath763 of the form does not fulfill @xmath776 in general .
by , @xmath777 one can check that @xmath778 ( otherwise @xmath779 would follow ) .
hence @xmath776 implies @xmath780 for all @xmath781 except for @xmath782 . summarizing the results so far
, we have @xmath783 therefore @xmath784 .
it is easy to show @xmath785 .
\(a ) if @xmath773 , then @xmath253 is the identity on @xmath754 ( more precisely , @xmath253 is an inner automorphism implemented by @xmath643 ) . by proposition [ prop : adj - obj - h - mod ] and
the previous lemma , we have @xmath786 ( b ) if @xmath751 , then @xmath787 ( @xmath788 is a basis of @xmath753 by the previous lemma .
the map @xmath789 of proposition [ prop : adj - obj - h - mod ] is represented by a diagonal matrix with respect to the basis @xmath790 , as follows : @xmath791 hence @xmath792 is given as stated .
two finite - dimensional hopf algebras @xmath689 and @xmath529 are said to be _ gauge equivalent _ if @xmath793 and @xmath794 are equivalent as @xmath12-linear monoidal categories . to conclude this paper , we discuss the possibility of applications of our results to the problem on gauge equivalence of hopf algebras .
namely , we investigate whether the following assertions are equivalent or not : it is obvious that ( 1 ) implies ( 2 ) .
however , it does not seem to be known whether the converse holds . in what follows ,
we write @xmath796 for simplicity . by theorems [ thm : adj - obj - invariance ] and
[ thm:1-r - th - ind - adj - book - hopf ] , we can give very partial answers to the above question .
for example , let @xmath797 be a primitive @xmath798-th root of unity .
then @xmath799 and @xmath800 are not gauge equivalent , since @xmath801 the invariant @xmath802 is not sufficient to solve the problem .
to see this , we use the gauge invariant @xmath803 , which we have mentioned in [ subsec : fs - ind - examples ] .
since the element @xmath804 is a non - zero left integral , we have @xmath805 by .
for simplicity , we define @xmath806 to be a unique integer such that @xmath807 table [ tab : book - hopf - gauge - inv ] displays the values of @xmath808 and @xmath809 for @xmath810 and all possible @xmath431 .
one finds in the table many pairs such that @xmath743 and @xmath811 are not gauge equivalent but @xmath812 for all @xmath14 .
c|ccccccccccccc @xmath431 & 1 & 2 & 4 & 5 & 7 & 8 & 10 & 11 & 13 + @xmath813 & 702 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath814 & 702 & 0 & 0 & 0 & 0 & 0 & 27 & 0 & 0 + @xmath815 & 702 & 0 & 27 & 0 & 27 & 0 & 27 & 0 & 27 + @xmath816 & 0 & 3 & 15 & 24 & 21 & 10 & 19 & 13 & 7 + + @xmath431 & 14 & 16 & 17 & 19 & 20 & 22 & 23 & 25 & 26 + @xmath813 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath814 & 0 & 0 & 0 & 27 & 0 & 0 & 0 & 0 & 0 + @xmath815 & 0 & 27 & 0 & 27 & 0 & 27 & 0 & 27 & 0 + @xmath816 & 7 & 13 & 19 & 10 & 21 & 24 & 15 & 3 & 0 the table also shows that @xmath817 and @xmath818 are gauge equivalent if and only if they are isomorphic as hopf algebras .
it is fortunate that we can obtain such a result ; for other @xmath819 of different orders , even the combination of @xmath820 and @xmath821 does not seem to be enough to conclude the problem on the gauge equivalence of @xmath743 s . b. day and r. street .
centres of monoidal categories of functors . in _ categories in algebra , geometry and mathematical physics _ , volume 431 of _ contemp .
_ , pages 187202 .
soc . , providence , ri , 2007
. s .- h .
ng and p. schauenburg .
higher frobenius - schur indicators for pivotal categories . in _
hopf algebras and generalizations _ ,
volume 441 of _ contemp .
_ , pages 6390 .
soc . , providence , ri , 2007 . | in this paper , we introduce the notion of the pivotal cover @xmath0 of a left rigid monoidal category @xmath1 to develop a theoretical foundation for the theory of frobenius - schur ( fs ) indicators in `` non - pivotal '' settings . for an object @xmath2 ,
the @xmath3-th fs indicator @xmath4 is defined by generalizing that of an object of a pivotal monoidal category .
this notion gives a categorical viewpoint to some recent results on generalizations of fs indicators .
based on our framework , we also study the fs indicators of the `` adjoint object '' in a finite tensor category , which can be considered as a generalization of the adjoint representation of a hopf algebra .
the indicators of this object closely relate to the space of endomorphisms of the iterated tensor product functor . |
we utilize an elastic rod model described by @xcite that has been parameterized to represent dna @xcite .
the equations of motion are as follows : @xmath3 in this system of equations , equation ( [ intro:1 ] ) represents the balance of force and linear momentum and equation ( [ intro:2 ] ) represents the balance of torque and angular momentum according to newton s laws
. equations ( [ intro:3 ] ) and ( [ intro:4 ] ) are continuity relations expressed in a non - inertial reference frame . here ,
@xmath4 and @xmath5 are independent variables representing time and fiducial arclength , respectively .
the functions that we wish to study are the four three - vector functions @xmath6 , @xmath7 , @xmath8 , and @xmath9 .
the matrices @xmath10 , @xmath11 and @xmath12 and the scalar @xmath13 are as follows : matrix @xmath10 is the linear density of the moment of inertia tensor
. matrices @xmath11 and @xmath12 represent the elastic properties of the rod ( @xmath14 : young s modulus , @xmath15 and @xmath16 : shear modulus , @xmath17 : torsional rigidity , @xmath18 and @xmath19 : bend stiffness ) . @xmath13 is the linear mass density and is equal to @xmath20 for an isotropic model of dna . the matrices have the following values for an isotropic model of dna : @xmath21 \label{intro:6}\\ c_{ik}&=&\left(\matrix{c_1 & 0 & 0 \cr 0 & c_2 & 0 \cr 0 & 0 & c_3 \cr}\right)\equiv \left(\matrix{8.16 & 0 & 0 \cr 0 & 8.16 & 0 \cr 0 & 0 & 21.6 \cr}\right)\times10^{-10}\left[\frac { km}{s^2}\right ] \label{intro:7}\\ d_{ik}&=&\left(\matrix{d_1 & 0 & 0 \cr 0 & d_2 & 0 \cr 0 & 0 & d_3 \cr}\right)\equiv \left(\matrix{2.7 & 0 & 0 \cr 0 & 2.7 & 0
\cr 0 & 0 & 2.06 \cr}\right)\times10^{-28}\left[\frac { km^3}{s^2}\right ] \label{intro:8}\end{aligned}\ ] ] for isotropic bending @xmath22 and shear @xmath23 we will denote @xmath24 and @xmath25 .
these definitions will be used below .
the above values have been adapted from @xcite , @xcite . in biological terminology the components of @xmath7
correspond to the three dna helical parameters describing translation @xmath26 and the components of @xmath6 correspond to the three dna helical parameters describing rotation @xmath27 .
if we attach a local coordinate frame , denoted by @xmath28 , to each base - pair the axes will point in the direction of the major groove , minor groove and along the dna helical axis , respectively .
the shape of the dna can be described by a three - dimensional vector function @xmath29 which is related to @xmath7 and @xmath6 by a suitable mathematical integration . in elastic rod terminology
the three - dimensional vector function @xmath29 gives the centerline curve of the rod as a function of @xmath4 and @xmath5 , showing only how the rod bends . to show the twist , shear and extension of the rod we must attach `` director '' frames made of the orthogonal triples @xmath28 at regular intervals along the rod .
the director frames are evenly spaced along the rod when the rod is not extended and are all parallel to each other ( with @xmath30 pointing along the rod ) when the rod is not bent or twisted .
any deformation of the rod will be indicated by a corresponding change in the orientation of the director frames .
the vector @xmath9 ( @xmath7 ) is the translational velocity of @xmath31 in time ( space ) : @xmath32 the directors are of constant unit length , so the velocity of each director @xmath8 ( @xmath6 ) is always perpendicular to itself , and it must also be perpendicular to the axis of rotation of the local frame . @xmath33 we will denote @xmath7 of the unstrained state by @xmath34 . in general an unstrained rod can have any shape .
the most simple case is when the rod is straight with unit extension , @xmath35 , because in the unstrained state : @xmath36 similarly , an unstressed elastic rod may have an intrinsic bend and/or twist , denoted by @xmath37 , which must be subtracted from the total @xmath6 to give the dynamically active part .
the simplest case is a rod with no intrinsic bend or twist @xmath38 and @xmath35 .
for short wave lengths ( section [ section:2 ] ) we will suppose that @xmath39 @xmath40 and @xmath41 . in section [ section:3 ] we will consider a straight rod with an intrinsic twist @xmath42 and @xmath43 .
the results from this section are valid for any shape in which the curvature of the rod is much greater than the wavelength of the disturbance being propagated regardless of the intrinsic shape of the rod .
let us search as usual for the equilibrium point of the system ( [ intro:1]-[intro:4 ] ) by setting @xmath9 , @xmath7 , @xmath8 and @xmath6 to constants .
it is easy to see that an equilibrium point of equations ( [ intro:1]-[intro:4 ] ) is @xmath44 , @xmath45 and @xmath46 we shall suppose that each variable differs slightly from the equilibrium value so we retain only the linear terms in the equations .
the linearized system ( [ intro:1]-[intro:4 ] ) is : @xmath47 in the system ( [ linear:1]-[linear:4 ] ) and further in this paper @xmath9 , @xmath7 , @xmath8 and @xmath6 denote the small deviations from the equilibrium values , e.g. @xmath48 . in the usual manner
we search for a solution to ( [ linear:1]-[linear:4 ] ) using harmonic analysis .
we assume that each variable depends on time and space as follows : @xmath49 where @xmath50 denotes one of the four vector variables from the system of equations ( [ linear:1]-[linear:4 ] ) and @xmath51 denotes its amplitude .
is a scalar quantity corresponding to a frequency of oscillation and that @xmath8 is a vector quantity that describes a rotation , so there should be no confusion between the two variables . ]
let us substitute ( [ linear:4a ] ) to the system ( [ linear:1]-[linear:4 ] ) : @xmath52 for short wavelength we search for solutions with @xmath53 , so in the limit @xmath54 we neglect all terms which contain cross products and the system ( [ linear:4b]-[linear:4e ] ) becomes : @xmath55 here division into longitudinal and transversal components has been obtained by defining @xmath56 and @xmath57 .
equations ( [ linear:4k]-[linear:4n ] ) represent four types of waves that can propagate in the elastic rod with velocity , @xmath58 ( according to the order of equations above ) : * shear waves ( @xmath59 ) : @xmath60\over 3.22 \cdot 10^{-15 } [ k / m]}\approx 5.03 \r a / ps\ ] ] * extension waves ( @xmath61 ) : @xmath62\over 3.22 \cdot 10^{-15 } [ k / m]}\approx 8.2 \r a / ps\ ] ] * bend waves ( @xmath63 ) : @xmath64\over 4.03 \cdot 10^{-34 } [ km]}\approx 8.2 \r a / ps\ ] ] * twist waves ( @xmath65 ) : @xmath66\over 2\cdot 4.03 \cdot 10^{-34 } [ km]}\approx 5.0 \r a / ps\ ] ] these results were obtained for a rod with arbitrary shape because all terms that define the rod shape in equations ( [ linear:4b]-[linear:4e ] ) were omitted .
so , if wavelength tends to zero ( is the least space parameter of the problem ) four wave types can propagate along the rod .
it is well known that the measurement of wave velocity is a usual method for determining elastic properties of solids .
so , wave velocity ( transversal and longitudinal ) and young s modulus ( shear modulus ) are uniquely determined by formulae similar to the ( [ linear:4p]-[linear:4s ] ) .
hakim et al . @xcite
have determined the velocity of sound in dna . based on elastic constants obtained from @xcite , @xcite and used in this work ,
the velocity is approximately two times less than that obtained by hakim et al .
@xcite .
experiments to measure elastic properties of single dna molecules have been reported using scanning tunnelling microscopy @xcite , fluorescence microscopy @xcite , fluorescence correlation spectroscopy @xcite , optical tweezers @xcite@xcite , bead techniques in magnetic fields @xcite @xcite , optical microfibers @xcite , low energy electron point sources ( electron holography ) @xcite and atomic force microscopy ( afm ) @xcite .
each method differs in the molecular properties probed , spatial and temporal resolution , molecular sensitivity and working environment , so each method gives close but not the same value of elastic constants .
so if we evaluate extension / compression velocity using constants obtained from the data reported in @xcite the result will be two times slower than the value reported here and based on data from @xcite and @xcite .
here we concentrate our attention on the case of the straight , twisted rod with no intrinsic bend or shear . in this section
we suppose that @xmath67 and @xmath68 . as we will see this case allows for a complete analytical analysis of extension / compression ( section [ sound ] ) , twist ( section [ twist ] ) , and bend / shear ( section [ coupled ] ) waves of arbitrary wave length
. equations ( [ linear:1]-[linear:4 ] ) in component form are : @xmath69 it is easy to see that equations ( [ linear:11 ] ) and ( [ linear:17 ] ) ( also ( [ linear:14 ] ) and ( [ linear:20 ] ) ) are independent from all other equations and describe extension ( twist ) waves .
these two wave types will be discussed in the next two sections . from equations ( [ linear:11 ] ) and ( [ linear:17 ] )
it is easy to obtain two wave equations for the extension / compression waves : @xmath70 these small amplitude waves have velocity ( [ linear:4q ] ) and these equations have an harmonic solution[multiblock footnote omitted ] : @xmath71 where @xmath72 are arbitrary constants ( wave amplitudes ) and @xmath73 @xmath74 .
using elastic constants obtained from data in @xcite and @xcite the velocity of sound in dna is equal to @xmath75 ( see equation ( [ linear:4q ] ) ) and dispersion is linear .
this is the velocity of propagation of harmonic waves in dna according to the linear approximation . in the linear approximation
the amplitude of the wave must be small , but it can have any wavelength . in the same manner equations ( [ linear:14 ] ) and ( [ linear:20 ] ) yield two wave equations which describe twist waves : @xmath76 linear solutions of these equations can be written as : @xmath77 where @xmath78 are the arbitrary constants ( wave amplitudes ) and @xmath79 @xmath80 .
these solutions describe twist waves with velocity @xmath81 ( see equation ( [ linear:4s ] ) ) . in the general case
, the solutions of these equations will be dalembertian waves described by ( [ sec1:11]-[sec1:12 ] ) .
again the dispersion law is linear . substituting ( [ linear:4a ] ) into the remaining equations of system ( [ linear:9]-[linear:20 ] ) yield : @xmath82 \label{linear:42}\\ -i\omega i\vec\omega_\bot & = & ikd\vec\omega_\bot+ d\left[\vec\omega_0\times\vec\omega_\bot\right]+ c\left[\vec\gamma_0\times\vec\gamma_\bot\right]\label{linear:43}\\ -i\omega\vec\gamma_\bot + \left[\vec\omega_\bot\times\vec\gamma_0\right ] & = & ik\vec\gamma_\bot + \left[\vec\omega_0\times\vec\gamma_\bot\right]\label{linear:44}\\ -i\omega\vec\omega_\bot + \left[\vec\omega_\bot\times\vec\omega_0\right]&= & ik\vec\omega_\bot \label{linear:45}\end{aligned}\ ] ] here , the definitions @xmath83 , @xmath56 , @xmath84 and @xmath57 are used .
this is an homogeneous system of linear equations with unknowns @xmath85 , @xmath59 , @xmath86 and @xmath63 .
solutions other than the trivial solution exist only if the determinant of the coefficients is zero .
this condition is satisfied if : @xmath87 let us discuss the untwisted rod ( @xmath88)-[linear:20 ] ) consists of two independent subsystems .
equations ( [ linear:9 ] , [ linear:13 ] , [ linear:15 ] , [ linear:19 ] ) and ( [ linear:10 ] , [ linear:12 ] , [ linear:16 ] , [ linear:18 ] ) present two different polarizations of bend / shear wave . ] . equation ( [ linear:46 ] ) yields : @xmath89 it will be convenient to use dimensionless variables ( marked by asterisk ) @xmath90 , @xmath91 in this section and define @xmath92 .
equation ( [ linear:47 ] ) becomes(asterisks are omitted ) : @xmath93 with solutions : @xmath94 ^ 2k^2+\gamma^2_0 } \pm \sqrt{[g-1]^2k^2+\gamma^2_0 } \right\}\ ] ] there are a total of four solutions .
the @xmath95 sign before the braces merely changes the direction of propagation and will not be discussed further .
the @xmath95 sign between the roots produces two different branches of solutions as presented in fig.[figure:1 ] . the `` upper '' branch .
] corresponds to bend waves and the `` lower '' branch to shear waves . for a rod of infinite length
, any value of @xmath96 can be chosen that satisfies equation ( [ linear:48 ] ) , but for a rod of finite length there will be an additional constraint that must be satisfied .
if the rod is fixed at both ends then the usual restriction of having a node at each end applies .
this introduces the constraint @xmath97 between the length of the rod , @xmath98 , and the possible wavelengths , @xmath99 , so that only discrete points along either sub - branch in figure 1 will be observed .
the primary results are that , in the linear approximation , four different types of waves can propagate through a uniform elastic rod .
these waves correspond to extension(compression ) , twist , bend and shear .
an extension or twist wave will propagate without exciting other modes or changing shape , and it has a linear dispersion relation .
bend and shear waves behave rather differently and have a nonlinear dispersion relation .
each type obeys a dispersion law that describes two additional sub - branches . utilizing constants suitable for dna we find that , in the limit of small wavelengths , extension and bend propagate with a velocity of approximately 8 / ps and twist and shear propagate with a velocity of approximately 5 / ps .
is is also significant that the dispersion relation for bend and shear waves is coupled with the inherent twist of dna .
we propose that these physical phenomena enable proteins interacting with dna to accomplish highly sophisticated tasks .
for example the difference in extension and twist velocities can be utilized to measure the distance between two points on the dna .
since the dispersion law for bend and shear waves depends on intrinsic twist , a mechanism for measuring dna topology exists because of the relation between twist , linking number and writhe .
other protein - protein communications can certainly be established to assist cellular mechanisms .
we suggest a molecular dynamics simulation experiment to check the correspondence between the linear theory of the propagation of waves in elastic rods and an all atom simulation of dna . in all - atom
molecular dynamics simulations of dna the number of base pairs that can be simulated for a significant length of time is from tens to less than hundreds of base pairs . for a simulation of dna with fixed ends
it is sufficient to apply a sharp impulse ( @xmath100 function excitation ) to one end of the dna and measure the time of propagation of this disturbance to the other end as a measure of the velocity of extension / compression waves in dna according to equation ( [ linear:4q ] ) .
the propagation of twist can be similarly tested by applying a torque . to measure dispersion
one must make some simple calculations before the simulation and then excite one end of the dna with a driving force of the appropriate frequency . in this manner
a standing wave can be established in the dna .
the wave frequency and wave number ( @xmath101 ) are related by the well known relation : @xmath102 where @xmath103 is the wave frequency , @xmath104 is the wave velocity and @xmath105 is the wavelength . in this case
the wave length ( wave number ) is easy to evaluate : @xmath106 and @xmath107 is the frequency of the driving oscillation .
the values obtained from simulation can be compared with the number obtained from formula ( [ linear:48 ] or [ linear:48 ] ) as a means of correcting our choice of constants .
we express our thanks to dr .
jan - ake gustafsson , dr .
iosif vaisman and dr .
yaoming shi for stimulating and helpful discussions .
we would also like to thank yaoming shi for a critical reading of the manuscript .
this work was conducted in the theoretical molecular biology laboratory which was established under nsf cooperative agreement number osr-9550481/leqsf ( 1996 - 1998)-si - jfap-04 and supported by the office of naval research ( no-0014 - 99 - 1 - 0763 ) .
30 john j. tyson , bela novak , garrett m. odell , kathy chen and dennis thron . chemical kinetic theory : understanding the cell - cycle regulation .
trends biochem .
21 , 89 - 96 , ( 1996 ) .
landau l.d .
, lifshitz e.m .
fluid mechanics .
pergamon press .
landau l.d .
, lifshitz e.m .
electrodynamics of continous media .
pergamon press .
j.c.simo , j.e.marsden , p.s.krishnaprasad , archive for rational mechanics and analysis , * 104 * , 125 - 184,(1988 ) .
a ) yaoming shi , andrey e. borovik , and john e. hearst .
elastic rod model incorporating shear extension , generalized nonlinear schrdinger equations , and novel closed - form solutions for supercoiled dna . j.
phys . * 103*,(8),3166 - 3183 , ( 1995 ) .
b ) martin mcclain , yaoming shi , t.c.bishop , j.e.hearst .
visualization of exact hamiltonian dynamics solutions .
( to be published ) .
j.d.moroz , p.nelson , entropic elasticity of twist - storing polymers , macromolecules , 31 , 6333 - 6347 , ( 1998 ) . c. bouchiat , m. m@xmath108zard .
elasticity model of a supercoiled dna molecule , phys .
* 80*,(7),1556 - 1559 , ( 1998 ) .
m.b.hakim , s.m.lindsay , j.powell .
the speed of sound in dna .
biopolymers , 23 , 1185 - 1192 ( 1984 ) guckenberger r. , heim m. , cevc g. , knapp h.f . , wiegr@xmath109be w. , hillebrand a. science * 266 * , 1994 , 1538 - 1540 .
yanagida m. , hiraoka y. , katsura i. , cold spring harbor symp .
. biol . * 47 * , 1983 , 177 .
wennmalm s. , edman l. , rigler r. proc .
sci . * 94 * , 1997 , 10641 - 10646 .
smith s.b .
, cui y. , bustamante c. science . * 271 * , 1996 , 759 - 799 .
wang d.w . , yin h. , landick r. , gelles j. , block s.m .
j. * 71 * , 1997 , 1335 - 1346 .
smith s.b .
, finzi l. , bustamante c. science . * 258 * , 1992 , 1122 - 1126 .
strick t.r . ,
allemand j.f .
, bensimon d. , bensimon a. , croquette v. science . *
271 * , 1996 , 1835 - 1837 .
, lebrun a. , heller ch
. , lavery r. , viovy j. , chatenay d. , caron f. science . * 271 * , 1996 , 792 - 794 .
fink h. w. , sch@xmath110nenberger ch .
science . * 398 * , 1999 , 407 - 410 .
hansma h.g . ,
sinsheimer r.l . , li m.q .
, hansma p.k .
nucleic acids res . *
20 * , 1991 , 3585 - 3590 .
bishop t. , zhmudsky o. currents in computational molecular biology 2001 .
les publications crm , montreal , 2001 .
information transmission along dna .
p. 105 - 106 .
isbn 2 - 921120 - 35 - 6 . | it is shown that information transmission inside a cell can occur by means of mechanical waves transmitted through dna .
the propagation of the waves is strongly dependent on the shape of the dna thus proteins that change the shape of dna can alter signal transmission .
the overall effect is a method of signal processing by dna binding proteins that creates a `` cellular communications network '' .
the propagation of small amplitude disturbances through dna is treated according to the mechanical theory of elastic rods . according to the theory four types of mechanical waves affecting extension(compression ) , twist , bend or shear
can propagate through dna .
each type of wave has unique characteristic properties .
disturbances affecting all wave types can propagate independently of each other . using a linear approximation to the theory of motion of elastic rods , the dispersion of these waves
is investigated .
the phase velocities of the waves lies in the range @xmath0 using constants suitable for a description of dna .
the dispersion of all wave types of arbitrary wave length is investigated for a straight , twisted rod .
based on these findings , we propose all - atom numerical simulations of dna to investigate the propagation of these waves as an alternative measure of the wave velocity and dispersion analysis . *
introduction * our treatment of dna is based on the hypothesis that such a well organized system as a cell must have a highly sophisticated system of communications .
for instance , the cell cycle is the result of a series of complex self organized processes that must be well orchestrated .
a model has been developed to demonstrate how these events may be initiated by critical concentrations of specific proteins which shift equilibrium to favor the advance of the cell cycle@xcite .
however , it is also known that the cell - cycle can be disrupted if conditions are not suitable .
similar arguments apply to virtually all cellular processes and in order to achieve the required checks and balances a method of communication is necessary .
is there the possibility that information is transmitted through dna electromagnetically instead of mechanically ?
in such case dna will function as a transmission line which requires total internal reflection ( tir ) of the radiation within the dna . to achieve tir
, the wavelength of the radiation must be 5 - 10 times less than the diameter of the transmission line .
since the diameter of the dna is close to 20 @xmath1 the radiation must have a wavelength close to 2 @xmath1 .
this wavelength is close to atomic dimensions so diffraction should dominate .
furthermore , the energy associated with this wavelength radiation is on the order of @xmath2 which is sufficient to destroy chemical bonds and therefore not easily managed biologically . for these reasons
, we believe that the communication will be achieved mechanically .
the remainder of the paper will demonstrate that the mechanical properties of dna have the necessary time and spatial dimension to support the propagation of information while interactions between proteins and dna provide a mechanism by which this information is processed .
this manuscript is based on a continuous medium model of dna .
such an approach is well known and widely used to describe solids , liquids and gases ( see , for example @xcite ) .
thus , when we speak of infinitely small elements of volume , we shall always mean those which are `` physically '' infinitely small , i.e. very small compared with wavelength or radius of curvature under consideration , but large compared with the distance between the atoms in dna . |
the production of @xmath0 mesons in two - nucleon collisions depends strongly on the final state interaction ( fsi ) as has been stressed in previous work @xcite .
notwithstanding this fact , the interaction between the final particles was treated in quite an approximate way .
only the @xmath1 interaction has been considered in a more precise manner .
the interaction of the @xmath0 meson either was not considered at all or simulated by means of a watson enhancement factor . in view of this apparent lack of a correct treatment of fsi
we want to study this reaction by using a rigorous three - body model .
the properties of the @xmath3 system , containing three free particles in the asymptotic region , were investigated in @xcite where fsi effects in @xmath4 reactions as well as in @xmath0 photoproduction on a deuteron were studied . as was shown in detail ,
the three - body aspects of fsi are very important , at least near the threshold region where @xmath5 waves give by far the largest contribution to the production process .
it is obvious that the three - body @xmath5 wave part is most sensitive to fsi in view of the dominance of the @xmath5 wave interaction in the @xmath1 and @xmath6 subsystems at low energies .
furthermore , it is intuitively clear that primarily the central region of the system , where @xmath5 waves dominate , is distorted by multiple scattering events . as a consequence , close to threshold
the @xmath3 interaction strongly enhances the @xmath7 cross section . at higher energies , where the large spatial extension of the deuteron comes into play , an essential fraction of higher partial waves contributes increasingly to the @xmath0 production mechanism
this rapidly reduces the relative @xmath5 wave contribution to the reaction amplitude , so that the role of fsi diminishes considerably with increasing energy . as a result , already in the region of c.m .
excess energies above 100 mev the behavior of the cross section is well explained in the plane wave approximation for the final state @xcite .
turning now to the process @xmath8 , firstly we would like to note that in this case the influence of fsi must be even more important . here
the characteristic radius @xmath9 of the reaction , determined by @xmath10 with @xmath11 as typical momentum transfer , is only about 0.3 fm .
this value is much smaller than in photoproduction , where not too close to the threshold the characteristic range is given by the deuteron radius @xmath12 , i.e. , @xmath13 2 fm .
therefore , one may expect that for energies for which @xmath14 is valid , with @xmath15 denoting the maximal c.m.momentum of the final particles , the final state will be dominated by @xmath5 waves , for which the role of the @xmath3 interaction is crucial .
indeed , as will be shown below , in the region from threshold up to an excess energy @xmath16 mev , the contribution of higher partial waves in the final state is of much minor importance compared to the @xmath5 wave contribution . in this connection
, the reaction @xmath17 seems in general to be more suited for studying the interaction in the @xmath3 system than @xmath0 photoproduction , where fsi has a smaller influence .
this advantage is however strongly reduced by a poor knowledge of the @xmath0 production mechanism in @xmath1 collisions . in the present paper
we investigate the reactions @xmath18 and @xmath2 in the region of excess energy up to @xmath16 mev which is covered in recent experiments as well as in theoretical studies ( see e.g. @xcite and references therein ) .
the two questions which we address here are : \(i ) to what extent are the three - body aspects essential for a quantitative description of fsi ?
as was already stressed in @xcite , the importance of a three - body treatment of fsi is dictated by the proximity of a virtual pole in the @xmath3 amplitude to the physical region @xcite . because of this property a perturbative approach to the @xmath3 dynamics , where only the interactions in the two - body @xmath1 and @xmath6 subsystems are taken into account , is of little use . in the present study we would like to demonstrate this fact for @xmath0 production in @xmath1 collisions .
\(ii ) what is the role of higher @xmath1 partial waves ? according to nakayama @xcite , the puzzling second bump in the @xmath19 invariant mass distribution of @xmath2 measured at @xmath20 mev in @xcite could be understood as arising mostly from a contribution of the @xmath21 state of the final nucleon pair .
however , the size of @xmath22 wave contribution appears to be comparable with that of the @xmath23 one ( see e.g. fig . 4 in @xcite ) .
such a large @xmath22 wave effect in the region where the @xmath1 c.m .
kinetic energy is limited to @xmath24 mev ( and in fact the maximum of the distribution of @xmath25 is located at even a considerably lower value because of the preference of higher @xmath0 momenta ) is quite surprising . therefore , we have investigated again the influence of @xmath5 , @xmath22 , and @xmath26 partial @xmath1 waves in both isospin states using the paris @xmath1 potential .
another point , closely connected to this , is the role of the short range behaviour of the @xmath1 interaction which is enhanced by quite large momentum transfers accompanying the @xmath0 production .
as will be shown below , this leads in particular to an important role of the tensor force in the final @xmath27 triplet state for @xmath18 because the @xmath28 state , excited in the @xmath0 production process , couples to the @xmath29 partial wave during subsequent @xmath1 rescattering .
first we review briefly in the next section the most essential formal ingredients on which our calculation is based . for the driving @xmath30 transition potential
we take a meson exchange model including @xmath31 , @xmath32 , and @xmath0 exchanges .
the two - body @xmath6 and @xmath1 interactions are parametrized in terms of separable @xmath33 matrices .
the dominant @xmath5 wave part of the @xmath3 wave function is calculated using three - body scattering theory . in sect . [ sect2 ] our results
are compared with experimental data and in sect .
[ sect4 ] the most important conclusions are summarized . in an appendix we list detailed expressions for the @xmath34 amplitude .
in the center of mass system the reaction to be studied reads in detail @xmath35 where the energies and momenta of the participating particles are explicitly defined for later purpose .
the reaction is governed by the transition matrix element @xmath36 where initial and final states are characterized by the total spin @xmath37 and isospin @xmath38 of the nucleon pair .
the calculation is done in the region of c.m .
excess energy @xmath39 up to 60 mev .
thus the use of nonrelativistic expressions for the energies of the final particles is justified , whereas for the initial state relativistic expressions are taken .
the c.m .
cross section for @xmath2 is then given by @xmath40 where the relative momentum of the two final nucleons @xmath41 is introduced .
the factor @xmath42 results from the average over initial spin states .
the corresponding @xmath18 cross section is obtained from ( [ 20 ] ) by the substitution @xmath43 for the @xmath0 production mechanism we have adopted the usual model where first an @xmath44 resonance is excited on one of the nucleons which subsequently decays into the @xmath6 channel .
the submatrix @xmath45 associated with the transition @xmath34 is expressed as a linear combination of eight pseudoscalar operators @xmath46 ( @xmath47 ) built from the vectors @xmath48 , @xmath49 , and spin operators @xmath50 and @xmath51 @xmath52 where @xmath49 denotes the relative momentum of the @xmath53 system .
the amplitudes @xmath54 are functions of only @xmath55 , @xmath56 and @xmath57 , and their specific form depends on the model for the @xmath0 production mechanism . for the operators
@xmath46 we have chosen the following forms @xmath58 } \otimes(\vec{p}^{\,\,(\prime)}\otimes(\vec{p}\times\vec{p}\,'))^{[2 ] } \right\}^{[0]}\,.\end{aligned}\ ] ] the latter two operators are tensor operators in two - nucleon spin space describing a spin transfer of 2 in the transition to the @xmath53 state .
for the impulse approximation ( ia ) , where fsi is ignored , the amplitude @xmath59 ( [ 15 ] ) is given by @xmath60 @xmath61 describes the propagation of the @xmath53 configuration followed by the decay @xmath62 . with respect to the spin structure of the amplitude ,
only three types of transitions are possible , namely @xmath63 , @xmath64 , and @xmath65 . the transition @xmath66 is excluded since it does not conserve parity .
it is easily seen that the operators @xmath67 through @xmath68 contribute to all three types of transitions .
the operators @xmath69 and @xmath70 , being antisymmetric under exchange of @xmath50 and @xmath51 , appear only in channels with a change of spin , i.e. @xmath63 and @xmath64 , whereas @xmath71 and @xmath72 , which require flipping of the spins of both nucleons , contribute only to @xmath65 . at low energies ,
the channel @xmath64 is strongly favored in the @xmath2 reaction since it allows the final nucleons to be left in a relative @xmath5 wave whose contribution is further increased by fsi . for the reaction @xmath18 both spin changing transitions are possible near threshold .
apart from a trivial isospin factor 3 appearing for @xmath73 , their relative importance depends crucially on the interference between @xmath31 and @xmath32 exchange @xcite . the transition @xmath65 , which requires either the nucleons to be in a relative @xmath22 state or the @xmath0 meson to be emitted at least in a @xmath22 state is suppressed at low energies . after these more formal considerations we will now turn to the outline of the @xmath0 production model used in the present work . the @xmath0 production mechanism via @xmath44 excitation through meson exchange was considered in a variety of previous studies @xcite . in the present work ,
we have adopted this approach including only @xmath31 , @xmath32 , and @xmath0 mesons whose contributions was investigated in greater detail than those of @xmath74 , @xmath75 and other mesons .
the corresponding lagrangians are taken in the customary form @xmath76 where @xmath77 denotes the four momentum of the exchanged meson and @xmath78 the relative meson nucleon three momentum with @xmath79 .
further , each meson nucleon vertex @xmath80 with @xmath81 contains a form factor of monopole type @xmath82 the corresponding @xmath83 transition vertices with @xmath84 are regularized by form factors of the form @xmath85 for the @xmath86 vertex the same form factor as for @xmath87 is used . with this meson exchange model
the functions @xmath54 appearing in ( [ 25 ] ) can be written as a sum over the participating mesons @xmath88 detailed expressions for @xmath89 @xmath90 are listed in the appendix . following the work of @xcite , the main calculation is based on the assumption that the @xmath34 transition is dominated by @xmath32 exchange .
the contribution of @xmath31 meson exchange to the amplitude is about three times smaller .
the @xmath0 meson with an @xmath3 coupling @xmath91 from @xcite provides only a tiny correction at the level of about 3 % .
the @xmath86 coupling constant @xmath92 was treated as a free parameter .
it was adjusted in such a way that the total cross section in both channels , @xmath19 and @xmath27 , is normalized to the corresponding data with the same normalization factor @xmath93 .
this factor takes effectively into account the damping of the initial beam via inelastic collisions other than @xmath94 .
the parameters appearing in ( [ 45 ] ) to ( [ 47 ] ) are listed in tables [ table1 ] and [ tab2 ] .
those for the @xmath80 couplings , used also in @xcite are given in @xcite , except for @xmath95 which was taken from @xcite .
the @xmath96 and @xmath97 parameters are taken from @xcite and @xcite as is explained below . with this set of parameters we obtain as normalization factor @xmath98 , which is appreciably smaller than the value @xmath99 of @xcite and also less than the values 0.23 and 0.2 obtained respectively in @xcite and @xcite . in view of some obvious limitations adopted here for the production potential , e.g. , neglect of direct emission of an @xmath0 meson by one of the nucleons as well as the contribution of other mesons ,
we do not put too much significance on this quite small value for the damping factor @xmath100 . clearly , the present model is only one of possible ways to parametrize the @xmath34 transition amplitude . as was discussed in @xcite ,
the isovector exchange with a dominance of @xmath32 over @xmath31 exchange is consistent with the observed ratio of about 6.5 of the @xmath18 to @xmath2 cross section @xcite as well as with the characteristic @xmath101-dependence of the @xmath0 angular distribution @xcite . for this model
the transition with @xmath73 strongly dominates over @xmath102 and , among others , results in a large probability of deuteron formation @xcite in agreement with experimental results @xcite .
however , as has been shown in @xcite and @xcite , more sophisticated treatments of the initial state interaction as well as inclusion of genuine two - body meson production mechanisms favour @xmath31 dominance .
therefore , in order to broaden our theoretical basis , we consider in addition the second type of @xmath34 potential , where the @xmath32 meson is neglected .
another purpose of using such an extreme approximation is to see to which extent the fsi effects depend on the details of the @xmath34 transition .
one might anticipate the result that due to the short - range nature of the @xmath34 interaction such a dependence can not be very significant .
however , in view of the basic theoretical uncertainties with respect to the production mechanism , this question has to be investigated quantitatively . indeed , as will be shown below , the sensitivity of fsi effects to the variation of the relative strength of @xmath31 to @xmath32 exchanges in the production potential , although not very dramatic , results in a variation of fsi effects of about 30 % .
our method of solving the three - body equations is based on separable representations of the driving two - body forces . for the @xmath6 interaction we use the isobar model of @xcite , where the resonance parametrization of the partial wave amplitudes is fitted to the @xmath103 , @xmath104 , and @xmath105 channels . in particular in the @xmath106 channel only the @xmath107 resonance was taken into account . apart from the original set of @xmath106 parameters presented in @xcite , which gives for the @xmath6 scattering length @xmath108 fm
, we considered another parametrization listed in table [ tab2 ] , which gives a value twice as large , namely @xmath109 fm , and at the same time allows one to describe well the above - listed reactions .
more details are given in @xcite .
these two sets of parameters are used to study the sensitivity of the result to the @xmath6 interaction model .
the @xmath1 interaction is included in @xmath5 , @xmath22 and @xmath26 waves in both isospin states , namely @xmath23 , @xmath110 ( @xmath111 ) , and @xmath112 waves for @xmath102 and @xmath29 , @xmath113 , and @xmath114 ( @xmath115 ) for @xmath73 .
the corresponding off - shell amplitudes were obtained using the separable representation of the paris potential from @xcite .
the remaining higher partial waves of the @xmath1 subsystem were taken as plane waves .
the faddeev equations for the @xmath3 system were solved only for the states of lowest angular momentum where a relative two - particle @xmath5 state , either @xmath116 or @xmath117 , is coupled with an @xmath5 state of the third particle relative to this pair , resulting in the partitions @xmath118 or @xmath119 , respectively .
as the explicit calculation shows , in the energy region of current interest , @xmath120 mev , these @xmath5 waves are almost totally overlapping .
the other partial waves were treated perturbatively up to first order in the @xmath6 and @xmath1 scattering amplitudes .
the separable representation of the two - body @xmath33 matrices allows one to reduce the original faddeev equations to a set of two integral equations in one variable only , presented schematically in fig .
their structure is identical to the usual lippmann - schwinger equation for the two coupled channels @xmath121 and @xmath122 , where the notation `` @xmath123 '' is used for the interacting @xmath1 subsystem .
the reaction amplitude @xmath124 is then expressed in terms of the auxiliary quasi - two - body amplitudes @xmath125 and @xmath126 as is presented in the same figure . with respect to the application of three - body scattering theory to reactions with three free particles in the final state
we refer to @xcite .
some technical points concerning the @xmath3 problem are described in @xcite .
we begin the discussion of our results with the total cross section shown in figs .
[ fig2 ] and [ fig3 ] for @xmath2 and @xmath18 , respectively .
as already noted , the three - body result is normalized to the data at one point , namely at @xmath127 mev to @xmath128b using promise - wasa data @xcite for @xmath19 and to @xmath129b with celsius data @xcite for @xmath27 .
of course , the ratio between the different curves is fixed by the model .
therefore , apart from the fact that the same normalization factor is used for both reactions , the only nontrivial result in figs . [ fig2 ] and [ fig3 ] is the energy dependence of the cross sections . for obvious reasons ,
detailed experimental data are available for the @xmath2 reaction only , which , however , is much more difficult for a theoretical study of the near threshold region because of the coulomb interaction between the protons . in the present work we have not included the coulomb force , which fact , of course , hampers to some extent a quantitative comparison with the data . nevertheless , the results shown in figs . [ fig2 ] and [ fig3 ] allows us to draw the following conclusions : the three - body model reproduces quite well the energy dependence of the experimental cross section in the @xmath19 channel .
the deviation very close to threshold must primarily be ascribed to the neglect of coulomb repulsion . as for the @xmath27 channel
, the agreement with the data of @xcite appears to be less satisfactory .
the theory exhibits a more flat form of the cross section .
clearly , for a meaningful conclusion about the quality of the present model a greater amount of data is needed , especially in the vicinity of the threshold .
the first - order calculation , where the fsi is reduced to only @xmath6 and @xmath1 rescatterings , does not reproduce the form of the @xmath2 cross section as predicted by the three - body treatment and also observed in the experiment . as was noted in the introduction ,
it is expected that the leading terms associated with @xmath6 and @xmath1 rescatterings do not constitute a reliable approximation of the whole multiple scattering series , due to the slow convergence of this series in the vicinity of the virtual @xmath3 pole .
the strong difference between the first order and the exact three - body calculation even at @xmath16 mev demonstrates the crucial importance of higher order terms .
we consider this result as a direct evidence for the fact that a three - body approach is definitely needed for a correct treatment of the @xmath3 interaction .
thus , the first - order approximation , although seemingly reasonable in view of the weak @xmath6 interaction as compared to the @xmath1 one , is in general not relevant .
the question of a correct treatment of fsi in @xmath0 production is closely related to the relative fraction of @xmath5 waves in the final state . as was mentioned above , in accordance with the short range hypothesis for the @xmath0 production mechanism ,
the @xmath5 waves in the @xmath3 system are expected to provide the major part of the cross section in the threshold region , thus making the fsi effect to be of particular importance in this reaction . in this context
it is instructive to compare the relative contribution of the @xmath5 waves to the total cross sections of @xmath130 and @xmath2 .
the results of the ia for both processes are presented in fig .
[ fig4 ] , where the @xmath131 contribution @xcite is shown separately . because of the relatively long range nature of the mechanism for @xmath0 photoproduction on the deuteron , the contribution of higher partial waves in the final state increases rapidly and already at the energy @xmath16 mev the @xmath5 wave provides only a 10 @xmath132 fraction of the total cross section .
in contrast to this , the @xmath131 configuration strongly dominates the reaction @xmath2 in the whole energy region , so that even at @xmath16 mev it produces about 77 % of the cross section . in conclusion ,
the role of the final state interaction depends not only on the phase space available for the final particles ( which is the same for both reactions ) but much more crucially on the range of the primary @xmath34 transition interaction . on the right panel of fig .
[ fig2 ] the dependence of the results on the model ingredients is exhibited .
firstly , we would like to note a visible sensitivity of the cross section to the choice of which meson exchange governs the production potential .
namely , @xmath31 dominance tends to make the form of the total cross section more convex as compared to the full curve where @xmath32 exchange prevails , whereas the ia gives in both cases practically indistinguishable results except for the normalization factor @xmath100 . using a weaker @xmath6 interaction from @xcite ,
we obtain , as expected , a weaker influence of fsi as is shown by the dash - dotted line in the right panel of fig .
[ fig2 ] . in fig .
[ fig5 ] we compare our calculation for the @xmath19 invariant mass distribution with the cosy-11 @xcite and tof @xcite data . at the excess energy @xmath133 mev the three - body result was normalized to the experimental integrated cross section @xmath134 @xmath135b of @xcite .
the distribution at @xmath20 mev is given in arbitrary units and the theoretical full curve was normalized to the area enclosed by the data points . for a lower excess energy @xmath136 mev one finds a reasonable agreement of the three - body calculation with the data , although one notes a systematic slight underestimation of the experimental cross section in the region above @xmath137 ( gev)@xmath138 .
at least some part of the deviation stems again from an overestimation of the @xmath19 interaction effect at the lower end of the spectrum due to the neglect of the coulomb force . in view of the overall normalization
, inclusion of the coulomb repulsion would reduce the @xmath23 peak near @xmath139 , and would enhance the other part of the distribution . however , at the higher excess energy @xmath140 mev the shape of the theoretical distribution is in strong disagreement with the data . in particular ,
the broad maximum around @xmath141 gev@xmath138 is not explained . of course , in order to reach at a definite conclusion , the coulomb interaction must be included in addition , but it is very unlikely that this would have such a strong effect as to explain the shape .
moreover , our results are at variance with those of @xcite where an appreciable contribution of higher angular momentum states in the @xmath19 subsystem is predicted , whereas our calculation does not indicate a sizeable influence from @xmath22 and @xmath26 partial @xmath1 waves . the corresponding contributions can be estimated by comparison of the full and the dash - dotted curves in fig .
[ fig5 ] . in the latter calculation
only the lowest @xmath131 configuration in the @xmath3 final state was included . on the other hand
, it is worth to underline the importance of the @xmath1 tensor force , which increases quite appreciably the transition to the @xmath29 state from an intermediate @xmath28 after @xmath0 production .
the relevance of this intermediate @xmath28 state is related to the strong momentum transfer accompanying @xmath0 production and emphasizing the high momentum components of the @xmath1 wave function .
one of the consequences of this fact is that the enhancement by the @xmath1 interaction for the @xmath73 transition is even slightly stronger than that for @xmath102 .
this result is at variance with @xcite , where a ratio of the enhancement factors of @xmath102 to @xmath73 of 1.85 at relative @xmath1 momenta @xmath142 @xmath143 was found . according to our results in fig .
[ fig6 ] , this ratio is reduced from 1.35 to only 0.98 at @xmath144 @xmath143 if the transition @xmath145 is included . in fig .
[ fig7 ] the calculated distribution with respect to the @xmath0 kinetic energy @xmath146 in @xmath2 is plotted together with the celsius data @xcite .
one readily notes quite a good agreement except for a very small region close to the maximal kinetic energy .
we do not consider this discrepancy as a serious defect because , as was noted in @xcite , some fraction of the events must have been lost .
furthermore , we would like to point out that the data of the @xmath19 mass distribution at @xmath140 mev in the right panel of fig .
[ fig5 ] appears to be in conflict with the data at @xmath147 mev in the right panel of fig .
the latter increases monotonically with meson energy except very close to the maximal kinetic energy @xmath146 , whereas the tof results for @xmath148 in fig .
[ fig5 ] exhibit a pronounced broad maximum at intermediate @xmath19 masses . in view of the simple relation @xmath149 one would expect on the other hand a similar behaviour for both distributions , because the small difference in the energies can hardly result in a very different form of the spectrum .
thus new experimental data for the @xmath146 and/or @xmath150 distributions in the region @xmath151 mev are needed in order to resolve this discrepancy . finally , we present in fig .
[ fig8 ] the angular distribution of an emitted proton for two excess energies .
the dominance of @xmath32 exchange in the production mechanism leads to a pronounced isotropy of the cross sections .
this becomes apparent by comparison with the angular distributions using a pure @xmath31 exchange .
they exhibit a much stronger angular dependence with a more pronounced peaking at forward and backward directions , which is consistent with a more periferal character of the @xmath31 exchange mechanism . in general , the proton angular distribution with a weak @xmath152 dependence is consistent with the almost isotropic form of the experimental results of @xcite , although the observed magnitude at @xmath140 mev is overestimated by about 20 % .
furthermore , one should keep in mind the pronounced disagreement between theory and the tof data for the @xmath153 spectrum at the same energy ( see fig .
[ fig5 ] ) . with respect to the @xmath0 angular distribution
, we did not analyze it here in view of its obvious sensitivity to higher @xmath6 resonances , predominantly @xmath154 , which were omitted in the present calculation .
in the present paper we have investigated @xmath0 production in @xmath1 collisions in the near threshold region .
the main attention was focussed on the understanding of the role of fsi , which is expected to strongly influence the reaction dynamics near threshold .
we believe that in conjunction with previous work @xcite the present study was necessary in order to establish a unified picture of this reaction .
in particular , it can be quite useful for the planning and analysis of future experiments . from our results
we draw the following conclusions : \(i ) the three - body treatment of the @xmath3 interaction provides a satisfactory , quantitative explanation of the energy dependence of the total cross section for @xmath2 .
however , the dependence of the results on the details of the @xmath34 transition introduces some uncertainty .
in particular , assuming a pion exchange dominance in the production potential leads to stronger energy dependence of the cross section and requires , therefore , a weaker @xmath6 interaction in order to bring the theory into agreement with the data .
\(ii ) there is a discrepancy between our results and the data of @xcite for the @xmath153 distribution at a c.m.excess energy @xmath140 mev .
namely , the broad peak of the data is not described by the theory . comparing to other theoretical work
, we do not reproduce the result of @xcite , where this broad structure is explained by the contribution of higher partial waves in the @xmath1 subsystem . on the contrary , according to our findings the contribution of @xmath22 and @xmath26 waves in the final @xmath19 subsystem plays a minor role only , at least in the region up to @xmath16 mev considered in the present work .
\(iii ) even close to threshold , where the final @xmath1 system is predominantly in an @xmath5 wave state , the tensor @xmath1 force in the @xmath155 channel must be taken into account .
as we have shown , the intermediate transition to a @xmath28 configuration with subsequent rescattering into the @xmath29 state enhances sizeably the @xmath1 interaction effect in the triplet state . before concluding
, we would like to note that the sensitivity of the fsi effect to the @xmath34 transition mentioned in ( i ) indicates that in principle the short range part of the final state wave function must be calculated more thoroughly than is done here .
obviously , the most straightforward way would be a three - body treatment of the whole @xmath94 process without restriction to the final state .
however it seems to be very difficult in view of the great amount of channels which have to be included in such an approach , not to mention the pure theoretical ambiguity in treating the @xmath1 system within the three - body formalism . on the other hand
, one can expect that short range mechanisms may be included perturbatively .
for example , the @xmath156 configurations could be considered in first order only ( apart from the driving meson exchange potential ) by including the additional mechanism where a pion , created in an @xmath1 collision , produces in turn an @xmath0 via rescattering on one of the nucleons .
clearly , such an approach , which would allow us to adopt the same fsi concept , is practically much easier to manage than a rigorous solution of the three - body coupled channel model .
this work was supported by the deutsche forschungsgemeinschaft ( sfb 443 ) .
in this appendix we give the explicit expressions for the amplitudes @xmath157 , @xmath158 for the transition @xmath34 determining the transition matrix element @xmath159 .
the latter was defined in the c.m.system with @xmath48 and @xmath49 denoting the momenta of the initial and final states , respectively . in the following expressions
we have introduced the notations : @xmath160 the expressions for @xmath32 exchange are obtained by using the on - shell gordon decomposition of the nucleon current @xmath161 u(\vec{p}\,)\ ] ] and the analogous relation for the @xmath86 coupling . in the present paper the invariant energy @xmath162 of the @xmath44 or @xmath6 subsystem was determined according to the spectator on - shell scheme @xmath163 which is consistent with the nonrelativistic three - body treatment of the @xmath3 system .
furthermore , this choice of @xmath162 justifies the use of the gordon decomposition for the tensor part in the @xmath87 coupling since in this case the nucleon remains on - shell after emission of a @xmath32 meson . the only approximation , whose validity is not justified is the gordon relation for the @xmath86 coupling .
we believe , however , that , since the coupling constant @xmath164 is treated here as a free parameter , it is a reasonable approximation .
\(ii ) @xmath0 meson exchange : since the formal expressions are the same as for @xmath31 exchange except for a replacement of coupling constants and meson mass , we do not need to repeat them here : @xmath167
\(iii ) @xmath32 meson exchange : @xmath168 where @xmath169\nonumber\\ & + & \frac{g_t}{2m_n\epsilon } \big(1-\frac{\vec{p}\cdot\vec{p}\,'}{\epsilon\epsilon'}\big ) \big[w+\epsilon'+\frac{p^{\prime\,2 } } { \epsilon^*}-\frac{p^\mu p^*_\mu}{\delta m^ * } \big]\,,\\ f_4 & = & \frac{g_v+g_t}{\epsilon^*}\big [ -1-\frac{\epsilon^*}{\epsilon'}-\frac{\vec{p}\cdot\vec r^{\,\prime}}{\epsilon } -\big(1+\frac{\vec{p}\cdot\vec{p}\,'}{\epsilon\epsilon'}\big ) \frac{p_0^*}{\delta m^*}-\frac{\vec r^{\,\prime}\cdot\vec p}{\delta m^ * } \big]\nonumber\\ & + & \frac{g_t}{2m_n\epsilon^ * } \big(1-\frac{\vec{p}\cdot\vec{p}\,'}{\epsilon\epsilon'}\big ) \big[w+\epsilon+\delta m^*+\frac{p^2}{\epsilon } + \frac{p^\mu p^*_\mu}{\delta m^*}\big ] \\ f_5 & = & -\frac{g_v+g_t}{\epsilon\epsilon ' } \big[2\epsilon'+\frac{\vec{p}\cdot\vec{p}\ , ' } { \epsilon}\big(1-\frac{\epsilon'}{\epsilon^*}\big ) + \frac{2p^{\prime\,2}}{\epsilon^ * } -\frac{2(w+m_n)}{(\delta m^*)}\,\vec p^{\,\prime}\cdot\vec q^{\,*}\big ] \nonumber\\ & - & \frac{g_t}{2m_n\epsilon\epsilon ' } \big[\frac{\vec{p}\cdot\vec{p}\,'}{\epsilon}\big(w+\epsilon'+\frac{p^{\prime\,2 } } { \epsilon^*}\big)+ \frac{p^{\prime\,2}}{\epsilon^*}\big(w+\epsilon'+\delta m^*+ \frac{p^2}{\epsilon}\big ) -\vec
p^{\,\prime}\cdot\vec q^{\,*}\,\frac{p^\mu p^*_\mu}{\delta m^*}\big ] \\
f_6 & = & \frac{g_v+g_t}{\epsilon\epsilon ' } \big[2\epsilon+\frac{p^2 } { \epsilon}\big(1+\frac{\epsilon'}{\epsilon^*}\big)- \frac{2(w+m_n)}{\delta m^*}\,\vec p\cdot\vec q^{\,*}\big ] \nonumber\\ & + & \frac{g_t}{2m_n\epsilon\epsilon ' } \big[\frac{\vec{p}\cdot\vec{p}\,'}{\epsilon^*}\big(w+\epsilon+\delta m^*+ \frac{p^2}{\epsilon}\big)+ \frac{p^2}{\epsilon}\big(w+\epsilon'+\frac{p^{\prime\,2 } } { \epsilon^*}\big ) -\vec
p^{\,\prime}\cdot\vec q^{\,*}\ , \frac{p^\mu p^*_\mu}{\delta m^*}\big ] \\
f_7 & = & \frac{g_v+g_t}{\epsilon^2\epsilon'}\big [ 1+\frac{\epsilon'}{\epsilon^ * } -\frac{2(w+m_n)}{\delta m^*}\big ] + \frac{g_t}{2m_n\epsilon^2\epsilon'}\big[w+\epsilon'+ \frac{p^{\prime\,2}}{\epsilon^*}- \frac{p^\mu p^*_\mu}{\delta m^*}\big ] \\
f_8 & = & \frac{g_v+g_t}{\epsilon\epsilon'\epsilon^*}\big[2 + \frac{2(w+m_n)}{\delta m^*}\big ] + \frac{g_t}{2m_n\epsilon\epsilon'\epsilon^ * } \big[w+\epsilon+ \delta m^*+\frac{p^2}{\epsilon}+ \frac{p^\mu p^*_\mu}{\delta m^*}\big]\end{aligned}\ ] ] left panel : total cross section for @xmath2 .
notation of curves : dotted : impulse approximation ; dashed : rescattering in two - body subsystems only ; dash - dotted : rescattering in @xmath1 subsystem only ; full : three - body calculation .
data from compilation @xcite .
open triangle : normalization point @xmath173 from @xcite .
inset shows near - threshold region .
right panel : dependence of three - body result on model ingredients .
notation of curves : solid identical to the one in the left panel ; dashed : without @xmath32 exchange in production potential ; dash - dotted : calculation with @xmath107 parametrization from @xcite with weaker @xmath6 interaction , @xmath174 fm , and @xmath31 dominated @xmath34 potential . ] distribution of squared @xmath19 invariant mass in @xmath175 .
notation of curves as in the left panel of fig .
the dash - dotted curve is obtained with a pure @xmath131 final state .
data at @xmath136 mev are from @xcite and at @xmath140 mev from @xcite . ]
enhancement factors associated with the @xmath1 interaction in @xmath73 and @xmath102 channels as function of the squared relative @xmath1 momentum . the dashed and dash - dotted curves are obtained with and without inclusion of the @xmath176 transition , respectively . ] | the production of @xmath0 mesons in @xmath1 collisions in the near threshold region is analyzed with special emphasis on the role of final state interaction which is calculated using three - body scattering theory .
the three - body aspects are shown to be very important in interpreting experimental results for the total cross section of @xmath2 .
furthermore , energy spectra and angular distributions of the emitted particles are calculated and compared to experimental data . |
figure [ figs1 ] shows all defect configurations examined in this work , which fall into eight categories : ( a ) se adatoms , ( b ) se@xmath27 dimers , ( c ) se@xmath35 antisites , ( d ) fe vacancies , ( e ) se interstitials , ( f ) se vacancies , ( g ) fe adatoms , and ( h ) fe interstitials .
their corresponding formation energies in the monolayer and bilayer fese supercells are plotted in fig .
[ figs2 ] .
se - rich configurations that are laterally centered at an fe site , consistent with the stm dumbbell signature , are enclosed in a red box in fig .
[ figs1 ] .
-symmetry about the fe site .
solid - line boxes mark the monolayer fese 4@xmath24 supercell boundaries . for ( b5 ) , ( e1 ) , ( e2 ) , ( f2 ) , and ( h1 ) , only a fraction of the bilayer fese 3@xmath23 supercell is shown for clarity .
fe atoms are violet and top / bottom se atoms are black / gray .
excess fe / se atoms are colored orange / red for distinction . ] 4 supercell monolayer and ( b ) 3@xmath23 supercell bilayer fese . reproduced from fig .
3 of the main text , but including additional defect structures from fig .
[ figs1 ] whose formation energy is less than 4 ev . ]
following ref .
@xcite , we estimate the experimental chemical potential of se using ideal gas approximations and tabulated thermodynamic quantities . under fixed temperature @xmath13 and pressure @xmath17 ,
the chemical potential of an ideal gas of se@xmath27 molecules is related to the gibbs free energy as @xmath62 where @xmath63 is the number of se@xmath27 molecules .
first , we compute @xmath64 at standard conditions ( @xmath65 = 298.15 k , @xmath66 = 1 bar ) , @xmath67 using enthalpy @xmath68 and entropy @xmath69 values derived from thermochemistry references .
next , we extrapolate to desired temperatures using reference heat capacity data , @xmath70,\ ] ] and desired pressures using ideal gas relationships , @xmath71 we make some practical remarks on estimating @xmath72 : \(1 ) to maintain consistency with dft calculations and total energies ( @xmath73 ) defined in vasp , we set @xmath74 for an isolated se _
\(2 ) based on the chosen reference , we compute @xmath68 in a two - step process : @xmath75 \(3 ) from p. 40 of ref .
@xcite , @xmath69 = @xmath76 jk@xmath77mol@xmath77 .
\(4 ) from p. 63 of ref .
@xcite , @xmath78 , where @xmath10 = @xmath79 jk@xmath77mol@xmath77 , @xmath11 = @xmath80 jk@xmath81mol@xmath77 , @xmath20 = @xmath82 jk@xmath83mol@xmath77 , @xmath16 = @xmath84 jmol@xmath77 , and @xmath85 = @xmath86 jkmol@xmath77 .
these values are valid from @xmath13 = 298 k to 1300 k. \(5 ) we choose @xmath13 to be the substrate temperature and @xmath17 to be the se partial pressure in the chamber , based on quasi - equilibrium growth assumptions suggested for similar iii - v semiconductor mbe processes @xcite . for @xmath13 = 400 @xmath12c and @xmath17 = 10@xmath87 t
, we find that @xmath34 = @xmath143.35 ev . for @xmath13 = 500 @xmath12c and @xmath17 = 10@xmath88 t ,
we find that @xmath34 = @xmath143.70 ev .
to simulate the stm topography of an fe vacancy site , we use dft to compute the real - space charge density , integrated from the fermi energy up to 50 mev .
we then trace the height variation ( @xmath89 ) for a given isosurface of charge density ( @xmath90 ) . as shown in fig .
[ figs3 ] , the value of @xmath90 chosen does not qualitatively affect the simulated topography , but we pick @xmath90 to be clearly in the regime of exponential decay . ) .
( f ) horizontal line cuts across the defect center .
@xmath91 is the distance measured from the fe plane .
the dark blue contour ( b ) depicts the isosurface chosen in fig .
4 of the manuscript .
( g ) plot of @xmath90 vs. the maximum height ( @xmath92 ) of the dumbbell impurity , demonstrating that our simulations are in the regime of exponential decay . ]
figure [ figs4 ] shows stm topographic simulations of an fe vacancy with different imaging biases @xmath93 , carried out by integrating the charge density from the fermi energy to @xmath94 .
we note that experimental fese bands exhibits orbital - dependent renormalization with an average factor @xmath95 = 1/6 relative to lda / gga - calculated bands @xcite , precluding a more detailed _ ab - initio _ analysis of stm defect signatures .
the classical probability distribution for an ensemble of identical , non - interacting particles moving in a 1d potential @xmath96 is given by @xmath97,\ ] ] where @xmath98 is the particle mass , @xmath99 is its velocity , and @xmath90 is a normalization constant .
we consider a finite spatial region in phase space , @xmath100 $ ] , with a potential barrier peaked at @xmath101 . the average transition rate @xmath102 is given by @xmath103 , where @xmath104 is the phase space current across the barrier and @xmath105 is the number of states in @xmath106 $ ] @xcite . from eq .
( [ eqdistr ] ) , we find that @xmath107\ ] ] and @xmath108 \bigg(\frac{k_b t}{2 \pi m}\bigg)^{1/2},\ ] ] which yields @xmath109 \bigg]^{-1 } \exp \bigg[-\frac{\phi(x_b)}{k_b t}\bigg].\ ] ] in our application to an fe atom hopping to a neighboring vacant site in fese , we take @xmath110 , and @xmath96 to be the dft total energy @xmath73 along the diffusion path , parameterized by the position @xmath111 of the moving fe atom @xcite .
we show that a @xmath1@xmath2@xmath1 ordering of fe vacancies in fese is thermodynamically stable under excess se flux . following the same reasoning in the main text , the formation energy of fe@xmath8se , where @xmath112 , is given by @xmath113 we consider two concentrations : se@xmath9 for two concentrations of fe vacancies : ( a ) @xmath111 = 0.03125 ( dilute ) , and ( b ) @xmath111 = 0.2 ( @xmath1@xmath2@xmath1 order ) .
( c ) assuming no condensation of bulk fe ( body - centered cubic ) or se ( se@xmath32 and se@xmath33 rings ) , chemical potential values @xmath34 are restricted between the dashed lines labeled bcc fe and se@xmath32/se@xmath33 .
@xmath34 is set to zero for an isolated se atom .
alternatively , shaded blue regions mark estimated @xmath34 values at typical substrate temperatures and se partial pressures , using ideal gas approximations and tabulated thermodynamic quantities . ]
figure [ figs5](c ) demonstrates that within a narrow range of viable @xmath45 values , the @xmath1@xmath2@xmath1 fe vacancy superstructure is more energetically favorable compared to a dilute fe vacancy .
9ifxundefined [ 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/physrevb.65.035406 [ * * , ( ) ] @noop _
_ , vol .
( , ) link:\doibase 10.1103/physrevb.59.15246 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.78.282 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.220506 [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.155106 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.026402 [ * * , ( ) ] \doibase doi:10.1016/0022 - 3697(57)90059 - 8 [ * * , ( ) ] link:\doibase 10.1103/physrevb.47.1659 [ * * , ( ) ] | the properties of iron - based superconductors ( fe - scs ) can be varied dramatically with the introduction of dopants and atomic defects . as a pressing example
, fese , parent phase of the highest-@xmath0 fe - sc , exhibits prevalent defects with atomic - scale `` dumbbell '' signatures as imaged by scanning tunneling microscopy ( stm ) .
these defects spoil superconductivity when their concentration exceeds 2.5% .
resolving their chemical identity is prerequisite to applications such as nanoscale patterning of superconducting / nonsuperconducting regions in fese , as well as fundamental questions such as the mechanism of superconductivity and the path by which the defects destroy it .
we use stm and density functional theory to characterize and identify the dumbbell defects .
in contrast to previous speculations about se adsorbates or substitutions , we find that an fe - site vacancy is the most energetically favorable defect in se - rich conditions , and reproduces our observed stm signature .
our calculations shed light more generally on the nature of se capping , the removal of fe vacancies via annealing , and their ordering into a @xmath1@xmath2@xmath1 superstructure in fese and related alkali - doped compounds .
fese , a member of the iron - based superconductors ( fe - scs ) with the simplest stoichiometry , lies at the vanguard of high-@xmath0 materials . on one hand
, its anomalous parent phase , with no static magnetic order @xcite , poses a fresh theoretical challenge @xcite . on the other hand , its plain , 2d - layered structure lends itself to bottom - up , nanoscale engineering of its electronic properties . as a striking example
, monolayer fese interfaced with srtio@xmath3 @xcite exhibits an order - of - magnitude enhancement in its transition temperature @xmath0 ( up to 109 k @xcite ) compared to its bulk value ( 8 k @xcite ) .
similar @xmath0 boosts up to 48 k have also been attained by depositing k adatoms @xcite , opening the door to all kinds of adatom modifications of fese .
more generally , defects in fe - scs are crucial to control @xmath0 @xcite , raise the critical current @xmath4 through vortex pinning @xcite , and also serve as microscopic probes of pairing symmetry @xcite .
furthermore , defect effects are typically enhanced in 2d systems .
an ultimate goal is to control precise placement of atomic defects , possibly through scanning probe lithography , as has been achieved with hydrogenated graphene @xcite , p dopants in si @xcite , and mn dopants in gaas @xcite . to similarly pattern nanostructures in fese , an atomistic understanding of defect formation in this material
is needed .
as an intriguing and urgent example , fese films grown by molecular beam epitaxy ( mbe ) exhibit prevalent defects with atomic - scale `` dumbbell '' signatures as imaged by scanning tunneling microscopy ( stm ) ( also called geometric dimers in ref .
@xcite ) .
they consist of two bright lobes on adjacent top - layer se sites [ figs .
[ fig1](d)-(f ) ] .
their concentration is highly tunable , increasing with excess se flux and decreasing with substrate temperature .
importantly , superconductivity emerges only when their concentration falls below 2.5% @xcite . despite the structural simplicity of fese , it is still unknown whether these dumbbell defects are se adsorbates , antisites , interstitials , or some other type of defect .
their identity is crucial to determine whether or not they can be engineered to define superconducting / nonsuperconducting regions in fese for nanoscale applications . here
we present an stm characterization of dumbbell defects and an exhaustive , first - principles investigation of candidate defect configurations . using density functional theory ( dft ) , we find that fe vacancies have the lowest formation energy . furthermore , our modeling shows that they perturb orbitals on neighboring se sites , producing dumbbell signatures when imaged by stm .
based on nudged elastic band calculations and 2d random walk simulations , we explain how fe vacancies can diffuse to the edge of terraces during vacuum annealing , consistent with experimental observations of reduced dumbbell density after annealing .
we further discuss implications for se capping of fese films for _ ex - situ _ applications .
we also connect our results to previous questions of vacancy ordering in fese and related alkali - doped compounds .
* methods . * films of fese were deposited via mbe on 6h - sic(0001 ) and srtio@xmath3(001 ) substrates , following established recipes @xcite .
the greater volatility of one element ( se ) over the other ( fe ) motivates two conditions for stoichiometric growth @xcite : first , by setting the substrate temperature between the source temperatures , @xmath5 , impinging fe with temperature @xmath6 will be adsorbed with sticking coefficient close to unity , while impinging se can stick only if they bind to free fe on the substrate [ fig .
[ fig1](a ) ] .
second , to compensate for high se losses and to mitigate excess fe clustering , typical molar flux ratios @xmath7 range from 5 to 20 .
post growth , the films were transfered _ in situ _ to a homebuilt stm and imaged at liquid nitrogen / helium temperatures . from figs .
[ fig1](d)-(f ) , we enumerate several characteristics of the dumbbell defects : first , their prevalence over any other kinds of defects suggests they are energetically favorable . in few - layer fese , this observation is independent of substrate , sic [ figs .
[ fig1](c)-(e ) ] or srtio@xmath3 [ fig .
[ fig1](f ) ] .
( we note that single - layer fese / srtio@xmath3 , with vastly different superconducting properties , exhibits a different set of defects @xcite . )
similar dumbbell defects have also been imaged in fese crystals grown by vapor transport @xcite and in li@xmath8fe@xmath9ohfese crystals grown by hydrothermal ion exchange @xcite .
second , the dumbbells are aligned along both the @xmath10 and @xmath11 axes of the 2-fe unit cell [ figs . [ fig1](d)-(f ) ] , pointing to their independence from a structural orthorhombic distortion @xcite and electronic nematic state in fese @xcite that break 90@xmath12 rotational symmetry .
third , our stm measurements up to @xmath13 = 84 k with bias voltages 10@xmath14100 mv demonstrate that the dumbbell signatures persist well above the superconducting state .
fourth , the dumbbell defects can be removed upon annealing , leaving behind pristine fese [ fig .
[ fig1](e ) ] . [ cols="^,^,^,^,^ " , ] we performed dft calculations using ` vasp ` @xcite .
we used the pbe exchange - correlation functional @xcite , and the projector augmented wave ( paw ) method , with fe 4@xmath15 , 3@xmath16 and se 4@xmath15 , 4@xmath17 electrons treated as valence .
an energy cutoff of 450 ev and methfessel - paxton smearing @xcite with @xmath18 = 0.1 ev were employed .
we modeled defects within freestanding monolayer and bilayer fese supercells ( details in table [ tab1 ] ) , with full relaxation of internal atomic coordinates ( corresponding to a magnitude of the force per atom @xmath19 0.025 ev / ) . to reproduce the experimental @xmath20-axis value
, we included van der waals corrections in the bilayer calculations using the dft - d2 method @xcite , with dispersion potential parameters taken from ref .
@xcite ( tested for bulk fese and fete ) .
* results . *
given the correlation of dumbbell defects with excess se flux , we examine candidate defects in which @xmath21 .
although the dumbbell signature is centered above an fe site , we explore all possible binding sites for completeness .
we begin with isolated se adatoms as the simplest class of se - rich defects . among three adsorption sites ( see supporting information ) , the hollow site in fese , directly above a bottom - layer se atom , is most stable [ fig .
[ fig2](a)-(b ) ] .
we compute the binding energy as @xmath22 where @xmath23 is the dft total energy of the system including the adatom , @xmath24 is the total energy of pristine fese within the same supercell , and @xmath25 is the energy of an isolated se atom .
we find that @xmath26 = @xmath143.14 ev ( @xmath143.02 ev ) for monolayer ( bilayer ) fese , which suggests chemisorption .
examining the relaxed structure [ fig .
[ fig2](b ) ] , we observe that the se adatom comes within bonding distance of neighboring fe atoms and induces local strain .
importantly , given that @xmath0 in fe - scs is highly sensitive to the fe - se / as height @xcite , our result points to a possible microscopic explanation of why amorphous se may be a poor capping material .
4 supercells ( a)-(e ) .
solid - line boxes mark the supercell boundaries . for the interstitial configuration ( f ) ,
only a fraction of the bilayer fese 3@xmath23 supercell is shown for clarity .
fe atoms are violet and top / bottom se atoms are black / gray .
excess se atoms are colored red for distinction . ]
we next examine adsorbed se@xmath27 dimers .
we find that among five possible adsorption geometries ( see supporting information ) , two are nearly degenerate , one of which has the se@xmath27 molecule centered above an fe site [ fig .
[ fig2](c ) ] . furthermore , with binding energy defined as @xmath28 where @xmath29 is the energy of an isolated se@xmath27 molecule , we calculate @xmath30 = @xmath140.39 ev ( @xmath140.69 ev ) for monolayer ( bilayer ) fese .
these values suggest that se@xmath27 dimers are weakly physisorbed and may have short adsorption lifetimes .
we contrast this result to the case of gaas(001)-(2@xmath24 ) , where surface dangling bonds can stabilize adsorbed as@xmath27 dimers or as@xmath31 tetramers with calculated binding energies up to @xmath141.6 ev @xcite .
such dangling bonds are absent in the top layer of fese . as a side note
, our dft calculations suggest that adsorption can be enhanced if two surface dimers cluster into se@xmath31 , but this would produce an unobserved four - lobe stm topographic signature . 4 supercell monolayer and ( b ) 3@xmath23 supercell bilayer fese .
capital letters correspond to labeled defects in fig .
[ fig2 ] . assuming no condensation of bulk fe ( body - centered cubic ) or se ( se@xmath32 and se@xmath33 rings ) ,
chemical potential values @xmath34 are restricted between the dashed lines labeled bcc fe and se@xmath32/se@xmath33 .
@xmath34 is set to zero for an isolated se atom .
alternatively , shaded blue regions mark estimated @xmath34 values at typical substrate temperatures and se partial pressures , using ideal gas approximations and tabulated thermodynamic quantities ( see supporting information ) . ] alternatively , some studies have proposed that a perturbation at the fe site ( either an unknown repulsive potential @xcite or se subsitution @xcite ) could affect the orbitals on neighbouring se atoms and generate a dumbbell signature .
we consider three possibilities : se@xmath35 antisites , fe - site vacancies , and se interstitials , perhaps binding to a surface - layer fe atom from below . as seen in fig .
[ fig2](d ) , the antisite configuration in the monolayer supercell produces pronounced distortions of nearby atoms .
fe atoms are pulled closer to the antisite , and se atoms are pushed away .
we note that the antisite could not be held in place in the bilayer supercell during structural relaxation .
figure [ fig2](e ) shows an fe vacancy .
figure [ fig2](f ) shows the most stable se interstitial configuration , where the excess se atom lies beneath a top - layer se site , not an fe site .
to compare formation energies @xmath36 among the aforementioned defects with variable stoichiometry , we include the energetic costs of incorporating @xmath37 ( @xmath38 ) additional fe ( se ) atoms from a reservoir into the defect : @xmath39 assuming quasi - equilibrium growth of fese and no bulk fe or se precipitation , we impose the following constraints on the chemical potentials : ( i ) @xmath40 ; ( ii ) @xmath41 ; ( iii ) @xmath42 .
eq .
( [ eqformation ] ) then yields @xmath43 where @xmath44 .
figures [ fig3](a ) , ( b ) show results for monolayer and bilayer fese supercells . in both cases the fe vacancy possesses the lowest formation energy by a margin of at least @xmath140.5
ev within estimated @xmath45 ranges . ) for comparison . set point : 50 mv , 500 pa ;
t = 5 k. ] having identified the fe vacancy as the lowest - energy , se - rich defect of fese , we considered whether it can produce a dumbbell signature .
figures [ fig4](a ) , ( b ) show a charge density isosurface , integrated from the fermi energy up to 50 mev . for improved accuracy , we increased the bz sampling to 8@xmath28@xmath21 and used tetrahedron smearing with blchl corrections @xcite .
due to the missing fe atom , orbitals on neighboring se atoms protrude further out .
we simulate an stm topography by tracing the height of the charge density isosurface . as seen in fig .
[ fig4](c ) , the two protruding se orbitals appear as bright lobes of a dumbbell , matching the experimental image [ fig .
[ fig4](d ) ] .
if the dumbbell defects are fe vacancies , they must also be capable of diffusing to the edge of typical film terraces at high temperatures , as implied by fig .
[ fig1](e ) . to elucidate this process , we performed nudged elastic band calculations to find the minimum energy path associated with fe vacancy hopping @xcite .
we used a smaller @xmath46 supercell with @xmath47 bz sampling .
we computed seven intermediate images , each relaxed with total force per atom ( tangential and chain ) @xmath19 0.025 ev / .
figure [ fig5 ] shows the relative energy along the diffusion path , with insets depicting initial , transition , and final states . in the transition state ,
two neighboring se atoms ( circled in blue ) are pushed above and below the plane , suggesting that vacancy diffusion may be easier on the surface than in the bulk , as expected .
we calculate the diffusion rate as @xmath48,\ ] ] where @xmath49 is the attempt frequency and @xmath50 = 1.69 ev is the barrier height . from vineyard transition - rate theory @xcite ( see supporting information ) , @xmath51 \bigg]^{-1},\ ] ] where @xmath52 is the mass of an fe atom and @xmath53 ( @xmath54 ) is the initial - state ( transition - state ) position of the hopping fe atom .
then for a random walk over a 2d lattice , the root - mean - square distance traveled after time @xmath55 is @xmath56 where @xmath57 . if we anneal at 450 @xmath12c for 2.5 h [ fig .
[ fig1](e ) ] , we estimate @xmath58 to be 950 .
this distance exceeds typical film island dimensions [ fig .
[ fig1](c ) ] , thereby explaining how dumbbell defects are removed upon annealing .
* discussion .
* we draw a final connection between dumbbell defects and fe vacancies . at large dumbbell concentrations ,
song _ et al . _
@xcite found that the defects ordered into a @xmath1@xmath2@xmath1 superstructure .
similarly , electron diffraction measurements of fese crystals , nanosheets , and nanowires have revealed various types of fe - vacancy order , including @xmath1@xmath2@xmath1@xmath21 @xcite .
given that the fe vacancy is the thermodynamically most stable defect , the closest packing of these vacancies would lead to a @xmath1@xmath2@xmath1 arrangement , because any closer packing would produce multi - vacancy defects ( two or more nearest neighbor fe atoms mising ) , which would likely destabilize the crystal altogether .
this argument provides an explanation of the @xmath1@xmath2@xmath1 pattern ( see supporting information for additional calculations ) .
the identification of the @xmath1@xmath2@xmath1 dumbbell superstructure with fe - vacancy order has further significance .
investigations of the related compound @xmath59fe@xmath60se@xmath27 ( @xmath61 = alkali metal ) , with enhanced @xmath0 up to 32 k @xcite , have been complicated by mesoscale phase separation into multiple fe - vacancy reconstructions @xcite .
our calculations suggest that fe vacancy order is not a pathological feature of @xmath59fe@xmath60se@xmath27 , but a phenomenon intrinsic to fese grown under excess se flux .
the crucial distinction is that in the latter case , fe vacancies can be removed upon annealing , while in the former , fe vacancy diffusion may be hindered by the buffer @xmath59 layers .
this additional flexibility in fese may afford better control of stoichiometric ( superconducting ) and ordered vacancy ( nonsuperconducting ) phases for nanoscale patterning . in conclusion ,
we have established the chemical identity of dumbbell defects that appear in mbe - grown fese under excess se flux and suppress superconductivity with concentrations greater than 2.5% .
our dft calculations show that fe vacancies ( 1 ) are energetically most favorable , ( 2 ) produce dumbbell signatures consistent with stm images , and ( 3 ) can diffuse to the edge of typical film islands with vacuum annealing .
these atomistic insights lay the foundation towards controlling precise placements of such defects .
we also reiterate that amorphous se may be a poor choice of capping material to perform _ ex - situ _ measurements due to induced distortions within the underlying fese . finally , we suggest a broader , microscopic connection between dumbbell defect phenomenology in fese and mesoscale phase separation in @xmath59fe@xmath60se@xmath27 .
* acknowledgements . *
we thank tetsuo hanaguri for suggesting to us the fe vacancy configuration .
we thank shiang fang , wei chen , matthew montemore and joerg rottler for useful conversations .
this work was supported by the national science foundation under grant no .
dmr-1231319 ( stc center for integrated quantum materials ) , and the gordon and betty moore foundation s epiqs initiative through grant no .
gbmf4536 .
computations were run on the odyssey cluster supported by the fas division of science , research computing group at harvard university .
e. k. acknowledge support by army research office ( aro - muri ) w911nf-14 - 1 - 0247 .
j. e. h. acknowledges support from the canadian institute for advanced research .
58ifxundefined [ 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:\doibase10.1038/nmat2491 [ * * , ( ) ] link:\doibase10.1038/nphys3434
[ * * , ( ) ] http://www.nature.com/doifinder/10.1038/nphys3456 [ * * , ( ) ] link:\doibase10.1103/physrevb.91.201105 [ * * , ( ) ] link:\doibase10.1103/physrevlett.115.116401 [ * * , ( ) ] http://stacks.iop.org/0256-307x/29/i=3/a=037402 [ * * , ( ) ] link:\doibase10.1038/nmat4153 [ * * , ( ) ] link:\doibase10.1073/pnas.0807325105 [ * * , ( ) ] http://www.nature.com/nmat/journal/v14/n8/full/nmat4302.html [ * * , ( ) ] link:\doibase10.1038/ncomms10840 [ * * , ( ) ] link:\doibase10.1021/ja800073 m [ * * , ( ) ] http://stacks.iop.org/0295-5075/84/i=3/a=37002 [ * * , ( ) ] link:\doibase10.1103/physrevlett.102.097002 [ * * , ( ) ] link:\doibase10.1103/physrevb.87.214519 [ * * , ( ) ] link:\doibase10.1103/revmodphys.78.373 [ * * , ( ) ] http://stacks.iop.org/0034-4885/63/i=10/a=202 [ * * , ( ) ] http://dx.doi.org/10.1021/nl902605t [ * * , ( ) ] http://science.sciencemag.org/content/335/6064/64 [ * * , ( ) ] link:\doibase10.1038/nature04971 [ * * , ( ) ] link:\doibase10.1103/physrevlett.109.137004 [ * * , ( ) ] link:\doibase10.1103/physrevb.84.020503 [ * * , ( ) ] @noop _ _ , ( ) link:\doibase10.1126/science.1202226 [ * * , ( ) ] link:\doibase10.1103/physrevlett.115.017002 [ * * , ( ) ] link:\doibase10.1103/physrevb.93.125129 [ * * , ( ) ] link:\doibase10.1073/pnas.1413477111 [ * * , ( ) ] link:\doibase10.1103/physrevx.5.031022 [ * * , ( ) ] link:\doibase10.1038/ncomms10565 [ * * , ( ) ] @noop ( ) , link:\doibase10.1103/physrevlett.103.057002 [ * * , ( ) ] link:\doibase10.1103/physrevb.90.121111 [ * * , ( ) ] link:\doibase10.1103/physrevlett.113.237001 [ * * , ( ) ] link:\doibase10.1103/physrevb.91.155106 [ * * , ( ) ] link:\doibase10.1103/physrevb.91.214503 [ * * , ( ) ] link:\doibase10.1103/physrevb.87.180505 [ * * , ( ) ] link:\doibase10.1016/0927 - 0256(96)00008 - 0 [ * * , ( ) ] link:\doibase10.1103/physrevb.54.11169 [ * * , ( ) ] link:\doibase10.1103/physrevlett.77.3865 [ * * , ( ) ] link:\doibase10.1103/physrevb.40.3616 [ * * , ( ) ] link:\doibase10.1002/jcc.20495 [ * * , ( ) ] link:\doibase10.1103/physrevb.87.184105 [ * * , ( ) ] link:\doibase10.1103/physrevb.81.205119 [ * * , ( ) ] link:\doibase10.1103/physrevlett.82.4886 [ * * , ( ) ] link:\doibase10.1103/physrevb.90.134520 [ * * , ( ) ] http://stacks.iop.org/0953-8984/26/i=26/a=265002 [ * * , ( ) ] link:\doibase10.1103/physrevb.49.16223 [ * * , ( ) ] link:\doibase10.1016/0039 - 6028(94)00731 - 4 [ * * , ( ) ] in @noop _ _ , ( , ) p. link:\doibase10.1016/0022 - 3697(57)90059 - 8
[ * * , ( ) ] link:\doibase10.1103/physrevlett.66.915 [ * * , ( ) ] http://www.pnas.org/content/111/1/63.abstract [ * * , ( ) ] link:\doibase10.1103/physrevb.82.180520 [ * * , ( ) ] link:\doibase10.1103/physrevb.83.060512 [ * * , ( ) ] link:\doibase10.1103/physrevb.83.140505 [ * * , ( ) ] link:\doibase10.1103/physrevb.84.060511 [ * * , ( ) ] link:\doibase10.1038/nphys2155 [ * * , ( ) ] link:\doibase10.1103/physrevlett.109.057003 [ * * , ( ) ] link:\doibase10.1038/ncomms2913 [ * * , ( ) ] * supporting information for : * * dumbbell defects in fese films : + a scanning tunneling microscopy and first - principles investigation * dennis huang , tatiana a. webb , can - li song , cui - zu chang , jagadeesh s. moodera , efthimios kaxiras , and jennifer e. hoffman |
many galactic black hole and neutron star sources in low mass x - ray binaries show both chaotic and quasi - periodic variability in their observed x - ray fluxes .
some of the quasi - periodic oscillations ( qpos ) are in the khz range and often come in pairs @xmath0 of _ twin peaks _ in the fourier power spectra ( e.g. , van der klis , 2000 ) . in all four black hole sources with twin peak khz qpos pairs , @xmath1 ( mcclintock & remillard , 2003 ) . for neutron stars
sources the ratio of twin - peak frequencies is mostly close to @xmath2 too ( abramowicz , bulik , bursa , kluniak , 2003 ) . based on these and other properties of qpos kluniak & abramowicz ( 2000 ) concluded that twin peak khz qpos are due to a resonance in the accretion disc oscillation modes and they noticed that the specific @xmath3 ratio could be a consequence of strong gravity . according to the standard shakura - sunyaev accretion disc model , matter spirals down into the central black hole along stream lines that are located almost on the equatorial plane @xmath4 , and that locally differ only slightly from a family of concentric circles @xmath5 .
the small deviations , @xmath6 , @xmath7 are governed , with accuracy to linear terms , by @xmath8 here the dot denotes the time derivative . for purely keplerian ( free ) motion @xmath9 , @xmath10 and
the above equations describe two uncoupled harmonic oscillators with the eigenfrequencies @xmath11 , @xmath12
. we shall start with an argument which appeals to physical intuition and which shows that the resonance now discussed is a very natural , indeed necessary , consequence of strong gravity .
assume that in thin discs , random fluctuations have @xmath13 thus , @xmath14 is a first order term in @xmath15 and should be included in the first order equation for vertical oscillations ( [ equation1 ] ) .
the equation now takes the form , @xmath16 \delta \theta = \delta a_{\theta},\ ] ] where @xmath17 is a known constant .
the first order equation for @xmath18 has the solution @xmath19 . inserting this in ( [ equation5 ] ) together with @xmath10 ,
one arrives to the mathieu equation ( @xmath20 is absorbed in @xmath17 ) , @xmath21 \delta \theta = 0,\ ] ] that describes the _
parametric resonance_. from the theory of the mathieu equation one knows that when @xmath22 the parametric resonance is excited ( landau & lifshitz , 1976 ) . the resonance is strongests for the smallest possible value of @xmath23 . because near black holes and neutron stars @xmath24 ( see figure [ fig1 ] ) , the smallest possible value for resonance is @xmath25 , which means that @xmath26 .
this is an example of a situation in which parametric resonance works in a thin accretion disc and it intuitively explains the observed 3:2 ratio ( kluniak and abramowicz , 2002 ) , because , obviously , @xmath27 of course , in real discs neither @xmath28 , nor @xmath10 exactly ; moreover in realistic situations , for purely geodesic motion ( @xmath29 ) the system does not show increasing amplitudes ( the higher terms prevent this from occurring ) .
+ however , one may expect that because these equations are approximately obeyed in thin discs , the parametric resonance will indeed be excited .
such a resonance was found in numerical simulations of oscillations in a nearly keplerian accretion disc by abramowicz et al .
( 2003 ) .
the equations roughly outlined in the previous section will now be derived .
the equations of the geodesic motion of a unit mass test - particle in spherical coordinates can be written in the form ( @xmath30 ) : @xmath31 where @xmath32 is the effective potential and @xmath33 is the energy . consider a geodesic on the equatorial plane @xmath34 and perturb it slightly , keeping constant the angular momentum .
@xmath35 here @xmath36 and @xmath37 denote small devations from the circular orbit , and to first order they describe two uncoupled harmonic oscillators with epicyclic eigenfrequencies : @xmath38 the ratio of the two frequencies is @xmath39 this does not happen in the newtonian case where the radial and vertical frequencies are always equal to the orbital frequency ( @xmath40 ) . when continuing to higher orders in the taylor expansion , the two motions become coupled , and this is why the so - called parametric resonance can arise .
the perturbed equations are nonlinear differential equations with constant coefficients which are the values of the different derivatives of the effective potential @xmath41 and the metric at equilibrium ( see appendix [ subsec : notation ] ) : @xmath42 in the case of schwarzschild s metric , @xmath43 the third order terms in the taylor expansion provide the damping which saturates the increasing amplitudes : this motivates stopping the expansion at this order of the perturbation .
+ in first approximation we used paczyski & wiita ( 1980 ) pseudo - newtonian model .
the perturbed equations in the model and in general relativity look the same , apart from those terms in @xmath44 which involve @xmath45 that here are absent . + on the geodesics @xmath46 ; however , in order to explore the features of the resonance , a small additional isotropic force , parametrized by @xmath47 $ ] , can be added ( @xcite ) ; this force at the moment has no physical meaning ( it may be a coupling due for example to pressure or viscosity ) , it is just a mathematical tool .
then @xmath48 when @xmath49 the previous equations are just the approximation of the geodesics , while when @xmath50 increases , they describe slight deviations from the free motion .
this method is a variation of the straightforward expansion which leads to a unformly valid approximate solution of systems of weakly nonlinear differential equations ; the underlying idea is to consider the expansion which represents the approximate solution to be a function of multiple independent variables , or scales , instead of a single variable .
the new `` time - like '' independent variables are defined : @xmath51 expressing the solution as a function of more variables , treated as independent , is an artifice to remove the secular terms , which would make the solution to explode unphysically .
+ by writing @xmath52 one makes a formal assumption of physical slow variation explicit .
indeed @xmath52 define progressively longer time scales , which are not negligible when @xmath53 is of the order of @xmath54 or longer .
the characteristic time scale of the orbital motion near a neutron star or a stellar mass black hole is short ( @xmath55ms ) : hence for @xmath56 of the order of @xmath57 the dynamical effects on the scale @xmath58 become important in the interval of @xmath59s .
+ the conditions to eliminate the secular terms give a simpler system of differential equations : for a set of initial conditions the solution to such a system is a sum of trigonometric functions .
the same solution can be found by assuming a priori that the asymptotic solution is the sum of trigonometric functions in which the dominant terms have frequencies not far from the original eigenfrequencies of the uncoupled harmonic oscillators ( method of lindstedt - poincar ) .
the corrections to the eigenfrequencies depend on the amplitudes and on the constant coefficients hence in a fourier transform of the geodesics perturbed in the vicinity of @xmath60 , two peaks centered near @xmath3 would be found , the position of which and the ratio depend on the perturbation ( @xmath61 ) and on the initial amplitudes .
this method was used to find the approximate solution to equations ( [ eqperturbategr ] ) ( both when @xmath62 , hence @xmath63 and not ) : @xmath64 the coefficients of the trigonometric functions ( @xmath65,@xmath66,@xmath67 .. ) are constants which depend on @xmath68 and the initial conditions
. for simplicity of notation here the initial phases are equal to zero .
the taylor expansion at constant angular momentum to third order near the stationary point of the effective potential leads to two nonlinear differential equations with constant coefficients : @xmath69 the polynomials of third degree ( @xmath70 ) which constitute the nonlinear part of the equation depend on the symmetry of the effective potential and of the metric .
+ the first part of the method of multiple scales consists in some algebraic calculations that highlight what are the so - called regions of resonance .
indeed the coefficients of the approximate solution are found to have in the denominators terms like @xmath71 .
these @xmath23 and @xmath72 can not take any value , but they depend on the symmetry of the effective potential , the metric and the perturbation via the @xmath73 which are present in the equations : the allowed @xmath23 and @xmath72 indentify the candidate regions of resonance .
the symmetry restricts further these regions : it could happen that the constant coefficients cancel and then there would be no resonance at all . this calculations performed for a general effective potential @xmath74 showed that the following resonances , usually present , can not appear in the * plane - symmetric * case : @xmath75 the * only * possible resonances to this degree of approximation are then : @xmath76 since in general relativity @xmath77 ( this is due to the curvature of the space ) , then the only possible resonance is the @xmath3 , which is indeed observed most of the times in black holes candidates . in the next approximation
a new resonance appears , @xmath78 , and one may find many other possible @xmath79 .
however a property of parametric resonance is that the region of instability becomes thinner when @xmath80 increases : the `` higher '' resonances are less probable ( @xcite ) .
this is why we do not observe them .
+ landau & lifshitz ( mechanics " 1976 ) , when studying systems similar to ours ( chapter @xmath81 par.@xmath82 ) with the non - geodesic terms @xmath62 , use the method of lindstedt - poincar and they conclude that these are the solution we are interested in , because in a closed system , without any source of energy , there can not be a spontaneus increment in the intensity of oscillations . in this sense
we asserted that _
no parametric resonance occurs for strictly geodesic motion"_(@xcite ) : the initial conditions of the perturbation of the geodesics are related to the energy ; even if the initially small two perturbations couple and exchange energy , their amplitudes can not become observable .
an approximate solution for schwarzschild s perturbed geodesics was obtained and the validity of the approximation was verified a posteriori for initial conditions of the order of @xmath57(which were used in the numerical integration ) .
+ in this solution two modes remain locked , with the two frequencies near to the @xmath3 ratio : the value of these frequencies is were we would see the twin peaks in the fourier power spectra of the observed signal .
the correction to the eigenfrequencies can be written in the following form ( higher order corrections can be neglected ) : @xmath83 @xmath84 = \nonumber \\ & & \frac{1}{8 \omega_r^2 \omega_{\theta } \{\omega_r^2- \omega_{\theta}^2)}\{2 [ c_{11}^2 \omega_r^2 - \tilde{c}_{21 } \omega_r^4 + \nonumber \\ & + & 4 \tilde{c}_{21 } \omega_r^2 \omega_{\theta}^2 + c_b^2 \omega_r^4 \omega_{\theta}^2 + c_{11}(- e_{20 } \omega_r^2 + \nonumber\\&+ & c_b \omega_r^4 + 4 e_{20 } \omega_{\theta}^2 - e_{r2 } \omega_r^4 + 4 e_{r2 } \omega_r^2 \omega_{\theta}^2 ) ] \nonumber \\ & & \tilde{\alpha}_r^2(0 ) + [ c_{11}(-3 e_{02 } \omega_r^2 + 8 e_{02 } \omega_{\theta}^2 + \nonumber \\ & - & \tilde{e}_{z2 } \omega_r^2 \omega_{\theta}^2 + 8 \tilde{e}_{z2 } \omega_{\theta}^4 ) + \omega_r^2 ( -3 \tilde{c}_{03 } \omega_r^2 + \nonumber \\ & + & 12 \tilde{c}_{03 } \omega_{\theta}^2 - 2 c_b e_{02 } \omega_{\theta}^2 + 2 c_b \tilde{e}_{z2 } \omega_{\theta}^4 ) ] \tilde{\alpha}_{\theta}^2(0)\}\:\ : , \end{aligned}\ ] ] @xmath85 = \nonumber \\ & & \frac{1}{24 \omega_r^3 ( \omega_r^2 - 4 \omega_{\theta}^2)}\{- ( \omega_r^2 -4 \omega_{\theta}^2)\nonumber \\ & [ & 10 e_{20}^2 + 9 e_{30 } \omega_r^2 + 10 e_{20 } e_{r2 } \omega_r^2 + \nonumber \\ & + & \omega_r^4 ( 3 e_{1re2 } + 4 e_{r2}^2 ) ] \tilde{\alpha}_r^2(0 ) + \nonumber \\&+ & 6 [ -e_{12}\omega_r^4 + 4 e_{12 } \omega_r^2 \omega_{\theta}^2 + \nonumber \\ & - & \tilde{e}_{1ze2}\omega_r^4 \omega_{\theta}^2 + 4 \tilde{e}_{1ze2}\omega_r^2 \omega_{\theta}^4 + \nonumber \\ & + & e_{02 } ( -2 e_{20}\omega_r^2 + 8 e_{20 } \omega_{\theta}^2 + \nonumber \\ & + & 4 c_b \omega_r^2 \omega_{\theta}^2 ) -2 e_{20 } \tilde{e}_{z2 } \omega_r^2 \omega_{\theta}^2 + \nonumber \\ & - & 2 c_b \omega_{\theta}^2 \omega_r^4 + 8 e _ { 20 } \tilde{e}_{z2 } \omega_{\theta}^4 + 4 c_b \tilde{e}_{z2 } \omega_r^2 \omega_{\theta}^4 + \nonumber \\ & + & 2 c_{11 } \omega_r^2 ( e_{02 } - \omega_{\theta}^2 \tilde{e}_{z2 } ) ] \tilde{\alpha}_{\theta}^2(0)\}\:\:.\end{aligned}\ ] ] the dependence on @xmath61 and @xmath68 is included in the constant coefficients : one can vary the initial conditions ( @xmath86 , @xmath87 ) and/or @xmath61 is included in the coefficients with tilde in the following way : @xmath88 @xmath89 and @xmath90 ( @xmath66 and @xmath65 in equations ( 14 ) ) are the initial values of the amplitudes of the zeroth order approximation ( the harmonic oscillators ) . ] , obtaining different values for the corrected frequencies and for their ratio . for small variations of these parameters ,
the dependence on them is almost linear . +
the analytic approximation in figure ( [ fig3 ] ) was obtained by fixing @xmath91 ( far from geodesics ) and @xmath92 : each point on the segment corresponds to the state reached by the system starting from differemt initial conditions ( @xmath93 $ ] and @xmath94 $ ] ) .
the same slope was matched numerically ( in the pseudo - newtonian case ) by fixing the initial conditions and by varying @xmath61 + the fact that the slope is smaller than @xmath3 depends on the choice of the initial conditions .
this analytic result is important in explaining the behaviour of the solution more than the numerical values : the fact that the observed ratios are not exactly in @xmath3 ratio is a feature of parametric resonance and the position of the centroid frequencies depends on the strength and features of the non - linear perturbation .
the taylor expansion of the relativistic geodesics to the third order leads to two coupled harmonic oscillators : the purely geodesics motion is stable , while when it is perturbed there are cases for which parametric resonance may occur . + using the method of multiple scales one can highlight that , owing to the curvature ( through the way it determines the effective potential ) , the first allowed resonance between the radial an the vertical epicyclic frequencies is the @xmath3 , in agreement with the numerical analysis .
+ moreover the deviation of the slope from @xmath3 is easily explained as a property of such a non - linear resonance + the analysis of this toy - model reinforces the theory that indeed the observed pairs of qpos may be due to parametric resonance , and finally to the strong gravitational field alone .
the work reported here is a part of my 2003 laurea thesis _
risonanza parametrica di dinamiche quasi geodetiche con campi gravitationali forti _ at the physics department , university of trieste .
i thank my external thesis supervisor ( _ correlatore _ ) , marek abramowicz , for suggesting the subject to me , the many discussions and his constant support .
i also thank my local supervisor ( _ relatore _ ) , fabio benatti , for his advice and help and wlodek kluniak for the help in revising this paper .
i was working on the problem at the university of trieste and several other institutions : at sissa ( trieste , italy ) , chalmers university ( gteborg , sweden ) , ukaff ( leicester university , england ) , and the cargese 2003 spring school on black holes ( cargese , france ) .
i thank all these institutions for hospitality .
this work have been supported by the eu grant in leicester , by ma s vr swedish goverment grant and by the polish kbn grant 2p03d01424 .
the equations studied in ( @xcite ) for the perturbed geodesics in the pseudo - newtonian potential were incorrect .
the correct ones are : @xmath95 this does not affect the numerical results : indeed the symmetry ( then the possible resonances ) is the same and the only difference is in the values of the initial conditions which lead to the resonance .
in equations ( [ eqperturbategr ] ) i named the constant coefficients in a way practical to remember their origin : for example @xmath96 is a coefficient in the equation @xmath97 ( it has the letter @xmath67 ) and it it is the coefficient of @xmath98 at the 2nd power and @xmath97 at the first power ( numbers @xmath99 ) .
every coefficient contains the derivatives of the effective potential and of the metric at equilibrium .
the explicit form of each coefficient is : @xmath100 @xmath101\:\:\:\ : , \nonumber\\ e_{20}&=&e^2[\frac{1}{2 r_0 ^ 2 } { \frac{\partial^{2 } u}{{\partial r}^2 } } + \frac{1}{4}(1-\frac{1}{r_0})\frac{\partial^3 u}{\partial{r^3 } } ] \:\:\:\ : , \nonumber\\ e_{z2}&=&r_0 - 1\:\:\:\ : , \nonumber\\ e_{30}&= & e^2 [ \frac{1}{4 r_0 ^ 2}\frac{\partial^3 u}{\partial{r^3}}-\frac{1}{2 r_0 ^ 3}{\frac{\partial^{2 } u}{{\partial r}^2 } } + \nonumber\\ & & + \frac{1}{12}(1-\frac{1}{r_0})\frac{\partial^4 u}{\partial{r^4 } } ] \:\:\:\ : , \nonumber\\ e_{1ze2}&=&1\:\:\:\ : , \nonumber\\ e_{12}&= & \frac{e^2}{4}[\frac{1}{r_0 ^ 2}\frac{\partial}{\partial{r } } { \frac{\partial^{2 } u}{{\partial \theta}^2 } } + \frac{\partial^2}{\partial r^2}{\frac{\partial^{2 } u}{{\partial \theta}^2 } } ( 1-\frac{1}{r_0 } ) ] \:\:\:\ : , \nonumber\\ e_{r2}&= & -\frac{1}{2 r_0}\frac{1}{r_0 - 1}\:\:\:\ : , \nonumber\\ e_{1re2}&= & \frac{1 - 2 r_0}{2 r_0 ^ 2 ( r_0 - 1)^2 } \:\:\:\ : , \nonumber\end{aligned}\ ] ] | using the method of multiple scales , one can derive an analytic solution that describes the behaviour of weakly coupled , non - linear oscillations in nearly keplerian discs around neutron stars or black holes close to the 3:2 orbital epicyclic resonance .
the solution obtained agrees with the previous numerical simulation .
such result may be relevant to the kilohertz quasi - periodic variability in x - ray fluxes observed in many galactic black hole and neutron star sources . with a particular choice of tunable parameters ,
the solution fits accurately the observational data for sco x-1 . |
the evidence for the existence of a dark sector in the universe has made the present era of cosmology fascinating and challenging . at present , the microscopic natures of dark energy and dark matter are still an open question , with both components only probed gravitationally @xcite . regarding dark matter , there are various particle physics candidates which may belong to the class of weakly interacting massive particles ( wimps ) or which may interact only gravitationally . on the other hand , for particle physics explanations of the observed dark energy ,
the main fundamental question is whether dark energy is a cosmological constant or a dynamical field . while in the former case
the equation of state parameter @xmath3 of dark energy is constant and equal to @xmath1 , in the latter case it is a function of redshift @xmath4 and in general differs from @xmath1 , providing a way to distinguish both fundamental types of dark energy experimentally .
such a deviation of @xmath3 from @xmath1 will be searched for in various future surveys , studying supernovae , baryonic acoustic oscillations , weak gravitational lensing effects , galaxy clusters , or other techniques @xcite .
one obstacle to relate dark energy more explicitly to particle physics is the fact that the amount of the observed energy density @xmath5 ev is _ much _ smaller than the `` theoretical expectation '' . in this paper
we will not address this `` cosmological constant '' problem , i.e. the suppression / cancellation of the various generic contributions to dark energy which are too large , but assume that it is resolved by some other mechanism .
our work is motivated by the intriguing observation that the energy density of dark energy is close to another very small scale in particle physics , namely the one of neutrino masses .
the discovery of flavour conversion of neutrinos from various sources , interpreted within the framework of neutrino oscillations , points to two mass eigenvalues of the light neutrinos above about @xmath6 ev and @xmath7 ev @xcite , while searches for neutrino masses from tritium @xmath8-decay and neutrinoless double @xmath8-decay yield an upper bound for each mass eigenvalue of roughly @xmath9 ev @xcite .
the proximity of the scale of dark energy to that of neutrino masses has inspired the proposal of the so - called mass varying neutrino ( mavan ) scenario @xcite . in this scenario
, the relic neutrino density contributes to the effective potential for the dynamical field which is called `` acceleron '' .
one consequence is that neutrino masses vary on cosmological time scales and , since the relation between neutrino masses and @xmath10 remains valid in earlier cosmic epochs , also dark energy behaves differently from a cosmological constant .
compared to other dynamical models of dark energy ( e.g. quintessence @xcite ) , in the mavan scenario one does not have to choose a mass for the dynamical field of the order of the present hubble parameter @xmath11 .
in fact it can be much larger , even of the order of the neutrino mass scale , due to the stabilising effect of the contribution of the relic neutrino density to the potential . as a result
, the field adiabatically tracks the minimum of the effective potential whose evolution is controlled by the time evolution of neutrino number density .
the mavan models have been investigated in many studies , e.g. regarding possible experimental signatures and constraints @xcite as well as model - building issues @xcite .
it has turned out that while there are many interesting possible signatures and attractive features , the scenario is tightly constrained by the requirement of consistency with late time structure formation . in particular , it has been pointed out @xcite that if neutrinos are non - relativistic today , due to fifth force effects the neutrinos would cluster at late time and finally form `` neutrino nuggets '' which would spoil the dark energy behavior of the neutrino fluid .
the goal of this paper is to investigate the phenomenology of a generalised scenario which we will refer to as mavan hybrid scenario , where in addition to a mavan sector a cosmological constant ( from a false vacuum ) is driving the accelerated expansion of the universe today . as we will show , for a generalised power law potential in the mavan sector the effective equation of state parameter @xmath0 as well as the fraction to which the dark energy is of dynamical nature is determined by the neutrino mass scale .
we will also analyse the possibility to realise neutrino dark energy with a `` high - scale '' seesaw mechanism , where the right - handed neutrino masses are close to the gut scale , and comment on how such models might help to suppress the formation of `` neutrino nuggets '' and resolve stability problems of mavan models with non - relativistic neutrinos . for the main part of the paper we will focus on neutrinos that are non - relativistic today .
the paper is organised as follows : in section 2 we will review the basics of the mavan scenario .
section 3 contains the definition of the phenomenological framework and the analysis is performed in section 4 . in section 5
we discuss the possibility of realising neutrino dark energy with a `` high - scale '' seesaw mechanism and in section 6 we comment on how this might help to solve stability issues for non - relativistic neutrinos .
section 7 contains our summary and conclusions .
in the basic mavan scenario a singlet fermion @xmath12 ( right - handed neutrino ) and a dynamical real scalar field @xmath13 ( called `` acceleron '' ) are introduced .
their dynamics gives rise to the acceleration of the universe and to neutrino masses which vary with time .
the basic form of the lagrangian is given by @xmath14 where @xmath15 is neutrino dirac mass , @xmath16 is a dimensionless coupling in dark sector and @xmath17 is the potential energy of the acceleron field .
the standard model neutrino @xmath18 as well as the singlet neutrino @xmath12 are both written as two - component left - chiral weyl spinor fields .
if @xmath19 , i.e. if the majorana mass of the right - handed neutrino is much larger than the neutrino dirac mass , one can effectively integrate out @xmath12 from the low energy theory which leads to the effective lagrangian @xmath20 the neutrino mass @xmath21 depends on the dynamics of @xmath13 .
taking into account the masses of the non - relativistic relic neutrinos we now include a term proportional to the relic neutrino number density @xmath22 to the effective potential , @xmath23 the effective potential @xmath24 for the acceleron @xmath13 has a minimum because the first term pushes @xmath13 to larger values whereas as the second term @xmath17 is assumed to favour smaller values of @xmath13 ( e.g.if @xmath25 ) . typically , in this type of models a sub - ev scale is introduced in order to realize the desired values of @xmath17 ( and finally of @xmath26 ) . as a consequence of the two contributions to the potential and since @xmath27 decreases with time , the field value of the acceleron field @xmath13 becomes smaller with time when it adiabatically tracks the minima of the potential ( if the mass of @xmath13 is above the hubble parameter ) .
the combined @xmath13 and neutrino fluid drives the accelerated expansion of the universe .
we remark that unlike in conventional quintessence models the scalar field does not roll down an extremely flat potential but it rather gets stabilised by the finite neutrino density . for this reason
, the slope of the potential can be much higher than in quintessence scenarios and the mass of the scalar field @xmath13 can even be as high as the neutrino mass scale , whereas in quintessence models it has to be as tiny as the hubble scale .
although the model can successfully explain the present acceleration of the universe with masses of the dynamical scalar field much larger than the hubble scale ( e.g. in the sub - ev range ) and provide a connection between dark energy and the physics of neutrino masses , it has also some shortcomings : for example , the potential is not stable under radiative corrections , unless new states appear with sub - ev masses , and it still has to be rather flat in order to be consistent with present constraints on the equation of state parameter .
furthermore , the mavan scenario with non - relativistic neutrinos has a problem with respect to the consistency with late time structure formation , as has been pointed out in ref .
if neutrinos are non - relativistic today , due to fifth force effects the neutrinos could cluster at late time and finally form `` neutrino nuggets '' which would spoil the dark energy behavior of the neutrino fluid .
however , it has been demonstrated that under certain conditions these problems might be overcome @xcite .
for example , in supersymmetric theories of neutrino dark energy the potential can be stable under radiative corrections ( for example in gauge mediated supersymmetry breaking where supersymmetry breaking is only transmitted to the dark sector via gravity such that the dark sector superpartners are light ) and stability problems with respect to neutrino density perturbations can be avoided if neutrinos are highly relativistic , although then the connection to neutrino masses is weakened since the two observed mass scales correspond to non - relativistic neutrinos .
after having reviewed the basics of the original mavan scenario in the last section , we will now set up the framework for our analysis .
it consists of a generalised mavan potential of power law type plus a constant vacuum energy contribution . in the following
we will specify it explicitly and comment on how it may be realised in classes of models as well as on the connection to the ( in principle ) experimentally measurable neutrino mass scale . in the following we will consider a rather general setup where neutrino dark energy is realised in a type of `` hybrid dark energy '' setup with the theory being in a false vacuum at present time .
this implies that in addition to the time varying contribution of the acceleron field there is an additional contribution of the energy density of the false vacuum .
motivated by the notion of providing a rather general framework for our study , we will therefore consider an effective potential for neutrino dark energy of the form @xmath28 where @xmath29 @xmath13 is the acceleron field as a real scalar field as in section [ sec : scenarios ] , unless stated otherwise .
] , @xmath30 is the energy density of the sm neutrinos today and @xmath31 is the scale factor .
the main ingredients are a constant term @xmath32 ( from a false vacuum ) , a potential for @xmath13 which is monotonically increasing with @xmath33 ( in our case a general power law potential @xmath34 ) and an effective term @xmath35 , where we assume @xmath36 in order to generalise the conventional seesaw relation .
we note that the potential of eq .
( [ potential ] ) is not the most general form of a potential that may arise in a realistic model . instead of the monomial potential that we are considering
, it can for instance also be a polynomial or logarithmic function of the field @xmath13 , or a combination of both . to motivate the appearance of the constant vacuum energy in the `` hybrid scenario '' by an example
, we may consider the following superpotential ( after ew symmetry breaking ) , @xmath37 where hats indicate superfields .
@xmath38 denotes the sm singlet ( right - handed neutrino ) superfield(s ) , and the most relevant contributions to the scalar potential stem from the f - terms @xmath39 .
the first term contributes the constant vacuum energy while the second term gives a mass term ( @xmath40 ) to the acceleron field .
together with the additional effective potential term @xmath41 , where @xmath42 is the vev of the higgs field which couples to the neutrinos , the potential for the acceleron has a minimum for a non - zero field value of @xmath13 .
@xmath43 also gives a time - varying mass to the right - handed neutrino , @xmath44 . while the above superpotential may serve as a simple example which could realise the generalised potential of eq .
( [ effective_potential ] ) with @xmath45 and @xmath46 , it should be kept in mind that that there is in principle a large variety of possible models which can give rise to other values of @xmath47 and @xmath8 ( as well as to other non - power - law forms of the potential ) . for example , different powers of @xmath13 may appear in the superpotential or the neutrino masses may originate from a different type of seesaw mechanism ( e.g. from a so - called double @xcite or multiple seesaw ) such that @xmath48 , and so on .
furthermore , one may anticipate that the parameters @xmath47 and @xmath8 in realistic values should be positive integers .
however , it is also possible that fractional values of the exponents appear , for example if the kinetic terms in a model are non - canonical and therefore a field transformation has to be performed in order to bring the kinetic terms back to canonical form .
we will discuss the possible origin of the generalised potential in more detail in the appendix . from neutrino oscillation experiments ,
two so - called mass squared differences are known : @xmath49 is known mainly from atmospheric neutrinos and @xmath50 from solar neutrino data and from the kamland experiment .
the experimental values are about @xmath51 , @xmath52 @xcite . with this nomenclature
we also know that @xmath53 , but @xmath54 can be the heaviest or the lightest of the mass eigenvalues .
for the sum @xmath55 defined as @xmath56 we know that there is a lower bound from oscillations ( saturated if @xmath57 and @xmath58 ev , @xmath59 ev ) and also an upper bound from neutrinoless double @xmath8 experiments of about @xmath60 ev ( with @xmath61 ev ) @xcite . finally , we have approximately @xmath62 \;.\end{aligned}\ ] ] on the other hand , the lightest of the neutrino mass eigenvalues ( @xmath63 or @xmath54 ) can be arbitrarily small .
one important question is whether this sum @xmath55 is relevant for the mavan setup , or whether the mass of only one of the light neutrinos matters . to answer this question , we have to go to the full three family scenario where we distinguish two possible cases . in one possible situation ( which we may call case i ) @xmath55 is indeed relevant .
this case applies when the vev of the field @xmath13 gives masses to all the right - handed neutrinos @xmath64 such that the part of the lagrangian relevant for neutrino masses reads for instance ( @xmath65 is a coupling matrix ) @xmath66 the seesaw suppressed mass matrix of the light neutrinos is then given by @xmath67 if we diagonalise @xmath68 , the effective potential reads @xmath69 where we have assumed that @xmath70 for all generations @xmath47 of neutrinos . on the other hand
, we could have the situation ( which we may call case ii ) that each @xmath64 gets its mass from a different scalar field , e.g. from the vevs of fields @xmath71 and @xmath72 and that we have a lagrangian of the form @xmath73 where @xmath74 are coupling constants .
now the light neutrino mass matrix would be given by @xmath75 in this example , only @xmath63 would be relevant for the effective potential for @xmath13 .. ] if consistency of the mavan scenario ( for a particular form of the potential ) would require a very light neutrino mass scale @xmath76 ev , this could be realised in case ii .
the relevant mass would then be @xmath63 for a so - called normal hierarchical spectrum or @xmath54 for an inverted hierarchical spectrum . in the following , we will mainly be interested in case i , since it has a particularly close link to the observed neutrino masses which manifests itself in the upper and lower experimental bounds on today s value of @xmath55 given in eq .
( [ eq : mbar_range ] ) .
we will therefore present our results using the notation for case i where the relevant neutrino mass scale is @xmath55 .
most of the formulae can readily be adopted to case ii , as long as neutrinos remain non - relativistic .
in this section we will investigate the dynamics and phenomenology of the mavan hybrid scenario with a general power law potential @xmath77 the observed small neutrino masses are generated by a version of the seesaw mechanism , and we will assume a general dependence of the neutrino masses on @xmath13 of the form @xmath78 where @xmath79 is the vev of the sm higgs field ( which is proportional to the dirac masses of the neutrinos ) .
the notation @xmath80 is introduced in section [ sec : connectiontomnu ] . as a first step of our analysis we consider the background dynamics of the scalar field @xmath13 in the mavan hybrid scenario defined in section [ sec : mavanhybrid ] , which leads to an effective potential for @xmath13 of the form @xmath81 where @xmath82 and where the index `` @xmath83 '' indicates todays values of the parameters .
the second term in the potential originates from the coupling between neutrino and the scalar field in the effective lagrangian . in the following discussions ,
if we do not mention the values of @xmath47 and @xmath8 , we will assume @xmath84 and @xmath85 which corresponds to the `` standard '' mavan scenario , i.e. where the potential for @xmath13 is quadratic and where the seesaw mechanism is of type i. we will also focus on the case where the neutrinos are non - relativistic today .
the friedmann equation in our scenario takes the form @xmath86 and the scalar field @xmath13 obeys the following modified klein gordon equation @xmath87 we note that while the dark matter density @xmath88 redshifts as @xmath89 , the time dependence of the neutrino energy density is non - trivial due to the time - dependence of @xmath13 .
now we use _ adiabaticity _ , i.e. the feature of mavan scenario that the acceleron field adiabatically follows the minimum of the effective potential since long time back in the cosmic history .
consequently , the condition @xmath90 determines the evolution of the scalar field to be @xmath91 the present value of @xmath92 can be found from the above equation and is given by @xmath93 which allows to derive the relation @xmath94 where @xmath95 has been used with @xmath4 being the redshift . from this expression
we can calculate the kinetic energy of the field compared to the total energy density carried by the field , @xmath96 we can see that if the field value of @xmath13 today is smaller than the planck scale @xmath97 , the kinetic energy of the field can consistently be neglected for the relevant time of the evolution of the dark energy is typically required with respect to the stability issues for non - relativistic neutrinos ( c.f .
discussion in section [ stability ] ) . ] .
in addition , a constant vacuum energy contribution further reduces the significance of the kinetic energy compared to the hubble constant during the evolution . we will discuss the adiabaticity condition of the acceleron field in section [ the mavan hybrid scenarios with high scale seesaw mechanism ] . using eq .
( [ sol1 ] ) and comparing the potential energy @xmath17 to the acceleron - dependent neutrino fluid density in eq .
( [ effectivevback ] ) , we find that the ratio of the cosmological density parameters of these two components is given by @xmath98 where we have substituted @xmath99 and @xmath100 .
we would like to highlight the result that the ratio of the energy densities from the interaction term and from the potential term @xmath17 is constant in time .
this feature , which is basically a by - product of the adiabaticity condition , will be useful to derive constraints on the model parameters from cosmological observations and it will lead to a particularly close connection to the neutrino mass scale in the mavan hybrid scenario with power law potentials . using the above results for the background evolution , we will now investigate how the parameters @xmath47 , @xmath8 and @xmath32 of the model are constrained by cosmological observations , in particularly regarding the dark energy equation of state , as well as by the present bounds on ( and eventually by a future measurement of ) the neutrino mass scale @xmath55 .
, as we will discuss below .
in fact , one `` smoking gun '' signal of mavan scenarios would be a `` cosmological upper bound '' on the sum of neutrino masses ( if they are assumed as constant in time ) which is incompatible with todays lower bound from terrestrial experiments .
] we will start by deriving the effective dark energy equation of state parameter @xmath0 .
the main point here is that when the dark energy equation of state is extracted from cosmological data sensitive to the hubble scale @xmath101 , it is always assumed that the dark matter component redshifts as @xmath102 .
massive neutrinos also contribute a small fraction to the dark matter today .
the total dark matter density today is then given by @xmath103 where @xmath104 is the amount of cold dark matter and where @xmath105 is calculated from the actual value of the neutrino mass @xmath106 ( which may be determined more precisely by future experiments ) , i.e. @xmath107 .
however , as we have already mentioned , the neutrino masses vary with time and therefore this contribution does not redshift as @xmath102 . consequently , an observer will not extract the intrinsic dark energy equation of state , but rather an effective one , @xmath108 .
more explicitly , the effective dark energy density ( which defines the effective equation of state @xmath108 ) is extracted from the hubble equation ( assuming for simplicity a flat universe ) @xmath109 where it is ( wrongly , but for a model - independent analysis unavoidably ) assumed that the neutrino density redshifts like ordinary matter .
we would like to add a few remarks to clarify the notation : @xmath105 ( i.e. the value of the quantity @xmath110 today ) is same as @xmath111 .
however , in an earlier epoch , the true neutrino energy density @xmath112 is _ not _ equal to the quantity @xmath113 . to calculate @xmath114 ,
following @xcite , we rewrite eq .
( [ friedmann1 ] ) as @xmath115 comparing this relation with eq .
( [ app ] ) yields @xmath116 from @xmath114 the observed equation of state @xmath108 can be calculated via the continuity equation @xmath117 taking the time derivative of eq .
( [ dark_energy_eff ] ) and using eq .
( [ kg ] ) and the expression for @xmath118 , we find @xmath119 we compare the above equation with eq .
( [ weffective ] ) to obtain @xmath120 this is the general result for the effective equation of state parameter @xmath0 for general @xmath13-dependent neutrino masses @xmath121 and for a general acceleron potential @xmath17 .
we now discuss some of the characteristic feature of @xmath108 : to start with , it can be seen from the above expression that at present time ( @xmath122 ) , the @xmath111-dependent term disappears from the denominator .
furthermore , in the absence of the constant vacuum energy term @xmath32 , the equation of state parameter at @xmath122 is simply given by the conventional equation of state parameter for the scalar field @xmath13 , @xmath123 since the slow - roll approximation can be applied for field values less than @xmath97 and since the kinetic energy @xmath124 is much smaller than @xmath17 in this case , @xmath125 is very close to -1 . regarding the effective equation of state parameter @xmath108 for @xmath122 , adding the constant vacuum energy term @xmath32 even makes its value closer to @xmath1 . for @xmath2 , however , the neutrino density term in the denominator of eq .
( [ eq : weffgeneral ] ) is non - zero .
furthermore , noting that in the mavan - type models of interest @xmath126 , the @xmath112-dependent term is negative which implies that @xmath108 becomes less than @xmath1 for @xmath2 . in summary , the effective equation of state parameter @xmath0
is defined by eq .
( [ eq : defweff ] ) , with @xmath114 related to the hubble parameter @xmath101 by eq .
( [ eq : wefffromhubble ] ) . by definition
, @xmath0 contains information about whether the effective dark energy density is dynamical or behaves like a cosmological constant at a given redshift @xmath4 . in order to extract @xmath114 from @xmath101 , additional information regarding the matter components
has to be used .
for example , sn data provides some information on @xmath101 .
however , @xmath127 is required as an input , in addition to the neutrino energy density @xmath128 which depends on the ( in principle ) measurable neutrino mass scale today .
one possibility to obtain @xmath104 and @xmath129 is to extract it from the cmb , where the neutrinos were relativistic and do not contribute to the matter density . on the other hand ,
if information on the matter density from other observations , e.g. from weak lensing , is used it may already contain @xmath130 if the neutrinos are clustered on the relevant scales .
we note that in order to confront the mavan hybrid scenarios with the whole variety of ( future ) data from cosmological observations , such issues have to be taken into account carefully . finally , in order to test a particular dark energy model of this type using all available cosmological data , one would have to go beyond the analysis of @xmath0 and fit all model parameters to the data .
we will now quantify the above general statements about @xmath0 for the more explicit scenario where the mavan potential has power law form , and investigate the further phenomenological consequences .
we start with rewriting the general form of the effective equation of state in terms of the parameters of the mavan hybrid scenario , which ( at this stage ) are @xmath131 and @xmath32 ( assuming further a fixed / measured value of @xmath55 ) . to do so , we use the results from our adiabatic background solution @xmath132 ( see eq . [ ratio ] ) and the relation @xmath133 which is plotted in fig .
( [ mnu_z ] ) for illustration . as the neutrino mass scale increases with time , neutrinos were relativistic at some earlier epoch .
for example , for @xmath134 , this happens around redshift @xmath135 .
we note that at this ( or larger ) redshift , the above - derived formulae do no longer apply since we have assumed non - relativistic neutrinos throughout .
the field @xmath13 remains stabilised when the neutrinos are relativistic , however at values which in general differ from those given in eq .
( [ eq : ainminimum ] ) . during cmb formation time
where @xmath136 , neutrinos were highly relativistic and furthermore their mass was much smaller than today , which implies that the effects of their mass density can be neglected during this epoch .
details about cosmological bounds from cmb data and large scale structure observations ( including neutrino masses ) can be found , e.g. , in @xcite .
using further the constraint from the observed dark energy density today , @xmath137 we find that @xmath138 where we have defined the parameter @xmath139 for the fractional contribution of @xmath32 to the total observed dark energy density today . after some straightforward calculations , we finally get the expression for the equation of state in terms of the parameters @xmath140 and as a function of redshift @xmath4 , @xmath141 we can see from the above expression that for @xmath142 , @xmath143 as this should be the case for a constant energy density . however for @xmath144 we find that @xmath0 deviates from @xmath1 for @xmath2 .
the amount of deviation depends on the choice of parameters @xmath145 .
we can also see that the value of the @xmath106 does not appear in eq .
( [ eq : weffgamma ] ) .
as we will show in the next subsection , @xmath106 is determined in this setup by the parameters @xmath146 and @xmath147 .
in fact , we use this to trade the parameter @xmath147 for the measurable quantity @xmath106 and re - express @xmath0 as a function of @xmath106 .
as mentioned above , the neutrino mass scale @xmath106 can be calculated as a function of the model parameters @xmath146 and @xmath147 . for non - relativistic neutrinos
( corresponding roughly to @xmath148 ev for @xmath149 ) the neutrino energy density is @xmath150 .
as we have mentioned in section [ sec : connectiontomnu ] , we are considering case i where the acceleron field provides the masses of all three right - handed neutrinos .
furthermore , the observed dark energy density today is given by @xmath151 .
using these two equations one can show that @xmath152 can be expressed in terms of @xmath147 as @xmath153 plugging in the observed value @xmath154 and the ( standard ) theoretical prediction for the neutrino number density today , @xmath155 @xcite , we find @xmath156 this simple expression relates the present value of the neutrino mass to the quantity @xmath147 which specifies which fraction of dark energy is constant vacuum energy . for fixed @xmath147 , the neutrino mass scale @xmath152
is predicted in this scenario .
one immediate consequence of the above relation is that for @xmath157 , i.e. without a constant energy density contribution @xmath158 , the predicted neutrino mass violates the the present experimental bounds , unless either @xmath47 is very small or @xmath8 is very large . in particular , for the standard case @xmath45 and @xmath46 , and for @xmath152 of the order @xmath159 ev , @xmath158 must contribute @xmath160% of todays dark energy of the total observed dark energy density ( c.f .
[ neutrino_constraint ] ) . on the other hand , for smaller @xmath161
the value of @xmath147 decreases and a larger fraction of dark energy is dynamical .
for example , for @xmath162 and @xmath163 ev , the dynamical component amounts about @xmath164% of the total dark energy density . , divided by their present value , as a function of redshift and of @xmath165 .
the right panel shows the value of @xmath166 as a function of the neutrino mass scale today and of @xmath165 .
, title="fig : " ] , divided by their present value , as a function of redshift and of @xmath165 .
the right panel shows the value of @xmath166 as a function of the neutrino mass scale today and of @xmath165 .
, title="fig : " ] [ neutrino_constraint ] using eq .
( [ m_nu ] ) we can express the effective equation of state parameter of eq .
( [ finaleos ] ) in terms of @xmath152 by eliminating the parameter @xmath147 , which yields @xmath167 in fig .
[ weffective_plot ] the @xmath4-dependent effective equation of state parameter @xmath108 is plotted for the case @xmath168 , @xmath85 and various values of @xmath152 .
the plots show the effective dark energy equation of state parameter @xmath108 as a function of redshift for different values of the neutrino mass scale @xmath169 .
the plots show the case @xmath170 and the left panel includes examples with @xmath171 ev whereas in the right panel values @xmath172 ev are presented.,title="fig : " ] the plots show the effective dark energy equation of state parameter @xmath108 as a function of redshift for different values of the neutrino mass scale @xmath169 .
the plots show the case @xmath170 and the left panel includes examples with @xmath171 ev whereas in the right panel values @xmath172 ev are presented.,title="fig : " ] for a dark energy equation of state which varies with time , a model independent extraction of @xmath0 is difficult from the currently available data . therefore , in most analyses of the present data , a constant @xmath3 is assumed .
one quantity which is used in the literature to compare with the best fit result under the assumption of a constant @xmath3 is the weighted ( or averaged ) equation of state parameter @xmath173 defined in the following way @xcite @xmath174 in fig .
[ w_av ] , @xmath175 is plotted as a function of @xmath152 for the case @xmath170 . the wmap 5 year
@xmath176 limit on the constant dark energy equation of state parameter is @xmath177 for zero curvature and @xmath178 when a nonzero value of the curvature is allowed @xcite . from the plot
we can see that even @xmath152 as large as @xmath60 ev is compatible with the present data .
averaged equation of state parameter @xmath173 as a function of neutrino mass scale @xmath169 .
the plot show the case where @xmath84 and @xmath85 . ]
one interesting result of the previous section is that with power law mavan potentials as in eq .
( [ effective_potential ] ) , the predictions for the observables do not depend on the value of @xmath13 today . with the masses @xmath179 of the right - handed neutrinos originating from terms in the potential of the form
@xmath180 this means that in principle the masses of the right - handed neutrinos today ( @xmath181 ) could be close to the gut scale and we might realise the mavan hybrid scenario with a `` high - scale '' seesaw mechanism .
compared to the usual mavan scenario where the right - handed neutrino masses are very small and where tiny neutrino yukawa couplings have to be postulated , in the `` high - scale '' seesaw the neutrino yukawa couplings can be of @xmath182 and the smallness of the neutrino masses is explained by the large masses of the right - handed neutrinos .
recalling eq .
( [ aoveranot1 ] ) , the scales @xmath92 and @xmath183 in the considered scenarios are connected to the neutrino mass scale @xmath152 by @xmath184 to check the adiabaticity condition for the case of large @xmath92 , we consider the square of the effective mass of the acceleron field today which is given by @xmath185 using eq .
( [ scales ] ) in eq .
( [ meffsq ] ) we can eliminate the dependence on @xmath183 and obtain @xmath186 now , for adiabaticity to hold , @xmath187 must be larger than the present value of the hubble scale @xmath188 . assuming for example @xmath168 and @xmath85 ( i.e. a simple mass term potential for @xmath13 and the standard seesaw relation ) , with @xmath92 around the gut scale @xmath189 gev and @xmath190 ev , @xmath187 is about @xmath191 ev , which is considerably heavier than the hubble scale . for the same choice of parameters and present vev of acceleron field , but with @xmath192 ev ; @xmath193 ev , still larger than the hubble scale .
this is in contrast to quintessence - type models , where the mass of the relevant scalar field has to be below the present hubble scale .
we would however like to note at this point that although the scenario has several attractive features , we have not attempted the construction of a viable model of this type .
one challenge is the radiative stability of the potential .
another challenge is the origin of the constant vacuum energy contribution .
both issues might be addressed in supersymmetric models ( for example in simple setups of the form of eq .
( [ eq : origin_1 ] ) ) , however even if hidden sector supersymmetry breaking is mediated to the sm fields by gauge interactions and to the dark sector only by gravity , this would typically induce a too large mass for @xmath13 of at least about @xmath194 ev . leaving these model - building questions aside
, we will now comment on how the `` high - scale '' seesaw might help with respect to stability issues for non - relativistic neutrinos .
in this section we will comment on the stability issue regarding the neutrino density perturbations , which are typically considered as a serious problem for mavan scenarios with non - relativistic neutrinos .
let us first briefly review the problem : soon after the introduction of the original mavan scenario @xcite , it has been pointed out @xcite that the original scenario faces a catastrophic instability for non - relativistic neutrinos due to the extra force carried between neutrinos and the acceleron field @xmath13 . with the mass @xmath195 of the field @xmath13 much larger than the hubble scale , it has been argued that the adiabatic perturbations at scales between @xmath196 are unstable .
let us now discuss how this problem might be overcome in the mavan hybrid scenarios with a `` high - scale '' seesaw mechanism . rather than attempting a full numerical analysis of the density perturbations of the neutrinos in this paper ( which we will leave for a future study ) , we will provide qualitative arguments and analytical estimates which will show how the mavan hybrid scenarios might help to suppress dangerous instabilities .
we will start with a general consideration regarding the coupling between the neutrinos and the acceleron field , and then turn to analytical criteria for the occurrence of the instabilities . one main difference between the mavan hybrid scenarios with power law type potentials ( where a `` high - scale '' seesaw mechanism is in principle possible ) and
the standard mavan scenario is that in the former the present value @xmath92 of the acceleron field can be much larger than the ev scale , even close to the gut or planck scale .
one can readily see that this might help with respect to the stability issue , when the heavy right - handed neutrino fields @xmath12 are integrated out the theory .
then , the interaction term between the light neutrino(s ) and the acceleron field becomes ( for , e.g. , @xmath197 ) @xmath198 where @xmath199 with @xmath200 denoting the quantum fluctuations around the classical value of the field @xmath92 . with @xmath201 , the term in eq .
( [ eq : stab_1 ] ) takes the form @xmath202 where @xmath203 . in the usual mavan scenario
the field value @xmath92 is not much larger than @xmath152 .
in contrast to this , as we have discussed in the previous section , in the mavan hybrid scenarios with power law potentials the present value of @xmath92 can in principle be as large as the gut scale ( or planck scale ) and the coupling between neutrinos and @xmath13 can be strongly suppressed ( by a factor @xmath204 ) .
let us now turn to the analytical criteria : first of all we note that when revisiting the stability issues of the mavan scenario , in ref .
@xcite bjaelde @xmath205 have argued that a negative value of the sound speed does @xmath206 always indicate the occurance of an instability . due to the dragging force of cold dark
matter , neutrino perturbations can remain stable even if the sound speed of the dark energy fluid becomes negative . more specifically , the dragging force due to the dark matter and baryons can stabilise the neutrino perturbations if the condition @xcite , @xmath207 is satisfied , where @xmath208 and where the coupling function @xmath209 is defined as @xmath210 with @xmath211 for the case @xmath46 . using the above definitions the stability condition of eq .
( [ compare ] ) can be rewritten as @xmath212 from this equation we can see that a small value of @xmath213 ( i.e. a small neutrino mass scale today ) as well as a larger value of @xmath214 compared to @xmath215 can help to stabilise the neutrino perturbations
. setting @xmath216 for simplicity and using eq .
( [ stabilitycondition ] ) , we find that , e.g. , for @xmath217 ev , @xmath209 has to be smaller than @xmath218 or for @xmath219 ev , @xmath209 has to be less than @xmath220 . in order to avoid the instability the present value of @xmath13 should therefore be larger than about @xmath221 for @xmath219 ev and @xmath222 for @xmath217 ev . for the case
@xmath219 ev this condition is well compatible with the adiabaticity condition in eq .
( [ kinoverdensity ] ) and the simple analytical consideration suggests that the instability problem might be cured . on the other hand , for @xmath223 the adiabaticity and the stability conditions can not safely be satisfied simultaneously and we conclude that whether the instability appears has to be checked numerically . and
it is therefore still negative in our model , however , as we have mentioned before we are following the argument of @xcite that despite a negative sound speed the instability can be avoided by the dragging force of dark matter . ] finally , we would like to note that , as we have discussed in section 4 , for the most standard types of potentials ( e.g. with @xmath224 ) the dynamical contribution to the total dark energy is rather small ( typically a few percent ) , and the main contribution arises from a constant vacuum energy density .
this means that even if the neutrinos would cluster , the observable effects on the smoothness of dark energy are much smaller than in the standard mavan scenario .
another related question is whether the scales on which neutrinos might cluster finally leads to any observable effect @xcite .
we conclude by remarking that although we have given some arguments which suggest that stability problems for non - relativistic neutrinos could be resolved in certain mavan hybrid scenarios with `` high - scale '' seesaw , this issue requires further investigations ( which are left for future studies ) .
motivated by the intriguing proximity of the energy density of dark energy and the neutrino mass scale we have studied the phenomenology of hybrid scenarios of neutrino dark energy , where in addition to a so - called mass varying neutrino ( mavan ) sector , a cosmological constant ( from a false vacuum ) is driving the accelerated expansion of the universe today . within the generalised framework we have focused on phenomenological issues such as on the connection to the neutrino mass scale and on its consequences for the dynamical nature of dark energy .
we have therefore calculated the effective equation of state parameter @xmath0 in the mavan hybrid scenario where the effective potential for the dynamical real scalar field ( the `` acceleron '' field @xmath13 ) has the following form @xmath225 we found that , for the case of a power law potential in the mavan sector , @xmath0 is determined by the neutrino mass scale and by the parameters @xmath47 and @xmath8 ( c.f .
( [ eq : weff_mnu ] ) ) . due to the interactions of the dark energy field with the neutrino sector , @xmath0 is predicted to become smaller than @xmath1 for increasing @xmath2 ( c.f .
[ weffective_plot ] ) , which could be tested in future cosmological observations . for the considered scenarios ,
we have also calculated how the neutrino mass scale determines which fraction of the dark energy is dynamical , and which originates from the `` cosmological constant like '' vacuum energy .
in particular , for the case of a mass term potential and a standard seesaw relation ( i.e. @xmath45 and @xmath46 ) we found that compatibility with the terrestrial neutrino mass bounds requires a large contribution of constant vacuum energy ( c.f .
( [ eq : mnuvsgamma ] ) ) . another interesting question , which we have investigated in the mavan hybrid scenario with power law potentials is whether it is possible to realise neutrino dark energy with a `` high - scale '' seesaw mechanism , where the right - handed neutrino masses are close to the gut scale .
we found that the field value of the `` acceleron '' field as well as the masses of the right - handed neutrinos can indeed be large and in principle a hybrid scenario of neutrino dark energy might be realised with a `` high - scale '' seesaw .
we have also commented on how the hybrid mavan scenarios with `` high - scale '' seesaw might allow to suppress the formation of `` neutrino nuggets '' and resolve stability problems of dark energy models with non - relativistic neutrinos . in summary , we have found that the considered hybrid scenarios of neutrino dark energy have several attractive features , in particular the close connection to the neutrino mass scale and an effective dark energy equation of state parameter @xmath0 which depends only on the parameters @xmath47 and @xmath8 of the potential and on the neutrino mass scale .
the prediction that @xmath226 for @xmath2 provides a `` smoking gun '' signal for such interacting dark energy scenarios , which could be observed in future surveys .
issues which are still open , and which are left for further studies , are whether a mavan hybrid scenario with `` high - scale '' seesaw can indeed solve the stability problems of conventional mavan scenarios with non - relativistic neutrinos and whether a consistent model can be constructed where this is realised .
we would like to thank neal weiner for useful discussions .
this work was partially supported by the cluster of excellence for fundamental physics `` origin and structure of the universe '' ( garching and munich ) .
sd is supported by us nsf career grant phy-0449818 and doe oji program under grant de - fg02 - 06er41417 .
sd acknowledges the hospitality of max - planck - institut fr physik ( werner - heisenberg - institut ) in munich where this work was initiated .
one possible origin of fractional powers in the potential in eq .
( [ potential ] ) is a non - canonically normalised real scalar field @xmath227 with the following lagrangian @xmath228 where @xmath229 . in the mavan hybrid scenario ,
the effective energy density of the field @xmath227 has the form @xmath230 where the last term arises form the smooth background energy density .
we have assumed @xmath231 .
now we would like to write it in terms of canonically normalised field @xmath13 , which satisfies @xmath232 integrating this equation we obtain the canonically normalised field @xmath13 in terms @xmath227 , and the effective potential in terms of the normalised field @xmath13 becomes , @xmath233 where @xmath234 which can be easily absorbed in the redefinition of @xmath183 . as a consequence of the non - canonically normalised scalar field we have obtained fractional powers in the mavan hybrid potential with @xmath235 and @xmath236 .
one motivation for introducing a constant term @xmath32 to the potential arises in supersymmetric `` hybrid - type '' models of dark energy . generalising the superpotential of eq .
( [ eq : origin_1 ] ) to three families , we may realise the situation that some of the `` waterfall fields '' @xmath64 are still in the false vacuum while others are in the true vacuum where @xmath237 .
a simple superpotential with this characteristic feature is the following @xmath238 in this case , the vacuum energy of the acceleron field does not account for the total dark energy , but only for part of it .
the other part stems from the @xmath64 which are in the false vacuum leading each to a contribution of energy density @xmath239 .
it is also interesting to note that the @xmath64 which have non - zero vevs generate a mass term for @xmath13 from the contribution @xmath43 to the scalar potential .
it is furthermore possible to generalise the setup of eq .
( [ eq : threefamilies ] ) such that higher powers of @xmath13 arise in the potential .
such higher powers can emerge from a superpotential of the form @xmath240 the most relevant parts of the scalar potential ( for a @xmath64 in the true vacuum ) would then be given by @xmath241 and @xmath242 .
this generalisation of the superpotential has two effects : firstly , @xmath243 leads to a higher power of a , explicitly @xmath244 , i.e.@xmath245 in the effective potential of eq .
( [ effective_potential ] ) and secondly , @xmath246 results in masses of the right - handed neutrinos of the form @xmath247 which leads to masses of the light neutrinos @xmath248 .
this implies @xmath249 in eq .
( [ effective_potential ] ) .
we note that another way to realise higher powers of @xmath8 would be to generate neutrino masses via a multiple seesaw ( see e.g.@xcite for a model of neutrino dark energy using the double seesaw mechanism ) .
50 m. colless _ et al . _ [ the 2dfgrs collaboration ] , `` the 2df galaxy redshift survey : spectra and redshifts , '' mon . not .
soc . * 328 * , 1039 ( 2001 ) [ arxiv : astro - ph/0106498 ] .
a. g. riess _ et al .
_ [ supernova search team collaboration ] , `` observational evidence from supernovae for an accelerating universe and a cosmological constant , '' astron .
j. * 116 * , 1009 ( 1998 ) [ arxiv : astro - ph/9805201 ] . s. perlmutter _ et al .
_ [ supernova cosmology project collaboration ] , `` measurements of omega and lambda from 42 high - redshift supernovae , '' astrophys . j. * 517 * , 565 ( 1999 ) [ arxiv : astro - ph/9812133 ] .
d. n. spergel _ et al . _
[ wmap collaboration ] , `` wilkinson microwave anisotropy probe ( wmap ) three year results : implications for cosmology , '' astrophys .
j. suppl .
* 170 * , 377 ( 2007 ) [ arxiv : astro - ph/0603449 ] .
a. albrecht _ et al .
_ , `` report of the dark energy task force , '' arxiv : astro - ph/0609591 .
r. n. mohapatra _ et al .
_ , `` theory of neutrinos : a white paper , '' rept .
phys . * 70 * , 1757 ( 2007 ) [ arxiv : hep - ph/0510213 ] .
r. fardon , a. e. nelson and n. weiner , `` dark energy from mass varying neutrinos , '' jcap * 0410 * , 005 ( 2004 ) [ arxiv : astro - ph/0309800 ] .
c. wetterich , `` cosmology and the fate of dilatation symmetry , '' nucl .
b * 302 * , 668 ( 1988 ) .
p. j. e. peebles and b. ratra , `` cosmology with a time variable cosmological constant , '' astrophys .
j. * 325 * , l17 ( 1988 ) .
r. r. caldwell , r. dave and p. j. steinhardt , `` cosmological imprint of an energy component with general equation - of - state , '' phys .
* 80 * , 1582 ( 1998 ) [ arxiv : astro - ph/9708069 ] .
d. b. kaplan , a. e. nelson and n. weiner , `` neutrino oscillations as a probe of dark energy , '' phys .
lett . * 93 * , 091801 ( 2004 ) [ arxiv : hep - ph/0401099 ] .
k. m. zurek , `` new matter effects in neutrino oscillation experiments , '' jhep * 0410 * , 058 ( 2004 ) [ arxiv : hep - ph/0405141 ] .
h. li , z. g. dai and x. m. zhang , `` testing mass varying neutrino with short gamma ray burst , '' phys .
d * 71 * , 113003 ( 2005 ) [ arxiv : hep - ph/0411228 ] .
v. barger , p. huber and d. marfatia , `` solar mass - varying neutrino oscillations , '' phys .
lett . * 95 * , 211802 ( 2005 ) [ arxiv : hep - ph/0502196 ] .
t. schwetz and w. winter , `` testing mass - varying neutrinos with reactor experiments , '' phys .
b * 633 * , 557 ( 2006 ) [ arxiv : hep - ph/0511177 ] .
a. ringwald and l. schrempp , `` probing neutrino dark energy with extremely high - energy cosmic neutrinos , '' jcap * 0610 * , 012 ( 2006 ) [ arxiv : astro - ph/0606316 ] . m. cirelli , m. c. gonzalez - garcia and c. pena - garay , `` mass varying neutrinos in the sun , '' nucl
. phys .
b * 719 * , 219 ( 2005 ) [ arxiv : hep - ph/0503028 ] . n. weiner and k. m. zurek , `` new matter effects and bbn constraints for mass varying neutrinos , '' phys .
d * 74 * , 023517 ( 2006 ) [ arxiv : hep - ph/0509201 ] .
a. w. brookfield , c. van de bruck , d. f. mota and d. tocchini - valentini , `` cosmology of mass - varying neutrinos driven by quintessence : theory and observations , '' phys .
d * 73 * , 083515 ( 2006 ) [ erratum - ibid .
d * 76 * , 049901 ( 2007 ) ] [ arxiv : astro - ph/0512367 ] .
a. e. bernardini and o. bertolami , `` stationary condition in a perturbative approach for mass varying neutrinos , '' phys .
b * 662 * , 97 ( 2008 ) [ arxiv:0802.4449 [ hep - ph ] ] .
r. fardon , a. e. nelson and n. weiner , supersymmetric theories of neutrino dark energy , jhep * 0603 * , 042 ( 2006 ) [ arxiv : hep - ph/0507235 ] .
r. takahashi and m. tanimoto , `` model of mass varying neutrinos in susy , '' phys .
b * 633 * , 675 ( 2006 ) [ arxiv : hep - ph/0507142 ] .
e. ma and u. sarkar , `` connecting dark energy to neutrinos with an observable higgs triplet , '' phys .
b * 638 * , 356 ( 2006 ) [ arxiv : hep - ph/0602116 ] .
r. d. peccei , `` neutrino models of dark energy , '' phys .
d * 71 * , 023527 ( 2005 ) [ arxiv : hep - ph/0411137 ] .
r. takahashi and m. tanimoto , `` false vacuum in the supersymmetric mass varying neutrinos model , '' phys .
d * 77 * , 045015 ( 2008 ) [ arxiv:0709.2965 [ hep - ph ] ] .
o. e. bjaelde and s. hannestad , `` neutrino dark energy with more than one neutrino species , '' arxiv:0806.2146 [ astro - ph ] . n. afshordi , m. zaldarriaga and k. kohri , `` on the stability of dark energy with mass - varying neutrinos , '' phys . rev .
d * 72 * , 065024 ( 2005 ) [ arxiv : astro - ph/0506663 ] .
l. amendola , m. baldi and c. wetterich , `` growing matter , '' phys .
d * 78 * , 023015 ( 2008 ) [ arxiv:0706.3064 [ astro - ph ] ] . c. wetterich ,
`` growing neutrinos and cosmological selection , '' phys .
b * 655 * , 201 ( 2007 ) [ arxiv:0706.4427 [ hep - ph ] ] .
j. r. bhatt , p. h. gu , u. sarkar and s. k. singh , `` left - right symmetric model of neutrino dark energy , '' phys .
b * 663 * , 83 ( 2008 ) [ arxiv:0711.2728 [ hep - ph ] ] .
s. das , p. s. corasaniti and j. khoury , `` super - acceleration as signature of dark sector interaction , '' phys .
d * 73 * , 083509 ( 2006 ) [ arxiv : astro - ph/0510628 ] .
k. ichiki and y. y. keum , `` neutrino masses from cosmological probes in interacting neutrino dark - energy models , '' jhep * 0806 * , 058 ( 2008 ) [ arxiv:0803.2274 [ hep - ph ] ] .
e. w. kolb and m. s. turner , `` the early universe .
reprints , '' _ usa : addison - wesley ( 1988 ) ( frontiers in physics , 70 ) _
e. komatsu _ et al .
_ [ wmap collaboration ] , `` five - year wilkinson microwave anisotropy probe ( wmap 1 ) observations : cosmological interpretation , '' arxiv:0803.0547 [ astro - ph ]
. o. e. bjaelde , a. w. brookfield , c. van de bruck , s. hannestad , d. f. mota , l. schrempp and d. tocchini - valentini , `` neutrino dark energy revisiting the stability issue , '' jcap * 0801 * , 026 ( 2008 ) [ arxiv:0705.2018 [ astro - ph ] ] .
r. takahashi and m. tanimoto , `` speed of sound in the mass varying neutrinos scenario , '' jhep * 0605 * , 021 ( 2006 ) [ arxiv : astro - ph/0601119 ] .
d. f. mota , v. pettorino , g. robbers and c. wetterich , `` neutrino clustering in growing neutrino quintessence , '' phys
. lett .
b * 663 * , 160 ( 2008 ) [ arxiv:0802.1515 [ astro - ph ] ] . | we study the phenomenology of hybrid scenarios of neutrino dark energy , where in addition to a so - called mass varying neutrino ( mavan ) sector a cosmological constant ( from a false vacuum ) is driving the accelerated expansion of the universe today . for general power law potentials
we calculate the effective equation of state parameter @xmath0 in terms of the neutrino mass scale . due to the interaction of the dark energy field ( `` acceleron '' ) with the neutrino sector , @xmath0 is predicted to become smaller than @xmath1 for @xmath2 , which could be tested in future cosmological observations . for the considered scenarios , the neutrino mass scale additionally determines which fraction of the dark energy is dynamical , and which originates from the `` cosmological constant like '' vacuum energy of the false vacuum .
on the other hand , the field value of the `` acceleron '' field today as well as the masses of the right - handed neutrinos , which appear in the seesaw - type mechanism for small neutrino masses , are not fixed .
this , in principle , allows to realise hybrid scenarios of neutrino dark energy with a `` high - scale '' seesaw where the right - handed neutrino masses are close to the gut scale .
we also comment on how mavan hybrid scenarios with `` high - scale '' seesaw might help to resolve stability problems of dark energy models with non - relativistic neutrinos .
mpp-2008 - 90 + arxiv:0807.4930 + |
we assume that the initial state of the computation is the standard state @xmath13 , and the preparation of the input state is included in the computational part .
therefore , we can assume without loss of generality that what alice wants to hide are the computation angles of the measurement - based quantum computation and the final output of the computation .
intuitively , a protocol is blind if bob , given all the classical and quantum information during the protocol , can not learn anything about alice s computational angles and the output @xcite . *
definition * : _ in this paper , we call a protocol is blind if _ * the conditional probability distribution of alice s computational angles , given all the classical information bob can obtain during the protocol , and given the measurement results of any povms which bob may perform on his system at any stage of the protocol , is equal to the a priori probability distribution of alice s computational angles , and * the conditional probability distribution of the final output of alice s algorithm , given all the classical information bob can obtain during the protocol , and given the measurement results of any povms which bob may perform on his system at any stage of the protocol , is equal to the a priori probability distribution of the final output of alice s algorithm .
@xmath14 intuitively , this means that bob s certainty " about alice s information does not changed even if bob does povm on his system . * theorem 1 * : _ protocol 1 satisfies ( b1 ) . _
* proof * : let @xmath15 be the random variable which represents alice s measurement angles , and @xmath16 be the random variable which represents the type of the povm which bob performs on his system .
let @xmath17 be the random variable which represents the result of bob s povm .
the two - party system is called no - signaling @xcite from alice to bob iff @xmath18 for all @xmath19 , @xmath20 , @xmath21 , and @xmath11 .
then , @xmath22 this means that bob can not learn anything about alice s measurement angles . @xmath14
* theorem 2 * : _ protocol 1 satisfies ( b2 ) .
_ * proof * : let @xmath23 be the random variable which represents the output of alice s algorithm , and @xmath16 be the random variable which represents the type of the povm which bob performs on his system .
let @xmath17 be the random variable which represents the result of bob s povm .
alice can change the output of her algorithm by changing the input .
( for example , since what is implemented in the quantum computation is a unitary operation , two input states which are orthogonal with each other become two mutually - orthogonal output states . )
because of the no - signaling principle , @xmath24 for all @xmath19 , @xmath25 , @xmath26 , and @xmath11 .
then , @xmath27 therefore , bob can not learn anything about the output of alice s algorithm .
@xmath14 0
* theorem 3 * : _ protocol 2 satisfies ( b1 ) .
_ * proof * : let us define @xmath28 where @xmath29 and @xmath30 are random variables , corresponding to the angles sent by alice to bob , alice s measurement angles , alice s secret computational angles , and the hidden binary parameters , respectively . from the construction of the protocol @xcite , the following relation is satisfied : @xmath31 where @xmath32 is alice s @xmath33th measurement result .
let @xmath34 be a povm which bob may perform on his system .
let @xmath35 be the random variable corresponding to the result of the povm .
let @xmath36 be the random variable which represents alice s message about the channel loss .
bob s knowledge about alice s computational angles is given by the conditional probability distribution of @xmath37 given @xmath38 , @xmath39 , and @xmath40 : @xmath41 where @xmath42 and @xmath43 . from bayes theorem , we have @xmath44 where @xmath45 and @xmath46 @xmath14 * proof * : let @xmath16 be the random variable which represents the type of the povm which bob performs on his system , and @xmath17 be the random variable which represents the result of the povm .
let @xmath36 be the random variable which represents alice s message to bob about the channel loss .
let @xmath15 be the random variable which represents alice s measurement angles .
bob s knowledge about alice s measurement angles is given by the conditional probability distribution of @xmath93 given @xmath94 , @xmath95 and @xmath39 : @xmath96 from bayes theorem , we have @xmath97 @xmath14 * proof * : let @xmath16 be the random variable which represents the type of the povm which bob performs on his system , and @xmath17 be the random variable which represents the the result of the povm .
let @xmath36 be the random variable which represents alice s message to bob about the channel loss .
let @xmath23 be the random variable which represents the output of alice s algorithm .
bob s knowledge about the output of alice s algorithm is given by the conditional probability distribution of @xmath98 given @xmath94 , @xmath95 , and @xmath39 : @xmath99 from bayes theorem , we have @xmath100 @xmath14 r. raussendorf and h. j. briegel , phys .
lett . * 86 * , 5188 ( 2001 ) .
r. raussendorf , d. e. browne , and h. j. briegel , phys .
a * 68 * , 022312 ( 2003 ) .
r. raussendorf , ph.d . thesis , ludwig - maximillians universitt mnchen , 2003 .
it is insightful to point out here that one great advantage of the measurement - based quantum computation which enables a simple blind protocol is the clear separation between the quantum stage ( resource preparation ) and the classical stage ( measurements and classical - signal feedforward ) . in order to see it ,
let us consider the following story .
alice and bob are in the same laboratory .
bob prepares the resource state , and he measures a particle .
then alice hammers bob , and bob loses his memory .
bob does the second measurement after waking up and again alice hammers bob .
if they repeat this process until the end of the computation , alice finally gets the output of the desired quantum computation . however , bob gets no information about this quantum computation since he loses his memory every time .
b. huttner , n. imoto , n. gisin , and t. mor , phys .
rev . a * 51 * , 1863 ( 1995 ) . c. bennett , f. bessette , g. brassard , l. salvail , and j. smolin , j. cryptology * 5 * , 3 ( 1992 ) .
g. brassard , n. ltkenhaus , t. mor , and b. c. sanders , phys .
. lett . * 85 * , 1330 ( 2000 ) .
n. ltkenhaus , phys .
a * 61 * , 052304 ( 2000 ) .
d. gross and j. eisert , phys .
* 98 * , 220503 ( 2007 ) .
d. gross , j. eisert , n. schuch , and d. perez - garcia , phys .
a * 76 * , 052315 ( 2007 ) .
d. gross and j. eisert , phys .
a * 82 * , 040303(r ) ( 2010 ) .
t. morimae , phys .
a * 83 * , 042337 ( 2011 ) .
t. morimae , phys .
a * 85 * , 062328 ( 2012 ) .
k. fujii and t. morimae , phys .
a * 85 * , 032338 ( 2012 ) .
t. morimae and k. fujii , sci . rep .
* 2 * , 508 ( 2012 ) .
it is interesting to note that measurement - based quantum computation is the sequence of teleportations @xcite , and therefore if alice does not send her measurement result , bob s state is the completely - mixed state .
the security of protocol 1 is easily understood from the following logic . in protocol 1 , alice can choose two strategies : ( 1 ) doing the correct measurements .
( 2 ) just discarding all particles sent from bob .
let @xmath101 @xmath102 be bob s state after alice finishing the strategy @xmath103 . of course , @xmath104 , since otherwise alice can transmit some message to bob . and
, obviously , @xmath105 does not contain any information about alice s input , output and algorithm , since no quantum computation was performed .
therefore @xmath106 does not contain any information about alice s computation .
protocol 2 runs as follows : ( 1 ) bob prepares the bell pair and sends the half of it to alice . ( 2 ) if alice does not receive it , she asks bob to try again , and goes back to ( 1 ) .
( 3 ) if alice receives the particle , she does the measurement in the @xmath54 basis . how to choose @xmath4 will be explained later .
after her measurement , bob has the state @xmath55 where @xmath56 , @xmath57 is alice s measurement result , and @xmath58 note that @xmath59 .
( 4 ) bob creates the state @xmath60 where @xmath61 is the c@xmath1 gate between the first and the second qubits and @xmath62 is any bob s state .
( 5 ) bob performs the measurement in the basis @xmath8 on the first qubit .
since @xmath63 commutes with @xmath61 , bob obtains @xmath64 if the measurement result is @xmath65 , where @xmath66 is the hadamard gate . 0 note that @xmath75 where @xmath76 is a pauli byproduct .
( be careful that different pauli byproducts are represented by the same character @xmath76 for simplicity . ) therefore , if alice and bob repeat the above ( 1)-(5 ) three times , @xmath77 can be deterministically implemented ( up to some pauli byproduct ) on any state @xmath62 . as is shown in ref .
@xcite , universal quantum computation is possible on the brickwork state @xcite if @xmath78 is possible .
note that @xmath79 furthermore , if @xmath80 or @xmath81 @xmath82 and if @xmath83 or @xmath84 @xmath85 therefore , if alice and bob repeat the above ( 1)-(5 ) for a constant number of times , they can perform the universal quantum computation .
note that @xmath86 where @xmath76 is a pauli byproduct .
( be careful that different pauli byproducts are represented by the same character @xmath76 for simplicity . )
this means that the operations @xmath87 can be done deterministically ( up to pauli byproducts ) if alice and bob repeat the above ( 1)-(5 ) three times .
therefore , if we consider the unit cell ( fig .
[ unitcell ] ) , the operations @xmath88 can be implemented deterministically up to some pauli byproducts as is shown in fig .
[ unitcell2 ] .
note that this set is universal set , since @xmath89 where @xmath90 is @xmath81 or @xmath9 , and @xmath91 is @xmath1 or @xmath84 . as is shown in fig .
[ tiling ] , the unit cell can be tiled to create the universal two - dimensional graph state which resembles the brickwork state @xcite . | blind quantum computation is a new secure quantum computing protocol which enables alice who does not have sufficient quantum technology to delegate her quantum computation to bob who has a fully - fledged quantum computer in such a way that bob can not learn anything about alice s input , output , and algorithm . in previous protocols , alice needs to have a device which generates quantum states , such as single - photon states .
here we propose another type of blind computing protocol where alice does only measurements , such as the polarization measurements with a threshold detector .
in several experimental setups , such as optical systems , the measurement of a state is much easier than the generation of a single - qubit state .
therefore our protocols ease alice s burden .
furthermore , the security of our protocol is based on the no - signaling principle , which is more fundamental than quantum physics .
finally , our protocols are device independent in the sense that alice does not need to trust her measurement device in order to guarantee the security . a first - generation quantum computer
will be implemented in the cloud " style , since only limited number of groups , such as governments and huge industries , will be able to possess it .
how can a client of such a cloud quantum computing assure the security of his / her privacy ?
protocols of blind quantum computation @xcite provide a solution .
blind quantum computation is a new secure quantum computing protocol which enables a client ( alice ) who has only a classical computer or a primitive quantum device which is not sufficient for universal quantum computation to delegate her computation to a server ( bob ) who has a fully - fledged quantum computer without leaking any alice s privacy ( i.e. , which algorithm alice wants to run , which value alice inputs , and what is the output of the computation ) to bob @xcite .
the first example of blind quantum computation was proposed by childs @xcite where the quantum circuit model was used , and the register state was encrypted with quantum one - time pad @xcite so that bob who performs quantum gates learns nothing about information in the quantum register . in this method
, however , alice needs to have a quantum memory and the ability to perform the swap gate .
the protocol proposed by arrighi and salvail @xcite is that for the calculation of certain classical functions , i.e. , not the universal quantum computation , and it requires alice to prepare and measure multi - qubit entangled states .
furthermore , it is cheat - sensitive , i.e. , bob can gain information if he does not mind being caught .
finally , in their protocol , bob knows the unitary which alice wants to implement .
aharonov , ben - or and eban s protocol @xcite requires a constant - sized quantum computer with a quantum memory for alice .
on the other hand , in 2009 , broadbent , fitzsimons and kashefi @xcite proposed a new blind quantum computation protocol which uses the one - way model @xcite . in their protocol ,
all alice needs are a classical computer and a primitive quantum device , which emits randomly rotated single - qubit states .
in particular , alice does not require any quantum memory and the protocol is unconditionally secure ( i.e. , alice s input , output , and algorithm are secret to bob whatever bob does ) .
recently , this protocol has been experimentally demonstrated in an optical system @xcite .
furthermore , this innovative protocol has inspired several new other protocols which can enjoy more robust blind quantum computation . in ref .
@xcite , two protocols which enable blind measurement - based quantum computation on the affleck - kennedy - lieb - tasaki ( aklt ) state @xcite have been proposed . in ref .
@xcite , a protocol of the blind topological measurement - based quantum computation @xcite has been proposed . due to the topological protection
, it is fault - tolerant @xcite .
the error threshold of the blind topological model has been shown to be comparable to that of the original @xcite ( i.e. , non - blind ) topological quantum computation @xcite . before starting the main part of this paper ,
let us quickly review the protocol of ref . @xcite . in this protocol ,
alice and bob share a classical channel and a quantum channel .
the protocol runs as follows : ( 1 ) alice prepares randomly - rotated single - qubit states @xmath0 is a random angle , and sends them to bob through the quantum channel . ( 2 ) bob creates a certain two - dimensional graph state , which is called the brickwork state @xcite , by applying the c@xmath1 gates among @xmath2 .
( 3 ) alice calculates the measurement angle on her classical computer , and sends it to bob through the classical channel . ( 4 ) bob performs the measurement in that angle , and returns the measurement result to alice .
( 5 ) they repeat ( 3)-(4 ) until the computation is finished .
if bob is honest , alice obtains the correct answer of her desired quantum computation @xcite .
furthermore , it was shown that whatever evil bob does , bob can not learn anything about alice s input , output , and algorithm @xcite .
the motivation of the blind quantum computation is to enable alice , who does not have any sophisticated technology and enough knowledge , to perform universal quantum computation .
therefore , there are two important goals .
one is to make alice s device as classical as possible , since alice is not expected to have any expensive laboratory which can maintain the coherence of complicated quantum experimental setups .
the other is to exempt alice from the precise verification of her device , since alice is not expected to have enough technology and knowledge to verify her device .
such a verification is important since she might buy the device from a company which is under the control of bob , and therefore the device might not work as alice expects .
for example , if alice is supposed to send a single - photon to bob , alice must confirm that more than two identical photons are not sent to bob , since otherwise bob might be able to gain some information by using , e.g. , the photon - number - splitting ( pns ) attack @xcite , which is a well - known technique in quantum key distribution ( qkd ) . in ref .
@xcite , a first step to the first goal , namely making alice s device more classical , was achieved .
they proposed an ingenious protocol of the blind quantum computation in which what alice needs to prepare are not single - photon states but coherent states .
since coherent states are considered to be more classical than single - photon states , their protocol allows alice s device to be more classical . in this paper
, we show that alice who has only a measurement device can perform the blind quantum computation . in several experimental setups , such as quantum optical systems , the measurement of a state , e.g. ,
the polarization measurement of photons with a threshold detector , is much easier than the generation of a single - qubit state , such as a single - photon state .
therefore , our results achieve the above mentioned first goal , namely making alice s device more classical . as we will see later
, our protocols can cope with the particle loss in the quantum channel between alice and bob and the measurement inefficiencies of alice s measurement device .
it also means that our protocols allow alice s device to be more classical .
we propose two new protocols , protocol 1 and protocol 2 .
importantly , the security of protocol 1 is based on the no - signaling principle @xcite , which is more fundamental than quantum physics @xcite .
therefore , even if bob does a super - quantum ( but no - signaling ) attack , alice s privacy is still guaranteed .
furthermore , the device - independence @xcite is attained for the security of protocol 1 .
hence the above second goal is also achieved .
the security based on the no - signaling principle and device - independence are important subjects in quantum key distribution , and much researches have been done within the decade @xcite .
however , protocol 1 can not cope with a high channel loss .
protocol 2 can tolerate any high channel loss , but the device - independence becomes weaker .
_ protocol 1_. our first protocol runs as follows : ( 1 ) bob prepares a resource state of measurement - based quantum computation . any resource state can be used for this purpose . for example
, the two - dimensional cluster state @xcite , the three - dimensional cluster state for the topological quantum computation @xcite , the thermal equilibrium states of a nearest - neighbour two - body hamiltonian with spin-2 and spin-3/2 particles @xcite or solely with spin-3/2 particles @xcite at a finite temperature for the topological measurement - based quantum computation , resource states for the quantum computational tensor network @xcite , the one - dimensional or two - dimensional aklt states @xcite , the tri - cluster state @xcite , and states in the haldane phase @xcite .
( 2 ) bob sends a particle of the resource state to alice through the quantum channel .
( 3 ) alice measures the particle in a certain angle which is determined by the algorithm in her mind .
they repeat ( 2)-(3 ) until the computation is finished .
obviously , at the end of the computation , alice obtains the correct answer of her desired quantum computation if bob is honest , since what alice and bob did is nothing but a usual measurement - based quantum computation .
( it is something like the following story : alice and bob are in the same laboratory .
the preparation and the maintenance of the resource state , which are boring routines , are done by a student bob , whereas the most exciting part of the measurement - based quantum computation , namely the measurements and the collection of data are done by his boss alice .
somehow , there is no communication between the boss and the student . )
it is also easy to understand that whichever states evil bob prepares instead of the correct resource state , and whichever states evil bob sends to alice , bob can not learn anything about alice s information , since alice does not send any signal to bob and therefore because of the no - signaling principle @xcite bob can not gain any information about alice by measuring his system @xcite : if alice could transmit some information to bob by measuring her system , it contradicts to the no - signaling principle @xcite .
( note that we assume there is no unwanted leakage of information from alice s laboratory .
for example , bob can not bug alice s laboratory .
it is the standard assumption in the quantum key distribution @xcite . ) in sec .
i of appendix , we give the mathematical proof of the security of protocol 1 based on the no - signaling principle .
protocol 1 has four advantages .
first , unlike ref .
@xcite , no random - number generator is required for alice . this is advantageous since it is not easy to generate completely random numbers , and the random - number generator might be provided by a company under the control of bob .
second , the security of the protocol is device independent in the sense that alice does not need to trust her measurement device in order to guarantee the security , since whatever alice s device does , bob can not gain any information about alice s computation ( due to the no - signaling principle ) as long as there is no unwanted leakage of information from alice s laboratory , which is the standard assumption in quantum key distribution @xcite .
third , the proof of the security is intuitive and very simple , and it is based on the no - signaling principle @xcite , which is more fundamental than quantum physics @xcite .
( even if quantum physics is violated in a future , protocol 1 survives as long as the no - signaling principle holds . ) finally , any model of measurement - based quantum computation ( such as the cluster model @xcite , the aklt models @xcite , and the topological model @xcite , etc . )
can be directly changed into a blind model : bob has only to let alice do measurements .
( on the other hand , in ref .
@xcite , many complicated procedures are required to make the aklt measurement - based quantum computation blind . )
since no modification is required to make a model blind , the advantage of a model is preserved when it is changed into a blind model .
for example , an advantage of doing the measurement - based quantum computation on the aklt states is that the quantum computation is protected by the energy gap of a physically natural hamiltonian @xcite .
if the aklt model is used in protocol 1 , bob who prepares and maintains the resource aklt state can enjoy that advantage , i.e. , bob s state is protected by the energy gap .
this is also the case for the models of refs .
@xcite : if these models are used in protocol 1 , bob can enjoy the advantage of these models , i.e. , bob does not need to keep his state in the ground state ; his state is allowed to be the equilibrium state at a finite temperature .
a disadvantage of protocol 1 is that the quantum channel between alice and bob must not be too lossy .
( throughout this paper , the channel loss " includes the detection inefficiency of alice s device , since the detection inefficiency behaves like the channel loss . ) on the other hand , in the previous protocols @xcite where alice sends randomly rotated particles to bob , the high loss rate of the quantum channel is not crucial , since if bob does not receive a particle due to the loss in the quantum channel , bob has only to ask alice to again generate and send another state with another random angle .
one way of overcoming that disadvantage of protocol 1 is to use a model which can cope with the particle loss .
for example , it was shown in ref .
@xcite that the topological measurement - based quantum computation @xcite can cope with the heralded particle loss if the loss probability is below the threshold .
if bob uses this model , alice and bob can perform protocol 1 without suffering from the particle loss as long as the loss rate of the quantum channel between alice and bob ( and that of alice s device ) is below the loss threshold calculated in ref .
@xcite .
_ protocol 2_. if we want to have a protocol which is tolerant against any high channel loss rate , we have to give up the perfect no - singling , since alice has to send some message to bob when a particle is lost .
one way of making protocol 1 tolerant against any high channel loss is to use the quantum teleportation .
let us consider the following protocol : ( 1 ) bob prepares a resource state .
( 2 ) he creates a bell pair , and sends a half of it to alice .
( 3 ) if the particle is lost , alice asks bob to send it again .
if alice receives the particle , alice lets bob know it .
( 4 ) bob teleports a particle of his resource state to alice by using the bell pair . ( 5 ) bob sends the measurement result of the teleportation to alice . ( 6 ) alice measures the teleported particle in an angle which is determined by her algorithm ( and bob s teleportation result ) .
this protocol is a modified version of protocol 1 where bob teleports a particle of his resource state instead of directly sending it to alice .
this protocol is loss tolerant , since if a half of a bell pair is lost in the channel , bob has only to send it again .
however , this protocol has a huge problem : alice has to have a single - particle quantum memory , since alice s measurement must be done after bob s teleportation ( otherwise alice can not correct byproducts created by bob s teleportation ) . such a quantum memory does not need to have a long coherence time since the quantum teleportation can be done quickly , but still the requirement of a quantum memory is disadvantageous to alice . here , we introduce protocol 2 , which can avoid such a quantum memory . this is our second main result of this paper .
the basic idea of protocol 2 is that alice prepares " rotated states which encode " algorithm in bob s place , and bob performs a layer - by - layer measurement - based quantum computation with these rotated states .
protocol 2 runs as follows : ( 1 ) bob creates a bell pair , and sends a half of it to alice through the quantum channel .
( 2 ) if alice does not receive it , because of the channel loss , alice asks bob to send it again and goes back to ( 1 ) .
( 3 ) if alice receives the particle , she measures it in the basis @xmath3 , where @xmath4 is a certain angle ( not a random angle ) determined by the algorithm which alice wants to run .
( @xmath5 for clifford gates , and @xmath6 for a non - clifford gate .
details will be explained in sec .
ii of appendix . ) in this way , alice can prepare " a state which encodes the angle of the algorithm in bob s place .
( 4 ) bob couples the half of the bell pair which he has to a qubit of his register state by using the @xmath7 gate , and measures the qubit in the register state in the @xmath8 basis .
this @xmath9-basis measurement implements the quantum gate .
( 5 ) bob sends the result of this @xmath9-basis measurement to alice through the classical channel .
( 6 ) they repeat ( 1)-(5 ) until the computation is finished . in sec .
iii of appendix , we show the blindness of protocol 2 : whichever states evil bob prepares and whichever states evil bob sends to alice , bob can not learn anything about alice s information .
one might think alice s measurement in step ( 3 ) has to be delayed until the end of step ( 5 ) so that she can feed - forward bob s measurement outcome as usual one - way model .
however , this is not the case . in protocol 2 , by properly choosing the measurement basis , we give a way to postpone alice s feed - forwarding until her subsequent measurements , and hence she does not have to wait for bob s measurement outcome .
this means that no quantum memory is required for alice . in sec .
ii of appendix , we give a detailed explanations about how alice should measure particles . in this way
, we can obtain a protocol which is loss tolerant .
however , as we have mentioned earlier , there is a trade - off between loss tolerance and no - signaling .
protocol 2 is no longer no - signaling . in protocol 1 ,
no signal is transmitted from alice to bob , and therefore the no - signaling is completely satisfied . however , in protocol 2 , the message whether alice receives a particle or not is sent from alice to bob , and therefore the no - signaling is no longer satisfied .
one might think that such a message is not directly related to alice s measurement angles , and therefore the situation is quasi " no - signaling .
however , if we want to show the device - independence security , a special care is necessary . in ref . @xcite it was shown that in quantum key distribution if alice and bob use the same measuring device many times , some secret information can be broadcasted through the legal " channel .
similar attack can be considered in our protocol 2 .
for example , let us assume that alice does the measurement in the angle @xmath10 .
then , the measuring device remembers the number 5 , and pretends to loss the particle fifth times .
then alice sends the message , the particle is lost " , to bob fifth times , and from that fact , bob can know the number 5 .
one way of avoiding such an attack is , as is explained in ref .
@xcite , to discard the measuring device after using it , and to use new device for every measurement .
the other way , which can be used in our protocol 2 , is that alice generates a random bit @xmath11 when bob sends a particle , and behaves as if the particle is lost ( arrived ) when @xmath12 ( 1 ) . in this case
, alice needs a random number generator , but the evil measuring device can no longer do that attack . _
discussion_. in this paper , we have proposed protocols of blind quantum computation for alice who does only measurements , such as the polarization measurement with a threshold detector . in quantum optics , for example ,
the state measurement is much easier than the single - qubit state generation .
therefore our scheme makes alice more classical than the previous protocols @xcite in certain experimental setups , such as optical systems . in the protocol of ref .
@xcite , bob is required to perform the non - demolition photon - number measurement , which is not easy with the current technology . in our protocols , on the other hand , bob is not required to have such an additional high technology .
we have proposed two protocols , protocol 1 and protocol 2 .
procotol 1 is simple , its security is based on the no - signaling principle , and it satisfies the device - independent security .
however , it can not tolerate a high channel loss rate . on the other hand ,
protocol 2 can tolerate any high channel loss rate , although it is more complicated than protocol 1 and no longer no - signaling .
appropriate one should be chosen depending on the situation .
finally , let us briefly discuss about the verification @xcite of blind quantum computing .
the verification is a way of alice checking whether bob is honestly following her protocol @xcite .
it is important for blind quantum computation , since evil bob can just destroy the computation and alice might accept wrong computational results .
methods of the verification for blind quantum computation of ref .
@xcite were already proposed in refs .
@xcite .
can we do the verification for our measuring alice blind quantum computation ?
one simple way of doing verification is that alice randomly chooses some subsystem of the resource state and measures the stabilizer operators in order to check whether bob correctly creates the resource graph state .
recently , more efficient way of doing verification for our measuring alice protocol has been proposed in ref .
@xcite , which uses the previous verification methods of aharonov , ben - or , and eban @xcite and fitzsimons and kashefi @xcite . by using that verification method
, alice can check whether bob is honestly doing computation or not .
tm was supported by jsps .
kf was supported by mext grant - in - aid for scientific research on innovative areas 20104003 .
we acknowledge matty j. hoban for bringing our attention to ref . @xcite . |
the nature and the time scale of the processes that shape galaxy properties in clusters and groups are still unclear .
the presence of a hot intercluster gas observed in x - rays should have a role in shaping some galaxy properties ( e.g. gunn & gott 1972 ) .
the window opened by the redshift dependence of the galaxy properties has been used to set constraints on the time scales of the processes ( e.g. butcher & oemler 1978 , 1984 ; dressler et al .
1997 ; stanford , eisenhardt & dickinson , 1998 ; treu et al .
2003 ) . however , the observational evidence of the environmental effect is still uncertain .
for example , the existence of a butcher oemler ( bo ) effect ( butcher & oemler 1984 ) , i.e. the fact that clusters at higher redshift have a larger fraction of blue galaxies , @xmath3 , is still controversial .
the controversy is raised by two criticisms concerning measurements and sample .
andreon , lobo & iovino ( 2004 ; hereafter ali04 ) analyse three clusters at @xmath4 without finding evidence of a high blue fraction with respect to @xmath5 .
they also show the drawbacks of the various definitions of @xmath3 adopted in the literature .
they conclude then that twenty years after the original intuition by butcher & oemler , we are still in the process of ascertaining the reality of the effect " . the same work put in a different perspective the results of rakos & shombert ( 1995 ) , clarifying the fact that even if all the galaxies in the universe are passively evolving , the blue fraction will be @xmath6 at @xmath7 in the rakos & shombert ( 1995 ) scale .
therefore , the very high fraction they found at high redshift does not require any special mechanism to account for the present day counterparts other than ageing .
ali04 introduce also a first discussion about the difficult task of measuring the error on @xmath3 , given the observations , showing that at least some previous works have underestimated errors and , by consequence , overstated the evidence for the bo effect .
the role of the inference , the logical step going from the observed data to the true value and its error , has been further elaborated in dagostini ( 2004 ) in the more general case of unknown individual membership for the galaxies .
kron ( 1994 ) claimed that all the high " redshift clusters known in the early 80 s ( @xmath8 ) were somewhat extreme in their properties , and this was precisely the reason why they were detected .
andreon & ettori ( 1999 ) quantify this issue , and show that many of the clusters compared at different redshifts have also different masses ( or x - ray luminosities ) , in such a way that we are comparing unripe apples with ripe oranges in understanding how fruit ripens " ( andreon & ettori 1999 ) . together with allison - smith et al .
( 1993 ) and andreon & ettori ( 1999 ) , ali04 show that the optical selection of clusters is prone to produce a biased - hence inadequate - sample for studies on evolution since at larger redshifts it naturally favours the inclusion in the sample of clusters with a significant blue fraction .
they show that clusters with a blue fraction as the observed ones are over - represented in optical cluster catalogs by a factor two , with respect to identical clusters but without a bursting population .
there is therefore a compelling need to study the properties of galaxies in clusters at intermediate redshift ( @xmath9 ) , avoiding the bias of an optical selection , by choosing clusters of the same mass as present day studied clusters to avoid an apple vs orange " issue .
this is the aim of this paper , where we present a bo style study of a small but representative sample of 7 clusters , x - ray selected , of low to average mass ( velocity dispersion ) and at intermediate redshift .
the layout of the paper is the following . in sect .
2 we present optical imaging and spectral data . in sect .
3 we show that the studied sample is both representative and x ray selected .
we revisit in section 4 the definition of the blue fraction , in order to account for the reduced age of the universe at higher redshift .
5 presents some technical details .
results are summarized in sect .
6 , whereas sect .
7 discusses relevant results published in the literature and some final conclusions .
appendixes present a bayesian estimate of cluster velocity dispersion , richness , and blue fraction .
we adopt @xmath10 , @xmath11 and @xmath12 km s@xmath13 mpc@xmath13 .
.the cluster sample [ cols= " < , > , < , > , > , > , > , > " , ] different physical mechanisms are thought to operate in different environments ( see treu et al .
2003 for a summary ) and thus , by identifying where the colour of galaxies starts to change , we can hope to identify the relative importance of such mechanisms .
for this reason , we studied the radial dependence of the blue fraction @xmath3 as usually done in the literature , by splitting the data in radial bins .
we arbitrarily choose @xmath14$],@xmath15 $ ] , @xmath16 $ ] and @xmath17 $ ] in units of @xmath18 , for simplicity .
in the outermost bin we were forced to drop xlssc 016 , because @xmath19 lies farther away than the mid - distance between xlssc 016 and the nearest cluster to it , as seen projected on the plane of the sky , and , therefore , this radial bin is potentially contaminated by galaxies belonging to the other cluster .
note that its inclusion , or exclusion , in the other radial bins does not affect the derived values , and therefore our conclusions .
table 3 lists the found values .
figure 8 ( solid points ) shows that the blue fraction increases with the clustercentric distance , from @xmath20 in the innermost bin , to @xmath21 and @xmath22 in the two outermost bins : galaxies at the center of clusters are found to have a suppressed star formation ( redder colours ) compared to those at larger clustercentric radii .
have we reached the field value of the blue fraction ? using the spectrophotometry listed in combo-17 ( wolf et al .
2004 ) , that encompasses @xmath23 deg@xmath24 of the chandra deep field region , we have selected the galaxies brighter than the ( same evolving ) absolute magnitude limit adopted in our work , and in the same redshift range ( @xmath25 ) .
there are 83 galaxies , of which 61 are bluer than an sa evolving template .
we , therefore , infer a blue fraction of @xmath26 , arbitrarily plotted at @xmath27 in fig . 8 . in the above calculation ,
we were forced , for lack of information , to neglect redshift errors and errors on the photometric corrections applied by the authors to compute absolute magnitudes .
the blue fraction is found to steadily increase from the cluster core to the field value .
the important point to note in fig 8 is that the influence of the cluster reaches large radii .
there are two possible explanations for the above result .
first , the mechanism affecting the galaxy colours reaches large radii .
in such a case , ram pressure stripping , tidal halo stripping and tidal triggering star formation ( just to mention a few , see e.g. treu et al .
2003 for definitions and other examples ) are ruled out as _ direct _ causes of the observed trend , because all of them are short - range mechanisms being effective in the cluster center only .
alternatively , the trend might be produced by a significant population of backsplash galaxies , i.e. satellite galaxies that once were inside the virial radius but now reside beyond it , as suggested on theoretical grounds by gill et al .
these authors claim that about the same number of infalling galaxies and backsplash galaxies should be at @xmath28 . under the reasonable assumption that infalling galaxies have a blue fraction equal to the field one , and rebounded galaxies have a blue fraction equal to the central one , the expected blue fraction @xmath3 at @xmath18 should be about 0.49 ( @xmath29 ) , in good agreement with the observed value , given support to the backsplash population alternative , in agreement with models and observations presented in balogh , navarro & morris ( 2000 ) . from a strict statistical point of view , this possibility is favoured because it provides a sharp prediction verified by the observations .
the kinematical predictions of gill et al .
( 2005 ) are also in qualitative agreement with observations of the coma cluster : blue spirals ( identified as the infalling population ) have a higher velocity , relative to the cluster center , than red spirals ( identified as rebounded objects ) and early type populations ( andreon 1996 ) .
the possible existence of a backsplash population , whose importance seems hard to quantify from theoretical grounds , requires to keep in stand - by our conclusion , as well other conclusions based on the ( often implicit ) hypothesis that the population observed at large radii is uncontaminated by rebounded galaxies ( e.g. mcintosh , rix , & caldwell , 2004 ) . for the very same reason
, one should keep in hold the interpretation of the morphology density ( or whatever density dependent trends in population properties , such as the strong emitter fraction ) , because it could either be the result of mechanisms operating at the studied density , but also the result of different degrees of contamination ( at different distances from the cluster center ) by the backsplash population .
we are not questioning the existence of the segregation , but the way one may interpret it .
comparison of our results with other ones requires to pay attention to the prescriptions adopted to define the blue fraction , to the way the cluster sample is built , and , sometime , to the adopted statistical approach .
the most similar work to ours is the seminal butcher & oemler paper , from which we modelled our prescriptions . by selecting a small subsample of clusters of richness similar to ours but located in the nearby universe ( @xmath30 , where bo and our prescriptions are identical ) they find @xmath3 values within @xmath31 , the radius that includes 30 per cent of cluster galaxies , in the @xmath32 to @xmath33 range , with a typical error of @xmath34 .
this range of values is not significantly lower , considering the various sources of uncertainties , than our central value @xmath35 , to claim that the two values are different at a high significance level , especially taking into account the fact that the butcher & oemler errors are sometimes optimistically estimated ( andreon , lobo & iovino 2004 ) .
de propris et al.s ( 2004 ) large sample of nearby clusters matches our sample in terms of richness : we find for their sample ] an average @xmath36 of 30 galaxies and a blue fraction inside @xmath37 of @xmath38 , taking into account that the blue fraction is a binomial deviate ( the authors assumed it to be a gaussian and find @xmath39 ) .
the error due to the sample size is negligible ( 0.01 ) because their sample is large .
however , the largest source of uncertainty in their work comes from their large photometric errors ( they use photographic plates ) .
such photometric errors induce a bias in the blue fraction that , as discussed in sect 3.1 , is difficult to correct for ( jeffreys 1938 , eddington 1940 ) , and is neglected by the authors .
after accounting for minor differences in the luminosity cuts between de propris et al . and
bo and for the mentioned malmquist bias , the estimated blue fraction within @xmath37 in the de propris et al.s ( 2004 ) sample becomes @xmath40 , but with an error hard to quantify .
inside @xmath37 we find @xmath41 , which identical to what found in the de propris et al.s large nearby sample . to summarize , our @xmath1 sample matches in terms of richness the nearby samples in butcher & oemler ( 1984 ) and , especially , de propris et al .
( 2004 ) and shows equal blue fractions within @xmath31 and @xmath37 , i.e. no butcher - oemler effect is seen .
it is to be noted that our sample has an almost identical size and redshift distribution as the high redshift clusters in the bo sample , and thus that our lack of detection of a bo effect is not due to a smaller or closer sample .
the compared clusters matches in terms of richness , but are constructed using different selection criteria , because the low redshift sample is an optically selected one , while our cluster sample is x - ray selected . as mention in sec .
1 , our x - ray selection is chosen to minimize the observational bias on @xmath3 , and hence to derive a fair measure of the blue fraction . at low redshift , we are not aware of any reason why @xmath3 should be biased _ at a fixed richness _ for an optically selected sample , such as the ones of butcher & oemler ( 1984 ) and de propris et al .
( 2004 ) : why clusters of a given richness and rich in blue galaxies should be over / underrapresented in cluster catalogs of the nearby universe ? therefore , even if the selection criteria used to build the compared cluster samples are different , the comparison of the blue fractions is safe , because both cluster samples provide unbiased values of @xmath3 .
there are hints that confirm the constancy of the blue fraction at even larger redshifts .
ali04 show evidence for a _ low _ blue fraction at @xmath4 .
recently , tran et al .
( 2005 ) also find a _ low _ blue fraction ( @xmath39 ) for a cluster at @xmath42 , computed inside a cluster portion that , if not rigorously identical to the one prescribed by butcher oemler , does support the _ non _ existence of a butcher oemler effect . both works adopted a non - evolving @xmath43 . if an evolving @xmath43 is used , the derived blue fraction at high redshift would even be lower than claimed , giving further support to our conclusion .
ali04 have also disproved all the reported literature evidence accumulated thus far for the existence of a butcher oemler effect , i.e for a change of the blue fraction inside @xmath31 .
all the above suggests that the fraction of blue galaxies , computed by separating the galaxies using a population formed by stars whose age increases at the same rate as the universe age increases , does not evolve . or
, if the reader prefers , there is no systematic drift from the blue to the red classes of galaxies as the look
back time evolves , between @xmath5 and @xmath44 . a result similar to the one depicted in fig .
8 is presented in lewis et al .
( 2002 ) based on nearby clusters .
our results are in qualitative agreement with theirs since we also find that the cluster affects the fraction of active galaxies up to the virial radius .
lewis et al . (
2002 ) have studied a _ nearby _
cluster sample composed of 440 member galaxies inside the virial radius ( vs our sample of 320 ) .
figure 8 shows that their fraction of galaxies with star formation rates , normalized to @xmath45 , larger than 1 solar mass per year ( triangles ) nicely compares with our derived fraction of blue galaxies .
our error bars for their points show the expected central 68 per cent credible intervals , only accounting for sampling errors , computed by us from a straightforward application of statistics . in lewis
et al . ( 2002 ) , the sample is split in classes very similar to ours and bo : in fact , our spectro - photometric sa has a star formation rate , normalized to @xmath45 , equal to their adopted threshold ( one solar mass per year per @xmath45 galaxy ) if @xmath46 , a value well inside the range of values usually observed ( e.g. blanton et al .
2001 ; norberg et al .
i.e. what is called blue by them is also called blue by us , on average .
it is not surprising , therefore that integrating the lewis et al.s ( 2002 ) blue profile within @xmath37 gives a blue fraction identical to the one observed in the propris et al.s ( 2004 ) sample ( @xmath47 vs @xmath48 ) , further supporting the similarity of the two classes ( blue by colour and blue by star formation rate ) .
their cluster sample has an overlapping , but different , range of masses ( velocity dispersion ) with respect to our sample : our richest clusters have a velocity dispersion typical of the average values of lewis et al .
( 2002 ) clusters . however , their profile is only marginally affected , if at all , by separating clusters in ( two ) velocity dispersion classes ( lewis et al .
furthermore , gomez et al . ( 2003 ) indirectly confirm that the radial profile is not too much affected by differences in cluster mass , by studying a sample of nearby clusters having velocity dispersions similar to our sample .
therefore , differences in the way cluster sample are built seems not to affect the derived blue " profile .
the agreement between lewis et al .
( 2002 ) and our blue fraction profiles is almost perfect ; however the however the studied clusters are located at quite different look back times : clusters in lewis et al .
are in the very nearby universe ( at @xmath49 ) , whereas our clusters have @xmath1 , implying a @xmath50 gyr time difference for the adopted cosmology .
as long as the separation of galaxies in classes by lewis et al .
( 2003 ) and in our work is similar , the agreement of the two radial profiles means that there is no evolution of the blue fraction between @xmath5 and @xmath1 , from the cluster center to the field value . considering a much more luminous x - ray ( and therefore massive ) sample of clusters at intermediate redshift , fairley et al .
( 2002 ) find an increasing blue fraction as a function of the clustercentric distance , up to 2 @xmath31 , in good agreement with the results we found over much larger clustercentric distances .
a quantitative comparison between the two pieces of work is however impossible , because there are too many uncontrolled variables that are allowed to change between these .
the authors find a blue fraction @xmath51 , in agreement with the value we observe in the cluster center @xmath35 .
however , we believe that this apparent agreement largely arises by chance . first , the authors used a non - evolving template to separate the galaxies in red and blue classes , and find two sets of ( different ) values for their two sets of available colours .
secondly , they considered higher redshift than we do , by observing clusters with comparable exposure times but with smaller telescopes ( 2.5 m vs 4.0 m ) . in spite of their expected larger errors ,
they neglect the effect of photometric errors on their blue fraction estimates ( sect 3.2 ) .
third , they do not adopt an evolving luminosity limit . and
, finally , a comparison of the values derived in the two works requires an extrapolation , because clusters with very different masses ( x ray luminosities ) are considered .
ellingson et al . ( 2001 ) performed a study quite different from ours , and adopt a galaxy separation that is the same irrespective of redshift ( i.e. of galaxy age ) , because they decomposed their spectra on non - evolving spectral templates .
their claim for a change in the population gradient is just a restatement of the fact that the blue fraction is higher everywhere in the cluster and in the field because galaxies are bluer when they were younger , i.e. is not informative about processes running in clusters or in the field , but just informative about aging .
these authors would observe an evolution of the gradient even if galaxies would be kept isolated from the surrounding environment and the infall in the cluster would be fixed ( i.e. no new galaxy falls in the cluster , and galaxies are kept fixed at their observed position ) : their old population " fraction increases going from high to low redshift because galaxies become older , and the effect is more marked at large clustercentric radii than in the centre , because in the cluster core the old population " fraction is already near to 1 and can not take values larger than 1 .
we , instead , choose to reduce by one the number of parameters , removing the age dependency by using an evolving ( sa ) template .
in summary , the referenced analysis do not reveal results in disagreement with our own work , although their interpretation may sometimes be different ( or even opposite to ours ) .
this paper revises the definition used to separate galaxies in two colour classes in a way that takes in to account the reduced age of the universe at higher redshift .
it is nowadays uninteresting to know whether the fraction of blue galaxies changes with redshift in a way that is different from the expectation of a model that we now know is unphysical ( that has the same age at all redshifts ) . if the model is unphysical
, there is no need to make observations to rule it out .
a stellar population whose age does not change in a universe whose age instead changes , as it is supposed by using a non - evolving spectral template ( or a fixed @xmath43 , i.e. the bo prescription ) , is clearly non - physical .
it was useful a long time ago to show that a universe with the same age at all redshifts is rejected by observations .
however , nowadays we can attack a more essential question : to know whether galaxies evolve differently from a reference evolution that is physically acceptable .
our measurements of evolution are , therefore , zero - pointed on the evolution of an object whose age increases as required by the current cosmological model .
we select a spectro - photometric sa to conform to the bo prescription in the local universe .
effectively , this is a change in perspective : we should no longer attempt to reject an unphysical universe , in which the age of the universe does depend on redshift , whereas the age of its content does not , but we should study whether the observed differences between the low and high redshift content are in agreement with differences of the universe age at the considered redshifts . furthermore , we have introduced in our specific domain the tools of bayesian inference ( see appendix ) , dramatically improving on previous approaches that led some authors to claim that they have observed unphysical values ( such as blue fractions outside the @xmath52 $ ] range or negative star formation rates ) .
such tools allow us to use all our data without rejecting blue fractions measured at large clustercentric radii , where the signal to noise is low , contrary to previous researchers obliged to discard such data ( or claiming that they have observed unphysical values ) .
the main result of this work is that _ we find that the cluster affects the properties of the galaxies up to two virial radii at @xmath1_. we have measured the blue fraction of a _ representative _ sample of clusters at intermediate redshift .
indeed , our sample is a random sampling of a volume complete x - ray selected cluster sample .
the x - ray selection has no cause - effect relationship on the cluster blue fraction , all the remaining parameters being kept fixed , to the best of our knowledge , and , therefore , the studied sample consists of an unbiased one ( from the blue fraction point of view ) .
our statement should not be over - interpreted , however , because we are only sampling a portion of the x - ray parameter space : very x - ray luminous clusters are missing in our sample because they are intrinsically rare , and clusters with fainter x - ray emission than the limiting flux are missing because they lie outside the sampled space .
studied clusters show a variety of values for the blue fraction , when the fraction is measured within @xmath18 .
the variety is too large to solely be accounted for by errors . at smaller radii , instead , the blue fractions are more homogeneous .
actually , there is some evidence that the blue fraction within @xmath18 increases with the cluster velocity dispersion , i.e. with the cluster mass , whereas the increase at smaller radii is much smaller , if present at all .
therefore , intermediate redshift clusters with the largest masses show the largest fractions of star - forming galaxies , when measured within @xmath18 . however , the evidence is good but not definitely conclusive and still requires an independent confirmation .
the radial dependence of the blue fraction is quite shallow : it smoothly and monotonically increases from the centre to the field .
the latter has been determined according to our prescriptions using combo-17 data .
the radial dependence ( i.e. the blue fraction at every computed clustercentric radius ) is equal to the one recently found in a comparable sample of clusters , but in a 3 gyr older universe , i.e. at @xmath5 ( lewis et al .
the agreement between the two derived profiles ( amplitude and shape ) , our blue fraction within @xmath31 and @xmath37 and the local similar determinations ( butcher & oemler 1984 , de propris et al .
2004 ) , the low blue fractions at high redshift ( ali04 , tran et al .
2005 ) , all of these suggest that there is no colour evolution beyond the one needed to account for the different age of the universe and of its content .
the above is found to hold from the cluster core to the field value .
previous controversial evidence from the literature assumed that the universe becomes older while its content does not , and overstated the significance of the evidence or compared heterogeneously measured blue fractions ( as shown in ali04 ) .
the interpretation of the observed radial trend is complicated by the possible existence of a backsplash population
. if the backsplash population represents a negligible fraction of galaxies at a given clustercentric radius , then the large clustercentric distance at which the cluster still produces some effect rules out short range scale mechanisms .
however , the predicted backsplash population is precisely what is needed to explain our observed fraction at @xmath18 , given the fraction at the cluster center and in the field , and also qualitatively accounts for different kinematics of galaxies having different star formation rates ( blue and red spirals ) in the coma cluster .
if this is the case , mechanisms efficient in the cluster center only come into play , because galaxies are affected when they reach the cluster core , and are then scattered at large clustercentric radii where they spend a lot of time and are observed .
the possible existence of the backsplash population does not offer us the possibility to draw a final inference about the nature and the time scale of the processes that shape galaxy properties in clusters .
the backsplash mechanism is a physical one : it affects the interpretation of measured radial ( or density ) trends drawn by us and other authors , and forces us to keep in hold their interpretations .
however , the infall pattern turns out not to have changed during the last 3 gyr , as measured by the identical blue fraction profiles at @xmath5 and @xmath1 , in spite of apparently contradictory previous claims , based on the use of a fraction definition that has one more ( uncontrolled ) dependence .
we warmly thank giulio dagostini , for his advices on the inference theory , allowing us to perform the @xmath3 measurements presented in this paper , as well as andrew jaffe , andrew liddle and mike west for useful discussions about bayesian statistics and model selection .
jon willis is thanked for his efforts in acquiring part of the ctio data used in this paper and for useful comments .
malcom bremer , pierre - alain duc and ivan valtchanov are thanked for their comments on an early version of this paper , and gianluca lentini for useful english hints .
gg acknowledges fondecyt grant # 1040359 and hq acknowledges partial support from the fondap centro de astrofisica ( conicyt ) .
part of this work was also performed in the framework of the iuap p5/36 project , supported by the dwtc / sstc belgian federal service .
this paper is based on observations obtained with xmm , with eso ( prop .
70.a-0733 ) , cerro - tololo interamerican observatory ( prog . 2000 - 0295 , 2001 - 0316 and 2002 - 0247 ) and magellan telescopes at las campanas observatory .
allington - smith , j. r. , ellis , r. , zirbel , e. l. , oemler , a. 1993 , apj 404 , 521 andreon , s. 1996 , a&a , 314 , 763 andreon , s. & ettori , s. 1999 , apj , 516 , 647 andreon , s. , lobo , c. , & iovino , a. 2004 , mnras , 349 , 889 ( ali04 ) andreon , s. , willis , j. , quintana , h. , valtchanov , i. , pierre , m. , & pacaud , f. 2004a , mnras , 353 , 353 andreon , s. , willis , j. , quintana , h. , valtchanov , i. , pierre , m. , & pacaud , f. 2004b , in `` exploring the universe .
contents and structure of the universe '' , xxxix rencontres de moriond ( astro - ph/0405574 ) andreon , s. , punzi , g. , grado , a. , 2005 , mnras , 360 , 727 balogh , m. , et al . 2004 , mnras , 348 , 1355 beers , t. c. , flynn , k. , & gebhardt , k. 1990 , aj , 100 , 32 bertin , e. , & arnouts , s. 1996 , a&as , 117 , 393 blanton , m. r. , et al .
2001 , aj , 121 , 2358 bruzual a. , g. , & charlot , s. 1993 , apj , 405 , 538 butcher , h. , & oemler , a. 1978 , apj , 226 , 559 butcher , h. & oemler , a. 1984 , apj , 285 , 426 .
( bo ) chiappetti , l. et al .
2005 , a&a , 439 , 413 coleman , g. d. , wu , c .- c . , & weedman , d. w. 1980 , apjs , 43 , 393 .
dagostini g. , 2003 , bayesian reasoning in data analysis . a critical introduction , world scientific publishing , 2003 dagostini g. , 2004 , physics/0412069 de propris , r. , stanford , s. a. , eisenhardt , p. r. , & dickinson , m. 2003 , apj , 598 , 20 de propris , r. et al .
2004 , mnras , 351 , 125 ( dp04 ) dickey , j. m. , & lockman , f. j. 1990 , araa , 28 , 215 dressler , a. et al .
1997 , apj , 490 , 577 eddington , a. s. 1940 , mnras , 100 , 354 ellingson , e. , lin , h. , yee , h. k. c. , & carlberg , r. g. 2001 , apj , 547 , 609 fairley , b. w. , jones , l. r. , wake , d. a. , collins , c. a. , burke , d. j. , nichol , r. c. , & romer , a. k. 2002 , mnras , 330 , 755 garilli , b. , maccagni , d. , & andreon , s. 1999 , a&a , 342 , 408 gill , s. p. d. , knebe , a. , & gibson , b. k. 2005 , mnras , 356 , 1327 g ' omez , p. l. , et al .
2003 , apj , 584 , 210 goto , t. , et al .
2003 , pasj , 55 , 739 goto , t. , 2005 , mnras , 356 , l6 gunn , j. e. , & gott , j. r. i. 1972 , apj , 176 , 1 helene , o. , 1983 , nucl . instr .
meth . , 212 , 319 jeffreys , h. 1938 , mnras , 98 , 190 jeffreys , h. 1961 , theory of probability , 3rd edn , oxford univ .
press , oxford kodama , t. , arimoto , n. , barger , a. j. , & aragon - salamanca , a. 1998 , a&a , 33 4 , 99 kraft , r. p. , burrows , d. n. , & nousek , j. a. 1991 , apj , 374 , 344 kron r , 1994 , the deep universe , saas
fee advanced course , eds . b. binggeli & r. buser , berlin , springer , p. 233
landolt , a. u. 1992 , aj , 104 , 340 .
lewis , i. , et al .
2002 , mnras , 334 , 673 liddle , a. r. 2004 , mnras , 351 , l49 loredo , t. , 1992 , in statistical challenges in modern astronomy , eds .
feigelson and g.j . babu ( eds . ) , springer - verlag , new york , 275 mauduit , j.c . ,
mamon , g.a . & hill , g.j . , 2005 ( in preparation ) mcintosh , d. h. , rix , h. , & caldwell , n. 2004 , apj , 610 , 161 norberg , p. , et al .
2002 , mnras , 336 , 907 oemler , a. , jr . 1974 ,
apj , 194 , 1 pierre , m. , et al .
2004 , journal of cosmology and astro - particle physics , 9 , 11 postman , m. , et al . , 2005 , apj , 623 , 721 press , w. , 1997 , in unsolved problems in astrophysics , eds .
j. bahcall & j. ostriker ( princeton : princeton university ) , 49 prosper h. , 1998 , phys .
rev , d , 37 , 1153 prosper h. , 2000 , in proceedings of the workshop on confidence limits , cern report ( hep - ph/0006356 also http://ph-dep.web.cern.ch/ph-dep/events/clw/papers.html ) rakos , k. d. , & schombert , j. m. 1995 , apj , 439 , 47 smith , j. a. , et al .
2002 , aj , 123 , 2121 stanford , s. a. , eisenhardt , p. r. , & dickinson , m. 1998 , apj , 492 , 461 schwarz g. , 1978 , annals of statistics , 5 , 461 tran , k. h. , van dokkum , p. , illingworth , g. d. , kelson , d. , gonzalez , a. , & franx , m. 2005 , apj , 619 , 134 treu , t. , ellis , r. s. , kneib , j. , dressler , a. , smail , i. , czoske , o. , oemler , a. , & natarajan , p. 2003 , apj , 591 , 53 wake , d. a. , collins , c. a. , nichol , r. c. , jones , l. r. , & burke , d. j. 2005 , apj , 627 , 126 weiner , b. j. , et al .
2005 , apj , 620 , 595 willis , j. , et al .
, 2005 , mnras , submitted wolf , c. , et al . 2004 , a&a , 421 , 913 zwicky , f. 1957 , morphological astronomy , berlin : springer
velocity dispersions and their uncertainties are computed according to statistical inference textbooks in a bayesian framework , from the observed values of the galaxy redshifts , while accounting for measurement errors .
we first numerically derive , using a monte carlo simulation , the likelihood of observing @xmath53 computed by using the scale parameter introduced by beers et al .
( 1990 ) , given the observed redshifts and redshift errors . then , using the bayes theorem and adopting a uniform prior we derive the probability that the cluster has a velocity dispersion @xmath53 , given the observed values of redshifts .
the posterior , for the chosen prior , turns out to be very well described by a gaussian .
velocity dispersions ( point estimates ) and uncertainties ( 68 per cent central credible intervals ) are quoted in table 1 and are robust to changes of priors : adopting a widely different prior ( @xmath54 ) , our point estimate of the cluster velocity dispersion changes by 2 to 5 per cent of its uncertainty .
derived velocity dispersions are corrected for the @xmath55 effect .
our velocity dispersions have the properties to be non negative , and their uncertainties do not include unphysical ( negative or complex ) values of the velocity dispersion . while the above properties seem useless to state , it should be noted that they are non trivial properties , because some unphysical velocity dispersions are still published . it should be noted that , for the velocity dispersions presented here , the frequentist and bayesian derivation of the value of the velocity dispersion turn out to be quite similar .
the cluster richness within @xmath18 is _ not _ naively derived using the common background subtraction ( e.g. zwicky 1957 , oemler 1974 ) : @xmath56 with obvious meanings for the symbols , using the _ observed _ number of galaxies , because it potentially leads to negative numbers of cluster galaxies , which is acceptable for the estimator described above , but not for the true value of the physical quantity aimed to be measured ( the cluster richness ) . furthermore , frequentist confidence intervals may have whatever size , including being empty or of vanishing length ( as actually occurs precisely for the above expression when the right - hand side of eq .
b1 is negative , e.g. kraft et al .
measurements derived from eq .
b1 , and their confidence intervals , have not the properties we would like richness and errors to have ( for example , richness to be positive , and errors to be be large when the uncertainty is large , and to become small when the uncertainty on nuisance parameters decreases , etc . ) .
these are well known and discussed with several degrees of approximation in both the frequentist and bayesian frameworks ( helene 1983 ; kraft et al .
1991 ; loredo 1992 ; prosper 1998 ; dagostini 2003 ) .
results derived from eq b1 , when @xmath57 are difficult to be understood and used ( say in computing averages , or when we need to propagate the uncertainty from @xmath58 on a derived quantity ) .
this situation occurs for one of our clusters ( xlscc 012 , if eq .
b1 is used , but we do not use it ) : its richness is @xmath59 and its confidence interval ( at whatever confidence level ) is empty ( kraft et al .
1991 ) .
`` if the results are to be supposed to have any relevance beyond the original data '' ( jeffreys 1938 ) , we believe that it is preferable to quote the point estimate of the cluster richness , given the data in hand , in place of the algebraic result of eq b1 .
we compute the ( posterior ) probability that the cluster has @xmath60 galaxies , and , when needed , we summarize it quoting , as we do for the velocity dispersion ( appendix a ) and the blue fraction ( appendix c ) , the median and the 68 per cent central interval .
specifically , we assume a uniform prior , taking advantage of the fact that the problem is mathematically worked out by kraft et al .
we checked that an almost identical result is obtained using a jeffrey prior ( the problem is worked out by prosper 1998 ) , once differences in the type of credible intervals are accounted for , i.e. that the result found is only marginally affected , if at all , by the prior choice .
it is comforting to find that xlssc 012 , which has at least 12 spectroscopic confirmed members ( table 1 ) , has some hot emitting gas , and hence _ does _ exist and has , as all clusters of galaxies , a positive number of galaxies , has a listed richness of 8 galaxies within @xmath18 ( brighter than an evolved @xmath61 mag ) , even if the ( naive , but widespread ) application of eq .
b1 attributes to it a negative number of galaxies ( @xmath59 ) and an empty confidence interval .
. b1 is routinely used in computing cluster luminosity functions in presence of a background , starting with zwicky ( 1957 ) , oemler ( 1974 ) .
andreon , punzi & grado ( 2005 ) update their use .
we conclude this section by reminding that , both in the frequentist and bayesian paradigms , the background subtraction ( marginalization ) does not require that the background in the cluster line of sight is equal to the _ average _ value or equal to the one observed in the control field , but only that it is drawn from the same parental distribution , contrary to some astronomical misconceptions .
many blue fractions published in astronomical papers can be dramatically improved : although , by definition , the blue fraction is hardly bounded in the [ 0,1 ] range ( otherwise part of a sample is larger than the whole sample ) , it is often claimed that the observed value of the blue fraction is outside the [ 0,1 ] range ( data points outside this range are present in several bo - like papers ) . in presence of a background ,
unphysical values frequently occur .
the reason is that the blue fraction is computed from : with obvious meanings for the symbols .
the unavoidable use of the _ observed _ number of galaxies , instead of the ( unknown ) true ones , in the above formula makes the result difficult to be understood , for the very same reasons already discussed for the richness estimator ( eq .
the use of the _ observed _ number of galaxies in eq .
c1 allows to find negative values for the blue fraction ( i.e. we would claim that there are more red galaxies than galaxies of all colours ) or blue fractions larger than one ( i.e. we would claim that there are more blue galaxies than galaxies of all colours ) , statements that are hard to defend .
this mainly occurs when poissonian fluctuations make background counts larger than counts in the cluster line of sight , or when there are more blue galaxies in the background than in the cluster line of sight .
eq c1 , adopted in postman et al .
( 2005 ) , forced these authors to discard two of three of their @xmath63 clusters in the determination of the spiral fraction .
bayesian inference allows not to discard data , to derive estimates that never take unphysical values , and ( credible ) intervals that have the properties we would like errors to have . as for the cluster velocity dispersion ( appendix a ) and richness ( appendix b )
, we compute the ( posterior ) probability that the cluster has a blue fraction @xmath3 , given the observed number of galaxies in the cluster direction ( total & blue ) and the expected number of background galaxies ( total & blue ) measured over a large control field .
all the mathematical aspects of the above computation have been worked out by dagostini ( 2004 ) , who provides all requested details and the exact analytic expression for the likelihood , to be used to derive the posterior , given our data .
such a posterior may be summarized by a few numbers : the median ( our point estimate of the cluster blue fraction ) and the 68 per cent central credible interval ( our estimate of the uncertainty ) , exactly as we did for the case of velocity dispersion and richness .
a trivial application of the bayes theorem allows to account for errors on @xmath18 ( due to the uncertainty on @xmath53 ) .
we verified that , properly accounting for @xmath18 errors , our results got unchanged , mainly because the blue fraction is a smooth and slowly varying function of @xmath18 .
it is interesting to note that a reasonable constraint on @xmath3 is achieved even for xlssc 012 , in spite of the naive expectation that , the cluster being poor ( eq . b1 would get @xmath59 galaxies ) , and the cluster richness appearing at the denominator of eq c1 , the error on the fraction is huge , and therefore the corresponding determined blue fraction is of low quality .
the correct inference takes , instead , a different approach and quantifies what is qualitatively apparent in figure 3 : to the left of the blue ( left ) arrow there is no evidence for an excess of blue galaxies in the cluster line of sight .
under such a condition , how is it possible that the cluster blue fraction gets large if almost no blue galaxy overdensity is observed ? given that almost no cluster blue galaxies are there , the cluster blue fraction is low , and the number of red galaxies sets how rich the cluster is and therefore how low " the fraction is . | the present paper focuses on the much debated butcher - oemler effect : the increase with redshift of the fraction of blue galaxies in clusters . considering a representative cluster sample made of seven group / clusters at @xmath0
, we have measured the blue fraction from the cluster core to the cluster outskirts and the field mainly using wide field ctio images .
this sample represents a random selection of a volume complete x - ray selected cluster sample , selected so that there is no physical connection with the studied quantity ( blue fraction ) , to minimize observational biases . in order to statistically assess the significance of the butcher oemler effect
, we introduce the tools of bayesian inference .
furthermore , we modified the blue fraction definition in order to take into account the reduced age of the universe at higher redshifts , because we should no longer attempt to reject an unphysical universe in which the age of the universe does depend on redshift , whereas the age of its content does not .
we measured the blue fraction from the cluster center to the field and we find that the cluster affects the properties of the galaxies up to two virial radii at @xmath1 .
data suggest that during the last 3 gyrs no evolution of the blue fraction , from the cluster core to the field value , is seen beyond the one needed to account for the varying age with redshift of the universe and of its content .
the agreement of the radial profiles of the blue fraction at @xmath2 and @xmath1 implies that the pattern infall did not change over the last 3 gyr , or , at least , its variation has no observational effect on the studied quantity .
[ firstpage ] galaxies : evolution galaxies : clusters : general galaxies : clusters : |
is a borel measure @xmath1 on a metric space @xmath0 fully determined by its values on balls ? in the context of general measure theory , such a question appears to be of extremely basic nature .
the answer ( when it is known ) strongly depends upon the interplay between the measure and the metric space . a clear overview on the subject is given in @xcite .
let us mention some known facts about this issue .
when @xmath5 equipped with a norm , the answer to the above question is in the affirmative .
the reason is the following : if two radon measures @xmath1 and @xmath6 coincide on every ball @xmath7 , then in particular they are mutually absolutely continuous , therefore by the radon - nikodym - lebesgue differentiation theorem one has @xmath8 for any borel set @xmath9 , where @xmath10 is the radon - nikodym derivative of @xmath1 with respect to @xmath6 ( defined for @xmath6-almost all @xmath11 ) . more generally , the same fact can be shown for any pair of borel measures on a finite - dimensional banach space @xmath3 . unfortunately , the differentiation theorem is valid on a banach space @xmath3 if and only if @xmath3 is finite - dimensional . of course , this does not prevent in general the possibility that borel measures are uniquely determined by their values on balls .
indeed , preiss and ti@xmath12 proved in @xcite that in separable banach spaces , two finite borel measures coinciding on all balls also coincide on all borel sets .
nevertheless , if this coincidence turns out to be satisfied only on balls of radius , say , less than @xmath13 , then the question still stands . in the case of
separable metric spaces , federer introduced in @xcite a geometrical condition on the distance ( see definition [ def : federer ] ) implying a besicovitch - type covering lemma that can be used to show the property above , i.e. , that any finite borel measure is uniquely identified by its values on closed balls .
when this condition on the distance is dropped , some examples of metric spaces and of pairs of distinct borel measures coinciding on balls of upper - bounded diameter are known ( see @xcite ) .
here we consider the case of a separable metric space @xmath0 where besicovitch covering lemma ( or at least some generalized version of it ) holds , and we ask the following questions : [ question_1]_how can we reconstruct a borel measure from its values on balls , and especially , what about the case of signed measures ? _ a classical approach to construct a measure from a given _ premeasure _ @xmath14 defined on a family @xmath15 of subsets of @xmath3 ( here the premeasure @xmath14 is defined on closed balls ) is to apply carathodory constructions ( method i and method ii , see @xcite ) to obtain an outer measure .
we recall that a premeasure @xmath14 is a nonnegative function , defined on a given family @xmath15 of subsets of @xmath3 , such that @xmath16 and @xmath17 . by method
i , an outer measure @xmath18 is defined starting from @xmath14 as @xmath19 for any @xmath20 .
but , as it is explained in @xcite ( section @xmath21 ) , method i does not take into account that @xmath3 is a metric space , thus the resulting outer measure can be incompatible with the metric on @xmath3 , in the sense that open sets are not necessarily @xmath18-measurable . on the other hand , method ii is used to define hausdorff measures ( see theorem [ thm_caratheodory_construction ] ) and it always produces a metric outer measure @xmath22 , for which borel sets are @xmath18-measurable .
as for a signed measure @xmath23 , the main problem is that , given a closed ball @xmath24 , it is impossible to directly reconstruct @xmath25 and @xmath26 from @xmath27 .
the idea is , then , to apply carathodory s construction to the premeasure @xmath28 ( here @xmath29 denotes the positive part of @xmath30 ) and check that the resulting outer measure is actually @xmath31 .
then , by a similar argument we recover @xmath32 .
now we consider the problem of reconstructing a measure @xmath1 from approximate values on balls .
we thus assume that a premeasure @xmath2 , defined on closed balls , is given and satisfies the following two properties : for some @xmath33 , @xmath34 , and @xmath35 , @xmath36 holds for all @xmath37 and all @xmath38 .
[ question_2 ] _ given a positive borel measure @xmath1 and a premeasure @xmath2 defined on balls and satisfying the two conditions in , is it possible to reconstruct an approximation up to constants of @xmath1 from @xmath2 ?
what about the case when @xmath1 is a signed measure ?
_ first notice that under the assumptions the best one can obtain is a reconstruction @xmath39 of @xmath1 such that @xmath40 for some constant @xmath41 .
we stress that , in the case @xmath42 in ( i ) , this can be easily obtained via carathodory method ii ( with @xmath43 ) while in the case @xmath44 carathodory construction does not provide in general such a measure @xmath39 . loosely speaking
, a loss of mass can happen in the recovery process , as the example presented in section [ section:2.1 ] shows . in order to recover @xmath1 , or at least some measure equivalent or comparable to @xmath1 ,
the choice of the centers of the balls in the collection , which are used to cover the support of @xmath1 , is crucial .
indeed they must be placed in some nearly - optimal positions , such that even the concentric balls with smaller radius have a significant overlapping with the support of @xmath1 . this led us to consider a packing - type construction .
the packing construction is mainly used to define the packing @xmath45-dimensional measure and its associated notion of packing dimension .
it is in some sense dual to the construction leading to hausdorff measure and dimension , and was introduced by c. tricot in @xcite .
then tricot and taylor in @xcite extended it to the case of a general premeasure . in our setting
we show that this packing construction can be formulated in a simpler and more manageable way ( see section [ sect : packtype ] ) . since the lower bound on @xmath46
is given in terms of @xmath47 , in the case @xmath44 and under some additional assumptions on the metric space @xmath0 we prove a suitable version of besicovitch covering lemma ( see proposition [ prop_diffusion_in_x ] and corollary [ bescivoic_with_doubling_balls ] ) , which represents a key ingredient in our construction and seems also of independent interest .
some further explanations about the assumption on @xmath46 are in order .
an example of @xmath46 satisfying is @xmath48 more generally one could consider @xmath49 where @xmath50 is a non - increasing probability density function .
notice that in @xmath51 this last expression corresponds to the convolution of @xmath1 with @xmath52 , while for a general metric space it may be understood as an extension of the convolution operation .
we also remark that the starting motivation of our study is related to the problem of rectifiability of a @xmath53varifold @xmath54 in @xmath51 obtained as the limit of `` discrete varifolds '' ( see @xcite ) .
the paper is organized as follows . in section @xmath55
, we explain how to reconstruct a positive measure and then a signed measure ( theorem [ prop_positive_case ] ) from their values on balls , thanks to carathodory s construction , answering question [ question_1 ] .
section @xmath56 deals with question [ question_2 ] , that is , the reconstruction of a measure starting from a premeasure satisfying . after explaining the limitations of carathodory s construction for this problem
, we prove our main result , theorem [ thm_main ] , saying that by suitable packing constructions one can reconstruct a signed measure equivalent to the initial one in any metric space @xmath0 which is directionally limited and endowed with an asymptotically doubling measure @xmath6 ( see hypothesis [ hypo1 ] in page ) .
let @xmath0 be a metric space .
* @xmath57 denotes the @xmath58algebra of borel subsets of @xmath3 .
* @xmath59 is the closed ball of radius @xmath60 and center @xmath61 .
* @xmath62 is the open ball of radius @xmath60 and center @xmath61 .
* @xmath15 denotes the collection of closed balls of @xmath3 and for @xmath63 , @xmath64 denotes the collection of closed balls of diameter @xmath65 . *
@xmath66 is the lebesgue measure in @xmath67 .
* @xmath68 is the set of all subsets of @xmath3 .
* @xmath69 is the cardinality of the set @xmath9 .
* @xmath70 .
we recall here some standard definitions and well - known facts about general measures , focusing in particular on the construction of measures from premeasures , in the sense of carathodory method ii @xcite .
let @xmath3 be a set , and let @xmath71 $ ] satisfying 1 .
2 . @xmath18 is monotone : if @xmath73 , then @xmath74 .
3 . @xmath18 is _ countably subadditive _ :
if @xmath75 is a sequence of subsets of @xmath3 , then @xmath76 then @xmath18 is called an _ outer measure _ on @xmath3 . in order to obtain a measure from an outer measure
, one defines the measurable sets with respect to @xmath18 .
let @xmath18 be an outer measure on @xmath3 .
a set @xmath20 is _
@xmath18measurable _ if for all sets @xmath77 , @xmath78 we can now define a measure associated with an outer measure .
thanks to the definition of @xmath18measurable sets , the additivity of @xmath18 among the measurable sets is straightforward , actually it happens that @xmath18 is @xmath58additive on @xmath18measurable sets .
[ thm_from_outer_measure_to_measure ] let @xmath3 be a set , @xmath18 an outer measure on @xmath3 , and @xmath79 the class of @xmath18measurable sets
. then @xmath79 is a @xmath58algebra and @xmath18 is countably additive on @xmath79 .
thus the set function @xmath1 defined on @xmath79 by @xmath80 for all @xmath81 is a measure .
we now introduce _ metric outer measures_. let @xmath82 be a metric space and @xmath18 be an outer measure on @xmath3 .
@xmath18 is called a _ metric outer measure _ if @xmath83 for any @xmath84 such that @xmath85 .
when @xmath18 is a metric outer measure , every borel set is @xmath18measurable and thus the measure @xmath1 associated with @xmath18 is a borel measure .
let @xmath18 be an outer measure on a metric space @xmath0
. then every borel set in @xmath3 is @xmath22-measurable if and only if @xmath18 is a metric outer measure .
in particular , a metric outer measure is a borel measure .
we recall two approximation properties of borel measures defined on metric spaces .
[ thm_approximation_borel_measure ] let @xmath0 be a metric space and @xmath1 be a borel measure on @xmath3 .
* _ approximation from inside _ : let @xmath86 be a borel set such that @xmath87 , then for any @xmath88 , there exists a closed set @xmath89 such that @xmath90 . * _ approximation from outside _ : assume that @xmath1 is finite on bounded sets and let @xmath86 be a borel set , then @xmath91 we can now introduce carathodory s construction of metric outer measures ( method ii , see @xcite ) .
let @xmath3 be a set and @xmath15 be a family of subsets of @xmath3 such that @xmath16 .
a nonnegative function @xmath14 defined on @xmath15 and such that @xmath92 is called a _
premeasure_. [ thm_caratheodory_construction ] suppose @xmath0 is a metric space and @xmath15 is a family of subsets of @xmath3 which contains the empty set .
let @xmath14 be a premeasure on @xmath15 .
for each @xmath63 , let @xmath93 and for any @xmath77 define @xmath94 as @xmath95 when @xmath96 , the limit @xmath97 exists ( possibly infinite ) .
then @xmath98 is a metric outer measure on @xmath3 .
let @xmath0 be a metric space and @xmath1 be a positive borel @xmath58finite measure on @xmath3 .
let @xmath15 be the set of closed balls and let @xmath14 be the premeasure defined on @xmath15 by @xmath99 \\ & & b & \mapsto & \mu ( b ) \end{array}\ ] ] let @xmath100 be the metric outer measure obtained by carathodory s metric construction applied to @xmath101 and then @xmath102 the borel measure associated with @xmath100 .
then , the following question arises .
[ question_positive_case ] _ do we have @xmath103 ?
in other terms , can we recover the initial measure by carathodory s method ii ? _
the following lemma shows one of the two inequalities needed to positively answer question [ question_positive_case ] .
[ lem_easy_inequality ] let @xmath0 be a metric space and @xmath1 be a positive borel measure on @xmath3 . then , in the same notations as above , we have @xmath104 .
let @xmath20 be a borel set , we have to show that @xmath105 .
this inequality relies only on the definition of @xmath106 as an infimum .
indeed , let @xmath107 be fixed , then for any @xmath108 there exists a countable collection of closed balls @xmath109 such that @xmath110 so that @xmath111 letting @xmath112 and then @xmath113 leads to @xmath114 .
a consequence of davies result @xcite is that the other inequality can not be true in general .
we need extra assumptions on the metric space @xmath0 ensuring that open sets are `` well approximated '' by closed balls , that is , we need some specific covering property . in @xmath67 with the euclidean norm , this approximation of open sets by disjoint unions of balls is provided by besicovitch theorem , which we recall here : [ thm_besicovitch ] let @xmath1 be a borel measure on @xmath67 and consider any collection @xmath115 of non degenerated closed balls .
let @xmath9 denote the set of centers of the balls in @xmath115 .
assume @xmath116 and that @xmath117 then , for every open set
@xmath118 , there exists a countable collection @xmath119 of disjoint balls in @xmath115 such that @xmath120 a generalization of besicovitch theorem for metric spaces is due to federer , under a geometric assumption involving the distance function .
[ def : federer ] let @xmath0 be a metric space , @xmath20 and @xmath121 , @xmath122 , @xmath123 .
the distance @xmath53 is said to be _ directionally @xmath124limited at @xmath9 _ if the following holds .
take any @xmath125 and @xmath126 satisfying the following property : let @xmath127 with @xmath128 and assume without loss of generality that @xmath129 , then for all @xmath61 such that @xmath130 and @xmath131 one has @xmath132 then @xmath133 .
let us say a few words about this definition . if @xmath134 is a banach space with strictly convex norm , then the above relations involving @xmath135 imply that @xmath136 hence in this case is equivalent to @xmath137 consequently , if @xmath3 is finite - dimensional , and thanks to the compactness of the unit sphere , for a given @xmath138 there exists @xmath139 such that @xmath140 is directionally @xmath124-limited for all @xmath141 .
hereafter we provide two examples of metric spaces that are not directionally limited .
consider in @xmath142 the union @xmath3 of a countable number of half - lines , joining at the same point @xmath143 .
then the geodesic metric @xmath53 induced on @xmath3 by the ambient metric is not directionally limited at @xmath144 .
indeed let @xmath145 , let @xmath146 and @xmath147 lying in two different lines , at the same distance @xmath148 of @xmath143 .
then @xmath61 such that @xmath149 and @xmath150 implies @xmath151 and thus @xmath152 but @xmath153 is not finite .
if @xmath3 is a separable hilbert space and @xmath154 a hilbert basis , @xmath155 and @xmath156 , the hilbert norm is strictly convex so that @xmath130 , @xmath157 uniquely define @xmath135 as @xmath158 for all @xmath159 and @xmath160 is infinite .
therefore @xmath161 is not directionally limited ( nowhere ) .
we can now state the generalized versions of besicovitch covering lemma and besicovitch theorem for directionally limited metric spaces .
[ thm_generalized_besicovic_federer_covering_lemma ] let @xmath0 be a separable metric space directionally @xmath124limited at @xmath20 . let @xmath162 and @xmath115 be a family of closed balls of radii less than @xmath163 such that each point of @xmath9 is the center of some ball of @xmath115 .
then , there exists @xmath164 countable subfamilies of @xmath115 of disjoint closed balls , @xmath165 such that @xmath166 in @xmath67 endowed with the euclidean norm it is possible to take @xmath167 and @xmath168 only dependent on @xmath138 and @xmath169 .
if we fix @xmath170 , then @xmath171 only depends on the dimension @xmath169 .
[ thm_generalized_besicovic_federer ] let @xmath0 be a separable metric space directionally @xmath124limited at @xmath20 .
let @xmath115 be a family of closed balls of @xmath3 satisfying @xmath172 and let @xmath1 be a positive borel measure on @xmath3 such that @xmath116 .
then , for any open set @xmath173 there exists a countable disjoint family @xmath119 of @xmath115 such that @xmath120 we can now prove the coincidence of the initial measure and the reconstructed measure under assumptions of theorem [ thm_generalized_besicovic_federer ] .
[ prop_positive_case ] let @xmath0 be a separable metric space directionally @xmath124limited at @xmath3 .
let @xmath1 be a positive borel measure on @xmath3 , finite on bounded sets .
let @xmath15 be the family of closed balls in @xmath3 and let @xmath14 be the premeasure defined on @xmath15 by . denote by @xmath100 the metric outer measure obtained by carathodory s metric construction applied to @xmath101 and by @xmath102 the borel measure associated with @xmath100
. then @xmath103 .
* step one . *
we first show that @xmath102 is a borel measure finite on bounded sets .
first we recall that @xmath100 is a metric outer measure by theorem [ thm_caratheodory_construction ] , then thanks to theorem [ thm_from_outer_measure_to_measure ] @xmath102 is a borel measure .
let us prove that @xmath102 is finite on bounded sets .
fix @xmath20 a bounded borel set and apply besicovitch covering lemma ( theorem [ thm_generalized_besicovic_federer_covering_lemma ] ) with the family @xmath174 to get @xmath175 countable subfamilies @xmath176 of disjoint balls in @xmath115 such that @xmath177 therefore , @xmath178 where @xmath179 is bounded , thus @xmath180 by assumption and hence for all @xmath181 @xmath182 whence @xmath183 .
the claim is proved since @xmath184 on borel sets .
* step two . *
we now prove that for any open set @xmath173 , it holds @xmath185 .
let @xmath63 be fixed . the collection of closed balls @xmath186 satisfies assumption
. we can apply theorem [ thm_generalized_besicovic_federer ] to @xmath102 and get a countable collection @xmath187 of disjoint balls in @xmath64 such that @xmath188 at the same time we have @xmath189 we fix any decreasing infinitesimal sequence @xmath190 and define @xmath191 .
we obtain @xmath192 and @xmath193 for any @xmath169 .
thus , owing to , we have @xmath194 this shows that @xmath195 , as wanted
. * step three . * since @xmath1 and @xmath102 are borel measures , finite on bounded sets , they are also outer regular ( see theorem [ thm_approximation_borel_measure ] )
. then for any borel set @xmath196 , and owing to step two , it holds @xmath197 coupling this last inequality with lemma [ lem_easy_inequality ] we obtain @xmath103 .
we recall that a borel signed measure @xmath1 on @xmath0 is an extended real - valued set function @xmath198 $ ] such that @xmath199 and , for any sequence of disjoint borel sets @xmath200 , one has @xmath201 notice that when @xmath202 is finite , its value does not depend on the arrangement of the @xmath203 , therefore the series on the right hand side of is commutatively convergent , thus absolutely convergent . in particular , if we write the hahn decomposition @xmath23 , with @xmath204 and @xmath32 being two non - negative and mutually orthogonal measures , then @xmath205 and @xmath206 can not be both @xmath207
. the question is now the following : _ let @xmath0 be a metric space , separable and directionally @xmath124limited at @xmath3 , and let @xmath1 be a borel signed measure , finite on bounded sets .
is it possible to recover @xmath1 from its values on closed balls by some carathodory - type construction ? _
the main difference with the case of a positive measure is that @xmath1 is not monotone and thus the previous construction is not directly applicable .
a simple idea could be to rely on the hahn decomposition of @xmath1 : indeed , @xmath31 and @xmath208 are positive borel measures , and since one of them is finite , both are finite on bounded sets ( recall that @xmath1 is finite on bounded sets by assumption ) . once again , we can not directly apply carathodory s construction to @xmath31 or @xmath208 since we can not directly reconstruct @xmath25 and @xmath26 simply knowing @xmath27 for any closed ball @xmath24 .
we thus try to apply carathodory s construction not with @xmath25 , but with @xmath209 , where @xmath210 ( resp .
@xmath211 ) denote the positive part @xmath212 ( resp . the negative part @xmath213 ) for any @xmath214 . to be more precise ,
we state the following definition . [
caratheodory_construction_signed ] let @xmath1 be a borel signed radon measure in @xmath3 and let @xmath15 be the family of closed balls in @xmath3 .
we define @xmath215 then according to carathodory s construction , we define the metric outer measure @xmath216 such that for any @xmath20 , @xmath217 similarly we define @xmath218 and then call @xmath219 and @xmath220 the borel measures associated with @xmath221 and @xmath218 . finally , we set @xmath222 .
[ thm_signed_case ]
let @xmath0 be a metric space , separable and directionally @xmath124limited at @xmath3 and let @xmath23 be a borel signed measure on @xmath3 , finite on bounded sets .
let @xmath223 be as in definition [ caratheodory_construction_signed ]
. then @xmath103 .
we observe that @xmath219 and @xmath220 are borel measures : indeed , by construction they are metric outer measures and carathodory criterion implies that they are borel measures .
furthermore , for any closed ball @xmath224 , if we set @xmath225 ( note that @xmath226 is the canonical premeasure associated with @xmath31 while @xmath227 is not a priori associated with any measure ) then @xmath228 thus by construction , @xmath229 and then one gets @xmath230 ; similarly one can show that @xmath231 .
thanks to proposition [ prop_positive_case ] , we have proven that @xmath232 in particular , @xmath219 and @xmath220 are finite on bounded sets , as it happens for @xmath31 and @xmath208 .
let now @xmath20 be a borel set .
it remains to prove that @xmath233 ( and the same for @xmath220 ) .
we argue exactly as in the proof of lemma [ lem_easy_inequality ] .
let @xmath107 , then for any @xmath108 there exists a countable collection of closed balls @xmath109 such that @xmath234 and @xmath235 , so that @xmath236 letting @xmath112 and then @xmath113 gives @xmath237 we recall that @xmath31 and @xmath208 are mutually singular , hence there exists a borel set @xmath238 such that for any borel set @xmath9 @xmath239 thanks to we already know that @xmath240 , therefore we get @xmath241 for any borel set @xmath9 .
thanks to , we finally infer that @xmath242 , @xmath243 , i.e. , that @xmath103 .
if @xmath1 is a vector - valued measure on @xmath3 , with values in a finite vector space @xmath244 , we can apply the same construction componentwise .
we now want to reconstruct a measure @xmath1 ( or an approximation of @xmath1 ) starting from approximate values on closed balls , given by a premeasure @xmath2 satisfying .
more precisely , we can now reformulate question 2 in the context of directionally limited metric spaces .
_ let @xmath0 be a separable metric space , directionally @xmath124limited at @xmath3 and let @xmath1 be a positive borel measure on @xmath3 .
is it possible to reconstruct @xmath1 from @xmath2 up to multiplicative constants ? can the same be done when @xmath1 is a signed measure ? _ in section [ section:2.1 ] below we explain with a simple example involving a dirac mass why carathodory s construction does not allow to recover @xmath1 from the premeasure @xmath2 defined in . then in section [ sect : packtype ] we define a _ packing construction _ of a measure , that is in some sense dual to carathodory method ii , and we show that in a directionally limited and separable metric space @xmath0 , endowed with an asymptotically @xmath245-bounded measure @xmath6 ( see ) it produces a measure equivalent to the initial one .
let us consider a dirac mass @xmath246 in @xmath67 and define @xmath247 it is easy to check that this particular choice of premeasure @xmath2 satisfies .
first of all , for any @xmath248 , @xmath249 if now @xmath250 is at distance @xmath138 from @xmath135 for some @xmath251 , we have @xmath252 therefore , @xmath253 as @xmath254 .
we can thus find a covering made by a single ball of radius less than @xmath255 for which @xmath256 is as small as we wish .
this shows that carathodory s construction associated with this premeasure produces the zero measure .
more generally , as soon as it is possible to cover with small balls such that the mass of the measure inside each ball is close to the boundary of the ball , one sees that carathodory s construction `` looses mass '' .
for instance , take @xmath257 , where @xmath258 is a curve of length @xmath259 and @xmath260 is the @xmath13-dimensional hausdorff measure in @xmath51 , then cover @xmath261 with a family of closed balls @xmath262 of radii @xmath163 with centers at distance @xmath138 from @xmath261 . assuming that no portion of the curve is covered more than twice , then @xmath263 { } 0 \ : , \end{aligned}\ ] ] with @xmath264 for instance .
the same phenomenon can not be excluded by blindly centering balls on the support of the measure @xmath1 .
indeed , take a line @xmath265 with a dirac mass placed on it at some @xmath266 , so that @xmath267 . then , by centering the balls on the support of @xmath1 , we may recover the hausdorff measure restricted to @xmath268 , but not the dirac mass , for the same reason as before .
we thus understand that the position of the balls should be optimized in order to avoid the problem .
for this reason we consider an alternative method , based on a packing - type construction .
taking into account the problems described in the examples of the previous section , one realizes the need to optimize the position of the centers of the balls in order to properly reconstruct the original measure @xmath1 . the idea is to consider a kind of dual construction , that is , a supremum over packings rather than an infimum over coverings . to this aim
we recall a notion of _ packing _ of balls .
[ def : packing ] let @xmath0 be a separable metric space and @xmath173 be an open set .
we say that @xmath115 is a packing of @xmath269 of order @xmath163 if @xmath115 is a countable family of disjoint closed balls whose radius is less than @xmath163 and such that @xmath270 [ dfn_packing_construction ] let @xmath0 be a separable metric space and let @xmath2 be a premeasure defined on closed balls .
let @xmath173 be an open set and fix @xmath107 .
we set @xmath271 and , in a similar way as in carathodory construction , we define @xmath272 and note that @xmath273 implies @xmath274 . then , @xmath4 can be extended to all @xmath20 by setting @xmath275 the main difference between definition [ dfn_packing_construction ] and carathodory s construction is that the set function @xmath4 is not automatically an outer measure : it is monotone but not sub - additive in general . in order to fix this problem we may apply the construction of outer measures , known as munroe method i , to the set function @xmath4 restricted to the class of open sets .
this amounts to setting , for any @xmath20 , @xmath276 one can check that @xmath277 is an outer measure .
we will prove in theorem [ thm_main ] that , for the class of set functions @xmath2 we are focusing on , and under additional assumptions on@xmath278 , @xmath4 is already a borel outer measure .
[ remk_sub_additivity_and_tilde_mu ] knowing that @xmath4 is sub - additive in the class of open sets is enough to show that @xmath279 .
indeed , the inequality @xmath280 comes directly from the fact that minimizing @xmath281 over @xmath269 open such that @xmath282 is a special case of minimizing @xmath283 among countable families of open sets @xmath284 such that @xmath285 .
assuming in addition that @xmath4 is sub - additive on open sets implies that for any countable family of open sets @xmath286 such that @xmath285 , @xmath287 by definition of @xmath277 , taking the infimum over such families leads to @xmath288 .
our construction of the measure @xmath4 is slightly different from the more classical packing construction proposed by tricot and taylor in @xcite . in particular , enforces outer regularity of @xmath4 , while for instance the classical @xmath45dimensional packing measure in @xmath67 associated with the premeasure @xmath289 is not outer regular for @xmath290 ( see @xcite ) .
a more specific comparison between our definition and the one by tricot and taylor will be carried out in section [ sect : blversustt ] , where it will be proved that , under the assumption that @xmath291 for every ball @xmath292 and for some constant @xmath293 and some radon measure @xmath1 , the two constructions actually produce the same measure .
[ thm_main ] let @xmath0 be a metric space satisfying hypothesis [ hypo1 ] , let @xmath1 be a positive borel measure on @xmath3 and let @xmath2 be a premeasure given on closed balls satisfying .
let @xmath4 be as in definition [ dfn_packing_construction ] .
then , the following hold : 1 . for any borel set @xmath20 , @xmath294 where @xmath295 and @xmath296 are the constants respectively appearing in hypothesis [ hypo1 ] and .
@xmath297 is countably sub - additive and is a metric outer measure ; 3 .
if moreover @xmath1 is outer regular , then @xmath1 and the positive borel measure associated with the outer measure @xmath4 ( still denoted as @xmath4 ) are equivalent , and more precisely @xmath298 .
we briefly sketch the proof of this result for the reader s convenience . 1 .
the first step ( see proposition [ prop_sub_additivity_open_case ] ) is to prove that @xmath297 is countably sub - additive on any sequence of open sets @xmath299 such that @xmath300
. this property does not require hypothesis @xmath13 and only relies on the upper bound (ii ) on @xmath2 .
2 . then , assuming moreover that @xmath1 is finite on bounded sets , we prove in proposition [ prop_sub_additivity_loc_finite_case ] the countable sub - additivity of @xmath297 .
now we show in proposition [ prop_equivalence_on_open_sets ] a crucial fact , i.e. that the finiteness of the measure @xmath1 can be dropped , thus we only require that @xmath1 is a positive borel measure ( not necessarily finite on bounded sets ) and that @xmath0 satisfies hypothesis [ hypo1 ] .
the heart of the proof is to use the lower bound ( i ) on @xmath2 and show that it can be transferred to @xmath297 , which gives .
4 . up to this last step
, we still do not know that @xmath297 is countably sub - additive .
this is proved in proposition [ cor_countable_sub_additvity ] and follows from the partial sub - additivity result of the first step and the lower bound in .
we stress that this last result does not require that @xmath1 is finite on bounded sets .
[ prop_sub_additivity_open_case ] let @xmath0 be a separable metric space and let @xmath1 be a positive borel measure on @xmath3 .
let @xmath2 be a premeasure defined on the class @xmath301 of closed balls contained in @xmath3 , such that ( ii ) holds .
then , for any countable family of open sets @xmath302 satisfying @xmath303 , one has @xmath304 in particular , if @xmath1 is finite , then @xmath4 is an outer measure .
let @xmath305 be a sequence of open subsets of @xmath3 such that @xmath306 .
let @xmath88 , then for all @xmath307 we define @xmath308 fix @xmath309 .
if @xmath24 is a closed ball such that @xmath310 and @xmath311 , then there exists @xmath312 such that @xmath313 .
indeed , @xmath314 and there exists @xmath312 such that @xmath315 and thus @xmath316 of course the inclusion @xmath317 remains true for any closed ball @xmath24 with @xmath310 .
therefore any packing @xmath318 of @xmath319 of order @xmath163 can be decomposed as the union of a countable family of packings @xmath320 , where @xmath321 is a packing of @xmath284 of order @xmath163 , whence @xmath322 by taking the supremum over all such packings @xmath318 of @xmath319 , we get @xmath323 then , taking the infimum over @xmath63 and then the supremum over @xmath88 gives @xmath324 we now want to prove that @xmath325 let @xmath318 be a packing of @xmath326 of order @xmath327 .
we have @xmath328 notice that since @xmath329 , for any @xmath330 , if @xmath331 then @xmath332 . since @xmath333 according to ( ii ) , we get @xmath334 owing to the fact that @xmath335 , and @xmath336 is decreasing in @xmath337 , we have that @xmath338 { } 0\ ] ] as soon as @xmath339 , which is true under the assumption @xmath340 .
therefore , by , and we infer that @xmath341 taking the supremum in over all packings @xmath318 of order @xmath163 of @xmath342 , we get @xmath343 then taking the limit as @xmath344 we obtain @xmath345 and finally , letting @xmath346 , we prove .
we now turn to the right hand side of . for fixed @xmath347 ,
@xmath348 is decreasing when @xmath349 , therefore @xmath350 provided that @xmath351 is finite for some @xmath63 , which is true since @xmath352 , @xmath353 for all @xmath347 , so that @xmath354 .
finally , thanks to , and we obtain the conclusion . [ prop_sub_additivity_loc_finite_case ] let @xmath0 be a separable metric space and let @xmath1 be a positive borel measure on @xmath3 , finite on bounded sets .
let @xmath2 be a premeasure defined on the class @xmath301 of closed balls contained in @xmath3 , such that ( ii ) holds .
then @xmath297 is countably sub - additive , thus it is an outer measure .
let @xmath200 be a countable family of disjoint sets such that @xmath355 .
we shall prove that @xmath356 let @xmath357 . by outer regularity of @xmath1 ( since @xmath1 a borel measure finite on bounded sets , it is outer regular by theorem [ thm_approximation_borel_measure ] ) and by definition of @xmath297 , let @xmath358 be a family of open sets such that , for any @xmath347 , @xmath359 hence @xmath360 and by proposition [ prop_sub_additivity_open_case ] we thus find @xmath361 let @xmath362 to get .
the case of a countable family @xmath200 such that @xmath363 is obtained from the case of disjoint sets in the standard way , by defining @xmath364 as @xmath365 the family @xmath366 is disjoint and @xmath367 , hence by one gets @xmath368 finally , let us consider the general case of @xmath200 being any countable family of sets .
let @xmath369 be an increasing family of bounded sets such that @xmath370 , then for all @xmath169 , @xmath371 is bounded , hence @xmath372 and therefore @xmath373 we let @xmath374 and we infer by monotone convergence that @xmath375 in order to have the countable sub - additivity of @xmath4 ( in the case where @xmath1 is not assumed to be finite on bounded sets ) it is enough to show that @xmath376 implies @xmath377 . if this is true , then either @xmath378 and the sub - additivity is given by proposition [ prop_sub_additivity_open_case ] , or @xmath377 and the sub - additivity is immediate .
now the core of the problem is to estimate @xmath4 from below by @xmath1 .
the main issue is that the lower bound ( i ) does not imply the stronger lower bound @xmath379 moreover , unless we know that the measure @xmath1 is doubling , there is no reason to expect that the inequality @xmath380 is verified for some @xmath381 ( where by @xmath382 we denote the ball concentric to @xmath24 with double radius ) and for any ball @xmath24 . nevertheless , by assuming that @xmath0 is directionally bounded and by comparing @xmath1 with an asymptotically doubling measure @xmath6 on @xmath3 , which we assume to exist , we are able to prove that a doubling property for @xmath1 actually holds for enough balls , so that we can choose packings among these balls .
before showing the result , we need to introduce the notion of _ asymptotically @xmath383-bounded measure_. we say that @xmath6 is an asymptotically @xmath383-bounded measure on @xmath0 if it is finite on bounded sets and strictly positive on any ball with positive radius , and for all @xmath38 it satisfies @xmath384 notice that if we assume that @xmath6 is asymptotically doubling on @xmath3 , that is , @xmath6 is finite on bounded sets and there exists a constant @xmath385 such that for all @xmath38 it holds @xmath386 then for any @xmath387 $ ] , taking @xmath388 as the unique integer such that @xmath389 , one can easily check that @xmath6 is asymptotically @xmath390-bounded .
we conveniently state some key properties on the metric space @xmath0 , that will be constantly assumed in the rest of this section .
[ hypo1 ] @xmath0 is a directionally limited metric space endowed with an asymptotically @xmath245-bounded measure @xmath6 satisfying for some constants @xmath391 $ ] and @xmath392 .
[ prop_diffusion_in_x ] assume that @xmath393 satisfy hypothesis and let @xmath1 be a positive borel measure on @xmath3 .
let @xmath394 then the following hold .
a. for all @xmath395 , either @xmath396 for all @xmath60 , or @xmath397 b. @xmath398 . in particular , for a fixed @xmath399 and for @xmath1almost any @xmath61 ,
there exists a decreasing infinitesimal sequence @xmath400 of radii ( depending on @xmath401 and @xmath135 ) , such that @xmath402 * proof of ( i ) .
* let @xmath395 .
by monotonicity , either @xmath396 for all @xmath60 ( and then is also trivially satisfied ) or there exists some @xmath403 such that , for all @xmath404 , @xmath405 . in this case
the function defined by @xmath406 is non - negative and finite for @xmath255 small enough .
moreover , since @xmath395 we have @xmath407 .
let us prove that @xmath408 assume by contradiction that @xmath409 , then necessarily @xmath410 and there exists @xmath411 and @xmath412 such that for all @xmath413 , @xmath414 .
consider now the sequence @xmath415 defined by @xmath416 . then @xmath417 and @xmath418 { } 0\ ] ] which contradicts @xmath407 and thus proves .
let us notice that @xmath419 for all @xmath60 since @xmath420 and that , by definition , we have @xmath421 since @xmath6 is asymptotically @xmath245-bounded , by we get @xmath422 * proof of ( ii ) .
* let us show that @xmath398 .
first assume that @xmath423 and let @xmath88 .
consider @xmath424 let @xmath425 be fixed . since @xmath426
, there exists @xmath427 such that @xmath428 . every point in @xmath429
is the center of some ball in @xmath430 , so that we can apply theorem [ thm_generalized_besicovic_federer_covering_lemma ] and obtain @xmath431 countable families @xmath432 of disjoint balls in @xmath430 , such that @xmath433 therefore @xmath434 hence @xmath398 if @xmath423 . otherwise , replace @xmath3 by @xmath435 for any fixed @xmath436 to obtain that for any @xmath437 , @xmath438 , then let @xmath439 to conclude that @xmath398 .
finally is an immediate consequence of the fact that @xmath440 coupled with ( i ) and ( ii ) .
[ bescivoic_with_doubling_balls ] assume that @xmath393 satisfy hypothesis and let @xmath1 be a positive borel measure on @xmath3 .
fix @xmath441 and for any @xmath63 define @xmath442 and , for any @xmath20 , @xmath443 then there exist @xmath444 and @xmath431 countable subfamilies of @xmath445 of disjoint closed balls , @xmath446 such that @xmath447 moreover , if @xmath116 , then for any open set @xmath173 there exists a countable collection @xmath119 of disjoint balls in @xmath445 such that @xmath448 thanks to ( see proposition [ prop_diffusion_in_x ] ) we know that for @xmath1-almost every @xmath61 there exists a decreasing infinitesimal sequence @xmath400 such that @xmath449 for all @xmath450 .
hence for @xmath1-almost any @xmath61 we have @xmath451 then , the conclusion follows from theorems [ thm_generalized_besicovic_federer_covering_lemma ] and [ thm_generalized_besicovic_federer ] .
we can now prove that @xmath4 and @xmath1 are equivalent on borel sets ( as set functions since we have not completely proved the sub - additivity of @xmath4 yet ) .
[ prop_equivalence_on_open_sets ] let @xmath393 satisfy hypothesis with constants @xmath383 and let @xmath1 be a positive borel measure on @xmath3 .
let @xmath2 be a premeasure satisfying with constants @xmath452 and @xmath296 .
let @xmath4 be as in definition [ dfn_packing_construction ] .
then for any borel set @xmath20 we have @xmath453 therefore if @xmath1 is outer regular then @xmath454 let @xmath173 be an open set , then the inequality @xmath455 is just a consequence of the definition of @xmath4 and of the second inequality in , i.e. , of the fact that , for any closed ball @xmath24 , @xmath352 .
now we prove the other inequality by splitting the problem into two cases .
_ case @xmath456 .
_ let @xmath399 and @xmath63 , then we can apply corollary [ bescivoic_with_doubling_balls ] ( besicovitch with doubling balls ) to get a countable family @xmath457 of disjoint balls of @xmath458 such that @xmath459 therefore by and by definition of @xmath460 , @xmath461 letting @xmath344 and then @xmath462 gives @xmath463 .
_ case @xmath464 .
_ let @xmath63 and @xmath465 be as in the previous case , then applying corollary [ bescivoic_with_doubling_balls ] ( besicovitch with doubling balls ) gives @xmath431 countable families @xmath466 of balls in @xmath467 such that @xmath468 then we get @xmath469 consequently there exists @xmath470 such that @xmath471 . therefore we have the same estimate as in the case @xmath472 : @xmath473 and we conclude @xmath474 .
[ cor_countable_sub_additvity ] let @xmath393 satisfy hypothesis .
let @xmath1 be a positive borel measure on @xmath3 and let @xmath2 be a premeasure satisfying .
let @xmath4 be as in definition [ dfn_packing_construction ] .
then @xmath4 is countably sub - additive .
let @xmath475 be a countable collection of subsets of @xmath3 . if @xmath476 then by proposition [ prop_equivalence_on_open_sets ] there exists @xmath477 such that @xmath478 for all @xmath169 , therefore @xmath479 whence the countable sub - additivity directly follows . recall that if @xmath480 and @xmath481 are open sets , then countable sub - additivity was proved in proposition [ prop_sub_additivity_open_case ] .
it remains to check the case @xmath480 for a generic sequence of borel sets @xmath481 .
fix @xmath357 , then by definition of @xmath297 there exists an open set @xmath482 for all @xmath169 , depending also on @xmath337 , such that @xmath483 and @xmath484 by sub - additivity on open sets we have @xmath485 letting @xmath362 concludes the proof .
our packing construction is very similar to the one introduced by taylor and tricot in @xcite for measures in @xmath67 . in that paper ,
starting from a given premeasure @xmath2 defined on a family of sets @xmath15 ( here , as in our construction , @xmath15 will be the family of closed balls ) a so - called _ packing premeasure _ is defined for any @xmath486 as @xmath487 where their notion of packing ( here specialized to families of closed balls , which we will refer to as @xmath488packing ) is slightly different from the one we introduced in definition [ def : packing ] .
[ def : tpacking ] let @xmath486 .
we say that @xmath115 is a @xmath488-packing of @xmath244 of order @xmath163 if @xmath115 is a countable family of disjoint closed balls whose radius is less or equal to @xmath163 and such that for all @xmath489 , @xmath490 .
we insist on the fact that such a packing is _ not _ included in @xmath244 ( as required in definition [ def : packing ] ) but only in some enlargement of @xmath244 .
then , from this packing premeasure @xmath491 , a packing measure @xmath492 is defined by applying carathodory s construction , method i , to @xmath493 on borel sets . to be precise , for any @xmath494 , @xmath495 we will now prove that these two constructions are equivalent when the premeasure @xmath2 is controlled by some radon measure @xmath1 , i.e. , when there exists @xmath496 such that for every closed ball @xmath24 , @xmath497 .
let @xmath2 be a premeasure defined on closed balls in @xmath51 .
we assume that there exists a radon measure @xmath1 and a constant @xmath293 such that @xmath498 for any closed ball @xmath292
. then @xmath499 .
let us prove that for any compact set @xmath500 @xmath501 fix an open set @xmath269 containing @xmath502 .
as @xmath502 is compact , there exists @xmath503 such that , for all @xmath504 , @xmath505 note that if @xmath318 is a @xmath488packing of order @xmath163 of @xmath502 then it is a packing of order @xmath163 of @xmath506 and thus of @xmath269 .
hence , for all @xmath507 @xmath508 let us pass to the limit as @xmath113 and obtain by definition @xmath509 the inequality is true for any open set @xmath510 , therefore @xmath511 since @xmath512 is trivially a covering of @xmath502 , then @xmath513 , which leads to .
it is easy to extend to any bounded borel set thanks to theorem [ thm_approximation_borel_measure ] and the fact that @xmath1 is radon and @xmath67 is locally compact . then for any borel set @xmath486 , @xmath514 let us now prove that for any bounded set @xmath486 we have the converse inequality @xmath515 given @xmath357 and @xmath516 we define @xmath517 which is obviously an open set
let @xmath318 be a @xmath163packing of @xmath518 , then @xmath519 with @xmath520 then , by definition [ def : tpacking ] , @xmath521 is a @xmath488packing of order @xmath163 of @xmath244 while any ball of @xmath522 is included in @xmath523 .
therefore , taking the supremum over all @xmath163packings of @xmath518 , we get @xmath524 then letting @xmath113 leads to @xmath525 as @xmath1 is radon and @xmath523 is bounded ( since @xmath244 is bounded ) , then @xmath526 , @xmath527 is increasing in @xmath337 and @xmath528 , so that @xmath529 { } 0 \ : .\ ] ] letting @xmath362 in , we get @xmath530 thanks to proposition [ prop_sub_additivity_loc_finite_case ] , we know that @xmath297 is countably sub - additive and for every countable family @xmath531 such that @xmath532 , we have by @xmath533 finally taking the infimum over all such @xmath531 gives , whence @xmath534 holds for any borel set @xmath486 .
this concludes the proof .
reading carefully the proof , one should note that the property of @xmath67 which is used is the local compactness of @xmath67 .
therefore , if we extend the definition of taylor and tricot @xcite to a metric space @xmath0 , then assuming @xmath0 locally compact the two packing constructions still coincide under the assumption @xmath535 for all closed balls @xmath24 , where @xmath1 is some given radon measure and @xmath496 .
our aim is to prove that the packing - type reconstruction applied to a signed measure @xmath1 , with premeasures @xmath536 satisfying @xmath537 and @xmath538 for some @xmath34 and @xmath33 , produces a signed measure @xmath539 whose positive and negative parts are comparable with those of @xmath1 . we notice that properties and are weaker than the following ( and apparently more natural ) ones : @xmath540 on the other hand we note that the premeasures defined as @xmath541 satisfy but not .
[ thm : signedboh ] let @xmath393 satisfy hypothesis [ hypo1 ] for some constants @xmath387 $ ] and @xmath542 , and let @xmath23 be a locally finite , borel - regular signed measure on @xmath3 .
let @xmath543 be a pair of premeasures satisfying and for some @xmath34 .
take @xmath544 as in definition [ dfn_packing_construction ] .
then the following properties hold : * @xmath545 , @xmath546 are metric outer measures finite on bounded sets ; * @xmath547 is a signed measure such that , for any borel set @xmath20 , @xmath548 whence in particular @xmath549 the countable sub - additivity of @xmath550 follows from the second inequalities in ( see proposition [ prop_sub_additivity_loc_finite_case ] ) .
then for any open set @xmath173 both inequalities @xmath551 are just a consequence of the definition of @xmath550 and of the second inequalities in . this proves ( i ) .
let now @xmath20 be a borel set .
we first derive an estimate concerning @xmath552 ( the estimate for @xmath553 can be obtained in the same way ) . if @xmath554 , we take an open set @xmath269 containing @xmath9 such that @xmath555 .
let @xmath556 be fixed , then apply corollary [ bescivoic_with_doubling_balls ] to @xmath204 and get a countable family @xmath557 of disjoint closed balls contained in @xmath269 with radii @xmath558 and satisfying @xmath559 for all @xmath560 , such that @xmath561 we have @xmath562 letting @xmath344 and then @xmath563 we find @xmath564 by definition of @xmath565 , there exists a sequence of open sets @xmath566 such that , for all @xmath347 , it holds @xmath567 and @xmath568 { } { { \hat \mu}}^{q_+ } ( a ) \ : .\ ] ] by outer regularity of @xmath208 ( which is borel and finite on bounded sets ) there exists a sequence of open sets @xmath569 such that , for all @xmath347 , we get @xmath570 and @xmath571 { } \mu^- ( a ) \ : .\ ] ] for all @xmath347 , let @xmath572 , then @xmath284 is an open set , @xmath573 and , by monotonicity , @xmath574 therefore @xmath575 { } { { \hat \mu}}^{q_+ } ( a ) \text { and } \mu^- ( u_k ) \xrightarrow[k \to \infty ] { } \mu^-(a ) \ : .\ ] ] evaluating on the sequence @xmath286 and letting @xmath347 go to @xmath207 , we eventually get @xmath576 owing to hahn decomposition of signed measures , there exists a borel set @xmath577 such that for all borel @xmath9 it holds @xmath578 finally , let @xmath9 be a borel set , then by applied to @xmath579 we get @xmath580 it remains to show that if @xmath581 , then @xmath582 .
this can be easily obtained by taking a sequence of open balls @xmath583 with fixed center and radius @xmath450 , and by considering the sequence @xmath584 for which @xmath585 and @xmath586 . by applying the same argument as before , we get @xmath587 thus the conclusion follows by taking the limit as @xmath588 .
this completes the proof of ( ii ) and thus of the theorem . if the approximate values @xmath46 are closer and closer to the actual values of @xmath589 when @xmath590 one obtains by theorem [ thm_main ] that the reconstructed measure @xmath4 coincides with @xmath1 . more precisely , we have the following stability result .
let us fix @xmath591 and @xmath400 such that @xmath592,\ { \gamma}_{n}\ge 1,\ c_{n}\ge 1,\ r_{n}>0
\\ & { \alpha}_{n},{\gamma}_{n},c_{n}\to 1\quad\text{and}\quad r_{n}\to 0\quad \text{as } n\to \infty\,.\end{aligned}\ ] ] let @xmath0 be a directionally limited metric space endowed with a sequence of asymptotically @xmath593-bounded measures @xmath594 satisfying .
let @xmath1 be a positive borel measure on @xmath3 and let @xmath2 be a premeasure defined on closed balls and satisfying @xmath595 for all @xmath38 , @xmath450 , and @xmath596
. then @xmath597 .
it is an immediate consequence of theorem [ thm_main ] .
the above corollary can be formulated for signed measures as well .
indeed under the same assumptions on @xmath3 and analogous assumptions on @xmath536 one obtains the same conclusion , i.e. the coincidence of the reconstructed signed measure with the initial measure , thanks to theorem [ thm : signedboh ] .
this work has been co - funded by gnampa ( indam ) and by palse program .
we also thank carlo benassi for fruitful discussions on the subject . | in a metric space @xmath0 we reconstruct an approximation of a borel measure @xmath1 starting from a premeasure @xmath2 defined on the collection of closed balls , and such that @xmath2 approximates the values of @xmath1 on these balls .
more precisely , under a geometric assumption on the distance ensuring a besicovitch covering property , and provided that there exists a borel measure on @xmath3 satisfying an asymptotic doubling - type condition , we show that a suitable packing construction produces a measure @xmath4 which is equivalent to @xmath1 .
moreover we show the stability of this process with respect to the accuracy of the initial approximation .
we also investigate the case of signed measures . |
divided differences can be viewed as a discrete analogue of derivatives and are commonly used in approximation theory , see @xcite for a survey .
recently , the second author and lyche established two univariate chain rules for divided differences @xcite , both of which can be viewed as analogous to fa di bruno s formula for differentiating composite functions @xcite .
one of these formulas was simultaneously discovered by wang and xu @xcite . in a follow - up preprint ,
the other chain rule was generalized to the composition of vector - valued functions of several variables @xcite , yielding a formula analogous to a multivariate version of fa di bruno s formula @xcite . in @xcite ,
the univariate chain rule was applied to find a formula for divided differences of the inverse of a function . in theorem
[ thm : main ] , the main theorem of this paper , we use the multivariate chain rule to prove a similar formula for divided differences of implicitly defined functions .
equation [ eq : inverse ] shows that the formula for divided differences of inverse functions in @xcite follows as a special case .
more precisely , let @xmath1 be a function that is defined implicitly by a function @xmath4 via @xmath5 and @xmath6 , for every @xmath2 in an open interval @xmath7 .
then the main theorem states that for any @xmath8 we can express @xmath9y$ ] as a sum of terms involving the divided differences @xmath10g$ ] , with @xmath11 . in section [ sec :
divideddifferences ] , we define these divided differences and explain our notation . in section [ sec : recursiveformula ] , we apply the multivariate chain rule to derive a formula that recursively expresses divided differences of @xmath1 in terms of divided differences of @xmath3 and lower order divided differences of @xmath1 .
finally , in section [ sec : formula ] , we solve this recursive formula to obtain a formula that expresses divided differences of @xmath1 solely in terms of divided differences of @xmath3 .
we end the section with applying the main theorem in some special cases .
let @xmath9f$ ] denote the _ divided difference _ of a function @xmath12 at the points @xmath13 , with @xmath14 .
if all inequalities are strict , this notion is recursively defined by @xmath15f : = f(x_0)$ ] and @xmath16f = \frac{[x_1,\ldots , x_n]f - [ x_0,\ldots , x_{n-1}]f}{x_n - x_0 } \qquad \textup{if}~n>0.\ ] ] if some of the @xmath17 coincide , we define @xmath18f$ ] as the limit of this formula when the distances between these @xmath17 become arbitrary small , provided @xmath19 is sufficiently smooth there . in particular ,
when @xmath20 , one can show that @xmath9f = f^{(n)}(x_0 ) / n!$ ] . for given @xmath21 satisfying @xmath22 , we shall sometimes shorten notation to @xmath23f : = [ x_{i_0},x_{i_1},\ldots x_{i_k}]f.\ ] ] the above definitions generalize to bivariate divided differences as follows .
let @xmath24 be defined on some @xmath25-dimensional interval @xmath26 suppose we are given @xmath27 and points @xmath28 satisfying @xmath29 and @xmath30 satisfying @xmath31 .
the cartesian product @xmath32 defines a rectangular grid of points in @xmath33 .
the _ ( bivariate ) divided difference _ of @xmath19 at this grid , denoted by @xmath34f,\ ] ] can be defined recursively as follows .
if @xmath35 , the grid consists of only one point @xmath36 , and we define @xmath37f : = f(x_0,y_0)$ ] as the value of @xmath19 at this point . in case @xmath38 , we can define ( [ eq : bivdd ] ) as @xmath39f - [ x_0 , \ldots , x_{m-1 } ; y_0 , \ldots , y_n ] f}{x_m - x_0},\ ] ] or if @xmath40 , as @xmath41f - [ x_0 , \ldots , x_m ; y_0 , \ldots , y_{n-1 } ] f}{y_n - y_0}.\ ] ] if both @xmath42 and @xmath43 the divided difference ( [ eq : bivdd ] ) is uniquely defined by either recursion formula . as for univariate divided differences , we can let some of the points coalesce by taking limits , as long as @xmath19 is sufficiently smooth .
in particular when @xmath44 and @xmath45 , this legitimates the notation @xmath46f : = \frac{1}{m!n ! } \frac{\partial^{m+n } f } { \partial x^m \partial y^n } ( x_0,y_0).\ ] ] similarly to equation [ eq : divdiff - short1 ] , we shall more often than not shorten the notation for bivariate divided differences to @xmath47f : = [ x_{i_0},x_{i_1},\ldots , x_{i_s};y_{j_0},y_{j_1},\ldots , y_{j_t}]f.\ ] ]
let @xmath1 be a function implicitly defined by @xmath48 as in section [ sec : introduction ] .
the first step in expressing divided differences of @xmath1 in terms of those of @xmath3 is to express those of @xmath3 in terms of those of @xmath1 .
this link is provided by a special case of the the multivariate chain rule of @xcite .
let @xmath49 be a composition of sufficiently smooth functions @xmath50 and @xmath3 . in this case
, the formula of @xcite for @xmath51 is @xmath52(g\circ \ff ) = \sum_{k=1}^n\ \sum_{0=i_0<i_1<\cdots < i_k = n}\ \sum_{s=0}^k\ ] ] @xmath53g\ ] ] @xmath54\phi \prod_{l = s+1}^k [ x_{i_{l-1}},x_{i_{l-1 } + 1 } , \ldots , x_{i_l}]\psi.\ ] ] now we choose @xmath55 to be the graph of a function @xmath1 , i.e. , @xmath56
. then the divided differences of @xmath57 of order greater than one are zero , implying that the summand is zero unless @xmath58 ; below this condition is realized by restricting the third sum in equation [ eq : smultivariatechainrule ] to integers @xmath59 that satisfy @xmath60 . since additionally divided differences of @xmath57 of order one are one , we obtain @xmath61g\big(\cdot , y(\cdot)\big ) = \sum_{k=1}^n\ \sum_{0=i_0<i_1<\cdots < i_k = n}\ \sum_{\substack{s = 0 \\s = i_s - i_0}}^k\ ] ] @xmath62 g \prod_{l = s+1}^k [ x_{i_{l-1}},x_{i_{l-1}+1},\ldots , x_{i_l}]y,\ ] ] where @xmath63 for @xmath64 . for example , when @xmath65 this formula becomes @xmath66g\big(\cdot , y(\cdot)\big ) = [ x_0 ; y_0 , y_1]g\ , [ x_0,x_1]y + [ x_0 , x_1 ; y_1]g,\ ] ] and when @xmath67 , @xmath68g\big(\cdot , y(\cdot)\big ) = & \ [ x_0;y_0,y_2]g\ , [ x_0,x_1,x_2]y \\ & + [ x_0;y_0,y_1,y_2]g\ , [ x_0,x_1]y\ , [ x_1,x_2]y \\ & + [ x_0 , x_1 ; y_1 , y_2]g\ , [ x_1,x_2]y \\ & + [ x_0 , x_1 , x_2 ; y_2]g.\\\end{aligned}\ ] ] in case @xmath1 is implicitly defined by @xmath48 , the left hand side of equation [ eq : cr1 ] is zero . in the case
@xmath65 , therefore , we see that @xmath69y = - \frac{[01;1]g}{[0;01]g},\ ] ] where we now used the shorthand notation from equations [ eq : divdiff - short1 ] and [ eq : divdiff - short2 ] . for @xmath70 , the highest order divided difference of @xmath1 present in the right hand side of equation [ eq : cr1 ] appears in the term @xmath71g\ , [ 01\cdots n]y$ ] . moving this term to the left hand side and dividing by @xmath72g$ ] ,
one finds a formula that expresses @xmath73y$ ] recursively in terms of lower order divided differences of @xmath1 and divided differences of @xmath3 , @xmath74y = - \sum_{k = 2}^n\ \sum_{0 = i_0 < \cdots <
i_k = n}\ \sum_{\substack{s = 0\\s = i_s - i_0}}^k\ ] ] @xmath75g}{[0;0n]g}\!\prod_{l = s + 1}^k [ i_{l-1 } ( i_{l-1 } + 1)\cdots i_l]y.\ ] ] we shall now simplify equation [ eq : recursionformulaa ] . by equation [ eq : recursionformulaz ] , the first order divided differences of @xmath1 appearing in the product of equation [ eq : recursionformulaa ] can be expressed as quotients of divided differences of @xmath3 .
to separate , for every sequence @xmath76 appearing in equation [ eq : recursionformulaa ] , the divided differences of @xmath3 from those of @xmath1 , we define an expression involving only divided differences of @xmath3 , @xmath77g}{[i_0 ; i_0 i_k]g } \prod_{\substack{l = s + 1\\ i_l - i_{l-1 } = 1}}^k \left ( - \frac { [ i_{l-1 } i_l ; i_l]g}{[i_{l-1 } ; i_{l-1 } i_l]g } \right).\ ] ] note that if a sequence @xmath78 starts with precisely @xmath59 consecutive integers , the expression @xmath79 will comprise @xmath59 terms . for instance , @xmath80g}{[0;03]g}\fg{2}{3 } , \\ \{013\}g = & \frac{[0;013]g}{[0;03]g}\fg{0}{1 } - \frac{[01;13]g}{[0;03]g } , \\
\{012\}g = & - \frac{[0;012]g}{[0;02]g}\fg{0}{1}\fg{1}{2 } + \frac{[01;12]g}{[0;02]g}\fg{1}{2 } - \frac{[012;2]g}{[0;02]g } .\end{aligned}\ ] ] the remaining divided differences @xmath81y$ ] in the product of equation [ eq : recursionformulaa ] are those with @xmath82 , and each of these comes after any @xmath59 satisfying @xmath60 .
we might therefore as well start the product of these remaining divided differences at @xmath83 instead of at @xmath84 , which has the advantage of making it independent of @xmath59 .
equation [ eq : recursionformulaa ] can thus be rewritten as @xmath85y = \sum_{k = 2}^n \sum_{0 = i_0 < \cdots < i_k = n } \{i_0\cdots i_k\}g \prod_{\substack{l=1\\ i_l - i_{l-1 } \geq 2}}^k [ i_{l-1}\cdots i_l]y.\ ] ] for @xmath86 this expression amounts to @xmath87y = & \ , \{012\}g,\label{eq : rec2}\\ [ 0123]y = & \ , \{0123\}g + \{023\}g\ , [ 012]y + \{013\}g\ , [ 123]y,\label{eq : rec3}\\ [ 01234]y = & \ , \{01234\}g + \{0134\}g\ , [ 123]y + \{034\}g\ , [ 0123]y \label{eq : rec4}\\ + & \ , \{0124\}g\ , [ 234]y + \{0234\}g\ , [ 012]y + \{014\}g\ , [ 1234]y\notag\\ + & \ , \{024\}g\ , [ 012]y\ , [ 234]y . \notag\end{aligned}\ ] ]
in this section we shall solve the recursive formula from equation [ eq : recursionformulab ] . repeatedly applying equation [ eq : recursionformulab ] to itself yields @xmath88y = \ , & \{012\}g,\\\label{eq : mainexamples2 } [ 0123]y = \ , & \{0123\}g + \{023\}g\,\{012\}g + \{013\}g\,\{123\}g,\\\label{eq : mainexamples3 } [ 01234]y = \ , & \{01234\}g + \{0134\}g\,\{123\}g + \{034\}g\,\{013\}g\,\{123\}g \\\notag + \ , & \{034\}g\,\{0123\}g + \{034\}g\,\{023\}g\,\{012\}g + \{0124\}g\,\{234\}g \\\notag + \ , & \{0234\}g\,\{012\}g + \{014\}g\,\{134\}g\,\{123\}g + \{014\}g\,\{1234\}g \\\notag + \ , & \{014\}g\,\{124\}g\,\{234\}g + \{024\}g\,\{012\}g\,\{234\}g.\end{aligned}\ ] ] examining these examples , one finds that each term in the right hand sides of the above formulas corresponds to a partition of a convex polygon in a manner we shall now make precise . with a sequence of labels
@xmath89 we associate the ordered vertices of a convex polygon .
a _ partition of a convex polygon _ is the result of connecting any pairs of nonadjacent vertices with straight lines , none of which intersect . we refer to these straight lines as the _ inner edges _ of the partition .
we denote the set of all such partitions of a polygon with vertices @xmath89 by @xmath90 .
every partition @xmath91 is described by its set @xmath92 of ( oriented ) faces .
each face @xmath93 is defined by some increasing sequence of vertices @xmath94 of the polygon , i.e. , @xmath95 .
we denote the set of edges in @xmath96 by @xmath97 .
let @xmath1 be a function implicitly defined by @xmath48 and @xmath98 @xmath99 be as in section [ sec : introduction ] .
equations [ eq : mainexamples1][eq : mainexamples3 ] suggest the following theorem .
[ thm : main ] for @xmath1 and @xmath3 defined as above and sufficiently smooth and for @xmath70 , @xmath100y = \sum_{\pi\in \ppp(0,\ldots , n)}\ \prod_{(v_0,\ldots , v_r)\in f(\pi ) } \{v_0\cdots v_r\}g,\ ] ] where @xmath101 is defined by equation [ eq : curlybrackets ] .
before we proceed with the proof of this theorem , we make some remarks .
for @xmath102 this theorem reduces to the statements of equations to prove theorem [ thm : main ] , our plan is to use equation [ eq : recursionformulab ] recursively to express @xmath73y$ ] solely in terms of divided differences of @xmath3 .
we have found it helpful to assign some visual meaning to equation [ eq : recursionformulab ] .
every sequence @xmath103 that appears in equation [ eq : recursionformulab ] induces a partition @xmath104 whose set of faces comprises an _ inner face _ @xmath76 and _ outer faces _
@xmath105 for every @xmath106 with @xmath107 .
we denote by @xmath108 the set of all partitions of the disjoint union of these outer faces .
an example of such a sequence @xmath109 , together with its inner face , outer faces , and partition set @xmath108 is given in figure [ fig : inner - outer ] .
we shall now associate divided differences to these geometric objects . to each outer face
@xmath105 we associate the divided difference @xmath110y$ ] , and to each inner face @xmath76 we associate the expression @xmath79 . for any sequence @xmath109 that appears in the sum of equation [ eq : recursionformulab ] , the corresponding inner face therefore represents that part of equation [ eq : recursionformulab ] that can be written solely in terms of divided differences of @xmath3 , while the outer faces represent the part that is still expressed as a divided difference of @xmath1 . , the sequence @xmath111 gives rise to the two outer faces @xmath112 and @xmath113 , which are drawn shaded in the figure .
the set @xmath114 contains in this case just @xmath115 partition , namely the union of the unique partitions @xmath116 and @xmath117 of the outer faces . ] repeatedly applying equation [ eq : recursionformulab ] yields a recursion tree , in which each node represents a product of divided difference expressions associated to inner and outer faces .
these recursion trees are depicted in figure [ fig : recursiontrees ] for @xmath118 .
equation [ eq : recursionformulab ] roughly states that the expression of any nonleaf vertex is equal to the sum of the expressions of its descendants .
, the figure depicts the recursion trees obtained by repeatedly applying equation [ eq : recursionformulab ] .
the top levels of these recursion trees correspond to equations [ eq : rec2][eq : rec4].,title="fig : " ] , the figure depicts the recursion trees obtained by repeatedly applying equation [ eq : recursionformulab ] .
the top levels of these recursion trees correspond to equations [ eq : rec2][eq : rec4].,title="fig : " ] this theorem is a generalization of theorem 1 in @xcite , and the proofs are analogous .
we prove the formula by induction on the order @xmath119 of the divided difference of @xmath1 . by the above discussion
, the formula holds for @xmath118 . for @xmath120 ,
assume the formula holds for all smaller @xmath119 .
consider the recursive formula from equation [ eq : recursionformulab ] .
for every sequence @xmath109 that appears in this equation , the corresponding outer faces have fewer vertices than the full polygon . by the induction hypothesis
, we can therefore replace each divided difference @xmath121y$ ] appearing in the product of equation [ eq : recursionformulab ] by an expression involving only divided differences of @xmath3 .
as before , let @xmath108 denote the set of all partitions of the disjoint union of the outer faces induced by @xmath109 .
then , by the induction hypothesis , the product in equation [ eq : recursionformulab ] is equal to @xmath122 for a given inner face @xmath109 , the set @xmath114 can be identified with @xmath123 by the bijection @xmath124 . substituting
the above expression into equation [ eq : recursionformulab ] then yields @xmath125y = & \sum_{\substack{\textup{inner faces}\\\ii=(i_0,\ldots , i_k ) } } \{i_0\cdots i_k\}g \sum_{\pi\in \ppp_\ii}\ \prod_{(v_0,\ldots , v_r)\in f(\pi ) } \{v_0\cdots v_r\}g \\ = & \sum_{\substack{\textup{inner faces}\\\ii=(i_0,\ldots , i_k)}}\ \sum_{\substack{\pi\in \ppp(0,\ldots , n)\\\ii\in f(\pi)}}\ \prod_{(v_0,\ldots , v_r)\in f(\pi ) } \{v_0\cdots v_r\}g \\ = & \sum_{\pi\in \ppp(0,\ldots , n)}\ \prod_{(v_0,\ldots , v_r)\in f(\pi ) } \{v_0\cdots v_r\}g.\qedhere\end{aligned}\ ] ] intuitively , this proof can be expressed in terms of the recursion tree as follows . as remarked in the previous section , equation [ eq : recursionformulab ]
states that the expression of any nonleaf vertex is equal to the sum of the expressions of its descendants . by induction , the expression @xmath73y$ ] of the root vertex
is therefore equal to the sum of the expressions of the leaves , which , by construction , correspond to partitions of the full polygon .
[ ex : maintheorem ] let us apply theorem [ thm : main ] to find a simple expression for divided differences of the function @xmath126 defined on the interval @xmath127 .
this function is implicitly defined by the polynomial @xmath128 . for any knots @xmath129
satisfying @xmath130 and corresponding function values @xmath131 , one finds @xmath132 g = x_a + x_b , \qquad [ x_a , x_b , x_c ; y_d]g = 1,\ ] ] @xmath133 g = y_b + y_c , \qquad \ [ x_a ; y_b , y_c , y_d]g = 1,\ ] ] and all other divided differences of @xmath3 of nonzero order are zero .
in particular , every divided difference of @xmath3 of total order at least three is zero , which means that the sum in equation [ eq : main ] will only be over _ triangulations _
( i.e. , partitions in which all faces are triangles ) . for a polygon with vertices @xmath89 , exercise 6.19a of @xcite states that the number of such triangulations is given by the catalan number @xmath134 consider , for a given triangulation @xmath91 , a face @xmath135 from the product in equation [ eq : main ] . as
any divided difference of the form @xmath136g$ ] is zero for this @xmath3 , equation [ eq : curlybrackets ] expresses @xmath137 as a sum of at most two terms .
there are four cases .
@xmath138 & a , b , c~\text{consecutive};\\ \displaystyle \frac { 1}{y_a + y_c}\cdot\frac{x_a + x_b}{y_a + y_b } & \text{only}~a , b~\text{consecutive};\\ \displaystyle \frac { 1}{y_a + y_c}\cdot\frac{x_b + x_c}{y_b + y_c } & \text{only}~b , c~\text{consecutive};\\ \displaystyle \frac{-1}{y_a + y_c } & \text{otherwise}. \end{array}\right.\ ] ] for example , when @xmath139 , our convex polygon is a quadrilateral , which admits @xmath140 triangulations @xmath141 and @xmath142 with sets of faces @xmath143 one finds @xmath144\sqrt{1-x^2 } = \{012\}g\ , \{023\}g + \{013\}g\,\{123\}g = \ ] ] @xmath145 \cdot \frac{x_2 + x_3}{y_2 + y_3 } + \ ] ] @xmath146 \cdot \frac{x_0 + x_1}{y_0 + y_1}.\ \!\\ ] ] next we show that theorem [ thm : main ] is a generalization of theorem 1 of @xcite , which gives a similar formula for inverse functions . to see this , we apply theorem [ thm : main ] to a function @xmath1 implicitly defined by a function @xmath147 . referring to equation [ eq : curlybrackets ]
, we need to compute @xmath148g$ ] for this choice of @xmath3 and various indices @xmath21 and @xmath149 .
applying the recursive definition of bivariate divided differences , one obtains @xmath150x \hfill = & \left\ { \begin{array}{ll } x_{i_0 } & \textup{if}~s = 0 , s = k;\\ 1 & \textup{if}~s = 1 , s = k;\\ 0 & \textup{otherwise } , \end{array } \right .
\\ [ i_0\cdots i_s;i_s\cdots i_k]h(y ) = & \left\ { \begin{array}{ll } [ i_s\cdots i_k]h & \textup{if}~s = 0;\\ 0 & \textup{otherwise}. \end{array } \right.\end{aligned}\ ] ] consider a face @xmath151 of a given partition @xmath152 in equation [ eq : main ] . since @xmath153 , the divided difference @xmath154\big(x - h(y)\big)$ ] is zero for @xmath155 . using this , equation [ eq : curlybrackets ] expresses @xmath101 as a single term @xmath156g}{[v_0;v_0v_r]g } \prod_{\substack{l=1\\v_l - v_{l-1 } = 1}}^r \left(-\frac{[v_{l-1}v_l;v_l]g}{[v_{l-1};v_{l-1}v_l]g } \right ) \\
= & -\frac{[v_0\cdots v_r]h}{[v_0v_r]h } \prod_{\substack{l=1\\v_l - v_{l-1 } = 1}}^r \frac{1}{[v_{l-1}v_l]h}.\end{aligned}\ ] ] taking the product over all faces in the partition @xmath96 , the denominators of the factors in the above equation correspond to the edges of the partition , while the numerators correspond to the faces of the partition . as there is a minus sign for each face in the partition , we arrive at the formula
@xmath157y = \sum_{\pi\in \ppp(0,\ldots , n ) } ( -1)^{\#f(\pi ) } \frac{\displaystyle \prod_{(v_0,v_1,\ldots , v_r)\in f(\pi ) } [ v_0v_1\cdots v_r]h}{\displaystyle \prod_{(v_0,v_1)\in e(\pi ) } [ v_0 v_1]h},\ ] ] which appears as equation 11 in @xcite .
note that the inverse of the algebraic function @xmath158 in example [ ex : maintheorem ] is again an algebraic function .
equation [ eq : inverse ] would therefore not have been of much help to find a simple expression for divided differences of @xmath1 .
in fact , example [ ex : maintheorem ] can be thought of as one of the simplest examples for which theorem [ thm : main ] improves on equation [ eq : inverse ] , as it concerns a polynomial @xmath3 with bidegree as low as ( 2,2 ) . in this example
we shall derive a quotient rule for divided differences .
that is , we shall find a formula that expresses divided differences of the quotient @xmath159 in terms of divided differences of @xmath160 and of @xmath161 .
let @xmath162 .
then , in equation [ eq : curlybrackets ] , @xmath163 g = \left\ { \begin{array}{rl } y_{i_s}[i_0\cdots i_s]q - [ i_0\cdots i_s]p & \textup{if}~s = k;\\ \,[i_0\cdots i_s]q & \textup{if}~s = k-1;\\ 0 & \textup{otherwise}. \end{array } \right.\ ] ] in equation [ eq : main ] , therefore , the only partitions with a nonzero contribution are those whose faces have all their vertices consecutive , except possibly the final one . in particular , the inner face with vertices @xmath164 should either be the full polygon , or should have a unique inner edge @xmath165 . by induction , it follows that the partitions with a nonzero contribution to equation [ eq : main ] are precisely those for which all inner edges end at @xmath119 .
these partitions correspond to subsets @xmath166 , including the empty set , by associating with any such @xmath167 the partition with inner edges @xmath168 .
equation [ eq : main ] becomes @xmath169\frac{p}{q } = \ , & \{0\cdots n\}g + \\\notag \ , & \sum_{r=1}^{n-2}\ \sum_{k=1}^r\ \sum_{0 = i_0 < i_1 < \cdots < i_k = r } \{r\cdots n\}g\ \prod_{j=1}^k \{i_{j-1}\cdots i_j n\}g,\end{aligned}\ ] ] where the dots represent consecutive nodes and an empty product is understood to be one .
a long but straightforward calculation involving equations [ eq : curlybrackets ] , [ eq : quotient1 ] , and [ eq : quotient2 ] yields @xmath170 \frac{p}{q } = \ , & \frac{[0\cdots n]p}{q_0 } + \\ \ , & \sum_{r=1}^n \frac{[r\cdots n]p}{q_r } \sum_{k=1}^r ( -1)^k \!\!\sum_{0 = i_0 < i_1 < \cdots < i_k = r}\ \prod_{j=1}^k \frac{[i_{j-1}\cdots i_j]q}{q_{i_{j-1}}},\end{aligned}\ ] ] where @xmath171 for @xmath172 .
alternatively , this equation can be found by applying a univariate chain rule to the composition @xmath173 , as described in section 4 of @xcite .
finally , we note that taking the limit @xmath174 in equations [ eq : recursionformulaz ] , [ eq : curlybrackets ] , [ eq : mainexamples1 ] , and [ eq : mainexamples2 ] yields @xmath175 where we introduced the shorthand @xmath176 these formulas agree with the examples given in @xcite , ( * ? ? ?
* page 153 ) and with a formula stated as equation 7 in @xcite .
we wish to thank paul kettler , whose keen eye for detail provided us with many valuable comments on a draft of this paper . | under general conditions , the equation @xmath0 implicitly defines @xmath1 locally as a function of @xmath2 . in this article
, we express divided differences of @xmath1 in terms of bivariate divided differences of @xmath3 , generalizing a recent result on divided differences of inverse functions . |
in ( * ? ? ? * th .
1 ) , the authors derived the coverage probability for the downlink transmission of a typical user in the single - tier multi - cell network by assuming that the desired channel undergoes rayleigh fading and that the base stations ( bss ) are spatially distributed according to homogeneous ppp as : @xmath1 where @xmath2 and @xmath3 are real non - negative quantities - tier systems as well @xcite . ] . here , @xmath4 is the path - loss exponent , @xmath5 is the intensity of the ppp , @xmath6 is the signal - to - interference - plus - noise ratio ( sinr ) threshold , @xmath7 is the constant transmit power , and @xmath8 is the noise variance .
the value of @xmath9 is given by the formula : @xmath10,$ ] where @xmath11 is gamma function while @xmath12 is the upper incomplete gamma function .
the @xmath13 $ ] is expectation taken with respect to interferers channel distribution @xmath14 .
the integral @xmath15 does not have closed - form expression for arbitrary values of @xmath4 .
for special cases : @xmath16 where @xmath17 is the q - function . in practice , the value of @xmath4 can range anywhere from 1.6 to 6.5 , depending on the environment .
the path - loss exponent for outdoor urban area is around 3.7 to 6.5 ; while for indoor single floor office buildings , it is around 1.6 to 3.5 ; for home environment , it is 3 ; whereas for stores , it can be from 1.8 to 2.2 @xcite .
( [ eqn : coverage ] ) can be directly evaluated by numerical integration techniques , but qualitative insight is lost in this process .
this can instead be replaced by approximations , which allows sufficiently accurate quantitative predictions , while at the same time offers qualitative insight into the relationship between various parameters .
thus , the motivation of this letter is to find simple expressions that allow us to approximate the value of coverage probability for these diverse path - loss exponents . in this letter
, we investigate a few approximations related to the integral ( [ eqn : coverage ] ) and discuss their convergence properties . while we focus on cases when @xmath18 , much of the analysis carries over to cases when @xmath19 as well .
there can be two limiting cases for the integral ( [ eqn : coverage ] ) : when @xmath20 and when @xmath21 .
although the solutions for these two cases are trivial , they have important physical significance .
physically , @xmath22 can occur when the bs intensity @xmath23 and hence the term @xmath24 in the integral can be neglected .
this is referred to as the noise - limited case .
similarly , @xmath21 can occur when @xmath25 , so that the term @xmath26 can be neglected .
this is referred to as the interference - limited case .
these cases can also be used as initial approximations when @xmath27 or @xmath28 .
for these two cases , the integral ( [ eqn : coverage ] ) assumes simple closed form solutions : @xmath29 to evaluate the integral for the case when @xmath22 , we have used the fact that @xmath30 , where @xmath31 , @xmath32 , @xmath33 ( * ? ? ?
. 3.326 , eqn .
2 , p. 337 ) .
we will use this formula frequently in the rest of the letter .
the simplicity of these expressions inspires the following simple , closed form of limiting approximation for the cases when both @xmath34 and @xmath35 are non - zero : @xmath36^{-1}. \label{eqn : limiting - approx}\ ] ] we see that ( [ eqn : limiting - approx ] ) reduces to one of the cases given in ( [ eqn : limit - a - zero ] ) or ( [ eqn : limit - b - zero ] ) as @xmath37 or @xmath38 , respectively , and is exact for @xmath39 .
the formula can be used when @xmath34 and @xmath35 are comparable to each other , but at the expense of accuracy . for the case when both @xmath34 and @xmath35 are positive
, one possible strategy of arriving at an integral approximation is to expand one of the exponential terms as @xmath41 where the lagrange form of the remainder is @xmath42 such that @xmath43 . since @xmath44
, absolute value of @xmath45 can be upper bounded by @xmath46 if we expand the term @xmath47 appearing in ( [ eqn : coverage ] ) and integrate term wise , we obtain @xmath48 where the remainder term @xmath49 .
neglecting the remainder term @xmath50 gives us our first approximation @xmath51 eq . ( [ eqn : first - approx - series ] ) reduces to ( [ eqn : limit - b - zero ] ) when @xmath52 , as such it is a refinement of ( [ eqn : limit - b - zero ] ) .
the authors in @xcite arrived at the first two terms of the series ( [ eqn : first - approx - series ] ) using integration by parts ; but they did not elaborate on its validity . if we take infinite number of terms , the series ( [ eqn : first - approx - series ] ) is not convergent for @xmath18 for arbitrary values of @xmath34 and @xmath35 ( see * appendix a * ) .
nevertheless , this form of approximation is appropriate when @xmath40 . to quantify the region
in which this approximation is valid , we first upper bound the remainder term in ( [ eqn : first - approx ] ) as @xmath53 the second inequality is from ( [ eqn : lagrange - bound ] ) where @xmath54 .
if we take @xmath55 terms of the approximating series , then for any given error tolerance @xmath56 , we require that the integral error be @xmath57 . using the upper bound for @xmath58 in this expression , we obtain the bound for @xmath35 in terms of @xmath34 as @xmath59 where @xmath60 . substituting the expressions for @xmath34 and @xmath35 in ( [ eqn : bound-1 ] ) , we get @xmath61 thus , we obtain the largest noise variance above which the error of approximation becomes unacceptable for given number of terms @xmath55 and error tolerance @xmath62 . however , it is not obvious as to what happens as @xmath55 is increased to infinity . for this
, we have the limit @xmath63 hence , for @xmath18 , we have the limit @xmath64 therefore , we have the following largest value of @xmath8 as @xmath65 below which the approximation will be valid for any given @xmath62 . in the integral ( [ eqn : coverage ] ) ,
if we consider expanding the term @xmath67 instead and perform term wise integration , we get @xmath68 where the remainder term @xmath69 . as before , if we neglect the remainder term @xmath50 , we obtain our second approximation as @xmath70 eq . ( [ eqn : second - approx - series ] ) reduces to ( [ eqn : limit - a - zero ] ) when @xmath52 , as such it is a refinement of ( [ eqn : limit - a - zero ] ) . the remainder term in ( [ eqn : second - approx ] ) can be bounded as @xmath71 the second inequality is from ( [ eqn : lagrange - bound ] ) where @xmath72 .
this form of approximation is suitable for cases when @xmath27 . to find the precise region for which this approximation is valid , consider again an error tolerance of @xmath56 and @xmath55 terms of the approximating series . since we require that the integral error be bounded by the error tolerance , @xmath57 , this leads us to a bound for @xmath35 in terms of @xmath34 as @xmath73 where @xmath74 .
substituting the expressions for @xmath34 and @xmath35 gives us the smallest value of noise variance @xmath75 below which the error becomes unacceptably large .
as in previous case , as @xmath55 tends to infinity , we have @xmath76 hence for @xmath18 , we have @xmath77 .
thus , the smallest value of @xmath8 above which the approximation will be valid for any @xmath62 is @xmath78 this result implies that the infinite series for ( [ eqn : second - approx - series ] ) is convergent when @xmath18 .
this is indeed the case ( see * appendix b * ) . here
, we would like to find an approximation for the case when @xmath34 and @xmath35 are comparable to each other , that is , when the system is neither noise limited nor interference limited .
we will try to obtain an approximation using laplace s method .
let @xmath79 .
the unique global minima of @xmath80 is at @xmath81 , which is where this method is usually applied @xcite .
if we take the first order taylor expansion of @xmath80 about @xmath81 , then it will merely result in the approximation @xmath82 .
thus , it can be advantageous to consider a point other than the global minima for the taylor expansion .
the taylor expansion of @xmath80 about @xmath83 is @xmath84 where @xmath85 .
the remainder term in lagrange form is @xmath86 where @xmath87 for some @xmath88 .
changing the variable in ( [ eqn : coverage ] ) from @xmath89 to @xmath90 , we obtain @xmath91 let @xmath92 , @xmath93 , and @xmath94 ; completing the square , we get @xmath95 now , let @xmath96 , and expanding its exponent , we have @xmath97 , where @xmath98 such that @xmath99 . its absolute value can be upper bounded as @xmath100 . thus the right integral
can be split as @xmath101 where @xmath102 and @xmath103 is the q - function of @xmath104 .
the remainder term is @xmath105 .
if we neglect this remainder term , then we obtain our required approximation as @xmath106 the expression is exact when @xmath107 .
unlike the previous two approximations , here the accuracy of approximation in general depends on the value of @xmath108 chosen .
we can bound the remainder term as @xmath109 the last step follows as such : since the function to be integrated is non - negative , @xmath110 .
when @xmath18 and @xmath111 , both @xmath32 and @xmath33 ; hence @xmath112 , and therefore , the maxima of @xmath113 occurs in negative axis .
so @xmath114 , where the last equality follows by changing the variable from @xmath90 to @xmath115 .
we have @xmath116 . when @xmath117 , @xmath118 is monotonically decreasing and its maximum value is attained at @xmath119 .
putting @xmath120 , we have @xmath121 putting @xmath122 , we obtain @xmath123 where @xmath124 and @xmath125 is the lower incomplete gamma function .
empirical considerations have led us to choose @xmath126 .
note that the form of this value is similar to the limiting approximation .
. , width=264 ] changes.,width=264 ] here we compare the four approximations to integral ( [ eqn : coverage ] ) with its value obtained via numerical integration . since both parameters @xmath34 and @xmath35 depend on @xmath127 and @xmath6 ,
we maintain them both at @xmath128 db .
we assume @xmath129 , i.e. , one bs on average per circular area of radius 500 meters .
we only vary @xmath8 .
the interferer s channel distribution @xmath14 is assumed to be exponential with mean @xmath7 .
1 plots the change in absolute error and coverage probability as snr = @xmath130 changes for @xmath131 .
for both the interference - limited and noise - limited cases , we take the summation up to four terms from ( [ eqn : first - approx - series ] ) and ( [ eqn : second - approx - series ] ) , respectively . for the interference - limited case ( [ eqn : second - approx - series ] ) , the error is near zero only for snr greater than certain value .
likewise , for the noise - limited case ( [ eqn : first - approx - series ] ) , the error is near zero only for snr smaller than some value . for intermediate case
, the laplace approximation gives an error less than 0.005 .
fig . 2 plots
the maximum absolute error for the limiting and laplace approximations . in general , the error amplitude tends to increase with @xmath4 , except for @xmath132 where the error is zero for laplace .
the error amplitude decreases as @xmath6 increases , while it remain unchanged when @xmath5 or @xmath127 is changed .
this is because when @xmath5 is increased , the system becomes interference limited at higher noise variance @xmath8 ( i.e. , at lower snr ) .
thus the coverage probability curve in fig .
[ fig : alpha3 ] shifts to the left .
this shift is quantitatively given by ( 10 ) and ( 11 ) .
however , the maximum coverage probability , given by @xmath133 , which occurs when @xmath25 ( and thus @xmath134 at high snr ) , remains unchanged .
similarly , since the transmit powers of all the bss are assumed to be equal in @xcite , changing the value of transmit power is equivalent to scaling the noise variance @xmath8 by @xmath127 in the sinr expression .
thus , the plot of the coverage probability versus the @xmath8 in db results in shifting of the curves towards the left as @xmath127 is increased , without changing its shape or amplitude .
since the coverage probability curve has only shifted and not changed its shape , the maximum error which occurs around the edge of the plateau of the coverage probability will also remain unchanged .
we have examined four different ways of approximating the coverage integral .
the limiting approximation is useful as an initial rough approximation .
we can use the approximation for interference - limited case or the noise - limited case so long as the parameters @xmath34 and @xmath35 satisfy some inequality relationship . for intermediate cases ,
we recommend the use of laplace approximation . to check the convergence of the series ( [ eqn : first - approx - series ] )
we will apply the ratio test : @xmath135 converges if @xmath136 . for our case , @xmath137 using the identity @xmath138 , we have @xmath139 using the identity : @xmath140 , we obtain @xmath141 for large @xmath89 and fixed @xmath90 , @xmath142 , so we have @xmath143 therefore , we finally have @xmath144 h.s .
dhillon , r.k .
ganti , f. baccelli , j.g .
andrews , `` modeling and analysis of @xmath157-tier downlink heterogeneous cellular networks , '' _ ieee j. sel .
areas commun . _ ,
3 , pp . 550560 , april 2012 . | this letter gives approximations to an integral appearing in the formula for downlink coverage probability of a typical user in poisson point process ( ppp ) based stochastic geometry frameworks of the form @xmath0 .
four different approximations are studied . for systems that are interference - limited or noise - limited ,
conditions are identified when the approximations are valid . for intermediate cases ,
we recommend the use of laplace approximation .
numerical results validate the accuracy of the approximations .
integral approximations , coverage probability , poisson point process |
the problem of a two - dimensional ( 2d ) bloch electron under a perpendicular magnetic field has attracted much interest for several decades , since not only it exhibits extremely rich energy band structures@xcite@xmath0@xcite but also it is closely related with various phenomena such as the quantum hall effect,@xcite the flux - state model for high-@xmath1 superconductivity,@xcite and the mean - field transition temperature of superconducting networks or josephson junction arrays.@xcite recently , by means of the remarkable advent of nanofabrication techniques , studies of finding the indication of the peculiar band structure and its effect on the transport and optical properties have been also extensively performed.@xcite@xmath0@xcite a traditional method to solve this problem is to start with energy dispersion @xmath2 without a magnetic field and then make the peierls substitution @xmath3 to construct an effective single - band hamiltonian @xmath4 , \label{eq:0}\ ] ] where @xmath5 is the momentum operator and @xmath6 is a vector potential .
an equivalent way to this method is to start directly with the tight - binding hamiltonian given by @xmath7 where @xmath8 is the hopping integral between the sites @xmath9 and @xmath10 , @xmath11 is a state of an atomiclike orbital centered at the site @xmath9 , and @xmath12 is the magnetic phase factor defined by @xmath13 @xmath14 being the magnetic flux quantum .
so far , a number of works have focused on the energy spectrum of a lattice with a square symmetry , and the magnetic subband structure of the lattice with isotropic hopping integrals between nearest - neighbor ( nn ) sites is well known as the hofstadter s butterfly.@xcite the effect of the anisotropy in the nn hopping integrals on the magnetic subband structure was also studied , and it was found that introducing the hopping anisotropy leads to the occurrence of band broadening and gap closing.@xcite@xmath0@xcite furthermore , the effect of next - nearest - neighbor ( nnn ) hopping integrals on the magnetic subband structure was also studied,@xcite and introducing the nnn hopping integrals was found to play a key role in removing the degeneracy which appears at the band center of the energy spectrum under isotropic nn hopping integrals . along with the case of the square lattice ,
magnetic subband structures of a triangular lattice have been intensively studied,@xcite@xmath0@xcite and it was found that the lattice with isotropic nn hopping integrals exhibits a recursive magnetic subband structure similarly to the case of the square lattice with isotropic nn hopping integrals.@xcite recently , the effect of the anisotropy in the nn hopping integrals on the magnetic subband structure was studied by gumbs and fekete.@xcite in studying the effect , they applied the peierls substitution for the energy dispersion given by @xmath15\right.\nonumber\\&&\left.+t_{c}\cos \left[(k_{x}-\sqrt{3}k_{y } ) \frac{a}{2}\right]\right\ } \label{eq:10}\end{aligned}\ ] ] to construct a hamiltonian matrix . and
, by diagonalizing the resultant matrix , they obtained magnetic subband structures for a few sets of ( @xmath16 ) and presented relevant arguments .
however , unfortunately , most of the results and arguments presented by them are erroneous ones that arise from a mistake in deriving the hamiltonian , as we shall show below .
we re - examine the same problem as that investigated by gumbs and fekete to put the correct formalism and numerical results , which is the first end of this paper .
the second end of this paper is to examine the effect of isotropic nnn hopping integrals on the magnetic subband structure . to our knowledge , few work has been devoted to this problem .
however , noting that the triangular lattice with nnn hopping integrals is topologically equivalent to the square lattice with nnn hopping integrals , it may be interesting to examine this problem and compare the results with the case of the square lattice .
we will show that the magnetic subband structure changes considerably depending on the strength of the isotropic nnn hopping integrals .
this paper is organized as follows : in sec .
ii , we derive a generalized harper equation that includes nnn as well as nn hopping integrals . in sec .
iii , we first point out the mistake and the misleading arguments presented in ref . 24 , and then present correct numerical results on the effects of both anisotropic nn and isotropic nnn hopping integrals on the magnetic subband structure .
we also discuss the symmetries of the magnetic subband structure obtained from the generalized harper equation . finally , sec .
iv is devoted to a brief summary .
we consider an electron on a 2d triangular lattice with a hexagonal symmetry in the presence of a uniform perpendicular magnetic field @xmath17 . under the landau gauge ,
the vector potential is given by @xmath18 and @xmath12 is given by @xmath19 where @xmath20 .
here , @xmath21 labels the lattice points , i.e. , @xmath22 , where @xmath23 , @xmath24 , and @xmath25 is the lattice constant . and , @xmath26 is the magnetic flux through the unit cell . by means of eqs .
( [ eq:1 ] ) and ( [ eq:3 ] ) , the tight - binding equation , @xmath27 , can be written as @xmath28 \nonumber \\ & & + t_{c}[e^{i\pi\phi(m+1/2)}\psi_{m+1,n-1 } + e^{-i\pi\phi(m-1/2)}\psi_{m-1,n+1 } ] \nonumber \\ & & + t_{a}'[e^{i\pi\phi(m-3/2)}\psi_{m-3,n-1 } + e^{-i\pi\phi(m+3/2)}\psi_{m+3,n+1 } ] \nonumber \\ & & + t_{c}'[e^{i\pi\phi(m+3/2)}\psi_{m+3,n-1 } + e^{-i\pi\phi(m-3/2)}\psi_{m-3,n+1 } ] \nonumber \\ & & + t_{b}'(e^{i2\pi\phi m}\psi_{m , n-2 } + e^{-i2\pi\phi m}\psi_{m , n+2 } ) , \label{eq:4}\end{aligned}\ ] ] where @xmath29 is the nn hopping integral along the @xmath30 direction , @xmath31 is the nn hopping integral along the direction which makes @xmath32 ( @xmath33 ) with the @xmath30 direction , and @xmath34 is the nnn hopping integral along the direction which makes @xmath35 ( @xmath36 , @xmath37 ) with the @xmath30 direction , respectively .
since eq .
( [ eq:4 ] ) is invariant under the transformation @xmath38 , @xmath39 can be written as @xmath40 thus , using eqs .
( [ eq:4 ] ) and ( [ eq : bc ] ) , we obtain a kind of generalized harper equation as follows : @xmath41 where @xmath42 with @xmath43 . denoting @xmath44 with relative primes @xmath45 and @xmath46 , it can be easily checked that @xmath47 where @xmath48 ( @xmath49 ) for an even ( odd ) @xmath45 .
thus , @xmath50 in eq .
( [ eq:5 ] ) satisfies the condition @xmath51 ( @xmath52 ) for an even ( odd ) @xmath45 , and the bloch condition along the @xmath30 direction can be written as @xmath53 using eqs .
( [ eq:5 ] ) and ( [ eq:8 ] ) , we obtain the eigenvalue equation , @xmath54 , where @xmath55 and @xmath56 with @xmath57 ( @xmath58 ) for an even ( odd ) @xmath45 . 2
before presenting our numerical results , we would like to point out the mistake made by gumbs and fekete in ref .
24 . in converting eq .
( [ eq:10 ] ) into eq .
( [ eq:0 ] ) [ i.e. , eq .
( 2 ) of ref .
24 ] , they made a mistake in operating @xmath59 to @xmath39 . according to them ,
the results of the operations are given by @xmath60 however , these are incorrect since the fundamental commutation relation @xmath61=i\hbar$ ] is ignored in the operation .
the correct results taking into account this relation should be expressed as@xcite @xmath62 we can easily check that the same equation as eq .
( [ eq:4 ] ) with @xmath63 can be obtained by the method of the peierls substitution if eq .
( [ eq:12 ] ) instead of eq .
( [ eq:11 ] ) is applied .
in fact , eq . ( 6 ) of ref . 24 , which was derived by means of eq .
( [ eq:11 ] ) , is a non - hermitian matrix , and thus it is unreasonable to obtain real energy eigenvalues from this matrix .
figure 1 shows the @xmath64 diagram obtained by diagonalizing eq .
( [ eq:9 ] ) for the parameters @xmath65 , @xmath66 , and @xmath67 with @xmath68 in the absence of the nnn hopping integral ( @xmath69 ) .
it is obvious that the energy spectrum is _ not _ symmetric about @xmath70 and there is _ no _ distinct array of energy eigenvalues parallel to the energy axis in the two largest energy gaps , contrary to the argument of gumbs and fekete [ see fig . 1 and the relevant arguments of ref .
in fact , fig . 1 of this paper is exactly the same as fig . 3 of ref . 21 presented by claro and wannier if @xmath71 is replaced by @xmath72 , which is quite natural because both of the two papers deal with the same problem .
the only difference between the two papers lies in choosing the @xmath30 and @xmath73 directions ; the axes taken in ref .
21 are turned by @xmath35 from those in this paper .
note that , even though the choice of the axes in ref .
21 yields a three - term difference equation [ see eq . ( 10 ) of ref .
21 ] while the choice in this paper yields a five - term difference equation in the absence of nnn hopping integrals , the resultant energy band structures are the same , as shown above .
figure 2 is the @xmath64 diagram for the parameters @xmath74 , @xmath66 , and ( a ) @xmath75 and ( b ) @xmath76 in the absence of nnn hopping integrals .
the increase of total bandwidths and the occurrence of gap closing except for a few large ones are clearly seen .
we argue that these behaviors may be generic effects of the hopping anisotropy , since the same phenomena were also observed in the square lattice.@xcite@xmath0@xcite note that fig . 3 of ref .
24 is the @xmath64 diagram plotted for the same parameters used in plotting fig .
2(a ) of this paper , from which gumbs and fekete argued that the bottom of the energy band is very flat near @xmath77 and that the spectrum is symmetric about @xmath70 .
however , it is evident that their arguments are incorrect . in regard to figs
. 2 @xmath78 4 of ref .
24 , we would like to point out that the @xmath64 diagrams for the parameters @xmath79 , @xmath80 , and @xmath81 should be exactly the same when all the values of @xmath82 in the magnetic brillouin zone ( mbz ; @xmath83 and @xmath84 ) are taken into account , which means that there is no preferable direction in the lattice plane under a uniform perpendicular magnetic field . in order to get more information on the anisotropy dependence of the subgap widths , we calculate the band structure as a function of the hopping anisotropy . note that the energy eigenvalues in the absence of a magnetic field satisfy the condition @xmath85 [ see eq .
( [ eq:10 ] ) ] and the total bandwidth increases linearly with the increase of @xmath29 .
assuming the situation also holds in the presence of a uniform magnetic field , we can introduce the rescaled energy defined by @xmath86 then , @xmath87 is always in the range of @xmath88 $ ] regardless of the strength of the hopping anisotropy , and we can examine more clearly the dependence of the subgap widths on the hopping anisotropy , in comparison with the isotropic case .
figure 3(a ) shows the @xmath29 dependence of the band structure for @xmath89 , where @xmath74 and all the values of @xmath82 in the mbz are taken into account . at isotropic case ( @xmath90 ) ,
there are five subbands and four subgaps .
however , introducing the hopping anisotropy ( @xmath91 ) changes the band structure considerably .
let us number the subgaps from the bottom of the energy spectrum .
then , in the regime of @xmath92 , the first and the second subgaps close for a small amount of anisotropy , while the third and the fourth subgaps survive even for large values of anisotropy .
however , since the widths of the third and the fourth subgaps decrease with increasing @xmath29 , we expect that these subgaps will also close in the limit of @xmath93 . and , in the regime of @xmath94 , the widths of the third and the fourth subgaps decrease with decreasing @xmath29 , as in the case of @xmath92 . meanwhile , the first and the second subgaps exhibit an interesting behavior .
as @xmath29 decreases , the width of the first ( second ) subgap decreases such that it closes at @xmath95 ( @xmath96 ) . and ,
as @xmath29 decreases further , the subgap reopens and its width increases such that the width of the first ( second ) subgap becomes the same as that of the fourth ( third ) subgap , resulting in a symmetric band structure about @xmath97 in the limit of @xmath98 . since the triangular lattice with the parameters @xmath99 is topologically equivalent to the square lattice with isotropic nn hopping integrals , it is natural of the band structure to be symmetric about @xmath97 . as an another example , we plot in fig .
3(b ) the @xmath29 dependence of the band structure for @xmath100 . at isotropic case ( @xmath90 ) , there might be six subbands and five subgaps .
however , due to the occurrence of band touching between the second and the third subbands , the second subgap is missing and the energy spectrum consists of four subgaps . and , in the regime of @xmath92 , we observe that the first subgap close even for a small amount of anisotropy and the third subgap close for a large value of anisotropy . besides ,
since the widths of the fourth and the fifth subgaps decrease with increasing @xmath29 , the closing of these subgaps is expected in the limit of @xmath93 . meanwhile , the @xmath29 dependence of the missing second subgap exhibits an interesting feature ; the subgap opens with increasing @xmath29 , even though its width is very small . in the regime of @xmath94
, we can see that the width of the first ( second ) subgap increases with decreasing @xmath29 such that the width becomes the same as that of the fifth ( fourth ) subgap , resulting in a symmetric band structure about @xmath97 in the limit of @xmath98 . and
, the width of the third subgap decreases with decreasing @xmath29 and closes at @xmath101 .
thus , there occurs a degeneracy at the band center of the energy spectrum , as in the case of the square lattice with isotropic nn hopping integrals . 2 in calculating energy eigenvalues with varying the nnn hopping integrals , we set @xmath65 and assume that the nnn hopping integrals are isotropic , i.e. , @xmath102 .
figure 4 shows the @xmath64 diagram for ( a ) @xmath103 , ( b ) @xmath104 , ( c ) @xmath105 , and ( d ) @xmath106 , respectively .
we can see that the fractal structure in the energy spectrum appears despite of introducing the nnn hopping integral , which may attribute to the isotropy of the nn and the nnn hopping integrals .
we can also see that , even though the total bandwidths for @xmath107 and @xmath108 increase monotonically with increasing @xmath109 , the bandwidths for generic values of @xmath110 depend crucially on @xmath109 , resulting in diverse shapes of @xmath64 diagrams . for small values of @xmath109 ,
the bandwidths become narrower than the case of @xmath111 and the resultant diagram exhibits a vary narrow structure [ see fig .
however , the bandwidths increase with increasing @xmath109 , resulting in a fat structure of @xmath64 diagram [ see figs .
4(b ) - 4(d ) ] .
figure 5 shows the @xmath109 dependence of the band structure for @xmath70 , which exhibits two critical values of @xmath109 .
the subgap width increases with increasing @xmath109 up to @xmath112 , and decreases up to @xmath113 such that gap closing occurs at @xmath114 . and
, then it increases again for @xmath115 .
it may be interesting to compare this behavior with the case of the square lattice , since the triangular lattice with @xmath116 is topologically equivalent to the square lattice with isotropic ( anisotropic ) nnn hopping integrals if @xmath117 ( @xmath118 ) . in the square lattice , introducing the nnn hopping integrals
removes the degeneracy appearing at @xmath97 for even values of @xmath46.@xcite however , in the triangular lattice , a degeneracy occurs at @xmath97 in the energy spectrum for @xmath70 when @xmath119 .
the @xmath109 dependence of the band structures for @xmath120 and @xmath121 also exhibit interesting features . in these cases ,
there are three subgaps at @xmath111 . and , with increasing @xmath109 , the widths of these subbands decrease up to a certain value of @xmath109 [ see fig . 4(a ) ] , and then increase such that the two subgaps close [ see figs .
4(b)-4(d ) ] for large values of @xmath109 .
particularly , the lowest ( highest ) subband edge for @xmath120 ( 2/3 ) goes downward ( upward ) with increasing @xmath109 such that @xmath122 [ @xmath123 in the limit of @xmath124 , where @xmath125 means the lowest ( highest ) subband edge . from this behavior
, it can be expected that the periodicity of the @xmath64 diagram in the limit of @xmath126 reduces by three times in @xmath110 .
figure 6 shows the @xmath64 diagrams for the lattice with @xmath127 and @xmath117 , which confirms this expectation .
note that it is nothing but an energy spectrum for the lattice with the isotropic hopping integral ( @xmath117 ) and the lattice constant @xmath128 . in order to see the @xmath109 dependence of the band structure for generic values of @xmath110 , we plot in fig . 7 the @xmath129 diagrams for ( a ) @xmath130 and ( b ) @xmath131 .
we can see that gap closing , gap reopening , and band crossing occur depending on the values of @xmath109 . before concluding this section
, we discuss the symmetry of the energy spectrum . denoting the set of energy eigenvalues for a wave vector @xmath82 under a flux
@xmath110 as @xmath132 and the energy spectrum taking into account all the val- ues of @xmath82 in the mbz as @xmath133 , we can see the following properties of the band structure for generic values of nn hopping integrals : ( i ) @xmath134 , ( ii ) @xmath135 , ( iii ) @xmath136 with @xmath137 , and ( iv ) @xmath138 . note that the property ( i ) can be easily checked in eq .
( [ eq:9 ] ) ; @xmath139 under the transformation @xmath140 . note also that the property ( ii ) means that there is no way for an electron to discern the direction of the magnetic field .
figures 1 and 2 show that the properties ( iii ) and ( iv ) hold in the presence of the nn hopping anisotropy . and , fig .
4 also shows that the properties ( i)(iv ) hold in the presence of nnn hopping integrals except for the case of @xmath127 ; when @xmath127 , the properties ( iii ) and ( iv ) are replaced by @xmath141 with @xmath142 and @xmath143 , respectively , as can be seen in fig .
in summary , the effects of anisotropic nn hopping integrals and of isotropic nnn hopping integrals on the magnetic subband structure of an electron on a triangular lattice were studied within the tight - binding approximation .
it was found that the phenomena of bandwidth broadening , gap closing , and gap opening are generic feature of introducing the nn hopping anisotropy .
it was also found that the magnetic subband structure changes considerably by introducing the nnn hopping integrals ; bandwidth broadening , gap closing , gap reopening , and band crossing occur depending on the strength of the isotropic nnn hopping integrals . in the study ,
the dependence of the subgap widths on the hopping anisotropy and the nnn hopping integrals was illustrated in detail , and the symmetries of the magnetic subband structures with and without the hopping anisotropy and the nnn hopping integrals were also discussed . in order to deepen the understanding on the band structure of the triangular lattice , a further study of the effect of , for example , nonuniform magnetic fields@xcite and an interaction between electrons@xcite on the magnetic subband structure of the lattice is required , and the work is on doing .
electronic address : [email protected] r. peierls , z. phys . * 80 * , 763 ( 1933 ) ; p. g. harper , proc .
london a * 68 * , 874 ( 1955 ) ; w. kohn , phys .
* 115 * , 1460 ( 1959 ) ; m. ya azbel , zh .
. fiz . * 46 * , 929 ( 1994 ) [ sov .
jetp * 19 * , 634 ( 1964 ) ] .
d. r. hofstadter , phys .
b * 14 * , 2239 ( 1976 ) .
g. h. wannier , phys .
status solidi b * 88 * , 757 ( 1978 ) ; g. h. wannier , g. m. obermaier , and r. ray , _ ibid .
_ * 93 * , 337 ( 1979 ) .
a. h. macdonald , phys .
b * 28 * , 6713 ( 1983 ) .
j. b. sokoloff , phys . rep . * 126 * , 189 ( 1985 ) .
chang and q. niu , phys .
b * 53 * , 7010 ( 1996 ) .
y. hatsugai , m. kohmoto , and y - s .
wu , phys .
b * 53 * , 9697 ( 1996 ) .
d. j. thouless , m. kohmoto , p. nightingale , and m. den .
nijs , phys .
. lett . * 49 * , 405 ( 1982 ) ; r. rammal , g. toulouse , m. t. jaekel , and b. i. halperin , phys . rev .
b * 27 * , 5142 ( 1983 ) .
p. w. anderson , phys .
t * 27 * , 60 ( 1989 ) ; p. lederer , d. poilblanc , and t. m. rice , phys .
. lett . * 63 * , 1519 ( 1989 ) .
s. alexander , phys .
b * 27 * , 1541 ( 1983 ) ; r. rammal , t. c. lubensky and g. toulose , _
ibid_. * 27 * , 2820 ( 1983 ) ; q. niu and f. nori , _ ibid_. * 39 * , 2134 ( 1989 ) .
b. pannetier , j. chaussy , r. rammal , and j - c .
villegier , phys .
lett . * 53 * , 1845 ( 1984 ) .
d. weiss , m. l. roukes , a. menshig , p. grambow , k. von klitzing , and g. wieman , phys .
rev . lett . * 66 * , 2790 ( 1991 ) .
a. lorke , j. p. kotthaus , and k. ploog , phys .
b * 44 * , 3447 ( 1991 ) .
r. r. gerhardts , d. weiss , and u. wulf , phys .
b * 43 * , 5192 ( 1991 ) ; r. r. gerhardts and d. pfannkuche , surf .
263 * , 324 ( 1992 ) .
m. kohmoto , phys .
11943 ( 1989 ) . y. hasegawa , y. hatsugai , m. kohmoto , and g. montambaux , phys .
b * 41 * , 9174 ( 1990 ) . s. n. sun and j. p.
ralston , phys .
b * 44 * , 13603 ( 1991 ) .
g. y. oh , j. jang , and m. h. lee , j. korean phys . soc .
* 29 * , 261 ( 1996 ) .
y. hatsugai and m. kohmoto , phys .
b * 42 * , 8282 ( 1990 ) ; j. h. han , d. j. thouless , h. hiramoto , and m. kohmoto , phys . rev .
b * 50 * , 11365 ( 1994 ) .
d. langbein , phys . rev .
* 180 * , 633 ( 1969 )
. f. h. claro and g. h. wannier , phys .
b * 19 * , 6068 ( 1979 ) .
d. j. thouless , phys . rev .
b * 28 * , 4272 ( 1983 ) .
o. khn , p. e. shlbmann , v. fessatidis , and h. l. cui , j. phys . : condens .
matter * 5 * , 8225 ( 1993 ) .
g. gumbs and p. fekete , phys .
b * 56 * , 3787 ( 1997 ) .
s. gasiorowicz , _ quantum physics _
( john wiley & sons inc . , 1974 ) ,
a. barelli , j. bellissard , and r. rammal , j. phys .
( paris ) * 51 * , 2167 ( 1990 ) ; g. y. oh and m. h. lee , phys . rev .
b * 53 * , 1225 ( 1996 ) ; q. w. shi and k. y. szeto , _ ibid . _
* 56 * , 9251 ( 1997 ) ; g. y. oh , _ ibid . _
* 60 * ( in press ) .
h. doh and s - h .
suck salk , phys .
b * 57 * , 1312 ( 1998 ) . | the effects of anisotropic nearest - neighbor ( nn ) and isotropic next - nearest - neighbor ( nnn ) hopping integrals on the magnetic subband structure of an electron on a triangular lattice are studied within the tight - binding approximation . a generalized harper equation that includes both the nn and the nnn hopping terms
is first derived .
then , the effects of anisotropic nn hopping integrals are examined and the results are compared in detail with those of an existing literature [ phys .
rev .
b * 56 * , 3787 ( 1997 ) ] .
it is found that the hopping anisotropy generically leads to the occurrence of band broadening and gap closing .
the effects of isotropic nnn hopping integrals are next examined , and it is found that introducing the nnn hopping integral changes considerably the magnetic subband structure ; band broadening , gap closing , gap reopening , and band crossing occur depending on the strength of the isotropic nnn hopping integral .
symmetries of the magnetic subband structures with and without both the hopping anisotropy and the isotropic nnn hopping integrals are also discussed . 2 |
* analytical model .
* detailed theoretical study of nonlinear magnetization dynamics is rarely to be accomplished in closed analytical form , however before we present the results of our numerical simulations we elaborate a simple toy model which allows us to include all relevant interactions and make predictions which will be shown to be qualitatively correct .
we examine the behaviour of a quasi - one - dimensional ferromagnet with lack of inversion symmetry located along @xmath7axis placed in the transverse magnetic field @xmath8 : the hamiltonian ( per a unit volume ) of such a system @xmath9 is known to incorporate the terms due to exchange @xmath10 , dzyaloshinskii - moriya @xmath11 , and zeeman @xmath12 interaction .
we have also included @xmath13 , a component which stands for easy - plane ( @xmath14plane ) anisotropy in the expression ( [ magnen ] ) . in the following
we assume the magnetisation is determined by a vector field @xmath15 , so that @xmath16 , while @xmath17 and @xmath18 are coordinates of space and time ; the parameters @xmath19 , @xmath20 , and @xmath21 represent exchange and dm couplings , and magnetic anisotropy respectively . in the absence of an external field dm favours the formation of a spiral structure along the @xmath7axis , which is incommensurate with the underlying crystalline structure , whereas the magnetic field @xmath22 , applied in the @xmath23direction , tends to align the spins in this direction .
the delicate interplay between these two interactions may lead to the formation of elongated domains of width @xmath24 ( see below in the text ) in which magnetisation is uniform .
there exists a critical field @xmath25 , where as long as @xmath26 the system is collinear , ferromagnetic .
the vector field @xmath15 allows a parametrisation with polar @xmath27 and azimuthal @xmath28 angles : @xmath29 .
there are many magnetic materials where the magnetic anisotropy is very high , so that @xmath30 @xcite .
for such systems one can use @xmath31 , where @xmath32 . if this is the case , the landau - lifshitz equation , combined with the hamiltonian in eqn.[magnen ] , leads to the following equations for the leading order terms for @xmath33 : @xmath34 @xmath35 where @xmath36 is the gyromagnetic ratio .
note that substituting ( [ ll1 ] ) into ( [ ll2 ] ) and introducing dimensionless variables @xmath37 and @xmath38 one arrives at the celebrated sine - gordon model ( see , e.g. @xcite ) : @xmath39 with the help of ( [ ll1 ] ) we can rewrite energy density ( [ magnen ] ) as follows : @xmath40+\tilde{q}\partial_{\tilde{z}}\phi+\left(1-\cos\phi\right),\ ] ] where the pitch vector @xmath41 coincides with its standard notation if we switch back to @xmath17 and @xmath18 , i.e. @xmath42 . depending on @xmath43 , stationary solutions to eqn.[sing ] are either a ferromagnetic state @xmath44 ( mod @xmath45 ) or a spiral structure @xmath46 , with @xmath47 .
note that @xmath48 is the amplitude of an elliptical integral , and @xmath49 is the modulus of the elliptic function .
it should be noted that @xmath49 can be thought of as a parameter of the theory and should be determined by minimising the total energy ( [ magnen1 ] ) of the system , resulting in @xmath50 equation ( [ min ] ) does not have any solution as long as @xmath51 , so that the ground state is ferromagnetic , whereas for @xmath52 the system is in a spiral state , i.e. @xmath53 changes by @xmath45 with period @xmath54 [ here we have used elliptical integrals of the first @xmath55 and second @xmath56 kind , respectively ] . in the latter case ,
the minimisation procedure leads to @xmath57,\ ] ] where @xmath58 . in general
, @xmath53 changes in a tiny region around @xmath59 with a characteristic lengthscale @xmath60 , so that @xmath53 can be approximated with a constant over large regions of space .
therefore , the ground state @xmath53 represents a lattice of @xmath61kinks separated by elongated regions @xmath24 ( an example is shown in fig . [ lattice ] ) . as we have just seen , under certain physically motivated conditions the landau - lifshitz equation for magnetically ordered systems can be reduced to the well - known sine - gordon problem , equation ( [ sing ] ) , where @xmath62 defines magnetic order in a structure .
all the solutions to equation ( [ sing ] ) with uniform asymptotic behaviour at @xmath63 have been analysed with the inverse scattering method @xcite .
interestingly , equation ( [ kink ] ) with nontrivial ground state configuration can not be integrated out with standard methods .
it is also reasonable to expect that the propagation of a soliton throughout the lattice of kinks differs from that with a homogeneous magnetisation @xcite .
the sine - gordon equation ( [ sing ] ) can be solved by means of bcklund auto - transformations leading in general to an @xmath49-solitonic solution @xcite .
these transformations are known to link one solution to another , often linear , equation .
it is noteworthy that the obtained solutions determine nonlinear superposition principle and in general allow to derive an infinite family of exact solutions , using straightforward algebra only .
the formation of solitons is associated with local translation of the magnetic structure by @xmath64 .
one can classify the general solution to belong to one of the following groups , discussed below .
first we consider a structure characterized by the following asymptotic behaviour : @xmath65 where @xmath66 : at @xmath67 an extra kink of the same chirality emerges and the background structure ( a lattice of @xmath61kinks ) is shifted by @xmath68 along the helical axis .
note that this shift is smaller than a period of the kink crystal @xmath24 . at @xmath69
the kink crystal grows by a period and an extra kink of the opposite chirality appears .
provided the expression ( [ bc1 ] ) takes place , one can expect the existence of an odd number of domain walls with mutually opposite chirality and an arbitrary number of breathers , which can be thought of as pulsatile soliton , i.e. a bound state of two kinks with opposite chiralities .
the magnetic structure in this case allows propagation of spin waves which does not lead to the local translation , i.e. spin waves do not affect the nontrivial background structure .
another scenario is realized for the following asymptotic behaviour : @xmath70 which allows an even number of kinks with mutually opposite chirality and an arbitrary number of breathers , while @xmath68 stands for the local translation due to soliton formation . to convince oneself that the expressions ( [ bc1 ] ) and ( [ bc2 ] ) do make physical sense one in general maps nonlinear differential equation to an equivalent system of differential equations ( in mathematics
this strategy is known as lax transformation ) . for the sine - gordon model ( [ sing ] )
one can construct a solution which corresponds to translations of the background structure ( see , e.g. , ref.@xcite ) .
when the analytical solution has been determined one can observe ( see supplementary movies ) that the formation of a soliton is associated with local translations of a lattice of @xmath61kinks which needs to be taken into account via asymptotic behavior ( [ bc1 ] ) and ( [ bc2 ] ) . in fact , a newly created soliton interacts with surrounding kinks pushing them away .
remarkably , in each case the background helical structure readjusts itself to counteract the formation of a soliton .
this causes the soliton to move through a kink crystal resulting in local translation of the spiral structure by @xmath68 in a direction opposite to that in which the soliton travels .
interestingly , the scalar field @xmath62 fluctuates when the soliton interacts with kinks of the background structure . in the vicinity of a commensurate - incommensurate phase transition ( @xmath71 )
the collision of two solitons with opposite chiralities leads to partial annihilation and local magnetisation fluctuations determined by @xmath73 .
summarising , a wide class of helical magnets with inhomogeneous ground state configuration ( spin spiral ) promotes propagation of multisolitonic structure , so that it shifts the spin spiral slightly past the original position .
this finding is as we will see in the following section , in good agreement with numerical simulations utilising an atomistic , discrete model . * numerical simulations *
meaningful information on the nature of the ground and excited states can be extracted from studying magnetisation dynamics by numerical methods .
this is typically done by finding solutions of phenomenological landau - lifshitz equation equipped with gilbert term @xcite , that stands for dissipation , and a slonczewski contribution responsible for current - driven phenomena .
both gilbert damping and spin - transfer torque have to be treated on equal footing and can be microscopically derived @xcite ; while spin - polarised electronic structure calculations , may be used to evaluate relevant parameters of an effective spin - hamiltonian @xcite . hence exchange and dm interaction
may be evaluated and used in an atomistic description of the spin - system , where also a dynamical description is possible . in the present study
, we are interested in the current induced soliton dynamics in a one - dimensional lattice . under the influence of an external magnetic field the spin structure is expected to change into kink crystal according to equation ( [ kink ] ) , as discussed above . to study the motion of the solitons in 1d structure in the presence of a spin polarised current , we have performed atomistic spin dynamics simulations based on the landau - lifshitz - gilbert ( llg ) equation with additional terms to describe the spin transfer torque effect . the investigation of dynamics of the solitons introduced in the system is performed in terms of atomistic spin dynamics method @xcite , as implemented in the uppasd package @xcite ( see method ) .
* stabilisation and soliton structure . *
using models which are defined in methods ( equations ( [ latham ] ) ( [ lali ] ) ) , we now turn to the task of describing the microscopic structure of the solitons considered earlier , as they result from numerical simulation based on an atomistic approach . in order to describe a soliton structure from the llg simulations , we consider a one - dimensional chiral magnet with 1000 magnetic moments along the @xmath7direction with periodic boundaries conditions . in the system
the heisenberg ferromagnetic exchange interactions and the dm interactions are considered between nearest neighbours only , so that in continuous limit the hamiltonian ( [ magnen ] ) is recovered . in the first calculation
the strength of the ferromagnetic exchange interaction is @xmath74 mry and the strength of the dm interaction is @xmath75 mry .
the magnetic moments are treated as classical vectors with fixed length and the magnitude 1 @xmath76 , and values of the dimensionless damping parameter were set to @xmath77 or @xmath78 . as discussed below , the choise of damping parameter influenced the results only marginally .
* schematic visualisation of the soliton ( this represents a zoomed - in picture of the system ) . *
the direction of the external magnetic field is on @xmath79-axis , the direction of the blue magnetic moments are parallel to it and the moments of the red region are antiparallel .
@xmath24 stands for the period of the soliton and @xmath80 is the centre of the period .
all the atomic magnetic moments are in @xmath81 plane ] the spin configuration in fig .
[ lattice ] is stabilised by applying an external magnetic field @xmath82 t in the @xmath23direction at zero temperature , and the period of the soliton @xmath24 is 174 nm .
the size is determined by the @xmath83 ratio and the strength of the applied external magnetic field , in full agreement with our analytical findings described in equation ( [ kink ] ) , which gives a period of 177.8 nm . in this work
the period of the soliton @xmath24 is determined by the distance between the magnetic moments with @xmath84 ( centre of the blue domains ) and @xmath85 is determined by the distance between the magnetic moments at the edges of the red region with @xmath86 ( see fig .
[ lattice ] ) . by performing atomistic spin dynamics simulations we can keep track of the characteristic size of the soliton s centre
: it is known that on approaching the commensurate - incommensurate phase transition it collapses .
the results of our simulations are shown in fig .
[ magnfield ] for two choise of damping parameter ( @xmath87 = 0.1 and @xmath87 = 0.01 ) at zero temperature . for finite temperature ( t = 2k )
, we only used one value of the damping parameter ( @xmath87 = 0.1 ) .
one can easily see that the systems with different damping are almost identical and upon increasing the external field , the size of the solitons decreases almost linearly with respect to the magnitude of the applied field . for the system which is held at finite temperature
we observe that fluctuations are more pronoucned , but it follows the tendency of the zero temperature simulations . for sufficiently large fields , the external field dominates over the internal exchange interactions and as a result , the soliton lattice breaks down and we observe a phase transition to a commensurate ferromagnetic ordering . the data in fig .
[ magnfield ] show that the soliton is stable also at elevated temperatures , albeit with slightly different characteristics compared to the soliton at t=0 k. in order to study the possibility of stabilizing solutions depicted in fig .
[ lattice ] at even higher temperatures we note that in agreement with theoretical predictions , increased values of @xmath88 and @xmath20 should not influence the size of the soliton , as long as the ratio @xmath83 ratio is fixed .
for this reason we increased the value of @xmath88 to 3.76 mry and @xmath20 to 0.16 mry , and investigated the stability as a function of applied field and temperature .
we find that the soliton is stable at temperatures at least up to 10 k , and that at this temperature it takes an applied field of @xmath89 t to destabilize the soliton .
parameters of exchange and dm interaction of this strength are not atypical @xcite , and hence this exercise shows that the soliton proposed here should be possible to study experimentally , not only at sub - kelvin temperatures . the period of spiral structure can be tuned by adjusting the external magnetic field .
furthermore , our simulations fit well with the experimental results on cr@xmath2nbs@xmath0 , where magneto - optical measurements were made in the presence of external field , perpendicular to the spiral axis @xcite .
these experiments are furthermore in line with the analytical results presented here .
* size of the centre of one soliton ( @xmath80 ) as a function of the external magnetic field .
* this calculations are performed with parameters @xmath74 mry , @xmath75 mry and @xmath90 fot two different temperatures ( 0 and 2 k ) and different values of damping ( 0.1 and 0.01 ) .
the dotted lines are used as guides for the size . ] * soliton dynamics and technology .
* soliton dynamics was studied for stable soliton configurations in the presence of a spin polarised current . in this case , the spin polarised current had a flux of angular momentum which gives rise to a net torque on the soliton that can be driven along the current direction through the spin transfer torque . in the llg equations we used a current density of @xmath91 a / m@xmath92 and the non - adiabatic parameter , which determines the strength of the non - adiabatic torque , was set to be @xmath93 . with these simulations we evaluated the speed of the solitons of fig .
[ lattice ] , as a function of time .
the relationship between the soliton velocity and time is shown in fig .
[ sigma_h ] . here , we observe that the velocity is increasing very fast during the first 0.2 ns , then it levels off up to 0.4 ns and after that is almost constant until the end of the simulation , at a time of 1 ns .
for the same set of parameters , we perform calculations by changing the non - adiabatic parameter @xmath94 ( @xmath95 , @xmath96 , @xmath97 and @xmath98 ) and investigate how the velocity is influenced by this change .
the results we extract from these calculations are presented in table [ distance ] and we observe that the soliton is moving even when @xmath99 , in other words it seems that @xmath94 is not essential for moving the soliton , although as table [ distance ] shows , the precise value o fthe velocity depends on @xmath94 .
.[distance ] velocities of the soliton with parameters @xmath74 mry , @xmath75 mry and @xmath90 by applying spin polarised current with different values of the non - adiabatic parameter @xmath94 .
the velocities are calculated for @xmath100 ns . [ cols="^,^,^",options="header " , ] we performed additional simulations for the time evolution of the spin s texture velocity , tuning the strength of the exchange interaction @xmath88 and the dzyaloshinskii - moriya interaction . in this way ,
two more systems were examined , the parameters chosen in the first set of simulations were @xmath101 mry and @xmath102 mry with @xmath83 ratio 0.025 and the spin texture is stabilised under the influence of an external magnetic field 0.2 t at zero temperature .
the parameters chosen in the second set were @xmath103 mry , @xmath104 mry and the @xmath83 ratio 0.034 , where an external magnetic field of 0.3 t is applied for stabilising the soliton at zero temperature .
the time evolution of the velocities are shown in fig .
[ sigma_h ] and in both cases the parameters of the spin polarised current are the same as above .
the trends of these two last systems are the same as in the first case ; the velocity increases very fast for 0.1 ns , then the increase is smoother and the only difference with the first case is that they reache a saturation earlier at approximately 0.5 ns . comparing the results obtained with the different choices of @xmath83 ratios we see that initially the dynamics differs a bit between the different systems .
it is observed that at the beginning , for @xmath105 ns , the velocities are different which implies that an increasing strength of the ferromagnetic nearest - neighbour interaction causes a reduced acceleration of the soliton motion .
eventually , the three different systems obtain a more steady - state motion with a constant velocity .
the observed acceleration phase of the solitons is close compared with what is normally observed in regular ferromagnetic domain wall motion @xcite .
* velocities as a function of time for three different @xmath83 ratios . *
the same parameters were used for all of the systems with @xmath106 a / m@xmath92 and @xmath107 . ]
one way of moving these kind of spin textures is to use a spin polarised current . after investigating the current - driven dynamics of isolated solitons we now turn our attention towards direct manipulation of interacting solitons . in our case after the stabilisation of two solitons at the beginning and at the end of our system ( fig .
[ two_solitons]a ) by applying a local magnetic field ( supplementary movie 1 , is an example of creating one soliton ) , two spin polarised currents are applied in the system in opposite directions and the solitons start moving towards each other . a suggestion for an experimental setup
is shown in fig .
[ experiment ] , and initially the two solitons are located in region a and region b ) . at the moment of their collision ( region c ) , the currents are removed from the system . in the absence of an external driving force ,
the repulsive interaction between the solitons dominate their dynamics and as a result the solitons move backwards , from each other , and relax after some time ( supplementary movie 2 and fig .
[ two_solitons]c ) .
if we apply larger spin currents into the system then the driving force overcomes the repulsive soliton - soliton interactions and as a result the solitons are forced to collide ( see fig . [ two_solitons ] d and e ) .
this results in an annihilation mechanism and during the annihilation process , spin waves are emitted ( supplementary movie 3 ) . at the moment of annihilation the spin currents are turned off and we find that after a while new solitons may emerge .
the emergence of new solitons is driven by the previously emitted spin waves which causes small fluctuations of the magnetisation .
since the simulations are performed at a small but finite temperature , stochastic fluctuations are also present in the system and as a result , the position of the newly created soliton / solitons is then not deterministically controllable
. however , fig .
[ two_solitons ] shows the interesting possibility to control the number of solitons in the system .
we started with the presence of two solitons ( fig .
[ two_solitons]a ) and after applying and removing the two opposite spin currents we end up with only one soliton ( fig .
[ two_solitons]e ) .
* schematic visualisation of a suggestion for applying two opposite currents in experiment . * with the presence of the grounding at the centre of the sample we are able to apply these opposite currents in order to accomplish the collision of solitons , shown in fig.4 and the supplementary videos .
regions a , b and c discussed in the text . ]
if now only one soliton is stabilised at the centre of the system ( region c , in fig.[experiment ] ) and the spin - currents are reversed , we find that a new soliton is created which results in two stable solitons . this process is shown in fig .
[ experiment2 ] and in supplementary movie 4 .
the reversed currents , with current density @xmath108 a / m^2^ and @xmath107 , are acting on the single soliton ( fig .
[ experiment2]a ) , dragging it in opposite directions ( fig .
[ experiment2]b ) and this results in the deformation of the soliton .
this is caused by the spin transfer torque imposed by the spin - current and the magnetic moments are forced to flip .
after very short time we stop the flow of the current and the system minimises its energy and in this process some of the atomic moments are flipping back .
the perturbation which is caused by the current is big enough for the system to stabilise two solitons ( fig .
[ experiment2]c ) when the system is recovering after the removal of the spin current . a suggested device for how to realize experimentally soliton collisions that are discussed around fig . [ two_solitons ] , is outlined in fig .
[ experiment ] .
other geometries and realizations of the results found in fig . [ two_solitons ] or fig
. [ experiment2 ] may be possible , but we discuss here some requirements that would be needed to realize an experiment based on the geometry of fig . [ experiment ] .
the solitons are initially located at region a and region b. the location of solitons at these regions can be realized by imposing local modifications of the materials parameters in these regions , e.g. by reducing the diameter of the nano - device in this region .
this is experimentally realized for domain walls . also , from figs .
[ magnfield ] and [ sigma_h ] we note that typical sizes and velocities of the solitons discussed here are in the range of 40 - 80 nm , respectively 2000 - 4000 m / s . a relevant time scale for a successful experimental realization of fig .
[ experiment ] would be set by the time a soliton would take before it leaves a region of space equal to its own size .
this puts a temporal limit of the electronics needed to control the spin - currents driving the colliding solitons , as discussed in fig .
[ two_solitons ] . with the size and speed mentioned above
we estimate that electronics in the frequency range 25 - 100 ghz is needed , which is within current technological abilities @xcite . for the event when one solition , located at region c , turns in to two solitons , by reversal of the current ( simulations in fig .
[ experiment2 ] ) , it is possible to ensure that the initial location of the singular soliton is at region c , by a local modification of the parameters of the nano - device , e.g. by changing the diameter , as discussed above . we have discussed here possible ways to both annihilate and create solitons with the use of spin polarised currents , in a controlled way , and we have proposed a potential devise which has technological relevance since it easily to create and destroy bits ( solitons ) .
a two state device can hence be realised , where the presence of one or two solitons in the device is simply detected by the resistance of the device .
our results are comparable with results of domain wall motion , albeit the soliton solution is shown here to be particularly useful for the device outlined in fig .
[ experiment ] .
for comparison we note that a micro magnetic study for domain wall motion by spin polarised current in a nanowire @xcite shows that the threshold current density for moving the domain wall is @xmath109a / m^2^ and in our case we are measuring the soliton velocities with current density @xmath91 a / m^2^. the domain wall in order to reach a velocity of 1400 m / s current density of @xmath110a / m^2^ is needed , on the other hand in our case with much smaller current density we reach velocities of approximately 4200 m / s .
thus , technology proposed in our work has advantage and happens because of soliton in our case propagates through kink crystal in contrast to domain walls which spreads through uniformly polarised ferromagnet . in the end of this section
we would like to point out that llg allows a certain class of analytical solutions , e.g. , propagating domain wall or walker s dynamical solution .
the latter is known to describe domain wall motion in an infinite medium that is characterized by uniaxial anisotropy @xcite .
such a solution was shown to survive under applied uniform magnetic field below a critical value which is purely determined by gilbert damping @xmath87 @xcite . the idea to use propagating domain walls for future spintronics applications looks promising and has been recently addressed from different perspective , e.g. , spin waves emission induced domain wall propagation in a magnetic nanowire with easy axis along the wire was studied for a finite transverse magnetic anisotropy @xcite . in the meanwhile , it turns out that walker s solution is not stable with respect to spin waves emission @xcite , making numerical simulations indispensable in studying soliton propagation .
interestingly , spin waves emitted by one domain wall traveling through another domain wall can lead to attraction between them via spin transfer @xcite . in line with our results
the authors @xcite showed that dm exerts an extra torque which rotates domain wall plane when spin waves pass through it , whereas the presence of the effective easy plane anisotropy suppresses this rotation resulting in a faster domain wall motion .
there basically exist two classes of independent solutions of the sine - gordon equation , describing nonlinear excitations of spiral magnetic structure .
they are stable solitons ( extra kinks or breathers ) and spin - wave pulses . in the former case ,
the individual motion of a soliton is always associated with a local translation of a magnetic spiral , as one can observe from our numerical simulations . a direct analytical relation between the soliton s parameters , and local translation ,
can be used to control solitons or for logical switch operation . in principle
, macroscopic translation of solitons in spiral structures might be visualised with magneto - optical measurements or magnetic force microscopy . in the proximity of a commensurate - incommensurate transition
, the solitonic motion can be easily controlled by tuning temperature or by an applied external magnetic field , which is perpendicular to the helical axis .
in addition , the presence of breathers can be investigated by applying microwave pulses corresponding to eigenfrequencies of the breather . the second class of nonlinear excitations is comprised of spin - wave pulses .
the formation of such pulses does not lead to macroscopic translation of the background structure .
in fact , magnetisation in propagating spin waves fluctuates around a nonuniform spiral state , which corresponds to an equilibrium magnetic structure . in this case
the interaction is not sufficient to stabilise the pulses , and pulses that initially are localised , spread in time thanks to dispersion .
small - amplitude pulses can be approximated with a superposition of linear spin waves corresponding to helical structure . by combining analytical and numerical work
, we have established stability criteria for these magnetic solitons . the parameter space relevant for the soliton solutions is , rewardingly , the same in the analytical and numerical work , and is an internal test of the robustness of our theory .
we have in addition shown that it is possible to annihilate and create solitons by means of a spin - polarised current , which couples to the magnetic sub - system via a spin - transfer torque .
concrete suggestions of magnetic devices have been proposed , and it is demonstrated that with experimentally realistic currents , two state solutions can be induced on a time - scale of 325 ps for the annihilation and of 255 ps for the creation . this creates a basis for new technologies in magnetic storage as well as for logical operations using magnetism .
our work hence establishes magnetic solitons as useful technology enablers , and as proposed here , can be useful in the emerging field of solitonics .
* model hamiltonian . *
the considered one - dimensional lattice is described by the following atomistic hamiltonian @xmath111 where @xmath112 and @xmath113 are atomic indices , @xmath114 is the classical atomic moment , @xmath115 is the strength of the exchange interaction , @xmath116 is the anisotropic dm interaction due to spin - orbit coupling , and the last term stands for the external applied magnetic field .
it should be noted at this point that the anisotropy term is neglected in the hamiltonian , due to the fact that this term is the less important compared with the other three terms ( exchange interaction , dm interaction and zeeman term ) . in the following we use model parameters .
* atomistic spin dynamics . * to study the motion of the solitons in 1d structure in the presence of a spin polarised current ,
we have performed atomistic spin dynamics simulations based on the landau - lifshitz - gilbert ( llg ) equation with additional terms to describe the spin transfer torque effect .
the resulting equation of motion becomes @xmath117 where @xmath87 is the gilbert damping coefficient , @xmath36 the gyromagnetic ratio , @xmath114 the magnitude of the magnetic moment , @xmath118 the effective field resulting from the spin - hamiltonian in equation ( [ latham ] ) , @xmath94 the so - called non - adiabatic parameter and @xmath119 the velocity term .
the resulting effective field is given by : @xmath120 where @xmath121 and @xmath122 refers to the hamiltonian ( [ latham ] ) .
the temperature effects on the spin system are taken into account through langevin dynamics ; the thermal fluctuations are represented here by @xmath123 which can be modelled by gaussian correlators @xmath124 where @xmath125 , @xmath126 are the indexes of cartesian coordinates .
the velocity term is in units of velocity ( m / s ) and is related to the magnitude and the direction of the current @xmath127 , is proportional to the applied current and is equal to : @xmath128 where @xmath113 is the current density , @xmath129 the spin polarisation of the current and @xmath130 the saturation magnetisation per unit cell .
we acknowledge support from the kaw foundation ( projects kaw 2013.0020 and kaw 2012.0031 ) as well as the swedish research council ( vr ) and essence .
calculations performed with support from snic / snac .
d.y acknowledges the support from rfbr project 16- 32 - 60040 .
the authors acknowledge useful comments from the two referees of this paper .
d.y designed the theoretical model and performed analytical , theoretical analysis .
k.k . performed the atomistic spin dynamics calculations .
, d.y . , o.e . and k.k .
analysed the data .
k.k . and d.y .
co - wrote the manuscript .
all authors discussed the data and commented on the manuscript .
the authors declare that they have no competing financial interests .
kishine , j. , bostrem , i. g. , ovchinnikov , a. s. & sinitsyn , v. e. coherent sliding dynamics and spin motive force driven by crossed magnetic fields in a chiral helimagnet .
b _ * 86 * , 214426 ( 2012 ) .
borisov , a. b. , kishine , j. , bostrem , i. g. & ovchinnikov , a. s. magnetic soliton transport over topological spin texture in chiral helimagnet with strong easy - plane anisotropy , _ phys . rev .
b _ * 79 * , 134436 ( 2009 ) .
liechtenstein , a. i. , katsnelson , m. i. , antropov , v. p. & gubanov , v. a. local spin density functional approach to the theory of exchange interactions in ferromagnetic metals and alloys .
_ j. magn .
mater . _ * 67 * , 65 ( 1987 ) . | in magnetic materials a variety of non - collinear ground state configurations may emerge as a result of competition among exchange , anisotropy , and dipole - dipole interaction , yielding magnetic states far more complex than those of homogenous ferromagnets .
of particular interest in this study are particle - like configurations .
these particle - like states , e.g. , magnetic solitons , skyrmions , or domain walls , form a spatially localised clot of magnetic energy . in this paper we address topologically protected magnetic solitons and explore concepts that potentially might be relevant for logical operations and/or information storage in the rapidly advancing filed of solitonics ( and skyrmionics ) .
an ability to easily create , address , and manipulate such structures is among the prerequisite forming a basis of `` -onics technology '' , and is investigated in detail here using numerical and analytical tools .
low - dimensional magnetic systems remain a challenge , representing an exciting part of modern condensed matter physics .
tremendous progress in fabrication and characterization of magnetic nanostructures @xcite allows to approach experimental investigations of nonlinear phenomena , which previously were considered to form a purely theoretical research area . from both fundamental scientific and technological perspectives , consistent studies of controllable motion of spin textures , that emerge in magnetic materials @xcite
, pose many of the most interesting scientific questions . depending on distance between neighbour spins , hybridization effects and details of the electronic structure as well as the crystalline symmetry
, various magnetic configurations may emerge .
these configurations range from collinear ferromagnetic and antiferromagnetic states , to non - collinear helimagnets and more complicated textures . in chiral magnets , where magnetic spiral ordering with a given sense of rotation is present , a non - collinear configuration may result from the competition between symmetric heisenberg exchange and anisotropic antisymmetric dzyaloshinskii - moriya ( dm ) coupling .
indeed , magnetocrystalline structures which support spiral ground states are characterized by the lack of inversion symmetry and , as a consequence , antisymmetric exchange dm interaction .
the latter takes place e.g. in metallic mnsi @xcite , insulating cub@xmath0o@xmath1 @xcite , and semiconducting cr@xmath2nbs@xmath0 @xcite . of particular interest for magnetic nano - objects
are magnetic structures that are localized in space , i.e. solitons .
an example of this is the domain wall , where a lot of knowledge has been obtained , both as concerns static @xcite and dynamic @xcite properties .
other magnetic solitons , like chiral spin - spirals and skyrmions @xcite have remained less explored , so far .
interestingly , the balance between heisenberg exchange and dm interaction allows for nontrivial topological spin - textures to exist as a ground state configuration of ferromagnets , that may develop towards two - dimensional chiral modulations , i.e. skyrmions .
the skyrmion spin - configuration , with a fixed azimuthal angle , influenced by the underlying crystalline structure , might be stabilized in a non - centrosymmetric magnet @xcite .
such a topological soliton may be stable both physically and topologically , i.e. it has a smooth defect - free core with a definite diameter that is characterized by material parameters only .
recently , the first experimental observation of a skyrmionic texture in a nano - sized , thin film structure were reported @xcite , for fege and fe@xmath3co@xmath3si .
experiments reported a stable skyrmion phase in a wide temperature window , well below the curie point , by means of lorentz transmission electron microscopy . in another , recent , work a soliton phase was reported , experimentally , in an antiferromagnetic chain compound , kcugaf@xmath4 @xcite .
in general , the class of quasi - one - dimensional magnetics includes a variety of systems ranging from metal - organic complexes , consisting of weakly coupled spin chains , to helimagnets , in which magnetisation varies in one spatial dimension . from practical aspects ,
the most interesting class of compounds is the one where there exists a possibility to manipulate non - trivial topological structures by means of an external field or current .
furthermore , chiral magnets attract a lot of attention not only for being interesting from a fundamental viewpoint but also as a promising candidate for future spintronic devices .
as mentioned , the compound cr@xmath2nbs@xmath0 @xcite is of interest , since a chiral soliton lattice has been proposed for its magnetic structure @xcite .
its hexagonal structure is built up from nbs@xmath0 layers intercalated by cr ions .
there are three basic magnetic interactions present in this structure : the ferromagnetic exchange within the cr layers @xmath5 , another ferromagnetic interaction between such cr layers , @xmath6 , and the dm interaction between the cr ions .
the latter two interactions belong to the two intercalating layers separated by nbs@xmath0 @xcite , and the competition between them forms a helical structure . there are two ways to drive the motion of a chiral soliton lattice : either by injecting a spin - polarised current into a sample ( incoherent method ) , or by twisting the phase angle of the magnetic order parameter which directly couples to a magnetic field ( coherent method ) .
for example , one may envision the latter scenario using a rotating magnetic field , that is applied to one end of the sample , and that is sufficiently strong to orient the magnetisation parallel to it . because of the stiffness of the spin system , the spin rotation at one end is transmitted to the other end of the sample , even if it is not subject to the direct effect of the applied rotating magnetic field .
transmission of the torque through the sample hence presents a spin current . the incoherent current injection method to drive the sliding dynamics
has already been proposed @xcite , whereas chiral helimagnets are promising candidates to realise the coherent method @xcite .
it has been pointed out @xcite that once a sliding motion of helimagnets is triggered , the soliton lattice maintains its persistent motion assisted by a generation of inertial mass .
another observable consequence of the coherent motion is the appearance of a spin motive force @xcite , when the time dependence of the longitudinal magnetic field manifests itself in the temporal regime of the spin motive force . remarkably , the chiral soliton lattice is a macroscopically ordered object , which contains macroscopic amounts of magnetic solitons ( kinks ) . due to the very large number of solitons , the spin motive force of this lattice
is expected to be strongly amplified as compared with that caused by a single magnetic domain wall in a ferromagnet @xcite . in this paper
we present a detailed and systematic analysis of a quasi - one - dimensional chiral soliton lattice , e.g. as has been recently observed in cr@xmath2nbs@xmath0 @xcite , and the soliton - soliton interaction .
we combine our numerical findings , obtained from atomistic spin dynamics simulations , with analytical results , performed for a one - dimensional helical structure which is known to support a soliton lattice as a ground state , when an external magnetic field perpendicular to helical axis is applied .
interestingly , in contrast to a uniformly polarized ground state configuration the formation and propagation of solitons on top of a kink crystal leads to a macroscopic translation of magnetic texture which should be experimentally detectable .
numerical results are found to be in good agreement with the analytical theory and this allows us to touch upon nonlinear dynamics of solitons and spin waves in spin - spiral structures , opening thus a way for solitonics . |
noise is usually considered as having a corrupting effect on meaningful signals .
there is however one well known counter example to this widespread belief ; the _ stochastic resonance _ ( sr ) phenomenon . in this case ,
addition of a random interference signal ( noise ) to a weak , subthreshold stimuli , may enhance its detection .
originally introduced in the framework of physics @xcite , stochastic resonance is now known to take place in several sensory systems from the cricket cercal sensory system @xcite , to crayfish mechanoreceptors @xcite , the somatosensory cortex of the cat @xcite and the human visual cortex @xcite in order to facilitate detection of weak signals ( for a review see @xcite ) .
in addition there is evidence that stochastic resonance may be used not only at the level of sensory processing , but also in central cognitive processes . in a psychophysical study by usher and feingold ,
the memory retrieval of arithmetical multiplication was found to be enhanced by the addition of noise @xcite . in this article
, we feed a simplified model of a cortical column with two time - varying input signals .
we first show that for a broad set of computation based on those input signals , addition of a small amount of noise enhances the computational power of the system .
the location where the noise strength is optimal lies approximately where the network reacts the strongest to a change in the mean input ( maximum susceptibility ) .
we then set the connectivity to zero ( with appropriate scaling of the statistics of the input to the neurons ) .
although the simplest task ( addition of both input signals ) can be achieved with a similar accuracy to that obtained with a connected network , a multiplicative task can only be solved if the population of neurons is connected .
the stochastic resonance effect is thus seen to take place at the _ system level _
@xcite rather than at the single cell level .
it is an emergent property of the neuronal assemblies .
we consider in our simulations networks of @xmath0 cells ( leaky integrate - and - fire neurons ) . the connectivity matrix is fixed and every neuron receives input from @xmath1 excitatory and @xmath2 inhibitory presynaptic neurons randomly chosen among the @xmath3 neurons in the network .
the system is made out of @xmath4 excitatory and @xmath5 inhibitory neurons , reflecting the ratio of pyramidal cells to interneurons in cortical tissue .
this excess of excitatory neurons is approximately balanced by the greater efficacy of synaptic transmission for inhibition ; in our model six times bigger than for excitation : @xmath6mv and @xmath7mv .
this approximate balance between excitation and inhibition is thought to take place at a functional level in cortical areas @xcite .
sparsely - connected networks of spiking neurons of this type have been fully described in terms of their dynamical behaviour @xcite .
the dynamics of the leaky integrate - and - fire neurons is described by the following equation : @xmath8 where @xmath9 describes the membrane potential of neuron @xmath10 with respect to its equilibrium value @xmath11 , @xmath12ms corresponds to its effective membrane time constant and @xmath13 is the effective input resistance of the neuron stimulated by a total input @xmath14 .
we add to this equation a threshold condition ; if the membrane potential of the neuron exceeds the critical value @xmath15mv , a spike is emitted and after a refractory period @xmath16ms , the integration start again from the reset potential @xmath17mv . every time the neuron @xmath10 receives an action potential from a presynaptic pyramidal cell @xmath18 ( resp .
interneuron ) , its membrane potential is depolarized according to the value of the synaptic efficacy for excitation @xmath19 ( resp . for inhibition
, the neuron is hyperpolarized by an amount @xmath20 ) .
the input a neuron @xmath10 will get from within the network can thus be written as : @xmath21 where @xmath22 is the ensemble of presynaptic neurons , @xmath23 the time neuron @xmath18 fires its @xmath24th spike and @xmath25ms is a short transmission delay .
we can now decompose the external stimulation @xmath26 into two contributions .
first , we model all noise sources , ranging from synaptic bombardment from neurons outside the network to different sources of noise diversely located at the level of synaptic transmission , channel gating , ion concentrations , membrane conductance to name but a few .
all these noise sources are grouped into a term @xmath27 with a mean depolarisation @xmath28 and a white noise component defined by its standard deviation @xmath29 .
we can define the _ susceptibility _ @xmath30 of the network as the sensitivity of the population spiking rate @xmath31 upon a change in the mean depolarisation @xmath28 : @xmath32 .
we constrained our simulation space to small mean depolarisation @xmath28 so that in absence of the noisy component @xmath33 , the network exhibits no spiking activity ( subthreshold regime ) . and @xmath34 .
in addition two randomly generated subpopulations , each composed of @xmath5 of the total number of neurons , receive test inputs @xmath35 and @xmath36 .
the readout @xmath37 _ sees _ all neurons in the network and is trained to match a function of the test input @xmath38 and @xmath39 ( see the methods section for details ) .
, width=309 ] then , we inject two test signals @xmath35 and @xmath36 , each to @xmath5 randomly chosen neurons in the network .
both inputs share the same statistical properties ; they are constant over a time interval @xmath40ms and then they switch to new randomly chosen values and remain constant for the next time interval @xmath41 . at each transition time , the new values are chosen uniformly over the interval @xmath42 $ ] .
we want to know how performance in a series of computational tasks based on both test signals @xmath35 and @xmath36 are affected by the noise level .
we adapt a paradigm introduced in the framework of liquid state machines @xcite or echo state networks @xcite : we consider a readout with a dynamics @xmath43 where the sums run over all firing times @xmath44 of all neurons in the network .
@xmath45ms is a short synaptic time constant .
the @xmath46 free parameters @xmath47 ( @xmath48 ) are chosen so as to minimize the signal reconstruction error between the readout and the target : @xmath49 ^ 2\rangle$ ] . in our simulations we fixed @xmath50 so that the transient period after a transition has vanished .
the set of functions @xmath51 that are under consideration in the present study are ; the addition @xmath52 , the multiplication @xmath53 which plays a crucial role in the transformation of object locations from retinal to body - centered coordinates @xcite and two polynomials of degree two @xmath54 and @xmath55 , which can be seen as the nonlinear xor paradigm @xcite .
parameters were optimized using a first simulation ( learning set ) lasting 100 seconds ( 100000 time steps of simulation ) and were kept fixed afterwards .
the performance measurements reported in this paper are then evaluated on a second simulation of 100 seconds ( test set ) .
we compare our results to a simple two parameter readout .
such a readout only adjusts to the mean of the time series @xmath56 .
the reconstruction error for such a readout equals the variance @xmath57 of the time series @xmath56 .
the performance with the full readout are therefore expressed as a gain ( in percent ) over the trivial prediction : @xmath58 , where @xmath59 is the error introduced above . in the simulation of fig.[control ] , all connections in the network were removed . in order to compensate the loss of input @xmath60 , we changed the background input @xmath27 so that the neurons receive an input with the same statistical properties ( mean and variance ) .
the adapted input in the network with no connection ( nc ) is thus : @xmath61 for the mean and @xmath62 for the variance term .
@xmath31 is the mean population rate in the connected network ( see @xcite ) .
simulation results were obtained using the simulation software nest .
we want to know how noise influences the ability of a recurrent network of spiking neurons to process information and to perform a series of computational tasks .
in particular we are interested in effects related to stochastic resonance .
we therefore relate the optimal noise level ( where the system has a maximum performance ) to dynamical properties of the network . from that perspective , we analyze how performance in a series of computational tasks based on both test signals @xmath35 and @xmath36 are affected by the noise level . in a first series of simulations , we measure the gain over the trivial prediction for three different functions of the test signals ; @xmath52 , @xmath63 and @xmath64 ( see figure [ gradient ] , respectively top right , bottom left and bottom right graphs ) .
note that the function @xmath64 can be seen as an implementation of the xor task , which can not be solved by a single layer neural network ( perceptron ) @xcite . in all three tasks ,
the network exhibits stochastic resonance ; for a given mean depolarisation @xmath28 , there is a non - monotonic dependence upon the noise level .
the maximum gain compared to trivial prediction reaches @xmath65 for the additive task and @xmath66 for polynomials of degree two .
a comparison to the map of the susceptibility of the network @xmath67 ( see top left graph of figure [ gradient ] ) indicated that the tasks are solved best when the sensitivity of the network upon changes in the mean input is highest .
addition of noise both increases the susceptibility of the network and the capacity of performing complex computation based on sparse inputs . in ( hz / pa ) of the network as a function of the statistical properties of the external drive ( mean and variance ) .
a high _ susceptibility _ ( light colors ) defines a network that is highly sensitive to a change in the mean of the drive .
top right and bottom : gain ( in percent ) over the trivial prediction as a function of the mean and variance of the external drive ; for the simple additive task @xmath68 ( top right ) , for a first polynomial of degree two @xmath69 ( bottom left ) and a second polynomial of degree two @xmath70 ( bottom right ) .
level curves are shown for the sake of clarity at @xmath71 and @xmath72 for the polynomials and at @xmath73 and @xmath74 for the addition .
the system exhibits stochastic resonance for all three different tasks .
in addition , the location where the performance peak approximately corresponds to the zone of high _ susceptibility_.,width=415 ] in order to know whether stochastic resonance displayed in the network is an emergent property of the population of neuron @xcite or a single cell effect , we removed all connections within the network . in this second series of simulations ,
we compare the performance achieved in networks with connectivity to networks with no connectivity .
the latter networks receive adapted version of the mean input @xmath75 and of the noisy input @xmath76 so that every neurons is stimulated with a mean and variance equivalent to that of the connected network ( see methods ) . in the first task ; the addition @xmath52 , the performance with and without connectivity are similar ( see figure [ control ] left ) .
since the readout unit performs a weighted _ sum _ that runs over all neurons , it can capture the essence of this simple computation by summing the averaged response of the groups of neurons receiving input @xmath38 with the averaged response of those receiving @xmath39 .
recurrent loops play here no significative role . in the second task
we train the network to perform the multiplication of the two test signals @xmath53 .
this arithmetical operation is thought to be essential to the brain in order to do coordinate transformation @xcite .
the computation of a multiplication by a recurrent neural network was shown to be achievable in a model of the parietal cortex @xcite . in this multiplicative task @xmath53
, the complex recurrent network outperforms the network with no connectivity ( see figure [ control ] right ) .
in fact , in absence of connections within the population of neurons , multiplication can not be solved by the simple addition of noise .
stochastic resonance displayed in the multiplicative task therefore takes place at a _ system _
level rather than at the level of single neurons . ; for the connected network ( solid ) and for the control network , when the connectivity is set to zero ( dashed ) .
the unconnected collection of neurons is capable of solving this simple task with a similar accuracy as the randomly connected neural network .
right : gain ( in percent ) over the trivial prediction for the multiplicative task @xmath78 ; for the reference network ( solid ) and when the connectivity is set to zero ( dashed ) . in absence of recurrence
, the network is no longer able to sustain complex computations .
stochastic resonance is thus a _ population - based _
effect rather than a _ single - cell _ phenomenon.,width=309 ]
complex networks of neurons fall in the class of non linear systems with a threshold ; systems that are known to exhibit stochastic resonance . from the experimental side , evidences have shown that the phenomenon helps in detecting sensory signals of small amplitude , and furthermore to favor high level cognitive processes such as arithmetical calculations .
our model has revealed the presence of stochastic resonance in a series of neural - based computation . whereas simple additive transformations of input signals can be solved by a collection of independent neurons ,
more complex computations need the massive recurrence typically observed in cortical tissue .
such complex tasks include the xor problem , a nonlinear benchmark test , and the arithmetical multiplication the brain is likely to use in order to achieve coordinate transformation .
stochastic resonance displayed is then an emergent property of the brain microcircuitry . | varied sensory systems use noise in order to enhance detection of weak signals .
it has been conjectured in the literature that this effect , known as _ stochastic resonance _ , may take place in central cognitive processes such as the memory retrieval of arithmetical multiplication .
we show in a simplified model of cortical tissue , that complex arithmetical calculations can be carried out and are enhanced in the presence of a stochastic background .
the performance is shown to be positively correlated to the susceptibility of the network , defined as its sensitivity to a variation of the mean of its inputs .
for nontrivial arithmetic tasks such as multiplication , stochastic resonance is an emergent property of the microcircuitry of the model network .
_ keywords : _ stochastic resonance , information processing , recurrent neural network + + _ contact author : [email protected]_ |
the most popular method for differential gene expression detection in two - sample microarray studies is to compute the t - statistics .
the differentially expressed genes are those whose t - statistics exceed a certain threshold .
recently , due to the realization that in many cancer studies , many genes show increased expressions in disease samples , but only for a small number of those samples .
the study of @xcite shows that t - statistics has low power in this case , and they introduced the so - called cancer outlier profile analysis " ( copa ) .
their study shows clearly that copa can perform better than the traditional t - statistics for cancer microarray data sets .
more recently , several progresses have been made in this direction with the aim to design better statistics to account for the heterogeneous activation pattern of the cancer genes . in @xcite
, the authors introduced a new statistics , which they called outlier sum .
later , @xcite proposed outlier robust t - statistics ( ort ) and showed it usually outperformed the previously proposed ones in both simulation study and application to real data set . in this paper
, we propose another statistics for the detection of cancer differential gene expression which have similar power to ort when the number of activated samples are very small , but perform betters when more samples are differentially expressed .
we call our new method the maximum ordered subset t - statistics ( most ) . through simulation studies we found the new statistics outperformed the previously proposed ones under some circumstances and never significantly worse in all situations .
thus we think it is a valuable addition to the dictionary of cancer outlier expression detection .
we consider the simple 2-class microarray data for detecting cancer genes .
we assume there are @xmath0 normal samples and @xmath1 cancer samples .
the gene expressions for normal samples are denoted by @xmath2 for genes @xmath3 and samples @xmath4 , while @xmath5 denote the expressions for cancer samples with @xmath3 and @xmath6 . in this paper , we are only interested in one - sided test where the activated genes from cancer samples have a higher expression level .
the extension to two - sided test is straightforward .
the usual t - statistics ( up to a multiplication factor independent of genes ) for two - sample test of differences in means is defined for each gene @xmath7 by @xmath8 where @xmath9 is the average expression of gene @xmath7 in normal samples , @xmath10 is the average expression of gene @xmath7 in cancer samples , and @xmath11 is the usual pooled standard deviation estimate @xmath12 the t - statistics is powerful when the alternative distribution is such that @xmath13 all come from a distribution with a higher mean . @xcite
argues that for most cancer types , heterogeneous activation patterns make t - statistics inefficient for detecting those expression profiles .
they defined the copa statistics @xmath14 where @xmath15 is the @xmath16th percentile of the data , @xmath17 is the median of the pooled samples for gene @xmath7 , and @xmath18 is the median absolute deviation of the pooled samples .
the choice of @xmath16 in ( [ copa ] ) depends on the subjective judgement of the user .
the use of @xmath19 and @xmath20 to replace the mean and the standard deviation in ( [ t ] ) is due to robustness considerations since it is already known that some of the genes are differentially expressed . in ( [ copa ] )
, only one value of @xmath21 is used in the computation .
a more efficient strategy would be to use additional expression values .
let @xmath22 be the outliers from the cancer samples for gene @xmath7 , where @xmath23 is the interquartile range of the data .
the os statistics from @xcite is then defined as @xmath24 more recently , @xcite studied ort statistics , which is similar to os statistics .
the important difference that makes ort superior is that outliers are defined relative to the normal sample instead of the pooled sample .
so in their definition , @xmath25 by similar reasoning @xmath19 in os is replaced by @xmath26 and @xmath27 by @xmath28 , where @xmath26 and @xmath29 are the medians of normal and cancer samples respectively . in both os and ort statistics ,
the outliers are defined somewhat arbitrarily with no convincing reasons . to address this question ,
we propose the following statistics that implicitly considers all possible values for outlier thresholds .
suppose for notational simplicity that @xmath30 are ordered for each @xmath7 : @xmath31 if the number of samples where oncogenes are activated were known , we would naturally define the statistics as @xmath32 when @xmath33 is not known to us , one would be tempted to define @xmath34 but this does not quite work since obviously @xmath35 for different values of @xmath33 are not directly comparable under the null distribution that @xmath36 . for example , when @xmath37 , we have @xmath38>0 $ ] while @xmath39=0 $ ] .
this observation motivates us to normalize @xmath35 such that each approximately has mean @xmath40 and variance @xmath41 .
this can be achieved by defining @xmath42 $ ] and @xmath43 where @xmath44 is the order statistics of @xmath1 samples generated from the standard normal distribution
. then we can define @xmath35 as : @xmath45 so that @xmath35 has mean and variance approximately equal to @xmath40 and @xmath41 respectively .
finally we can define our new statistics ( called most ) as @xmath34 with most , we practically consider every possible threshold above which @xmath5 are taken to be outliers . in this formulation , the number of outliers is implicitly defined as @xmath46 [ cols="^,^ " , ]
some simulations are carried out to study most , and compare its performance to os , ort , copa , and t - statistics .
for copa , we choose to use the @xmath47th percentile in its definition as in @xcite .
we generate the expression data from standard normal with @xmath48 . for various values @xmath49 , which is the number of differentially expressed cancer samples
, a constant @xmath50 is added for differentially expressed genes .
we simulated @xmath51 differentially and non - differentially expressed genes , and calculated the roc curves from them by choosing different thresholds for gene calls .
figure [ roc1 ] and [ roc2 ] plots the roc curves for some combinations of @xmath33 and @xmath50 . for @xmath52 and @xmath33 small ,
all five statistics behave similarly with t - statistics performing the worst .
as @xmath33 increases , t becomes better and os and copa begin to lose power . for @xmath53 and medium to large @xmath33 ,
the performance of most is only worse than t and better than other statistics .
smaller @xmath33 in this case basically leads to roc curve that is close to a @xmath54 line for all statistics since the signal @xmath53 is too weak in this case , so we do not show these results . for @xmath55 and small @xmath33 ,
most is better than ort , copa and t , and in this situation only os is competitive with most .
larger @xmath33 in this case will produce nearly perfect roc curves for all statistics , and thus those results are also omitted . besides roc curves , we have also tried examining the possibility of using ( [ k ] ) for estimating the number of differentially expressed samples @xmath33 , but so far have been unable to get a reasonable estimate out of it . from the above simulations
, we judge that our new estimate most is at least as good as other previously proposed statistics , sometimes much better .
thus it is a valuable tool for detecting activated genes in many situations . as an example of real data application ,
the data from @xcite is publicly available from http://data.cgt.duke.edu/west.php .
the microarray used in the breast cancer study contains 7129 genes and 49 tumor samples , 25 of which with no positive lymph nodes identified and the other 24 with positive nodes . similar to @xcite , we take the log transformation of the expressions after normalizing the data .
we apply most to the data and compare it to the t - statistics by computing the fdr using the sam approach ( @xcite ) .
figure [ fdr ] plots the fdr versus the number of genes called significant . for this example , most seems to perform a little better than t - statistics , although the difference is too small to be of any significance .
significance analysis of microarrays applied to transcriptional responses to ionizing radiation .
_ proceedings of the national academy of sciences of the united states of america _ * 98 * , 511621 . | we propose a new statistics for the detection of differentially expressed genes , when the genes are activated only in a subset of the samples .
statistics designed for this unconventional circumstance has proved to be valuable for most cancer studies , where oncogenes are activated for a small number of disease samples .
previous efforts made in this direction include copa ( @xcite ) , os ( @xcite ) and ort ( @xcite ) .
we propose a new statistics called maximum ordered subset t - statistics ( most ) which seems to be natural when the number of activated samples is unknown .
we compare most to other statistics and find the proposed method often has more power then its competitors .
cancer ; copa ; differential gene expression ; microarray . |
in special and general relativity the spin of a gyroscope is represented by a four - vector @xmath0 . assuming that we move the gyroscope without applying any torque to it ( in a system comoving with the gyroscope )
, the spin vector will obey the fermi transport equation @xmath1 here @xmath2 is the four - velocity of the gyroscope .
for a trajectory in a given spacetime , and a spin vector specified at some point along this trajectory , we can integrate ( [ rall ] ) to find the spin at any point along the trajectory .
the fermi transport equation is however deceivingly simple since we have not inserted explicitly the affine connection coming from the covariant differentiation . also , even when we have a flat spacetime and inertial coordinates ( so that the affine connection vanishes ) the equation is more complex than you might think . as an example we consider motion with fixed speed @xmath3 along a circle in the @xmath4-plane , with an angular frequency @xmath5 . letting the groscope start at @xmath6 at the positive @xmath7-axis
, we get a set of coupled differential equations @xmath8 here @xmath9 and @xmath10 is the spatial part of @xmath0 . for initial conditions @xmath11 the solutions ( see @xcite p. 175 - 176 ) can be written as @xmath12 + ( \gamma-1)\sin[\omega \gamma t ] \sin[\omega t]\right ) \quad \quad
\label{hm1}\\ s^y&=&s \left(\sin[(1-\gamma)\omega t ] -(\gamma-1)\sin[\omega \gamma t]\cos[\omega t]\right ) \label{hm2}\\ s^z&=&0 , \quad \quad s^0=-s r \omega \gamma \sin[\omega \gamma t ] \label{hm3}\end{aligned}\ ] ] looking at @xmath13 and @xmath14 , we note that ( written in the particular form above ) the first terms in respective expression corresponds to a rotation around the @xmath15-axis , but then there is also another superimposed rotation with time dependent amplitude . to find this solution directly from the coupled differential equations that are the fermi equations ,
seems at least at first sight quite difficult , even for this very symmetric and simple scenario . to get a simpler formalism we may consider , not the spin vector @xmath0 itself , but the spin vector we _ would _ get if we momentarily would stop the gyroscope ( relative to a certain inertial frame ) by a pure boost ( i.e. a non - rotating boost )
this object we will call the _ stopped _ spin vector .
while being a four - vector it is effectively a three - dimensional object ( having zero time component in the inertial frame in question ) and we will show that the spatial part of this object undergoes pure rotation with a constant rate for the example of motion along a circle in special relativity . knowing that there is a simple algebraic relation between the stopped and the standard spin vector , the stopped spin vector can be used as an intermediate step to easily find the standard spin vector .
there is however also a direct physical meaning to the stopped spin vector , apart from being the spin vector we would get if we stopped the gyroscope .
the stopped spin vector _ directly _ gives the spin as perceived in a comoving system , see section [ tttt ] for further discussion on this . in this article
we will also consider more general reference frames than inertial ones .
for instance we will consider a rotating and accelerating reference frame .
this allows us to apply the formalism , via the equivalence principle , to describe in a simple three - dimensional manner how a gyroscope orbiting for instance a rotating black hole will precess relative to the stationary observers . in figure [ fig1 ]
we illustrate how a gyroscope spin vector precesses relative to a vector parallel transported with respect to the spatial geometry . given a reference congruence of timelike worldlines
, we first derive a general spacetime transport equation for the stopped spin vector ( stopped relative to the congruence in question ) .
we then consider a spacetime equation corresponding to _ spatial _
parallel transport with respect to the spatial geometry defined by the congruence . for the case of a rigid congruence
, we easily derive such a transport law . considering a shearing congruence we use the formalism derived in @xcite .
having both the transport equation for the stopped spin vector and the equation for parallel transport , we can put them together and thus get an equation for how fast the stopped spin vector precesses relative to the local spatial geometry connected to the reference congruence .
as is the case for the inertial congruence , we will see that the precession corresponds to a simple law of three - rotation .
the general scheme as outlined here has already been pursued by jantzen et .
( see @xcite ) , although the angle of approach and the algebraic formalisms are different .
the explicit use of the three - dimensional formalism of this paper also appears novel .
this article is complementary to a companion paper @xcite , where the formalism of relativistic spin precession in three - dimensional language is derived in a very intuitive manner .
this paper verifies , through a more formal derivation , the result of @xcite for the particular case of a rigid congruence as assumed in @xcite .
let us denote the local four - velocity of our reference congruence by @xmath16 .
we introduce a _ stopped _
spin vector @xmath17 as the spin vector that we get if we make a pure boost of the spin vector such that it is at rest with respect to the local congruence line . in figure [ fig2 ]
we illustrate in 2 + 1 dimensions how the two spin vectors are related to each other .
it follows readily from the lorentz transformation that we get the stopped vector by removing the @xmath16-part of @xmath0 , and shortening the part parallel to the spatial direction of motion by a @xmath18-factor .
note that the resulting stopped vector is not in general parallel to the spatial part of @xmath0 .
letting @xmath19 be a normalized vector orthogonal to @xmath16 in the direction of motion , we can express the stopped spin vector as @xmath20 s^\alpha .\end{aligned}\ ] ] here we have adopted the spatial sign convention @xmath21 as we will throughout the article . knowing a little about thomas precession we may guess that for the simple case of motion along a circle in an inertial frame as discussed earlier , there is a simple law of three - dimensional rotation for this object . indeed in the following discussion we will show this , and at the same time consider the effects of rotation coming from having non - inertial reference frames ( connected to @xmath16 ) .
we also need an explicit expression for the standard spin vector in terms of the stopped spin - vector @xmath17 .
the relationship between the two vectors follows readily from the lorentz - transformation : @xmath22 \label{s1b } .\end{aligned}\ ] ] this we may now insert into the fermi transport equation to derive an expression for the _ stopped _
spin vector .
in this section we consider gyroscope transport relative to an arbitrary reference congruence @xmath16 . for a spin vector @xmath0 transported along a worldline of four - velocity @xmath2
, we have the fermi transport law @xmath23 using ( [ s1a ] ) in ( [ fermi ] ) readily yields @xmath24 .\end{aligned}\ ] ] we need now the inverse of @xmath25 to get an explicit transport equation for the stopped spin vector . through a general ansatz .
the ansatz is of the form @xmath26 . ] , we find @xmath27 that this is indeed the inverse of @xmath25 is easy to verify we are free to add terms containing @xmath28 , since these anyway die when multiplied by @xmath29 . if we instead would have defined @xmath30 we would get the inverse @xmath31 . here
the perfect symmetry in @xmath32 and @xmath33 is transparent .
there however does not appear to be any particular advantages of this gauge . ] .
so we have @xmath34{{k^{-1}}^\nu}_\mu .\end{aligned}\ ] ] here we have the desired expression . in [ simp ]
we expand and simplify this to find @xmath35_\perp -t_\alpha \left [ \frac{d}{d\tau}\left(u^\mu + \eta^\mu \right)\right]_\perp \right ) \nonumber\\ & & + \eta^\mu \bar{s}^\alpha \frac{d\eta_\alpha}{d\tau } .\end{aligned}\ ] ] by the perpendicular sign @xmath36 we here mean that we should select only the part orthogonal to both @xmath19 and @xmath16 .
note that @xmath37 means covariant differentiation along the gyroscope worldline .
equation ( [ kott1 ] ) then tells us how the stopped spin vector deviates from a parallel transported vector relative to a freely falling system .
in fact we notice from the antisymmetric form of ( [ kott1 ] ) that ( excepting the @xmath16 term ) it corresponds to a spatial rotation ( see section [ gotland ] for a more detailed argument ) .
that seems very reasonable since it insures that the norm of the stopped spin vector will be constant ( consider the rotation with respect to a freely falling system locally comoving with the congruence ) .
we also see that only if @xmath38 changes along the gyroscope worldline , with respect to a freely falling system , do we get a net rotation relative to this freely falling system . introducing the wedge product defined by @xmath39 and the projection operator @xmath40
, we can put ( [ kott1 ] ) in a more compact form @xmath41_\perp \right ) .\end{aligned}\ ] ] incidentally we may note that , as regards @xmath19-components within the bracketed expression , we do not need the @xmath36 sign .
any @xmath19 components within the bracketed expressions will cancel due to the anti - symmetrization as is easy to see .
we however keep the @xmath36 sign to indicate orthogonality to @xmath16 .
the simple form of ( [ kottfinal ] ) appears to be novel .
while we have yet to put the formalism in its final form , some applications and discussion may be useful already at this point for the simple case of an inertial reference congruence in special relativity . as a particular example
, consider a flat spacetime with an inertial congruence . for this case
it is not hard to show , see e.g @xcite , that the spatial curvature of a trajectory depends on the four - acceleration as @xmath42_\perp=\gamma^2 v^2\frac{n_\alpha}{r } .\end{aligned}\ ] ] here @xmath43 is the spatial curvature and @xmath44 is a normalized four - vector , orthogonal to the inertial congruence @xmath16 , pointing in the direction of spatial curvature . using this in ( [ kottfinal ] )
we get @xmath45 as we will see in the following section , this differential equation corresponds to a three - dimensional rotation .
choosing inertial coordinates adapted to the inertial congruence in question so that @xmath46 , @xmath47 and @xmath48 we get from ( [ kottfinal2 ] ) @xmath49 .\end{aligned}\ ] ] the expression within the brackets is a vector triple product and we may write it as a double cross product . letting @xmath50 we get @xmath51 rather than using @xmath52 we could use local time @xmath53 ( the time as experienced by observers at rest relative to the inertial congruence in question ) in which case we get a gamma factor less on the right hand side .
@xmath54 this is the famous thomas precession , in stopped spin vector three - formalism . introducing @xmath55 as the precession vector , around which the stopped spin vector rotates , we can alternatively write ( [ aye ] ) as @xmath56 looking at ( [ ayea ] ) component - wise , it is a set of coupled differential equations , just like the standard fermi equations . unlike the fermi - equations however , the new equations correspond to a simple law of rotation ( precession ) . as a specific example we may consider , as in the introduction , the precession of a gyroscope transported at constant speed @xmath3 around a circle of radius @xmath43 in the @xmath57 plane . assuming a motion with a clockwise angular velocity @xmath58 , the counterclockwise angular velocity @xmath59 for the precession of the stopped spin vector
is then according to ( [ ayeb ] ) given by @xmath60 consider then for instance the net precession after one lap .
the local time per lap is simply @xmath61 and hence the net precession angle ( in radians ) around the plane normal is given by @xmath62 . if the circular motion is counter - clockwise , the precession is clockwise and vise versa .
we can also trivially find the solution for the standard ( projected ) spin vector for the case of circular motion with constant speed with initial conditions as listed in the example in the introduction .
we know that the standard ( projected ) spin vector is related to the stopped spin vector through a lengthening of the stopped spin vector in the forward direction of motion @xmath63 by a @xmath18-factor .
we have then @xmath64 using the notation of the previous subsection we have then trivially for the case at hand @xmath65 using these expressions in ( [ hpp ] ) we immediately get the desired solution . using elementary rules for manipulating the trigonometric functions we can write it in the form of ( [ hm1])-([hm3 ] ) .
if we are interested also in @xmath66 , it is given by the orthogonality of the standard spin vector and the four - velocity as @xmath67 .
note that by use of the stopped spin vector formalism there is effectively no differential equation solving involved for this simple case . for motion in a circle with constant velocity ,
the fermi equation can be solved without use of the stopped spin vector formalism , although the solution is a bit complicated .
what about if we consider motion with constant velocity along some other curve , say a part of a parabola or some more irregular curve ? then the fermi equation would likely appear to be very complicated to solve analytically in the general case . using the method with the stopped spin vector the solution can however trivially be found for arbitrary curves .
first let us state a small theorem that we will then easily prove .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the stopped spin vector of a gyroscope transported with constant speed @xmath3 along a smooth curve in a spatial plane in a flat spacetime will rotate a net clockwise angle around the normal of the plane given by @xmath68 where @xmath69 is the net counterclockwise turning angle of the tangent direction of the curve . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ note that the parameter @xmath69 may be larger than @xmath70 . for a simple closed curve ( one that is not crossing itself ) , assuming the gyroscope to be transported once around the curve
, we have @xmath71 .
this theorem is easily proven by dividing an arbitrary smooth curve into infinitesimal segments within which we can consider the local curvature radius to be constant . letting @xmath5 denote the counter - clockwise angular velocity of the forward direction of motion @xmath63 ( so @xmath72 ) ,
we have according to ( [ how ] ) the clockwise angular velocity as @xmath73 .
thus the net angles of the gyroscope precession and the turning of the forward direction of the curve , along the segment in question , are related through @xmath74 .
adding up the precession contributions from all the segments of the curve we get @xmath75 thus the theorem is proven .
note that while the motion is assumed to be in a plane , the spin vector may point off the plane .
we can draw a conclusion from the above proven theorem ( also knowing that there is a simple algebraic relation between the stopped and the standard spin - vector ) that can be expressed in terms of the standard spin vector , without reference to the stopped spin vector .
consider then a smooth simple closed curve and let a certain point along this curve be the initial position for the gyroscope . for given initial spin vector , initial direction of motion and constant speed @xmath3 , the final spin vector ( after one lap around the loop ) is independent of the shape of the loop as illustrated in figure [ xf1 ] .
but of course , the theorem is stronger than this . given an arbitrary , not necessarily closed but smooth curve along which we transport the gyroscope with constant velocity , we can trivially find the standard spin vector at any point along the curve .
we take the spatial part of the initial spin vector and shorten the part parallel to the direction of motion by a @xmath18-factor to form the initial stopped spin vector . for any given curve @xmath76
we then calculate the initial direction of the curve together with the direction of the curve at the point in question .
then , modulo a winding number times @xmath77 we only get the turning angle @xmath69 up to a term @xmath78 , where @xmath79 is an integer , from the local quantity @xmath80 , we can trivially find the corresponding @xmath69 and thus through ( [ hh ] ) the corresponding stopped spin vector at the point in question . lengthening the parallel part of the stopped spin vector by a factor @xmath18 , we get the spatial part of the standard spin vector at the point in question . if we are interested in the zeroth component of the standard spin vector it is given by @xmath81 . thus solving a possibly very complicated differential equation
is reduced to performing a few algebraic steps . for motion in a plane where the velocity is not constant ,
the procedure is analogous to that described in section [ concon ] except that we need to integrate ( a single integral which may or may not be complicated to solve analytically ) to find @xmath82 .
for the most general motion , not necessarily confined to a plane and with a speed that may vary , it is however not just a matter of ordinary integration . given an arbitrary motion @xmath83 along a smooth curve we can however solve a differential equation , given by ( [ ayea ] ) and ( [ ayeb ] ) , for @xmath84 .
likely this differential equation will be simpler to solve than the fermi equation . to gain further intuition on the meaning of the stopped spin vector it may be useful to explore how it is related to other vectors of physical interest connected to the gyroscope spin . in particular we may consider the gyroscope intrinsic angular momentum , and the momentary direction of the gyroscope axis as perceived in the reference system in question ( where the observers are integral curves of @xmath16 ) .
consider then a gyroscope moving along a straight line in the @xmath4-plane in special relativity ( using inertial coordinates ) with constant speed .
the gyroscope axis is assumed to lie in the plane of motion and to be tilted somewhere between the forward and the sideways direction . in 2 + 1
dimensions we can easily visualize the worldsheet of the gyroscope central axis as well as various vectors of interest , see figure [ xf2 ] . we note that there are ( at least ) three different spatial directions of relevance for the gyroscope .
it is easy to realize ( length contraction ) that the direction of the gyroscope axis is simply related to the direction of the stopped spin vector through a gamma factor . given any of these directions the other two can thus easily be found .
furthermore one can show , at least for an idealized scenario as considered in [ spinapp ] , that the the intrinsic angular momentum , that we will denote @xmath85 , is in fact given by @xmath86 .
the various vectors involved are illustrated in figure [ xf3 ] .
consider a spatial vector @xmath87 that connects the base of the gyroscope to the tip of the gyroscope , as perceived in the reference system connected to @xmath16 .
we understand that this vector evolves according to a simple rule of rotation given by ( [ aye ] ) modulated by a contraction by a factor of @xmath18 in the direction of motion .
it is a short exercise to show that this means that @xmath87 in fact obeys a rather compact differential equation @xmath88.\end{aligned}\ ] ] we can perform a corresponding analysis for the projected spin vector to find @xmath89.\end{aligned}\ ] ] the equations for the stopped spin vector can be written in the form @xmath90 from ( [ rall5 ] ) , letting @xmath91 , we readily find @xmath92 - \gamma^2 v \frac{dv}{dt } { \bf s}{_{{\textrm{\scriptsize l}}}}.\end{aligned}\ ] ] comparing the four differential equations we see that they are all quite compact , although the equation for the stopped spin vector , corresponding to a pure rotation , is more likely to be simple to solve ( as we have seen for the example of motion on a circle ) .
one might argue that the object of physical interest is the intrinsic ( spin ) angular momentum of the gyroscope which is given by @xmath93 , or perhaps the observed direction of the gyroscope central axis . from this point of view
the stopped spin vector is in a sense a means to an end . by using the stopped spin vector as an intermediate step we can find the solutions to otherwise quite complicated differential equations for the objects of physical interest . from a mathematical point of view
this is certainly sufficient to motivate the use of the stopped spin vector .
there is however more to the stopped spin vector than this .
in particular we note that the stopped spin vector _ directly _ gives us the spin as perceived in a comoving system .
for instance , if the stopped spin vector is at a @xmath94 angle with respect to the forward direction
so it will be with respect to a system comoving with the gyroscope .
this is contrary to the standard spin vector which only gives the spin direction with respect to the comoving system after a lorentz transformation .
consider the following example .
a gyroscope is suspended inside a satellite such that no torque is exerted on the gyroscope as seen from the satellite .
the satellite is assumed to be orbiting along some predetermined smooth simple closed curve , on a plane in special relativity . ] , using it s jet engines to stay on the path .
suppose then that we wish to measure , from the satellite , the precession angle of the gyroscope ( as predicted by relativity ) after a full orbit ( or maybe several full orbits ) .
we note that the direction of the gyroscope relative to the satellite itself is not a good measure . assuming that we have a couple of fixed stars
, we can however use the direction of these stars ( as perceived from the satellite ) as guidelines to set up a reference system within the satellite for this scenario the stopped spin vector is exactly the physical object that we are interested in .
it exactly represents the gyroscope direction relative to the star - calibrated reference system of the satellite .
thus if the stopped spin vector turns a certain angle , that is precisely the turning angle of the gyroscope relative to the star - calibrated reference system of the satellite .
while we are here focusing on spinning gyroscopes , it should also be noted that the formalism of the stopped spin vector is immediately applicable to describe the resulting rotation of _ any _ object which has zero _ proper _ ( comoving ) rotation . in conclusion , the stopped spin vector may be used as an intermediate step to simplify the calculation of the evolution of the intrinsic angular momentum ( spin ) of a gyroscope , or the perceived direction of the gyroscope axis . the stopped spin vector is however also of direct physical importance since it gives us the spin as perceived in a comoving system .
so far we have only given examples that apply to flat spacetime , and inertial reference frames .
as we will see in the following sections the stopped spin vector can be just as useful also for curved spacetimes and non - inertial reference frames .
the transport equation ( [ kottfinal ] ) tells us how the stopped spin vector deviates from a vector that is parallel transported with respect to the spacetime geometry .
this by itself is however not really what we are after if the reference congruence is non - inertial . to get a truly three - dimensional formalism , we in stead want an expression telling us
how fast the stopped spin vector deviates ( rotates ) from a vector that is parallel transported with respect to the _ spatial _ geometry determined by the congruence .
as is demonstrated in @xcite and in @xcite , it is possible derive a _
spacetime _ transport law corresponding to a _ spatial _ parallel transport . for the simple , and perhaps most useful , case of a rigid congruence
the issue is sufficiently simple that we will briefly review it in the coming subsection .
suppose then that we have a rigid congruence with nonzero acceleration @xmath95 , nonzero rotation tensor @xmath96 but with vanishing expansion - shear tensor @xmath97 , the acceleration vector @xmath98 , the shear tensor @xmath99 and the rotation tensor @xmath100 .
furthermore we employ what we denote the expansion - shear tensor @xmath101 . ] . in figure [ fig3 ]
we show an illustration of the spacetime transport of a vector orthogonal to the congruence .
it is easy to show that in the coordinates of a freely falling system @xmath102 , locally comoving with the congruence , the velocity of the congruence points ( assuming vanishing @xmath97 ) is to first order given by @xmath103 knowing that the velocity of the congruence is zero to lowest order , relative to the inertial system in question , we need not worry about length contraction and such .
it is then easy to realize that the proper spacetime transport law of a spatial vector @xmath104 corresponding to standard spatial parallel transport is @xmath105 here @xmath106 can easily be determined from the orthogonality of @xmath104 and @xmath16 follows ( differentiate @xmath107 along the gyroscope worldline ) that @xmath108 .
contracting both sides of ( [ init ] ) by @xmath109 gives @xmath110 . ] .
here we have then a spacetime transport equation corresponding to spatial parallel transport , for the case of a non - shearing ( non - expanding ) congruence . for a more complicated congruence that is shearing and expanding , it is not quite so obvious how to define the spatial parallel transport .
indeed as discussed in e.g @xcite and @xcite , there are several ways of doing this .
we will here follow the approach of @xcite , and consider two different such parallel transports .
these transports are connected to two different ways of defining a spatial curvature for a test particle worldline , with respect to the congruence @xmath111_\perp = [ a^\mu]_\perp+2v ( { \omega^\mu}_\alpha t^\alpha + [ { \theta^\mu}_\alpha t^\alpha]_\perp ) + v^2 \frac{n{_{{\textrm{\scriptsize ps}}}}^\mu}{r{_{{\textrm{\scriptsize ps } } } } } \\[3 mm ] \textrm{new : } & & \frac{1}{\gamma^2 } \left[\frac{du^\mu}{d\tau}\right]_\perp = [ a^\mu]_\perp+2v { \omega^\mu}_\alpha t^\alpha + v^2 \frac{n{_{{\textrm{\scriptsize ns}}}}^\mu}{r{_{{\textrm{\scriptsize ns } } } } } .\end{aligned}\ ] ] here @xmath112 and @xmath113 are the curvature and the curvature direction that we get if we project the the spacetime trajectory down along the congruence onto a local timeslice ( orthogonal to the congruence at the point in question ) .
the suffix ps stands for projected straight. the curvature @xmath114 and the curvature direction @xmath115 are defined with respect to deviations from a certain ( new ) notion of a spatially straight line .
the latter is defined as a line that with respect to variations in the projected curvature , leaves the integrated spatial distance ( as defined by the congruence ) unaltered ( to first order in the variation ) . as it turns out , a straight line with respect to this definition , has in general a non - zero projected curvature when the congruence is shearing .
the suffix ns stands for new - straight. this particular curvature is connected to fermat s principle , and optical geometry @xcite . for brevity
we let the suffix s denote either ps , or ns. introducing @xmath116 , @xmath117 we can then express both curvatures jointly as @xmath118_\perp & = & [ a^\mu]_\perp+2v ( { \omega^\mu}_\alpha t^\alpha + c{_{{\textrm{\scriptsize s}}}}[{\theta^\mu}_\alpha t^\alpha]_\perp ) + v^2 \frac{n{_{{\textrm{\scriptsize s}}}}^\mu}{r{_{{\textrm{\scriptsize s}}}}}.\end{aligned}\ ] ] from these two curvature measures one can introduce corresponding equations for spatial parallel transports @xcite .
a joint expression for the parallel transport of a vector @xmath104 is given by @xmath119 here @xmath120 is the covariant derivative along the ( gyroscope ) worldline in question .
notice that for vanishing shear expansion tensor , the two transports both correspond to ( [ init ] ) .
having defined two types of parallel transport according to ( [ ps ] ) , we can define corresponding covariant differentiations along a curve as @xmath121 these derivatives then tells us how fast a vector deviates from a corresponding parallel transported vector ( momentarily parallel to the vector in question ) . substituting @xmath122 and using ( [ kottfinal ] ) we get the equations for how fast the stopped spin vector precesses relative to a spatially parallel transported vector ( of the two types ) .
first we however rewrite ( [ kottfinal ] ) .
we saw in the preceding section how the kinematical invariants of the congruence entered naturally in the definition of spatial parallel transport .
we can also expand @xmath123 , in the transport equation ( [ kottfinal ] ) for the stopped spin vector , in terms of the kinematical invariants of the congruence .
first of all we have @xmath124 also we know that ( see e.g @xcite p. 566 ) @xmath125 using ( [ katta ] ) , we have then @xmath126 using this together with ( [ jcurve ] ) in ( [ kottfinal ] ) , also adding the proper @xmath16-term enabling the removal of the projection operator in ( [ kottfinal ] ) , we readily find @xmath127 + \eta^\mu \bar{s}^\alpha \frac{d\eta_\alpha}{d\tau } .\end{aligned}\ ] ] notice that we have omitted the perpendicular signs ( @xmath36 ) on @xmath128 and @xmath129 since these objects are already orthogonal to @xmath16 and any @xmath19 components die due to the anti - symmetrization .
now it is time to put together the results of the preceding two sections . what we want
is the net rotation of the stopped spin vector relative to a spatially parallel transported vector . using ( [ delettps ] ) and ( [ transproj ] ) ( setting @xmath130 )
, we then readily find @xmath131 .\end{aligned}\ ] ] here @xmath116 and @xmath117 .
so this gives us how fast a gyroscope stopped spin vector deviates from a corresponding ( spatially ) parallel transported vector . in particular considering the expression in a freely falling system locally comoving with the congruence , we understand that the expression within the brackets on the right hand side is simply the effective rotation tensor relative to the spatial geometry
. it could be practical with an expression corresponding to ( [ finp ] ) but where the proper four - acceleration is explicit . using ( [ kottfinal ] ) , ( [ transproj ] ) and ( [ deta ] )
we readily find @xmath132_\perp { \hspace{-0.5mm}}{\hspace{-0.5mm}}{\hspace{-0.5mm}}{\hspace{-0.5mm}}+ { \hspace{-0.5mm}}\gamma a_\alpha { \hspace{-0.5mm}}+ { \hspace{-0.5mm}}\gamma v \omega_{\alpha \rho}t^\rho { \hspace{-0.5mm}}+ { \hspace{-0.5mm}}(2\gamma c{_{{\textrm{\scriptsize s } } } } { \hspace{-0.5mm}}-{\hspace{-0.5mm}}1)\frac{\gamma { \hspace{-0.5mm}}+ { \hspace{-0.5mm}}1}{\gamma v}\theta_{\alpha \rho}t^\rho\right ] \nonumber\\ & & \hspace{-0 mm } -\gamma { \omega^\mu}_\alpha \bar{s}^\alpha .\end{aligned}\ ] ] notice that the expression for the four - acceleration here ( naturally ) is independent of what curvature measure that we use . still ( [ fin2p ] ) depends on what curvature measure we are using ( manifesting itself in the occurrence of @xmath133 ) assuming non - zero @xmath128 , since the transport laws for the two types of spatial parallel transport differs .
we can rewrite ( [ finp ] ) and ( [ fin2p ] ) as purely three - dimensional equations . for any specific global labeling of the congruence lines ( i.e. any specific set of spatial coordinates adapted to the congruence )
we can locally choose a time slice orthogonal to the congruence so that @xmath134 .
this then uniquely defines the three - vector @xmath84 at any point along the gyroscope trajectory .
analogous to what we did in going from ( [ kottfinal2 ] ) to ( [ e2 ] ) , for a set of vectors @xmath17 , @xmath19 and @xmath104 orthogonal to the congruence , we let @xmath135 . of two
three - vectors @xmath136 and @xmath137 is @xmath138^{-\frac{1}{2 } } \epsilon^{ijk}a_j b_k$ ] where the indices have been lowered with the local three - metric ( again assuming local coordinates orthogonal to the congruence ) .
notice that in general ( for congruences with rotation ) there are no global time - slices that are orthogonal to the congruence .
the local three - metric corresponding to local orthogonal coordinates is however well defined everywhere anyway . for a shearing ( expanding ) congruence
it will however be time dependent ( whatever global time slices we choose ) . ]
also we let @xmath139 in coordinates locally comoving with the congruence , we have @xmath140 , where @xmath141 $ ] and @xmath142 is @xmath143 , @xmath144 or @xmath145 for @xmath146 being an even , odd or no permutation of @xmath147 respectively . ] .
for simplicity , let us assume that the congruence has vanishing shear and expansion .
for this case the two different approaches to spatial curvature radius coincide and we will drop any instances of subscripts ns or ps. introduce then @xmath148 , where @xmath149 and @xmath53 are the inertial coordinates of a system locally comoving with the congruence naturally transforms as a three - vector . ] . also denoting the acceleration of the reference congruence relative to an inertial system locally comoving with the reference congruence by @xmath150
, we get from ( [ fin2p ] ) corresponds to covariant differentiation with respect to the three - metric . ] @xmath151\times { \bf v } \right ) \times { \bf \bar{s}}-\gamma { \mbox{\boldmath$\omega$ } } \times { \bf \bar{s } } .\end{aligned}\ ] ] this is a perfect match with the result of the intuitive derivation performed in @xcite .
analogously we may study ( [ finp ] ) for the particular case of vanishing shear , thus considering a rigid congruence .
the three - dimensional version of this equation then becomes @xmath152 \times \bar{\bf s } .\end{aligned}\ ] ] we may simplify this expression a bit by introducing @xmath153 , where @xmath154 and @xmath36 means parallel respectively perpendicular to @xmath63 . also using @xmath50
we readily find @xmath155 \times \bar{\bf s } .\end{aligned}\ ] ] again this is a perfect match with the intuitive formalism of @xcite .
letting @xmath156 denote local time for an observer comoving with the congruence we can write ( [ fin3 ] ) and ( [ fin5 ] ) respectively as @xmath157 here @xmath158 is given by ( [ fin3 ] ) @xmath159 this form is practical if the gyroscope is freely falling , in which case @xmath160 .
alternatively we can get @xmath158 from ( [ fin5 ] ) @xmath161 this form is practical if the gyroscope follows some predetermined path while being acted on by forces . from the simple relation ( see section [ spax ] ) between the stopped spin vector and the projected spin vector and the gyroscope axis respectively , we can use the law of rotation for the stopped spin vector to derive corresponding differential equations for @xmath10 and * x * @xmath162 + { \mbox{\boldmath$\omega$}}_{{{\textrm{\scriptsize e } } } \parallel } \times [ { \bf s}]_\perp + { \mbox{\boldmath$\omega$}}_{{{\textrm{\scriptsize e } } } \perp } \times \left ( \frac{1}{\gamma } [ { \bf s}]_\parallel + \gamma [ { \bf s}]_\perp \right ) \label{combo1}\\ \frac{d { \bf x}}{d\tau_0}{\hspace{-0.5mm}}{\hspace{-0.5mm}}{\hspace{-0.5mm}}{\hspace{-0.5mm}}{\hspace{-0.5mm}}&=&{\hspace{-0.5mm}}{\hspace{-0.5mm}}{\hspace{-0.5mm}}{\hspace{-0.5mm}}-\gamma^2 \frac{d { \bf v}}{d\tau_0 } \left[{\bf x } \cdot { \bf v } \right]+ { \mbox{\boldmath$\omega$}}_{{{\textrm{\scriptsize e } } } \parallel } \times [ { \bf x}]_\perp + { \mbox{\boldmath$\omega$}}_{{{\textrm{\scriptsize e } } } \perp } \times \left ( \gamma [ { \bf x}]_\parallel + \frac{1}{\gamma } [ { \bf x}]_\perp \right ) . \label{combo2}\end{aligned}\ ] ] here we have for brevity introduced @xmath163 note that the @xmath164 entering ( [ combo1 ] ) and ( [ combo2 ] ) is the velocity derivative relative to the reference frame ( not relative to a freely falling frame ) .
we note that these differential equations are more complicated than the ones for the stopped spin vector .
we conclude that if we are interested in @xmath10 or @xmath87 , it is likely wise to first solve the equation for the stopped spin vector and then ( as in section [ revhep ] ) use the result to find @xmath10 or @xmath87 .
as a first example of how one may use the derived formalism , consider a train moving along a straight spatial line in some static geometry .
in fact , to be specific , we may consider the train to be moving relative to an upwards accelerating platform in special relativity . on the train
we have suspended a gyroscope so that there are no torques acting on it in the comoving system .
we thus consider gyroscope motion along a straight line , with respect to a non - rotating but accelerating reference frame .
letting @xmath165 and @xmath166 , ( [ fin5 ] ) is immediately reduced to @xmath167 we understand that a gyroscope initially pointing in the forward direction will tip forward as depicted in figure [ fig12 ] .
note that the stopped spin vector with respect to the platform corresponds precisely to the spin vector as perceived relative to the train .
for example , if the stopped spin vector points @xmath94 down from the horizontal direction , the gyroscope as seen from the train points @xmath94 down from the horizontal direction . to express
the gyroscope precession with respect to coordinates comoving with the train we therefore just let @xmath168 in ( [ perfa ] ) and we have the precession explicitly in terms of the time @xmath52 on the train .
relative to the train , the gyrocope thus precesses at a steady rate given by @xmath169 .
this means that the train in fact has a proper rotation , but more on this is given in @xcite .
we can parameterize the gyroscope trajectory by the distance @xmath170 along the platform rather than the time @xmath53 .
then ( [ perfa ] ) can be expressed as @xmath171 assuming the train velocity to be low , the tipping angle per distance traveled is thus independent of the velocity .
we have simply @xmath172 .
thus on a stretch of length @xmath173 we get a net rotation @xmath174
@xmath175 if we want to express @xmath173 and @xmath176 and in si units we must divide the right hand side by @xmath177 ( expressed in si units ) . setting @xmath178 and @xmath179
we get simply @xmath180 this is quite a small angle , and we understand that the relativistic effects of gyroscope precession for most cases here at earth are small .
notice how simple this calculation was in the stopped spin vector three - formalism .
the equations ( [ fin3 ] ) and ( [ fin5 ] ) both describe how the gyroscope rotates with respect to a coordinate frame that is parallel transported with respect to the spatial geometry .
suppose then that we consider motion in the equatorial plane of some axisymmetric geometry . as a specific example we may want to know the net rotation of the gyroscope relative to its initial configuration after a full orbit ( not necessarily a circular orbit ) .
we must then take into consideration that a parallel transported frame in general will be rotated relative to its initial configuration after a complete orbit .
we can however introduce a new reference frame , that rotates relative to local coordinates spanned by @xmath181 and @xmath182 , in the same manner as a parallel transported reference frame does on a plane .
in other words , if we for instance consider a counterclockwise displacement ( @xmath183,@xmath184 ) , then relative to the local vectors @xmath181 and @xmath185 , the new reference frame should turn precisely @xmath186 clockwise .
such a reference frame would always return to its initial configuration after a full orbit . to find the rotation of the new reference frame with respect to a parallel transported frame
, we first investigate how a vector that is parallel transported with respect to the curved axisymmetric geometry rotates relative to the local coordinates spanned by @xmath181 and @xmath182 .
the line element for a two - dimensional axisymmetric spatial geometry can be written in the form - terms , we can not simply select the spatial terms ( without @xmath187 ) to get the spatial line element .
there is however an effective spatial geometry also for this case .
we may derive the form of this geometry by for instance sending photons back and forth between nearby spatial points and checking the proper time that passes . ] @xmath188 as depicted in figure [ fig14 ] we can imagine an embedding of the geometry , where we cut out a small section and put it on a flat plane .
what we want is an expression for how much a vector that is parallel transported , for example along the depicted straight dashed line , rotates relative to the local coordinates @xmath181 and @xmath182 .
we understand that the rotation angle corresponds to the angle @xmath174 as depicted .
using the notations introduced in figure [ fig14 ] we have simply @xmath189 eliminating @xmath190 and using @xmath191 it follows readily that @xmath192 so this tells us how a parallel transported vector turns relative to the local @xmath181 and @xmath182 , for a small displacement in @xmath193 and @xmath194 . on a flat plane ,
the corresponding expression to ( [ rdone ] ) is of course simply @xmath195 a reference frame that with respect to @xmath181 and @xmath182 rotates as if we had a flat geometry would then according to ( [ rdone ] ) and ( [ rdonex ] ) rotate relative to a parallel transported reference frame with an angular frequency ( never mind the sign for the moment ) @xmath196 note also that we have @xmath197 thinking about the sign for a second , we realize that with respect to the would - be - flat reference frame , a parallel transported reference frame will have a rotation given by @xmath198 knowing that infinitesimal rotations can simply be added ( to lowest order ) , using ( [ katt ] ) together with ( [ rjx2 ] ) and letting @xmath199 , we get the gyroscope rotation relative to the would - be - flat-grid as ) a little by introducing a @xmath200 sign in the @xmath201-term .
if @xmath181 , which by definition is taken to point away from the center of symmetry , points in the direction of increasing @xmath194 , we choose the positive sign , otherwise the negative sign should be chosen .
note that @xmath201 is zero for radii where the sign changes , so there is no discontinuity in @xmath202 . ]
@xmath203 we can integrate this equation to find the net precession of a gyroscope transported along any path in the axisymmetric geometry .
as a particular application of ( [ katt2 ] ) , we can consider the net precession of a gyroscope transported along some closed orbit .
since the would - be - flat reference frame returns to its original configuration after a full turn , we just integrate the effects of the infinitesimal rotations following from ( [ katt2 ] ) to calculate the net turn .
notice however that to do this straightforwardly , we need @xmath202 in the coordinate base of the reference frame ( i.e. the would - be - flat frame ) .
in general we however only have @xmath202 in the coordinates adapted to the stationary observers . for most cases where we would be interested in motion in an axisymmetric geometry , like motion in the equatorial plane of a kerr black hole , this however presents no problem .
then all rotations are in the plane of motion and the rotation vector @xmath202 is constant ( in the @xmath204-direction ) in the coordinate basis of the reference frame .
notice that the @xmath53 implicitly entering these equations in the @xmath202 is the proper time for a stationary observer .
if we instead want to express the precession in global ( schwarzschild ) time , we just multiply by the time dilation factor . even assuming all rotations to be in the plane of motion ,
we must still in general integrate to get the net precession of the gyroscope , we understand that time dilation , @xmath43 , @xmath205 , @xmath3 and @xmath206 all depends on @xmath207 . assuming all rotations to be in the plane of motion it is effectively a single ( scalar ) integral ( of the net rotation angle around the @xmath15-axis ) . ] .
for the particular case of circular motion with constant speed , assuming the time dilation ( i.e. the lapse ) , @xmath205 and @xmath208 to be independent of @xmath193 ( as is the case for the equatorial plane of a kerr black hole ) , there is however no need to integrate at all since all the terms of ( [ katt2 ] ) are constant .
the result follows immediately , assuming that we know @xmath205 , @xmath209 and @xmath210 .
incidentally it follows from ( [ pr1 ] ) and ( [ pr2 ] ) that @xmath211 for circular motion .
the reference background ( fixed to the stars ) around a spinning planet , like the earth , is both accelerating and curved .
also there is frame dragging due to the planet rotation , giving a non - zero rotation of the stationary reference observers .
if we have the spacetime metric , we can easily find @xmath205 , @xmath209 and @xmath210 . if we do not have an exact spacetime metric however , as is the case for a spinning planet , we need some approximate method ( like the post - newtonian approximation ) to estimate @xmath205 , @xmath209 and @xmath210 .
once this is done , assuming the approximation to be valid , ( [ katt2 ] ) gives an accurate description of the precession even considering relativistic speeds . in the case of a rotating ( kerr )
black hole , we do know the metric , and the precession relative to the stationary observers can readily be calculated .
notice that within the ergosphere , there are no stationary ( timelike ) observers .
still we can in principle use the formalism of this paper also within the ergosphere . to do this
we consider coordinates that rigidly rotate around the black hole sufficiently fast to be timelike in the region in question . indeed for the particular case of circular motion
there is a paper @xcite , that uses this technique . as a simple example , consider a freely falling gyroscope ( @xmath212 ) orbiting in the equatorial plane of a schwarzschild black hole . using the static observers as our reference congruence
, we have @xmath213 .
then it follows from ( [ rjx3 ] ) that we have @xmath214 the schwarzschild line element in the equatorial plane is given by @xmath215 here @xmath216 . for convenience
we now consider large @xmath194 , so that @xmath217 is small .
we have then the acceleration of the freely falling frames @xmath218 ( counted positive in the inwards direction ) .
it follows readily , using ( [ katt ] ) , that for this case we have @xmath219 for the large @xmath194 in question the velocities are low and we may set @xmath220 . using ( [ ospace ] ) together with @xmath221 and @xmath222 in ( [ yesrotny ] ) gives @xmath223 this result was derived by w. de sitter in 1916 ( although in a quite different manner than that described here , see @xcite p. 1119 ) .
we may note that one third of the net effect comes from the spatial geometry . using a little bit of newtonian mechanics
it is easy to derive that for a satellite orbiting the earth at a radius @xmath224 , inserting the proper factor of @xmath225 to handle si - units , ( [ h3 ] ) amounts to @xmath226 knowing that the exterior metric of the earth is approximately schwarzschild , we have then an approximation of the effective rotation vector for a gyroscope orbiting the earth . we can refine this approximation by considering an appropriate non - zero @xmath205 , as discussed earlier .
note that , as discussed in section [ tttt ] , the derived precession is the precession with respect to a star - calibrated reference system on the satellite . in @xcite p. 1117
- 1120 , a similar explicitly three - dimensional formalism of spin precession is derived .
it is only valid in the post - newtonian regime however .
the precession given by ( [ katt2 ] ) is however exact ( assuming an ideal gyroscope ) .
for instance , considering the above example of free circular motion in a static geometry , we can easily calculate the exact expressions for @xmath176 and @xmath3 , and thus express the gyroscope precession rate arbitrarily close to the horizon .
we have seen how we in a covariant manner can derive an effectively three - dimensional spin precession formalism in a general spacetime . in particular the simple form of ( [ kottfinal ] )
seems novel . in @xcite p. 1117
a similar approach is taken where they consider only the standard spin vector , but expressed relative to a _ boosted _ set of base vectors .
they however only apply it to the post - newtonian regime . as mentioned earlier , jantzen et .
( @xcite ) have already pursued the same general idea , although the specific approach and final form of the results differ .
in particular they have not employed the explicit 3-dimensional formalism . while the general formalism is derived assuming a general congruence
, it seems to have its greatest use as a simple three - dimensional formalism assuming a non - shearing congruence .
then we have a fixed spatial geometry and the spatial parallel transport is unambiguous . for this particular case , the derived three - dimensional formalism verifies the result of the intuitive derivation of @xcite .
we have also given examples of how the three dimensional formalism can be used to easily find results of physical interest .
in the expansion of the second term within the brackets of ( [ done ] ) there will according to ( [ kk ] ) be terms of the type @xmath227 .
these can be rewritten in terms of @xmath228 and @xmath229 since we have @xmath230 dealing with @xmath228 rather than @xmath227 is convenient since the former readily can be expressed in terms of spatial curvature and velocity changes relative to the congruence , see @xcite .
also @xmath228 has a direct physical relevance . using the identity @xmath231
, it is then easy to derive an alternative form of @xmath25 @xmath232 using this in the second term within the brackets of ( [ done ] ) we have @xmath233 .\end{aligned}\ ] ] as we expand this expression there will be terms containing @xmath16 , @xmath28 and @xmath234 .
these we will disregard for the following reasons . terms containing @xmath28 will anyway die when multiplied by @xmath29 ( as they are in ( [ done ] ) ) .
terms containing @xmath16 we will disregard since we for the moment only are interested in @xmath235 . when we have a neat expression for this we can find the @xmath16-part a posteriori using the orthogonality of @xmath29 and @xmath28 .
we will disregard terms of the type @xmath234 since we know that these must cancel anyway for @xmath29 to stay normalized ( as we know it must by construction of the fermi transport and the relation to the stopped spin vector ) . for the particular case where @xmath236 momentarily
, it follows that any net term of the form @xmath237 in the right hand side of ( [ done ] ) must vanish .
since the parameter @xmath238 does not depend on @xmath239 it follows that it must vanish entirely .
the point is that the form of ( [ come ] ) is such that , when expanded it can be written as a sum of tensors of the type @xmath240 . letting the suffix @xmath36 indicate that only the part orthogonal to both @xmath16 and @xmath19
should be selected , each such term can be written in the form @xmath241_\perp)(t_\alpha t^\sigma b_\sigma+ [ b_\alpha]_\perp)$ ] . adding up the resulting terms of the type @xmath234 ( including the terms of this type coming from the first term within the brackets of ( [ done ] ) ) into a single term @xmath242 we _ know _ that @xmath238 must be zero . ] . note however that in principle , we should contract with the inverse @xmath243 ) that the effect of contracting @xmath244 with a contravariant vector is that it shortens the @xmath19-component of the vector by a @xmath18-factor , while the rest of the on - slice ( orthogonal to @xmath16 ) part of thee vector is unaffected .
] , before disregarding the terms of the described types ( see ( [ done ] ) ) .
the form of the inverse is however such that we can carry out the effective cancellations prior to applying the inverse we could not cancel @xmath16 terms directly within the brackets of ( [ done ] ) .
that the inverse is not containing any such terms is a benefit of the particular gauge choice in choosing @xmath245 where we had a freedom to include any terms containing @xmath109 . ] .
we then readily find @xmath246_\perp t^\mu+ t_\alpha \left[\frac{d u^\mu}{d\tau } + \frac{d \eta^\mu}{d\tau } \right]_\perp \right ) .\end{aligned}\ ] ] by the perpendicular sign @xmath36 we here mean that we should select only the part orthogonal to both @xmath19 and @xmath16 . by @xmath247
we indicate that the equality holds excepting terms of the type @xmath16 , @xmath28 and @xmath234 . in an analogous manner
we readily find for the first term within brackets of ( [ done ] ) @xmath248_\perp .\end{aligned}\ ] ] now use ( [ part1 ] ) and ( [ part2 ] ) in ( [ done ] ) .
shorten the @xmath19 components by a @xmath18 factor ( according to the effect of the inverse ) , and neglect the @xmath16-term .
we readily find @xmath249_\perp -t_\alpha \left [ \frac{d}{d\tau}\left(u^\mu + \eta^\mu \right)\right]_\perp \right ) .\end{aligned}\ ] ] now that we have this compact expression we may also find the @xmath16 term that we earlier omitted
. from orthogonality , @xmath250 , follows that @xmath251 which gives @xmath252_\perp
-t_\alpha \left [ \frac{d}{d\tau}\left(u^\mu + \eta^\mu \right)\right]_\perp \right ) \nonumber\\ & & + \eta^\mu \bar{s}^\alpha \frac{d\eta_\alpha}{d\tau } .\end{aligned}\ ] ] so here we have the transport equation for the stopped spin vector ( giving the rotation relative to inertial coordinates ) .
as an idealized scenario we consider a special relativistic gyroscope which we model as an isolated system of point particles undergoing four - momentum conserving internal collisions . following the discussion in @xcite p. 87
- 90 , we define the angular momentum tensor with respect to the spacetime origin as @xmath253 here the summation runs over events @xmath254 and four - momenta @xmath255 for the various particles at a particular time slice @xmath256 .
the ( pauli - lubanski ) spin vector can be written as @xmath257 here @xmath258 is the four - velocity of the center of mass and @xmath259 is the levi - civita tensor ( density ) where @xmath260 .
furthermore we introduce an angular momentum four - vector @xmath261 , where @xmath262 is the standard ( relativistic ) angular momentum three - vector , with respect to our reference coordinates .
defining @xmath16 as a purely time directed normalized vector with respect to the reference coordinates , we can write @xmath263 letting @xmath264 be the velocity of the center of mass , @xmath18 the corresponding gamma factor and setting @xmath265 with respect to the reference coordinates , we can decompose the four - velocity of the center of mass as @xmath266 . using this in ( [ britt ] ) together with ( [ bratt ] ) , it follows that @xmath267 it is a short exercise to show that in three - formalism this amounts to @xmath268 here @xmath262 is the net angular momentum of the system of point particles , @xmath269 is the center of mass ( center of energy ) , @xmath18 is the gamma factor for the velocity of the center of mass , @xmath270 is the net relativistic three - momentum and @xmath10 is the spatial part of the spin vector . note that the intrinsic angular momentum is not given by @xmath10 but by @xmath93 .
note incidentally also that there is a difference between the center of mass and the _ proper _ center of mass ( see @xcite p. 87
- 90 ) . as pointed out e.g. in @xcite , the gyroscope center of mass
does not in general lie on the gyroscope central axis .
a real gyroscope moving under the influence of forces is neither ( simply ) consisting of point particles nor is it isolated .
a more detailed analysis would likely assume a general energy momentum tensor @xmath271 and allow for external forces acting on the elements of the gyroscope .
for the purposes of this article the simple derivation outlined above will however suffice .
+ + * references * + | we derive an effectively three - dimensional relativistic spin precession formalism .
the formalism is applicable to any spacetime where an arbitrary timelike reference congruence of worldlines is specified .
we employ what we call a _ stopped _
spin vector which is the spin vector that we would get if we momentarily make a pure boost of the spin vector to stop it relative to the congruence .
starting from the fermi transport equation for the standard spin vector we derive a corresponding transport equation for the stopped spin vector . employing a spacetime transport equation for a vector along a worldline , corresponding to spatial parallel transport with respect to the congruence ,
we can write down a precession formula for a gyroscope relative to the local spatial geometry defined by the congruence . this general approach has already been pursued by jantzen et .
al .
( see e.g. jantzen , carini and bini 1997 _ ann .
phys . _ * 215 * 1 ) , but the algebraic form of our respective expressions differ .
we are also applying the formalism to a novel type of spatial parallel transport introduced in jonsson ( 2006 _ class .
quantum grav . _
* 23 * 1 ) , as well as verifying the validity of the intuitive approach of a forthcoming paper ( jonsson 2007 _ am .
journ .
phys . _ * 75 * 463 ) where gyroscope precession is explained entirely as a double thomas type of effect .
we also present the resulting formalism in explicit three - dimensional form ( using the boldface vector notation ) , and give examples of applications .
+ + pacs numbers : 04.20.-q , 95.30.sf |
the computer used is an ibm thinkpad laptop .
the experiments were done using an adaptive bulirschstoer algorithm @xcite in c++ .
the simulation was done in three dimensions and i checked to make sure that the orbit stays in the plane and that the total energy is conserved ; both to an accuracy of @xmath37 .
the data was then extracted and the plots were made in matlab ( figure 2 ) .
the theoretical plots were all generated in matlab . as stated before , we investigated the case where @xmath22 , @xmath23 , and @xmath24 . for numerical purposes
it is convenient to rewrite the equation @xmath124 in terms of geometrical constants of the orbit , namely eccentricity @xmath82 and @xmath125 .
hence , @xmath126 and are given by the energy and the angular momentum , which are in turn function of the initial conditions .
therefore eq .
[ eq : delke_gravapp ] reads @xmath127 , where @xmath125 and @xmath128 . here
@xmath129 and @xmath130 denote the magnitude of the angular momentum and the energy in the center of mass frame . a gravitational model for
this choice of parameters is one in which ( in mks units ) @xmath131 , @xmath132 @xmath133 .
therefore we have @xmath134 and @xmath135 .
the big mass @xmath8 was initially at the origin and @xmath7 had the coordinates @xmath136 with initial velocity @xmath137 .
the initial velocity of @xmath8 was @xmath138 ; therefore , @xmath139 from which we can calculate the constants of the orbit .
the angular momentum is conserved and is given by , @xmath140\hat{\mathbf{k}}$ ] , and the energy is given by , @xmath141 .
therefore the orbit constants @xmath83 and @xmath82 are specified and so is [ eq : final ] , which is plotted in figure 4 as a function of @xmath29 and compared to the experimental case .
the experiments were done by integrating newton s equations of motions in the lab frame subject to the same initial conditions .
namely , i integrated @xmath142 using bulirsch - stoer integrator .
i then saved the state of both particles ( position and velocity ) as well as the change in the energy of the whole system to make sure it was constant . by conservation of angular momentum motion
must take place in a plane , which i took to be the @xmath143 plane and monitored the @xmath117 component of the position of both bodies .
accuracies were @xmath37 .
furthermore i calculated , in the c++ code , the kinetic energies of both particles for values of @xmath144 , where @xmath145 .
these were written in .txt files that in turn were uploaded in a matlab code which plotted the asymptotic value of the kinetic energy of @xmath7 vs. @xmath29 .
the result is figure 2 .
the gain in the repulsive force field with @xmath24 is done similarly .
let @xcite : @xmath146 consequently , @xmath147 and @xmath148 .
as for the attractive case the motion of @xmath8 in the lab frame taken to be @xmath149 .
[ eq : delke ] becomes : @xmath150 @xmath151 for small angle scatterings @xmath152 ; therefore , @xmath153 .
the final form becomes : @xmath154
lastly , the impact parameter @xmath91 is determined by the angular momentum and @xmath167 via @xmath170 . with this ,
the problem is fully specified in the center of mass frame from the quantities given in the lab frame .
this is more standard and is done in most textbooks .
see for example ( * ? ? ?
* section 14 ) .
though , what we said applies to all central forces , recall that @xmath24 corresponds to gravitational and coulomb attraction . for @xmath171 the integrals can be integrated in terms of elementary functions @xcite .
[ [ subformulation - entirely_inlabformulation - of - our - results - entirely - in - the - lab - frame ] ] [ sub : formulation - entirely_inlab]formulation of our results entirely in the lab frame ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ from section 8.3 , we can obtain all the quantities in the lab frame , in particular we no loss of generality we take , @xmath174 , @xmath175 .
this specifies @xmath176 .
in addition , @xmath177 . hence , from eqs .
[ eq : delke_gravapp ] and [ eq : final ] we have , note that the expression depends on the square of the angular momentum .
more importantly note that the pre - factor to the @xmath179 depends on @xmath29 .
the angular momentum expressed in lab quantities reads , @xmath180\\ & \simeq & m_{1}\left[v_{2}\left(x_{2}-x_{1}\right)\sin\theta - v_{1}\left(y_{1}-y_{2}\right)\right]\end{aligned}\ ] ] which when substituted into [ eq : delke1 ] gives @xmath181 | we show that in attractive force fields a high speed particle with a small mass speeding through space , statistically loses energy by colliding softly and transversely with large masses that move slowly and randomly .
furthermore , we show that the opposite holds in repulsive force fields : the light particle statistically gains energy .
we consider the large masses to be distant from one another such that a series of two - body scatterings would suffice for the statistical analysis .
the difference of this process from equilibration is discussed .
[ sec : dynamical - description]_dynamical description- _ in addition to the well known gravitational and coulomb , nearly all other interactions in nature such as intermolecular forces and interaction of vortices in superconductors are central . in conservative fields ,
the central force on each particle can be derived from a potential function by @xmath0 where @xmath1 is the distance between the two bodies and @xmath2 ; here @xmath3 is the strength of the interaction depending on the parameters of the problem , @xmath4 defines the range of the interaction @xcite .
lastly , @xmath5 and @xmath6 correspond to attractive and repulsive force fields respectively @xcite .
in this paper we study a fast small mass passing through a dilute system of randomly moving central forces , where changes in the state of the small mass can be well approximated by a series of two - body scatterings . for the purposes of this section , it is best to think of the scatterings in the lab frame .
consider a scattering between two interacting particles @xmath7 and @xmath8 where @xmath7 is much lighter ( @xmath9 ) yet much faster than the other ( @xmath10 ) but nevertheless @xmath11 .
for example one can visualize a small comet ( @xmath7 ) undergoing a small angle scattering in the gravitational field of the planet jupiter ( @xmath8 ) .
for the sake of concreteness take the potential to be attractive for now . in a typical scattering
both masses are initially moving .
the questions we are interested in investigating are : what _ statistically _ invariant features are shared by such scatterings in central force fields ? would many such small angle scatterings have a net effect on the energy of @xmath7 ? while there is an active frontier of numerical work on many - body simulations @xcite , there are still interesting statistical inferences that can be derived from close analysis of two - body collisions .
further , such inferences in the cases where the collisions are head on have been extensively studied as in the fermi acceleration mechanism@xcite .
in contrast the main contribution to the effect presented in this paper is from collisions that are transverse ; i.e. , the trajectory of the massive particle , roughly speaking , is perpendicular to that of the small particle .
the organization is as following : 1 ) we give a heuristic description of the effect , 2 ) we provide a numerical verification 3 ) we prove the result by giving a mathematical derivation , 4 ) lastly we discuss this effect in a larger statistical context and discuss its difference from the equilibration process .
let us consider two extreme cases that would convey the gist of what underlies this work .
in the first case the massive particle , @xmath8 , slowly and _ transversely _ veers away from the trajectory of @xmath7 that is speeding by . in the second case , @xmath8 slowly and _ transversely _ approaches the trajectory of @xmath7 ( see figure [ fig : general_scattering(a , b)]-a ) . in the first case where @xmath8 is moving away
, @xmath7 falls into the potential well of @xmath8 and so long as it is approaching the point of minimum distance it gains kinetic energy . after passing this point
, @xmath7 starts climbing up the potential well and pays back the gained kinetic energy by restoring it into the potential energy of the two - body system .
however , on the way out of the well it climbs a well that is effectively smaller than the one it fell into as @xmath8 is on average farther away from it .
therefore , in the case that the big mass is moving _ transversely away _ from the particle , the small particle emerges with a _ gain _ in the kinetic energy i.e. @xmath12 .
the exact opposite effect holds in the second case , where @xmath8 is moving towards @xmath7 . in this case , @xmath7 enters the potential well set up by @xmath8 and , as in the previous case , gains kinetic energy so long as it is approaching the minimum distance between the two masses .
however , on the way out of the well it faces a more demanding climb as @xmath8 is on average closer to it and the potential well is steeper and deeper than before .
therefore , in the case that the big mass is _ transversely approaching _ the particle ( figure [ fig : nonlinearity ] ) , the small particle emerges with a _ loss _ in the kinetic energy i.e. @xmath13 .
the point however is that the two cases are not symmetric .
_ the decreasing of the magnitude of the force with distance breaks the symmetry between the two cases_. this is shown in figure [ fig : nonlinearity ] : in an attractive force field , @xmath7 has a greater loss ( in magnitude ) of energy when @xmath8 approaches it than a gain ( in magnitude ) when @xmath8 recedes away from it .
this asymmetry , deduced from dynamical principles , has consequences for the statistical mechanics of @xmath7 that we will explore . a plot of potentials of type @xmath2 .
@xmath14 , originally sitting at @xmath15 , is the central force that @xmath7 scatters from .
when @xmath8 moves to the right ( approaches @xmath7 ) by @xmath16 , it will set up a potential that is @xmath17 deeper for @xmath7 to climb out of than the one it fell into , whereby @xmath7 loses @xmath17 of energy to the central force as the result of the scattering .
whereas , when @xmath8 is moving to the left ( away from @xmath7 ) by the same amount , @xmath16 , its potential is @xmath18 shallower and consequently @xmath7 gains @xmath18 of energy .
note that the nonlinear dependence of @xmath19 , on @xmath1 makes the magnitude of the energy transfer in the two cases asymmetric : @xmath20 . ] in repulsive force fields , @xmath6 , the potential in figure [ fig : nonlinearity ] flips about the horizontal axis .
therefore the phenomenology is the exact opposite .
namely , the particle has a larger gain than loss .
so far we have described a purely dynamical phenomena where @xmath8 collides softly and transversely with @xmath7 .
clearly if a very fast small particle ( e.g. , an electron ) zips through a dilute soup of big masses ( e.g , massive ions or stars in a galaxy ) that randomly either approach it or move away from it in the manner described above , the small mass statistically loses ( gains ) energy to ( from ) the big masses when the force fields are attractive ( repulsive ) .
the fact that a high speed particle with low energy can lose or gain energy by scattering from larger masses with higher energies is not new . in the gravity
assist or the slingshot effect , a satellite can be carefully steered so that , by elastically scattering from a planet or a star that is moving along or opposite to the direction of its asymptotic motion ( @xmath21 ) , it acquires a boost in energy or brakes to save fuel @xcite .
the intuition in gravity assist , as in this work , is that to boost the energy of the satellite we have to throw it behind the approaching planet and vice versa @xcite .
slingshot effect , however , is not a net statistical effect resulting from an underlying random process .
we proceed to examine a scenario where the large masses move in random directions .
[ sec : numerical - verification]_a numerical experiment- _ we simulate a process where @xmath22 , @xmath23 , and @xmath24 .
the initial conditions for @xmath7 and initial position of @xmath8 are taken to be the same in all the simulations ( see appendix [ sec : appendix :- numerical - details ] for the numerical details ) . to single out the effect of the _ direction _ of the motion of the large mass on the statistical energetics of @xmath7 , we take the velocity of @xmath8 to be : @xmath25 and @xmath26 , where @xmath27 is kept constant , @xmath28 and @xmath29 is uniformly distributed on a circle @xmath30 .
we find that @xmath7 on average loses energy to @xmath8 ( figure [ fig : numerics ] ) .
it is instructive to look at the points corresponding to the two extreme cases discussed in the previous section . when @xmath8 moves towards the trajectory of @xmath7 along the path of minimum approach ( @xmath31 ) , maximum loss is seen ; furthermore , when @xmath8 veers away from the path of minimum distance ( @xmath32 ) , the maximum gain is seen .
both are congruent with our expectations . in figure
[ fig : numerics ] , the @xmath29-average of the change in the kinetic energy of @xmath7 is shown in red , which is seen to be less than the kinetic energy of @xmath7 when @xmath8 is pinned down ( i.e. @xmath33 for all times ) . in the latter case ,
used as a calibration here , the asymptotic kinetic energy of @xmath7 is conserved ( shown with a black line in figure [ fig : numerics ] ) .
the black line corresponds to the case that @xmath8 is pinned down and @xmath7 , considered alone , undergoes a conservative elastic scattering : @xmath34 the red line corresponds to the @xmath35average of the asymptotic change in the kinetic energy of @xmath7 , with @xmath8 moving in twenty different directions chosen uniformly on a circle .
we find that @xmath36 .
error bars on the energies are of @xmath37 ; see the appendix . ] _ analytical derivation- _ we are considering an standard elastic collision between two particles @xcite .
let @xmath38 and @xmath39 be the velocities of the two bodies in the lab frame ; the velocities in the center of mass ( @xmath40 system denoted by the suffix @xmath41 ) are related to the velocities in the lab by @xmath42 , @xmath43 , where @xmath44 . in the @xmath40 system
the collision simply rotates the velocities , which remain opposite in direction and unchanged in magnitude .
if we denote by @xmath45 the unit vector in the direction of the velocity of @xmath7 after the collision , then the velocities of the two particles after the collision ( distinguished by primes ) are @xmath46 and @xmath47 . in order to obtain the final velocities in the lab frame , denoted by @xmath48 and @xmath49
, we add the velocity of the center of mass @xmath50 to @xmath51 and @xmath52 , @xmath53 no further information about the collision can be obtained from the laws of conservation of momentum and energy .
the direction of the unit vector @xmath45 depends on the particular law of interaction and positions during the collision .
we assume the massive particles are far enough from one another that a sequence of two - body scatterings would be adequate to describe the process @xcite .
the change in the kinetic energy of @xmath7 before and after a collision is @xmath54 which reads @xmath55\equiv\mu v\mathbf{v.}\mathbf{n}\label{eq : delke}\ ] ] where , @xmath56 is the reduced mass , @xmath57 and @xmath58 denote the unit vectors pointing in the direction of motion of @xmath7 before and after the collision in the @xmath40 frame and @xmath59 is their difference . before calculating eq .
[ eq : delke ] as a function of @xmath29 , once again , we look at the two special cases discussed above .
first consider an attractive force field .
suppose @xmath60 lies on the line of the minimum distance as shown in figure [ fig : general_scattering(a , b)]a .
clearly if @xmath8 recedes away from @xmath7 then @xmath60 points in the same direction as @xmath61 and the dot product on the right hand side of [ eq : delke ] evaluates to be positive , whereas if @xmath60 and @xmath61 point in opposite directions the right hand side is negative . in the case of repulsion
the signs would be the opposite ( see figure [ fig : general_scattering(a , b)]b ) .
unbounded trajectory in a ) an attractive force field and b ) repulsive force field , title="fig:"]unbounded trajectory in a ) an attractive force field and b ) repulsive force field , title="fig : " ] what we now show is that the two cases are not symmetric .
that is the kinetic energy loss ( gain ) in the approaching case is larger than the gain ( loss ) in the receding case for an(a ) attractive ( repulsive ) potential .
the reason is that if we keep the setting the same but _ only _ change the direction of @xmath60 , both @xmath62 and @xmath61 will change . in the attractive case ( figure [ fig : general_scattering(a , b)]a ) , if @xmath8 moves towards @xmath7 the angle between the asymptotes , @xmath63 , is smaller than it would be if @xmath8 moved away
therefore , @xmath64 is larger in the approaching case .
for very high speed encounters , @xmath65 , we can approximate @xmath62 to be the same in the two cases , therefore eq . [ eq : delke ] becomes : @xmath66 we next consider the case where @xmath8 moves in _
random directions_. we do so by applying the above to the cases of gravitational and coulomb interactions , where the equations of motion are particularly simple to integrate @xcite and analytical results can be compared to the numerical experiments of the previous section .
we first parametrize : @xmath67 which is equal to @xmath60 in the limit where @xmath14 .
the orbits of scatterings in @xmath68 ( here @xmath3 is taken to be positive ) are hyperbolas parametrized by @xcite : @xmath69 , @xmath70 , @xmath71 and @xmath72 , where @xmath73 ( hence time ) varies from @xmath74 to @xmath75 and @xmath76 we are interested in the change in the kinetic energy of the smaller mass in the lab frame @xcite . since , @xmath77 ; we have for @xmath78 , which reads @xmath79 . here
@xmath80 and @xmath81 corresponds to the approaching case . at first sight
it seems like averaging over @xmath29 would yield zero but as mentioned above , the eccentricity @xmath82 , the semi - axis @xmath83 and consequently @xmath84 depend on the initial velocities as they are functions of the angular momentum and the energy in the center of mass frame .
we can rewrite the formula by noting that @xmath85 and @xmath86 .
for small angle scatterings , @xmath87 , we have @xmath88 , where @xmath89 is the relative velocity before collision and is a function of the initial angle the two velocities make with one another .
therefore the previous equation can be rewritten as @xmath90 with @xmath91 being the impact parameter , which also depends on the initial angle between the velocity of the two masses .
the strength of interaction for gravitational and coulomb attractions are @xmath92 and @xmath93 respectively .
for the two cases discussed above we have for the approaching case : @xmath81 and @xmath94 . and for the receding case @xmath95 and @xmath96 .
it is easy to see that @xmath97 ; therefore we yet again arrive at the same conclusion . in figure
[ fig : theoretical - prediction ] we plot eq . [ eq : final ] as a function of @xmath29 and see that our theoretical treatment and numerical experiment figure [ fig : numerics ] agree .
theoretical prediction of a net loss of energy for attractive fields agrees with the numerical experiment . here
, @xmath29 was shifted by @xmath98 to be comparable to the experiment shown in figure [ fig : numerics ] . as before @xmath99 corresponds to the approaching case and @xmath100 to the receding case .
the small discrepancy between the theoretical and experimental values of @xmath101 is due to the finite box size in the experiments . ]
a formal statistical treatment depends on the particular laws of interaction as well as the statistical distribution of the large masses .
for @xmath102 it remains to show that under the assumption of microscopic chaos in the direction of motion of the large masses , the effect holds for any generic choice of initial conditions .
equation [ eq : final ] can be written entirely in terms of initial conditions in the lab frame as ( see appendix [ sub : formulation - entirely_inlab ] ) @xmath103 where , @xmath104 and @xmath105 encode the dependence on the initial conditions . now let @xmath106 and @xmath107 denote the @xmath29- average , we have @xmath108 where @xmath109 encodes the dependence on the choice of initial conditions . figure [ fig : fvsz ] establishes the existence of the effect irrespective of particular choices of initial conditions in the scattering process as long as the assumptions leading up to eq .
[ eq : final ] are intact .
one can use mechanics of similarity to calculate @xmath110 when the law of interaction is coulombic @xcite .
if we fix the corresponding paths we can related the change in the kinetic energy of the two : @xmath111 , where we have explicitly labeled the quantities corresponding to the gravitational and the coulomb interactions by @xmath112 and @xmath113 respectively .
furthermore , the corresponding time of travel between two fixed points on the path are related by @xmath114 . consequently , for _ repulsive _ coulomb interaction our final result reads , @xmath115 where @xmath116 as before .
we see that the equation we get is the negative of eq .
[ eq : final ] which , as expected , corresponds to a net gain in energy when averaged over @xmath29 .
a calculation from first principles yields the same result ( see eq.[eq : finalrep ] in appendix [ sec : appendix :- coulomb scattering ] ) .
plot of @xmath109 vs. @xmath117 given by eq .
[ eq : f ] . ] _ statistical context- _ a phenomena worth considering is dynamical friction @xcite . dynamical friction , however , is like brownian motion @xcite as a big mass enters a medium of many smaller particles and slows down as a result .
but it is distinct from brownian motion as the interactions are long ranged .
it is found that in dynamical friction `` only stars with velocities less than the one under consideration contribute to the effect '' @xcite and @xcite .
the main requirement in our work is that the particle scatters from the _ time - dependent _ field _ _ of other massive particles , that are moving randomly and slowly in space , through a series of two - body scatterings .
in addition , it is required that during the effective scattering the scatterer remains on one side of the scattering particle .
this is equivalent to the assumption of the slow motion , in the dilute medium , of the large masses
. it would be interesting to analyze the effect of micro - dynamics of the structure in the universe on the frequency shift of photons coming from distant sources .
we expect a small loss of energy for photons that undergo weak lensing @xcite in the interstellar media of galaxies , clusters and super clusters @xcite .
clearly the potential fluctuations , derived within this framework , in energy of the photons will be due to the dynamics of the lenses and disjoint from the essential time dependence of large scale space - time metric in the general theory of relativity as in sachs - wolfe and rees - sciama effects @xcite . at first sight
the effect we are proposing seems to violate the equipartition of energy because a smaller and lighter particle _ `` heats up '' _ the medium of much larger particles that have higher average energies .
but note that we are working with a non - equilibrium process in an open system @xcite .
in attractive potentials the small particle starts from non - equilibrium initial conditions and through a series of scatterings it statistically loses energy till a final scattering where the energy in the center of mass is negative . there on the small particle would have a bounded orbit about that final scatterer .
this corresponds to the breakdown of small angle scattering assumption of this paper . in plasma physics
this is known as shielding and in astrophysics it corresponds to capturing of a comet by a center of force .
in repulsive potentials the energy of the small particle grows till the relativistic effects become significant and the transfer of energy between the small particle and the scatterers becomes of order unity with respect to the initial energy ( * ? ? ?
* the second to last paragraph of section 13 ) .
hence the effect is not an equilibration process and is applicable to systems where the assumptions of small angle scattering , as well as , @xmath9 , @xmath10 but nevertheless @xmath11 hold .
the former assumption is bound to break on time scales required for relaxation to equilibrium .
it is our hope that this small effect helps with better understanding of rapid structure formation @xcite , high energy cosmic rays , fermi acceleration @xcite and redshift problem .
i thank peter w. shor , richard v. lovelace , jeffrey goldstone and jack wisdom for discussions and @xcite .
i also like to thank mehran kardar , frank wilczek , otto e. rossler , reinhard nesper and eduardo cuervo - reyes .
this work was partly supported by the national science foundation through grant number ccf-0829421 .
10 there are other potentials , like the yukawa potential @xmath118 or @xmath119 that are not of eq .
1 type but nevertheless can be included in this work as they have similar qualitative nonlinearities in their potential .
rossler , o.e .
, movassagh , r. , journal of nonlinear sciences and numerical simulation 6(4 ) , 337 - 338 ( 2005 ) @xmath120 , where @xmath121 is the dimension of the space is considered to be long range .
l.d .
landau and e.m lifshitz , `` statistical mechanics '' third edition , section 74 ( 1980 ) .
douglas heggie , piet hut , `` the gravitational million - body problem : a multidisciplinary approach to star cluster dynamics '' , cambridge university press ; first edition ( february 24 , 2003 ) fermi , e. , physical review , vol 75 , no . 8 ( 1949 ) johnson , r.c . , `` the slingshot effect '' , http://www.dur.ac.uk/bob.johnson/sl/ ( 2003 ) bartlett , a.a . ,
hord , c.w . , the physics teacher , pp .
466 - 473 nov .
( 1985 ) landau , l.d . ,
lifshitz , e.m .
, `` mechanics '' ( 1976 ) chandrasekhar , s. , `` principles of stellar dynamics '' .
dover phoenix edition goldstein , h. , poole , c.p . ,
safko , j.l .
, `` classical mechanics '' third edition . pp . 88 - 89 ( 2002 )
. orbits corresponding to other power law potentials can be expressed in terms of hypergeometric functions . according to this reference
the orbits corresponding to potentials with @xmath122 can be integrated in terms of trigonometric functions and @xmath123 in terms of elliptic functions .
the equations can be written in more general form as in equation 2.416 in @xcite , where the orbital and the fundamental plane make a nonzero angle with one another .
einstein , a. , annalen der physik 17 : 549560 ( 1905 ) further the assumption of homogeneity can break down if the interactions are not short ranged .
see goldenfeld , n. , `` lectures on phase transitions and the renormalization group '' westview press ; illustrated edition .
( 1992 ) chandrasekhar , s. , , 97 : 255262 ( 1943 ) chandrasekhar , s. , , 97 : 263273 ( 1943 ) chandrasekhar s. , annals of new york academy of sciences , volume xlv ( 1943 ) sachs , r.k . , wolfe , a.m. , , vol .
147 , p. 73
( 1967 ) rees , m.j . ,
sciama , d.w . ,
nature , vol .
217 , feb . 10
( 1968 ) weinberg , s. , `` cosmology '' , oxford university press , usa ( 2008 ) wittman , d. m. , tyson , j. a. , kirkman , d. , dellantonio , i. , bernstein , g. , nature vol 405 , issue 6783 , pp .
143 - 148 ( 2000 ) movassagh r. , in preparation .
l.d .
landau and e.m .
lifshitz , vol ii , the classical theory of fields , 4th extended english edition .
a few of these are succinctly summarized by peebles , p.j.e . and nusser , adi , in nature vol 465 , pp .
565- 569 ( 2010 ) a numerical package in c++ with a standard adaptive step size bulirschstoer algorithm built into it was provided to me by jack wisdom . |
interaction and collisional properties between matter and antimatter is of fundamental importance in physics @xcite .
the anti - hydrogen atom @xmath9 , which is a bound state of an anti - proton @xmath10 and a positron @xmath11 , is the simplest representative of an antimatter atom .
this is a two - particle system , which , however , may possess very different interactional and dynamical properties compared to its matter counterpart the h atom @xcite . by now much effort has been exerted in various experiments to build and store @xmath9 at cold and ultra - cold temperatures @xcite .
new experiments are planned or in progress to test the fundamental laws and theories of physics involving antiparticles and antimatter in general @xcite .
for example , it follows from the cpt symmetry of quantum electrodynamics that a charged particle and its anti - particle counterpart should have equal and opposite charges and equal masses , lifetimes and gyromagnetic ratios .
the cpt symmetry predicts that hydrogen and anti - hydrogen atoms should have identical spectra . in this way
experimentalists plan to test whether in fact h and @xmath9 have such properties .
specifically , a starting point would be to compare the frequency of the 1s-2s two - photon transition in h and @xmath9 .
also , one of the important practical applications of antihydrogen has been mentioned in @xcite , where the authors considered controlled @xmath9 propulsion for nasa s future plans in very deep space .
researchers at cern @xcite and from other groups @xcite are interested to trap and study @xmath9 at low temperatures , e.g. , @xmath12 k , when the @xmath9 atom will be almost in its rest frame .
the study of lamb shift and response of antihydrogen to gravity at ultra - low energies should allow them to test more precisely the predictions of two fundamental theories of modern physics : quantum field theory and einstein s general theory of relativity @xcite .
it has been pointed out that the main cause of loss of @xmath9 atoms confined in a magnetic gradient trap is due to @xmath13h@xmath0 and @xmath13he collisions .
therefore , the @xmath13h@xmath0 scattering cross - sections and corresponding rotational - vibrational thermal rate coefficients , in the case of h@xmath0 , would be very helpful to gain a practical understanding of the slowing down and trapping of @xmath9 .
hence the investigation of the possibility of cooling of @xmath9 atoms by colliding them with colder h@xmath0 is of significant practical interest @xcite .
( similar collision between trapped fermionic atoms with cold bosonic atoms has been fundamental in cooling the fermionic atoms and thus leading them to quantum degeneracy @xcite . )
such investigation of @xmath9 interaction with h and h@xmath0 can reveal the survival conditions of @xmath14 in collisions with h and , even more importantly , with h@xmath0 @xcite .
further , cooling occurs by energy transfer in elastic collisions of @xmath9 with h@xmath0 .
however , during the collision the rearrangement process may lead to the formation of protonium ( @xmath15 ) and positronium ( @xmath16 ) exotic atoms and the destruction of @xmath14 atoms .
they are formed as matter - antimatter bound states , which then annihilate .
( there have also been many studies of scattering of positronium atoms @xcite , the lightest matter - anti - matter atom ) .
thus one can conclude , that the effectiveness of cooling @xmath9 is determined from a comparison of the cross sections for direct scattering and rearrangement . by now a series of theoretical works have been published , in which the properties of interaction between @xmath9 and h , he , h@xmath0 have been investigated @xcite .
some theoretical studies have been carried out for the @xmath14+h system at thermal energies using quantum - mechanical methods @xcite .
also , discussions on the importance and applications for this system , especially , in connection with bose - einstein condensation @xcite , ultra - cold collisions @xcite , and its static and dynamic properties @xcite , can be found in the literature . in this work we present results for the collision of an ultra - cold @xmath9 atom with h@xmath0 , where h@xmath0 is treated as a rigid rotor with a fixed distance between hydrogen atoms .
the elastic , rotational state - resolved scattering cross sections for the @xmath9-h@xmath0 scattering and their corresponding thermal rate coefficients are calculated using a non - reactive quantum - mechanical close - coupling approach .
the potential interaction between @xmath9 and the hydrogen atoms is taken from ref .
@xcite . in the next section we present the quantum - mechanical formalism used in this work .
the results and discussion are presented in sec .
conclusions are given in sec .
in this section we describe the close - coupling quantum - mechanical approach we used to calculate the cross sections and collision rates of a hydrogen molecule h@xmath0 with an anti - hydrogen atom @xmath9 . atomic units ( @xmath17 )
are used in this section , where @xmath18 and @xmath19 are charge and mass of an electron .
three - body jacobi coordinates @xmath20 for the @xmath9+h@xmath21 system used in this work are shown in fig . 1 .
the two h atoms are labeled 2 and 3 and the @xmath22 atom is labeled 1 , @xmath23 is the center of mass of the h@xmath0 molecule , @xmath24 is the polar angle between vector @xmath25 connecting the two h atoms in h@xmath0 ( labeled 2 and 3 ) and vector @xmath26 connecting the center of mass of the h@xmath0 molecule to the @xmath22 atom ( labeled 1 ) .
next , @xmath27 and @xmath28 are angular momenta corresponding to the vectors @xmath25 and @xmath26 , respectively .
the quantities @xmath29 and @xmath30 are the distances between the @xmath22 atom labeled 1 and the h atoms labeled 2 and 3 , respectively .
the schrdinger equation for an @xmath31 collision in the center of mass frame , where @xmath32 @xmath33 is an atom and @xmath34 @xmath35 is a linear rigid rotor , is @xcite @xmath36 where @xmath37 is the relative momentum between @xmath32 and @xmath34 , @xmath38 is the reduced mass of the atom - molecule ( rigid rotor in this model ) system @xmath31 : @xmath39 , @xmath40 is the reduced mass of the target : @xmath41 , @xmath42 is the angle of orientation of the rotor @xmath43 , @xmath44 is the potential energy surface ( pes ) for the three - atom system @xmath45 , and @xmath46 is the total energy of the system .
the eigenfunctions of the operator @xmath47 in eq .
( [ eq : schred ] ) are the spherical harmonics @xmath48 . to solve eq .
( [ eq : schred ] ) , the following expansion is used @xcite @xmath49 where channel expansion functions are @xmath50 here @xmath51 is the total angular momentum of the system @xmath45 , and @xmath52 is its projection onto the space fixed @xmath53 axis , @xmath54 and @xmath55 are projections of @xmath5 and @xmath56 respectively , @xmath57 are the clebsch - gordan coefficients , and @xmath58 s are the appropriate radial functions .
substitution of eq .
( [ eq : expan ] ) into eq .
( [ eq : schred ] ) provides a set of coupled second order differential equations for the unknown radial functions @xmath59 @xmath60 to solve the coupled radial equations ( [ eq : cpld ] ) , we apply the hybrid modified log - derivative - airy propagator in the general purpose scattering program molscat @xcite .
additionally , we tested other propagator schemes included in molscat .
our calculations reveal that other propagators can also produce quite stable results .
the log - derivative matrix is propagated to large intermolecular distances @xmath61 , since all experimentally observable quantum information about the collision is contained in the asymptotic behavior of functions @xmath62 .
the numerical results are matched to the known asymptotic behavior of @xmath59 relating to the the physical scattering @xmath63-matrix @xcite @xmath64 where @xmath65^{1/2}$ ] is the channel wave - number of channel @xmath66 , @xmath67 is rotational channel energy and @xmath46 is the total energy in the @xmath45 system .
this method was used for each partial wave until a converged cross section was obtained .
it was verified that the results have converged with respect to the number of partial waves as well as the matching radius , @xmath68 , for all channels included in our calculations .
cross sections for rotational excitation and relaxation phenomena can be obtained directly from the @xmath63-matrix . in particular , the cross sections for excitation from @xmath69 summed over the final @xmath70 and averaged over the initial @xmath71 are given by @xcite @xmath72 the kinetic energy is @xmath73 , where @xmath74 is the rotation constant of the rigid rotor @xmath34 , i.e. the hydrogen molecule .
the relationship between the rate coefficient @xmath75 and the corresponding cross section @xmath76 can be obtained through the following weighted average @xcite @xmath77 where @xmath78 is pre - collisional translational energy at temperature @xmath79 , @xmath80 is boltzman constant , and @xmath81 is the minimum value of the kinetic energy needed to make @xmath82 levels accessible . in the following section
, we will present our results for rotational quantum transitions in collision between h@xmath0 and an anti - hydrogen atom @xmath22 , that is @xmath84 here h@xmath0 is treated as a vibrationally averaged rigid monomer rotor .
the bond length was fixed at 1.449 a.u . or 0.7668 .
the rotation constant of the h@xmath0 molecule has been taken as @xmath85@xmath86 .
the h@xmath0 rigid rotor model has been already applied in different publications @xcite . for the considered range of kinetic energies
the model can be quite justified in this special case when only pure rotational quantum transitions at low collisional energies are of interest as in h@xmath0(@xmath5)+@xmath22 , and when the energy gap between rotational and vibrational energies is much larger than kinetic energy of collision .
in such a model the quantum mechanical approach is rather simplified .
next we consider an important physical parameter in atomic and molecular collisions , e. g. , the pes between the atoms .
there is no global potential energy surface available for the three - atom @xmath22-h@xmath0 system .
however in ref .
@xcite , the author calculated the values of interaction energy between h and @xmath22 , i. e. , the h-@xmath22 energy curve using the rayleigh - ritz variational method .
further , the microhartree accuracy of born - oppenheimer energies of the system has been achieved in that work . to construct the h@xmath0-@xmath22 interaction potential
we take the h-@xmath22 energy data from ref .
@xcite and make a cubic spline interpolation through all 46 points taken from table 1 of that paper .
these data have been tabulated from @xmath87 a.u .
to @xmath88 a.u
. inter - atomic distances .
because in the current work we use the rigid rotor model for h@xmath0 , we do not need the interaction energy between hydrogen atoms in h@xmath0 .
the interaction potential between a hydrogen molecule and @xmath22 is taken by sandwiching the two @xmath22-h potential energy curves : @xmath89 where distances between atoms are written as follow ( cf .
1 ) : @xmath90 the functions @xmath91 with @xmath92 are represented as cubic spline interpolation functions for any value of @xmath93 or @xmath94 as follows : @xmath95 where @xmath96 , @xmath97 , @xmath98 perform the spline interpolation and where @xmath99 , in each sub - interval @xmath100 $ ] , @xmath101 . the coordinates @xmath102 and corresponding values of the h-@xmath103 potential energy data have been taken from table 1 of ref .
@xcite .
the calculated pes is shown in fig .
it is clear that the potential has a singular value when the distance between @xmath22 and h@xmath0 is equal to zero .
additionally , the @xmath104-h@xmath0 pes which we obtain from eq .
( [ eq : pesmodel ] ) is shown in fig .
specifically , this potential was used in our calculations of @xmath22 + h@xmath0 collisions . again , as seen in fig .
2 the potential energy curve between @xmath22-h has a coulomb type singularity at small distances . in our calculations we needed to make additional test runs to achieve convergence in our results . in the next section
we will briefly demonstrate the numerical convergence of the results when calculating total elastic scattering cross sections .
these results depend on various numerical and quantum - mechanical scattering parameters .
numerous test calculations have been undertaken to insure the convergence of the results with respect to all parameters that enter in the propagation of the schrdinger equation .
these include the atomic - molecular distance @xmath61 , the total angular momentum @xmath105 , the number of total rotational levels to be included in the close - coupling expansion and others , see fig . 1 .
particular attention has been given to the total number of numerical steps in the propagation over the distance @xmath61 of the schrdinger equation ( [ eq : cpld ] ) . specifically , the parameter r ranges from 0.75 a.u . to 20.0 a.u .
we used up to 50000 propagation points .
we also applied and tested different mathematical propagation schemes included in molscat .
the rotational energy levels of @xmath106-h@xmath21 and the corresponding angular momenta @xmath5 are shown in table i. the goal of this work is to get new results for @xmath9 + @xmath106-h@xmath0 thermal rate coefficients @xmath75 at ultra - low temperatures : specifically @xmath107 k@xmath108 k. the corresponding cross sections have been calculated for collision energies varying from @xmath109 @xmath86 to @xmath110 @xmath86 .
these energy values are very small . however , despite this fact , to reach convergence of the results we needed to include in expansion ( [ eq : expan ] ) a significant number of rotational levels of the h@xmath0 molecule , specifically up to @xmath111 . below in table
ii we present these results . also , we found that , at lower energies , for the numerical solution of eq .
( 4 ) a much larger number of propagation ( integration ) points are needed than at higher energies .
specifically , at higher energies we need 500 propagation points , but for lower energies 50,000 points are needed to achieve comparable precision .
convergence has been achieved for elastic scattering cross sections for various scattering parameters .
below in table iii we present these results .
then we used this data in our calculation for rotational energy transfer , elastic scattering cross section and thermal rate coefficient . now in table ii we present results for the total elastic cross sections for two collisional energies : 0.1 @xmath86 and 0.01 @xmath86 the cross sections are shown for a number of different maximum values of the rotational angular momentum @xmath112 in the h@xmath0 molecule included in expansion ( [ eq : expan ] ) .
this is also the jmax parameter in molscat @xcite .
other scattering parameters have also been treated correctly in the calculation .
one can see that jmax should be at least 56 .
the other scattering parameter in table ii is mxsym @xcite .
it reflects the number of terms in the potential expansion over angular functions @xcite .
it is therefore evident that in this calculation we need to keep at least 24 terms in the expansion . in table iii
we also present results for total elastic scattering cross sections for few more selected energies .
however in this table the convergence has been reached by increasing the total angular momentum @xmath105 and by increasing the total number of the propagation steps in the propagation over the coordinate @xmath61 of the schdinger equation ( [ eq : cpld ] ) .
as expected , for lower energy collision we needed smaller values for the maximum @xmath105 .
for example , for collision energy @xmath113 a.u .
it is enough to have @xmath114 , however @xmath115 should be taken for @xmath116 a.u . in regard to the total number of the propagation steps in the solution of eq .
( [ eq : cpld ] ) one can see , that we need to include many more propagation points for low energy calculations .
all test calculations in table iii have been done with jmax = 56 and mxsym = 24 .
the obtained results concerning the numerical and scattering parameters have been used in our calculation for total elastic @xmath117 and rotational quantum state transfer @xmath118 cross sections and corresponding thermal rate coefficients @xmath75 .
now we present computational results for process ( [ eq : proc1 ] ) , namely , for elastic scattering ( @xmath119 ) and for low quantum number rotational transition between levels with @xmath120 and 4 : @xmath121 , @xmath122 , @xmath123 , and @xmath124 . from the results of table iii we see that to reach numerical convergence , for example , for the elastic scattering cross section , we need to include a large number of h@xmath0 rotational levels , specifically up to 60 .
the results for the elastic scattering cross sections @xmath2 for @xmath9 + h@xmath0 @xmath125 h@xmath0 + @xmath9 are shown in fig . 4 together with the corresponding results of variational calculations of gregory and armour @xcite .
it can be seen , that basically the two sets of cross sections are close to each other , although in our calculation we use larger number of collision energy points , specifically up to 200 . in our calculation
a shape resonance is found at energy @xmath126 hartree . as in ref .
@xcite our @xmath2 tends to reach a constant value at lower energies with @xmath127 .
this result allows us to calculate the @xmath22+h@xmath0 scattering length , which is @xmath128 the gregory - armour scattering length @xcite obtained with a variational method is @xmath129 .
the two results are in reasonable agreement with each other . in fig . 5 ( a ) and ( b ) we show the total state - resolved cross sections @xmath130(v ) vs. velocity v and the corresponding thermal rate coefficients @xmath131(t ) vs. temperature t for the hydrogen molecule rotational relaxation process
it is seen , that when the collision energy increases the de - excitation cross section decreases .
it can be explained in the following way : at low relative velocities ( kinetic energies ) between h@xmath132 and @xmath22 , the rotationally excited h@xmath0 molecule has more time for interaction and consequently , it has higher quantum - mechanical probability to release its internal rotational energy to @xmath22 .
the resulting corresponding rate coefficients have been calculated for a temperature range from @xmath133 and are also presented in fig . 5 below the cross section results .
next , in fig . 6 ( a ) and ( b ) we present results for the total state - resolved cross sections @xmath134(v ) vs. velocity v and the corresponding thermal rate coefficients @xmath135(t ) vs. temperature t for the hydrogen molecule rotational excitation process .
it is quite understandable , as we find from fig .
6 ( a ) , that when the collision energy ( relative velocity v ) is increases the quantum - mechanical probability and corresponding cross section of the rotational excitation of h@xmath21 are also increases . figure 6 ( b ) depicts the corresponding results for the thermal rate coefficient . in figs . 7 ( a ) and ( b ) we present results for cross sections and rates for rotational relaxation process , as in figs . 5 ( a ) and ( b ) , but now connecting the states @xmath136 and @xmath137 . finally , in figs . 8 ( a ) and
( b ) we present results for cross sections and rates for rotational relaxation process but now connecting the states @xmath136 and @xmath138 .
an unexpected result has been found in fig . 7 ( a ) in the rotational transition cross section @xmath139 ,
i.e. , when the h@xmath0 quantum angular momentum has been changed from @xmath136 to @xmath137 .
one can see , that the values of these cross sections at very low collision energies are almost from 5 to 10 times larger then other cross sections considered in this work , compare with the results from figs . 5 , 6 and 8 .
a quantum - mechanical study of the state - resolved rotational relaxation and excitation cross sections and thermal rate coefficients in ultra - cold collisions between hydrogen molecules h@xmath0 and anti - hydrogen atoms @xmath22 has been the subject of this work .
a model pes for h@xmath0-@xmath22 has been constructed by sandwiching two h-@xmath22 interaction potentials for two different hydrogen atoms taken from ref .
this h-@xmath22 interaction potential is shown in fig . 2 .
the h@xmath0-@xmath22 pes is presented in fig .
calculation for total elastic scattering cross section and for low quantum rotational transition states have been performed .
we considered only the following quantum transitions : 2@xmath1250 , 0@xmath1252 , 4@xmath1252 , and 4@xmath1250 .
a test of the numerical convergence was undertaken .
these results are presented in tables ii and iii .
our results reveal that it is necessary to set the rotational angular momentum @xmath140 in the h@xmath0 molecule to a relatively large number , i.e. in the expansion ( [ eq : expan ] ) we needed to include up to 60 terms .
the calculation was performed using the molscat program @xcite .
different propagation schemes included in the molscat program have been used and tested .
additionally , the mxsym potential parameter in that program also needed to have a relatively large value to obtain good convergence as can be seen from table ii .
the numerical convergence has also been tested over the total number of the propagation steps over coordinate @xmath61 in the solution of the schrdinger equation ( [ eq : cpld ] ) .
we have found , that at low energies we need a much larger number of integration points than at higher energies , cf .
table iii .
our results for the h@xmath21+@xmath22 total elastic scattering cross section are in reasonable agreement with the corresponding results from gregory and armour @xcite .
the authors of this paper used a different pes , which is still unpublished , and applied a quantum - mechanical variational approach .
unfortunately , the rotational transitions in the h@xmath21+@xmath22 collisions have not been calculated in that work @xcite .
one of the interesting results of the present work is that the cross section of the rotational transitions from h@xmath141 at ultra - low energies are approximately 5 - 10 times larger than other transition state cross sections . to the best of our knowledge we do not know of any other calculation of the rotational transitions in the h@xmath21+@xmath22 collision .
these results can help to model energy transfer processes in the hydrogen - anti - hydrogen plasma , and perhaps to design new experiments in the field of the anti - hydrogen physics .
finally , we believe , that in the future work it should be useful to include vibrational degrees of freedom of the h@xmath0 molecules , i.e. to carry out quantum - mechanical calculations for different rotational - vibrational relaxation processes : @xmath142 where @xmath143 and @xmath144 are the vibrational quantum numbers of h@xmath0 before and after the collision , respectively . p. k. biswas and s. k. adhikari , chem .
phys . lett . *
317 * , 129 ( 2000 ) ; j. phys .
b * 33 * , 1575 ( 2000 ) ; * 31 * , l315 ( 1998 ) ; s. k. adhikari and p. k. biswas , phys . rev .
a * 59 * , 2058 ( 1999 ) ; s. k. adhikari , p. k. biswas and r. a. sultanov , _ ibid .
_ * 59 * , 4829 ( 1999 ) .
l. labzowsky , d. solovyev , v. sharipov , g. plunien , and g. soff , j. phys .
b 36 , l227 ( 2003 ) ; c. y. hu and d. caballero , _ ibid . _ 35 , 3879 ( 2002 ) ; e. cubero , m. orozco , p. hobza , and f. j. luque , j. phys . chem . a 103 , 6394 ( 1999 ) ; p. hobza , v. spirko , z. havlas , k. buchhold , b. reimann , h. d. barth , and b. brutschy , chem .
299 , 180 ( 1999 ) ; p. hobza and z. havlas , _ ibid .
_ 303 , 447 ( 1999 ) .
a.m. arthurs and a. dalgarno , proc .
soc . a 256 , 540 ( 1963 ) .
sultanov , d. guster , chem . phys . 326 , 641 ( 2006 ) .
( note that in eq .
( 1 ) of this work the momentum operators should be squared and the equation should look as follows : @xmath145 as eq .
( 2 ) in @xcite .
lcccccccc @xmath46 ( @xmath86 ) & + & & + & 0 & 2 & - & & 500 & 1000 & 10000 & 50000 + 0.1 & 1.59(3)&1.59(3 ) & & & 3.64(2)&2.67 & 1.60(3)&1.59(3 ) + 0.01 & 6.62(3)&6.62(3 ) & & & 3.62(2)&1.75(-1)&6.63(3)&6.62(3 ) + + & 4 & 6 & 8 & & 500 & 1000 & 5000 & 7000 + 10.0 & 5.85(1)&5.96(1)&5.96(1 ) & & 6.23(1)&6.00(1)&5.96(1)&5.96(1 ) + 1.0 & 1.69(2)&1.69(2 ) & - & & 1.77(2)&1.70(2)&1.70(2)&1.70(2 ) + + & 8 & 10 & 12 & & 500 & 750 & - & - + 100.0 & 1.65(2)&1.72(2)&1.72(2 ) & & 1.72(2)&1.72(2 ) & - & - vs. velocity @xmath148 .
lower plot ( b ) : corresponding thermal rate coefficients @xmath149 vs. temperature @xmath150 for the hydrogen molecule rotational relaxation process h@xmath151 in @xmath9-h@xmath0 collision.,title="fig : " ] + vs. velocity @xmath148 .
lower plot ( b ) : corresponding thermal rate coefficients @xmath149 vs. temperature @xmath150 for the hydrogen molecule rotational relaxation process h@xmath151 in @xmath9-h@xmath0 collision.,title="fig : " ] vs. velocity v. lower plot ( b ) : corresponding thermal rate coefficients @xmath152 vs. temperature t for the hydrogen molecule rotational excitation process h@xmath153 in @xmath9-h@xmath0 collision.,title="fig : " ] + vs. velocity v. lower plot ( b ) : corresponding thermal rate coefficients @xmath152 vs. temperature t for the hydrogen molecule rotational excitation process h@xmath153 in @xmath9-h@xmath0 collision.,title="fig : " ] vs. velocity @xmath148 .
lower plot ( b ) : corresponding thermal rate coefficients @xmath154 vs. temperature @xmath150 for the hydrogen molecule rotational relaxation process h@xmath155 in @xmath9-h@xmath0 collision.,title="fig : " ] + vs. velocity @xmath148 . lower plot ( b ) : corresponding thermal rate coefficients @xmath154 vs. temperature @xmath150 for the hydrogen molecule rotational relaxation process h@xmath155 in @xmath9-h@xmath0 collision.,title="fig : " ] vs. velocity @xmath148 .
lower plot ( b ) : corresponding thermal rate coefficients @xmath156 vs. temperature @xmath150 for the hydrogen molecule rotational relaxation process h@xmath157 in @xmath9-h@xmath0 collision.,title="fig : " ] + vs. velocity @xmath148 .
lower plot ( b ) : corresponding thermal rate coefficients @xmath156 vs. temperature @xmath150 for the hydrogen molecule rotational relaxation process h@xmath157 in @xmath9-h@xmath0 collision.,title="fig : " ] | in this work we report the results concerning calculations for quantum - mechanical rotational transitions in molecular hydrogen , h@xmath0 , induced by an ultra - cold ground state anti - hydrogen atom @xmath1 .
the calculations are accomplished using a non - reactive close - coupling quantum - mechanical approach .
the h@xmath0 molecule is treated as a rigid rotor .
the total elastic scattering cross section @xmath2 at energy @xmath3 , state - resolved rotational transition cross sections @xmath4 between states @xmath5 and @xmath6 and corresponding thermal rate coefficients @xmath7 are computed in the temperature range 0.004 k @xmath8 4 k. satisfactory agreement with other calculations ( variational ) has been obtained for @xmath2 . |
one can perfectly well view @xmath3 decay as arising from the @xmath4 quark - level @xcite linear @xmath5 model ( l@xmath5 m ) instead of through the @xmath6 chiral anomaly @xcite .
in such a description it is straightforward to predict many other @xmath4 radiative decays through quark triangle graphs . the purpose of this paper is to investigate how well this quark loop picture accords with experimental data @xcite ; in this way we avoid discussing anomaly predictions of these processes and notions of strong cp violation with consequential ( low mass ) axions , since these effects have yet to be discovered in the laboratory .
we also consider the situation for rare @xmath4 decays that are associated with tiny quark mixings ( such as @xmath7 ) and with isospin - changing corrections ( such as @xmath8 ) . in section
ii we show that the dominant l@xmath5 m quark triangle @xmath4 predictions for @xmath9 , @xmath10 and @xmath11 all fit the data , within experimental errors ; the only parameters at our disposal are @xmath12 and @xmath13 , the couplings of the pseudoscalar and vector mesons respectively to the quarks , the constituent quark masses @xmath14 and a couple of well - known mixing angles .
we pursue this approach in section iii , for the rarer isospin - conserving interactions where the data gives valuable information about the small amount of admixture of light quarks in the heavy @xmath15 states , and vice versa ; here the agreement is less good , but the experimental information is also sparser and may be expected to change with time . in section iv , we show that a simple picture of @xmath16 meson - meson transitions , associated with the @xmath17 quark mass difference , nicely provides the rates for isospin changing decays .
the last section contains some preliminary work on three - body decays in which at last one pair of pions is produced in a p - state ; we show that @xmath18-meson poles essentially saturate the amplitudes for those processes , given the two - body amplitudes determined previously .
the reasonable agreement between naive theory and experiment suggests that this quark loop picture is a viable alternative to the field - theoretic anomaly - inspired evaluation of such amplitudes .
however the tantalizing picture which emerges is that internal mass dependence of the box diagrams is largely negated by the form factors which must be present at the meson - quark vertices .
such are our conclusions in section vi .
for the past thirty years , the @xmath19 decay has been understood as an @xmath6 chiral anomaly @xcite combined with the partial conservation of axial current ( pcac ) . writing this process and all other @xmath4 amplitudes in the conventional form @xmath20 , where @xmath21 and @xmath22 refer to vector particle momentum and polarization respectively ,
one derives @xmath23 with @xmath24 mev and a quark color number @xmath25 , the theoretical result is accurately consistent with data @xcite : @xmath26 the question then arises as to which theoretical models also predict the same result while simultaneously demanding that @xmath27 .
one such favored model is the quark - level @xcite linear @xmath5 model ( l@xmath5 m ) .
this model also provides @xmath28 and @xmath29 quark triangle graphs but with the meson interacting through the pseudoscalar @xmath30 vertex , rather than through the divergence of an axial vector .
no accompanying l@xmath5 m @xmath31 loop is allowed because the @xmath32 coupling vanishes ; consequently the @xmath6 anomaly result immediately follows in the spirit of steinberger @xcite .
( in carrying out the evaluation of @xmath33 , we go to the soft limit , because the pion is essentially massless , relative to other hadronic masses and even the constituent quarks . )
the fact that this l@xmath5 m also demands @xmath34 @xcite is tied to the lee null tadpole condition @xcite and ensures dynamical generation of masses , in contrast to the original nucleon - level l@xmath5 m @xcite .
the pion - quark coupling @xmath35 needed obeys the quark - level goldberger - treiman relation ( gtr ) @xmath36 where the constituent ( nonstrange ) quark mass @xmath37 is dynamically generated to the expected @xmath38 value in the chiral limit , since @xmath39 is required now .
our aim in this paper is to investigate how well the sigma model picture fits the data for @xmath4 processes .
we take the naive viewpoint that the @xmath4 amplitudes , just like @xmath40 can be estimated by working out @xmath41 in the limit of _ zero _ external momentum , even though this is a _ long way _ from the mass shell for the _ heavy _ mesons particularly .
the reason why we do so is because many of these processes are quite close to particle thresholds and would otherwise be highly dependent on binding energies and the like for which there is no evidence .
in such a soft limit and including the crossed diagram , it is very simple to work out that @xmath42 where @xmath43 are the internal quark masses , in an obvious cyclic notation .
the feynman parametric integral is especially easy to work out when all quark masses are equal to @xmath44 , when @xmath45 it is not much harder to determine the integral ( 3 ) if two masses are equal , say @xmath46 , when @xmath47 thus @xmath48 in the equal mass limit @xmath49 .
see figure 1 to see how @xmath50 varies with @xmath51 .
it only remains to substitute the couplings of the mesons to the quarks in order to derive the theoretical values for @xmath41 .
to do so we will adopt the naive perspective that all strong couplings are related by @xmath0 symmetry .
therefore , in the absence of any isospin breaking and neglecting the third generation of quarks , let us see how far one can go with the effective lagrangian ( suppressing the @xmath30 matrix between spinors ) , @xmath52.\end{aligned}\ ] ] the universal coupling @xmath53 above is fixed by the @xmath54 coupling , which in turn is determined by gtr ; thus @xmath55 .
we must also relate the physical neutral pseudoscalar states to the isospin 0 combinations in eq .
( 4 ) via certain mixings . for the purpose of this and the next section
, we shall disregard electromagnetic effects in the strong sector and take ( see appendix ) @xmath56 with @xmath57 unmixed .
( later on we will examine rare decays , where the small residual mixings become rather important . )
the strong vector meson interactions are likewise written as @xmath58,\end{aligned}\ ] ] suppressing vector indices and the @xmath59 matrix between quark spinors . here
the coupling constant is fixed by that of the @xmath18-meson which is itself determined by the leptonic decay rate : @xmath60 .
these vector states also undergo the semistrong mixing via @xmath61 ( at this level we may identify @xmath62 with the @xmath63 particle . ) finally , it is trivial to consider radiative decays , by substituting @xmath59 for the meson and including the proper electromagnetic coupling of the photon to the quarks , @xmath64 .
the magnitudes of the @xmath41 can be found directly from experimental decay rates , as they occur @xcite .
for @xmath65 decays , the rate is given by @xmath66 where @xmath67 and @xmath68 is the magnitude of the three - momentum of one of the vector mesons in the rest frame of the decaying @xmath69 .
( we must be careful to multiply the right - hand - side of ( 6 ) by a factor of 1/2 when the two vector mesons are identical . )
a very similar formula applies to the decay rate @xmath70 , @xmath71 where @xmath72 refers to the mass of the initial vector and @xmath73 is the momentum of one final particle in the decay rest frame .
it is then a simple matter to extract the @xmath74 from the measured @xcite rates @xmath75 , for subsequent comparison with theory .
we shall do this constantly , without further elaboration .
we mentioned at the start that the @xmath76 amplitude is theoretically determined in the sigma model by ing the @xmath28 and @xmath29 loop contributions and equals @xmath77 , fitting experiment admirably .
indeed many of the processes involve neutral mesons and require the interaction terms , @xmath78,\ ] ] @xmath79.\ ] ] instead of treating each process one at a time , let us consider just three examples to explain what we are doing , before presenting our results for non - rare decays in a table .
consider first the decay @xmath80 , which contains a charge coupling , one vector coupling and one scalar coupling .
the sum of the loops of nonstrange quarks and the mixing angles implied by ( 8) and ( 9 ) gives ( @xmath81 ) @xmath82 using the values quoted previously , we obtain the rough theoretical magnitude , @xmath83 gev@xmath84 , which can be compared with the experimental result , @xmath85 our second example is the semistrong process @xmath86 , which is governed by nonstrange quark loops but is rather sensitive to the vector meson mixing angle @xmath87 .
the theoretical result here is @xmath88 inserting the accepted value of @xmath89 , the prediction is that @xmath90 gev@xmath84 , and this is in agreement with the upper bound obtained by experiment : @xmath91 gev@xmath84 ; in this connection we note that the 1996 pdg compilation @xcite cites a specific value for this particular channel ( which we prefer ) , but this has been retracted in the 1998 pdg compilation , resulting in less predictive power .
finally we take a look at the decay @xmath92 , which is governed by a charm loop and involves one photon .
this time we get @xmath93 which we estimate to be about 0.19 gev@xmath84 , for a constituent quark mass ratio of @xmath94 .
this is close to average experimental value of @xmath95 gev@xmath84 , obtained from the decay rate . in this manner
we can work out theoretical estimates of all dominant @xmath4 processes .
we have tabulated the results and the experimental magnitudes for comparison in table 1 .
the near agreement of naive theory with data is actually quite astonishing , because we are dealing with sizeable external masses in many instances , particularly for heavy charm quarks , and so are a very long way from the chiral limit .
we will return to this point towards the end of the paper .
it should be noted that the last five decay channels in table i correspond to unequal mass quarks running round the loop , in contrast to all the previous cases , and the absolute value of the amplitudes are perhaps not as impressively predicted are the others ( although the ratio @xmath96 is reasonably close to experiment ) . on the whole
the fit is quite satisfactory , considering the fact that we have no free parameters at our disposal .
there are many processes involving the heavier quarks which conserve isospin .
they tend to be rarer than the cases considered above and measure the smaller admixture of lighter quarks in the meson composites .
in fact these rare processes provide much valuable information about such small mixings and , emboldened by the success of the soft momentum predictions of @xmath4 amplitudes , we shall apply the same method to probe the mixing angles of neutral vectors and pseudoscalars .
let us write the meson which couples to the @xmath97 as the combination @xmath98 where the two @xmath99 are very small quantities ( which is why we have not modified the normalization factor in front of @xmath57 in first approximation ) .
it follows then that the non - strange and strange pseudoscalars are ( in the same limit ) , @xmath100 consequently the interactions ( 8) of the neutral pseudoscalars get modified to @xmath101.\end{aligned}\ ] ] much the same idea can be applied to the neutral vector interactions ( with mixing parameters @xmath99 replaced by @xmath102 ) : @xmath103.\end{aligned}\ ] ] we are now in a position to tackle rare @xmath4 processes involving @xmath57 and @xmath104 .
these are evaluated in the same way as before , but now using the neutral couplings ( 14 ) and ( 15 ) .
the results are summarized in table ii ; theoretical values are only worked out to first order in @xmath99 or @xmath102 , since these are already small quantities .
we need to give some words of explanation as to what entry in the last column is being predicted and what is not , in contrast to table i , which is essentially parameter - free . in our parametrizations ( 14 ) and ( 15 ) , the cleanest predictor of the amount of nonstrange and strange quark components in @xmath57 are the two decays @xmath105 and @xmath106 ; they determine separate combinations of the mixings @xmath107 and @xmath108 ( which represent the amount of charm quarks contained in @xmath109 and @xmath110 ) .
we may derive a reasonable fit by choosing the values @xmath111 and @xmath112 ; this permits us to _ predict _ the decay rates for @xmath113 and @xmath114 in the table . while the latter prediction is good , the former is decidedly not in fact the experimental value is slightly embarrassing and difficult to understand on any sensible theoretical basis
; for u(2 ) symmetry between @xmath28 and @xmath29 entails that the process @xmath115 should be almost equal to @xmath116 , which is a far cry from what seems to be observed ! with respect to the @xmath102 parameters (
because @xmath87 is so small ) the process @xmath117 tells us about the nonstrange - charm mixing @xmath118 almost at once , while @xmath119 provides the mixing between strange and charm states , namely @xmath120 .
we find a fair fit ( including sign ) with the values , @xmath121 .
this enables us to _ predict _ the amplitudes for the other entries in table ii involving @xmath102 .
they are all roughly correct , except for @xmath122 and @xmath123 .
we do not understand the reasons for this ; blaming the discrepancy on the effect of mass extrapolation from the shell to zero - momentum would undermine the other reasonable answers ; but in any case it should be noted that the vector meson dominance ( vmd ) prediction is a factor of 250 larger than the data which heightens our suspicions .
it is possible that the experimental results may shift a little over time and bring the results into closer agreement with theoretical expectations , or possibly the experimental decay rate also includes the contribution from @xmath124 .
here we want to take a look at four processes , @xmath125 because they involve isospin breaking interactions of mesons , without being accompanied by photons .
( interestingly , there seems to be no comparable data in the @xmath57 decay sector . )
all four processes are smaller by at least a factor of three compared to processes that conserve isospin ; eg @xmath126
. we may ascribe the existence of these small amplitudes to electromagnetic effects and the @xmath17 mass difference . with regard to the quark triangle loop ,
the @xmath17 difference gives a correction of order @xmath127 , which is insufficient to explain the observed magnitudes .
we shall find below that the experimental results can be reasonably well obtained by @xmath16 mixings between mesons ( including the well - known @xmath128 transition ) , which are themselves dominated by the @xmath17 difference in quark - loop self - energy contributions . before going further ,
let us estimate two strong @xmath4 amplitudes that can not be directly measured from decays , because of the masses of the participating mesons ; they are nonetheless important in a subsidiary role for what follows .
firstly , there is the interaction @xmath129 ( which actually enters @xmath130 , via pole dominance ) , that we estimate to be @xmath131 next we shall need the amplitude @xmath132 finally we shall require amplitudes which we have estimated previously fairly well ( or which we can take directly from experiment ) , namely @xmath133 gev@xmath84 , @xmath134 gev@xmath84 and @xmath135 gev@xmath84 .
other quantities we will need are the @xmath136 and @xmath137 elements , where unprimed states refer to @xmath138 composites while primed states correspond to @xmath139 mesons .
these elements include the most well - studied case of the @xmath128 transition , which is the sum of two pieces @xcite : @xmath140 in similar vein , we may estimate the @xmath141 transition to be @xmath142 we must be a little more careful with the @xmath143 transition , which is dominated by the photon exchange contribution @xmath144 : @xmath145 last , but not least , we shall require the @xmath146 and @xmath147 transitions .
this arises from a quark self - energy bubble plus an @xmath148 tadpole term .
the latter equals the former contribution @xcite because @xmath149 by su(6 ) and chiral symmetry @xcite . noting the gap equation @xcite stemming from the gtr , the total result sums to @xmath150\nonumber \\ & \simeq & 2(m_u^2-m_d^2 ) \simeq -5400 { \rm~mev}^2 \quad { \rm for~~ } m_d - m_u\simeq 4 { \rm ~mev}.\end{aligned}\ ] ] we thereby estimate the two neutral pseudoscalar elements , @xmath151 actually the last element leads to very small corrections because each of the mesons is rather far from the mass shell of the other .
we are now in a position to estimate these rare amplitudes .
( remember that we may ignore the negligible correction induced by @xmath17 mass differences within the triangle loop . )
the first case is @xmath152 upon substituting the @xmath16 transition elements found above and the vertex amplitudes determined previously .
similarly , we find @xmath153 @xmath154 @xmath155 the experimental values of these amplitudes , deduced from the measured decay rates , are , in units of gev@xmath84 , latexmath:[\[\begin{aligned } these answers are in the `` right ball park '' and we venture to conclude that this @xmath16 , pole - dominated mechanism captures the main features of such feeble decays .
here we shall have a look at a few three - body decays , in which pions are produced in a p - state and are known to be dominated by @xmath18-meson poles ; specifically we shall examine the three processes @xmath157 and @xmath158 to see how well they tie up with the earlier @xmath33 amplitudes . before doing
so we shall require more accurate values of the on - shell couplings @xmath159 .
we recall that the quark loop alone gives @xmath160 and @xmath161 mev@xmath84 .
however , in a sigma model , besides the non - strange quark loop we have to add a meson loop associated with sigma exchange .
the effect of this is to enhance @xcite the value of @xmath162 by a factor of 6/5 , to about 6.1 , which is very close to the experimental value coming from the @xmath18 decay width .
correspondingly , the @xmath18-@xmath163-@xmath164 coupling is enhanced @xcite by a factor of ( 6/5)@xmath165 to @xmath166 gev@xmath84 .
the amplitude for the @xmath167 process @xmath130 may be written in the covariant form @xmath168 , where the scalar amplitude @xmath33 is dominated by @xmath18 mesons @xcite in each of the three possible two - body channels : @xmath169\simeq 1480{\rm ~gev}^{-3},\ ] ] in deriving this result , we have averaged over the mandelstam variables : @xmath170 gev@xmath165 . if we then follow thews phase space analysis @xcite , we get @xmath171 with the constant matrix element giving @xmath172 in this way , we predict @xmath173 mev , very close to the observed rate of @xmath174 mev .
our next object of study , @xmath175 , is dominated by a @xmath176 meson pole , but just in the two - pion channel @xcite .
the process also takes the same covariant form as the previous case . here
we estimate @xmath177 upon substituting the @xmath109-@xmath18-@xmath59 amplitude found earlier ( table i ) and going to the soft pion limit .
once more , we find @xmath178 for a phase space factor @xcite of @xmath179 .
the prediction ( 31 ) compares well with the observed rate of @xmath180 ev .
finally we consider the @xmath181 decay .
unlike the previous two examples , the @xmath176 pole is now well within the physical region .
the latest particle data tables state that @xmath182 , including nonresonant contributions ) = @xmath183 kev , whereas our prediction ( see table i ) is that latexmath:[$\gamma(\eta'\rightarrow\rho\gamma ) = strongly suggests that any nonresonant p - wave contact contributions are quite small and that the @xmath18 poles accurately match the data .
incidentally it also confirms that the pseudoscalar mixing angle is @xmath185 , rather than @xmath186 , coupled with the anomaly , as the latter angle gives a rate for @xmath187 which is a factor of 20 less than the observed rate @xcite .
taking stock of the results , we appear to have succeeded in estimating some 34 @xmath188 amplitudes with reasonable accuracy ( see the tables ) barring two or three rare decays involving @xmath110 and @xmath163 mesons which are themselves extracted from one or two experiments that surely deserve repeating in the future .
the success extended to a few abnormal parity three body decays which are known to be dominated by p - wave pion pairs .
we do not regard the coupling constants @xmath53 and @xmath189 which provide the overall scales to be ` parameters ' , because they are predicted in [ 1 ] or are derived from other processes and have well - established magnitudes ; nor do we think that the quark masses are variables at our disposal , since they are also fixed by other means .
the mixing angles @xmath190 and @xmath87 are parameters but they too can be found through processes other than the ones which we have considered in this paper .
the only ( four ) variables which we were truly able to adjust were the mixing angles describing the non-@xmath191 content of @xmath57 and @xmath104 .
perhaps the most surprising aspect of the analysis is the fact that the amplitudes , which we estimated in the soft limit , have reasonable magnitudes even for external heavy meson @xmath191 composites .
we really did not expect this and we think that the conclusions are rather deep and point to some underlying heavy mass scale , way beyond the usual qcd scale of about 300 mev . to put this deduction into perspective , consider some field - theoretic model which includes meson interactions with spinors ; assume it is renormalizable for definiteness .
then the equation for the proper vertex function @xmath192 ( where @xmath193 is the meson momentum and @xmath194 is the relative fermion momentum ) will take the form @xmath195 where @xmath196 is the appropriate interaction kernel for the model in question , @xmath197 stands for the spinor propagator and @xmath198 is the renormalized contact term . if one specialises to pseudoscalar mesons and ignores fermion dressing for simplicity , the equation has the generic scalar form , @xmath199.\ ] ] the subscript @xmath193 has been attached to the renormalization constant @xmath200 to remind ourselves that the renormalization procedure is often momentum dependent , although the infinite part of @xmath200 ( or a @xmath201 pole term in the dimensional method ) is of course insensitive to @xmath193 .
the limit as @xmath202 of the equation above is readily taken and describes the meson vertex in the ( off - shell ) soft limit .
if the pseudoscalar field is actually a composite of the fermions , then @xmath200 must vanish on the meson mass shell ( @xmath203 ) in order that the vertex equation reduces to a homogeneous bethe - salpeter equation .
the results which we found indicate that the products of vertex functions and quark propagators running round the loop are rather insensitive to external momentum since they suggest that an amplitude like @xmath204 above depends very little on @xmath193 , _
i.e. _ the @xmath194-dependence on the right hand side has compensating effects from form factors and propagators .
( it is striking that the propagator denominators would appear to depend sensitively on the binding energy near the constituent threshold @xmath205 in the infrared region of @xmath206 , and there is no sign of this . )
therefore it indicates that @xmath207 depends very little on @xmath193 and that the homogeneous bethe - salpeter equation can very likely be extrapolated to @xmath208 with relatively little change .
our guess is that such loop integrals are really dominated by the ultraviolet region and a mass scale @xmath209 of at least 10 gev much higher than the standard qcd scale associated with the ratio @xmath210 .
indeed , the seemingly weak dependence of our results on @xmath193 are the most extraordinary part of this work and trying to understand the issue properly is an exciting avenue of future research ; our discussion above does scant justice to the problem .
this research was supported by the australian research council .
mds appreciates the hospitality of the university of tasmania where a substantial proportion of the research work was carried out .
since we are using quark loops to predict @xmath4 amplitudes involving @xmath109 or @xmath110 pseudoscalar mesons , it is imperative that we determine the most important @xmath211 mixing angle @xmath190 from processes other @xcite than @xmath4 decays .
here we shall consider two other ways of fixing this angle .
first we consider tensor @xmath212 decays , @xmath213 .
given the respective percentage branching ratios , from the pdg tables , @xmath214 , and remembering the @xmath215 phase space , we may deduce the two independent ratios , @xmath216 @xmath217 inserting the known magnitudes @xmath218 mev , into this pair of equations , we may conclude from the first that @xmath219 , and from the second that @xmath220 .
likewise the measured @xmath221 branching ratio of 0.003 requires @xcite @xmath222 .
as well , there are the tensor decays @xmath223 , but then the tensor angle @xmath224 enters the analysis ; yet @xmath225 survives the results of the analysis .
a second way of extracting @xmath190 is through the qcd quark - annihilation diagram , involving at least two gluon exchange .
diagonalization of the ns - s meson mass matrix leads @xcite to the @xmath211 mixing angle , @xmath226^{1/2 } = 41.9^o.\ ] ] the close agreement between @xmath190 determined by @xmath227 data and qcd theory justifies our use of @xmath185 in tables i and ii . in the singlet - octet basis ,
the mixing angle @xmath228 is instead given by @xcite @xmath229 @xmath230 thus for the pseudoscalar mesons , @xmath231 , which is midway between the original gell - mann - okubo value of @xmath232 and most recent determinations of @xmath233 . in this connection , it is interesting to compare the quark triangle predictions , both ( a.6ab ) exhibit u(3 ) symmetry in the limit @xmath240 .
however for the actual constituent quark mass ratio ( obtained from the gtr or magnetic moments ) of @xmath241 , the numerical values in ( a.6 ) follow .
note that the anomaly version@xcite estimates @xmath242 via chiral perturbation theory but also finds @xmath243 , which is close to our prediction ; we agree with their justification that @xmath244 being near unity is what one expects if the singlet state and the pion have the same wave function .
e - mail : [email protected] e - mail : [email protected] permanent address : physics department , university of arizona , tucson , az 85721 ; e - mail : [email protected] ; r. delbourgo and m.d .
scadron , mod .
a10 * , 251 ( 1995 ) ; a. bramon , riazzuddin and m.d .
scadron , j. phys . *
g24 * , 1 ( 1998 ) .
these are archived in hep - ph/9807505 and hep - ph/9709274 respectively .
adler , phys . rev . * 177 * , 2426 ( 1969 ) ; j.s .
bell and r. jackiw , nuovo cim . * 60 * , 47 ( 1969 ) .
particle data group , c. caso et al , eur .
j. * c3 * , 1 ( 1998 ) .
all the data in our tables refers to the average values appearing in this reference ; these averages are obtained from many , many experiments .
j. steinberger , phys .
76 * , 1180 ( 1949 ) .
lee , _ chiral dynamics _ ( gordon and breach , ny , 1972 ) p. 12 .
m. gell - mann and m. levy , nuovo cim .
* 16 * , 705 ( 1960 ) .
barnett et al , phys .
d54 , 1 ( 1996 ) .
coon and r.c .
barrett , phys . rev . * c36 * , 2189 ( 1987 ) ; s.a .
coon , p.c .
mcnamee and m.d .
scadron , nucl phys . *
a249 * , 483 ( 1975 ) ; _ ibid _ * a287 * , 381 ( 1977 ) .
r. delbourgo and m.d .
scadron , int . j. mod
13a * , 657 ( 1998 ) .
coon and m.d .
scadron , phys . rev . * c51 * , 2923 ( 1995 ) . p.g.o . freund and s. nandi ,
32 * , 181 ( 1974 ) ; s. rudaz , phys . lett . * b145 * , 281 ( 1984 ) . m. gell - mann , d. sharp and w. wagner , phys .
* 8 * , 261 ( 1962 ) .
thews , phys . rev . *
d10 * , 2993 ( 1974 ) ; see his appendix . m.s .
chanowitz , phys .
lett . * 35 * , 977 ( 1975 ) ; _ ibid _ * 44 * , 59 ( 1980 ) ; m. benayoun et al , z. phys . *
c58 * , 31 ( 1993 ) ; _ ibid _ * c65 * , 399 ( 1995 ) .
venugopal and b.r .
holstein , phys .
rev . * d57 * , 4397 ( 1998 ) . a. bramon and m.d .
scadron , phys .
* b234 * , 346 ( 1990 ) ; phys . rev . * d40 * , 3779 ( 1989 ) .
a. bramon , r. escribano and m.d .
scadron , eur .
j * c * , in press ; see also hep - ph/9711229 . h.f .
jones and m.d .
scadron , nucl .
* b155 * , 409 ( 1979 ) ; m.d .
scadron , phys . rev . *
d29 * , 2076 ( 1984 ) .
.dominant strong and radiative decays that are somewhat insensitive to mixing angles . under the column of theoretical formulae , @xmath41
, we use the abbreviations , @xmath245 , @xmath246 , @xmath247 , in units of gev@xmath84 , and @xmath248 for the constituent quark mass ratio .
the latter are determined by the assumed values ( in gev ) @xmath249 [ cols="^,^,^,^,^ " , ] | we show that all abnormal parity three - body meson interactions can be adequately described by quark loops , evaluated at zero external momentum , with couplings determined by @xmath0 symmetry .
we focus primarily on radiative meson decays which involve one pseudoscalar .
the agreement with experiment for non - rare decays is surprisingly good and requires very few parameters , namely the coupling constants @xmath1 and @xmath2 and some mixing angles .
this agreement extends to some three - body decays that are dominated by pion pairs in a p - wave state .
= 10000 |
brown dwarfs link planetary and stellar astrophysics , with compositions like stars but the temperatures of planets .
they form the tail of the initial mass function and , too low in mass to have core temperatures high enough to fuse hydrogen , they cool over time through the brown dwarf spectral sequence . as they cool , different molecules and condensates form and carve their spectra . with the discovery of very cool brown dwarfs we are able to investigate for the first time the physical and chemical processes that occur in atmospheres with effective temperature ranges that would be suitable for a warm beverage . while brown dwarfs are free - floating
, they should share many of the same physical processes as the giant planets that will be uncovered by future surveys .
the proposed spectral class y encompasses brown dwarfs that have cooled below @xmath0@xmath1500 k. about 17 objects have been classified as y dwarfs to date .
many of those have now been found using the wide - field infrared survey explorer ( wise ) cushing11 , kirkpatrick12 , liu12 , tinney13 , kirkpatrick13 .
additional objects have been discovered as wide - separation companions .
liu11 found a very cool ( @xmath1y0 ) companion to a late t dwarf ; luhman12 discovered a @xmath1300350 k object orbiting a white dwarf . at these temperatures ,
nh@xmath3 absorption features begin to appear in their near - infrared spectra , and sodium and potassium wane in importance in the optical as they condense into clouds .
recent follow - up studies have aimed to characterize the y dwarf population .
there has been a large effort to measure parallaxes of y dwarfs : marsh13 present results for 5 y dwarfs and 3 late t dwarfs using a compilation of data from the ground and space .
dupuy13 present results using only the _ spitzer _
space telescope for 16 y and t dwarfs .
beichman14 present results from a compilation of data from keck ii , the _ spitzer _ space telescope , and the _ hubble _ space telescope for 15 y and t dwarfs .
groups have also been collecting followup observations to better understand the spectral energy distributions of y dwarfs .
leggett13 present followup near - infrared photometry for six y dwarfs and a far - red spectrum for wisepc j205628.90 + 145953.3 .
lodieu13 observed 7 y dwarfs in the _ z _ band using the gran telescopio de canarias .
a number of available models for brown dwarfs include models cold enough to represent y dwarfs allard12 , saumon12 , morley12 , but do not yet treat the effects of water clouds .
the first models to incorporate the effects of water clouds into a brown dwarf atmosphere are those of burrows03b .
these models generally find that water clouds do not strongly affect the spectrum of y dwarfs , but there have been few followup studies .
hubeny07 also includes simple water clouds with a fixed mode particle size of 100 . in section [ burrows ]
we will discuss how our results compare to these early models .
a number of studies have included water clouds in exoplanetary atmospheres ; marley99 and sudarsky00 both modeled the effect of water clouds on the albedos of giant exoplanets ; they find the formation of water clouds significantly increases the planetary albedos .
burrows04 also consider water clouds in exoplanets , using a similar approach as burrows03b but for irradiated planets ; sudarsky03 and sudarsky05 calculate the thermal emission of exoplanets that include water clouds and find that they have a strong effect on the emergent spectrum .
clouds have posed the greatest challenge for brown dwarf modeling since the first l dwarfs were discovered . as brown dwarfs cool along the l sequence from 2500 k to 1300 k , refractory materials like corundum , iron , and silicates condense to form thick dust layers lunine86 , fegley96 , burrows99 , lodders02 , lodders03 , lodders06 , helling06 , visscher10 which thicken as the brown dwarf cools .
these dust clouds shape the emergent spectra of l dwarfs [ see , e.g.][]tsuji96 , allard01,marley02,burrows06,helling08,cushing08,witte11 . for field
brown dwarfs these clouds clear over a very small range of effective temperature around 12001300 k and around the same temperature , methane features begin to appear in the near - infrared .
the brown dwarf is then classified as a t dwarf and , for many years , mid to late t dwarfs were considered to be cloud - free .
however , it has long been recognized that other somewhat less refractory materials such as sulfides and salts should condense in cooler t dwarfs lodders99 . as
late t dwarfs ( 500900 k ) were discovered and characterized , a population of objects redder in the near - infrared ( e.g. @xmath5 , @xmath6 colors ) than the predictions of cloud - free models emerged .
morley12 included the clouds predicted to form by condensation of the sulfides and alkali salts and showed that by including thin layers of these clouds , the colors and spectra of these redder observed t dwarfs can be matched .
as these t dwarfs are further characterized , variability has been observed in mid - late t dwarfs buenzli12 ; the sulfide clouds may play a role in this variability .
spectra of directly - imaged planets are also strongly influenced by the opacity of clouds .
the first multi - planet directly - imaged system , hr 8799 , has four planets , all of which have infrared colors that indicate cloudy atmospheres , much like l dwarfs marois08 . other planetary - mass objects also appear to have spectral properties similar to l dwarfs including @xmath7 pictoris b bonnefoy13 and 2m1207b barman11b .
in fact , at similar effective temperatures planetary - mass objects appear to be even more cloudy than their brown dwarf counterparts madhu11c , barman11 , which has been used to suggest that the breakup of the iron and silicate clouds at the l / t transition may be gravity dependent metchev06 , marley12 .
nonetheless the transition to a methane - dominated atmosphere and cloud - depleted near - infrared spectrum must happen at some effective temperature , with the resulting objects appearing as low - gravity ` t ' and ` y ' dwarfs .
the first such object discovered is gj 504b kuzuhara13 which is currently the coldest directly - imaged planet ( @xmath0@xmath1500 k ) and has colors very similar to t dwarfs ; followup observations probing the methane feature at 1.6 suggest that , as expected from thermochemical equilibrium calculations , methane is present in the atmosphere janson13 . in a cold solar composition atmosphere , water clouds will be a massive cloud and an important opacity source . unlike the refractory clouds which have been extensively studied by a number of groups am01 , helling06 , allard01 , tsuji96 , burrows06 , helling08
, the same attention has not been paid to volatile clouds in brown dwarfs . in this work
, we aim to predict the effects that water clouds will have on brown dwarf atmospheres .
we calculate pressure temperature profiles , spectra , and colors for the coolest brown dwarfs .
we study the signatures of the water clouds and their optical properties , estimate their likely particle sizes , and determine at which effective temperatures the cloud will become optically thick in the photosphere .
we end by considering the observability of y dwarfs with the four major instruments being built for the _ james webb space telescope ( jwst ) _ and the detectability of cool giant planets with new and upcoming ground - based instruments .
we calculate 1d pressure temperature profiles , which are self - consistent with both the chemistry and clouds , of atmospheres in radiative convective equilibrium .
the thermal radiative transfer is determined using the `` source function technique '' presented in toon89 .
the gas opacity is calculated using correlated - k coefficients to increase calculation speed ; our opacity database incorporates published data from both laboratory experiments and first - principles quantum mechanics calculations and is described extensively in freedman08 .
the opacity database includes two significant updates since freedman08 , which are described in saumon12 : a new molecular line list for ammonia yurchenko11 and an improved treatment of the pressure - induced opacity of h@xmath2 collisions richard12 .
the cloud opacity is included as mie scattering of spherical cloud particles in each atmospheric layer .
the atmosphere models are more extensively described in mckay89 , marley96 , burrows97 , mm99 , marley02 , saumon08 , fortney08b .
the cloud model calculates the vertical locations , heights , and mode particle sizes of clouds as they condense in the atmosphere .
the calculation is coupled with the radiative transfer calculations , so a converged model will have a temperature structure that is self - consistent with the clouds . the cloud code is a modification of the am01 cloud model .
this model has successfully been used to model the effects of the iron , silicate , and corundum clouds on the spectra of l dwarfs saumon08,stephens09 as well as the sulfide and chloride clouds that likely form in the atmospheres of t dwarfs morley12 . here , we modify it to include the effects of water clouds .
the am01 approach avoids treating the highly uncertain microphysical processes that create clouds in brown dwarf and planetary atmospheres .
instead , it aims to balance the advection and diffusion of each species vapor and condensate at each layer of the atmosphere .
it balances the upward transport of vapor and condensate by turbulent mixing in the atmosphere with the downward transport of condensate by sedimentation .
this balance is achieved using the equation @xmath8 where @xmath9 is the vertical eddy diffusion coefficient , @xmath10 is the mixing ratio of condensate and vapor , @xmath11 is the mixing ratio of condensate , @xmath12 is the convective velocity scale , and @xmath13 is a parameter that describes the efficiency of sedimentation in the atmosphere . solving this equation allows us to calculate the total amount of condensate in each layer of the atmosphere .
we calculate the modal particle size using the sedimentation flux and by prescribing a lognormal size distribution of particles , given by @xmath14}\ ] ] where @xmath15 is the total number concentration of particles , @xmath16 is the geometric mean radius , and @xmath17 is the geometric standard deviation .
we fix @xmath17 at 2.0 for this study ( see discussion in am01 ) .
we calculate the falling speeds of particles within this distribution assuming viscous flow around spheres ( and using the cunningham slip factor to account for gas kinetic effects ) .
we calculate the other parameters in equation [ advdiff ] ( @xmath9 and @xmath12 ) using mixing length theory and by prescribing a lower bound for @xmath9 of 10@xmath18 @xmath19/s , which represents the residual turbulence due to processes such as breaking gravity waves in the radiative regions of the atmosphere .
this process allows us to calculate the mode particle size in each layer of the atmosphere using calculated or physically motivated values for all parameters except for the free parameter @xmath13 . in general
, we find larger particles ( which have higher terminal velocities ) in the bottom layers of a cloud and smaller particles ( which have lower terminal velocities ) in the upper layers .
a high sedimentation efficiency parameter @xmath13 results in vertically thinner clouds with larger particle sizes , whereas a lower @xmath13 results in more vertically extended clouds with smaller particles sizes . as a result
, a higher @xmath13 corresponds to optically thinner clouds and a lower @xmath13 corresponds to optically thicker clouds .
the am01 cloud model code computes the available quantity of condensible gas above the cloud base by comparing the local gas abundance ( accounting for upwards transport by mixing via @xmath20 ) to the local condensate vapor pressure @xmath21 . in cases where the formation of condensates does not proceed by homogeneous condensation
we nevertheless compute an equivalent vapor pressure curve .
the condensation of water vapor into water ice clouds poses some unique problems for our self - consistent equilibrium approach .
water vapor is the most dominant source of opacity in a y dwarf atmosphere .
after it condenses , if the cloud opacity is somehow removed , there is very little gas opacity left in the atmosphere and the layers beneath can radiate efficiently .
however , the cloud opacity can not just be removed from the atmosphere ; it must condense into a cloud , and because oxygen is one of the most abundant elements , the water cloud that forms is quite massive and optically thick .
this means that the dominant vapor opacity source condenses into a dominant solid opacity source . in practice ,
when aiming to calculate a solution in radiative convective and chemical equilibrium that is self - consistent with the water cloud , we find that we are not able to find a self - consistent solution for a range of model effective temperatures from @xmath1225450 k. as the water cloud forms in the model , it significantly warms the atmosphere below it because it prevents flux from escaping .
this warming causes the cloud to evaporate , removing the opacity source , and allowing flux to escape and cool the atmosphere again ; a cloud forms again . an equilibrium solution is not found .
to solve this problem , we borrow phenomenological ideas from our own solar system . when water clouds are observed in the solar system on earth , jupiter , and saturn , they never form in a globally homogeneous layer .
instead , they form into patchy clouds .
for example , on jupiter , there are 5 hot spots though which flux emerges from deep within the atmosphere ; these are believed to be `` holes '' or thin areas of the deep jovian water cloud westphal69 , westphal74 , orton96 , carlson94 .
saturn has similar mid - infrared heterogeneity baines05 . because water clouds _ never _ appear to form globally homogeneous , uniform clouds in the solar system planets , models that include clouds for these planets do not attempt to find a 1d steady - state equilibrium solution in a self - consistent way
instead , for clouds in earth s atmosphere , the evolution of clouds is either modeled over time using a time - stepping model or the clouds are modeled in 2d or 3d .
in fact a method sometimes used in 3d circulation models on earth inspires our approach , described below . in these circulation models ,
clouds form on scales smaller than the grid scale ; cloud opacity is implemented using a two - column sub - grid approach .
other previous efforts for exoplanets and brown dwarfs either did not iterate or used a highly specified cloud parametrization for particle size and cloud height [ e.g][]marley99 , sudarsky00 , hubeny07 .
our approach here is not to include a specified cloud , but instead to iteratively solve for a cloud profile that is self - consistent with the atmosphere structure .
we rely on a model which successfully reproduces cloud particle sizes and distributions on jupiter and earth am01 .
perhaps fortuitously , the particular cloud model employed by burrows03b , who also solve the problem self - consistently , produced somewhat large particles with small infrared optical depth .
more general cases in which some clouds have smaller particle sizes have a far greater optical depth and are more challenging to converge . in this work
, we make the assumption that patchy water clouds also form in brown dwarfs . theoretical motivation for this patchiness has not yet been well - developed , but recent highly idealized models suggest that the rotation and internal heating of brown dwarfs could drive jet - like or vortex - like circulation zhang14 .
such weather patterns in the atmosphere may create rising and sinking parcels of air and maintain inhomogeneous clouds .
other evidence for cloud patchiness in the silicate cloud decks of warmer brown dwarfs has been seen in the variability observations by , e.g. , radigan12 and the spatial mapping of the l / t transition brown dwarf luhman 16b crossfield14 . in this model , even though the water cloud layer forms a thick opacity source , the flux can emerge from holes in the cloud deck .
this assumption allows us to calculate a temperature structure in radiative convective equilibrium that is self - consistent with the clouds because flux is able to emerge through the holes even when the clouds become optically thick .
this means that the temperature structure can remain cool enough to have a condensed water cloud layer . .
@xmath22 represents a fully clear atmosphere ; @xmath23 represents a fully cloudy atmosphere . for the models in the grid presented here , @xmath24 .
, width=360 ] we calculate patchy clouds following the approach of marley10 , who implemented patchy clouds in an attempt to understand a mechanism that could reproduce the l / t transition , in which clouds break up progressively and more flux emerges from holes in the clouds . following this prescription ,
we calculate flux separately through both a cloudy column ( with the cloud opacity included ) and a clear column ( with the cloud opacity not included ) with the same pressure temperature profile .
this calculation is shown schematically in figure [ cartoon ] .
we can change the cloud - covering fraction by varying @xmath25 , the fractional area of the atmosphere assumed to be covered in holes : @xmath26 using this summed flux @xmath27 through each atmospheric layer , we iterate as usual until we find a solution in radiative convective equilibrium .
using @xmath280.40.5 greatly improves convergence . with a hole fraction of at least 4050% , flux escapes from warm , deep layers through the cloud - free column and a consistent @xmath29 profile with water clouds can be calculated . to model clouds and their radiative effect in the atmosphere , we need three pieces of information about the material .
the first is the optical properties the real and imaginary parts of the refractive index of the condensed solid or liquid . in y dwarfs
, water always condenses in the solid phase burrows03b , so we use the optical properties of water ice warren84 .
the second property is the material s density ; we use 0.93 g/@xmath30 for water ice .
lastly , we need the saturation vapor pressure of water ice , which tells us where the cloud will form and how much material is available to form it .
we assume that all material in excess of the saturation vapor pressure condenses to form a cloud .
we use the equation from buck81 to describe the saturation vapor pressure of water ice : @xmath31\ ] ] where @xmath32 is the temperature in degrees celsius and @xmath33 , @xmath34 , @xmath35 , and @xmath36 are constants ( 6.1115@xmath37 , 23.036 , 279.82 , and 333.7 , respectively ) .
the other clouds included in these models are cr , mns , na@xmath2s , kcl , and zns .
the thermochemical models that describe the formation of these clouds are described in visscher06 .
using these models , fits to the saturation vapor pressure as a function of pressure , temperature , and metallicity were presented in morley12 , section 2.4 .
sources of the optical properties used in the mie scattering calculations are also presented in table 1 of morley12 .
our grid of models encompasses the full range of y dwarfs and extends the grid from morley12 to lower temperatures .
this grid includes models from 200450 k in increments of 2550 k. we include surface gravities that range from giant planets to brown dwarfs , from log @xmath38 of 3.0 to 5.0 in increments of 0.5 .
we run all models on this main grid with @xmath13=5 and @xmath25=0.5 . of course
, it is likely that @xmath25 and @xmath13 vary , and could in fact be different for the underlying sulfide / salt clouds and the higher altitude water clouds .
in fact , models at the l / t transition generally need to include non - uniformity in cloud properties across the transition to match its shape marley10 .
we do not aim to fully model all parts of this parameter space here .
however , we do explore some parts of this space ; we run additional models in which we vary @xmath13 from 37 at a single surface gravity ( log @xmath38=4.0 ) and in which we vary @xmath25 from 0.2 to 1.0 ( see figure [ diffcloudcover ] ) . we assume solar metallicity composition for all models , using elemental abundances from lodders03 .
the grid was deliberately chosen to be square , but includes some unphysical combinations of temperature and surface gravity , because higher mass brown dwarfs can not have cooled enough during the age of the universe to reach very low temperatures .
for the coolest models in our grid , @xmath0=200 k , the maximum expected log @xmath38 ( that of a 10 gyr brown dwarf ) is @xmath39 . for 300 k
, maximum log @xmath38 is @xmath404.6 ; for 450 k , @xmath14.75.0 saumon08 .
the ranges come from using a cloudy or cloud - free atmospheric boundary condition for the evolution models , see figures 4 and 5 from saumon08 .
absolute fluxes and magnitudes are calculated from our model spectra by applying the evolutionary radii of saumon08 . those cooling sequences provide the radius of the brown dwarf as a function of @xmath0 and @xmath41 . here we have used the radii from the evolution sequences computed with cloudless atmospheres as the surface boundary condition . a fully self - consistent calculation of the evolution would use a surface boundary condition defined by the corresponding model atmospheres .
this is becoming increasingly difficult for brown dwarfs as the evolution of clouds during the long cooling time of the very cool objects considered here is rather complex .
the sequence of transitions from cloudy l dwarfs to mainly clear mid - t dwarfs , to late - t dwarfs veiled with sulfide clouds ( mns , na@xmath2s , zns , cr , kcl ) , which may also clear out in early y dwarfs before water clouds appear , has yet to be understood properly , both empirically and theoretically .
the use of a uniform , cloudless surface boundary condition has the virtue of simplicity .
we can estimate the uncertainty in the radius thus obtained by comparing the surface boundary condition extracted from the present atmosphere models to those used by saumon08 .
we find that for our nominal @xmath42 , @xmath24 partly cloudy sequence , the entropy at the bottom of the atmosphere ( which gives the entropy of the matching interior model ) is quite close to the cloudless case at @xmath43k .
as the object cools the entropy decreases , and for @xmath44k the entropy is close to the entropy of cloudy atmospheres used in saumon08 .
this nicely corresponds to the transition of the partly cloudy models from optically thin to optically thick water clouds .
figures 4 and 5 of saumon08 show that for given @xmath0 and @xmath41 , the radii between the cloudless and cloudy evolution sequences vary by at most 12% below 500@xmath45k .
thus , the inconsistency between the surface boundary condition used in the evolution sequences and the model atmospheres presented here causes at most a 4% error in the absolute fluxes .
we present results for the grid of models discussed in section [ modelgrid ] .
where appropriate , we incorporate warmer models of t dwarfs from previous studies saumon12 , morley12 for comparison . in section [ cloudprops ] ,
we present the model cloud properties . in section [ pt - struc ]
we present the temperature structures of the models . in section [ specs ]
we present the model spectra , including effects of disequilibrium chemistry . in section [ colors ]
we present model photometry and compare to the growing collection of very cool objects with known distances dupuy13
, beichman14 .
lastly , in sections [ jwst ] and [ gpi ] we will make predictions for the characterizability of y dwarfs with _ jwst _ and the detectability of cool planets with new instruments like gpi , sphere , and the lbt .
and q@xmath46 ) for water clouds of three particle sizes are shown .
these results are for single particle sizes , not a distribution of sizes .
all three show similar general properties , with low q@xmath46 in the optical rising into the infrared and the strongest absorbing feature around 3 .
larger particles are more efficient at both absorbing and scattering for most wavelengths . , width=360 ]
the models presented here include the effects of both sulfide / chloride clouds ( first included in model atmospheres in morley12 ) and of water clouds .
we will mainly focus on the properties of water clouds as the former set of clouds are more thoroughly examined in morley12 .
figure [ miescattering ] shows the optical properties calculated using mie theory for water ice particles with sizes of 0.1 , 1 , and 10 .
the scattering efficiency , q@xmath47 , is the ratio of the scattering cross section of the particle to the geometric cross section .
the absorption efficiency , q@xmath46 , is the ratio of the absorbing cross section to the geometric cross section . in general , the larger particle sizes both scatter and absorb more efficiently than smaller particles for most wavelength ranges .
the locations of features are similar for different particle sizes with the strongest feature in q@xmath46 at 3 . in general
, q@xmath46 rises from optical to infrared wavelengths and remains fairly high through the infrared .
the persistence of these features over a large range in particle sizes suggests that water ice features may be observable for a relatively optically thick cloud even if it contains a range of particle sizes .
figure [ h2oabs ] shows both the absorption efficiency of water ice and the cross section of water vapor ; the wavelength range at which water absorbs shifts as it condenses from vapor to solid phase . in particular , water ice absorbs strongly within the major water vapor opacity windows in the mid - infrared . of water ice particles of three particle sizes ( 0.5 , 5 , and 50 )
is shown ( left axis ) .
these results are for single particle sizes , not a distribution of sizes .
the absorption cross section of water vapor is shown on the right axis .
the phase change of water substantially changes the wavelengths at which it strongly absorbs , filling in many of the regions where water vapor is transparent .
, width=336 ] line is shown to guide the eye .
thin water clouds form in all three models , but only become optically thick in the two coolest models .
mode particle sizes are small ( 35 ) for @xmath0=275 k and larger ( 520 ) for the 200 k model .
the sulfide / salt clouds form and become optically thick in the photosphere of the 400 k model but are optically thick below the photospheres of the cooler two models as they form more deeply .
, title="fig:",width=360 ] line is shown to guide the eye .
thin water clouds form in all three models , but only become optically thick in the two coolest models .
mode particle sizes are small ( 35 ) for @xmath0=275 k and larger ( 520 ) for the 200 k model . the sulfide / salt clouds form and become optically thick in the photosphere of the 400 k model but are optically thick below the photospheres of the cooler two models as they form more deeply . , title="fig:",width=360 ] line is shown to guide the eye .
thin water clouds form in all three models , but only become optically thick in the two coolest models .
mode particle sizes are small ( 35 ) for @xmath0=275 k and larger ( 520 ) for the 200 k model .
the sulfide / salt clouds form and become optically thick in the photosphere of the 400 k model but are optically thick below the photospheres of the cooler two models as they form more deeply .
, title="fig:",width=360 ] water clouds form in brown dwarfs cooler than @xmath0=400 k , but initially in thin , tenuous layers .
they first become relatively optically thick in the photospheres of brown dwarfs cooler than @xmath1350375 k. figure [ cloudcolopdreff ] shows the cloud properties ( mode particle size and geometric column optical depth ) of a representative set of cloudy models .
geometric column optical depth is the equivalent optical depth of particles that scatter as geometric spheres ; since water clouds are very much non - gray absorbers , this is a poor approximation at wavelengths where the particles scatter much more strongly than they absorb .
@xmath0=400 k model atmospheres have sulfide and salt clouds in the photosphere and a thin water cloud in the upper atmosphere .
the cloud properties for an example @xmath0=400 k , log g=4.5 model is shown in the upper panel of figure [ cloudcolopdreff ] .
the water cloud does forms at a pressure level of 4@xmath48 bar , in the upper atmosphere , in an optically thin layer .
the sodium sulfide cloud becomes optically thick ( @xmath49 ) at 20 bar , near the bottom of the photosphere .
mode particle sizes of this dominant na@xmath2s cloud are around 2030 . as a brown dwarf cools below @xmath0=400 k , the water cloud forms deeper in the atmosphere and more material is available to condense , making the water cloud much more optically thick .
the middle panel shows a cooler model , @xmath0=275 k , in which the cloud has become geometrically optically thick near the top of the photosphere .
mode particle sizes in this high cloud layer are fairly small around 15 .
the feature in the column optical depth around 10@xmath50 bar is caused by the fact that the pressure temperature profile becomes warmer than the water condensation curve at that pressure . in this model ,
the sulfide and chloride clouds become optically thick much deeper in the atmosphere , around 100 bar , which is below the photosphere . for a brown dwarf that has cooled to @xmath0=200 k
, the water cloud is very optically thick and forms within the photosphere .
the bottom panel shows a @xmath0=200 k model ; the base of the water cloud is at 2 bar and the column optical depth is @xmath160 .
mode particle sizes of water ice in the photosphere are 420 .
the cloud opacity is the dominant opacity source through cloudy columns of the model atmosphere .
overall , and in agreement with marley99 and burrows03b , we see that water clouds begin to form with small particle sizes high in the atmospheres of objects around 400 k. they become marginally optically thick for objects cooler than 350375 k. for very cold objects , @xmath1200250 k , the water cloud is a dominant opacity source through cloudy columns of the atmosphere . s cloud . for models with @xmath0=200 k and 275 k ,
the single scattering albedos of both the water and na@xmath2s cloud are shown for a single atmospheric layer .
the water cloud forms high in the atmosphere ( 2 bar and 0.03 bar for the layers shown from the 200 and 275 k models , respectively ) and the na@xmath2s cloud forms deeper ( 200 and 60 bar , respectively ) .
the sulfide cloud single scattering albedo is relatively uniform , rising from @xmath10.6 in the optical to 0.9 at 10 .
the water cloud single scattering albedo has many more features , which vary with particle size ( the mode particle size is @xmath120 for the 200 k model and @xmath15 for the 275 k model ; the single scattering albedo is calculated for the distribution of particle sizes calculated using the cloud code ) . in the optical
the single scattering albedo is 1.0 , which means that the water clouds do not absorb efficiently at short wavelengths .
, width=336 ] however , even if a cloud is geometrically optically thick , depending on the optical properties of the absorbing and scattering particles , the cloud may not dramatically affect the spectrum .
if the absorption efficiency is very low at a given wavelength , photons from layers below the cloud will have a very low probability of being absorbed by the cloud ; thus the spectrum will appear essentially as it would in a cloud - free atmosphere at that wavelength . the single scattering albedo quantifies the importance of scattering and absorption by cloud particles .
it is the ratio of the scattering coefficient to extinction coefficient ( including both scattering and absorption ) at a given wavelength .
a single scattering albedo of 1.0 indicates that the cloud particles are entirely scatterers ; a single scattering albedo of 0.0 indicates that the cloud particles are entirely absorptive .
figure [ ssa ] shows the single scattering albedo for the clouds in the @xmath0=200 and 275 k models also shown in figure [ cloudcolopdreff ] .
we show both the single scattering albedo for the sodium sulfide cloud , which is deep within the atmosphere at @xmath1 100 bar , and the water ice cloud , which is in the photosphere around 0.11.0 bar .
the single scattering albedo of the sulfide cloud is almost identical for each of these two model atmospheres ; it rises from 0.6 in the optical to 0.95 at 12 , indicating that the cloud becomes less efficient at absorbing as the wavelength increases .
in contrast , the single scattering albedo of the water ice cloud has strong absorption features throughout the near and mid - infrared .
the two models have slightly different features because the particle sizes are different , and the scattering and absorption properties depend fairly strongly on particle size ( see figure [ miescattering ] ) . for the 200 k object ,
the water ice mode particle size at 2 bar is about 20 . for the warmer 275 k model , the mode particle size at 0.03
bar is about a factor of four smaller , @xmath15 .
however , the strongest features are evident for both models .
the most obvious is the sharp decrease in single scattering albedo at 2.8 , indicating that the cloud becomes more strongly absorbing at that wavelength .
the feature at 10 is very evident for the warmer model , and much more muted in the cooler model .
the presence of features indicates that water ice is not a mostly gray absorber ( like many of the more refractory clouds are ) and , when present and optically thick , may cause spectral features , including ones that depend on particle size .
examples of model pressure temperature profiles for the model grid are shown in figure [ pt - profile ] .
in general , cloud opacity in a brown dwarf increases the temperature of the atmosphere at all points in the atmosphere .
this is because in a cloudy atmosphere , the overall opacity is slightly higher , so the temperature structure of a converged model with the same outgoing flux will be slightly warmer . for a cloud - free model atmosphere ,
once the water has condensed out of the atmosphere , there are very few opacity sources left : mainly ch@xmath4 , nh@xmath3 , and collision - induced absorption from molecular hydrogen .
this means that the brown dwarf is quite transparent in those layers and flux is able to emerge from deeper layers . in contrast , if we assume , as we do in our cloudy models , that water in excess of the saturation vapor pressure condenses to form a water ice cloud , that cloud provides a large opacity source , preventing the brown dwarf from efficiently emitting from layers underneath the cloud , and significantly warming the atmosphere . for model atmospheres between 400 and 500 k , even though the cloud - free p t profile may cross the water condensation curve , the converged cloudy models ( including the effect of sulfide clouds ) are somewhat warmer and do not cross the water condensation curve , so the model water cloud does not form .
for the warmest effective temperature brown dwarfs in which we find that water clouds form and can exist in radiative equilibrium , ( @xmath51400 k ) , the water clouds are at low pressures within the radiative upper atmosphere .
they remain above the photosphere and optically thin .
this means that the clouds affect the spectra very little , but the converged cloudy model s p
t profile is warmer by @xmath120 - 50 k in the upper atmosphere than a corresponding cloud - free model .
figure [ pt - profile ] shows the models of pressure temperature profiles of brown dwarfs ( log g=5.0 , upper panel ) and planet - mass objects ( log g=3.5 , lower panel ) , with effective temperatures of 200 , 300 , and 450 k. the photospheres , shown as thick black lines , show that the observable region of the atmospheres tends to be below the radiative upper atmospheres which are prone to numerical challenges in the models .
the convective regions are also shown as colored shaded regions ; higher gravity objects at these temperatures have multiple convection zones .
this well - known result is because their p
t profiles cross regions of parameter space where the opacity of a solar composition equilibrium gas is low , and radiative energy transport becomes efficient . in lower pressure regions in the atmosphere ,
the opacity increases and radiative transport is once again inefficient and the model becomes unstable to convection [ e.g.][]marley96 , burrows97 .
note that the coolest models ( @xmath0=200 k ) also cross the nh@xmath3 condensation curve , indicating that for very cold y dwarfs we will also need to consider the effects of the ammonia cloud . like the h@xmath2o cloud
, it will first form as a thin cloud high in the atmosphere , and become more optically thick as the object cools further .
the opacity of this cloud is not currently included in the models .
the spectra of y dwarfs are dramatically different from blackbodies at the same effective temperatures , with strong molecular absorption features where the thermal emission would peak and opacity windows at shorter wavelengths where a blackbody would be faint .
a 450 k object is 6 orders of magnitude brighter in j band than its blackbody counterpart ; a 300 k object is 10 orders of magnitude brighter ; a 200 k object is 15 orders of magnitude brighter .
the spectra of y dwarfs are dominated by the opacity of h@xmath2o , ch@xmath4 , nh@xmath3 , and h@xmath2 collision - induced absorption ( cia ) .
figure [ molopac ] shows the molecular opacity and collision induced absorption at representative locations in the photospheres of objects with effective temperatures of 900 ( t6.5 ) , 450 ( y0 ) , and 200 k ( y2 + ) .
this progression from mid - t to late - y is marked by the increase in ammonia absorption relative to water and methane .
features from co and co@xmath2 have been found in t dwarfs yamamura10 , tsuji11 .
the strongest band of co is the dominant opacity source at 4.55 , even assuming equilibrium chemistry , for a mid - t dwarf .
co@xmath2 is also an important opacity source in the mid - infrared for warmer objects , especially if its mixing ratio is increased by vertical mixing . as an object cools ,
co and co@xmath2 are strongly disfavored in equilibrium , but vertical mixing could increase their abundance in the atmosphere by several orders of magnitude , so the strongest absorption bands may still prove to be important for late t and early y dwarfs . for very cold objects , the effects of disequilibrium carbon chemistry should become less important .
species such as ph@xmath3 ( phosphine ) and h@xmath2s , which have been observed in jupiter s atmosphere prinn75 , will also be present in y dwarfs .
phosphine is likely observable in the mid - infrared ; the strongest ph@xmath3 feature is at 4.3 and is the dominant opacity source at that wavelength in the photosphere of a @xmath0=450 k y dwarf .
while equilibrium models find little phosphine in the photospheres of @xmath0=200 k objects , visscher06 predict that phosphine will be in disequilibrium in giant planet and t dwarf atmospheres and could be orders of magnitude more abundant than equilibrium models predict .
the effect of phosphine on y dwarf spectra may therefore be underestimated in these models and even more pronounced in real y dwarfs .
h@xmath2s affects the spectra mostly in _
band , where it acts largely as a continuum absorber , depressing the _ h _ band peak , and to a smaller degree the _ y _ band peak .
= 5 , @xmath25=0 ( 900 , 600 k ) and @xmath25=0.5 ( 450 , 300 k ) .
spectra are rescaled such that the flux at the peak of _ j _ band is the same for all models .
note the change in the shape of the near - ir spectral windows .
_ j _ and _ h _ bands narrow as ammonia and methane increase in abundance .
ammonia absorption in _
y _ band causes the band shape to bifurcate for the coolest model . , width=336 ] model spectra of objects at two different representative gravities ( log g=5.0 , 4.0 ) from @xmath0=450 to 200 k are shown in figure [ spec - log - bands ] .
the models shown assume @xmath13=5 and @xmath24 , as described in section [ partlycloudymodels ] , and include both the salt / sulfide clouds ( na@xmath2s , kcl , zns , mns , cr ) and water ice clouds .
as a brown dwarf cools over the y dwarf sequence , the near - infrared flux dramatically declines . by the time
the object has cooled to 200 k , almost all flux emerges in the mid - infrared , between strong molecular absorption features .
figure [ spec - log - bands ] also shows the locations of the dominant absorption bands . for objects at these temperatures ,
absorption is dominated by water , methane , and ammonia , with contributions from ph@xmath3 in the mid - infrared , including the feature at 4.3 for the 450 k model spectrum .
features from co and co@xmath2 are not present here due to the low mixing ratios of these molecules at these temperatures in chemical equilibrium .
minor differences exist between models at the two shown gravities .
we explore the gravity - dependence of features more in the discussion of figure [ gravsigs ] .
figure [ irspec - zoom ] shows how the changes in molecular abundances and absorption changes the shape of the near - infrared spectra over the t to y sequence .
the wide spectral windows in the water absorption , typical of warmer brown dwarfs , narrow as ammonia and methane increase in abundance in the near - infrared photosphere . as ammonia increases in abundance ,
the weaker ammonia feature between 1 and 1.1 begins to carve away the center of _ y _ band .
for the 300 k model shown in figure [ irspec - zoom ] , the _ y _ band is bifurcated into two peaks by this absorption band .
the appearance of this split _ y _ band will depend on the nitrogen chemistry ; if ammonia is less abundant due to disequilibrium chemistry , this change may occur at a lower temperature .
the decline of the alkali absorption with temperature ( see figure [ alkalis ] ) will also affect the underlying continuum absorption in _
y _ band and therefore the appearance of this split .
figure [ spec - log - bands ] shows the summed flux through the cloudy and clear columns of the model atmosphere .
however , the flux through each of those columns is not equal ; since the opacity of the cloud increases the total opacity through the column , less flux always emerges through the cloudy column than through the clear column .
figure [ cloudyspecs ] shows models at the same effective temperatures , now showing the flux from the cloudy and clear columns . at effective temperatures above 400 k , the water cloud has not yet formed or is extremely thin , so the cloudy and cloud - free columns look quite similar ; they differ only substantially in the @xmath52 and @xmath53 bands , which is the region that the deep sulfide clouds affect . at effective temperatures between 300375 k ,
the water cloud gradually becomes more optically thick .
it first forms quite high in the atmosphere , and influences the mid - infrared from 2.85 where water ice particles absorb most efficiently ( see figure [ ssa ] ) . for
cold y dwarfs between 200 - 300 k the water cloud thickens as the object cools . at effective temperatures of 200
k , it has become quite optically thick : most regions of the near- and mid - infrared have significantly less flux emerging from the cloudy column .
in fact , about 10@xmath54 times less flux emerges at 4.5 from the cloudy column than the cloud - free column .
this picture , where flux is emitted almost entirely through clearer columns of the atmosphere , is similar to jupiter and saturn s deep water clouds , which appear to have holes in the clouds ( the so - called ` 5-micron hot spots ' in jupiter ) through which most of the mid - infrared flux emerges .
interestingly , water ice particles of this size ( 120 ) are very inefficient at absorbing photons with wavelengths shorter than 1.4 , so the @xmath53 and @xmath52 bands are not strongly affected , even in a column with a geometrically optically thick cloud .
it is also instructive to look at models that have the same total amount of flux emitted through model atmospheres with different cloud - covering fraction @xmath25 and different sedimentation efficiency @xmath13 , to understand the sensitivity of our results to our choice of those parameters .
the models presented in figure [ diffcloudcover ] are separate from our main grid , which was run with @xmath24 and @xmath13=5 for all models ; these models instead are run with @xmath25=1.0 , 0.7 , 0.4 , 0.3 , and 0.2 and with @xmath13=3 , 5 , and 7 respectively .
all models have the same amount of total flux emitted as a 200 k blackbody . in both panels ,
the models look very similar to each other at wavelengths shorter than about 2 , where the water clouds do not strongly absorb . as we increase the cloud fraction ( decrease @xmath25 ) , more flux emerges between 2 and 3.6@xmath55 m and beyond 5.5@xmath55 m .
this additional flux comes at the expense of the peak flux at @xmath56 m .
in essence , increases in cloudiness redistribute this peak flux to other wavelengths .
when we vary @xmath13 , as expected , lower values of @xmath13 have somewhat more cloud opacity ; this effect is largest in _ k _ band .
disequilibrium carbon chemistry was predicted by fegley96 and confirmed in spectra by noll97 and saumon00 .
it is known to be important for brown dwarfs of many temperatures saumon06 , hubeny07 and may be even more significant for young planets konopacky13 .
the atmospheres of cool brown dwarfs should be methane - dominated .
however , the chemical reaction that leads to methane formation is strongly temperature sensitive and becomes very slow at cold temperatures .
if the timescale of mixing in the atmosphere is faster than the timescale for this reaction to occur , carbon will remain in co instead of being converted to the ch@xmath4 favored by equilibrium chemistry ; for very cold objects this will become less important as the region where methane is thermochemically favored will extend very deeply into the atmosphere .
this process has been explored in detail in a number of papers saumon06 , hubeny07 , visscher11 , moses11 .
disequilibrium chemistry also affects other abundant molecules , such as the conversion of n@xmath2 to nh@xmath3 and co@xmath2 to ch@xmath4 .
other elements such as phosphorous are also out of chemical equilibrium in cold planets like jupiter ; in equilibrium , phosphorous would be in the form p@xmath4o@xmath57 but it is instead observed as ph@xmath3 because the pathways for forming p@xmath4o@xmath57 are kinetically inhibited visscher06 . here , we use the approach developed in smith98 and techniques presented in saumon06 to approximate the effect of co / ch@xmath4 and n@xmath2/nh@xmath3 disequilibrium in the atmospheres of y dwarfs . using this approach
, we calculate a quench point in the atmosphere where the mixing timescale is equal to the chemical reaction timescale , which is governed by the slowest step of the fastest pathway of the reaction . above the quench point
, we assume that the mixing ratio of the molecule is constant .
this has been shown using full kinetics models to be a good approximation in substellar atmospheres visscher11 .
note that in the disequilibrium chemistry calculation we calculate @xmath9 in the convective zone using mixing length theory and vary @xmath9 in the radiative zone as a free parameter , between 10@xmath5810@xmath59 @xmath19/s . in contrast , in the cloud code , we calculate @xmath9 using mixing length theory with a minimum @xmath9 of 10@xmath18 @xmath19/s ; the clouds and chemistry are thus not strictly self - consistent in the radiative region .
the results of these calculations are shown in figure [ diseq ] for three test cases .
for the 450 k model , the disequilibrium model is generally slightly brighter across the near - infrared than the equilibrium model .
this is because in equilibrium , ammonia is strongly thermochemically favored ; out of equilibrium , there is slightly more n@xmath2 , which is not a strong absorber , and slightly less nh@xmath3 , which absorbs strongly across the near - infrared ( see figure [ molopac ] ) . in the mid - infrared ,
the disequilibrium model is slightly brighter around 4 and slightly fainter around 4.6 .
this is due to the increase in co and decrease in ch@xmath4 ; this shift increases absorption from the most prominent co band at 4.5 to 4.9 and decreases absorption from both ch@xmath4 and nh@xmath3 between 3 and 4.4 .
the increase in flux beyond 8 is because of the decrease in nh@xmath3 which is the strongest absorber at those wavelengths . for the 300 k and 200 k
objects shown in figure [ diseq ] , disequilibrium chemistry of these particular gases becomes less important as the objects cool .
the atmospheres of these colder objects favor ch@xmath4 and nh@xmath3 strongly in equilibrium over a progressively wider proportion of their atmospheres .
this means that even if deeper layers were mixed upwards to the photosphere , those layers would still be dominated by ch@xmath4 and nh@xmath3 .
the shape of the @xmath60 band has been used as an indicator of the increased abundance in ammonia through the t sequence to the y dwarfs as they cool ( see figure 5 from cushing11 ) .
we show a modeler s version of the same sequence in figure [ hbandammonia ] with the three sets of spectral indices used to classify t dwarfs burgasser06a , delorme08 .
the blue side of @xmath60 band narrows as the object cools , due largely to increased ammonia absorption from 1.5 to 1.6 . in equilibrium ,
the shape changes from 900450 k as the ammonia absorption increases .
disequilibrium chemistry changes the progression somewhat because the ammonia appears more gradually from 900300 k. the shape of the disequilibrium @xmath60 band at 450 k is very similar to the shape of the equilibrium @xmath60 band at 600 k. = 10@xmath54 @xmath19/s ) narrow more slowly on the blue side where ammonia absorbs because disequilibrium chemistry decreases the amount of nh@xmath3 and increases the amount of n@xmath2 .
the locations of spectral indices used to classify t dwarfs are shown burgasser06a , delorme08.,width=360 ] because of the high densities in brown dwarf atmospheres , sodium and potassium bands at optical wavelengths are extremely pressure - broadened in brown dwarf spectra tsuji99 , burrows00b , allard05 , allard07 .
there are few other optical absorbers in brown dwarf atmospheres , so these strong pressure - broadened features shape the optical spectra of most brown dwarfs , but for y dwarfs these atoms begin to wane in abundance . as is shown in figure [ alkalis ] , as sodium and potassium condense into na@xmath2s and kcl solid condensates , the depths and widths of the alkali features decrease .
because the slope of the optical spectrum for warmer objects is largely controlled by the pressure - broadening of these features , as they decrease in strength , the overall slope of the optical also decreases , making y dwarfs somewhat bluer in these colors . of great interest to the community in the coming years
is the detection of cold planetary mass objects , either orbiting stars or free - floating marois08 , liu13 .
figure [ gravsigs ] shows potential gravity signatures for 450 k objects predicted by our models in the near - infrared ( @xmath52 , @xmath53 , @xmath60 , @xmath61 ) and from 312 . for the near - infrared bands
, the inset figure shows the shapes of the bands with the peak flux in the bands normalized to the same relative height .
these broad gravity signatures are largely caused by the higher pressure photospheres of higher gravity objects . in @xmath52 and @xmath53 bands ,
the wings of the alkali bands , especially potassium , extend into these bands .
these broad wings are due to the extreme pressure - broadening of the alkali lines .
the higher pressures probed in the higher gravity photosphere cause less flux to emerge in both @xmath52 and the blue side of @xmath53 band .
as opacity sources change due to increased pressure , the temperature profile of the atmosphere adjusts ; for y dwarfs , this brightens _
band , causing more flux to emerge from that window . in @xmath61 band ,
the collision - induced absorption of h@xmath2 is a major opacity source .
this feature is quite pressure - dependent , so much like the broadened alkali bands in @xmath52 and @xmath53 bands , the higher gravity atmosphere with the higher pressure photosphere has more total opacity in @xmath61 band .
this decreases the amount of flux that emerges and broadens the @xmath61 band shape . the major gravity dependent feature at wavelengths longer than 3 is between 3.5 and 4.7 .
at these wavelengths , the lower gravity objects have an additional absorber which changes the shape of that mid - infrared feature .
this region of the spectrum is in a window between major methane and water absorption features and is the brightest peak in the near- mid - infrared spectra .
ph@xmath3 is a strong absorber from 44.6 and somewhat gravity dependent at those wavelengths , absorbing more strongly in the lower gravity models . additionally at these wavelengths , due to the different p
t profile at high gravity , we probe somewhat deeper , hotter layers , allowing more flux to emerge from the higher gravity model . in order to compare to observations of the y dwarf population ,
we calculate model photometry .
the radii used to calculate absolute magnitudes were interpolated using the cloud - free evolution models from saumon08 .
color magnitude diagrams ( cmds ) of the model photometry are shown for two different gravities ( log g=5.0 and log g=4.5 ) in figures [ cmd - logg5 ] and [ cmd - logg45 ] respectively .
each set of cmds shows l and t dwarfs as open grey circles and y dwarfs as green points with error bars .
models from this work and models from saumon12 and morley12 are shown .
the first panel ( @xmath62 vs. m@xmath63 ) shows that y dwarfs are significantly bluer in this color than the slightly warmer t dwarfs , with a 0.25 magnitude gap in color between the coolest t dwarfs and warmest y dwarfs .
this bluer @xmath62 color is expected and indeed predicted by the models , and is due to the condensation of the alkalis into na@xmath2s and kcl clouds .
as they condense out of the gas phase , the broadened alkali lines decrease in strength marley02 , burrows03b .
since those lines had been suppressing @xmath52 band flux more than @xmath53 band flux , as they decrease the effect is to make @xmath62 appear bluer .
however , three of the six y dwarfs with measured @xmath52 and @xmath53 photometry are appreciably bluer than even the bluest models predict , and cloudy models are significantly redder .
the second cmd shows @xmath6 vs. m@xmath64 .
these colors are most sensitive to the sulfide and chloride clouds , which suppress the @xmath52 and @xmath53 band flux .
since clouds tend to suppress the flux in @xmath53 band , the clouds make these colors redder , matching the observations of the redder y dwarfs .
the sulfide / chloride clouds wane in importance as objects cool from 450 to 325 k. for objects cooler than 325 k , the water clouds become increasingly important .
counterintuitively , the water clouds tend to make @xmath6 colors bluer ; this is because of water ice is strongly scattering ( and a poor absorber ) in @xmath53 band but becomes absorbing in @xmath60 band ( see single scattering albedo plot , figure [ ssa ] ) . around 325 k ,
the cloudy color and cloud - free colors are the same . on the @xmath6 cmd
, a line showing the effect of disequilibrium chemistry is also shown .
interestingly , this also makes the @xmath6 colors redder , mostly due to decreased absorption from nh@xmath3 across the near - infrared . in reality
, a combination of condensates and disequilibrium may be affecting y dwarf spectra .
the third panel shows @xmath65 vs. @xmath66 .
somewhat puzzlingly , all the y dwarfs except one cluster around a color of 0.0 and m@xmath67 of 20.5 .
this behavior is quite different from the late t dwarfs , which have a large spread in @xmath65 colors .
the location of this cluster is somewhat redder and brighter in @xmath60 band than the models predict .
this could be due to a number of factors ; a major contributor is the incompleteness of the methane line list used in the current study .
we expect from preliminary results that the new line list from yurchenko14 will redden these colors .
the last panel shows the y dwarfs where they emit the most energy : the mid - infrared .
brown dwarfs get quickly redder in @xmath68 $ ] color as they cool and the peak of the planck function moves redward .
generally the colors of the objects seem to match the model colors relatively well , with the exception of the two reddest objects , which appear to be brighter than the models at 4.5 microns by up to 2 magnitudes .
y dwarfs emit most of their flux in the mid - infrared .
this will make their characterization from the ground extremely challenging , especially for the coldest objects .
no y dwarfs had yet been discovered during the _ spitzer _ space telescope s cryogenic mission , which ended in 2009 , so the coolest brown dwarf to have a mid - infrared irs spectrum is spectral type t7.5 saumon06 .
_ jwst _ will have unprecedented sensitivity in the near - infrared , and it will be the main tool with which we can study the coldest brown dwarfs
. the two most important instruments for spectroscopic characterization will be the near - infrared spectrograph ( nirspec ) and the mid - infrared instrument ( miri ) .
nirspec is the most sensitive near - infrared spectrograph on _ jwst _ and will be capable of moderate resolution spectroscopy ( r@xmath11000 or r@xmath12700 ) from 15 in 3 bands ( 1.01.8 , 1.73.0 , and 2.95.0 respectively ) .
figure [ jwst - nirspec ] shows the sensitivity for each of these channels ; these sensitivity lines represent the faintest continuum flux observable with an integration time per channel of 10@xmath18 seconds ( 27.8 hours ) and a signal - to - noise ratio ( snr ) of 10 .
we also show example spectra of brown dwarfs spanning a full range of y dwarf temperatures and located 5 pc from the earth . in the bottom panel
we zoom into the band 3 spectral region , where even our coolest models ( @xmath0=200 k ) would be observable .
the dotted lines show models with no clouds from saumon12 ; the solid lines show the models from this work including water and sulfide / salt clouds .
the warmer y dwarfs discovered to date , @xmath0=400500 k , will be observable across the near - infrared bands , with the exception of the deepest absorption bands
. we will be able to detect the presence or absence of strong features from water , ammonia , methane , phosphine , carbon monoxide , and carbon dioxide using this instrument , allowing us to constrain disequilibrium carbon , nitrogen , and phosphorous chemistry in y dwarf atmospheres .
the y dwarfs cooler than 300 k will be too faint to observe at wavelengths from 13.5 , as the flux from the near - infrared collapses
. however , we will be able to detect these objects at high snr between 3.5 and 5.0 .
the shape of this region is controlled by a variety of absorbers : ammonia and methane on the blue side , water on the red side , and possibly h@xmath2s , ph@xmath3 ( if phosphorous chemistry is in disequilibrium ) , and co and co@xmath2 ( if carbon chemistry is in disequilibrium ) .
the prominent feature at 4.2 is an ammonia absorption feature and could be useful for determining the effective temperatures of cold brown dwarfs .
this window region of the opacity will likely provide the most information about these otherwise very faint objects .
miri will be capable of low ( r@xmath1100 ) resolution spectroscopy from 514 and moderate resolution ( r@xmath13000 ) spectroscopy from 528.3 micron .
it is the only _ jwst _ instrument that will observe wavelengths longer than 5 and will be 50 times more sensitive than the _ spitzer _ space telescope .
the miri moderate resolution spectrograph has four channels : channel 1 from 5.07.7 , channel 2 from 7.711.9 , channel 3 from 11.918.3 , and channel 4 from 18.328.3 .
figure [ jwst - miri ] shows the sensitivity for each of these bands ; like the nirspec sensitivity limits , these sensitivity lines represent the faintest continuum flux observable with an integration time per channel of 10@xmath18 seconds ( 27.8 hours ) and a signal - to - noise ratio ( snr ) of 10 , and we show example spectra of the same brown dwarf models located 5 pc from earth .
the miri sensitivity at 6.4 is about 10 times less sensitive than the nirspec sensitivity at 5 .
this is due to a combination of instrumental effects , including miri s higher dark current , higher intrinsic spectral resolution , finer spatial sampling , lower quantum efficiency , and additional optics . to obtain the full 528 spectrum without gaps , three separate observations ( using three different settings of the mrs spectrograph )
are needed , so the actual observing time to acquire these spectra is three times as long .
many of the current suite of y dwarfs discovered to date using the _ wise _ data will be easily observable using this instrument .
most of these objects have temperatures between 400500 k and are within 10 pc of the earth dupuy13 , beichman14 .
the @xmath0=450 k model shown in figure [ jwst - miri ] is well above the sensitivity limit for 518.3 .
this will allow us to probe parts of the spectrum where the opacity is dominated by different molecules : water from 57.2 , methane from 7.28.5 , and ammonia from 8.518 .
channel 4 is much less sensitive so a 450 k object is only marginally detectable from 1828.3 with a snr of 10 .
the coldest brown dwarfs will push the sensitivity limits for this instrument .
no objects colder than 300 k have been discovered to date , but if we find such objects , they will be quite challenging to observe .
the highest snr spectra will be from channels 2 and 3 , from 7.718.3 .
this wavelength range is shaped by methane , ammonia , and h@xmath2 cia .
objects below 300 k will be only marginally detectable in channel 1 and not detectable in channel 4 at r@xmath11000 .
binning the spectra to r@xmath1300 would improve sensitivity by a factor of @xmath13 , improving the detection limit for @xmath0=250 k objects and still providing adequate resolution for identifying prominent molecular bands .
the two other main science instruments on _ jwst _ are the near - infrared camera ( nircam ) and near - infrared imager and slitless spectrograph ( niriss ) .
these instruments are somewhat less well - suited to the spectral characterization of the coolest brown dwarfs .
niriss is optimized for high contrast imaging , high resolution imaging of extended sources , and transiting exoplanet measurements ; it also has a wide - field r@xmath1150 slitless spectroscopy mode designed for detecting high redshift emission lines .
however , for the characterization of cool brown dwarfs , sensitivity is the most important feature .
nircam does have a grism mode that will be capable of 2.45 r@xmath12000 slitless spectroscopy .
however , its sensitivity at those wavelengths will be somewhat lower than nirspec s sensitivity .
the lower sensitivity is because nircam uses a slitless grism that is sensitive to sky background across a large field .
this mode is more optimized for precision photometry and stability , making it a powerful instrument for , e.g. , exoplanet transmission spectra and secondary eclipses . for the sensitivity - limited work needed to characterize y dwarfs
, nirspec will be a more suitable instrument .
observing directly - imaged giant planets at the temperatures of y dwarfs ( 200450 k ) will push the limits of current technology .
nonetheless , some current or soon forthcoming instruments specialized for high - contrast imaging will be capable of detecting such planets .
if planets with masses between 1 and 10 m@xmath64 are quite common , systems of a variety of ages will have planets with these temperatures .
depending on the mechanism of formation , a 10 m@xmath64 planet will reach @xmath0=450 k at an age of 12 gyr .
a 5 m@xmath64 planet will reach @xmath0=450 k in 300600 myr ; a 1 m@xmath64 planet in 2030 myr fortney08b .
the current and forthcoming instruments most capable of detecting such planets are the gemini planet imager ( gpi ) , the spectro - polarimetric high - contrast exoplanet research ( sphere ) , and the large binocular telescope adaptive optics system ( lbtao ) .
gpi macintosh06 and sphere beuzit06 are instruments designed for 8-meter class telescopes and optimized for studying young hot giant planets around bright stars
. both will have advanced adaptive optics systems and hope to achieve planet - star flux contrasts between 10@xmath69 and 10@xmath70 .
they will target young stars to find self - luminous planets and expect to find up to dozens of planets in their initial campaign surveys mcbride11 .
predicted gpi contrast curves suggest that it will be capable of @xmath71 contrast for very bright stars ( @xmath72 ) and 10@xmath69 contrast for fainter ( @xmath73 ) targets . in figure
[ gpi - plot ] , a representative value of 1.9@xmath74 is shown , which is the predicted contrast at 1 arcsec separation from a 7th magnitude g0 dwarf . the models shown
are binned to the resolution of gpi in _
band , @xmath75 at 1.65 and are shown as the flux ratio compared to a blackbody with the temperature and radius of a g0 dwarf .
the @xmath0=450 k planet is detectable above the contrast limits in the spectral regions where the planet is bright .
cooler planets ( @xmath0=350 , 250 k ) will be too faint to observe around a g dwarf .
the lbtao system includes a high - contrast imaging instrument optimized for the mid - infrared .
it is capable of 4.8@xmath76 contrast in _ l _
band skemer14 for a bright star and has six narrow band filters spanning 3.04 to 3.78 .
this spectral region is particularly useful for two reasons : first , the planet - star flux ratio is much higher for cool planets in _
l _ and _ m _ bands .
second , this spectral region spans the most prominent methane feature at 3.3 , allowing for the characterization of a prominent atmospheric component . for example
skemer13 used this instrument to observe the hr 8799 system and find that the planets do not have strong methane absorption , inferring that methane must be in disequilibrium .
similar to gpi , a 450 k planet around a g dwarf would be detectable with lbtao but a 350 or 250 k planet around a g dwarf would not be .
note that , because it operates in the mid - infrared where the sky is bright , the lbtao system is also background limited ; it can only observe objects brighter than @xmath118th magnitude in l@xmath77 .
this limit approximately corresponds to a 350 k object at 10 pc .
more distant objects will therefore be background limited ; closer objects will be contrast limited .
the leech campaign is currently using this instrument to survey stars in the solar neighborhood and discover new planets skemer14 .
discrepancies between models and observations may indicate that the physics or ingredients in the model are either incorrect or incomplete .
for example , there are many sources of uncertainty in the molecular opacity databases .
alkali opacity uncertainties may affect the optical and near - infrared .
the handful of y dwarfs that have observed @xmath62 color suggest that these objects are bluer in this color than the models .
this spectral region is controlled in part by the decline of the strongly pressure - broadened alkali opacity , so this mismatch could be due to the treatment used for the alkali opacity . for these calculations , we use the line broadening treatment outlined in burrows00b , which is somewhat _ ad hoc _ and potentially creates some inaccuracies in the model flux in @xmath52 and @xmath53 bands . a calculation of the molecular potentials for potassium and sodium in these high pressure environments , as is carried out in allard05 , allard07 , would improve the accuracy of these models .
subsolar metallicities also make @xmath62 colors bluer mace13a , burrows06 , so it is important to model the opacity correctly to interpret metallicity measurements using this spectral region .
another known source of opacity uncertainty is in the methane line list ; the list used in this study is known to be incomplete .
the new line list from yurchenko13 and yurchenko14 has over nine billion lines and will vastly improve accuracy of the treatment of methane .
it is possible that we may be missing important physical processes in our models that occur in actual y dwarfs . for example
, an assumption we make when calculating the spectra is that the atmospheres are in radiative convective equilibrium .
it is inevitable that real brown dwarf atmospheres have higher levels of complexity than these simple assumptions .
the upper atmospheres of brown dwarfs could be heated by a similar mechanism to that which creates jupiter s thermosphere , in which energy is deposited high in the atmosphere by dissipation of gravity waves young97 .
if indeed the upper atmosphere is hotter than radiative convective equilibrium models , this would change the observed spectrum . using mid - infrared spectra of l dwarfs , sorahana14
show that several l dwarf spectra can be fit significantly better using a model that allows for this upper atmospheric heating .
in fact , in our equilibrium models the [ 3.6]@xmath78[4.5 ] color is redder by @xmath11 magnitude at some @xmath0 ; simple preliminary models in which we change the p
t profile of the upper atmosphere show that heating high in the atmosphere increases the flux within the methane band centered at 3.3 , which makes the [ 3.6]@xmath78[4.5 ] color bluer , and closer to the observed colors .
another limitation of this study is that the models here include only solar abundances .
we expect y dwarfs and exoplanets to have a range of metallicities and potentially a range of other abundance ratios such as non - solar c / o ratios .
future work will be needed to analyze the effect of abundances on y dwarf spectra and colors .
we consider disequilibrium chemistry of n@xmath2/nh@xmath3 and co / ch@xmath4 , but other molecules such as ph@xmath3 and co@xmath2 may also be out of chemical equilibrium . such unmodeled chemistry could change abundances of molecules we do include , or create additional molecules that we do not include in the calculations . as 1d models , these calculations naturally do not include the effect of 3d dynamics on the cloud structure .
the breakup of the iron and silicate clouds at the l / t transition is a source of continued study ; due to dynamical processes , those clouds dissipate at higher temperatures than 1d cloud settling models predict .
as the sulfide and salt clouds sink more deeply into the atmosphere as the brown dwarf cools , they may also break up and disappear from spectra at higher temperatures than our treatment predicts .
further study , including dynamical effects , will be necessary to understand this phenomenon across the brown dwarf spectral sequence . to our knowledge ,
the only previous comprehensive set of models analyzed and published for y dwarfs including water clouds using a cloud model were those in burrows03b .
our approaches and results are overall very similar ; we assume chemical equilibrium , radiative convective equilibrium , incorporate a cloud model , and publish spectra and colors of the coolest brown dwarfs . like burrows03b
, we predict the growing importance of methane and ammonia absorption over the t to y sequence , the weakening of the alkali absorption , and a reversal in the blueward trend in near - ir colors of the t dwarfs around 400 k. our spectra differ in details due largely to changes in the line lists over the past decade ; we have continuously improved our opacity database over the last ten years freedman08 .
most relevant here , we are using updated treatments for both ammonia ( which becomes very important for y dwarfs ) and h@xmath2 collision - induced absorption .
one of the conclusions presented in burrows03b is that the water clouds do not significantly affect the spectra of y dwarfs and it is on this point that we differ most significantly .
the differences lie in our treatment of the cloud model for the water clouds .
burrows03b uses the cloud model presented in cooper03 , which results in a uniform distribution of very large particles ( 20150 ) within a single pressure scale height .
since our cloud particles are much smaller ( @xmath11 - 20 ) , we have far more particles to create a cloud of the same mass .
this means that when our water cloud forms , it is much more optically thick . for a more detailed comparison of the cloud models themselves
see marley03 which describes the challenges of modeling clouds in brown dwarfs and the problems inherent to different approaches .
wisepa j182831.08 + 265037.8 ( hereafter wisep j1828 + 2650 ) is a particularly interesting object .
its near - infrared colors are inconsistent with the models and with the other y dwarfs .
the peculiarities of the near - infrared colors and comparisons to models led cushing11 and kirkpatrick12 to classify it as a > y0 and >
y2 respectively ; comparisons to models gave cushing11 a temperature estimate of @xmath0@xmath79300 k. however , for a brown dwarf that cold to have the measured mid - infrared luminosity , it would need to have an unphysically large radius , leading leggett13 to suggest that it is an unresolved binary .
dupuy13 revised the parallax measurement and found that based on its luminosity , wisep j1828 + 2650 likely has a temperature closer to the late t dwarfs , 500600 k , and that even if it is a binary , those components must still be 400500 k. using the models presented here , it is not possible to fit all of the near - infrared colors of wisep j1828 + 2650 using a lower gravity model .
for example , the models predict that a brighter _ j _ band flux and corresponding bluer @xmath6 color at lower gravity , due to the decreased strength of the alkali absorption in a lower gravity photosphere , whereas for this object we observe an extremely red @xmath6 color .
additional spectroscopic data at near- and mid - infrared wavelengths will be required to determine whether wisep j1828 + 2650 is indeed a prototypical @xmath0@xmath79300 k y dwarf , or a peculiar version of the t8y0 spectral classes .
as brown dwarfs approach the effective temperatures of the solar system s planets , volatile clouds will form in their atmospheres .
the first and most massive type of volatile cloud that forms is water ice clouds .
water ice clouds form in objects cooler than effective temperatures of @xmath1400 k. in order to converge atmospheric temperature structures self - consistently with both clouds and chemistry , we calculate models in which , like water clouds in the solar system planets , the clouds heterogeneously cover the surface ( `` patchy '' clouds ) . our model grid covers the y dwarf spectral class as well as giant planets with the same effective temperatures , from @xmath0=200450 k and log g=3.05.0 . \1 . while water condenses high in the atmospheres of all objects below @xmath0@xmath1400 k , these clouds do not become optically thick until the object has cooled to 350375 k. this result means that for the current set of y dwarfs warmer than 400 k , water clouds will not strongly effect their spectra .
water clouds , unlike other clouds in brown dwarf atmospheres , are very much non - gray absorbers . using the am01 cloud model , water ice particle sizes range from @xmath1120 . for these particle sizes
, the ice particles are strongly scattering in the optical through _ j _ band and do not change the spectra significantly at those wavelengths .
the ice particles absorb strongly in the infrared with prominent features , the strongest of which is at 2.8 .
h@xmath2o , nh@xmath3 , ch@xmath4 , and h@xmath2 cia are the dominant opacity sources in y dwarf atmospheres .
less abundant species such as ph@xmath3 may also be observable at 44.6 , as well as h@xmath2s in _ h _ and _ y _ bands and the alkalis in the optical .
jwst s miri and nirspec instruments will be well - suited to characterizing cool brown dwarfs . @xmath0=400500
k objects will be observable across their near- and mid - infrared spectra , and even @xmath0=200 k objects will be observable in the spectral window region between 3.8 and 5.0 and at some wavelengths between 8 and 17 .
existing and upcoming ground - based instruments such as gpi , sphere , and lbtao will be capable of directly - imaging @xmath0=400500 k planets around nearby g dwarfs .
the authors acknowledge gregory mace , michael liu , and the anonymous referee for comments that improved the paper .
we also thank andy skemer for information on the observing capabilities of the lbt .
we also acknowledge the database of ultracool parallaxes maintained by trent dupuy .
jjf acknowledges the support of nsf grant ast-1312545 , ds acknowledges the support of nasa astrophysics theory grant nnh11aq54i , and msm and kl acknowledge the support of the nasa astrophysics theory and origins programs . ,
b. , et al .
2006 , in presented at the society of photo - optical instrumentation engineers ( spie ) conference , vol .
6272 , advances in adaptive optics ii .
edited by ellerbroek , brent l. ; bonaccini calia , domenico .
proceedings of the spie , volume 6272 , pp .
62720l ( 2006 ) . | the formation of clouds affects brown dwarf and planetary atmospheres of nearly all effective temperatures .
iron and silicate condense in l dwarf atmospheres and dissipate at the l / t transition .
minor species such as sulfides and salts condense in mid - late t dwarfs . for brown dwarfs below @xmath0@xmath1450
k , water condenses in the upper atmosphere to form ice clouds .
currently over a dozen objects in this temperature range have been discovered , and few previous theoretical studies have addressed the effect of water clouds on brown dwarf or exoplanetary spectra . here
we present a new grid of models that include the effect of water cloud opacity .
we find that they become optically thick in objects below @xmath0@xmath1350375 k. unlike refractory cloud materials , water ice particles are significantly non - gray absorbers ; they predominantly scatter at optical wavelengths through _ j _ band and absorb in the infrared with prominent features , the strongest of which is at 2.8 .
h@xmath2o , nh@xmath3 , ch@xmath4 , and h@xmath2 cia are dominant opacity sources ; less abundant species such as may also be detectable , including the alkalis , h@xmath2s , and ph@xmath3 .
ph@xmath3 , which has been detected in jupiter , is expected to have a strong signature in the mid - infrared at 4.3 in y dwarfs around @xmath0=450 k ; if disequilibrium chemistry increases the abundance of ph@xmath3 , it may be detectable over a wider effective temperature range than models predict .
we show results incorporating disequilibrium nitrogen and carbon chemistry and predict signatures of low gravity in planetary - mass objects .
lastly , we make predictions for the observability of y dwarfs and planets with existing and future instruments including the _ james webb space telescope _ and _ gemini planet imager_. |
the first and most radical departure of string theory from a theory of elementary point particles is in the nature of its elementary degrees of freedom .
the rest of the structure of string theory rather conservatively follows established principles : relativistic invariance ( generalized to include supersymmetry ) , quantum mechanics , locality of interaction , and internal mathematical consistency .
together , these result in new symmetries and properties that open up conceptual possibilities inconceivable in elementary point particle theories . in this note
we focus on a few such properties that make it possible to imagine a resolution of the initial singularity problem in cosmology . as will be evident , this possibility is directly traceable to the basic degrees of freedom of a string . in the next section
we begin with a detailed discussion of the degrees of freedom in string theory , contrasting them with elementary point particle theory , and describing the nature of ` particles ' in string theory .
the ` string uncertainty principle ' ( which modifies the heisenberg uncertainty principle to state that there is a minimum observable size in a world whose fundamental constituents are strings ) is introduced in terms of string wavefunctions .
the ` t - duality ' symmetry ( indistinguishability of very large and very small universes ) is described .
section iii discusses some features of the thermodynamics of an ideal gas of strings at very high energy densities .
this suggests the possibility that temperature and pressure will have finite limiting values in a string universe . in section
iv the preceding material is applied to string cosmology .
we discuss a thought experiment for measuring the size of a very small universe in the context of string theory .
this prompts a shift of perspective regarding what one means by a small universe , and from the new vantage point the initial singularity problem disappears ( brandenberger and vafa 1989 ) .
we conclude with some cautionary remarks and some questions of possible relevance to observations . while much of the material presented here is a review of existing literature , there are some points which do not seem to have appeared elsewhere .
one is the expression and physical interpretation of the pressure of an ideal string gas .
another is the derivation of the string uncertainty principle from string wavefunctions .
* classical string modes in flat euclidean space * : the classical configuration of a scalar particle is described by specifying its position , a point in space , say @xmath0 .
the classical configuration of a closed bosonic string is similarly described by specifying its position , in this case a closed curve embedded in space .
the latter can be described by specifying a continuous map @xmath1 from the unit circle into space , @xmath2 . if @xmath3 $ ] is a coordinate on @xmath4 and @xmath5 are coordinates of @xmath0 , the map @xmath1 is specified by @xmath6 continuous functions @xmath7 that are periodic , @xmath8 .
the point @xmath9 traces out a closed curve in space as @xmath10 traverses from @xmath11 to @xmath12 .
it is convenient to fourier decompose the functions @xmath7 : @xmath13 the infinite - tuple @xmath14 equivalently describes the map @xmath1 or the classical string configuration .
@xmath15 is a ` centre of mass ' coordinate describing the average position of the string .
@xmath16 describe the extension of the string in space .
it is evident that a string has an infinitely richer repertoire of classical configurations than a point particle .
e.g. , in the configuration where all @xmath16 are zero the string has no extension and becomes a point particle at the point @xmath17 in space .
if @xmath18 and all other @xmath19 are zero , the string configuration is a circle of radius @xmath20 in a plane parallel to the @xmath21 plane with centre @xmath17 .
if some of the @xmath19 with higher @xmath22 are also non - zero , the configuration will become a ` wiggly ' circle , and wiggles become finer with increasing @xmath22 ( typical radius of curvature of a wiggle caused by the @xmath23 mode is @xmath24 ) .
the length of the string in a classical configuration specified by a map @xmath1 is given by @xmath25 , and it is evident that this can be expressed in terms of the @xmath16 .
0.3 cm * string wavefunctions ; particle like states * : in the quantum theory a state can be described by a wavefunction , a complex valued function over the configuration space . for a point particle
the configuration space is just @xmath0 ( real space ) , hence a wavefunction is a map @xmath26 . for a string
let us take the configuration space to be the space @xmath27 of infinite - tuples @xmath14 or , equivalently , the set of maps @xmath28 .
then a string wavefunction is a map @xmath29 , assigning , to every string configuration @xmath1 or infinite - tuple @xmath14 , the complex number @xmath30\equiv \psi({\bf x}_0,\{{\bf x}_n,\tilde{\bf x}_n\})$ ] .
( we ignore for the moment the subtlety that strictly @xmath31 ought to be a map from @xmath27 modulo diffeomorphisms of the circle . ) just as for a point particle @xmath32 stands for the probability density of finding the particle at the point @xmath1 in space when it is in the state @xmath31 , so similarly @xmath33 represents the probability density ( in the infinite dimensional space @xmath27 ) of finding the string in the configuration @xmath14 when it is in the state @xmath31 .
it is instructive to consider a few sample string wavefunctions to see what the string would look like in the corresponding states .
0.1 cm * 1*. @xmath30 = \delta^d({\bf x}_0-{\bf x}_0 ) \prod_{n=1}^{\infty } \delta^d({\bf x}_n-{\bf x}_n)\delta^d(\tilde{\bf x}_n-\tilde{\bf x}_n)$ ] .
this wavefunction has support only over a single classical configuration described by the infinite - tuple @xmath34 .
thus an ` observation ' of the components of the infinite - tuple @xmath14 in this state ( assuming such an experiment can be devised , a point we return to later ) would yield a definite answer for each component , with the conclusion that in this quantum state the string looks like a classical string at the configuration @xmath34 .
in particular if the @xmath35 are all zero , the string would appear to be a point particle ( with no extension ) localized at @xmath36 .
0.1 cm * 2*. @xmath30 = \delta^d({\bf x_0}-{\bf x}_0 ) [ \prod_{i=1}^d \phi_i(x^i_1)\tilde{\phi}_i(\tilde{x}^i_1 ) ] [ \prod_{n=2}^{\infty}\delta^d({\bf x}_n)\delta^d(\tilde{\bf x}_n)]$ ] .
case i ) @xmath37 for all @xmath38 and @xmath39 is a function that has support only in a small region around the origin ( e.g. , @xmath40 ) .
in such a state , the fourier components for @xmath41 have a spread of order @xmath42 , and consequently the string will appear to have an average position still @xmath36 but will now have an indefinite extension in space of order @xmath42 .
if @xmath42 is small compared to the probes , it will still appear particle - like ( a fuzz of size @xmath42 around @xmath36 ) .
if @xmath43 ( as in the example ) , the fuzz will appear spherically symmetric.case ii ) the @xmath44 still have support in a small region of order @xmath42 but are now different for different @xmath38 .
then the ` fuzz of size @xmath42 around @xmath36 ' will no longer appear isotropic .
( the spin of the photon , graviton , etc .
, in string theory is due to such internal spatial structure of the corresponding states .
there @xmath42 is of order planck length . )
it should also be evident that the visual picture is not qualitatively very different if some of the higher modes are also given a fuzz around the origin . since these states are not eigenstates of the @xmath16 , they are also not eigenstates of the length operator . when higher modes are allowed to be nonzero , the expectation value of @xmath45 increases ( more wiggles in the classical configurations over which the wavefunction has support ) .
0.1 cm * 3*. @xmath30 = e^{-i{\bf p.x_0 } } [ \prod_{i=1}^d \phi_i(x^i_1)\tilde{\phi}_i(\tilde{x}^i_1 ) ] [ \prod_{n=2}^{\infty}\delta^d({\bf x}_n)\delta^d(\tilde{\bf x}_n)]$ ] .
if @xmath44 are as in 2 this wavefunction represents the same particle - like string state , but now with a definite centre of mass momentum @xmath46 rather than a definite centre of mass position @xmath36 . the above discussion made no reference to string dynamics .
we only discussed the degrees of freedom in string theory and how particle like states can be imagined out of strings .
it turns out that the most natural dynamics for strings in fact makes such states appear as eigenstates of the energy .
the latter turn out to be plane waves for the centre of mass coordinate and wavefunctions localized around the origin ( in fact harmonic oscillator wavefunctions ) for the wiggle modes .
0.3 cm * dynamics : * as a single string moves , its trajectory traces out a two dimensional surface embedded in space .
this can be described by introducing a fictitious parameter @xmath47 taking values in the interval @xmath48 $ ] of the real line and letting @xmath1 be a map from @xmath49 into space .
the image of @xmath49 under the map is a string trajectory . to specify the dynamics for a single string we need to introduce an action for every such trajectory .
it is rather cumbersome to introduce a relativistic invariant action in terms of the space variables @xmath50 alone .
it is more convenient to let the time coordinate @xmath51 also be a function of @xmath47 and @xmath10 and to write an action for @xmath52 , which describes a string trajectory , or ` worldsheet ' , in @xmath53 dimensional minkowski spacetime .
the action is taken to be proportional to the area of the worldsheet ( in analogy with the action of a relativistic point particle s trajectory , which is proportional to the length of the corresponding worldline ) . @xmath54=-{1
\over { 2\pi\alpha'}}\times { \rm area\ of\ worldsheet } = - { 1 \over { 2\pi\alpha'}}\int_{\tau_1}^{\tau_2 } d\tau\int_0^\pi d\sigma \sqrt{-\det \gamma},\ ] ] where @xmath55 is the metric on the worldsheet induced from the spacetime metric @xmath56 , and @xmath57 refer to worldsheet coordinates @xmath58 .
this dynamics has relativistic invariance in that one can define the generators of spacetime translations @xmath59 , and of rotations and boosts @xmath60 ( see scherk ( 1975 ) for a review ) whose poisson brackets satisfy the poincare algebra .
the @xmath59 are canonically conjugate to the centre of mass coordinates @xmath61 . the proportionality constant @xmath62 is called the string tension or mass per unit length since the energy @xmath63 of any static classical configuration @xmath64 turns out to be @xmath65 , where @xmath45 is the length .
the action ( [ action1 ] ) is nonlinear in the derivatives @xmath66 .
but it has an infinite local symmetry corresponding to the reparametrizations @xmath67 which can be used to bring @xmath68 into the form @xmath69 , with @xmath70 . in this ` conformal gauge ' @xmath71 and the action reduces to the free scalar field form @xmath72 $ ] . substituting the mode expansion @xmath73 $ ] in this action yields @xmath74 with @xmath75,\ ] ] where the dot denotes derivative w.r.t .
thus the area law dynamics automatically prescribes a free particle like role for the centre of mass mode and a simple harmonic oscillator like role for the wiggle modes @xmath76 with frequency @xmath77 .
it is then evident that the quantum states @xmath31 of the system in question ( a single free string in minkowski spacetime ) will be specified by the set of quantum numbers @xmath78 where @xmath79 is a momentum conjugate to the centre of mass mode ( and is the eigenvalue of the spacetime translation generator @xmath80 ) and @xmath81 are harmonic oscillator excitation level quantum numbers for the modes @xmath76 .
the conformal gauge does not fix the freedom of reparametrizations completely .
all the @xmath82 are not independent variables .
one can show that the independent variables can be taken to be @xmath83 where @xmath84 goes only over the ` transverse ' spatial indices @xmath85 .
classically , once these are known as functions of @xmath47 , all others , @xmath86 and @xmath87 , are determined as functions of @xmath47 by the constraints and hence the string worldsheet is determined .
( roughly speaking , diffeos of @xmath88 eat up two spacetime coordinates @xmath89 excepting the zero mode of @xmath90 . ) 0.3 cm * the spectrum * : quantum mechanically , this means that string states are characterized by the set of quantum numbers @xmath91 , the other quantum numbers being determined in terms of them .
in particular the quantum number @xmath92 , eigenvalue of the energy @xmath63 , is given by @xmath93.\ ] ] the closure of the quantum lorentz algebra fixes @xmath94 .
this defines the spectrum of the free closed string in @xmath0 .
the wavefunction of this state in the basis of independent coordinate variables then follows from inspection of ( [ lagrangian ] ) : @xmath95},\ ] ] where @xmath96 is the @xmath97 hermite polynomial . on the oscillators coming from the fact that there is no preferred point in the @xmath10 direction along the string
the form of @xmath98 is more complicated than the usual @xmath99 because the @xmath76 defined here do not represent the left and right moving modes respectively .
we henceforth assume that @xmath91 in ( [ spectrum1 ] ) and ( [ wave1 ] ) are such that this constraint is satisfied . ]
it is interesting to compare this with the third wavefunction discussed earlier .
the @xmath100 there are replaced by harmonic oscillator wavefunctions whose width is order @xmath101 .
the @xmath102-functions of the higher @xmath22 modes are also replaced by harmonic oscillator wavefunctions of width @xmath103 .
thus a string in a state with quantum numbers @xmath91 in which the @xmath104 are not too large , when observed via probes of energy @xmath105 , will effectively appear to be a particle with some internal structure of size @xmath106 and momentum @xmath46 .
( a large value of @xmath107 would elongate the size in the @xmath108 direction to @xmath109 .
these are sometimes referred to as the ` really stringy ' states . )
it is natural to define the mass @xmath110 of a state as @xmath111 .
( [ spectrum1 ] ) then gives the mass formula in terms the oscillator excitations of the state .
thus different states in the spectrum of a single string can be identified with various particle species having different masses ( which come in units of @xmath112 ) and momenta .
in addition to the length scale @xmath113 , string theory has a dimensionless coupling constant @xmath114 , which represents the amplitude that two strings touching each other will fuse into a single string , or the reverse process . between such joinings or splittings
, strings travel freely according to ( [ action1 ] ) .
these rules essentially specify string perturbation theory completely .
the effective interactions of massless particles in the string spectrum ( gravitons , dilatons and antisymmetric tensor particles in the bosonic string and also photons or gauge particles in the heterotic string ) can be determined from these considerations ( see , e.g. , green , schwarz and witten 1987 ) .
in particular the gravitational constant in @xmath6 spatial dimensions is given by @xmath115 .
thus newton s constant ( @xmath116 in @xmath117 ) is given by @xmath118 ( assuming higher dimensions compactify to radii @xmath119 ) .
in other words , string theory reproduces classical einstein gravity at low energies if we choose its two parameters @xmath120 and @xmath114 to satisfy @xmath121 ( planck length ) . note that the ` string length scale ' @xmath113 is @xmath122 if @xmath123 ( strong coupling ) and is much larger than @xmath124 if @xmath125 ( weak coupling ) .
0.3 cm * the size of strings ; string uncertainty principle * : what is the size of the string in the state ( [ wave1 ] ) ? consider the transverse ` mean - square spread ' operator ( mitchell and turok 1987 ; karliner , klebanov and susskind 1988 ) @xmath126 .
( [ wave1 ] ) is not an eigenstate of this , but we can ask for its expectation value . consider the ground state of all the oscillator modes , @xmath127 ( the scalar tachyon ) .
the expectation value of @xmath128 in this state is @xmath129 which diverges logarithmically because each of the @xmath130 modes makes a finite contribution .
this divergence is empirically unobservable because an experiment does not observe @xmath128 or the @xmath131 directly .
a typical experiment involves scattering a probe off the string . in order for the probe to ` see ' the @xmath131 mode it must interact with it and excite it from its ground state .
this would cost energy @xmath132 from ( [ spectrum1 ] ) . a probe with a finite energy @xmath133
would only excite a finite number of oscillator modes ; therefore the infinite sum in @xmath128 should be cutoff at a finite value of @xmath22 depending upon the energy of the probe . for @xmath134 , none of the oscillator modes will be excited and the string will effectively look like a point particle .
probes with @xmath135 will see the state as a fuzz of size @xmath113 .
for probes with energy @xmath136 the fuzz size will increase .
the maximum fuzz size is obtained if all the energy of the probe goes into exciting only the @xmath41 mode .
its excitation level is then @xmath137 from ( [ spectrum1 ] ) , and the consequent root mean square spread in space of the target string wavefunction ( through its @xmath138 and @xmath139 modes ) is @xmath140 .
the size _ grows _ with the energy of the probe .
this is a new term that must be added to the usual uncertainty in position @xmath141 coming from heisenberg s uncertainty principle . setting @xmath142 ( in @xmath143 dimensions ) and putting back units we get the ` string uncertainty principle ' [ stringup ] x ~c e + g_n e . minimizing this w.r.t .
@xmath133 , one finds the smallest observable length scale in string theory @xmath144 .
here we assumed that all the energy goes into exciting only the @xmath41 modes .
if the energy is shared with the higher modes whose wavefunctions are more strongly localized , the spread will be smaller .
for example if one assumes that each mode @xmath22 upto some maximum @xmath145 is excited to its first excited level ( @xmath146 for @xmath147 and zero thereafter ) , then one finds @xmath148 and the sum in @xmath128 should be cutoff at this value .
then , instead of @xmath149 , one gets @xmath150^{1/2}$ ] , modifying the second term in ( [ stringup ] ) .
different choices putting in less energy into the higher modes than the second case would yield @xmath151 between these two values .
for all these choices it remains true that @xmath152 .
these two forms of the string uncertainty principle were conjectured , respectively , by gross ( 1989 ) and amati , ciafaloni and veneziano ( 1989 ) from studies of string scattering amplitudes at high energies at all loops ( gross and mende 1988 ; amati , ciafaloni and veneziano 1988 ) .
it is interesting that both forms as well as intermediate ones can be derived by elementary considerations of string wavefunctions using different assumptions of how the energy is distributed among the oscillator modes of the target .
one can ask , what determines the actual distribution of the probe kinetic energy among the various oscillator modes of the target ?
this needs further investigation .
the analysis of amati , ciafaloni and veneziano ( 1989 ) suggests that the scattering angle plays a role in determining it . from the above it
is evident that the smallest observable length of any object in a string universe ( where everything , objects and probes , is ultimately made of strings ) is of the order of @xmath113 .
this is a direct consequence of the new wiggle degrees of freedom of strings .
0.3 cm * gravitational collapse , black holes , random walks * : in the above scattering experiment if too much energy is deposited in the higher @xmath22 modes of the target string , its size can become smaller than its schwarzchild radius and it can suffer gravitational collapse .
for example , an interesting choice of energy distribution among the modes of the target is to assume that it is thermal .
that is , assume that all oscillator states of the target having a total energy @xmath133 are equally likely .
( to avoid a violation of unitarity , the probe , also stringy , carries away all the correlations . )
what is @xmath153 in such an ensemble ?
this question has been investigated by mitchell and turok ( 1987 ) and aharonov , englert and orloff ( 1987 ) in a different context .
it turns out that @xmath154 .
taking @xmath155 to be the size of the string , such a string would suffer gravitational collapse if its schwarzchild radius exceeded @xmath20 , or its energy exceeded @xmath156 ( in @xmath143 dimensions ) .
the entropy @xmath157 of this ` black hole ' , given by @xmath158 ( see next section ) would be @xmath159 , proportional to its ` area ' .
the expectation value of the length of the string in the thermal state is @xmath160 , as for a classical string configuration ( or a cosmic string ) .
thus in this state the string resembles a random walk in space with step length @xmath113 , since @xmath161 ( mitchell and turok 1987 ; aharonov , englert and orloff 1987 ) .
.3 cm * string spectrum in a compact space * : spacetime itself is characterized by a metric ( at least on scales familiar to us ) , to determine which we must measure lengths . if the smallest measureable length is in principle @xmath162 , this must ultimately reflect on the smallest conceivable size of the universe in string theory . to study this
more precisely , we now consider strings in a finite sized space .
consider a toroidal compactification of space with a radius @xmath163 , i.e. , the coordinate @xmath164 of space ( @xmath165 ) is identified with @xmath166 with @xmath167 an integer . while we describe this special case for simplicity and clarity , many of the consequences for cosmology discussed later are valid for a much larger class of compactifications .
then classical string configurations have another mode , modifying ( [ fourier ] ) to @xmath168 where @xmath169 with @xmath170 an integer .
as @xmath10 runs from @xmath11 to @xmath12 , @xmath164 runs from @xmath171 to @xmath172 , the string therefore winds around the universe in the @xmath173 direction @xmath170 times .
this adds a term @xmath174 to ( [ lagrangian ] ) , and @xmath175 to ( [ spectrum1 ] ) . in compact space @xmath176
is also quantized ( @xmath177 integer ) , and the spectrum is now given by ( green , schwarz and brink 1982 ) @xmath178.\ ] ] this spectrum maps into itself under the transformation [ duality ] r /r .
this is evident from the fact that the state with quantum numbers @xmath179 in a universe of radius @xmath163 has the same energy as the state @xmath180 in a universe of radius @xmath181 .
the interchange @xmath182 together with the transformation ( [ duality ] ) does not alter the r.h.s . of ( [ spectrum2 ] ) .
thus at the level of the free spectrum , string theory does not distinguish between a universe of size @xmath163 and a universe of size @xmath181 .
this symmetry is also respected by string interactions : the amplitude of a process in a universe of size @xmath163 with a given set of external states is the same as the amplitude in a universe of size @xmath181 of the ` dual ' set of states ( obtained by interchanging @xmath183 and @xmath184 quantum numbers for each state in the first set ) .
this symmetry , known as target - space duality or ` t - duality ' was found by kikkawa and yamasaki ( 1984 ) , sakai and senda ( 1986 ) , nair , shapere , strominger and wilczek ( 1987 ) , sathiapalan ( 1987 ) , and ginsparg and vafa ( 1987 ) .
0.3 cm * a new periodic spatial coordinate and wavefunctions of winding states * : in addition to the coordinate @xmath185 which is compact with period @xmath186 , there exists another spatial coordinate @xmath187 in string theory with period @xmath188 ( sathiapalan 1987 ) .
this is just the conjugate variable to the operator @xmath189 whose eigenvalue is @xmath190 ( just as @xmath185 is conjugate to @xmath191 whose eigenvalue is @xmath176 ) .
formally , define @xmath192 and @xmath193 , where the sum goes over @xmath194 and @xmath195 denotes @xmath196 for brevity .
it follows that @xmath197 = i\delta^{ij}$ ] .
since @xmath184 is quantized on an integer lattice , it is easy to see that @xmath198 for any @xmath199 .
i.e. , the points @xmath187 and @xmath200 in this ` dual space ' are physically indistinguishable .
the wavefunctions are now given by @xmath201 , where the r.h.s differs from that of ( [ wave1 ] ) by the factor @xmath202 being replaced by @xmath203 .
the physical significance of this ` dual position coordinate ' will be discussed in the last section .
* the partition function and the density of states * : in order to study the very early universe in the context of string theory , it is important to know how a very hot gas of superstrings behaves .
consider the thermal partition function of a string gas : [ z ] z(,r)= _
e^-e_(r ) . here
@xmath204 labels a state with @xmath205 strings , the quantum numbers of the @xmath206 string being given by @xmath207 .
each @xmath207 in turn stands for the full set of quantum numbers @xmath208 for the @xmath206 string .
@xmath209 is the energy of the multi - string state @xmath210 in a universe of radius @xmath163 , and in the ideal gas approximation is given by the sum of the individual single string energies : @xmath211 where @xmath212 is given by ( [ spectrum2 ] ) for closed bosonic strings .
( for superstrings the formula for @xmath213 is modified by additional degrees of freedom but retains the essential character needed for subsequent discussion . )
the sum over @xmath210 includes a sum over all individual string states for a fixed @xmath205 and a sum over @xmath205 from zero to infinity .
this partition function has a number of interesting properties .
first , it has singularities in the complex @xmath214 plane ( other than the usual @xmath215 singularity ) even at finite volume . in point particle field theories , singularities , which are usually signatures of phase transitions ,
arise only in the thermodynamic limit .
in the string case they arise at finite volume because even a single string has infinite degrees of freedom .
the location of the right most singularity , @xmath216 ( @xmath217 , where @xmath218 is known as the hagedorn temperature ( hagedorn 1965 ; huang and weinberg 1970 ) ) , is proportional to the only length scale in the theory , @xmath219 . the proportionality constant is independent of the size of the box ( or universe ) and other details of compactification ( antoniadis , ellis and nanopoulos 1987 ; axenides , ellis and kounnas 1988 ) but dependent only on the type of string theory , bosonic ( @xmath220 ) , type ii superstring ( @xmath221 ) , or heterotic ( @xmath222 ) .
as long as space is compact , the singularity is universally a simple pole ( brandenberger and vafa 1989 ; deo , jain and tan ( djt ) 1989a ) .
there is a representation of @xmath223 due to obrien and tan ( 1987 ) ( see also maclain and roth 1987 ; mcguigan 1988 ) which is useful in determining its analytic structure in the complex @xmath214 plane .
it turns out that there is an infinite number of singularities to the left of @xmath216 ( djt 1989a ) whose locations in general depend upon the radius of universe . for universes much larger than @xmath101 ( and also , by duality , for universes much smaller than @xmath101 ) , a number of these singularities approach @xmath216 .
second , since the spectrum exhibits duality , so do the partition function and density of states @xmath224 : [ zduality ] z(,r)=z ( , ) ( e , r)= ( e , ) .
this follows from the fact that for every @xmath210 there exists an @xmath225 ( obtained from @xmath210 by interchanging the momentum and winding numbers of every string in the state @xmath210 ) such that @xmath226 .
third , the behaviour of @xmath223 near @xmath216 is such that at temperatures close to the hagedorn temperature , fluctuations are large and invalidate the use of the canonical ensemble for deducing the thermodynamic properties of the string gas .
one is forced to use the more fundamental microcanonical ensemble , defined by @xmath227 ( frautschi 1971 ; carlitz 1972 ; mitchell and turok 1987 ; turok 1989 ) . finally ,
since @xmath228 and @xmath229 are related by a laplace transform , the leading large energy behaviour of @xmath230 is controlled by the behaviour of @xmath223 near its singularities , and can be determined by a contour deformation technique ( djt 1989a , 1991 ) . at large radius ( @xmath231 ) and at energy densities above the ` hagedorn energy density ' @xmath232 , the density of states is given by ( deo , jain , narayan , tan 1992 ( djnt ) ) [ density ] ( e , r ) _ 0 e^_0 e + a_0 v [ 1-(e , r ) ] , ( e , r)= ( _ 0 e)^2|d - 1 e^-(_0-_1)(e-_0 v ) .
here we use the notation that @xmath6 represents the total number of spatial dimensions , all of them compact ( @xmath94 for bosonic strings and @xmath233 for superstrings and heterotic strings ) .
@xmath234 is the number of spatial dimensions that have large radius @xmath231 ; the remaining @xmath235 dimensions are assumed to have radii @xmath106 .
@xmath236 is the volume of the large dimensions .
@xmath237 is a constant of order @xmath238 .
@xmath239 is the singularity of @xmath240 closest to @xmath216 ; @xmath241 .
the formula ( [ density ] ) is valid for @xmath242 and for energy density @xmath243 greater than @xmath244 .
@xmath245 should be large enough ( greater than @xmath246 ) so that @xmath247 .
0.3 cm * thermodynamic properties ; physical interpretation in terms of degrees of freedom * : the thermodynamic properties of the gas are determined by ( [ density ] ) .
the entropy @xmath248 is given by [ entropy ] s(e , r ) _ 0 e + a_0 v + ( 1- ) , from which one finds the temperature @xmath249^{-1}$ ] to be [ temp ] t(e , r ) t_h(1-_0-_1 ) , and the pressure @xmath250 [ pressure ] p(e , r ) t_h a_0 ( 1- [ 1+_0_0 ( 2_0 + |d - 2 ) ] ) . thus both the temperature and pressure of the string universe reach asymptotic values determined by the string length scale @xmath120 at energy densities above hagedorn ; corrections to these asymptotic values are exponentially suppressed above these energy densities .
the physical reasons for this are as follows .
the leading contribution to the density of states of a string gas grows as the exponential of a linear function of @xmath133 , unlike for a gas of point particles where it grows exponentially with a sublinear function .
this is because the number of oscillator states at a fixed large value of @xmath251 , grows as @xmath252 .
this is just the hardy - ramanujam asymptotic formula for the number of partitions of a large positive integer @xmath253 into non - negative integers , a result in number theory .
thus , even for a single string the density of states grows exponentially @xmath254 with energy ( since @xmath255 from ( [ spectrum2 ] ) ) .
by contrast the contribution to the density of states from the momentum and winding modes is very small .
e.g. , for a single particle , for which only momentum modes contribute , the density of states grows only as a power @xmath256 .
thus at large energies it is entropically favourable for the energy to go into oscillator modes rather than momentum or winding modes .
a term in the entropy of the gas that is linear in energy gives rise to a constant , i.e. , energy independent temperature .
the form of the subleading corrections ( which is due to an interplay between oscillator , momentum and winding modes ) tells us that @xmath218 is an upper limiting temperature .
figure 1 displays the behaviour of temperature as a function of energy for an ideal gas of strings and contrasts it with an ideal gas of point particles .
the reason for the asymptotic pressure is as follows : the leading term @xmath257 in the entropy does not contribute to the pressure because it is independent of the volume at constant energy ; this is because the oscillator mode contribution to the energy ( [ spectrum2 ] ) is volume independent .
the second term , @xmath258 , should be interpreted as the contribution of momentum modes .
it is proportional to volume just like for a gas of point particles , due to the translational degrees of freedom .
( a pure winding mode gas by contrast will give a contribution proportional to @xmath259 , which is subleading for large radii . ) for an ordinary point particle gas the entropy also depends upon the energy density @xmath260v$ ] , from which the usual expression @xmath261 with @xmath262 follows .
in the string case above , the coefficient of @xmath263 is just a constant , @xmath237 .
the physical reason is that above hagedorn energy density , the energy density in momentum modes is a constant independent of the total energy density .
if more energy is pumped into the box , it goes primarily into oscillator modes , which are entropically favoured , than into momentum modes .
conversely , if some energy is taken out of the box ( keeping the total density still above hagedorn ) , it is primarily extracted from the oscillator modes keeping the energy in the momentum modes essentially the same .
equivalently , if one expands the volume slightly keeping energy the same ( this is what is implied by the derivative @xmath264 ) , energy flows from the oscillators to the momentum modes to keep the energy density in the latter constant . thus the energy density in momentum modes ( which are the contributors to pressure , consisting of small strings bouncing around like particles )
is independent of the volume or the total energy density ( as long as the latter is above hagedorn ) and hence is the pressure .
the above argument seems to be consistent with our present picture of how the total energy of the gas is distributed among various strings . in the energy and radius domain under discussion ,
the string gas can be considered to be consisting of broadly speaking two ` components ' , of energies @xmath265 and @xmath266 with @xmath267 .
the first component consists of a few ( @xmath268 $ ] ) large strings which capture most of the energy of the gas ( @xmath269 provided @xmath270 ) . `
large ' strings are those whose energies are @xmath271 or greater .
most of their energy is due to oscillator modes and the wavefunctions of these strings spread across the whole universe ( recall from the previous section that the size of a thermal string of energy @xmath213 is @xmath272 , hence spread is @xmath273 for @xmath274 ) .
if one adopts a classical picture , the universe is stuffed with space filling brownian walks ( see salomonson and skagerstam 1986 ; mitchell and turok 1987 ) .
the second component has fixed total energy @xmath275 and consists of many ( @xmath276 ) small strings ` small ' ranges in size from @xmath277 to @xmath278 , and in energy from zero to @xmath279 ) .
a crucial property of the gas is that as more energy is pumped into the box , it goes into the first component , leaving @xmath266 fixed .
this was qualitatively anticipated by frautschi ( 1971 ) , carlitz ( 1972 ) , mitchell and turok ( 1987 ) , aharanov , englert and orloff ( 1987 ) , and bowick and giddings ( 1989 ) , and made quantitatively explicit in djt ( 1989b , 1991 ) and djnt . this picture is unaltered by the introduction of conservation laws for the total winding number and momentum , even though additional subleading terms arise in the density of states .
0.3 cm * duality ; thermodynamics in small spaces * : we have so far discussed the case of large @xmath133 and large radius @xmath280 . what happens at very small radii ?
this is immediately answered by duality . at @xmath281 , the r.h.s . of ( [ density ] ) has the same form but with @xmath263 replaced by @xmath282 ( now @xmath283 ) .
the same is therefore true of temperature and pressure ( we now define @xmath284 ) . at @xmath285
we find that the leading behaviour of the density of states is still given by ( [ density ] ) , but now @xmath263 is replaced by a slowly varying function of @xmath163 of order @xmath286 , and @xmath287 ( djt 1989a , djnt ) .
the temperature as a function of @xmath133 is still given by ( [ temp])with these replacements .
however the pressure needs to be appropriately defined and interpreted in this domain ( since @xmath288 has to have an extremum at the duality radius , both definitions @xmath250 and @xmath284 imply that @xmath79 passes through a zero at @xmath289 ) .
0.3 cm * inconsistency of string thermodynamics in non - compact spaces * : finally we remark that string thermodynamics seems to be internally consistent only in a compact space .
the reason is that in a noncompact space to define the density of states we have to consider an artificial box of large volume @xmath263 to confine the gas and later take the thermodynamic limit .
this is problematic in string theory because strings are extended objects , they can in principle extend from one wall to another , and render the entropy inextensive .
one can see the problem explicitly at high energy densities @xmath290 when there exist a few strings in the gas whose individual energy is a significant fraction of the total energy @xmath133 .
the spread of their wavefunction is therefore @xmath291 .
let the number of large dimensions ( of radius @xmath292 ) be @xmath234 . then since @xmath293
, these strings have a size @xmath294 .
thus for @xmath242 these strings have a spread much greater than @xmath163 , the size of the universe itself . in a compact universe
this is not a problem ; the string can wrap around the universe many times .
but if the universe were to be noncompact in these @xmath234 directions , then we find that these strings hit the walls of the artificial box with nowhere to expand , leading to an inconsistency of interpretation ( see also djnt ) .
* absence of a temperature singularity * : we now discuss how the above considerations might impinge on cosmology . let us follow our present universe ( assumed compact in all dimensions but with three large dimensions ) backwards in time according to the standard model of cosmology . at the epoch where the energy density in the large dimensions is above @xmath295 but the radius is still much greater than @xmath113 ( this is quite natural in the standard model at early epochs ) , let us assume that the standard model physics is replaced by string theory , and use the ideal gas approximation ( [ density ] ) .
then as we proceed to smaller radii and hence higher energy densities , the temperature and pressure being governed by equations ( [ temp ] ) and ( [ pressure ] ) no longer increase indefinitely ( as they would in any point particle theory ) but flatten out .
the temperature remains flat as @xmath163 approaches @xmath101 and well into the domain @xmath281 ( as long as @xmath296 or @xmath297 ) .
as @xmath163 declines further ( i.e. , @xmath298 increases ) the temperature _
falls_. this is shown in figure 2 .
the behaviour of temperature as a function of radius ( at fixed energy or fixed entropy ) is symmetric about @xmath299 . at very small radius
it does not diverge as it does for a universe made of elementary point particles , but behaves just as for a very large universe .
the string universe has no temperature singularity
. 0.3 cm * physical interpretation of a small universe * : what is the physics of this bizarre behaviour ?
this was discussed by brandenberger and vafa ( 1989 ) , even before the precise expression ( [ density ] ) for the density of states was known .
they asked the question : how would one measure the size of the universe if it were very small ? for a large periodic box one can imagine sending a light signal ( a localized photon wave - packet ) and measuring the time it takes to come back .
but this experiment would fail in very small box .
the energy of a momentum mode goes as @xmath300 , and a superposition of many such modes is needed to create a localized wave - packet , thereby making it more and more energetically difficult to send a wave - packet in smaller universe . in string theory
the photon is a massless state with some momentum quantum number @xmath301 , winding number zero , and a single oscillator excitation ( the term @xmath302 $ ] in ( [ spectrum2 ] ) , or its analogue for heterotic strings , is zero ) . in today
s universe ( assumed large ) these are easily excited , but it would be energetically very difficult to create photons and send them around in a universe of size @xmath281 ( see ( [ spectrum2 ] ) ) . on the other hand , in a very small universe , particles ` dual ' to the photon , with quantum numbers @xmath303 , some winding number * w * , and the same oscillator quantum numbers as the photon would be easily excited . indeed these would constitute the ` light ' particles of the very small universe .
an observer in this very small universe would hardly think of sending photons to measure the size of his universe ( just as we would not contemplate sending winding modes around ) ; he would use a superposition of the ` dual photon ' modes . by sending such modes he would be measuring the extent of the ` dual position coordinate ' @xmath304 ( recall that @xmath304 is to winding modes
what position @xmath164 is to momentum modes ) .
but , as discussed earlier , that extent is just @xmath305 ; hence observers in a universe of size @xmath281 would find its radius to be not @xmath163 but @xmath306 . indeed in a universe with @xmath281 all momentum modes would be energetically difficult to excite .
everything - signals , apparatus , observer - would be made from particles that have zero * m * quantum number ( in our present large universe everything is made of zero * w * quantum number ) . since string theory has duality as a symmetry of the spectrum as well as the interactions
, the dual particles would interact with each other exactly the way normal particles do in our present universe .
the observers in a very small universe would not therefore know that they are in a universe much smaller than @xmath101 , their physics would be identical to ours ( and for that matter nor do we know whether our universe is very large or very small compared to @xmath113 ) .
it is therefore no surprise that temperature has the behaviour shown in figure 2 . as radius goes much smaller than @xmath101 , the universe actually _ expands , as seen by the modes that are excited in it_. this also makes it evident that there are no physical singularities in the energy density , pressure or curvature as @xmath307 . in a very small universe ,
the physical energy density is not @xmath308 but @xmath309 ( which goes to zero and not infinity as @xmath307 ) , since the physical volume of the universe is @xmath310 . in string
theory the smallest _ physical _ size of the universe is @xmath113 . note that the arguments leading to the string uncertainty principle - that the smallest observable size of an elementary string is @xmath113 - and the arguments leading to the same minimum physical size of the universe both make essential use of probes in thought experiments .
also note the difference : while the former argument uses the oscillator modes , the latter rests on the duality between momentum and winding modes ( although the limiting temperature and pressure depend again on the oscillators ) .
all these modes are simultaneously forced upon us as soon as we accept strings as the elementary constituents of nature , and all are governed by a single scale parameter that appears in ( [ action1 ] ) .
0.3 cm * a cosmological scenario without initial singularities * : brandenberger and vafa sketch the following scenario .
let us assume that at some point in the future our universe stops expanding and starts contracting and heating up . as the energy density increases to the hagedorn energy density , stringy effects will take over and the temperature will flatten out . if it continues to contract through the duality radius and comes out the ` other side ' , then dual ( analogues of winding ) modes will take over .
the universe will cool and ` expand ' and give rise to dual nucleons , galaxies , stars , planets , life , etc .
what appears to us to be the ` big crunch ' will be a ` big bang ' for the dual observers .
the process could repeat giving rise to an oscillatory universe . `
our own ' big bang was just one such periodic occurrence .
of course much more work is needed to justify any such _ dynamical _ scenario .
we have been concerned with just those aspects which hinge only on the _ degrees of freedom_. a body of literature now exists which also deals with the time evolution of the metric and other low energy modes in string theory in the cosmological context ( see tseytlin and vafa ( 1992 ) , gasperini ( 1997 ) , the contribution by bose ( 1997 ) to these proceedings , and references therein ) .
perhaps it would be worthwhile to revisit some of this in the light of the expression for the pressure of a string gas presented here , since pressure as part of the energy momentum tensor is a source in the field equations .
nevertheless the above scenario is important in that it at least allows us to _ imagine _ how initial singularities might be avoided in string theory .
it is important to emphasize that singularities are avoided not by recourse to quantum gravity ( spacetime has all along been treated classically ) but simply by a reinterpretation of what it means to talk of a small universe in the light of string theory . in point particle theories
, classical imagination fails at @xmath311 .
this is an example of how the new degrees of freedom in string theory allow ( or rather , necessitate ) a new perspective on our ideas of spacetime , in this case specifically on our notion of the size of the universe .
it should be mentioned that while we have explicitly discussed the case of a toroidal compactification for simplicity , the t - duality symmetry which makes this reinterpretation possible holds for a much larger class of string models ( and is expected to be a symmetry in a nonperturbative formulation of string theory ) .
a limiting temperature and pressure in the ideal gas approximation also seem to be a universal feature of strings in compact spaces . 0.3 cm * cautionary remarks * : at this point some caveats are in order .
thermodynamics in the presence of gravity must take into account the jeans instability . at constant energy density
a sufficiently large volume will be susceptible to gravitational collapse .
this places an upper limit on the value of @xmath163 for which our thermodynamic considerations are valid .
second , the results are based on an ideal gas approximation , used in a regime of high energy densities , greater than the string energy scale itself .
this is justified only if the coupling is weak ( @xmath125 ) . even for a fixed weak coupling
the approximation can be expected to break down at sufficiently high energy densities , at which point non - perturbative effects will need to be taken into account .
this places a lower limit on @xmath163 for the validity of the approximation .
thus there is possibly a window @xmath312 , @xmath313 ( and the ` dual window ' @xmath314 ) in which one can expect this to be valid .
the window expands in both directions as coupling becomes weaker ( see atick and witten ( 1988 ) for related arguments ) .
in addition to string interaction and nonperturbative effects in the region of @xmath163 close to @xmath113 , we face the uncertainty of interpretation of spacetime itself at such small scales . for sufficiently large ( or sufficiently small ) @xmath163 , spacetime may be treated classically , as we have done .
but this is questionable near the duality radius .
this is the regime where the universe as well as its elementary constituents have the same ` size ' .
this problem awaits a better understanding of spacetime in string theory . 0.3 cm * possible observational consequences , further questions * : assuming that there was an era in the past where the ideal string gas approximation was valid
, could there be some observable relic ?
from the picture of how energy is distributed in the string gas it seems likely that density fluctuations would have a different character in the stringy era , and as seeds for later structure formation could have observable consequences .
second , it would be interesting to look for signatures of compactness at very large scales in the universe .
apart from a resolution of the singularity problem , string thermodynamics also seems to be internally consistent only in compact spaces .
a compact universe is even otherwise natural in string theory , since the extra dimensions in any case have to be compact . at a more theoretical level
, it may be worthwhile to investigate dynamical mechanisms based on string modes ( see brandenberger and vafa ( 1989 ) for a proposal ) for why only three spatial dimensions are large .
also it is of interest to study how the recent progress in our understanding of some non - perturbative aspects of string theory affects the above considerations .
note added : after this writeup was submitted for the proceedings , i became aware of other papers ( yoneya 1989 ; konishi , paffuti and provero 1990 ; kato 1990 ; susskind 1994 ) which attempt to derive the string uncertainty principle .
the argument presented in the present article is different from those given in these papers .
other recent references of related interest are li and yoneya 1997 , which argues that the string uncertainty principle is consistent with d - brane dynamics , as well as barbon and vazquez - mozo 1997 and lee and thorlacius 1997 , which attempt to include d - branes within superstring statistical mechanics .
for literature on a minimal length in the context of quantum gravity without invoking string theory see references in the review by garay 1995 , as well as padmanabhan 1997 .
i thank n. deo , c - i tan and c. vafa for discussions in which most of my understanding of string thermodynamics and cosmology was developed , s. kalyana rama for getting me interested in the pressure of an ideal string gas and pointing out some recent references , and the participants and organizers of the conference on big bang and alternative cosmologies for a stimulating meeting .
aharonov , y. , englert , f. , and orloff , j. 1987 , _ phys .
lett . _ * b199 * , 366 .
+ amati , d. , ciafaloni , m. , and veneziano , g. 1989 , _ phys .
lett . _ * b216 * , 41 .
+ amati , d. , ciafaloni , m. , and veneziano , g. 1988 , _ int
phys . _ * a3 * , 1615 .
+ antoniadis , i. , ellis , j. , and nanopoulos , d. v. 1987 , _ phys .
lett . _ * b199 * , 402 .
+ atick , j. j. and witten , e. 1988 , _ nucl .
phys . _ * b310 * , 291 .
+ axenides , m. , ellis , s. d. , and kounnas , c. 1988 , _ phys .
_ * d37 * , 2964 .
+ barbon , j. l. f. and vazquez - mozo , m. a. 1997 , _ dilute d - instantons at finite temperature _ ( hep - th/9701142 , preprint no .
cern - th/96 - 361 ) .
+ bose , s. 1997 , _ solving the graceful exit problem in superstring cosmology _ , these proceedings ( hep - th/9704175 , preprint no .
iucaa-26/97 ) .
+ bowick , m. j. and giddings , s. b. 1989 , _ nucl .
phys . _ * b325 * , 631 .
+ brandenberger , r. h. and vafa , c. 1989 , _ nucl .
phys . _ * b316 * , 391 .
+ carlitz , r. d. 1972 , _ phys .
_ * d5 * , 3231 .
+ deo , n. , jain , s. , narayan , o. , and tan , c - i 1992 , _ phys .
rev . _ * d45 * , 3641 .
+ deo , n. , jain , s. , and tan , c - i 1989a , _ phys . lett .
_ * b220 * , 125 .
+ deo , n. , jain , s. , and tan , c - i 1989b , _ phys . rev . _ * d40 * , 2626 .
+ deo , n. , jain , s. , and tan , c - i 1991 , in _ modern quantum field theory _ , edited by s. das et al ( world scientific , singapore ) .
+ frautschi , s. 1971 , _ phys . rev _ * d3 * , 2821
. + garay , l. j. 1995 , _ int .
phys . _ * a10 * , 145 .
+ gasperini , m. , 1997 , _ birth of the universe in string cosmology _ ( gr - qc/9706037 ) .
+ ginsparg , p. and vafa , c. 1987 , _ nucl .
phys . _ * b289 * , 414 .
+ green , m. b. , schwarz , j. h. , and brink , l. 1982 , _ nucl .
phys . _ * b198 * , 474 .
+ green , m. b. , schwarz , j. h. , and witten , e. 1987 , _ superstring theory _ , cambridge university press .
+ gross , d. j. 1989 , _ phil .
lond . _ * a329 * , 401 .
+ gross , d. j. and mende , p. f. 1988 , _ nucl .
* b303 * , 407 .
+ hagedorn , r. 1965 , _
nuovo cim .
suppl . _ * 3 * , 147 .
+ huang , k. and weinberg , s. 1970 , _ phys .
lett . _ * 25 * , 895 .
+ karliner , m. , klebanov , i. , and susskind , l. 1988 , _ int .
phys . _ * a3 * , 1981 .
+ kato , m. 1990 , _ phys .
lett . _ * b245 * , 43 .
+ kikkawa , k. and yamasaki , m. 1984 , _ phys .
lett . _ * b149 * , 357 .
+ konishi , k. , paffuti , g. , and provero , p. 1990
lett . _ * b234 * , 276 .
+ lee , s. and thorlacius , l. 1997 , _ strings and d - branes at high temperature _ ( hep - th/9707167 , preprint no .
pupt-1708 ) .
+ li , m. and yoneya , t. 1997 , _ phys .
_ * 78 * , 1219 .
+ maclain , b. and roth , b. d. b. 1987 , _ comm .
_ * 111 * , 539 .
+ mcguigan , m. 1988 , _ phys .
rev . _ * d38 * , 552 .
+ mitchell , d. and turok , n. 1987 , _ nucl .
phys . _ * b294 * , 1138 .
+ nair , v. p. , shapere , a. , strominger , a. , and wilczek , f. 1987 , _ nucl .
phys . _ * b287 * , 402 .
+ obrien , k. h. and tan , c - i 1987 , _ phys .
rev . _ * d36 * , 1184 .
+ padmanabhan , t. 1997 , _ phys .
lett . _ * 78 * , 1854 .
+ sakai , n. and senda , i. 1986 , _ prog .
phys . _ * 75 * , 692 .
+ salomonson , p. and skagerstam , b. s. 1986 , _ nucl .
phys . _ * b268 * , 349 .
+ sathiapalan , b. 1987 , _ phys .
lett . _ * 58 * , 1597 .
+ scherk , j. 1975 , _ rev .
_ , * 47 * , 123
. + susskind , l. 1994 , _ phys .
rev . _ * d49 * , 6606 .
+ tseytlin , a. a. and vafa , c. 1992 , _ nucl .
* b372 * , 443 .
+ turok , n. 1989 , _ physica _ * a158 * , 516 .
+ yoneya , t. 1989 , _ mod .
. lett _ * a4 * , 1587 . | in a universe whose elementary constituents are point particles there does not seem to be any obvious mechanism for avoiding the initial singularities in physical quantities in the standard model of cosmology .
in contrast in string theory these singularities can be absent even at the level where spacetime is treated classically .
this is a consequence of the basic degrees of freedom of strings in compact spaces , which necessitate a reinterpretation of what one means by a very small universe .
we discuss the basic degrees of freedom of a string at the classical and quantum level , the minimum size of strings ( string uncertainty principle ) , the t - duality symmetry , and string thermodynamics at high energy densities , and then describe how these considerations suggest a resolution of the initial singularity problem
. an effort has been made to keep this writeup self - contained and accessible to non - string theorists .
= 10000 |
undoubtedly , gamma - ray bursts ( grbs ) have remained to be one of the most exiting , intriguing , and enigmatic astrophysical phenomena since their mysterious discovery in the past several decades ( for a recent expository review of grbs the interested reader is referred to the excellent work by j. i. katz @xcite ) .
although no two grbs resemble each other and each one has its own peculiarities which makes the problem of modeling grbs very difficult , the whole of grbs reveals several interesting features .
since the publication of the batse data @xcite which included the observation of over @xmath4 grbs and revealed an almost uniform distribution of the location of grbs in the sky , combined with the deficiency of faint grbs , the association of grbs with the galactic plane has been ruled out .
the successive publications confirmed the figure more and more @xcite .
however , since the observation of afterglows in x - ray @xcite , optical @xcite and radio spectrum @xcite and the advert of robotic optical transient search experiment ( rotsie ) telescope ( * ? ? ?
* ; * ? ? ?
* e.g. ) which has revealed the red shifts for several grbs , their cosmological origin is widely accepted .
+ the batse data @xcite has also revealed another equally important overall feature of the grbs .
the distribution of time duration in observed grbs shows a double heap distribution , which the smaller one peaks around 0.2 sec and the larger one peaks around 20 sec ( fig.1 ) .
this two peaked distribution which apparently separated grbs into the so called short duration and long duration ones was referred to as @xmath5 @xcite and led some investigators to believe that there are two distinct populations of grbs .
+ it has been widely believed that whatever the central engine is , the radiation reaching us originates from the space surrounding the central engine .
it is also believed that during the collapse the energy release streams out in relativistic confined ejectas ( not isotropically ) and thus the total energy release during each event is far less than the unbelievable amount that one might obtain by assuming isotropic radiation @xcite .
+ the aim of this paper is to present a rather simple model , based on these general ideas , and show that there is no need for assuming two distinct populations of grbs , and to show that once the geometrical considerations and cosmological effects are fully accounted for a genetic standard candle ejecta , crossing an amorphous dense cloud , the so called @xmath5 can be deduced . + in
2 the model and its general formulation is introduced . in
3 the computational results and the fit of model parameters with batse data will be presented as well . 4
is devoted to a discussion on the results .
some needed calculations and discussions are presented in appendices .
grbs are modeled as a central engine with an instantaneous ultra - relativistic jet of material surrounded by an amorphous dense cloud . the central engine and its jet
are taken as a standard candle with a total release energy of @xmath6 , and an initial lorentz factor @xmath7 and initial opening angle @xmath8 of the jet for all grbs , whereas the clouds are considered to have a distribution both in thickness as well as in density .
for the sake of illustration and brevity we take both of these distributions to be gaussian .
+ we want to calculate the distribution of logarithm of time duration of observed grbs according to the above model .
we should however explore the ejecta evolution since the observed gamma ray emission originates in the shock front of ejecta+shocked medium .
the equations describing the ejecta evolution are presented here , based on the notations of paczynski & rhoads @xcite .
we consider the cloud to be at a distance @xmath9 from the source , which may be negligible in comparison to the cloud thickness @xmath10 .
the ratio of swept - up mass to ejecta mass @xmath11 has the form below : @xmath12 where , @xmath13 is the ejecta distance from the source .
the cloud density @xmath14 is taken to be independent of @xmath13 . furthermore , @xmath15 $ ] , in which @xmath16 represents the opening angle of the ejecta at radius @xmath13 .
so , equation(1 ) can be written as : @xmath17r^{2},\ ] ] where @xmath6 and @xmath7 are the initial kinetic energy and initial lorentz factor of the ejecta ( @xmath18 ) , and @xmath19 denotes the number density of the cloud ( @xmath20 ) .
paczynski & rhoads @xcite derived the relation between @xmath21 and @xmath22 from conservation of energy and momentum . here , we use their relation in a form suitable for our computations : @xmath23{\gamma^{2}-1}}~.\ ] ] the ejecta s opening angle @xmath16 increases with increasing @xmath13 as a result of lateral spreading of the cloud of ejecta+swept - up matter in the comoving frame at the sound speed @xmath24 , which has been derived by rhoads @xcite to be as below : @xmath25 where @xmath26 denotes the time from the event , measured in the ejecta comoving frame . substituting @xmath27 , and @xmath28
, we have : @xmath29 now , eliminating @xmath21 between equations(2 ) and ( 3 ) yields : @xmath30 \frac{\sqrt[3]{\gamma^{2}-1}}{\sqrt{\gamma_{0}^{2}-1}}r^{2}~.\ ] ] let s rewrite equations(5 ) and ( 6 ) in a non - dimensional form as below : @xmath31{\gamma^{2}-1}}{\sqrt{\gamma_{0}^{2}-1 } } [ 1-\cos\zeta(\eta)]\eta^{2 } \end{array}\right
. \nonumber\\,\ ] ] where , the non - dimensional parameter @xmath32 is defined as : @xmath33 in which : @xmath34 these coupled first order differential equations can be solved numerically by introducing the initial conditions : @xmath35 where @xmath36 .
+ noting that @xmath37 , we have : @xmath38 in equation(11 ) we used equation(8 ) and a non - dimensional time parameter @xmath39 defined as : @xmath40 the numerical results of equation(11 ) are used in appendix c , where we consider the effect of burster geometry on the observable time duration .
we begin with introducing the probability density for a collimated burst to occur in a direction through the cloud with a thickness @xmath10 and a number density @xmath19 .
we assume the cloud thickness to have a gaussian distribution in various directions from the central engine . by assuming a gaussian distribution for the cloud density as well we have : @xmath41\:\frac{1}{\sqrt{2\pi}\sigma_{_{l } } } \:\exp\left[\frac{-(l-\overline{l})^{2}}{2\sigma_{_{l}}^{2}}\right]~.\ ] ] the quantities @xmath42 and @xmath43 denote the mean thickness of the cloud and its dispersion , respectively .
likewise , @xmath44 and @xmath45 are defined in a similar manner . denoting the angle between the ejecta symmetry axis and the line of sight by @xmath46 ( fig.2 ) , and considering the independence of the probability density from azimuth angle in equation(13 ) ,
we can write : @xmath47\:\frac{1}{\sqrt{2\pi}\sigma_{_{l } } } \:\exp\left[\frac{-(l-\overline{l})^{2}}{2\sigma_{_{l}}^{2}}\right]~.\ ] ] the synchrotron emission is a fast process for our model ( see appendix a ) , so @xmath48 ( the time duration of a grb as measured by an observer cosmologicaly near to the source and located on the line of sight ) , is attributed to the time that the shock front takes to cross the dense cloud . as explained in appendix c
, the cloud thickness @xmath10 can be expressed as a function of @xmath46 , @xmath19 , and @xmath48 ( eqn.[c16 ] ) : @xmath49 so we write : @xmath50 or : @xmath51 notifying that by this substitution(eqn.[15 ] ) , the bursts that do not manage to cross through the cloud with a thickness @xmath10 ( and stop in it ) are practically omitted ( see appendix c ) . using equation(14 ) in equation(17 ) , it is seen that : @xmath52 \exp\left[\frac{-(l(t_{rec},n,\theta)-\overline{l})^{2}}{2\sigma_{_{l}}^{2}}\right].\end{aligned}\ ] ] integrating over @xmath46 and @xmath19 yields : @xmath53 \exp\left[\frac{-(l(t_{rec},n,\theta)-\overline{l})^{2}}{2\sigma_{_{l}}^{2}}\right]\nonumber\\ & \times&\left(\frac{dl}{d\log t_{rec}}\right)_{_{n,\theta}}\sin\theta\:d\theta\:dn~.\end{aligned}\ ] ] the effect of red shift is not considered yet .
equation(19 ) only gives the probability density for an observed burst to have a specified logarithm of time duration , as measured by an observer near to it .
we now investigate the relation between @xmath54 and @xmath55 where @xmath56 stands for the time duration of a grb measured at earth .
to obtain the later , the former must be integrated over red shift z , using a weight function @xmath57 , so that @xmath58 represents the probability for occurring a grb in a red shift between @xmath59 and @xmath60 . to show this ,
let s consider an observer located on the line from us to an occurred grb which is ( cosmologicaly ) near to it .
the probability density for the grb to occur in a red shift z ( with respect to us ) , and to have a specific @xmath61 ( measured by the observer near to the grb ) , is clearly as below : @xmath62 to obtain @xmath63 , which is the probability density for observing a grb occurred at a red shift z and observed to have a specific @xmath56 , we write : @xmath64~,\ ] ] noting that @xmath65 , the second term in the the bracket equals one .
now , using equation(20 ) we have : @xmath66 after integrating equation(22 ) over z , the final form of the observable probability density will be as below : @xmath67 the explicit form of @xmath57 is needed .
this form is obtained in appendix d ( eqn.[d4 ] and eqn.[d5 ] ) .
the second term in the integrand is given by equation(19 ) , in which the implicit form of @xmath68 appears .
appendix c is devoted to the procedure of obtaining this function . for evaluation of @xmath69 ( eqn.[23 ] ) ,
we need the values of eight parameters .
four of them are the cloud parameters @xmath42 , @xmath70 , @xmath44 , and @xmath71 , and the fifth one is the index @xmath72 corresponding to the grb occurrence rate ( see eqn.[d5 ] ) .
the later three ones are the ejecta parameters , @xmath6 , @xmath7 , and @xmath8 , which are the initial kinetic energy , the initial lorentz factor , and the initial opening angle of the ejecta , respectively .
the @xmath73 software is used in the numerical computations . in the procedure
we begin with solving the coupled differential equations(7 ) and ( 11 ) which govern the ejecta evolution , and in which @xmath7 and @xmath8 are the only free parameters . furthermore , we take @xmath74 . for fixed values of these parameters ,
the functions @xmath75 , @xmath76 and @xmath77 ( see eqn.[c3 ] ) can be uniquely obtained .
obviously @xmath22 decreases with increasing @xmath32 ( fig.3 ) , and reaches to @xmath78 when @xmath32 approaches a certain value .
this in fact takes an infinite time and results in an infinite non - dimensional time duration @xmath79 ( see eqn.[c4 ] and eqn.[c5 ] ) .
so , to circumvent the difficulty , we consider an effective lower limit @xmath80 for the lorentz factor of shocked matter , which yields an effective upper limit for non - dimensional time @xmath39 .
we adopted @xmath81 as a lower cut - off .
then , following the procedure explained in appendix c , for a specific cloud thickness @xmath10 , and correspondingly a specific @xmath82 ( eqn.[c13 ] ) , the non - dimensional time duration @xmath83 of a grb can be calculated for every @xmath46 and @xmath82 ( fig.4 ) .
let s denote the radius corresponding to @xmath80 by @xmath84 so that @xmath85 , and recall that in the model , for @xmath86 , the emitted photons would be completely scattered by the electrons of not - swept part of the cloud ( that they have to pass through , before entering free space ; see appendix c ) , and so , the whole phenomenon may be called a `` @xmath87 '' .
but , if @xmath88 , the shocked matter succeeds to go out of the cloud and , as explained in 4.3 , due to a suppression process that intensively decreases the cross - section of compton scattering,(the major part
of)the emitted photons finally succeed to get released from the shocked medium and enter free space , provided that @xmath89^{-1}$ ] ( see appendix b ) .
it is really for this reason that @xmath79 happens to be a function of @xmath82 , and practically independent of @xmath84 .
then , solving @xmath83 for @xmath82 numerically , we can obtain the function @xmath90 which is the equivalent non - dimensional form of expression ( 15 ) .
we rewrite equation(19 ) in a non - dimensional form suitable for numerical computations : @xmath91 \exp\left[\frac{-\left(\frac{l(\overline{n}\nu / e)}{\overline{l}}\eta_{_{l}}(\tau_{rec},\theta)-1\right)^{2}}{2\varsigma^{2}}\right]\nonumber\\ & \times&\frac{l(\overline{n}\nu / e)}{\overline{l}}\left(\frac{d\eta_{_{l}}}{d\log \tau_{rec}}\right)_{_{\nu,\theta}}\sin\theta\:d\theta\:d\nu,\end{aligned}\ ] ] in which , equation(c13 ) and the definitions below : + @xmath92 , + @xmath93 , + and + @xmath94 + are used .
so , after choosing the quantities @xmath42 , @xmath43 , @xmath44 , @xmath71 , and of course @xmath6 , we can evaluate the integral appearing in equation(24 ) . then , after choosing a value for @xmath72 , the integral of equation(23 ) can be performed to obtain the observable quantity @xmath95 . + as seen in equation(24 ) , the initial kinetic energy @xmath6 and the cloud mean density @xmath44 appear only in the form @xmath96 , and therefore they can not be found separately in a fitting process , and all that can be obtained is only their ratio .
but , the observed fluence of grbs reveals that the released energy in grbs is of order of @xmath97 for a isotropic burst , which reduces to @xmath98 if the bursts were confined to a cone with a solid angle @xmath99 .
we used this amount of isotropic energy to relate @xmath6 to @xmath8 ( @xmath100 ) , and reduce the free parameters of the model to seven as @xmath101 , @xmath71 and @xmath72 , which hereafter are called @xmath102 parameters .
these parameters must be chosen so that the results make the best fitting to the observed distribution of grbs . to achieve the task
, one must search in the seven dimensional @xmath102 space , and find the point in which the statistical quantity @xmath103 takes the smallest value @xmath104 .
we used the batse 4th catalogue @xcite of 1234 grbs and adopted the bins @xmath105 in a range @xmath106 , and used the @xmath107 technique to move toward the best point in which @xmath103 gets minimized .
the obtained fitted values are : @xmath108 with a corresponding value @xmath109 ( per degree of freedom ) .
as can be seen in fig.5 , the deviation from a more perfect coincidence to data seems for the structure in the observed distribution located around @xmath110 .
the structure has been noted before @xcite .
we put aside the the data of the noted structure , which are ones with durations between @xmath111 , and again repeated the numerical searching process in the parametric space .
the new obtained values of the fitted parameters are : + @xmath112 which slightly differ from what previously obtained ( eqn.[25 ] ) . but
this time , @xmath104 reduces to @xmath113 ( fig.6 ) .
so the mentioned structure may be interpreted as a result of an independent phenomenon or effect which was not considered in our model .
though the original shapes of equations ( 13 ) and ( 14 ) are naturally normalized , the resulting final equations ( 19 ) and ( 23 ) are not , because : + 1)the bursts that their symmetry axes make an angle @xmath114 can not be detected ( see appendix c ) , + 2)the bursts for which @xmath86 were not considered in the numerical computations ( because they do not manage to cross out the cloud ) . + the total probability of observing the occurred bursts , @xmath115 , is obtained to be @xmath116 , using our best fit parameters(eqn.[26 ] ) .
this small probability must be interpreted to be due to the above two reasons .
the first reason describes the suppression of observed grbs by the term @xmath117 , while the second is responsible for the remaining factor of @xmath118 .
so , only about one percent of the bursts manage to produce a real grb , and only about @xmath119 of these grbs occur in our line of sight . of course
, the lateral spreading of the ejecta+swept mass may modify these two factors , increasing the first and decreasing the second .
+ with the values for the fitted parameters in equation(25 ) or in equation(26 ) , the mean mass of the clouds @xmath120 is about @xmath121 , which is of the order of the envelop mass in massive stars .
such amorphous clouds seem strange in stars , but in close neutron star - supergiant binaries where the neutron star orbits around the core and accrete the envelope , the spherical symmetry is likely removed , as pointed out by podsiadlowski et al .
terman , taam , & hernquist @xcite show that the system would emerge to form a red supergiant with a massive thorne - zytkow object ( tzo ) @xcite .
podsiadlowski et al .
@xcite also estimated a tzo birth - rate of @xmath122 in the galaxy . considering equation(d4 ) and the obtained total probability of observing a `` real '' grb , which is @xmath123 in our model ( see above ) , it is seen that the grb observation rate ( @xmath124 events per day ) implies a total ( real+failed ) grb rate of the order of @xmath125 .
this is not too far from what podsiadlowski et al .
@xcite theoretically estimated for tzo birth - rate ( @xmath126 ) .
qin et al .
@xcite introduced aicns ( accretion - induced collapse of neutron stars ) scenario as grb engines .
katz @xcite introduces a dense cloud model to explain the observed @xmath127 gamma - rays in a number of grbs ( * ? ? ?
* ; * ? ? ?
* e.g. ) , and suggests the amorphous envelops of tzos for these clouds .
+ the initial opening angle of the ejecta in our model ( @xmath128 ) is obtained during the fitting process ( 3 ) .
such a small opening angle for an ejecta or a jet might be explained by attributing it to the collimating process of an ultra - relativistic ejecta with @xmath129 in a sufficiently high magnetic field @xcite .
the initial lorentz factor of the ejecta , as obtained in our model ( @xmath130 , see eqn.[25 ] ) provides the necessary condition of @xmath129 . aside from these theoretical justifications for the idea of a highly collimated ejecta at the source
, there are some pieces of evidences supporting this idea @xcite . at the mean time the observed bimodality must be interpreted as a result of the second reason expressed in 4.1 .
though in our model one may expect only one heap in the @xmath131 distribution associated with the directions in the clouds having both the most probable @xmath10 and @xmath19 ( which are equal or near to @xmath42 and @xmath44 ) , but as the numerical computation shows , such directions in the clouds are too thick and too dense to be crossed out by the ejecta , and therefore a real grb would not be produced .
so , we are left with four other regions with high probabilities : + 1 ) @xmath132 and @xmath133 ( short grbs ) + 2 ) @xmath132 and @xmath134 ( no grbs ) + 3 ) @xmath135 and @xmath136 ( no grbs ) + 4 ) @xmath137 and @xmath136 ( long grbs ) + now , we show that the directions trough the clouds associated with region(1 ) produce the short duration grbs , and ones associated to regions ( 2 ) and ( 3 ) produce no grb , while the others in the forth region produce the long duration ones .
+ as to the equation(9 ) the quantity @xmath138 ( which is of the order of the sedov length ) is @xmath139 for @xmath140 ( see eqn.[25 ] ) . as seen in fig.7 ,
the time duration of grbs associated to such directions is of the order of the time duration of short grbs ( case ( 1 ) above ) .
fig.8 shows that in our model the calculated time duration of grbs for @xmath141 is of the order of the duration of long grbs . using equation(9 )
, we see that in this case ( @xmath141 ) , @xmath138 is about @xmath142 .
so in our model the long duration grbs are due to the passing of ejecta through the directions where @xmath143 and @xmath144 ( case ( 4 ) above ) .
furthermore , since @xmath145 when @xmath132 , we see that in cases ( 2 ) and ( 3 ) above , the ejecta would stop in the dense cloud and the produced photons can not scape from the optically thick cloud . in fig.9 ,
@xmath146 is plotted for a number of densities , when @xmath147 has the highest permitted value @xmath84 .
+ these general features of our calculations result from the general features of our model , and therefore we speculate that any distributions for the clouds thickness and density which are picked around a mean value could explain the general features of the duration distribution .
our chose of gaussian distributions for thickness and density was only for the few parameters needed to describe them .
+ in `` the dense cloud model '' at this stage we have simply omitted the bursts which ejectas can not go out of the cloud and stop in it ( because the emitted photons would be scattered by the dense cloud ) , and we have claimed that the produced photons by all bursts that succeed to cross out the dense cloud can finally enter the free space .
+ at the first glance the model might appear to have a serious problem in opacity , namely , in this model the best fit mean density and mean thickness of the clouds are found to be of the orders of @xmath148 and @xmath149 .
so , as to the relation @xmath150 , ( @xmath151 ) one would expect an optical depth of the order of @xmath152 .
but there are two factors that remedy the situation : + ( i ) as the numerical calculation shows ( see 4.2 ) , the most probable directions characterized by @xmath132 and @xmath136 are dynamically too thick to be crossed out by the ejecta and the radiation produced in this case would be completely scattered by the dense cloud . on the other hand , as discussed in
4.2 , the long duration grbs are due to the crossing of ejecta through directions where @xmath153 and @xmath136 .
since the density is reduced by the factor @xmath154 , the optical depth drops to the order of @xmath155 . as explained in 4.2 , the short duration grbs are due to the directions in the cloud where @xmath132 and @xmath156 . in this case
the optical depth of the cloud reduces to @xmath157 , which is yet too high .
but , + ( ii ) most of photons emitted off the shock front have the chance to be overtaken by the moving shock ( appendix b ) .
moreover , it has been shown @xcite that in high temperatures @xmath158 the cross section of compton scattering for instance for @xmath159 photons effectively drops to @xmath160 , so such a high temperature plasma is much more transparent than what may seem at first .
the same is true for the case of a power law distribution of electrons ( see eqn.[a4 ] ) .
most of the electrons in such a distribution have energies of the order of @xmath161 , in which we have used equation(a5 ) , taking @xmath162 , @xmath163 and @xmath164 . for this case
the compton cross section suppresses by the factor of @xmath165 for @xmath166 photons ( the energy of the emitted photons in the shocked medium comoving frame is less than their observed energy by the factor of @xmath22 ) , and therefore , the optical depth drops to the order of @xmath155 .
so the photons overtaken by the shocked medium may remain in it without being scattered until the shocked medium crosses up the dense cloud .
briefly , due to this second factor , the optical depth in short duration case diminishes to @xmath167 , and the optical depth in long duration case diminishes to @xmath168 .
+ the variability seen in the light curves of long duration grbs maybe attributed to the heterogeny of cloud s density in the ejecta s trajectory .
this maybe the case in long duration grbs for which @xmath168 , but in short duration grbs for which @xmath167 , we expect that such a heterogeny would not be appeared .
+ the lateral expansion of the shocked medium can reduce its density and therefore its optical depth .
this effect of course may cool the shocked medium and cause an increase in effective compton cross section in it .
the consideration of dynamical feedback makes our modeling much more complicated and it has been neglected here . in `` the dense cloud model '' at this stage we have simply omitted the bursts which ejectas can not go out of the cloud and stop in it , but have claimed that the produced photons by all bursts that succeed to cross out the dense cloud can finally enter free space . +
a complete study needs to include flux computations .
an exact comparison with observed time duration data would be possible only when the theoretical time duration appearing in this work were exactly evaluated in the same manner as the observable duration @xmath169 is defined ( as the time interval over which @xmath170 percent to @xmath171 percent of the burst counts accumulate ) .
moreover , the batse s triggering mechanism made it less sensitive to short grbs than to long ones and therefore short grbs were detected to smaller distances @xcite , so a smaller number of them have been observed .
lee & petrosian @xcite studied this effect and corrected the number of short grbs . in a more exact work this correction must be considered too .
we acknowledge the anonymous referee for valuable comments .
f. s. acknowledges mr .
mehdi haghighi for his guides on computational methods .
this research has been partly supported by grant no .
nrci 1853 of national research council of islamic republic of iran .
in this part we use the review paper of piran ( 1999 ) to estimate the synchrotron cooling time in our model .
the synchrotron cooling time in the comoving frame is : @xmath172 where @xmath173 is the thompson cross section and @xmath174 is the lorentz factor of the emitting electron , while @xmath175 stands for the energy density of magnetic field which in grb literature is assumed to be proportional to @xmath176 , the comoving internal energy density of the shocked matter : @xmath177 so that @xmath178 represents the share of magnetic field in @xmath176 , which is given by : @xmath179 in which , @xmath19 is the number density of surrounding medium ( ism , or a dense cloud as assumed in our model ) , and @xmath22 denotes the lorentz factor of the shocked matter .
+ the electrons in shocked media are assumed to develop a power law distribution of lorentz factors : @xmath180 the convergence of total energy of the electrons requires the power index @xmath181 to be greater than @xmath124 , while the assumed lower limit @xmath182 is to prevent the divergence of electron number density and is obtained to be : @xmath183 where @xmath184 represents the share of the electrons in the internal energy of shocked matter .
+ furthermore , the time interval between emission of two photons from the same point in the comoving frame , @xmath185 , and the time interval @xmath186 , which represents the time interval of their successive arrival to a cosmologically distant observer ( at the earth ) , are related as below : @xmath187 equations(a1-a6 ) are adopted from piran ( 1999 ) .
now , by substituting equations(a2),(a3 ) , & ( a5 ) in equation(a1 ) , and then using equation(a6 ) , the synchrotron cooling time measured by a terrestrial observer turns out to be : @xmath188 which is clearly much less than all observed time durations of grbs .
so , the synchrotron emission in a dense cloud model must be considered as an instantaneous process , and therefore the shock front must be regarded as the emitting surface .
this allows us to attribute time duration of a grb merely to the time that the shock front takes to cross the dense cloud .
here , we want to find the relation between @xmath22 and @xmath189 which are respectively the lorentz factors of the shocked matter and of the shock front ( which is the emitting surface in our model ; see appendix a ) . in fig.10 , @xmath190 and @xmath191
correspond to shocked matter and shock front speeds , both measured in the ( central ) source frame . in the shocked matter frame ( fig.11 ) the dense cloud which is seen to have a density @xmath192 , moves toward the shocked medium with a speed @xmath190 , while being compressed to a density equal to @xmath193 .
consequently , as illustrated in fig.11 , the shocked medium expands towards right with a speed @xmath194 which is in fact the speed of the shock front in the shocked matter frame ( compare fig.10 and fig.11 ) .
considering the conservation of nucleon number , it is seen that : @xmath195 considering the relativistic summation of velocities , we have : @xmath196 noting the relations @xmath197 and @xmath198 , the substitution of equation(b1 ) in equation(b2 ) finally yields : @xmath199 the distance from the origin to the shock front , denoted by @xmath200 , can be obtained by integrating @xmath191 over @xmath201 : @xmath202 the results are shown in fig.12 . as seen , the difference between @xmath13 and @xmath200 never exceeds one percent . this result justifies the use of symbol @xmath13 ( instead of @xmath203 ) for the location of the shock front through out our formulation .
+ now , let s consider a photon that is radiated from the shock front ( the emitting surface ) moving with the lorentz factor @xmath204 ( eqn.[b3 ] ) , in a direction which makes an angle @xmath205 with the velocity vector of the shocked matter ( fig.10 ) . if : @xmath206 the photon would be overtaken by the shock front .
the majority of the emitted photons fulfill this condition when @xmath207 and @xmath208 ( see eqn.[25 ] ) .
here the effect of burster geometry on the observed time duration is investigated . at first , as shown in fig.13 , we consider a radiating segment on the shock front .
the symmetry axis is denoted by @xmath209 . noting the definition of @xmath205 in the figure , it is seen that in addition to a radial component @xmath210 , the velocity vector of the segment must have a lateral component @xmath211 which is equal to : @xmath212 as measured in the source frame . in equation(c1 ) the non - dimensional radius @xmath32 , as defined in equation(8 ) , is used instead of @xmath13 , .
equation(c1 ) is obtained simply by using equation(5 ) and assuming that the lateral speed of the segment in a frame moving ( only radially and ) instantaneously along with the segment , is the fraction @xmath213 of sound speed @xmath24 ( which is the lateral speed of the emitting surface at its edges ) , and noting that the lateral speed in the source frame is less than its corresponding value in the ( radially instantaneous ) comoving frame by the factor @xmath214 .
it is well known that the radiation emitted by the segment is almost confined to an angle @xmath214 . in fig.13 the axis of the radiation cone emitted by the segment
is denoted by @xmath215 , which is parallel to the velocity vector of the segment and , as seen in the figure , makes an angle @xmath216 with @xmath217 axis , where @xmath218 $ ] .
so the radiation angle @xmath219 ( as depicted in the figure ) can be written as below : @xmath220 and the total radiation angle , defined as the radiation angle at the edge of the emitting surface , can be written as : @xmath221 this angle represents the cone of space illuminated by the emitting surface . using the results of equation(7 ) , the total radiation angle can be evaluated for every `` @xmath32 '' .
having defined the total radiation angle , we explore the effect of burster geometry on its observed time duration @xmath56 .
as shown in fig.2 , the problem is studied in a spherical coordinate system in which the central engine is taken as the origin , and the line - of - sight as the polar axis @xmath59 .
consider a photon emitted off a point on the emitting surface ( the shock front ) with radial coordinate @xmath13 and polar coordinate @xmath222 ; and at an instance t , measured in the source frame ( the point is not shown in the figure ) .
the relation between @xmath201 and the photon arrival time @xmath223 to a ( cosmologically ) near observer is obtained by granot , piran , & sari @xcite .
making suitable for our model , it is adjusted to the form below : @xmath224 in this equation the instance @xmath225 is defined as the time that the ejecta collides with the dense cloud at @xmath226 ; while @xmath227 is the time that the ( cosmologically ) near observer receives the photon emitted at @xmath225 from the point with coordinates @xmath226 and @xmath228 . defining : @xmath229 and noting equation(8 ) and equation(12 ) , we rewrite equation(c4 ) in the form below : @xmath230 multiplying equation(c4 ) by the cosmological time dilation term @xmath231 , results in the arrival time as maybe observed at the earth : @xmath232 or equivalently : @xmath233 in which we used the non - dimensional time duration @xmath234 defined as below : @xmath235 now , as shown in fig.2 , we consider a situation where the ejecta s symmetry axes makes an angle @xmath46 with the line of sight .
the necessary condition that at least some photons of the emitting surface are detected by the near observer is : @xmath236 here we denote the inverse of function @xmath77 by @xmath237 , which gives the radius corresponding to @xmath238 . as seen in fig.2 , for @xmath46 s larger than @xmath239 ,
the first photons reaching the detectors are those emitted at @xmath240 .
so , we use @xmath76 ( which is obtainable by solving equation(11 ) ) to write : @xmath241 where @xmath242 represents the starting time ( in the source frame ) that the emitted photons can reach the ( cosmologicaly near ) observer .
now , using the numerical results of equation(11 ) we can obtain the function @xmath243 .
then , considering equation(c6 ) , the non - dimensional time @xmath244 corresponding to @xmath242 would be as below : @xmath245+\eta_{0}&\mbox : $ \:\:\theta > \zeta_{rad}(\eta_{0})$ } ~.\ ] ] now , we are to find the time @xmath246 after which no photons can be detected by the near observer . in fig.2 , the photons emitted from the edge point @xmath247 can reach us at all times greater than @xmath242 .
let s remind that in our model the emission process terminates at the time that the shocked matter goes out of the cloud ( appendix a ) . as is seen in equation(c6 ) , the closer to the point @xmath248 is a point on the emitting surface , the later its emitted photons would reach the near observer , of course , provided that the observer line - of - sight remains in the radiation cone of the emitting point . + defining : @xmath249 and recalling equation(c2 )
, we can solve the equation : @xmath250 to find the function @xmath251 , which specifies the furthest point ( on the shock front at the radius @xmath82 ) which its radiation reaches us , of course , if it remains smaller than @xmath252 .
+ now , using equation(c6 ) , we can find the instance that the last photons reach the near observer : @xmath253)+\eta_{0}~.\ ] ] finally , the non - dimensional time duration of a grb , @xmath83 , will be equal to @xmath254 and , as to equation(c12 ) and equation(c15 ) , besides being a function of parameters of the ejecta and the cloud , it is also a function of the inclination angle @xmath46 .
so , the time duration of a grb as measured by an observer cosmologivally near to it would be a function of @xmath10 , @xmath19 , and @xmath46 : @xmath255 where : @xmath256 the expression(c16 ) needs some explanations . in our model
the ejecta parameters @xmath8 , @xmath6 and @xmath7 and the radius @xmath9 are assumed to be the same in all grbs .
so , these parameters do not appear in equation(c16 ) explicitly , though it is implicitly a function of them too . at the mean time ,
the cloud s density @xmath19 and its thickness @xmath10 are assumed to be different in different directions , and therefore they do appear explicitly in the expression ( c16 ) .
furthermore , if a cloud thickness @xmath10 is much more than its associated @xmath257 length @xmath258 , the ejecta may not cross it up , and finally stops in it . in such a situation the time duration of a grb would not be a function of @xmath10 , while it still remains dependent on @xmath19 and @xmath46 .
this is why in expression ( c16 ) , the quantity @xmath10 is distinguished from @xmath19 and @xmath46 by a semicolon .
the rearrangement of the expression(c16 ) in the form @xmath259 would be meaningful only when we are dealt with the situation where the ejecta succeed to cross up the cloud ( see eqn.[15 ] ) .
here the relation between @xmath57 appearing in equation(23 ) and the grb occurring rate @xmath260 ( in units of @xmath261 ) is derived , so that by adopting a cosmological model for the occurring rate , the integral in equation(23 ) can be evaluated .
+ the number of grbs that their effects could reach us in a time interval @xmath262 ( which is very much less than the present comoving time @xmath263 ) , and from a spherical volume element @xmath264 ( disregarding various effects , such as the geometrical ones discussed in appendix c , or those related to detectors threshold which affect the number of detected grbs ) is as below : @xmath265 where @xmath266 denotes the non - dimensional radial parameter of the source that its effects reach us with a red shift z , and @xmath267 , the scale factor at this red shift . in einstein - de sitter model
we have : @xmath268 where @xmath269 is the hubble constant .
furthermore , in frw metrics : @xmath270 where @xmath271 .
now , using equations(d2 ) and ( d3 ) in equation(d1 ) we obtain : @xmath272 what is remained is the explicit form of @xmath260 . the high variability seen in grb light curves
has convinced the investigators to relate grbs to stellar objects , and consequently their rate @xmath260 to the star formation rate @xmath273 .
the simplest model is of course a proportional model @xmath274 .
the proportional model may be correct if grbs are attributed to the evolution of massive stars whose lifetime is negligible in comparison with the cosmological time scale , but in ns - ns mergers model the proportionality may not be valid ( because of the delay time from the star formation to ns - ns mergers )
. wijers et al .
@xcite claimed that there is a good consistency between the proportional model and the observed grb brightness distribution , while petrosian & lloyd @xcite concluded that none of the ns - ns and the proportional model are in agreement with the observed @xmath273 .
totani @xcite ascribed this discrepancy to the uncertainties in sfr observations .
anyway we simply assume the grb rate to be as below : @xmath275 and treat @xmath72 as a free parameter that its best value should be obtained during the fitting procedure . in fig.14
, @xmath57 is plotted for a number of @xmath72 s .
clearly the case @xmath276 corresponds to a universe where the changes in the rates of astrophysical phenomena are only due to its expansion ( non - evolutionary universe ) .
it can be seen that @xmath57 takes its maximum at @xmath277 , which is not very sensitive to the magnitude of @xmath72 .
akerlof , c. w. , et al .
2003 , pasp , 115 , 132a begelman , m. c. , blandford , r. g. , & rees , m. j. 1984 , reviwes of modern physics , 56 , 255 cohen , e. & piran , t. 1995 , apj , 444 , l25 costa , e. , et al .
1997 , nature , 387 , 783 dingus , b. l. , et al .
1993 , aas , 182 , 7404 fishman , g. j. , et al .
1993 , a&as , 97 , 17 frail , d. , kulkarni , s. r. , nicastro , s. r. , feroci , m. , taylor , g. b. 1997 , nature , 389 , 261 gisler , g. r. , et al .
1999 , in aip conf .
499 , small missions for energetic astrophysics : ultraviolet to gamma - ray , ed .
s. p. brumby & n.y .
melville ( aip ) , 82 granot , j. , piran , t. , & sari , r. 1999 , apj , 513 , 679 granot , j. 2003 , apj , 596 , l17 jones , b. b. , et al .
1996 , apj , 463 , 565 katz , j. i. 2002 , the biggest bangs ( oxford u.press ) katz , j. i. , & canel , l. m. 1996 , apj , 471 , 915 katz , j. i. 1994 , apj , 432 , l27 kouveliotou , c. , et al .
1993 , apj , 413 , l101 kulkarni , s. r.,et al .
1999 , nature , 398 , 389 lamb , d. q. , donaghy , t. q. , graziani , c. 2004 , newar , 48 , 459 lee , t. t. , & petrosian , v. 1996 , apj , 470 , l479 mao , s. , narayan , r. , & piran , t. 1994 , apj , 420 , 171 meegan , c. a. , et al .
1996 , apjs , 106 , 65 norris , j. p. , nemiroff , r. j. , kouveliotou , c. , fishman , g. j. , meegan , c. a. , & paciesas , w. s. 1993 , aas , 183 , 2904 paciesas , w. s. , et al . 1999 ,
apjs , 122 , 465 paczynski , b. , & rhoads , j. e. 1993 , apj , 418 , l5 petrosian , v. , lloyd , n. m. 1998 , in aip conf .
428 , gamma - ray bursts , ed . c. a. meegan , r. d. preece , & t. m. koshut ( aip ) , 35 piran , t. 1999 , phr , 314 , 575 podsiadlowski , p. , cannon , r. c. , rees , m. j. 1995 , mnras , 274 , 485 qin , b. , wu , x. , chu , m. , fang , l. , & hu , j. 1998 , apj , 494 , l57 rhoads , j. e. 1999 , apj , 525 , 737 shekh - momeni , f. , & samimi , j. 2004 , preprint ( astro - ph/0402194 ) terman , j. l. , taam , r. e. , hernquist , l. 1995 , apj , 445 , 367 thorne , k. s. , & zytkow , a. n. 1977 , apj , 212 , 832 totani , t. 1999 , apj , 511 , 41 van paradijs , j. , et al .
1997 , nature , 386 , 686 waxman , e. 2003 , nature , 423 , 388 wijers , r. m. j. , bloom , j. s. , bagla , j. s. , natarajan , p. 1998
, mnras , 294 , l13 yu , w. , li , t. , wu , m. 1998 , tx19.confe , 90y ( astro - ph/9903126 ) denotes the radius where the radiation angle @xmath238 becomes equal to @xmath46 ( see appendix c ) . as shown in the figure ( for @xmath46 s larger than @xmath239 ) the first photons
reaching us are those emitted from the edge of the shock front at this radius .
if @xmath278 , then the last photons reaching us are those emitted from @xmath251 on the shock front ( point s in the figure ) .
the point s would be the furthest visible point on the shock front , since the line of sight fells out of the radiation cone of the points farther than it ( points between s and b ) .
if @xmath279 , then the point b would be the last visible point ( see eqn.[c15]).,width=453 ] .
it is plotted versus the inclination angle @xmath46 for a number of possible thicknesses @xmath82 .
as seen , the order of time durations are of the order of ones in short duration grbs ( see 4.2 ) .
, width=302 ] .
it is plotted versus the inclination angle @xmath46 for a number of possible thicknesses @xmath82 .
as seen , the order of time durations are of the order of ones in long duration grbs ( see 4.2 ) .
, width=302 ] is plotted versus the inclination angle @xmath46 for a number of densities , with @xmath283 equal to the highest permitted value @xmath84 .
the number near each curve is @xmath284 ( @xmath285 ) ( see 3 and
, width=302 ] , measured in the shocked frame . in this frame the cloud has a density of @xmath287 and moves toward the shocked medium with a speed @xmath190 , and becomes compressed to a density of @xmath288 .
so clearly we have @xmath289 ( see appendix b).,width=302 ] ( the probability density of observing a grb at a redshift @xmath59 ) is plotted for a number of @xmath72 s ( the value of @xmath72 is written near to the peak of its corresponding curve ) . as seen , all curves have a maximum at @xmath294 . this explains why the most of grbs have redshifts @xmath295 ( see eqn.[d5]).,width=453 ] | in this model a collimated ultra - relativistic ejecta collides with an amorphous dense cloud surrounding the central engine , producing gamma - rays via synchrotron process .
the ejecta is taken as a standard candle , while assuming a gaussian distribution in thickness and density of the surrounding cloud .
due to the cloud high density , the synchrotron emission would be an instantaneous phenomenon ( fast cooling synchrotron radiation ) , so a grb duration corresponds to the time that the ejecta takes to pass through the cloud .
fitting the model with the observed bimodal distribution of grbs durations , the ejecta s initial lorentz factor , and its initial opening angle are obtained as @xmath0 , and @xmath1 , and the mean density and mean thickness of the surrounding cloud as @xmath2 and @xmath3 .
the clouds maybe interpreted as the extremely amorphous envelops of thorne - zytkow objects . in this model the two classes of long and short duration grbs are explained in a unique frame . |
in order to access the energy exchange rates among quasiparticles , we have measured the distribution function @xmath2 in a stationary out - of - equilibrium set - up @xcite , described in fig .
[ figsetup ] .
we consider a mesoscopic metallic wire placed between two ideal reservoirs . in the absence of interactions , the population of quasiparticles at a given energy interpolates linearly between the distribution functions in the contacts , leading , if @xmath3 to a double - step - shaped distribution function . in the opposite `` hot electron '' regime , in which the typical interaction time is much shorter than the diffusion time @xmath4 equilibrium is reached locally at each position along the wire : the energy distribution function @xmath5 is a fermi function ( see dotted curves in fig.@xmath6[figsetup ] ) .
if heat is only carried out by electrons , the local temperature is @xmath7 where @xmath8is the lorenz number @xcite .
our experiments focus on the intermediate regime , in which interactions lead to a significant redistribution of the energy among quasiparticles , but not to a complete thermalization .
the distribution function is obtained from the differential conductance @xmath9 of a tunnel junction between the wire and superconducting electrodes relax , rowell . when the temperature of the superconductor lies well below the superconducting transition temperature , and if the density of states in the normal electrode wire is taken as energy independent on the probed energy range , the differential conductance of the junction is simply proportional to the convolution product of the density of states in the superconductor @xmath10 with @xmath11 the gap of the superconductor , and of the energy derivative of the distribution function in the wire , @xmath12@xmath13with @xmath14 the tunnel resistance of the junction .
the samples were fabricated by deposition with an electron - gun , at several angles , through a pmma suspended mask patterned using e - beam lithography .
the substrate was thermally oxidized silicon , as in our experiments on cu and ag .
we fabricated on the same chip two au wires : wire # 1 , with length @xmath15 and a single probe junction placed at @xmath16 ; and wire # 2 , with length @xmath17 and two probe junctions , at @xmath18 and @xmath19 we first deposited a @xmath20-thick aluminum film , which was then oxidized .
this layer defines the superconducting probe electrodes .
the wires and the pads were obtained by the subsequent evaporation from a 99.99% purity gold target , at a pressure of @xmath21 , at @xmath22 . the thickness and width of the wires are @xmath23 and @xmath24 @xmath25 .
the electrodes at the ends of the wires are @xmath26-thick pads , with an area of about @xmath27 . from the low temperature wire resistances @xmath28 and @xmath29 we deduce from einstein s relation , assuming rectangular cross - sections , the diffusion constant @xmath30 and the diffusion times @xmath31 and @xmath32 the samples were mounted in a copper box thermally anchored to the mixing chamber of a dilution refrigerator .
electrical connections were made through filtered coaxial lines @xcite .
measurements proceed as follows . in a first step ,
the voltage @xmath33 is set to zero . from the comparison of the measured differential conductance of the tunnel junctions with the calculated convolution product of the bcs density of states and of the derivative of a fermi function at the temperature of the mixing chamber
( see eq .
( 1 ) ) , we deduce the gap of the superconductor ( @xmath34 ) and the tunnel resistances @xcite ( @xmath35 ( wire # 1 ) , @xmath36 and @xmath37 ( wire # 2 ) ) . in a second step ,
the voltage @xmath33 is set to a dc value . from the differential conductance of the tunnel junctions
, we deduce by numerical deconvolution the distribution functions in the wire at the position of the junctions , using eq .
( 1 ) and the values of the gap and tunnel resistance determined in the first step .
we show in fig .
[ f(e ) ] the distribution functions measured with the three junctions at @xmath38 @xmath39 @xmath40 and @xmath41 @xmath42 heatings .
as expected from the difference in the diffusion times , the distribution functions are much more rounded in wire # 2 than in wire # 1 .
the distribution function at the middle of wire # 2 is close to the hot electron prediction , shown with a dotted line in the central panel for @xmath43 however , the distribution functions measured on the same wire close to the left electrode ( right panel ) are clearly different from fermi functions : they display a strong slope near @xmath44 reminiscent of the fermi step at this energy in the closest pad ( a ) which was grounded .
the larger the voltage @xmath45 the wider the interval of energy in which @xmath46 varies from 1 to 0 , as expected in both limiting regimes ( see fig .
figsetup ) .
in order to remove this dependence , we have replotted the same data in the bottom of fig .
[ f(e ) ] , but with the reduced energy @xmath47 on the horizontal axis .
remarkably , all scaled curves superimpose .
this scaling property , which was also found in all the experiments on copper wires , is not generic : in experiments on silver wires , the slope of @xmath2 at @xmath48 was found to increase with @xmath33 @xcite@xmath49 the distribution function can be calculated by solving the stationary boltzmann equation in the diffusive regime @xcite : @xmath50where @xmath51 and @xmath52 are the rates at which quasiparticles are scattered in and out of a state at energy @xmath53 by inelastic processes .
the observation of the hot - electron regime in the middle of wire # 2 , with a temperature close to the expected one , indicates that the energy is mainly redistributed among the quasiparticles .
in particular , phonon emission can be neglected . assuming that the dominant inelastic process is a two - quasiparticle interaction which is local on the scale of variations of the distribution function , @xmath54where the shorthand @xmath55 stands for @xmath56 the out - collision term @xmath57 has a similar form .
the kernel function @xmath58 is proportional to the averaged squared interaction between two quasiparticles exchanging an energy @xmath59 we have neglected the possible dependence of @xmath60 on the energies of the initial and final states and on the position along the wire .
the theory of electron - electron interactions in diffusive conductors in the @xmath61 regime @xcite predicts @xmath62 a regime which was observed in silver wires @xcite . in gold and copper wires , the scaling property implies , by a simple change of variables in eq .
( 3 ) , that @xmath63 is a function of @xmath64 only @xcite .
if futhermore @xmath58 does not depend on @xmath45 one obtains @xmath65 with @xmath66 a typical interaction rate .
we nevertheless tried to fit our data at @xmath67 with @xmath68 , and obtained @xmath69 , which is three orders of magnitude larger than the theoretical prediction @xmath70 with @xmath71 the density of states in gold and @xmath72 the cross - section of the wire kamenev , relaxag .
moreover , the shape of the distribution functions on wire # 2 is not properly reproduced and the calculated curve for wire # 1 at @xmath73 with the same parameter is significantly more rounded than the experimental data .
figure [ fitsexp2 ] shows the best fit of the data with @xmath74 obtained with @xmath75 . by construction , the scaling property of the data
is exactly reproduced . whereas the shape of the distribution functions on wire # 2 is well accounted for , the experimental data on wire # 1 are more rounded than we calculate .
the opposite discrepancy was observed in experiments on @xmath76 @xmath77long copper wires @xcite .
we attribute the strong rounding in the distribution functions on wire # 1 to the large heating of the reservoirs associated with the low resistance of wire # 1 : at a given value of voltage @xmath78 the heat power @xmath79 is the highest in the less - resistive wires .
the solution of the heat equation in the contacts , assumed to have the shape of an angular sector , with angle @xmath80 , and neglecting phonon emission , gives the reservoir temperature at the end of the wire @xcite : @xmath81 here , @xmath82 is the temperature of the quasiparticles at a large distance @xmath83 from the contact to the wire , and @xmath84 with @xmath85 the sheet resistance of the reservoir , estimated from the resistivity of the wires and the thickness ratio of the wires and of the reservoirs , @xmath86 a typical equilibration length between electrons and phonons @xcite , and @xmath87 the smallest radius for the radial approximation to be valid : @xmath88 in the experiment , the left contact ( labelled a in fig .
[ dessin ] ) , opens with an angle @xmath89 whereas the two other contacts ( b and c in fig . [ dessin ] ) have a smaller angle : @xmath90 we have fitted the distribution functions separately from @xmath91 to @xmath92 with the temperatures of the reservoirs as fit parameters .
the results are shown in fig .
[ tempreservoirs ] . at voltages below @xmath93
the fit temperature is nearly constant and identical in both reservoirs : @xmath94 in reasonable agreement with the temperature @xmath95 indicated by the thermometer on the mixing chamber . at such low voltages ,
scaling is not obeyed since temperature produces a significant rounding of the steps and @xmath96 varies with @xmath97 from the fit of the data on fig .
[ tempreservoirs ] , we deduce @xmath98 for the left contact , and @xmath99 for the right contact .
theory gives @xmath100 in excellent agreement with the ratio found from the fits .
our estimated values of @xmath101 and @xmath102 are about a factor of 2 larger than those obtained from fig .
[ tempreservoirs ] , possibly because the sheet resistance of the reservoirs is less than our estimate based on the electronic mean free path in the wire . at voltages larger than @xmath103 ,
the temperature of the reservoirs is proportional to @xmath45 leading to a scaling in the rounding of the steps .
in contrast , we want to stress that the slope of the distribution function near @xmath104 can not be explained by heating alone .
the fits of the distribution functions in this regime , taking into account both @xmath105 ( with @xmath106 ) and heating is shown on the left panel of fig .
[ tempreservoirs ] .
a consequence of reservoir heating is that the distribution functions are rounded at the scale of @xmath107 and the sharp features that were observed on copper samples can not not be resolved @xcite .
the distribution functions can also be well accounted for by a model in which quasiparticles are in local equilibrium with two - level systems distributed uniformly along the wire .
the relevance of tls on phase relaxation has recently been suggested by several authors imry , zawa , kroha .
agreement is found if one assumes , in a simple model ( described in @xcite ) , that the quasiparticles are weakly coupled to the tls with a density inversely proportional to the spacing @xmath108 between the two levels .
such a density is obtained if the two - level systems are the two lowest energy levels in symmetric double - well potentials , and if the distribution of the barrier heights is uniform . in glasses the distribution of potential well asymmetries
is usually taken as white glasses , but one might argue that in metals , symmetric double - wells could result from crystalline symmetries @xcite .
we note that a similar assumption is made in calculations based on a strong coupling to tls kroha .
complementary information on interactions was obtained from resistance measurements . we have fabricated long gold wires in the same deposition machine as for the energy relaxation experiment .
to enhance adhesion of the au film on the substrate , we used two different methods . for a first sample ( au1 ) @xcite , we evaporated first @xmath109 of aluminum , and oxidized it . on a second sample ( au2 ) ,
the surface of the sample was ion - milled just before gold deposition .
a more complete set of data was taken on sample au2 , and we report only here the results on this sample .
its length , width and thickness are @xmath110 @xmath111 and @xmath112 respectively .
the low temperature resistance was @xmath113 assuming einstein s relation and a rectangular cross - section , we deduce the diffusion constant @xmath114 @xmath115rrrin another fabrication run conducted at michigan state university , we have joule evaporated another sample , called aumsu , with very pure gold ( 99.9999% , _ i.e. _ @xmath116 of impurities ) , with @xmath117 @xmath118 , @xmath119 and @xmath114 @xmath115 .
all the samples were measured in the same top - loading dilution refrigerator .
we show in fig .
[ rvst ] the temperature dependence of the resistance for samples au2 and aumsu .
the contribution of weak localization for au2 , smaller than @xmath120 ( see below ) , can be neglected . for aumsu
, this contribution , up to 3@xmath121 in @xmath122 has a well understood temperature dependence ( see below ) , and was substracted .
the contribution of electron - electron interactions to the resistivity , @xmath123 waves , with @xmath124 and @xmath125 plotted as dotted lines in fig .
[ rvst ] , accounts well for the data of aumsu @xcite , where the fit parameter @xmath126 is very close to the calculated value @xmath127 in contrast , the variations observed in au2 are stronger and have a different temperature dependence .
a good fit of the data can be obtained with : @xmath128 with @xmath129 the logarithmic term is characteristic of the kondo effect @xcite , and it was in particular studied in great detail in dilute alloys of fe in au giordano , bergmann , chandrasekhar , mwkondo , where @xmath130 was found to be proportional to the fe concentration and dependent on the cross section of the wire as well as on disorder .
assuming that the logarithmic dependence of the resistance of au2 is due to fe impurities , we get , by comparison with a sample with similar width , thickness and diffusion constant ( aufe2 in mwkondo ) , an impurity concentration @xmath131 which is compatible with the purity of the gold source used for evaporation ( 99.99% , _
i.e. _ @xmath132 of impurities ) @xcite .
we present in the left panel of fig .
[ rvsb ] the magnetoresistance of au2 and aumsu . around zero magnetic field , the resistance presents a dip , as expected from weak antilocalisation aa .
a large - scale negative magnetoresistance is found on au2 , which is similar to measurements performed on dilute alloys of fe in au giordano , mwkondo .
it can be attributed to the freezing of the magnetic moment of the impurities with the magnetic field @xcite . in order to extract the phase coherence time from the magnetoresistance of au2 , we subtracted out the large scale magnetoresistance , which was , at each temperature , fitted with @xmath133 the remaining magnetoresistance was then fitted with the predictions of the weak localization theory @xcite . in the right panel of fig .
[ rvsb ] , we show the phase coherence time @xmath134 of samples au1 , au2 and aumsu as a function of temperature . upon cooling ,
the phase coherence time of au1 and au2 remains unchanged from @xmath0 to @xmath135 : @xmath136 then it increases roughly as @xmath137 .
a simple interpretation of the very low value of @xmath138 and of the desaturation below @xmath1 is found by comparing with similar measurements performed on gold wires in which a small amount of iron was purposely introduced @xcite . the same behavior of @xmath139 was found in these experiments , and it was interpreted as an effect of spin - flip scattering by the impurities , the rate of which presents a maximum at the kondo temperature of fe in au ( @xmath140 ) .
the value @xmath141 of @xmath134 near the kondo temperature was found to be roughly given by @xmath142 with the concentration of impurities @xmath143 expressed in ppm ( parts per million ) mw , mwkondo . in this interpretation
, @xmath144 is obtained for @xmath145 this provides us with another estimation of @xmath143 which is of the same order of magnitude as what we deduced from the resistance vs. temperature measurements .
sample aumsu does not show any magnetoresistance at large scale , and the phase coherence time presents no saturation down to @xmath146 ( see fig .
rvsb ) . a good fit of the data could be obtained with the theoretical dependence @xcite @xmath147 with @xmath148 and @xmath149 the theoretical value from the theory of electron - electron interactions is @xmath150 @xcite@xmath49 to allow for a comparison with the results of mohanty _ et al .
_ , we have also plotted on the same figure the phase coherence time of samples au-3 and au-6 in @xcite , which have similar size and diffusion constants as aumsu ( au-3 : @xmath151 , @xmath152 and @xmath153 @xmath115 ; au-6 : @xmath154 , @xmath155 and @xmath114 @xmath115 ) .
it is clearly seen that our data do not display the saturation found in @xcite .
we had already found no saturation in measurements on silver wires @xcite , which were made from very pure ( 99.9999% ) ag .
we suspect that energy exchanges in samples fabricated with very pure au would be consistent with the theory of ref .
@xcite , as in very pure ag @xcite , but this experiment remains to be performed .
it is noteworthy that the low - temperature saturation of @xmath156 reported in cu wires @xcite was identically found in other samples obtained from 99.9999% cu , supporting the hypothesis that surface oxide could play a major role in cu haesendonck .
the data presented in section 2 provide strong evidence that some of the gold wires we have studied contain several tens of ppm of iron .
the question arises as to whether the energy exchange measurements can also be explained by the presence of these magnetic impurities .
kaminski and glazman glazman recently pointed out that @xmath157 is obtained in second order perturbation theory from the interaction of two quasiparticles with a magnetic impurity .
the effective matrix element @xmath158 of the second order process is proportional to @xmath159 where @xmath160 is the coupling parameter between the quasiparticles and the magnetic impurity and @xmath161 is the energy difference between the intermediate ( virtual ) state and the initial state . in the intermediate state ,
one quasiparticle is promoted to an energy state higher by @xmath108 than the initial state , whereas the magnetic impurity has been reversed .
one thus obtains @xmath162 which has the energy dependence observed in the experiment , but the bare coupling parameter @xmath163 is too small@xmath49 if one assumes that we have fe impurities in au , @xmath164 is of the order of the energies probed in the energy exchange experiment , and the coupling constant is expected to be renormalised by the kondo effect .
this effect has been treated in @xcite , and the proposed expression @xmath165 has now the right order of magnitude but too strong a @xmath33-dependence .
further work is clearly needed to clarify this issue .
s. wind , m.j .
rooks , v. chandrasekhar , and d.e.prober , phys .
lett . * 57 * , 633 ( 1986 ) ; p.m. echternach , m.e .
gershenson , h.m .
bozler , a.l . bogdanov and b. nilsson , _ phys .
b _ * 48 * , 11516 ( 1993 ) .
f. pierre , h. pothier , d. esteve , and m. h. devoret , _ j. low temp . phys . _ * 118 * , 437 ( 2000 ) .
the measurements in all ag wires can be accounted for by the effect of electron - electron interactions together with phonon emission , which we had incorrectly neglected for the @xmath166-long wires in this work .
the predicted intensity of the electron - electron interaction @xcite was between 3 and 12 times smaller than the experimental determination .
we generally find that the current voltage characteristic of the junction presents a small hysteresis at voltages @xmath167 , that we attribute to surface effects on the probe electrode .
when applying a small magnetic field ( @xmath168 ) perpendicularly to the surface of the chip , this singularity disappears .
the magnetic field is kept at the same value during all the experiment .
we are grateful to l. glazman , y. imry , h. kroha , a. zawadowski and h. grabert for useful discussions and comments , to jian huang for fabricating sample aumsu and to p.f .
orfila for technical assistance .
this work has been partly supported by the bureau national de la mtrologie and by nsf grant dmr-9801841 .
travel between saclay and east lansing was supported by nato grant crg-970273 . | we have measured in gold wires the energy exchange rate between quasiparticles , the phase coherence time of quasiparticles and the resistance vs. temperature , in order to probe the interaction processes which are relevant at low temperatures .
we find that the energy exchange rate is higher than expected from the theory of electron - electron interactions , and that it has a different energy dependence .
the dephasing time is constant at temperatures between 8 k and 0.5 k , and it increases below 0.5 k. the magnetoresistance is negative at large field scales , and the resistance decreases logarithmically with increasing temperatures , indicating the presence of magnetic impurities , probably fe . whereas resistivity and phase coherence measurements can be attributed to magnetic impurities , the question is raised whether these magnetic impurities could also mediate energy exchanges between quasiparticles .
several recent experiments have demonstrated that the low - energy properties of quasiparticles in metallic thin films are sample - dependent . on the one hand ,
the power - law increase of the phase - coherence time with decreasing temperature , predicted by the theory of electron - electron interactions in diffusive wires @xcite , has been observed in several experiments @xcite .
the energy exchange rates between electrons was found to be in agreement with this theory in experiments on silver wires @xcite . on the other hand ,
the dephasing rate of quasiparticles was found to saturate at low temperature in a series of gold wires @xcite , and the energy exchange rate between quasiparticles in copper wires to display an energy dependence different from the predicted one @xcite .
we present here measurements on gold samples , in which the energy exchange rates have the same energy dependence as was observed in copper and have a magnitude even higher . in order to investigate the origin of this effect
, we have performed resistance measurements on samples fabricated similarly .
the logarithmic dependence of the resistance , the negative magnetoresistance at large field and the temperature dependence of the phase coherence time , which is constant between @xmath0 and @xmath1 and increases at lower temperature , suggest the presence of magnetic impurities , which might mediate electron - electron interactions . |
the central region of our galaxy is dominated by a bar @xcite , which consists of an x - shaped box / peanut bulge at @xmath9 @xcite and a vertically thin part , the _ long bar _ , extending to @xmath10 @xcite .
recently , the apache point observatory galactic evolution experiment ( apogee ) survey @xcite has revealed that the line of sight velocity @xmath1 distributions of @xmath11 k / m giant stars towards the galactic centre at galactic coordinates @xmath12 degrees and @xmath13 degrees exhibit distinct shoulders at high velocities @xmath2 @xcite .
this result was followed up by other surveys , which yielded differing results : on the one hand , @xcite analysed the @xmath1 distribution of @xmath14 bar red clump stars at @xmath15 degrees and @xmath16 degrees and found hints of a distinct high @xmath1 component at @xmath17 at positive @xmath3 and additionally of a distinct low @xmath1 component at @xmath18 at negative @xmath3 .
on the other hand , @xcite failed to detect distinct high or low @xmath1 components in a survey of @xmath19 red clump stars in the central galaxy .
however , they were probing different sight - lines , most of them at @xmath20 degrees and thus further away from the plane than the other authors , or pointing more closely towards the galactic centre at @xmath21 .
bars are rotating triaxial structures which are expected to be supported mostly by stars on prograde orbits parented by the @xmath7 family of closed long - axis orbits @xcite .
these orbits have proven essential in explaining the @xmath22 diagram of gas towards the galactic centre @xcite and are thus expected to be an important ingredient for explaining any observation of the kinematics of bar stars . @xcite
have recently shown that they could recover a distinct high @xmath1 peak in an @xmath0-body simulation of a barred galaxy if they analysed only 2:1 resonant orbits elongated along the long bar axis , i.e. @xmath7 orbits .
they did not recover the feature when taking into account all stars .
previously , @xcite and @xcite had failed to recover the observational feature in @xmath0-body simulations of milky way ( mw ) like galaxies .
this led @xcite to conclude that the high @xmath1 peak is either a spurious feature due to low numbers of stars or not connected to stars on bar orbits .
here we analyse a controlled simulation of a growing mw - like disc galaxy with respect to @xmath1 distributions towards the barred galactic centre .
the model galaxy features a live dark halo and is continuously fed with stellar and gas particles on near circular orbits .
this type of simulation was introduced by @xcite for a 2d disc but has scarcely been used since ( e.g. @xcite for a specific problem ) .
the disc develops a bar early on and becomes comparable in structure to the mw at later stages , which allows us to create @xmath1 distributions and compare them to the apogee data . with the help of this simulation
, we explore the origin of the high @xmath1 structures and fit the detections and non - detections into a consistent picture .
we find that these features are associated with a kinematically cool population of stars running alongside the bar and likely recently captured from the galactic disc as a consequence of bar growth .
our paper is organised as follows : basic information about the data and simulation are found in section 2 ( survey data , selection function ) and section 3 ( simulation setup ) .
the general evolution of disc and bar in our simulation is discussed in section 4 .
detailed kinematics are analysed in section 5 with focus on the high @xmath1 peak . in section 6
we apply the apogee selection function and directly compare the model to the data .
we also explore some variation of assumed parameters , e.g. latitude @xmath23 and bar angle @xmath24 . in section 7
, we use the successfully identified high @xmath1 peak for a detailed study of the contributing orbits .
finally , we conclude in section 8 . and age @xmath25 at a fixed metallicity of @xmath26 .
_ top panel _ : contour plot as a function of @xmath27 and @xmath25 . _
bottom panel _ : the age dependence at a typical bar distance of @xmath28 .
note the strong bias towards stars with ages @xmath29 gyr at larger distances .
, width=302 ]
for kinematic data from the central bar / bulge region of the mw we use data from the apache point observatory galactic evolution experiment ( apogee , @xcite ) , an infrared , high - resolution survey , which is part of the sloan digital sky survey iii ( sdss * ? ? ?
the fibre - spectrograph allows for @xmath30 spectra per plate covering a spectral range of @xmath31 in the @xmath32 band . as an infrared survey
, apogee has a unique capability to obtain information from stars in the high - extinction regions within the galactic plane and towards the central region of the mw .
we use the apogee data publicly available via the cas for sdss dr10 @xcite . throughout this work ,
we correct all observed stellar velocities from apogee to the galactic standard of rest , using @xmath33 , in concordance with @xcite and @xcite .
another advantage of the survey is the benign selection function described in @xcite .
stars in the target magnitude range with dereddened colours @xmath34 are targeted evenly . in the case of the central galaxy fields used within this work ,
stars were targeted for @xmath35 . to model the selection function
we employ the population synthesis machinery from the @xcite model .
we apply a standard salpeter imf with an exponent of @xmath36 , neglecting binarity of stars .
the population synthesis uses a dense grid of b.a.s.t.i .
stellar models ( @xcite , for their colour calibrations see @xcite ) . at each distance @xmath27 and age @xmath25
we count the number of stars per mass of the initial population that would end up within the colour and magnitude selection of apogee .
the result is shown in figure [ selfu ] for a fixed metallicity of @xmath37 .
the use of a single , slightly super - solar metallicity is a simplification , as the inner mw has a relatively wide metallicity distribution around this value @xcite .
however , all relevant metallicities show a similar preference for the detection of young stars .
as the detailed age - metallicity distributions for the relevant sightlines are not well known and the detailed connection between model galaxy and mw in terms of structure and formation history is unclear , we choose to keep the model simple and use one typical metallicity .
the band of high selection probabilities at very small distances derives from main - sequence dwarf stars .
the stellar models are not fully reliable in this region for such cool dwarfs . however , this foreground is strongly suppressed by the geometric factor ( @xmath38 ) of the pencil beam survey .
the number of stars currently in the giant stages of interest to us declines strongly ( by roughly one order of magnitude ) with age , implying a strong bias in favour of young stars within the apogee survey .
we apply this selection function to our @xmath0-body simulation by weighting each particle with its respective selection probability .
apogee uses dereddened magnitudes for targeting , however there are residual effects from reddening uncertainties , which we neglect here .
in this paper we present a controlled collisionless+hydrodynamic simulation of a growing galactic disc embedded in a non - growing dark matter halo .
the simulations were all carried out with the tree smoothed particle hydrodynamics ( treesph ) code gadget-3 , last described in @xcite .
the applied gravitational softening lengths are @xmath39 for stellar and gaseous disc particles and @xmath40 for halo particles .
we use an opening angle for the tree code of @xmath41 . an adaptive time - stepping scheme which assigns time steps in bins @xmath42 according to @xmath43 to individual particles
is applied . here
@xmath44 is an accuracy parameter and @xmath45 is the gravitational acceleration at the position of particle @xmath46 with softening @xmath47 .
the base timestep @xmath25 is appropriately chosen by the code and a maximum of 10 myr is allowed for timesteps .
the hydrodynamical timestep is based on a courant - like condition @xmath48 , where @xmath49 is the courant parameter , @xmath50 is the smoothing length and @xmath51 is the soundspeed .
we apply an isothermal equation of state @xmath52 with @xmath53 and use @xmath54 neighbours for the smoothing kernel . for further code details
we refer to @xcite . to generate initial conditions ( ics ) for our numerical experiments we use the publicly available galic code @xcite , which produces near - equilibrium ics of multi - component collisionless systems with given density distributions using an iterative approach .
the initial system consists of a dark halo with a mass of @xmath55 represented by @xmath56 particles and an embedded thin galactic disc with a mass of @xmath57 represented by @xmath58 particles , so that the disc particle mass is @xmath59 .
the dark halo has a hernquist profile with @xmath60 the inner profile is adjusted so that it is similar to an nfw profile with concentration @xmath61 and virial radius @xmath62 .
this leads to a scale radius @xmath63 .
the halo initially has a spherical density profile and radially isotropic kinematics , i.e. equal velocity dispersion in the principal directions , @xmath64 and consequently a value of the anisotropy parameter of @xmath65 .
the disc is set up with a mass profile @xmath66 with an exponential scalelength of @xmath67 and an isothermal vertical profile with a radially constant scaleheight of @xmath68 .
the vertical velocity dispersion @xmath69 thus declines with radius .
we assume @xmath70 , so that toomre s @xmath71 shows a minimum value of @xmath72 at @xmath73 . for further information regarding the ics we refer to aumer et al .
( in prep . ) .
we note that here the halo is rather poorly resolved with @xmath56 and a rather short softening length ( @xmath74 ) is applied .
the low @xmath75 is due to limited computational resources and a preference to use them to resolve the disc .
as expected ( e.g. @xcite ) , purely collisionless ( no sph ) tests with a higher @xmath76 show reduced disc heating and less bar slowdown .
yet the disc heating for the simulation studied here is comparable to the mw , with similar velocity dispersions for 10 gyr old solar neighbourhood stars , since heating by massive halo particles compensates the lack of heating by giant molecular clouds or other sources such as satellite galaxies .
as far as bar slowdown due to angular momentum transfer to the halo is concerned , the radius and rotation frequency of bars differ by less than @xmath77 20 per cent at late times ( 6 - 10 gyr ) , the time relevant for this work .
+ to model the continuous growth of galactic discs via star formation and gas accretion we , over a timespan of 10 gyr , add new particles with @xmath59 to the existing disc every 5 myr .
we assume a mass growth rate history given by @xmath78 with an exponential decay timescale @xmath79 , which is motivated by the findings of @xcite , and a normalisation @xmath80 , so that a total of @xmath81 are fed over 10 gyr .
the added disc mass is radially distributed as @xmath82 , with an exponential scale - length that increases with time as @xmath83 this increase from @xmath84 to a scale - length at 10 gyr @xmath85 is motivated by the findings of @xcite .
the particles are randomly scattered over azimuth @xmath24 and placed at @xmath86 .
the coordinate system is regularly updated to be centred on the centre of mass of the system .
the particles are assigned near circular orbits .
we determine the new particles rotational velocities as @xmath87 for a particle @xmath46 at cylindrical radius @xmath88 and the component of the gravitational acceleration pointing towards the centre , @xmath89 . for stellar particles ,
we add random velocity components in all three directions @xmath24 , @xmath90 and @xmath91 , drawn from gaussian distributions with @xmath92 .
after a disc goes bar unstable , there are no stable circular orbits in the bar region .
so we impose an inner cutoff radius @xmath93 , within which no particles are introduced .
we use the radius where the logarithm of the @xmath94 fourier amplitude @xmath95 drops below @xmath96 as an inner cutoff radius . in disc galaxies ,
bars slow down and grow in size due to the transfer of angular momentum to the dark halo @xcite .
this loss is to some extent balanced by the transfer of angular momentum from the gas to the bar @xcite .
therefore , it is desirable to include a gas component in barred disc galaxy models . we do not wish to introduce a full model containing the accretion of gas onto the galaxy , its cooling and the subsequent star formation and feedback processes ( see e.g. @xcite for an isolated galaxy model ) , as uncertainties regarding the modelling of the physical processes are high
( see e.g. @xcite ) . to keep the model simple ,
we add a constant fraction of gas to our galactic disc and model it with an isothermal equation of state @xmath52 with @xmath53 .
we simulate the growth of the disc by continuously adding stellar and gas particles .
the initial gas disc is created by turning 5 per cent randomly chosen stellar particles into sph particles .
this mild disturbance of the ics is unimportant compared to the rapid onset of disc instability in the growing galaxy . during disc growth we aim at a gas mass fraction in the disc of 10 per cent as observed in modern day mw like disc galaxies .
we achieve this by adding 10 per cent of new particles as gas when the mass fraction of gas in the galaxy is above 10 per cent and 20 per cent of new particles as gas whenever it is lower .
computationally , high gas densities in the central galaxy are very expensive .
therefore we include a number of measures to limit the central density : i ) prior to the bar forming and the inner cutoff on particle insertion coming into play , we create no gas particles at @xmath97 .
ii ) we include a model for star formation in the very central galaxy .
we turn sph particles with hydrogen number densities @xmath98 and specific angular momentum @xmath99 into star particles with a probability @xmath100 , where @xmath101 is an efficiency parameter , @xmath102 is the local dynamical timescale and @xmath103 is the timestep of the particle . to mimic the effect of a central galactic outflow @xcite
we only turn one out of five particles into a star and remove the other 4 out of five from the simulation .
note that only gas particles in the inner bar region are affected by these prescriptions .
in this section we discuss the structure of our model of a barred disc galaxy .
note that we are comparing a specific - and in no way fitted - model to a specific galaxy , the milky way . though our model should not provide a perfect match to our galaxy
, we expect the model to qualitatively share important characteristics .
note that due to computational limitations we currently have only one model of this type including hydrodynamics available .
+ in figure [ bar - evo ] we present the evolution of bar parameters in our simulation .
we determine a bar radial extent @xmath104 by first calculating the @xmath94 fourier amplitude profile @xmath105 where n(r ) is the number of particles in the radial bin centred on @xmath90 and we only take into account stellar particles .
then we define @xmath104 as the radius where @xmath106 drops below -1.5 or as zero if it never reaches the value in the central area .
in the left panel of figure [ bar - evo ] we see that the bar forms in the first gyr , but its radius and strength are varying significantly . from @xmath107 gyr on the bar grows almost linearly in time to reach @xmath108 at @xmath109 gyr .
our definition for @xmath104 allows no exact comparison to the values cited for the length of the mw long bar . @xcite
determine a length of 5 kpc and give 4 kpc as a lower limit , not in disagreement with our model .
+ in the middle panel of figure [ bar - evo ] we present the evolution of the bar rotational frequency @xmath110 determined by following the angle of the bar major axis from snapshot to snapshot .
the bar initially rotates with a frequency of @xmath111 , but as the loss of angular momentum to the halo is not balanced by the transfer of angular momentum from the inflowing gas component , @xmath110 steadily declines to reach @xmath112 at the end of the simulation . for the mw bar this quantity is little constrained .
while our value is lower than most estimates for the mw @xcite , the value is in concordance with a recent estimate by @xcite ( see their section 8.3 for a discussion of deduced mw bar pattern speeds ) .
our model is thus at the lower end of values inferred for the mw . in figure [ images ] we present surface density projections of the whole galaxy model and the bar region at a simulation time @xmath113 gyr .
the lower panels show face - on views of the stellar and the gaseous components . apparently , the galaxy is dominated by the bar within @xmath114 and shows flocculent spiral structure at outer radii .
the upper panels give close ups of the bar region .
the face - on view shows two distinct radial regions in terms of minor axis extent . at @xmath9 ,
the bar shows a mildly elongated structure , at @xmath115 a thinner , long bar structure is visible . the edge - on view reveals , that the region @xmath9 is dominated by an x - shaped structure with the tips of this structure at @xmath116 , very similar to the structure of the mw bulge / bar region inferred by @xcite .
this view also shows that the outer bar is significantly thinner . around the tips of the bar we measure a vertical exponential scale - height of 180 pc , similar to the long bar scale - height inferred by @xcite , but without a super - thin 45 pc component as found by these authors .
such a component is likely too thin to be resolved in our simulation , considering our softening length of 30 pc . at a solar neighbourhood like radius @xmath117
, we find that our model at 7.3 gyr has a stellar exponential scale - length of 3.5 kpc , which grows mildly towards 10 gyr .
the disc is thus likely more extended than that of the mw , for which estimates usually range at 2 - 3 kpc .
the model has a total baryonic mass of @xmath118 at 7.3 gyr , it has grown from a mass of @xmath119 at @xmath120 . in the right panel of figure [ bar - evo ]
we plot the circular velocity curve @xmath121 in the plane of the model galaxy at 7.3 and 9.3 gyr .
the contributions from dark matter and baryons are shown in red and blue .
we measure @xmath122 by calculating the component of the gravitational acceleration pointing towards the centre @xmath123 on an @xmath124 grid at @xmath86 and at each @xmath90 average over the equally distributed values of @xmath24 .
we note that the circular velocity curve in the centre of the mw is not well constrained and that it is uncertain how the observations of gas velocities in the mw bar are connected to our definition of @xmath121 .
there is only mild evolution from 7.3 to 9.3 gyr visible in @xmath121 .
the circular velocity curves show features at the bar radius , deriving from the baryonic matter distribution .
the reason for these features is that the bar redistributes matter towards the centre , whereas by construction of the model new baryons are added outside the bar region , which gives rise to the baryonic dip at @xmath125 . in our model ,
the baryonic contribution to the circular velocity is greater than the dark matter equivalent only within @xmath126 .
it is unclear how this relates to the mw , but our model agrees with two important constraints : a ) at @xmath117 and 7.3 gyr we recover @xmath127 , in good agreement with what @xcite finds from local kinematics .
b ) the total dark matter mass within a sphere of @xmath128 is @xmath129 in good agreement with the dark matter mass within a sphere of solar galactic radius that @xcite have deduced . at the evolution stages relevant for this paper , the model has a slow bar with @xmath130 , where @xmath131 is the bar corotation radius .
this is likely connected to the fact that our galaxy model is _ sub - maximum _ , i.e. the dark matter contribution to the circular velocity curve is significant in the bar / disc region , as fast bars with @xmath132 are thought to require _ maximum _ discs with baryon - dominated circular velocity curves @xcite .
in the previous section we have shown that the bar emerging in our galaxy model has comparable structural parameters to the mw bar .
we now extract line of sight velocity @xmath1 distributions from our model and qualitatively compare them to the results of @xcite . in this section ,
we limit ourselves to the general understanding of the observed phenomena .
we will attempt a quantitative comparison to observed data in the following section .
as we expect stars moving along the major axis of the bar to be the source of the stars that comprise the high @xmath1 peak , we focus here on the central region of the model , defined by galactocentric @xmath133 .
the full model including the selection function will be discussed in the next section . to create @xmath1 distributions we place ourselves at a solar radius of @xmath134 @xcite and orient the bar to an angle of @xmath24 between its major axis and the sun - galactic centre sightline .
values found for @xmath24 in the literature range between 15 and 40 degrees . for most of this paper we choose @xmath135 degrees and show in the next section that our conclusions are unchanged for the full range of cited values .
we usually define sightlines which are 2 degrees wide in @xmath3 and @xmath23 and centred on the given values .
to curb poisson noise we use the @xmath136 ( @xmath94 ) symmetry of the model and use two sightlines per snapshots , 180 degrees apart in azimuth . for each plot
we stack 50 snapshots , separated in time by 1 myr and rotate the bar to achieve the same value of @xmath24 for each snapshot . in figure
[ vlos - all ] , we show @xmath1 histograms at latitudes @xmath137 degrees at a simulation time of 7.3 gyr for eight different longitudes @xmath3 between @xmath138 and @xmath139 degrees .
the figure shows histograms for all bar stars with @xmath133 in blue and for the subset of young star particles with ages @xmath140 gyr in red .
the @xmath1 distributions for all stars ( blue lines ) do not reveal significant features at high velocities .
this is consistent with recent examinations of simulations in @xcite , @xcite and @xcite , who also detected no features when all stars are taken into account .
+ @xcite reported peaks of high velocity stars with @xmath141 of varying extent for longitudes
@xmath3 between @xmath138 and @xmath142 degrees and @xcite reported such peaks for @xmath143 and @xmath144 degrees and a peak at @xmath145 at @xmath146 degrees .
closer inspection of the blue histograms at @xmath147 and @xmath148 degrees reveals an excess of stars at the expected @xmath1 values .
moreover , the position of the main peak in the @xmath1 distribution decreases with decreasing @xmath3 as observed by @xcite .
these main peaks of the distributions are , however , wider in the model than in the observation .
the fact that the apogee selection function ( figure [ selfu ] ) favours young stars motivates to examine only stars with ages @xmath140 gyr shown as red curves . note that stars formed from low angular momentum central gas do not play a role for this analysis and are excluded here .
the main @xmath1 peaks are much narrower for young stars than for all stars , indicating that these populations are cooler .
the red histograms clearly show distinct components at @xmath149 for @xmath150 and @xmath151 degrees and at @xmath152 for @xmath153 and @xmath139 degrees , in agreement with observations .
interestingly , we do not recover a separate component at @xmath154 degrees , where @xcite also claimed to find a distinct high @xmath1 peak , a finding which we will return to in the next section . as the model young populations show the observed signature , it is interesting to ask whether this is specific to a unique time or stage of bar evolution or a generic feature of young stars . in figure
[ vlos - young ] we examine the evolution of young stellar populations with time plotting the populations which at respective times are less than 2 gyr old .
we consider simulation times 7.3 , 8.3 and 9.3 gyr .
we notice several points : i ) there is very little change in the general shape of the histograms and in the position and width of the main @xmath1 peaks at all longitudes @xmath3 .
ii ) the extreme values of @xmath1 increase with time by around @xmath155 for all sightlines which do not cross the galactic centre .
iii ) the high @xmath156 peaks identified in figure [ vlos - all ] become slightly more distinct in time and also move to higher velocities .
iv ) for 9.3 gyr additional features / peaks become apparent at the opposite side of the main peaks , e.g. at @xmath157 at @xmath158 degrees or at @xmath159 at @xmath146 degrees .
the high @xmath1 features at positive longitudes and low @xmath1 features at negative longitudes are thus stable over time and not connected to any temporary phenomenon . due to the lack of opposite peaks in the apogee data we will focus our analysis on the simulation time 7.3 gyr . so far we have shown that young stars show high @xmath156 peaks , whereas the whole population of bar stars does not .
it is thus interesting to study in more detail how the @xmath1 distribution evolves with age .
we attempt this in figure [ vlos - age ] , where we concentrate on one sightline , @xmath160 degrees , for which both our simulation and apogee data display a peak at high @xmath1 .
panel e shows how the @xmath1 distribution varies when stars are sorted in four different age bins . whereas the high @xmath1 population is very distinct for stars with ages below 1 gyr , this population becomes less distinct at higher ages .
for stars with ages between 1 and 2 gyr , it is still clearly visible , at 2 - 4 gyr only a small feature remains and for older stars even this weak signature vanishes .
the vanishing of the distinct high @xmath1 population goes hand in hand with broadening of the main @xmath1 peak indicating a heating of the population . + to see how this evolves with simulation time , we plot in panels a - d stars in these age bins at three different simulation times , 7.3 , 8.3 and 9.3 gyr .
we notice that there is much less evolution with simulation time than there is with age .
the level of distinction of the high @xmath1 signature remains similar at all considered simulation times for each age bin . to strengthen these findings we plot in panel f of figure [ vlos - age ] the population of stars born between 6.3 and 7.3 gyr at three different times representing mean population ages 0.5 , 1.5 and 2.5 gyr .
this underlines that the evolution of the @xmath1 distribution with age shows the broadening of the main @xmath1 peak due to heating and the gradual loss of the high @xmath1 peak .
we again notice that the young populations of stars at 9.3 gyr show additional features at the opposite side of the main peak .
closer inspection reveals hints of this feature also at earlier times .
additionally , the main peaks move to higher @xmath1 .
we will provide an explanation for these changes in section 7 .
we note , however , that these additional peaks are only visible if we select young bar stars .
when applying the selection function on all stars along the line of sight , as described in the next section , these features , unlike the conventional high @xmath1 peaks , are too small to be detected .
in the previous section , we have shown that populations of young bar stars are expected to display high @xmath156 features as observed by @xcite and @xcite . in reality ,
not only stars with @xmath133 will contribute to the apogee @xmath1 distributions , but also disc stars in front of or behind the bar . in the following ,
we directly compare the @xmath0-body simulation to the apogee data , applying the selection function described in section 2 .
+ we start with figure [ compare ] , where we compare the @xmath1 histograms for stars with galactocentric radii @xmath133 ( see previous section ) to @xmath1 histograms obtained for the full sightlines .
we do so for three longitudes , @xmath161 and 14 degrees .
the blue curves represent all stars within 4 kpc and the red curves show the subset of young stars , both are taken from figure [ vlos - young ] .
the velocity distributions for all stars along the sightline ( equally weighted , green ) are strikingly different .
their main @xmath1 peaks are significantly narrower and shifted by more than @xmath162 towards zero .
this is due to the dominant contribution by fore- and background disc stars , which are kinematically cold and are observed with the local azimuthal direction nearly perpendicular to the line of sight .
we noticed above that no distinct high @xmath156 feature is found for all stars at @xmath133 .
interestingly this is not true for all stars along the line of sight . in this population , the feature is less distinct than the one for young bar stars , but lies at the same @xmath1 range .
the histograms obtained when weighting all stars along the line of sight according to the selection function presented in section 2 are shown as black lines .
overall , they resemble more the ones for all stars along the line of sight than to the ones for stars at @xmath133 .
when present , the high @xmath156 feature is strengthened by the selection function and the main peak is shifted to slightly larger @xmath156 . from these comparisons
we conclude that the @xmath1 distribution is not only shaped by stars in the bar region , but by all stars along the line of sight .
this is illustrated by the selection function weighted ( sfw ) distribution of stars along the line - of - sight @xmath163 as a function of distance from the sun @xmath27 in the left panels of figure [ sightline ] ( dashed lines ) .
clearly , the sfw distribution functions do not simply reflect the density distributions along the pencil beams , but are shaped by the age and distance dependent selection function modulated by the beam area increasing as @xmath38 .
of the three shown sightlines , only the one which crosses the bar most centrally ( @xmath158 degrees ) has its absolute maximum coincident with the bar distance .
the loci of young stars , the highest densities of which are by construction at a galactocentric radius @xmath164 and thus at distances @xmath165 and @xmath166 are visible as local maxima . moreover ,
spiral structure creates minor peaks .
the high @xmath1 feature appears connected to young bar stars . to verify this assumption , we plot in the left panels of figure [ sightline ] the contribution to the high @xmath167 features of sfw stars for the three sightlines of figure [ compare ] as a function of distance from the sun @xmath27 ( solid lines ) .
the intervals we use as definitions for these features are indicated by dotted vertical lines in figure [ compare ] . for @xmath161 degrees ,
we were guided by the distinctly visible feature , for @xmath154 degrees , where we find no feature , we use the estimate by @xcite .
+ at each sightline in figure [ sightline ] the regions of contributions to the high @xmath167 stars are much narrower than the sfw distribution of all stars in the pencil beam . moreover , the maxima of contribution to the high @xmath167 stars are ( apart from @xmath158 degrees ) offset from the maxima of the sfw distribution of all stars .
the distance distribution of contributing stars @xmath163 is narrowest for @xmath158 degrees and widest for @xmath154 degrees .
the main peaks are at @xmath168 on the rear side of the bar ( front side for @xmath146 degrees ) . to visualise where these contributions arise in relation to the bar , we use the right panel of figure [ sightline ] , where we show the position of the sun in yellow , the considered sightlines in blue , red and orange and the surface density contours of the stellar mass in the bar region at 7.3 gyr in black ( see figure [ images ] ) .
the 95 per cent contribution regions of high @xmath167 stars for each sightline are shown in magenta for stars moving away from the sun and in cyan for stars moving towards the sun at negative longitudes .
we highlight the regions of @xmath169 around the peak distances in darker shades . for @xmath161 degrees ,
these regions are fully within the bar region and strictly on one side of the bar ( rear side for @xmath170 , front for @xmath171 ) , as the fast stars travelling alongside the bar have to lose angular momentum on the way in and gain it on the way out . the contribution region is narrower for @xmath158 degrees as this sightline crosses the bar at a greater angle than the one for @xmath146 degrees .
this indicates that the high @xmath156 features found by @xcite correspond to stars which are associated with the bar region and not with disc orbits .
this confirms the ideas of @xcite and @xcite and rejects the notion of @xcite that the observed features are not related to the bar . at @xmath154 degrees
there is a tentative feature in observed data , but not in our simulations . if we select the highest @xmath1 stars in the simulation we find that the main contributions come from the edge of the bar , but extend far into the disc region behind the bar . at these longitudes ,
the inner parts of the disc align with the line of sight and start to contribute to the high @xmath1 window .
should the high @xmath1 feature at @xmath154 degrees prove to be real , a clear attribution may only be possible with detailed analytical models .
+ we now directly compare our model with apogee data .
remember that we have done nothing at all to fit this specific model galaxy to the mw
. there may be significant differences in length , mass , age and detailed history of the bar etc .
hence the goal is to qualitatively identify features and not to do a full quantitative analysis . in figure [ apogee ]
we compare @xmath1 histograms for all stars along the sightlines , equally weighted ( ew , green ) or selection function weighted ( sfw , black ) to the apogee data described in section 2 for various sightlines .
we display in - plane data at @xmath137 degrees in the upper two rows , and sightlines at higher @xmath172 in the bottom two rows . despite its limitations ,
the simulation fits the data surprisingly well .
the sfw models provide a better fit to apogee data than the ew models .
their high @xmath1 features are more distinct and the ew histograms are offset from the data at large @xmath3 as well as at large @xmath23 .
the sightlines with @xmath173 degrees are increasingly dominated by disc stars . there
, the parts aligned with the line of sight would bury any distinct velocity feature from the bar . in concordance with this expectation
, there is no discernible feature in the corresponding in - plane apogee data , and also not in the model .
in contrast , at @xmath174 degrees , the high velocity feature is prominent both in the mw data and our simulation .
taking into account the significant poisson noise in the observations , there is little difference at @xmath175 degrees . at @xmath143 degrees ,
the feature is less distinct in the simulation . interestingly , at @xmath176 degrees
the sfw histogram provides a worse fit than the ew histogram , as it misses the position of the main peak , which at other longitudes is well reproduced .
when we look slightly above and below the plane at @xmath177 degrees , asymmetries at high @xmath1 are present in apogee data at @xmath178 degrees , but again not at greater @xmath3 .
the tentative @xmath154 peak in the data was highlighted by @xcite , but has a very weak counterpart in the simulation .
the sfw model histogram here drops less steeply than the data on the high velocity side of the main peak , as is the case for @xmath176 degrees ( mentioned above ) .
these are the strongest discrepancies between model and observations
. there may be various explanations for this : \i ) we know that the galactic potential will not be perfectly matched . in particular for disc contributions ,
the location of the main @xmath1 peak is determined by the radial potential gradient ( in idealisation the circular speed ) .
if this differs , simulation and data will show a systematic difference .
ii ) we know that the structure in particular in this region will show differences between simulation and observed data , e.g. caused by a difference in bar length .
a longer bar would mean that we see regions further away from the bar tips .
we can achieve such an effect also by lowering the solar radius .
indeed , if we do so , the asymmetry becomes more prominent .
iii ) another explanation could be the dense ring of molecular clouds at @xmath179 in the mw @xcite .
this ring is associated with an intense radial peak in star formation , which is not matched by our simulations .
the selection function of apogee is very sensitive to young stars .
this feature will have no strong consequence when the sightline is perpendicular to circular orbits at @xmath180 ( at small @xmath3 ) or outside the region ( @xmath181 ) , but sightlines at @xmath182 degrees will be strongly affected . as they pass through @xmath183 at an intermediate angle to the local azimuthal direction , the enhancement of these stars will drive up stellar densities at intermediate @xmath1 and shift the main peak .
overall , the complicated contribution function of stars at these sightlines prevents any strong conclusions to be drawn from this discrepancy . moving even further away from the plane
, we consider latitudes @xmath184 degrees in the lower row of figure [ apogee ] .
there is no discernible feature , either in observations or simulation .
the slight asymmetry is linked to the rotation of the bar .
so our model provides reasonable fits to the data in and out of the plane .
+ we now examine the behaviour at different latitudes .
figure [ diff - b ] shows simulation data for two sightlines @xmath161 degrees and varying latitude .
we do this both for sfw histograms for all stars along the lines of sight ( left panels ) and for ew young central stars ( right panels ) .
the difference between these two types of histograms is even more striking at higher @xmath172 where for young central stars they are significantly broader .
this indicates that young bar stars play a small role in observations at higher latitude .
note also that the number of young bar stars decreases strongly with increasing latitude in the simulations , which explains the noisier nature of the corresponding histograms .
we find that the exponential scaleheight @xmath185 for stars with ages @xmath140 gyr , which contribute to the high @xmath1 features is @xmath186 , compared to @xmath187 for all stars at the ends of the bar .
all panels show that distinct high @xmath156 features are only present close to the plane . at @xmath188 degrees
a mild feature is still present , while at greater @xmath23 there is nothing left .
this explains why @xcite did not discover such features in their data , which are mostly at @xmath189 degrees . their only field at smaller latitude is at @xmath21 , where we do not expect a peak anyway
this can be explained by the fact that the high @xmath1 populations are dynamically cool and thus also live at small vertical distances from the galactic plane .
so far , we have assumed a bar angle of @xmath135 degrees . in the literature
, estimates range from 15 to 40 degrees .
@xcite recently claimed that the high @xmath1 features observed by @xcite could help to constrain @xmath24 and would favour low values @xmath190 degrees . in figure [ phi
] , we thus attempt to study the influence of @xmath24 on @xmath1 distributions .
we consider angles between 10 and 55 degrees for three sightlines and both sfw stars ( upper row ) and young stars with @xmath133 ( lower row ) .
+ we see that both for positive and negative @xmath3 , the high @xmath156 features are less distinct for large @xmath24 .
this can be explained by the larger angle between the sightline and the bar . as the high velocity stars move along the bar , their projected @xmath1 decreases . at larger @xmath24
, stars on different orbits moving at greater angles to the bar can be seen at higher @xmath1 .
as they are less numerous , the feature becomes less prominent .
this is especially clear for young stars at @xmath133 .
if one tried to fit multiple gaussians to the @xmath1 distribution , as did @xcite and @xcite , one would find that the high @xmath1 peak becomes broader and centred at lower @xmath1 as reported by @xcite . in the upper panels of figure [ phi ] ,
apogee data are plotted over the sfw model .
we should keep in mind the uncertainties in the observations and the fact that the structure and age distribution of the model bar differ in an unknown extent from the mw bar .
we thus conclude that the apogee data put only mild constraints on the bar angle .
they favour small angles , but only strongly disfavour angles @xmath191 degrees .
the estimate by @xcite is also intrinsically overly confident , as they ascribe the high @xmath1 feature to one orbit family only , whereas , as we will show in the next section , several families can contribute .
moreover , degeneracies with the potential , the pattern speed , the bar structure etc complicate their determination of @xmath24 .
to understand better the origin of the high and low @xmath1 peaks discussed above , we focus in this section on the orbits producing these features .
we created snapshots of the simulations at a frequency of 1 per 1 myr from simulation time 7.3 gyr onwards .
we first consider the following exercise : what would the @xmath1 distribution look like if it were only made up of stars that at some point contribute to the high @xmath1 peak .
we therefore select all stars with ages @xmath140 gyr and @xmath133 that at 7.307.35 gyr contribute to the high @xmath1 peak at @xmath160 degrees and @xmath137 degrees .
we then follow these particles for a further 500 myr and construct a @xmath1 distribution at the same @xmath3 and @xmath23 ranges only from these stars .
we do the same 2 gyr later for a new set of stars selected in the same way .
figures [ vlos - age ] and [ vlos - young ] have already shown that the maximum ( minimum ) @xmath1 observed increase ( decrease ) by about @xmath155 within these 2 gyr .
this is because of a growth of both the potential well and the bar and also because of the slowdown of the bar which allows orbits to have lower angular momentum and reach positions closer to the centre and thus cover a larger potential difference .
consequently , we increase the high @xmath1 window from @xmath192 to @xmath193 .
+ these @xmath1 distributions are shown in figure [ vlos - peak ] .
unsurprisingly they have their main peaks at the selection positions .
both distributions clearly show a distinct peak at negative @xmath1 made up by orbits returning along the other side of the bar .
this peak was already visible in figure [ vlos - age ] for 9.3 gyr .
the increase in maximum absolute velocities between 7.3 and 9.3 gyr widens the gap between the peaks and shifts the main peak of the overall distributions for stars within @xmath133 to higher @xmath1 .
consequently the negative @xmath1 peak at 9.3 gyr for figure [ vlos - age ] becomes visible .
the negative peak comprises approaching stars while the positive peak features stars moving away from the sun .
figure [ vlos - peak ] reveals a third peak for 7.3 gyr indicating that a group of orbits has a more complicated structure than alternately contributing to the two peaks .
in the right panel of figure [ vlos - peak ] we visualise the origin of the high and low @xmath1 peaks at @xmath193 and @xmath157 for young bar stars at 9.3 gyr from spatially separate components along the line of sight . here
we plot the spatial positions of all stars in the positive @xmath1 peak at 9.3 gyr in black and of all stars in the negative @xmath1 peak at 9.3 gyr in red .
the components clearly separate to the expected sides .
one orbit family that would contribute to both peaks are the @xmath7 orbits . in the rotating frame of the bar , these orbits move along opposing sides of the bar major axis and
turn around at the bar tips .
they would obviously produce the two peak feature in the left panel of figure [ vlos - peak ] , as was e.g. discussed by @xcite .
we overplot in green one such orbit extracted from the simulation which alternately contributes to both peaks .
however , as we will show below , the high @xmath1 features in our simulation are made up of a more complicated population of orbits from different orbit families .
+ to get a better understanding for the orbits that contribute to the peak , we track all orbits that contribute to the high @xmath1 peak at @xmath160 and @xmath137 degrees at 7.30 - 7.35 gyr defined by @xmath192 for young stars with ages @xmath140 gyr and radii @xmath133 .
if an orbit is similar to a closed orbit , we will expect it to return to the same region in phase space periodically .
our selection of sightline plus @xmath1 range can be seen as a rough proxy for a phase space region .
considering that the bar grows and the potential changes , a more sophisticated definition is complicated .
therefore we record the time interval needed by an orbit to return to the sightline and @xmath1 window .
the histogram for the first return time @xmath194 is plotted in the panel a of figure [ hist ] .
we see that this distribution shows three distinct peaks at circa 110 ( red ) , 190 ( green ) and 280 myr ( blue ) , which contain approx .
23 , 13 and 28 per cent of the stars .
there are also two broader peaks at 450 - 650 myr ( 14 per cent , purple ) and 650 - 800 myr ( 7 per cent , orange ) .
note that @xmath77 13 per cent of orbits never return or have return times @xmath195 myr .
were all these orbits simple @xmath7 orbits , we would expect a simpler structure . in panels
b - f of figure [ hist ] , we look at histograms of the second return time @xmath196 for stars selected to have first return times @xmath194 in the five peaks found in the @xmath194 histogram . to some degree , all these @xmath196 histogram resemble the @xmath194 histogram , as they reproduce one to three of the previously found three narrow peaks . above 450 myr , the @xmath196 histograms all seem dominated by noise , but all of them contain a similar fraction of such orbits , varying between 25 and 31 per cent .
a first clear conclusion is , that for each histogram , the fraction of orbits that remain in the same return time peak is below 50 per cent .
this shows that most orbits do not show a steady periodicity in our definition , but rather combine a low number of possible return times . to connect the shape of these histograms to types of orbits , we extracted random sets of orbits with similar combinations of return times @xmath194 and @xmath196 , for example contributors to the green peak for @xmath194 and to the blue peak for @xmath196 .
in addition we extracted a set of orbits that do not contribute to return time peaks .
from these we select by eye 42 typical orbits .
they are shown in figure [ orbits ] .
note that due to different prominence of the peaks this figure does not represent orbit types at their underlying contribution fractions .
the stars with shortest return times @xmath194 ( red peak ) show a high fraction ( 45 per cent ) of stars with @xmath197 .
additionally , 30 per cent of the stars have @xmath196 in the area of the green or blue peak , but these peaks are not separately visible .
the first row of figure [ orbits ] reveals that these stars are essentially all on @xmath7 like orbits .
if they have @xmath198 , it is mostly because the orbits are not exactly closed .
they can miss the high @xmath1 window as their directions of motion and velocity vary when they cross the sightline . moreover , @xmath0-body noise can cause the integrals of motion of the particles to fluctuate . stars from the second ( green ) @xmath194 peak show three @xmath196 peaks at the locations of the three narrow @xmath194 peaks which are equally populated ( @xmath199 per cent each ) .
most stars are again on @xmath7-like orbits , but there are also the first objects from more complex closed orbit families with pretzel - like structures @xcite , such as the 202 - 213 or 216 - 229 objects . at higher return times ,
@xmath7 orbits are a minority .
if they have such return long return times , it is because they miss @xmath1 due to the fluctuations mentioned above more often than not , e.g. because they have lower energies and only in exceptional cases reach high enough @xmath1 .
pretzel - like morphologies are the most frequent pattern found at longer return times . for blue @xmath194 stars ,
there is again a high fraction ( 46 per cent ) of @xmath197 stars , which tend to shows such patterns , as seen in the fourth row of figure [ orbits ] .
generally , the structures become more complicated and more irregular as @xmath194 increases .
clearly some orbits show 3:1 characteristics with a changing orientation ( objects 421 - 402 or 506 - 862 ) , but there are also more complicated , but apparently regular structures , such as the object 273 - 176 , which was labelled q in @xcite , and objects like 176 - 324 or 424 - 103 which also combine 3:1 and @xmath7 features .
other orbits appear to change structure with time , an indication of chaotic behaviour .
if we combine the results from above , we estimate that the fraction of stars in the high @xmath1 peak that are on @xmath7 orbits at the given time is between 40 and 50 per cent .
if we determined the full population of stars , which over a longer timescale ( @xmath200 gyr ) periodically contributed to the high @xmath1 peak , the fraction of @xmath7 stars would be even lower , as they have the shortest return times .
this invalidates the original assumption of @xcite that the high @xmath1 feature is only caused by 2:1 orbits .
there are other major contributors , most prominently the pretzel - like orbits , but also 3:1 or more complex orbits .
we note that @xcite have recently updated their paper in response to our work and now also find that higher order orbit families ( they name 3:1 and 5:2 ) can make significant contributions to the high @xmath1 features .
we note that we have attempted to analyse the vertical structure of orbits in connection to the in - plane one and experimented with frequency analysis , but concluded that the results were to diverse or noisy to help with the understanding of the high @xmath1 features . .
we exclude orbits that never return to the high @xmath1 window or have very long return times @xmath201 0.8 gyr .
we record each passage of the sightline @xmath160 degrees ( total number @xmath202 ) and each contribution to the high @xmath1 peak @xmath192 ) during such a passage ( total number @xmath203 ) .
we plot the fraction @xmath204 .
, title="fig:",width=302 ] + in our discussion of figure [ hist ] , we already noted that around 10 per cent of stars which contribute to the high @xmath1 peak at a given time and sightline , need 1 gyr or longer to achieve this again .
the @xmath196 histograms have also shown that @xmath7725 per cent of stars that initially had return times @xmath205 myr need more than 500 myr to return again .
so orbit populations slowly change in a way that reduces the proportion of their time during which they contribute to the high @xmath1 peak . to understand why the high @xmath1 peak vanishes with age , we conducted the following exercise : we tracked all orbits from the histogram in panel a of figure [ hist ] .
as the orbits that never return or have very long return times ( defined as @xmath206 0.8 gyr ) are likely random contributors , we exclude them .
we record each passage of the sightline ( total number @xmath202 ) and if during this passage the orbit contributes to the high @xmath1 peak ( total number @xmath203 ) .
the fraction @xmath207 gives a measure of how much the population of orbits is contributing to the peak at time @xmath208 .
a priori one would expect a ratio @xmath209 for @xmath7 orbits as they are approximately axisymmetric along the bar passing once through the sightline while approaching and once while moving away and contributing to the high @xmath1 peak .
consequently , smaller values for @xmath210 are expected for more complex orbits .
this is slightly complicated by the geometry and the slower absolute velocity of stars while approaching : the front side of the bar is geometrically disfavoured , as the pencil beam covers a larger area at the rear side . on the other hand ,
the approaching stars are closer to the tip and hence slower , taking more weight in the observations .
moreover , it is also possible that stars moving along the rear side have high enough @xmath1 only for a fraction of the time they spend in the sightline .
if we select @xmath7 orbits which at every second crossing of the sightline contribute to the high @xmath1 peak , we empirically find a mean value of @xmath211 , close to the simple estimate .
+ from the return time analysis shown above , we would expect @xmath212 , as most orbits exhibit a more complex behaviour .
@xmath213 is shown in figure [ frac ] .
initially , @xmath210 fluctuates periodically .
this can be understood as most orbits return to the high @xmath1 peak after specific characteristic times ( see figure [ hist ] ) and they have been selected at correlated phases . the average value of @xmath210 over the covered 2 gyr drops from @xmath214 to @xmath215 .
this shows that the fraction of orbits contributing to the high @xmath1 peak continuously decreases .
most of this decrease is caused by diffusion of stars away from the contributing orbits , but it is not clear , if our simulation handles this diffusion correctly .
sources for this diffusion are chaotic orbit behaviour , two - body noise and fluctuations in bar parameters .
our resolution is likely not sufficient to cover chaotic orbital behaviour correctly .
for the two - body noise , it is unknown if the level of @xmath0-body noise is realistic compared to the mw ( bearing e.g. giant molecular clouds near the bar tips , and dark matter substructure ) . in addition , our simulation will not have sources for bar fluctuations like tidal torques exerted by satellites .
the important point to take away is that the vanishing of the high @xmath1 peak in older stars is not only due to the broadening of the main peak , but also due to active loss of stars from high @xmath1 orbits . by design , new - born star particles in our simulation
are released on near circular orbits , where the circular speed was determined to balance the local radial component of the gravitational force .
obviously , the young stars contributing to the high @xmath1 peak discussed above are not on circular orbits .
just as obviously , there are no stable circular orbits near the bar .
this is why we set an inner radius limit for new born stars , so that no new star particles are inserted within the main bar region .
this cutoff is , of course , to some degree arbitrary and young stars born close to the bar can be captured by the bar , resulting in the young bar populations studied in the previous sections .
bars in real galaxies can also grow by capturing stars and thus changing their orbits .
so the young stars that end up in the high @xmath1 peak are predominantly the result of continuous bar growth . in figure
[ capture ] , we show six stars that at 8.3 gyr contribute to the high @xmath216 feature and were born within 1 gyr before that time .
we find that the shortest time it takes to appear in the high @xmath1 peak is similar to the shortest return time peak , 110 myr .
the capture times are rather equally distributed between this lower limit and the maximum 1 gyr due to the selection criteria .
the shown star particles were selected at different _ capture time _ intervals .
they start out on non - bar orbits and at some point transition to @xmath7-like or pretzel - like orbits .
clearly stars that end up in the high @xmath1 features are born at radii that correspond to the tip of the bar , so they belong to the innermost inserted particles . to visualise this point , we plot in figure [ radbirth ] the distribution of birth radii @xmath217 of high @xmath1 stars . for this figure we select stars which were born after 7.3 gyr and contribute to the high @xmath1 feature at @xmath158 degrees at 8.3 - 8.5 gyr and/or at 9.3 - 9.5 gyr .
the union of the samples found at these two time intervals comprises @xmath218 stars , for which we determine birth radii .
the histogram for all stars is shown in blue , the one for the subsample of stars born after 8.5 gyr is shown in red .
we notice that the majority of stars is born just outside the bar at @xmath219 .
as seen in figure [ bar - evo ] the bar radius fluctuates mildly , which allows smaller birth radii .
the subsample of stars which were born at later times has its lower cutoff slightly further out due to the continuing growth of the bar .
the outer tail is wider for older stars , indicating that stars which are older at capture can have larger birth radii .
this is because they had more time to migrate radially inwards .
window at @xmath158 degrees .
the union of the samples found at these two time intervals comprises @xmath218 stars .
the distribution of birth radii @xmath217 of all @xmath218 stars is shown in blue .
the histogram of stars born after 8.5 gyr and contributing at 9.3 - 9.5 gyr is shown in red .
the vertical dashed lines mark the positions of the inner 4:1 ( ultra - harmonic ) resonance at 7.3 and 9.3 gyr .
, title="fig:",width=302 ] + resonances can play a role in capturing orbits ( e.g. @xcite ) .
the corotation resonance of our bar at the analysed time intervals is at @xmath220 , but no stars from these regions are captured .
we find that the ultra - harmonic resonance ( uhr ) , @xmath221 , in our model lies just outside the bar .
this is likely not coincidental , as the uhr can limit the radial extent of stable @xmath7 orbits and thus the size of the bar @xcite .
the resonance may play a role in shaping the spatial distribution from which stars are captured by the bar , as while the bar grows , the resonance moves outwards , and stars on orbits just outside the bar get scattered , absorbed or released by it .
these processes can transport orbits to the regions of phase space occupied by high velocity bar orbits .
however , the exact role of the resonance in the capture process is unknown , as the bar is strong , evolving and the potential is not smooth .
the capture of orbits can be understood in simpler terms ( similar to @xcite ) , in that stars come in at the right angle to the bar to lose angular momentum and get trapped .
a detailed study of capture processes is beyond the scope of this work .
capture of stars onto these orbits along the bar is mostly drawn from kinematically cool populations that are close to the galactic plane , which have a small random energy compared to the potential disturbance and are at low altitude , where the disturbance is strongest .
while these stars drastically change the in - plane geometry of their orbits during capture by the bar , their vertical energy and hence scale height remain small .
consistently , both the mw and our simulation display a discernible high @xmath1 feature only at observed latitudes @xmath222 degrees .
above , we have learned that for populations of stars which at a given time contribute to the high @xmath1 feature , the contribution fraction slowly declines with time .
as the simulation evolves , the high @xmath1 feature , however , remains detectable .
this is because of ongoing star formation close to the bar region and the continuous capture of young stars on @xmath7-like or pretzel - like orbits .
the young populations on such orbits then slowly evolve towards more complex orbit populations with a decreasing fraction of high @xmath1 stars .
we note that another process which could produce young stars on high @xmath1 orbits is not included in our model : in real galaxies , molecular clouds can exist on bar orbits @xcite .
stars born in these clouds would make up kinematically cool populations born directly on high velocity bar orbits and could thus contribute to the high @xmath1 features .
in this paper , we have studied how an @xmath0-body+sph model of a growing milky way ( mw ) like barred disc galaxy compares to the mw in terms of line of sight velocities @xmath1 of stars seen from the earth and in the direction of the bar .
the model uses an approach that has not yet been used in this context .
we start with a light ( one tenth of the final mass ) and short scale - length disc of gas and stars in equilibrium within a live dark halo . this disc is grown by continuously adding star and gas particles on near circular orbits in a controlled way .
the mass and size growth histories are matched to expectations for the mw , and indeed at simulation time 7 - 10 gyr the simulated bar is similar in structure to the mw bar .
our simple prescription for the evolution of the galaxy in our @xmath0-body simulation lowers the computational cost and also circumvents the uncertainties in the parametrisation of sub - grid model approaches to star formation and stellar feedback .
the general dynamics of gas and stars are expected to be reasonably represented , but we do not cover varying gas fractions , supernova induced turbulence and effects of multiple gas phases
. the simulation does not produce any sensible interplay between gas and stars : there is e.g. no enhanced star formation at places with high gas densities .
external galaxies show indications for star formation associated with the gas flows along the edges of a bar ( the equivalent of the mw s @xmath223 arms ; * ? ? ?
* ) , which depend strongly on the dynamical state of the bar ( see e.g. * ? ? ?
* ) . however , these processes are not covered in our simulation . on the positive side
, our approach allows us to set a reasonable star formation history , to keep spurious sources of heating under control , and hence to observe the dynamical effects of bar formation with less scatter .
we use this simulation to interpret line of sight velocity ( @xmath1 ) data for the central and inner disc regions of the mw from the apogee survey . to enable a robust comparison
, we apply a simple population synthesis to model the apogee selection function . apart from minor discrepancies at a minority of sightlines , the model fits the data surprisingly well without any scaling or adjustment .
the simulation not only reproduces the positions and dispersions of the main peaks in the @xmath1 distributions at each sightline , but also displays the high @xmath1 features found by @xcite within the central bar region .
our simulation clearly identifies these high @xmath1 features with stars travelling near the plane alongside the bar . by showing that these are kinematically cool features with preferentially young stars in the plane ,
we naturally explain why the high @xmath1 structure was seen by @xcite and @xcite in observation fields at @xmath224 , but is absent in the higher - latitude fields studied by @xcite . by the strong preference of the apogee selection function for young stars ,
the high @xmath1 features get additionally enhanced and have comparable strength and shape in our simulation and the apogee observations .
we also determine at which velocities low @xmath1 peaks at negative @xmath3 , produced by stars moving towards the sun along the front side of the bar , should be visible .
so far , only hints of these features have been seen by @xcite .
we show that , unlike for the high @xmath167 features , the position and width of the main peaks of the @xmath1 distributions are dominated by fore- and background stars , which are not associated with the bar .
we also study a variation of our parameters , in particular the bar angle @xmath24 .
other than @xcite we feel unable to put strong constraints on the observed bar angle : while large @xmath225 are clearly disfavoured as the majority of high velocity stars get a too large projection angle with the line of sight , no reliable constraint can be found for smaller values of @xmath24 , since change in velocities is degenerate with our general uncertainty about the bar length , pattern speed and strength . additionally , we show that the high @xmath1 feature is detectable at simulation times between 7 and 10 gyr and is thus not connected to specific events
. the feature changes mildly over this period due to a growth of the bar in length and mass , but the change is too small to allow any conclusions about the properties of the mw bar .
we note that on two sightlines , namely @xmath226 and @xmath227 degrees , the simulation shows a mild difference to the observations , missing the tentative high @xmath1 peak at @xmath226 and predicting the main peak of the velocity distribution at too high velocities at both sightlines .
possible origins for the discrepancies are a longer bar length for the mw or the @xmath228 molecular ring in the mw .
a decisive answer on this question would require detailed analytical models that can cover the full space of possible parameters .
a detailed analysis of the stars participating in the high @xmath1 features in our model reveals that ( contrary to the original view of * ? ? ?
* ) only about half of the contribution is made by stars on @xmath7 orbits , with additional large contributions from pretzel type @xcite and more complicated orbits .
this finding results from a detailed study of these orbits guided by return times into the high velocity peak at the sightline @xmath229 degrees .
the return times display several peaks that are associated with the different orbit families ( and some scatter by stars missing the observation window ) and help with the quantification of the contribution of different orbit families .
we note that , in response to our work , @xcite have updated their paper and now also conclude that higher order orbit families ( they name 3:1 and 5:2 ) can make significant contributions to the high @xmath1 features .
another interesting statistic is the ratio of the number stars in the high @xmath1 region of a certain sightline divided by all stars passing through the sightline at arbitrary @xmath1 values for all stars that have been observed once in the high velocity peak . while we find that @xmath7 orbits with stable short return times show an average ratio of @xmath211 , the stars in the simulation start with a ratio of @xmath230 , as a consequence of the significant contribution of more complex orbits to the velocity peak . within @xmath231 gyr of the initial detection , this ratio drops steadily down to @xmath232 .
due to changes in the bar potential , length , and pattern speed , as well as orbital diffusion , stars diffuse out of these orbit families weakening the high @xmath1 peak . however , the high @xmath1 feature in our simulation persists , as the growing bar captures new stars from the surrounding disc .
it is unclear , how realistically our model represents the precise balance of orbit capture and diffusion , which is determined by chaotic diffusion , two - body interactions and fluctuations in bar parameters .
in addition , observations of enhanced star formation at the bar tips and on bar orbits along the leading bar edges ( see above ) , imply a source of young stars on high velocity bar orbits , which is not mirrored by our model .
overall , our result provides an indication for the recent growth of the galactic bar .
quantitative constraints on the bar dynamics and history would demand further exploration of model parameters and also a detailed age and distance tomography of the observed high @xmath1 stars .
we note that our bar is a slow bar with @xmath130 in a disc with a high dark matter contribution to the circular velocity curve .
as discussed in section 4 , observations
do not constrain the pattern speed and the length of the mw bar well and we currently can exclude neither a slow bar as ours nor a fast bar with @xmath233 .
as far as the dark matter fraction is concerned , our disc agrees with recent estimates on the local circular velocity in the mw and the corresponding dm contribution .
further work is necessary to understand how faster bars and/or lower dark matter fractions would alter the processes discussed in this paper .
@xcite point out that a nuclear stellar disc with a radial extent of @xmath234 could explain high velocity features at @xmath235 degrees .
although this idea is interesting , they do not currently have a model which can explain the velocity distributions at the full apogee range of bar longitudes , as our model can .
the simulations of @xcite advocate against such a giant @xmath236 disc for the current mw centre , but the extent of such a disc may have varied over time . to safely determine whether this idea can explain the data , a model , which self - consistently predicts velocity distributions at longitudes @xmath237 degrees would be needed
. a higher number of observed stars in this region with reasonable measurements of distances , ages and metallicities would certainly help to ( in)validate our model or other attempts to explain the observed velocity distributions .
as far as the nuclear stellar disc in the mw is concerned , its existence so far has only reliably been detected in the central @xmath238 @xcite and the short scalelength inferred for this structure currently seems to disfavour a large nuclear disc as proposed by @xcite . to summarise , we have shown that the high @xmath1 features seen towards the bar in apogee are likely due to preferentially young stars on orbits moving along the major axis of the bar .
they move close to the plane , which is why the feature disappears at high latitudes .
the most important orbit family contributing to the feature are @xmath7 orbits , but a variety of more complex orbits also make significant contributions . despite its simplicity , our model is the first shown to match the detailed kinematics in the bar region , and hence its late stages can serve as a blueprint for further and more detailed studies .
it is a pleasure to thank james binney for helpful discussions and comments on the manuscript .
we are grateful to mattia sormani and evgeny vasilyev for helpful comments .
the simulations presented here were run on the dirac facility jointly funded by stfc , the large facilities capital fund of bis and the university of oxford .
this work was supported by the uk science and technology facilities council through grant st / k00106x/1 and by the european research council under the european union s seventh framework programme ( fp7/2007 - 2013)/erc grant agreement no . 321067 .
ahn c. p. et al . , 2014 ,
apjs , 211 , 17 allende prieto c. et al . , 2008 , an , 329 , 1018 antoja t. et al . , 2014 ,
a&a , 563a , 60 aumer m. , binney j. j. , 2009 , mnras , 397 , 1286 babusiaux c. et al . , 2014 ,
a&a , 563a , 15 bedin r.l . , cassisi s. , piotto g. , castelli f. , salaris m. , momany y. , pietrinferni a. , 2005 , mnras , 357 , 1038 bensby t. et al . , 2013 , a&a , 549a , 147 berentzen i. , shlosman i. , martinez - valpuesta i. , heller c. h. , 2007 , apj , 666 , 189 berrier j. c. , sellwood j. a. , 2015 , apj , 799 , 213 binney j. , gerhard o. e. , stark a. a. , bally j. , uchida k. i. , 1991 , mnras , 252 , 210 binney j. , tremaine s. , 2008 , galactic dynamics : second edition , princeton university press , princeton , nj usa bland - hawthorn j. , cohen m. , 2003 , apj , 582 , 246 blitz l. , spergel d. n. , 1991 , apj , 379 , 631 bovy j. , rix h .- w . , liu c. , hogg d. w. , beers t. c. , lee y. s. , 2012 , apj , 753 , 148 contopoulos g. , papayannopoulos t. , 1980 , a&a , 92 , 33 dame t.m .
et al . , 1987 , apj , 322 , 706 debattista v. p. , sellwood j. a. , 2000 , apj , 543 , 704 debattista v. p. , ness m. , earp s. w. f. , cole d. r. , 2015 , arxiv e - prints , arxiv:1507.01433 dehnen w. , 2000 , aj , 119 , 800 dubinski j. , berentzen i. , shlosman i. , 2009 , apj , 697 , 293 eisenstein d.j .
et al . , 2011 ,
aj , 142 , 72 fraternali f , binney j. j. , 2008 , mnras , 386 , 935 friedli d. , benz w. , 1995 , a&a , 301 , 649 hill v. et al . , 2011 ,
a&a , 534a , 80 li z .- y .
, shen j. , rich r. m. , kunder a. , mao s. , 2014 , apj , 785l , 17 lynden - bell d. , 1979 , mnras , 187 , 101 majewski s. r. et al . , 2015 , arxiv e - prints , arxiv:1509.05420 mcmillan p. j. , 2011 , mnras , 414 , 2446 molloy m. , smith m. c. , evans n. , shen j. , 2015 , arxiv e - prints , arxiv:1505.04245 nelson d. , vogelsberger m. , genel s. , sijacki d. , kere d. , springel v. , hernquist l. , 2013 , mnras , 429 , 3353 nidever d. l. et al .
, 2012 , apj , 755l , 25 patsis p. a. , athanassoula e. , quillen a. c. , 1997 , apj , 483 , 731 pietrinferni a. , cassisi s. , salaris m. , castelli f. , 2004 , apj , 612 , 168 pietrinferni a. , cassisi s. , salaris m. , percival s. , ferguson j.w . , 2009 ,
apj , 697 , 275 piffl t. et al . , 2014 , mnras , 445 , 3133 portail m. , wegg c. , gerhard o. , martinez - valpuesta i. , 2015 , mnras , 448 , 713 portail m. , wegg c. , gerhard o. , 2015 , mnras , 450l , 66 rokar r. , debattista v. p. , quinn t. r. , wadsley j. , 2012 , mnras , 426 , 2089 schnrich r. , binney j. , 2009 , mnras , 396 , 203 schnrich r. , 2012 , mnras , 427 , 274 schnrich r. , aumer m. , sale s. e. , 2015 , arxiv e - prints , arxiv:1507.02695 sellwood j. a. , carlberg r. g. , 1984 , apj , 282 , 61 sheth k. , vogel s.n . , regan m.w . , teuben p.j . ,
harris a.i .
, thornley m.d . , 2002 ,
aj , 124 , 2581 sormani m.c . , binney j. , magorrian , j. , 2015 , arxiv e - prints , arxiv:1507.03078 springel v. , 2005 , mnras , 364 , 1105 voglis n. , harsoula m. , contopoulos g. , 2007 , mnras , 381 , 757 wegg c. , gerhard o. , 2013 , mnras , 435 , 1874 wegg c. , gerhard o. , portail m. , 2015 , mnras , 450 , 4050 yurin d. , springel v. , 2014 , mnras , 444 , 62 zasowski g. , et al . , 2013 , aj , 146 , 81 zoccali m. , 2014 , a&a , 562a , 66 | we analyse a controlled @xmath0-body + smoothed particle hydrodynamics simulation of a growing disc galaxy within a non - growing , live dark halo .
the disc is continuously fed with gas and star particles on near - circular orbits and develops a bar comparable in size to the one of the milky way ( mw ) .
we extract line of sight velocity @xmath1 distributions from the model and compare it to data recently obtained from the apogee survey which show distinct high velocity features around @xmath2 . with an apogee like selection function , but without any scaling nor adjustment , we find @xmath1 distributions very similar to those in apogee . the stars that make up the high @xmath1 features at positive longitudes @xmath3 are preferentially young bar stars ( age @xmath4 ) which move away from us along the rear side of the bar . at negative @xmath3 , we find the corresponding low @xmath1 feature from stars moving towards us . at @xmath5 degrees
the highest @xmath1 stars are a mixture of bar and background disc stars which complicates the interpretation of observations .
the main peak in @xmath1 is dominated by fore- and background stars . at a given time , @xmath6 per cent of high @xmath1 stars occupy @xmath7 like orbits , but a significant fraction are on more complex orbits .
the observed feature is likely due to a population of dynamically cool , young stars formed from gas just outside the bar and subsequently captured by the growing bar .
the high @xmath1 features disappear at high latitudes @xmath8 degrees which explains the non - detection of such features in other surveys .
[ firstpage ] galaxy : bulge ; methods : numerical - galaxy : evolution - galaxy : kinematics and dynamics - galaxy : structure ; |
single - transverse spin asymmetries ( ssa ) in hadronic processes have a long history @xcite .
recent experimental measurements of ssas in polarized semi - inclusive lepton - nucleon deep inelastic scattering ( sidis ) @xcite , in hadronic collisions @xcite , and in the relevant @xmath0 annihilation process @xcite , have renewed the theoretical interest in ssas and in understanding their roles in hadron structure and quantum chromodynamics ( qcd ) .
there are several approaches to understanding ssas within the qcd framework @xcite .
transverse - momentum - dependent ( tmd ) parton distributions and fragmentation functions , and their relevance for semi - inclusive dis , the drell - yan process , di - hadron production in @xmath0 annihilations , and the single - inclusive hadron production at hadron colliders have been investigated .
two important contributions from these tmd parton distributions and fragmentation functions have been mostly discussed in the last few years : the sivers quark distribution and the collins fragmentation function .
the sivers quark distribution @xcite represents a distribution of unpolarized quarks in a transversely polarized nucleon , through a correlation between the quark s transverse momentum and the nucleon polarization vector .
the existence of the sivers function requires final / initial - state interactions @xcite , and an interference between different helicity fock states of the nucleon .
the collins function represents a correlation between the transverse spin of the fragmenting quark and the transverse momentum of the hadron relative to the `` jet axis '' in the fragmentation process . like the sivers function
, it vanishes when integrated over all transverse momentum .
one of the most nontrivial properties associated with the sivers and collins functions are their universality properties .
although they both belong to the so - called naive - time - reversal - odd " functions , they do have different universality properties .
for the quark sivers function , because of the initial / final state interaction difference , they differ by signs for the sidis and drell - yan processes .
this non - universality has also been extended to other processes , such as the dijet - correlation in hadronic reactions , where it was found that both initial and final state interactions contribute to the ssa , and there exists non - trivial relation between this and those in the sidis and drell - yan processes @xcite , and a standard tmd factorization breaks down @xcite . on the other hand ,
there have been several studies showing that the collins function is universal between different processes , primarily in the sidis and @xmath0 annihilation @xcite . in these discussions , the gauge links in the fragmentation functions
do not play a crucial role to leading to a nonzero collins function though they are important to retain the gauge invariance , whereas it has been well understood that the gauge links in the parton distributions play very important roles to obtain non - zero quark sivers function . the collins effect in the fragmentation process and its universality has been recently extended to the hadron production in @xmath1 collisions @xcite , where the azimuthal distribution of hadrons inside a high energy jet can probe the collins fragmentation function and the quark transversity distribution @xcite in the single transverse polarized nucleon - nucleon scattering . in this paper
, we will give the detailed derivation of these results , and argue that the universality is in general and model - independent .
we are interested in the hadron production from the fragmentation of a transversely polarized quark which inherit transverse spin from the incident nucleon through transverse spin transfer in the hard partonic scattering processes .
as shown in fig . 1 , we study the process , @xmath2 where a transversely polarized proton with momentum @xmath3 scatters on another proton with momentum @xmath4 , and produces a jet with momentum @xmath5 ( transverse momentum @xmath6 and rapidity @xmath7 in the lab frame ) .
the three momenta of @xmath3 , @xmath4 and @xmath5 form the so - called reaction plane . inside the produced jet ,
the hadrons are distributed around the jet axes .
a particular hadron @xmath8 will carry certain longitudinal momentum fraction @xmath9 of the jet , and its transverse momentum @xmath10 relative to the jet axis will define an azimuthal angle with the reaction plane : @xmath11 , shown in fig . 1 .
thus , the hadron s momentum is defined as @xmath12 .
the relative transverse momentum @xmath10 is orthogonal to the jet s momentum @xmath5 : @xmath13 .
similarly , we can define the azimuthal angle of the transverse polarization vector of the incident polarized proton : @xmath14 .
scattering.__,width=340 ] -0.4 cm the leading order contribution to the jet production in @xmath1 collision comes from @xmath15 sub - processes , where two jets are produced back - to - back in the transverse plane . for the reaction process of ( 1 ) , one of the two jets shall fragment into the final observed hadron . in this paper , we study the physics in the kinematic region of @xmath16 .
the unpolarized cross section contribution from the partonic @xmath15 process @xmath17 where the final state quark @xmath18 fragments into final observed hadron @xmath8 , can be written as @xmath19 where @xmath20 represents the phase space for this process , @xmath7 and @xmath21 are rapidities for the jet @xmath5 and the balancing jet , respectively , @xmath6 is the jet transverse momentum , and the final observed hadron s kinematic variables @xmath9 and @xmath10 are defined above . here ,
@xmath22 and @xmath23 are the momentum fractions carried by the parton @xmath24 " and@xmath25 " from the incident hadrons , respectively . in the above equation , @xmath26 and @xmath27
are the associated parton distributions , and @xmath28 is the tmd quark fragmentation function .
the hard factors @xmath29 are equal to the partonic differential cross section for the relevant subprocess : @xmath30 .
similarly , the differential cross section for the transverse - spin dependent scattering process can be written as @xmath31\times h_{qb\to qb}^{\rm collins}\ , \end{aligned}\ ] ] @xmath32 with convention @xmath33 , and @xmath34 is the hard factor for the partonic channel @xmath35 . here ,
@xmath36 ( also noted by @xmath37 , @xmath38 and @xmath39 in the literature ) is the quark transversity distribution , and @xmath40 the collins fragmentation function @xcite ( also noted as @xmath41 or @xmath42 in the literature ) . it was argued that the collins function is universal between the above process and other processes such as @xmath0 annihilation and sidis @xcite .
as an example , we will demonstrate this universality for the particular partonic channel @xmath43 contribution to our process , and all other channels will follow accordingly . for convenience , we list the hard factors for this channel , @xmath44 for the unpolarized and single - transverse - spin polarized cross sections , respectively . here
@xmath45 , @xmath46 , and @xmath47 are the usual partonic mandelstam variables .
the rest of this paper is organized as follows . in sec.ii
, we calculate the differential cross sections for the unpolarized and single - spin dependent scattering processes from @xmath43 channel contributions , and demonstrate the universality of the collins function .
especially , we will present a detailed calculation for one - gluon exchange diagrams which are essential for the universality argument .
an extension to two - gluon exchange diagrams is presented in sec.iii .
we summarize our paper in sec.iv .
we follow the model used in ref . @xcite to calculate the quark fragmentation into a pion . as shown in fig .
2(a ) , a quark ( with momentum @xmath48 ) fragments into a pion ( with momentum @xmath49 ) by the vertex from a model described in @xcite
. a simple calculation will give the unpolarized quark fragmentation function @xcite , @xmath50 where @xmath51 is the coupling between the quark and pion , @xmath52 is the quark mass .
scattering in a model described in @xcite ( b)._,width=340 ] -0.4 cm we can also use this model to calculate pion production in hadronic process of ( 1 ) . in fig .
2(b ) , we show the feynman diagram for the typical partonic channel @xmath43 contribution , where the initial quarks have momenta @xmath3 and @xmath4 , respectively . in the final state
, the produced pion has momentum @xmath49 , and the associated final state quark has momentum @xmath53 , whereas the balancing jet has momentum @xmath54 .
we further introduce a light - like momentum @xmath55 : @xmath56 , which represents the dominant component of the fragmenting quark s momentum .
it can be parameterized as follows , @xmath57 where @xmath6 is the transverse momentum for the fragmenting quark in the lab frame , @xmath45 , @xmath46 and @xmath47 as mentioned above , are the usual partonic madelstam variables for this partonic process : @xmath58 , @xmath59 , and @xmath60 . in our discussions ,
the jet s transverse momentum @xmath6 ( in the lab frame ) is the large momentum scale at the same order as @xmath45 , @xmath46 and @xmath47 .
of course , the full momentum of the fragmenting quark @xmath61 is off - shell in this diagram .
however , its off - shellness is much smaller than @xmath6 . in order to formulate the final state hadron s momentum ,
we introduce a conjugate light - like vector @xmath62 : @xmath63 and @xmath64 .
it is convenient to define this momentum in the center of mass frame of the two incident momenta @xmath3 and @xmath4 . in this frame
, we have @xmath65 which happens to be the momentum of the balancing jet .
from above , we have @xmath66 and @xmath67 . in the following calculations
, we will work in this particular frame .
we emphasize that our results do not depend on the frame . with the above two momenta
, we can formula the final state pion s momentum as @xmath68 where @xmath69 is the momentum fraction of the fragmenting quark carried by the pion in the final state , @xmath10 is the transverse momentum relative to the fragmenting quark momentum @xmath55 : @xmath70 and @xmath71 . in the above parameterization , we have neglect the pion mass , which is not relevant in our calculations .
similarly , we can formulate the associated final state quark momentum @xmath53 as , @xmath72 where we have kept the quark mass , because it will be relevant for the nonzero single spin asymmetry discussed below . with the above decompositions for the relevant momenta
, it is straightforward to calculate the feynman diagrams for this process in fig .
2(b ) . in the calculations
, we will utilize the power counting method to keep the leading order contributions , and neglect all higher order corrections of @xmath73 or @xmath74 . by doing that
, we can separate the short distance physics ( at the scale of @xmath6 ) from the long distance physics ( at the scale of @xmath10 and @xmath52 ) . finally , the cross section contribution from fig .
2(b ) will be , @xmath75 in the limit of @xmath16 .
this result is indeed factorized into the hard factor @xmath76 in eq .
( [ e8 ] ) times the fragmentation function in eq .
( [ unf ] ) calculated from fig .
2(a ) .
( a ) , deep inelastic scattering ( b ) , and @xmath1 scattering ( c ) , when we have dressed quark propagator associated with the fragmenting quark in these processes .
the universal collins function can be calculated from the diagram in ( d ) .
the blobs in the diagrams represent the dressed quark propagator in this model .
_ , width=340 ] -0.4 cm now , we turn to discuss the ssa in this process .
we need to generate a phase from the scattering amplitudes to have a non - vanishing ssa .
as suggested in @xcite , the dressed quark propagator in this model may contribute to such a phase .
similarly , the vertex correction to the quark - pion vertex can also contribute a phase .
if the phase comes from the above sources , it is easy to argue the universality of the collins function between our process and the sidis/@xmath0 process , because they are the same .
for example , as we show in fig . 3 , the dressed quark propagator associated with the fragmenting quark can contribute to a nonzero phase @xcite , which will contribute the same to the collins function in all these three processes : @xmath0 annihilation , sidis , and hadron production in @xmath1 scattering .
this propagator can be parameterized as : @xmath77 @xcite , where @xmath78 and @xmath79 are complex numbers . following the above calculations for the unpolarized cross section
, we will find the single - spin dependent cross section for process ( 1 ) from fig .
3(c ) can be written as @xmath80\frac{\alpha_s^2\pi}{\hat s^2}\frac{n_c^2 - 1}{4n_c^2}\frac{4\hat s\hat u}{-\hat t^2 } \frac{g^2}{16\pi^2}\frac{z_h^2(1-z_h)2m{\rm im}(a^*b)}{(p_{ht}^2+z_h^2m^2)^2 } \ , \ ] ] where again we only keep the leading order contribution in the limit of @xmath16 and @xmath81 . in the derivation of the above result
, the following identity has been used to simplify the final expression , @xmath82 = 0\ , \ ] ] which holds in our working frame .
the above differential cross section can be factorized into the collins function calculated from the dressed quark propagator from fig .
3(d ) @xcite and the hard factor @xmath83 from eq .
( [ e8 ] ) in this partonic channel @xmath43 , and this collins function will be the same as that in @xmath0 and sidis processes in fig .
3(a ) and 3(b ) .
similarly , the vertex corrections contributions to the collins function can be analyzed accordingly , and the same factorization and universality of the collins function will follow . collisions .
the short bars indicate the pole contributions to the phase needed for a non - vanishing ssa .
the additional two cuts in ( d ) cancel out each other.__,width=377 ] -0.4 cm the main issue of the universality discussion concerns the extra gluon exchange contribution between the spectator and hard partonic part @xcite .
for example , in our case , because the hadron is colorless while the quark is colored , the remanet in the fragmentation process will be also colored .
thus the gluon exchanges between the remanet and the other parts of the scattering amplitudes become essential . in fig . 4 , we have shown all these interactions , including the gluon attachments to the incident quarks ( a , c ) , and final state balancing quark ( d ) and the internal gluon propagator ( b ) .
these diagrams are much more complicated than those discussed in @xcite for sidis and @xmath0 processes , where there is only one diagram contribution in both cases .
therefore , the universality argument for the collins function is not straightforward .
however , the dominant contribution to the fragmentation function comes from the kinematic region where the exchanged gluon is parallel to the final state hadron @xcite .
otherwise , their contributions will be power suppressed in the limit of @xmath16 or belong to a soft factor . for these collinear gluon interactions
, we can use eikonal approximation and ward identity to sum them together to form the gauge link in the definition of the fragmentation function @xcite .
meanwhile , we notice that the contributing phases of the diagrams in fig . 4
come from the cuts through the internal propagators in the partonic scattering amplitudes . in fig . 4 , we labeled these cut - poles by short bars in the diagrams . from our calculations
, we find that all these poles come from a cut through the exchanged gluon and the fragmenting quark in each diagram , and all other contributions either vanish in the leading order contribution or cancel out each other .
for example , in fig .
2(d ) , we show two additional cuts , which contribute however opposite to each other and cancel out completely . to see this cancellation more clearly
, we can write down the momentum integral of the exchange - gluon , @xmath84 where @xmath85 represents the denominators coming from the scattering amplitude . by power counting analysis ,
the dominant contribution to the fragmentation comes from the kinematic region of @xmath18 being parallel to the final state hadron s momentum @xmath86 . from this fact
, we can parameterize @xmath18 in terms of @xmath87 and @xmath62 , and define @xmath88 and @xmath89 .
thus the integral of momentum @xmath18 becomes @xmath90 , where @xmath91 is the transverse momentum relative to the jet momentum @xmath55 . because @xmath18 is parallel to @xmath87
, @xmath92 will be order 1 , whereas @xmath93 will be order of @xmath94 .
when we perform the integrals of @xmath95 , we need to take two poles from the above propagators to obtain a nonzero collins asymmetry .
these poles will form a cut through the fenyman diagram . physically
, these cuts represent the kinematic allowed final state re - scattering in the diagram . by examining the behaviors of the propagators in the above kinematic region ,
we further notice that the @xmath96channel gluon propagator @xmath97 does not contribute to a pole .
this is because this propagator is far off - shell : @xmath98 .
if we take a pole from this propagator , we have to constrain the momentum of @xmath18 being proportional to @xmath54 and @xmath4 , whose contribution will be power suppressed .
thus , we shall calculate the pole contributions from other propagators . in fig . 4d
, we show three possible cuts which are kinematic allowed for this diagram .
two of them are associated with the propagator @xmath99 .
this propagator involves large momentum @xmath54 , and can be simplified by using the eikonal approximation , @xmath100 the pole contribution from this propagator is proportional to @xmath101 . with this delta function , the integral over @xmath93 vanishes , because the rest poles are in the same half plane of @xmath93 , @xmath102 where we have used the fact that @xmath103 and @xmath104 .
this means that the two cuts associated with the propagator @xmath99 cancel out each other .
the above result depends on the momentum flow of @xmath18 in this diagram and the time - like process in the fragmentation region requiring that @xmath105 and @xmath106 .
therefore , the only contribution to the nonzero ssa associated with the collins effect comes from the cut going through the fragmenting quark and the exchange - gluon , as we labeled by short bars in this diagram . summarizing the above analysis
, we find that the contribution from this diagram can be written as @xmath107 where we have made the eikonal approximation for the propagators @xmath99 and @xmath97 .
similar analysis can be done for all other diagrams , and we find that their contributions come from the same poles of the fragmenting quark and the exchange - gluon .
therefore , their contributions will have the similar expression as eq .
( [ eikonal ] ) with the same delta functions in the integral : @xmath108 and the propagator @xmath109 .
thus the contributions from all these diagrams can be summed together . in this sum , we notice that the different diagrams have different color - factors , @xmath110\times { \rm tr}[t^at^b]=c_f\times \frac{n_c^2 - 1}{4n_c^2}+\frac{1}{n_c^2}\frac{if_{abc}}{2}{\rm tr}[t^at^bt^c ] \ , \nonumber\\ { \rm 4(b ) } : & & \frac{1}{n_c^2}{\rm tr}[t^at^ct^b]\times { \rm tr}[t^at^d]if_{dbc}=-\frac{1}{n_c^2}\frac{if_{abc}}{2}{\rm tr}[t^at^bt^c ] \ , \nonumber\\ { \rm 4(c ) } : & & \frac{1}{n_c^2}{\rm tr}[t^at^ct^b]\times { \rm tr}[t^at^bt^c ] \ , \nonumber\\ { \rm 4(d ) } : & & \frac{1}{n_c^2}{\rm tr}[t^at^ct^b]\times { \rm tr}[t^at^ct^b ] \
.\end{aligned}\ ] ] we further find that the contributions ( without the color - factors ) from the diagrams ( c ) and ( d ) are opposite to each other .
thus , their total contribution will be the difference on the color - factor , which is @xmath111 $ ] .
that means the contributions from all these four diagrams can be grouped into two terms with different color factors : one with @xmath112 , and one with @xmath113 $ ] .
the latter one vanishes in the leading order of @xmath73 after we sum all diagrams contributions , and thus we are left with the first color - factor contribution . after summing over all diagrams contribution ,
the spin - dependent differential cross section coming from the collins effect will be @xmath114 \frac{g^2}{(2\pi)^3}c_fg_s^2\int\frac{dq^+dq^-d^2q_t}{(2\pi)^4 } \nonumber\\ & & \times \left(q^+p_{ht}^{\beta'}-z_hq_t^{\beta'}\right ) \frac{1}{(k'-q)^2}\delta(q^2 ) \delta\left((p_h+k'-q)^2-m^2\right ) \ . \label{xs}\end{aligned}\ ] ] where @xmath115 is the strong coupling . from the above result
, we find a clear separation of the short distance physics at the scale @xmath6 and long distance physics at the scale @xmath10 .
the short distance part is just the hard factor @xmath83 for the spin - dependent cross section , which can be calculated from the partonic process with both initial and final state quarks transversely polarized @xcite , as we show in the left panel of fig .
the long distance part of the above result can be factorized into the collins fragmentation function calculated from the right panel of fig . 5 .
in this part , because the @xmath116 integral is proportional to @xmath117 , we can combine the two terms in the integral into one expression contained in the collins function .
therefore , the spin - dependent cross section eq .
( [ xs ] ) can be re - written as @xmath118 \frac{\delta \hat q(z_h , p_{ht})}{m_h } \ , \end{aligned}\ ] ] where the collins function @xmath119 is calculated from the feynman diagram in the right panel of fig .
5 , @xmath120 where the index @xmath121 is not understood as a sum .
this final result demonstrates that we do have a factorization for the spin - dependent cross section into the hard factor @xmath83 times the collins fragmentation function , and the collins function is the same as that in @xmath0 and sidis processes @xcite .
therefore , by using the ward identity at this particular order , the final results for all the diagrams of fig . 4 will sum up together into a factorized form as shown in fig . 5 , where the cross section is written as the hard partonic cross section for @xmath122 subprocess multiplied by a collins fragmentation function . the exchanged gluon in fig .
4 is now attaching to a gauge link from the fragmentation function definition @xcite as shown in the right panel of fig .
5 . into the hard partonic cross section multiplied by the universal collins fragmentation function .
the short bars indicate the pole contribution to the collins function.__,width=377 ] -0.4 cm the key steps in the above derivation are the eikonal approximation and the ward identity .
the eikonal approximation is valid when we calculate the leading power contributions in the limit of @xmath123 .
the ward identity ensure that when we sum up the diagrams with all possible gluon attachments we shall get the eikonal propagator from the gauge link in the definition of the fragmentation function . the most important point to apply
the ward identity in the above analysis is that the eikonal propagator does not contribute to the phase needed to generate a nonzero ssa .
this is what we have shown for the collins asymmetry in the above calculations , and the reason , as we mentioned above , is due to the momentum flow of the exchanged gluon and the kinematic constraints in the fragmentation process .
we will show in the next section , that for the two - gluon exchange diagrams the eikonal propagators do not contribute to the phase for the nonzero collins ssa in this process .
therefore , we conjecture that the above conclusions are valid to higher order contributions too
. this argument can not apply to the ssa associated with the parton distributions , where the eikonal propagator does contribute to the phase to generate a nonzero ssa .
that is the reason we have sign differences for the sivers functions in sidis and drell - yan processes .
as we discussed in the last section , to demonstrate the universality of the collins function , we have to apply the ward identity to sum up all gluon exchange contributions into the gauge link from the definition of the fragmentation function . in order to use this argument
, the eikonal propagator should not contribute to the phase needed to generate nonzero ssa associated with the collins effects .
this has been explicitly demonstrated in the last section for the one - gluon exchange contribution . in this section
, we will extend the discussions to the two - gluon exchange contributions .
especially , we will show that these eikonal propagators do not contribute to the phase for the ssas .
the reason , again , is due to the time - like feature and the momentum flow in the fragmentation process . -0.4
cm we will focus our discussions on some representative diagrams from the two - gluon exchange contributions .
all other diagrams will follow accordingly .
we show these diagrams in figs .
6(a , b , c ) .
the contribution from fig .
6(a ) will depend on the following integral of the exchange gluons momenta @xmath124 and @xmath125 , @xmath126 where @xmath127 is the fragmenting quark s momentum and @xmath128 represents the numerators depending @xmath124 and @xmath125 , especially their transverse momentum components . following the arguments used in the last section ,
the first two propagators in the above expression can be further simplified by using the eikonal approximation , and then we will obtain the following expression @xmath129 where @xmath130 follow the definitions in the last section .
the normalization of the above integral has been changed for convenience .
this normalization is not relevant for our discussions , because we want to show that the eikonal propagators do not contribute to the phase needed for a nonzero ssa , not the actual contribution from this diagram
. we will show if we take pole contributions from these two eikonal propagators , the final integral will vanish .
because of the existence of two eikonal propagators , the analysis will be more complicated than that in the last section .
we discuss their contributions separately . 1 . pole contribution from @xmath131 .
+ if we take pole of this eikonal propagator , @xmath132 and @xmath133 will be constrained : @xmath134 , and the integral of ( [ eq1 ] ) will become , @xmath135 where @xmath136 are some quantities depending on the transverse momenta of @xmath137 , @xmath53 and @xmath49 .
the following analysis does not depend on the details of these numbers . in deriving the above equation ,
we have used the constraint of @xmath134 to simplify the expression .
we further notice that @xmath105 and @xmath138 .
thus , the poles of the second and third factor in the integral of @xmath139 and @xmath140 are both in the upper half plane .
if @xmath141 , which means that @xmath142 , the pole of the fifth factor will be also in the upper half plane of @xmath140 .
therefore , the poles of the three factors ( the second , third and fifth ) depending on @xmath140 are all in the upper half plane of @xmath140 , and the integral over @xmath140 will vanish , and so will the above integral .
similarly , if @xmath143 , the pole of the fourth factor will be in the upper half plane of @xmath139 .
meanwhile , we will also have @xmath144 , and the pole of the first factor will be in the upper half plane too . therefore , the poles of the four factors ( the first , second , third and fourth ) depending on @xmath139 are all in the upper half plane of @xmath139 .
the integral over @xmath139 will vanish , and so will the above expression . in conclusion , in any case of @xmath141 or @xmath143 , the above integral vanishes , and we do not have contribution from the pole of @xmath131 .
pole contribution from @xmath145 + because @xmath146 , we can simplify the integral of ( [ eq1 ] ) as follows , @xmath147 again , the normalization has been changed for convenience . because @xmath148 , we will analyze the contributions of the above equation by classifying the different regions of @xmath133 : ( a ) @xmath149 ; ( b ) @xmath150 ; ( c ) @xmath151 . in the region of ( a )
, we will have @xmath152 and @xmath153 .
therefore , the poles of the three factors ( the second , third and fourth ) are all in the lower half plane of @xmath140 , and the integral over @xmath140 vanishes . in the region of ( b ) , we have @xmath154 , @xmath155 and @xmath156 .
thus , the poles of the three factors ( the first , second and third ) depending on @xmath139 are all in the upper half plane , and the integral over @xmath139 vanishes . in the region of ( c ) , we have @xmath157 , @xmath153 and @xmath156 .
therefore , the @xmath140 integral will pick up the pole of the third factor , which actually determines the value of @xmath158 . after substituting this back into the equation
, we will find the second factor does not depend on @xmath139 any more .
the only dependence comes from the first factor .
obviously , this integral over @xmath139 will vanish . in conclusion , in any case of ( a , b ,
c ) , the above integral vanishes , and there is no contribution from the pole of @xmath145 . in summary , there is no contribution to the ssa from the pole of the eikonal propagators in the diagram of fig .
similarly , the contribution from fig .
6(b ) depends on the following integral , @xmath159 where we have made the eikonal approximations for the two propagators along the incident quark line @xmath3 . comparing with eq .
( [ eq1 ] ) , we find the only difference is the third factor @xmath160 .
again , we can show that none of the two eikonal propagators will contribute to the phase needed for a nonzero ssa .
we will discuss their contributions separately . 1 .
pole contribution from @xmath131 .
+ after taking this pole , the integral of ( [ eq2 ] ) will become , @xmath161 which is the same as eq .
( [ eq3 ] ) if we interchange @xmath162 and @xmath163 .
thus , the above integral will vanish by the same arguments we have used for eq .
( [ eq3 ] ) .
pole contribution from @xmath145 + because @xmath146 , we can simplify the integral of ( [ eq2 ] ) as follows , @xmath164 again , we classify three different regions of @xmath133 in the above equation : ( a ) @xmath149 ; ( b ) @xmath150 ; ( c ) @xmath151 .
the contributions from ( a ) and ( b ) regions vanish by the same reasons as we have shown for eq .
( [ eq4 ] ) in the above . in region ( c )
, we have @xmath153 and @xmath156 .
therefore , the @xmath139 integral will pick up the pole of the third factor , which again actually determines the value of @xmath158 . after substituting this back into the equation
, we will find the second factor does not depend on @xmath140 any more .
the only dependence comes from the first and last factors .
obviously , this integral over @xmath140 vanishes because @xmath153 and @xmath157 , and the poles of these two factors are both in the lower half plane . in conclusion ,
in any case of ( a , b , c ) , the above integral vanishes , and there is no contribution from the pole of @xmath145 . similarly , the contribution from fig .
6(c ) will depend on the following integral , @xmath165 where again we have made the eikonal approximations .
there are two eikonal propagators in the above equation , and as above we will discuss their contributions separately . 1 . the pole contribution from @xmath145 + after taking this pole , @xmath146 , the above equation eq .
( [ eq5 ] ) will reduce to eq .
( [ eq4 ] ) . according to the same arguments we used there
, there will be no contributions from this pole .
2 . the pole contribution from @xmath166
+ this pole contribution means that @xmath167 , and the integral of eq .
( [ eq5 ] ) become @xmath168 which will be identical to eq .
( [ eq6 ] ) if we interchange @xmath162 to @xmath163 . using the same arguments there , the above integral vanishes .
the above three examples are typical diagrams we encounter for the two - gluon exchange contributions for this channel .
all these diagrams can be analyzed by a similar manner , and we will find that the eikonal propagators do not contribute to the phase needed to a nonzero ssa .
because of this fact , all these diagrams can be summed together to form the contributions from the gauge link in the fragmentation function , where the two gluons attach to the gauge link similar to the diagram we have shown in fig .
since there is no contributions from these eikonal propagators , the collins function calculated from these diagrams will be the same as that in @xmath0 and sidis processes , and the universality preserved .
we have also drawn some other diagrams at this order in fig .
6(d , e , f ) , which contribute to a real gluon radiation in addition to the gluon exchange .
the analysis of these diagrams also show that we do not get contribution from the pole of the eikonal propagators .
for example , the contribution from fig .
6(d ) depends on @xmath169 where @xmath170 is the momentum for the radiated gluon .
we have two eikonal propagators in the above equation . however
, none of them contributes to the phase needed to a nonzero ssa . 1 . the pole contribution from @xmath171 + this pole contribution means that @xmath172 , and the integral of @xmath93 will reduce to @xmath173 because @xmath174 and @xmath105 , the poles of the above two factors are both in the lower half plane of @xmath93 , and the integral vanishes .
2 . the pole contribution from @xmath175 + after taking this pole , we will have the following @xmath93 integral @xmath176 because the pole constrains that @xmath177 and @xmath178 , the poles of the above three factors are all in the lower half plane .
the integral over @xmath93 vanishes . in summary
, there is no contribution from the pole of the eikonal propagators in the diagram of fig .
the contribution from fig .
6(e ) will depend on the following integral , @xmath179 because @xmath180 , we only have one possible pole contribution from the eikonal propagator @xmath181 , which vanishes by the same reason as above for diagram fig .
similarly , if the gluon with momentum @xmath18 attaches to the radiated gluon instead of the incident quark line with momentum @xmath3 ( we did not show this diagram in fig .
6 ) , the contribution vanishes by the same reason .
the contribution from fig .
6(f ) depends on the following integral @xmath182 after eikonal approximation .
if we take the pole contribution from the eikonal propagator , the above integral will reduce to @xmath183 again , because @xmath184 , the poles of the above three factors are all in the lower half plane , and the integral over @xmath93 vanishes .
thus , there is no contribution from the pole of the eikonal propagator for this diagram . in summary , for the gluon radiation diagrams
, there is no contributions from the pole of the eikonal propagators . because of this fact
, we can use ward identity to sum all these diagrams together to form the gauge link contribution from the fragmentation function , similar to the diagram in fig . 5 with an additional gluon radiation .
concluding the analysis of the two - gluon exchange diagrams in fig .
6 , the eikonal propagators do not contribute to the phase needed for the nonzero ssa associated with the collins effect .
therefore , we can apply the ward identity at this order to sum all these diagrams plus other similar ones .
this sum will lead to the gauge link contribution from the fragmentation function definition , and the fragmentation function will be the same as that in @xmath0 and sidis processes .
in this paper , we have shown that the collins function in hadron production in single - transverse - spin polarized @xmath1 scattering is the same as that in @xmath0 and sidis processes .
this universality is a general and model - independent observation , and depends on the fact that the eikonal propagators do not contribute to the phase needed for a nonzero ssa .
we have demonstrated this by explicit calculations for one - gluon exchange diagrams which corresponds to one eikonal propagator in the amplitudes , and two - gluon exchange diagrams which correspond to two eikonal propagators .
although our calculations were based on a model , the analysis and arguments are quite general . the results , as we emphasized , depend on the momentum flow and kinematic constraints in the fragmentation process .
this observation is very different from the ssas associated with the parton distributions , where the eikonal propagators from the gauge link in the parton distribution definition play very important role .
it is the pole of these eikonal propagators contribute to the phase needed for a nonzero ssa associated with the naive - time - reversal - odd parton distributions , which also predicts a sign difference for the quark sivers function between the sidis and drell - yan processes .
more complicated results have been found for the ssas in the hadronic dijet - correlation @xcite , where a normal tmd factorization breaks down @xcite .
the reason is that the eikonal propagators from the initial and final state interactions in dijet - correlation process do contribute poles in the cross section @xcite .
because of this , the ward identity is not applicable , and the standard tmd factorization breaks down , although a modified factorization may be valid if we modify the definition of the tmd parton distributions to take into account all the initial and final state interaction effects @xcite . in the fragmentation process , as we discussed in our paper , the eikonal propagators do not contribute to an imaginary part , and the ward identity is applicable .
we have shown this in our explicit calculations including one - gluon and two - gluon exchange contributions .
there has been discussion about the twist - three quark - gluon correlation contribution in the fragmentation function , especially for the collins effects .
it will be interesting to further understand these contributions following the analysis in this paper , and discuss the universality issues in a more general ground .
we thank j. collins , l. gamberg , r. jaffe , x. ji , j. qiu , and w. vogelsang for useful discussions .
especially , we thank a. metz for his valuable comments and many useful discussions . this work was supported in part by the u.s .
department of energy under contract de - ac02 - 05ch11231 .
we are grateful to riken , brookhaven national laboratory and the u.s .
department of energy ( contract number de - ac02 - 98ch10886 ) for providing the facilities essential for the completion of this work .
see , for example : d. l. adams _ et al .
_ [ e581 and e704 collaborations ] , phys .
b * 261 * , 201 ( 1991 ) ; d. l. adams _ et al .
_ [ fnal - e704 collaboration ] , phys .
b * 264 * , 462 ( 1991 ) ; k. krueger _
_ , phys .
b * 459 * , 412 ( 1999 ) .
a. airapetian _ et al . _
[ hermes collaboration ] , phys .
lett . * 84 * , 4047 ( 2000 ) ; phys .
lett . * 94 * , 012002 ( 2005 ) . v.
y. alexakhin _ et al . _ [ compass collaboration ] , phys .
lett . * 94 * , 202002 ( 2005 ) .
f. videbaek [ brahms collaboration ] , aip conf
. proc . * 792 * , 993 ( 2005 ) ; j. h. lee [ brahms collaboration ] , talk presented at the `` 17th international spin physics symposium ( spin 2006 ) '' , kyoto , japan , october 2 - 7 , 2006 . for reviews ,
see : m. anselmino , a. efremov and e. leader , phys .
rept . * 261 * , 1 ( 1995 ) [ erratum - ibid .
* 281 * , 399 ( 1997 ) ] ; v. barone , a. drago and p. g. ratcliffe , phys .
rept . *
359 * , 1 ( 2002 ) .
a. v. efremov and o. v. teryaev , sov . j. nucl .
phys . * 36 * , 140 ( 1982 ) [ yad .
fiz . * 36 * , 242 ( 1982 ) ] ; a. v. efremov and o. v. teryaev , phys .
b * 150 * , 383 ( 1985 ) .
j. qiu and g. sterman , phys .
lett . * 67 * , 2264 ( 1991 ) ; nucl .
b * 378 * , 52 ( 1992 ) ; phys .
d * 59 * , 014004 ( 1999 ) ; c. kouvaris , j. w. qiu , w. vogelsang and f. yuan , phys
. rev .
d * 74 * , 114013 ( 2006 ) .
ralston and d. e. soper , nucl .
b * 152 * , 109 ( 1979 ) .
j. c. collins and d. e. soper , nucl .
b * 193 * , 381 ( 1981 ) [ erratum - ibid .
b * 213 * , 545 ( 1983 ) ] ; nucl .
b * 197 * , 446 ( 1982 ) .
j. c. collins and d. e. soper , nucl .
b * 194 * , 445 ( 1982 ) .
j. c. collins , d. e. soper and g. sterman , nucl .
b * 250 * , 199 ( 1985 ) .
d. w. sivers , phys .
d * 41 * , 83 ( 1990 ) ; phys .
d * 43 * , 261 ( 1991 ) .
j. c. collins , nucl .
b * 396 * , 161 ( 1993 ) .
m. anselmino , m. boglione and f. murgia , phys .
b * 362 * , 164 ( 1995 ) ; m. anselmino and f. murgia , phys .
b * 442 * , 470 ( 1998 ) ; m. anselmino , m. boglione , u. dalesio , e. leader and f. murgia , phys . rev .
d * 71 * , 014002 ( 2005 ) .
j. mulders and r. d. tangerman , nucl .
b * 461 * , 197 ( 1996 ) [ erratum - ibid .
b * 484 * , 538 ( 1997 ) ] ; d. boer and p. j. mulders , phys .
d * 57 * , 5780 ( 1998 ) .
s. j. brodsky , d. s. hwang and i. schmidt , phys .
b * 530 * , 99 ( 2002 ) ; nucl .
b * 642 * , 344 ( 2002 ) .
j. c. collins , phys .
b * 536 * , 43 ( 2002 ) .
x. ji and f. yuan , phys .
b * 543 * , 66 ( 2002 ) ; a. v. belitsky , x. ji and f. yuan , nucl .
b * 656 * , 165 ( 2003 ) .
d. boer , p. j. mulders and f. pijlman , nucl .
b * 667 * , 201 ( 2003 ) .
d. boer , r. jakob and p. j. mulders , nucl .
b * 504 * , 345 ( 1997 ) .
a. v. efremov , k. goeke and p. schweitzer , phys .
d * 73 * , 094025 ( 2006 ) .
m. anselmino , m. boglione , u. dalesio , a. kotzinian , f. murgia , a. prokudin and c. turk , phys .
d * 75 * , 054032 ( 2007 ) .
x. ji , j. p. ma and f. yuan , nucl .
b * 652 * , 383 ( 2003 ) .
x. ji , j. w. qiu , w. vogelsang and f. yuan , phys .
* 97 * , 082002 ( 2006 ) ; phys .
d * 73 * , 094017 ( 2006 ) ; phys .
b * 638 * , 178 ( 2006 ) . c. j. bomhof , p. j. mulders and f. pijlman , phys .
b * 596 * , 277 ( 2004 ) ; eur . phys .
j. c * 47 * , 147 ( 2006 ) ; jhep * 0702 * , 029 ( 2007 ) ; a. bacchetta , c. j. bomhof , p. j. mulders and f. pijlman , phys
. rev .
d * 72 * , 034030 ( 2005 ) ; c. j. bomhof and p. j. mulders , arxiv:0709.1390 [ hep - ph ] .
a. metz , phys .
b * 549 * , 139 ( 2002 ) .
j. c. collins and a. metz , phys .
* 93 * , 252001 ( 2004 ) .
d. amrath , a. bacchetta and a. metz , phys .
d * 71 * , 114018 ( 2005 ) ; l. p. gamberg , g. r. goldstein and k. a. oganessyan , phys .
d * 68 * , 051501 ( 2003 ) .
m. stratmann and w. vogelsang , phys .
b * 295 * , 277 ( 1992 ) .
u. d alesio , talk given at trento workshop on transverse momentum , spin and position distributions of partons in hadrons 11 - 15 jun 2007 , trento , italy . | we study the collins effect in the azimuthal asymmetric distribution of hadrons inside a high energy jet in the single transverse polarized proton proton scattering . from the detailed analysis of one - gluon and two - gluon exchange diagrams contributions , the collins function
is found the same as that in the semi - inclusive deep inelastic scattering and @xmath0 annihilations .
the eikonal propagators in these diagrams do not contribute to the phase needed for the collins - type single spin asymmetry , and the universality is derived as a result of the ward identity .
we argue that this conclusion depends on the momentum flow of the exchanged gluon and the kinematic constraints in the fragmentation process , and is generic and model - independent . |
constraining the galactic potential optimally has been a very difficult yet important aspect of studies of the milky way and the local group . understanding the various components of the galactic potential , for instance separating the potential contributions from the baryonic disc and that from the dark matter halo , is fundamental to understanding the history and formation of our own galaxy .
furthermore , understanding the dark halo potential near the sun is a crucial step to pin down the density and thus the scattering cross - section of the dark matter , which in turn is crucial input for interpreting any dark matter annihilation signal at the centre of milky way ( e.g. * ? ? ?
* ) . in the past few decades , constraining the galactic potential
has mostly relied on the jeans equation ( for a summary , see * ? ? ?
* ) despite the known problems of this approach .
for instance , one only keeps the velocity dispersion moments up to second order in the jeans equation . this predicts a gaussian ellipsoid velocity distribution that does not capture the observed and expected asymmetries in the angular velocity @xmath0 ( e.g. * ? ? ? * ) .
one of the ways to alleviate these problems is through explicitly modelling the stellar population distribution function ( df ) , either via integrals of motions or actions instead of just taking position
velocity moments , as is the case of the jeans equation @xcite .
for such modelling , it is crucial that relatively simple analytic dfs , e.g. based on the actions , are reasonably good approximations to the actual orbit distributions for at least subpopulations of stars in the galactic disc .
how well such df approximations work in practice is not yet clear . to explore this question
, we consider in particular stellar subpopulations of a particular age or abundance pattern @xcite .
recently , @xcite and @xcite ( hereafter refer all of these references as bo12 ) showed , using g - dwarfs from the sloan extension for galactic understanding and exploration ( segue ) survey @xcite , that mono - abundance , stars that have nearly the same [ fe / h ] and [ @xmath1/fe ] , populations have very simple @xmath2 phase - space distribution properties : spatial density distributions that vary exponentially , both in the radial and vertical directions ; furthermore , each mono - abundance population shows isothermal velocity dispersion in the vertical direction and exponentially in the radial direction @xcite .
this simplicity in @xmath2 space gives hope that one could describe mono - abundance populations with simple action - based dfs which is one of the purposes of this study .
since stars are the most abundant tracers of the galactic potential , simple action - based dfs will naturally provide a statistical and rigorous method to constrain the galactic potential because the conversion of position velocity variable ( configuration space ) to the action angle variable depends on the parameters of the potential ( see @xcite for generic ground work on the topic ) .
the main idea is to determine the likelihood of the observational data , given a joint set of parameters for both the df and the gravitational potential ; subsequent marginalization over the df parameters provides then a rigorously derived constraint on the potential .
such an approach appears to be a precondition to fully exploit the dynamical information content of large galactic surveys that will provide us spatial and velocity distribution , along with elemental abundances , including the apache point observatory galactic evolution experiment @xcite , galactic archaeology survey with hermes ( galah ; * ? ? ?
* ; * ? ? ?
* ) and european southern observatory-_gaia _ @xcite . in the context of dynamically modelling the galactic disc with action - based dfs ,
the goals of the paper are three - fold : explore how well ` mono - abundance populations ' can be approximated by simple action - based analytic dfs ; lay out a formalism that provides constraints from a set of discrete stellar positions and velocities on the gravitational potential , after marginalizing over the df ; forecast what constraints on the _ shape _ of the potential we can expect given existing sample sizes .
this paper is organized as follows : in section[sec : df ] , we discuss the choice of parametrization of the df and summarize the adiabatic approximation that we assume to calculate the action variable efficiently . in section[sec : df - match ] , we will show that our choice of df exhibits all the basic phase - space distribution properties of mono - abundance stars , with the only caveat that disc potential has to be included in order for the vertical spatial profile to fit . in section[sec : constrain_potential ] , we show that by studying the likelihood of the df potential parameters , one can recover the potential parameters , even under a restrictive and realistic spatial selection function . throughout this study
, we denote the cylindrical coordinate to be @xmath3 .
we assume that velocities are measured in the inertial galactocentric frame with axisymmetric potential , @xmath4 . in practice
, the conversion from the heliocentric frame to the galactocentric frame requires the knowledge of the solar motion . from the study of sgra * , the solar motion is now accurate to a few @xmath5 ( e.g. * ? ? ? * ) and the solar radius @xmath6 is accurate to @xmath7kpc ( e.g. * ? ? ?
* ; * ? ? ?
we will also discuss the uncertainty in estimating the potential parameters due to the uncertainty of this conversion in section[sec : constrain_potential ] .
we follow the df advocated by @xcite and @xcite closely , but note that we rearrange some of the terms in the df to facilitate physical interpretation of the df . for circular orbits in the equatorial galactic plane
, the df has to be of the form @xmath8 where @xmath9 determines the angular momentum distribution for an infinitely cold planar disc , @xmath10 and @xmath11 are the radial and vertical action , respectively . in the cold disc limit , the circular radius @xmath12 coincides with the observed radius , where @xmath13 and @xmath14 is the circular velocity , @xmath15 thus , we have for an exponential disc , up to a multiplicative normalization constant , @xmath16 the df parameter @xmath17 determines the radial spatial distribution . in a more general setting ,
we relax the @xmath18 constraint and propose the df to be @xmath19\notag\\ \bigg[\frac{\nu(l_z)}{d_r(l_z)}\exp\bigg(-\frac{\nu ( l_z ) j_z}{d_r(l_z)}\bigg)\bigg ] , \label{eq : df}\end{aligned}\ ] ] where @xmath20 @xmath21 with @xmath12 the circular radius given angular momentum @xmath22 , assuming equatorial in - plane movement ; @xmath23 and @xmath24 are the radial and vertical frequency , respectively , under epicycle approximation .
the df parameter @xmath25 determines the radial exponential decay of the velocity dispersion , whereas @xmath26 and @xmath27 control the total radial and vertical dispersion , respectively . and
@xmath27 are _ different _ from the velocity dispersions obtained as moments of the df , which correspond to the observable dispersion .
these parameters are named as such because they govern the velocity dispersions .
they differ in their value quite significantly from the velocity dispersions in the @xmath2 space . ]
the @xmath28 and @xmath29 terms are necessary ( e.g. * ? ? ?
* ) for two reasons . 1 .
a finite increment in energy @xmath30 can lead to an infinite increment of actions in the unbound case .
one would love to have a term that will tend to zero at large @xmath22 and couple this term with @xmath10 .
one of the possible choices is the radial frequency @xmath31 . at large @xmath12 ,
the effective potential is less concave at the minimum point and therefore @xmath31 tends to zero .
similarly for the vertical frequency .
2 . qualitatively , in @xmath12 region where @xmath28 changes more drastically , the increment of @xmath10 is also more drastic , and therefore the scalelength of @xmath10 has to decrease in proportion to compensate for this effect .
similarly for the vertical oscillation .
importantly , we only measure @xmath32 but not the actions . in order to study an action - based df
, we employ the adiabatic approximation in order to calculate the radial and vertical actions efficiently .
, @xmath33kpc , @xmath34kpc , the vertical dashed line marks approximately the solar galactocentric radius , the horizontal dashed line indicates the common accepted circular velocity @xmath35 at the solar radius and the shaded region shows the region where the circular velocity is approximately constant in this choice of potential ; the middle and bottom panels show the variation of actions calculated using the adiabatic approximation in the course of trajectory .
we plot the case of @xmath36 in table[table : adiabatic ] .
the results show that actions calculated with the adiabatic approximation are almost conserved along the orbit with less than @xmath37per cent variation .
the orbit was integrated about 10 gyr.,title="fig:",width=307 ] , @xmath33kpc , @xmath34kpc , the vertical dashed line marks approximately the solar galactocentric radius , the horizontal dashed line indicates the common accepted circular velocity @xmath35 at the solar radius and the shaded region shows the region where the circular velocity is approximately constant in this choice of potential ; the middle and bottom panels show the variation of actions calculated using the adiabatic approximation in the course of trajectory .
we plot the case of @xmath36 in table[table : adiabatic ] .
the results show that actions calculated with the adiabatic approximation are almost conserved along the orbit with less than @xmath37per cent variation .
the orbit was integrated about 10 gyr.,title="fig:",width=316 ] it has been shown that the vertical action @xmath11 calculated in a fixed radius @xmath38 approximation is almost conserved in the axisymmetric system @xcite , and in discs including bar / spiral structure perturbation and radial migration @xcite . more precisely , @xmath39 this reduces the calculation to one - dimensional integration that can be numerically calculated easily . furthermore , since @xmath40 , to calculate the approximated vertical action , we have @xmath41 where the subscript ` init .
' stands for the observed value in practice .
similarly , for the radial action , we can consider the radial component at the equatorial plane , ignoring the vertical motion . under this assumption , @xmath42 ccccccc + @xmath43 & @xmath44 & @xmath22 & @xmath45 & @xmath46 & @xmath47 & @xmath48 + ( kpc ) & ( kpc ) & @xmath49 & @xmath50 & @xmath50 & ( per cent ) & ( per cent ) + + 3 & 0 & 450 & 90 & 60 & 4 & 3 + 6 & 0 &
1250 & 60 & 40 & 5 & 3 + 9 & 0 & 1970 & 30 & 20 & 4 & 2 + we deduce @xmath51 @xmath52 where @xmath53 and @xmath54 are the apo- and pericentric radius , respectively .
fig.[fig : adiabatic ] and table[table : adiabatic ] illustrate the variation of the approximated actions in the course of a trajectory .
we assume a miyamoto nagai potential @xcite , @xmath55 with parameters @xmath56 , @xmath33kpc , @xmath34kpc .
we choose these parameters because we have almost a flat rotation curve at the galactocentric radius @xmath57kpc and circular velocity @xmath58 at the solar radius @xmath59kpc as shown in the top panel of fig.[fig : adiabatic ] .
we apply slight perturbation on circular orbits that mimic the radial velocity dispersion of @xmath60 at the solar galactocentric radius , radial - to - vertical velocity dispersion ratio of @xmath61 @xcite and velocity scalelength of @xmath62kpc ( bo12 ) .
we define the variation of action @xmath63 to be @xmath64 over the mean value of the action for each time step .
the results of @xmath63 are listed in table[table : adiabatic ] .
the results show that the approximated actions are almost conserved in all cases .
we also examine the potential due to a double - exponential disc ( i.e. with density decaying both radially and vertically ) ( see appendix a , equation a10 , * ? ? ?
* ) , and also the flattened isothermal potential : @xmath65 the results are qualitatively similar , with the variation of actions less than @xmath37 per cent in all cases .
@xcite studies a more precise method of calculating the actions , but it is computationally more demanding . as we will discuss in section[sec : constrain_potential ] , we find that the error in using the adiabatic approximation will introduce a systematic error of about @xmath66per cent in estimating the potential parameters .
the adiabatic approximation is sufficient for our current study since the purpose of this study is to introduce a new statistical method to constrain the galactic potential .
we only focus on the miyamoto
nagai potential and the flattened isothermal potential in this study .
we do not consider the potential due to a double - exponential disc because the analytic form of this potential ( by solving the poisson equation ) involves an improper integration of a bessel function @xcite .
this is computationally very expensive .
moreover , the miyamoto
nagai potential is good enough as a first approximation for a realistic disc potential .
in this section , we will show that the df in equation([eq : df ] ) is a good representation of the observed position velocity distribution of mono - abundance subpopulations of the milky way disc , namely that they exhibit an exponential spatial density , both radially and vertically , and an isothermal velocity dispersion in the vertical direction and exponentially decaying in the radial direction .
@xmath67{fig2a.ps } & \includegraphics[scale=0.27]{fig2b.ps } \end{array } $ ] + @xmath68{fig2c.ps } & \hspace{0.2 cm } \includegraphics[scale=0.23]{fig2d.ps } & \hspace{0.2 cm } \includegraphics[scale=0.23]{fig2e.ps } \end{array }
$ ] in order to have the conversion of df between canonical position velocity variable and the action variable , one has to take care of the jacobian term carefully .
note that @xmath69 where @xmath70 is the canonical momentum associated with @xmath71 .
since @xmath72 , @xmath73 are canonical variables and measure is invariance under canonical transformation , we have @xmath74 furthermore , in a cylindrical coordinate , @xmath75 , we deduce @xmath76 in summary , for an axisymmetric potential , in order to convert an action - based df to @xmath2 space distribution , it suffices to multiply the jacobian ( @xmath77 ) term .
we create mock data stellar tracers with @xmath32 attributes , such that it has a density scalelength of about @xmath78kpc , dispersion scalelength of about @xmath62kpc with a gaussian velocity dispersion of @xmath79 at the solar galactocentric radius , @xmath80 , @xmath81 ( * ? ? ?
we assume that the mean velocity of the gaussian for each spatial point is @xmath82 @xmath83 where ( see @xcite for details ) @xmath84
we assume that all of the mixed moments ( @xmath85 , @xmath86 and @xmath87 ) vanish .
thus , we assume that both the vertex deviation and the tilt of the velocity ellipsoid are zero , even though these were not measured by bo12 .
the tilt of the velocity ellipsoid in particular is not expected to be zero at heights @xmath88kpc above the plane ( e.g. * ? ? ?
* ) . however , as discussed by @xcite , the adiabatic approximation is unable to capture a non - zero tilt as it assumes that the radial and vertical motions are independent .
thus , even though the quasi - isothermal df of equation([eq : df ] ) has a non - zero tilt when used with correctly calculated actions ( for example , using the torus machinery ; * ? ? ?
* ) , it does not when using the adiabatic approximation . for this reason , we assume that the tilt of the velocity ellipsoid is zero in what follows . , but for the isothermal potential with the axis ratio @xmath89 and circular velocity @xmath35 .
we also tried various choices of @xmath90 , but the fits are equally unsatisfactory : the df family in equation([eq : df ] ) can not produce a vertically exponential tracer density profile , if stars orbit in a flattened isothermal sphere . ]
we search for the best - fitting df parameters by maximum likelihood .
for this section , we fix a particular choice of miyamoto nagai potential parameters : @xmath91 , @xmath33kpc , @xmath34kpc .
we examine various choices of the potential parameters .
it does not change the results qualitatively .
we emphasize that the mock tracer population mimics a mono - abundance population .
note that the probability that we observe @xmath92 , according to our df model with df parameters @xmath93 , is @xmath94 given this probability , we can define the log likelihood to be the sum over @xmath95 mock data , and search for the best df parameters by maximizing the log likelihood , @xmath96.\end{aligned}\ ] ] we obtain the best - fitting parameters by performing a nested - grid search on the multidimensional @xmath93 space and finding the set of parameters that maximizes the likelihood .
after obtaining the best - fitting parameters , we calculate from @xmath97 the stellar profile in configuration space @xmath2 and compare it to the one of the mock data to check whether our assumption on the functional form of the df represents well the mock data .
we consider a grid of position velocity configuration space @xmath32 and calculate the df in configuration space using equation([eq : configuration_space ] ) .
we then calculate the stellar profile of each position velocity component by marginalizing over the other spatial - velocity components .
note that , given the functional form of the df in equation([eq : df ] ) , the predicted vertical profile is fixed by the vertical velocity dispersion and the potential , but independent of the scaleheight of the mock data .
we find that the vertical profile is predicted satisfactorily with the miyamoto
nagai potential ( as shown in fig.[fig : nm_fit ] ) but not with the flattened isothermal potential ( as shown in fig.[fig : log_fit ] ) .
this can be explained by the following . in the case where @xmath98 , the vertical restoring force of the miyamoto nagai potential and thus @xmath99 is about constant .
this implies that the vertical action is proportional to @xmath100 , and the vertical profile in ansatz equation([eq : df ] ) goes down exponentially . on the other hand ,
the vertical restoring force of the flattened isothermal potential and thus @xmath99 is proportional to @xmath100 .
this implies that the vertical action is proportional to @xmath101 , and the vertical profile will go down too drastically [ proportional to @xmath102 .
also note that , in the upper panel of fig.[fig : nm_fit ] , the df predicts a shallower vertical density profile near the galactic plane . as segue does not observe the galactic plane , we can not decide whether this prediction is accurate or one should revise the df to better fit the vertical density profile at this point .
nonetheless , the density distribution is not expected to be exponential all the way to @xmath103 and @xmath104 .
our aim here is to match the exponential profiles in the range of @xmath38 and @xmath100 where segue has data .
we conclude that the quasi - isothermal df of equation([eq : df ] ) provides a good representation of the df of mono - abundance subpopulations of the milky way disc , namely that the df predicts ( 1 ) an exponential spatial density , both radially and vertically ; ( 2 ) an isothermal velocity dispersion in the vertical direction and ( 3 ) exponentially decaying in the radial direction .
more importantly , the action - based dfs predict a low-@xmath0 tail . as discussed in section[sec : introduction ] , this coincides better with the observation ( e.g. * ? ? ?
* ) . in the following section
, we will generate mock samples by rejection sampling _ directly from the df_.
in this section , we will show that , given a functional form of the potential and the df , we can recover the optimal potential parameters by marginalizing over the df parameters .
note that in practice , we only observe a very small part of the galaxy . in this study , we assume that the observed volume is a cylinder around the sun with a finite width and height .
we assume that the location of the sun in the galactocentric cartesian coordinate is @xmath105 and include the spatial selection function @xmath106 with this selection function in place , given any parameters of the df and the potential , @xmath107 , we carefully normalize the df through monte carlo integration , more precisely @xmath108 where @xmath109 to our best knowledge , this is the first attempt that has never before been implemented to study the joint distribution of df and potential parameters in the context of constraining the galactic potential .
first , we consider the two - dimensional in - plane miyamoto nagai potential which contains two potential parameters : @xmath110 and @xmath111 .
we consider the case where @xmath112kpc , @xmath113kpc and @xmath114 in equation([eq : df ] ) , which roughly mimics the observed velocity distributions from our calculation in section[sec : df - match ] .
we define the likelihood of the potential parameters , @xmath115 , by marginalizing over @xmath93 .
we estimate the one - sigma significance of the best - fitting potential parameter by finding the demarcation line at @xmath116 and two - sigma significance at @xmath117 .
we create data with @xmath56 , @xmath33kpc , @xmath34kpc in the observed volume @xmath66kpc radially from the sun and @xmath118kpc .
we reiterate the process for various sample sizes for the mock data , @xmath119 for the same observed volume . as shown in fig.[fig : nm_2d_opti_potential ] , the uncertainty goes down as @xmath120 .
we use the results with the sample size @xmath121 to be our reference , because the uncertainty of the potential parameters is @xmath122per cent in this case .
we show that this reference value coincides with the true value , indicating that , assuming a potential model , we can recover the true potential parameter with observation restricted to a small volume .
we show that we can find the best - fitting potential parameters within one- to two - sigma significance from the true value regardless of the sample size .
more importantly , the data constrain a highly degenerate combination of @xmath110 and @xmath111 : since @xmath123 , we find that @xmath124 as seen in fig.[fig : nm_2d_opti_potential ] . on top of that , we show that , with the sample size of @xmath125 , we recover both parameters with a precision of about @xmath126per cent for a @xmath127per cent significance ( two - sigma ) .
the uncertainties are larger than the one in the study of a flattened isothermal potential below because the parameters are highly correlated .
we extend this result to the three - dimensional case by considering the flattened isothermal potential with two parameters ( equation[eq : isothermal_potential ] ) : @xmath14 and @xmath90 .
we choose to constrain only two potential parameters because ( 1 ) the main purpose of this study is to show that we can pin down the two upmost important parameters of the galactic potential around the solar neighbourhood , namely the total mass and the vertical density scaleheight . in this regard , two parameters are sufficient .
( 2 ) while our method is general and could be used to constrain any number of potential parameters , it is computationally expensive to add another dimension in the nested - grid optimization method that we use .
we could have implemented markov chain monte carlo ( mcmc ) and have increased the efficiency of higher dimensional search , but the purpose of this paper is to illustrate the new method .
we shall leave the mcmc implementation to a later study .
we consider the warm - disc case where @xmath112kpc , @xmath113kpc , @xmath114 and @xmath128 for df in equation([eq : df ] ) .
we create the mock data from the potential with @xmath35 and @xmath129 . as shown in fig.[fig : log_3d_opti_potential ] , one can recover both the circular velocity @xmath14 and the axis ratio @xmath90 successfully with our proposed method . note that , in most cases , without a galactic dust map , observations in the equatorial galactic disc are less reliable
. therefore , we also tested the more realistic scenario , where the observed volume is restricted to the cylindrical height of @xmath130kpc without the mid - plane .
the result as shown in fig.[fig : log_3d_opti_potential_without_midplane ] shows that we can about equally well constrain the axis ratio @xmath90 and the circular velocity @xmath14 , given the same sample size
. this could be probably due to the fact that our three - dimensional flattened isothermal potential is too restrictive ; therefore , the axis ratio is not too sensitive to the vertical observed volume restriction . in both cases , we show that , with the sample size of @xmath125 , we recover the circular velocity with a precision of about @xmath66per cent for a @xmath127per cent significance ( two - sigma ) and of about @xmath131per cent for the axis ratio .
it is important to note that , in this study , we perform the potential parameter estimation by marginalizing over the df parameters instead of maximizing over the df parameters .
it has been argued that this is the best way to constrain the potential in this type of study @xcite . to the best of our knowledge ,
our study is also the first successful study in implementing this marginalization .
interestingly , we find that by performing marginalization , the results we obtain are almost exactly the same as maximizing over the df parameters .
since marginalization will take at least an order of magnitude more in terms of computational time , it might be worthwhile to explain these results . to illustrate these results
, we consider a two - dimensional parameter space @xmath132 and we marginalize over one of the dimensions .
it is trivial to generalize the arguments to higher dimensions .
if the posterior joint distribution is a two - dimensional joint gaussian probability distribution , assuming flat prior , the posterior probability distribution is given by @xmath133 since this is a monotonic function , the contour of the posterior probability distribution will resemble the contour of @xmath134 .
note that the general formula for a two - dimensional posterior gaussian probability distribution can be written as @xmath135.\end{aligned}\ ] ] one can show that the maximum of @xmath136 is obtained at @xmath137 .
therefore , @xmath138 differing only by a multiplicative constant , this is the same as @xmath139 when considering the difference in @xmath134 with respect to the minimum point , the multiplicative constant will drop out .
hence , with all the arguments above , we show that in the marginalized @xmath140 , to obtain the boundary point at which @xmath141 , it is the same as maximizing over parameter @xmath142 , with the assumption that the posterior is a multidimensional gaussian distribution .
to better address the concern of systematic error arising from the use of the adiabatic approximation in the potential parameter estimation , we also perform an identical study with mock data generated using the torus machinery ( mcmillan & binney , private communication ) .
the torus machinery @xcite , compared to the adiabatic approximation , is an more accurate method of calculating actions .
we find that the best - fitting parameters ( recovered assuming the adiabatic approximation ) , in this case , are slightly shifted ( about @xmath66per cent ) compared to the input parameters ( generated using the torus machinery ) .
this is probably due to the error in calculating actions with the adiabatic approximation .
we conclude that more accurate and efficient ways of calculating actions are very valuable in improving our analysis ; however , our analysis with the adiabatic approximation only introduces a non - significant error . , but with the vertical range of available tracer constraints restricted to @xmath143kpc the mid - plane is cutout : this is a more realistic selection function since in most cases , without a galactic dust map , observations in the equatorial galactic disc are less reliable . ] we also test our method by shifting the galactocentric frame velocities by @xmath144 , mimicking the uncertainty due to the solar motion .
we find that all the results presented in this paper remain to be qualitatively similar , with the only difference that the estimated circular velocity of the galactic potential will be shifted by the same amount .
more importantly , when the galactocentric frame is shifted by a larger amount , the maximum likelihood decreases more , accordingly .
therefore , if we consider the solar motion to be another free parameter in the fit , we could study the most probable solar motion with our proposed method and put an independent constraint on the true solar motion in the galactocentric frame .
bo12 show that the galactic disc , when decomposed into mono - abundance populations , contains cold populations ( i.e. smaller velocity dispersion ) as well as warmer populations .
to compare the constraints which we can obtain from different populations , we study the three - dimensional miyamoto nagai potential .
we consider a warm population with @xmath114 , @xmath128 as before , and a cold population with @xmath145 , @xmath146 .
we fit for @xmath147 ( the scaleheight ) and @xmath110 ( the total mass ) , but fix @xmath111 ( the scalelength ) to keep the computational cost down ( see above ) .
we find that cold populations constrain the total mass @xmath110 better and also constrain slightly better the scaleheight @xmath147 .
the result is expected for @xmath110 because it is related to the circular velocity .
one should be able to pin down the circular velocity better if the dispersion of the velocity is smaller .
we interpret the result for the scaleheight as follows : there are two competing effects for the scaleheight estimation : ( 1 ) the warm populations probe a wider range in height , given the same selection function ; ( 2 ) the cold populations can pin down the trajectories / orbits better .
we find that in our case , the latter effect slightly dominates the former .
however , the results should be viewed with caution .
the results could be biased because ( 1 ) our model might be too restrictive , ( 2 ) the selection function that we choose ( @xmath148kpc ) is larger than the galactic scaleheight ; therefore , there is not much gain in probing a wider range in height . with these caveats in mind ,
we find that colder populations are more valuable in the study of the galactic potential with our proposed method .
we demonstrate a way to constrain the galactic potential from mono - abundance stellar populations .
we show that the phase - space distribution properties of mono - abundance stars can be fitted very well with a simple action - based df . assuming that the vertical and radial velocity dispersion scalelengths are the same , these properties are fully determined by only four df parameters .
we further show that , assuming that the proposed df is the true representation of the mono - abundance populations and certain parametrization of the galactic potential , we devise a statistical and rigorous method to measure the galactic potential parameters .
this is achieved by calculating the likelihood of observational data , given a joint set of parameters for both the df and the gravitational potential and subsequently marginalizing over the df parameters .
@xcite perform a rigorous test on the method we propose in this paper and confirm that our method has significant advantages on both the uncertainty in the potential parameter estimation and the computational time .
we create a mock data sample from the df with various potentials , including the miyamoto nagai in - plane potential and a flattened isothermal potential .
we show that even with a mono - abundance sample of the size of a few thousands , given a two - parameter gravitational potential , we can pin down the potential parameters to a few per cent . in this study , we assume that the measurement is perfect without uncertainty in the position and velocity measurements
. the inclusion of measurement uncertainty should be conceptually straightforward .
although a more realistic potential , for instance a combination of halo , bulge and disc potential , will contain more than two parameters and hence post a challenge to the estimation , different mono - abundance stars are tracing the same galactic potential and therefore their joint constraints with the method we propose are still very promising . in the current available segue survey , considering an elemental abundance bin of [ fe / h]@xmath149 and @xmath150=0.05 $ ] , we have about @xmath151@xmath152 tracers per bin . in the future ,
when galah is fully operational , we expect to have about @xmath153 tracers per bin . in practice ,
when dealing with multiple mono - abundance populations , one can combine the constraints from different populations by multiplying the individual likelihoods marginalized over the df parameters .
the assumptions of axisymmetry , of dfs that are smooth in the actions and no time variation remain important limitations in our current study . to tackle some of these aspects ,
we have started analysing data from the large @xmath154-body simulation described in @xcite , which has strong , time - dependent spiral structure . as a preliminary result
, we find that , even with realistic levels of non - axisymmetry and time variation of the potential , the method still works , but we plan to describe this in a later paper for a more concrete and rigorous analysis of this problem .
we thank the anonymous referee for helpful comments .
we also thank elena donghia for providing data from the simulation in @xcite .
yst and jb are grateful to the max - planck - institut fr astronomie for its hospitality and financial support during the period in which this research was performed .
jb was supported by nasa through the hubble fellowship grant hst - hf-51285.01 from the space telescope science institute , which is operated by the association of universities for research in astronomy , incorporated , under the nasa contract nas5 - 26555 .
jb was also partially supported by sfb 881 funded by the german research foundation dfg .
99 binney j. , 2010 , mnras , 401 , 2318 binney j. , 2011 , pramana , 77 , 39 binney j. , 2012 , mnras , 426 , 1324 binney j. , mcmillan p. , 2011
, mnras , 413 , 1889 binney j. , tremaine s. , 2008 , galactic dynamics .
princeton univ . press ,
princeton , nj bovy j. , tremaine s. , 2012 , apj , 756 , 89 bovy j. , rix h.w . ,
, 2012a , apj , 751 , 131 bovy j. , rix h.w . , liu c. , hogg d.w . , beers t.c .
, lee y.s . , 2012b , apj , 753 , 148 bovy j. , rix h.w . , hogg d.w . , beers t.c . , lee y.s . ,
zhang l. , 2012c , apj , 755 , 115 donghia e. , vogelsberger m. , hernquist l. , 2013 , apj , 766 , 34 dehnen w. , 1999 , aj , 118 , 1201 eisenstein d. et al . , 2011 , aj , 142 , 72 freeman k.c . , 2010 , in block d.l .
, freeman k.c . , puerari i. , eds , galaxies and their masks .
springer - verlag , berlin , p. 319
freeman k.c . , 2012 , in aoki w. , ishigaki m. , suda t. , tsujimoto t. , arimoto n. , eds , asp conf .
galactic archaeology : near - field cosmology and the formation of the milky way .
pac . , san francisco , p. 393
freeman k.c . ,
bland - hawthorn j. , 2002 , ara&a , 40 , 487 fuchs b. et al .
, 2009 , aj , 137 , 4149 ghez a.m. et al . , 2008 , apj , 689 , 1044 gillessen s. , eisenhauer f. , trippe s. , alexander t. , genzel r. , martins f. , ott t. , 2009 , apj , 692 , 1075 gilmore g. et al .
, 2012 , the messenger , 147 , 25 kuijken k. , gilmore g. , 1989 , mnras , 239 , 571 magorrian j. , 2006 , mnras , 373 , 425 magorrian j. , 2013 , arxiv:1303.6099 mcmillan p.j . , 2011 , mnras , 414 , 2446 mcmillan p.j . , binney j. , 2012 , mnras , 419 , 2251 mcmillan p.j . , binney j. , 2013 , mnras , arxiv:1303.5660 minchev i. , famaey b. , quille a.c . , dehnen w. , martig m. , siebert a. , 2012 , a&a , 548 , a127 miyamoto m. , nagai r. , 1975 , pasj , 27 , 533 reid m. j. , brunthaler a. , 2004 , apj , 616 , 872 siebert a. et al . , 2008 , mnras , 391 , 793 solway m. , sellwood j.a . ,
schnrich r. , 2012 , mnras , 422 , 1363 su m. , finkbeiner d.p . ,
2012 , arxiv:1207.7060 ting y .- s .
, freeman k.c . , kobayashi c. , de silva g.m . , bland - hawthorn j. , 2012 , mnras , 421 , 1231 yanny b. et al . , 2009 ,
apj , 137 , 4377 | we present a rigorous and practical way of constraining the galactic potential based on the phase - space information for many individual stars . such an approach is needed to dynamically model the data from ongoing spectroscopic surveys of the galaxy and in the future _ gaia_. this approach describes the orbit distribution of stars by a family of parametrized distribution function ( df ) proposed by mcmillan and binney , which are based on actions .
we find that these parametrized dfs are flexible enough to capture well the observed phase - space distributions of individual abundance - selected galactic subpopulations of stars ( ` mono - abundance populations ' ) for a disc - like gravitational potential , which enables independent dynamical constraints from each of the galactic mono - abundance populations .
we lay out a statistically rigorous way to constrain the galactic potential parameters by constructing the joint likelihood of potential and df parameters , and subsequently marginalizing over the df parameters .
this approach explicitly incorporates the spatial selection function inherent to all galactic surveys , and can account for the uncertainties of the individual position velocity observations . on that basis , we study the precision of the parameters of the galactic potential that can be reached with various sample sizes and realistic spatial selection functions . by creating mock samples from the df ,
we show that , even under a restrictive and realistic spatial selection function , given a two - parameter gravitational potential , one can recover the true potential parameters to a few per cent with sample sizes of a few thousands .
the assumptions of axisymmetry , of dfs that are smooth in the actions and of no time variation remain important limitations in our current study .
[ firstpage ] galaxy : disc galaxy : fundamental parameters galaxy : halo galaxy : kinematics and dynamics galaxy : structure . |
it is widely believed that the release of gravitational energy from an disk of accreting matter onto a super massive black hole ( smbh ) generates a huge amount of energy in active galactic nuclei ( agns ) .
energy and matter can be ejected in the form of jets , which have relativistic velocities and emit non - thermal radiation .
the jets of radio galaxies , develop from the central engine beyond their host galaxy and are characteristically narrow .
very long baseline interferometry ( vlbi ) images of agns often reveal a `` radio core , '' a region of peak intensity generated by a compact component at the apparently upstream end of the jets .
the radio cores of blazars or bl lacertae objects manifest as standing shock points up to @xmath3@xmath4 @xmath2 ( schwarzschild radius ) distant from the black hole @xcite . in other objects , such as the galaxy
, the central engine is located within a few tens of @xmath2 from the 43 ghz radio core @xcite .
the result was obtained by measuring core shifts , defined as the positional change of the core at different frequencies caused by absorption of the base of jet .
the infinite - frequency extrapolation of the core position can be interpreted as the origin of the jet . @xcite
estimated the position of the central engine of by measuring the core shifts at varying frequency and observing an exterior source as a phase reference .
the absorption processes that cause core shifts are synchrotron self - absorption ( ssa ) and free free absorption ( ffa ) . in ssa
, the core is located at the position of unity opacity .
ssa opacity depends on physical conditions of jet , such as the luminosity , structure , velocity , and inclination , etc . of the jet , as well as the observed emission frequencies .
besides , this type of core shift has been measured in @xcite , @xcite , @xcite , among others @xcite .
core shifts arising from ffa potentially provide information on the size , density , and temperature of the absorbing material .
ffa absorbers existing within sub - pc distances have been suggested in some objects such as @xcite , @xcite , @xcite and @xcite .
core shifts have also been reported in and @xcite .
measurements of core shifts and studies of the distributions of a spectral index require an unambiguous reference position when comparing core positions at different frequencies and overlaying images collected at different frequencies .
previous studies performed properly using optically - thin jet components as reference positions ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) or an exterior source by phase - referencing technique ( e.g. , * ? ? ?
* ; * ? ? ?
however , some aspects of core shifts remain elusive . although the extrapolation of core positions to infinite frequency is supposed to locate a central engine , it is a working hypothesis
. only we can claim rigidly is that a central engine and a black hole exist somewhere upstream of observed cores , as far as measurements perform on single - sided jet sources .
previously , core shift measurements both in approaching and counter jets have been attempted for ngc 1052 under the lack of a common reference position over observing frequencies ; however , multiple candidates for the location of central engine are discussed @xcite .
a core position was frequently assumed as the reference position in previous studies , for example , when spectral index maps were made .
however , as discussed above , this may alter with frequency because of absorption . to resolve these problems
, we supplement the analysis of a two - sided jet source with phase - referencing observations .
is a suitable agn candidate on which this technique is applied .
( 3c 270 ) is a nearby fanaroff
riley class i ( fr - i ) radio galaxy .
it is characterized by nearly symmetric kiloparsec - scale two - sided jets @xcite .
the galaxy is 31.6 mpc distant from the earth @xcite , corresponding to 0.15 pc / mas .
this proximity presents an advantage for studying regions within 1 pc of the central engine .
the nucleus is known to contain a central black hole of mass @xmath5 @xcite .
hubble space telescope ( hst ) observations have also revealed a disk of gas and dust , having a diameter of 300 pc , in the nucleus @xcite .
another feature of , revealed by vlbi studies , is its prominent parsec - scale two - sided jet along large - scale jets , running parallel to the rotation axis of the dust disk @xcite .
the west- and east - side jets are approaching and receding to observers , respectively ; the latter is known as a counter jet .
the brightness of both jets is slightly affected by relativistic beaming . @xcite estimated the speed of the jets by very long baseline array ( vlba ) monitor observations at 8 ghz . from the jet - to - counter jet brightness ratio and apparent speed
, they obtained a gentle intrinsic jet velocity of about @xmath6 , inclined at @xmath7 from the line of sight .
is further characterized by an apparent discontinuity in the structure of the counter jet .
this region , called the `` gap '' , has been interpreted as obscuration by an edge - on , geometrically thin , cold parsec - scale disk with a temperature of @xmath8 k @xcite . as an alternative to the ffa disk model , the gap may arise from resistive radiatively inefficient accretion flow @xcite . according to @xcite ,
the low temperature of the standard disk model @xcite precludes the existence of ffa plasma gas in a gap region so distant from the central black hole .
the present study aims to determine the position of the central engine in and investigate the structure of the absorber . in section 2
, we explain the phase - referencing method from multi - frequency observations and core shift measurements .
section 3 is devoted to core shift results and spectral index maps .
the position of the central engine , the interpretation of the core shift , the absorption process and the disk model of are discussed in section 4 .
section 5 concludes the paper .
multi - frequency observations of were conducted using the vlba @xcite at four frequency bands ( 1.4 , 2.3 , 5.0 , and 8.4 ghz ) on july 3 , 2003 , and at three frequency bands ( 15 , 22 , and 43 ghz ) on june 26 , 2003 .
the two sets of observations were separated by only 7 days .
each observation period was approximately 12 hours . at 1.4 ghz , dual - circular polarizations at two intermediate frequencies ( if ) were received , covering a total bandwidth of 16 mhz .
at both 2.3 and 8.4 ghz , two ifs covered a total bandwidth of 16 mhz in right - hand circular polarization .
other frequency bands collected four ifs , covering a total bandwidth of 32 mhz in left - hand circular polarization . digital sampling of 2-bit was applied at all observational frequency bands .
correlations were processed using the vlba correlator in socorro , new mexico .
the radio core positions of were measured by the nodding - style phase - referencing technique , relative to a nearby reference source , j1222 + 0413 , separated by less than @xmath9 from ; -t - c1-t - c1-t- , where t and c1 represent the target and primary calibrator , respectively .
switching between the target and calibrator was repeated after scanning for 20 s. observations were conducted for approximately 20 minutes at each frequency in turn . under this measurement condition ,
uv_-coverage among frequencies was almost identical , thereby minimizing the residual systematic errors in astrometry .
the total on - source time on the target was approximately one hour at each frequency .
the antenna parallactic angles and earth orientation parameters were corrected .
furthermore , ionospheric dispersive delays were corrected using the ionospheric model from gps data provided by the jet propulsion laboratory @xcite .
instrumental delays / phases and bandpass were calibrated using data . assuming a point source model
, we performed fringe - fitting on j1222 + 0413 as a phase calibrator c1 .
subsequently , we calibrated the visibility amplitude and phase using a source structure model available in the difmap software @xcite procedures .
this model is based on iterative deconvolution and self - calibration algorithms . applying these solutions to the visibilities of
, we obtained phase - referenced images of . on the other hand , we also made self - calibrated images for ngc 4261 using the difmap ( figure [ fig : all_cln ] ) .
the source position in the phase - referenced images was measured using jmfit in aips @xcite , adopting a single - ellipse gaussian profile model .
however , we consider that , because the core and jets were blended , the core position potentially drifted toward the jet side from its true position .
the jet contribution was evaluated by comparing results by two methods of measuring the core position in final ( self - calibrated ) images of ngc 4261 and j1222 + 0413 ; one - component jmfit and fitting of visibility data to multi circular gaussian components using modelfit in difmap ( figure [ fig:8 g ] ) .
the final core position was defined as the positions between the core in the phase - referenced image of ngc 4261 with the correction of jet contribution on its final image and the core of j1222 + 0413 in the final images with the correction of jet contribution . following the method of @xcite
, we estimated the error in our astrometric measurements .
the error incorporates ionospheric residuals , tropospheric residuals , antenna positional error , earth orientation error , a priori source coordinate error , core identification error , and thermal noise .
the error contributions are listed in table [ tab : eb ] .
based on gps data , we assume a total electron content of @xmath10 , which is reducible to 1/4 equivalent following ionospheric residual correction by tecor in aips @xcite . the zenith excess path error in tropospheric residuals is assumed to be 2 cm @xcite .
we adopted proper geometric errors such as antenna location error ( 3 mm ) , earth orientation parameter error ( vlba memo 69 ) , and a priori source coordinate error that is determined from source declination @xcite .
the core identification error was estimated from the difference in core positions yielded by the above - described two fitting methods .
the thermal noise error was defined by the beam size over the dynamic range of phase - referenced maps .
figure [ fig : all_cln ] shows the ( self - calibrated ) cleaned images from uniformly weighted visibilities at each frequency ; the corrections of relative positions in right ascension , which were determined by phase - referencing ( section [ section : corepositionmeasurement ] ) , have been applied in figure [ fig : maps ] .
the beam sizes and image qualities are listed in table [ tab : ip ] .
approaching jets and counter jets , extending west and east from the brightest region ( core ) , respectively , appear in the images at @xmath11 15 ghz clearly .
the counter jet is not so prominent in the image recorded at 22 ghz , while no obvious jet - like structures are visible in the 43 ghz image .
however , both approaching and counter jets were clearly detected as model components of modelfit in difmap .
the lower the frequency , the more extended and symmetric the jets .
the determined core positions at 1.443 ghz are plotted on the 5 ghz self - calibrated image of ngc 4261 in figure [ fig : cp ] . the core appears to shift northwestward as the frequency decreases .
the jet of the reference source j1222 + 0413 is oriented nearly perpendicular to that of ( toward the south , whereas the ncg 4261 jet tends to the west ; see figure [ fig : calib ] ) . for this reason ,
the observed core shifts comprise the sum of core shifts of the both sources .
therefore , we decomposed the observed core shift into a westward shift for and a northward shift for j1222 + 0413 .
the east west structure of s jet appears slightly fluctuating in north
south direction . on the approaching side
, the jet gradually inclines toward south beyond @xmath12 mas from the 43 ghz core .
we analyzed the ridge line in the 5-ghz image and measured its deviation to be @xmath13 mas southward at @xmath14 mas where the core at 1 ghz is found .
this southward deviation is used to correct the core position at 1 ghz for the final image to be used for making a spectral index map ( section [ sec : spix ] ) . on the other hand , in the innermost region
, the jet shows an s - shaped structure at 1543 ghz images .
the measured position angle of jet is @xmath15 in the 43 ghz image , while @xmath16 up to 1.5 mas from the core in the 15 ghz image .
the amplitude of the s - shape is only @xmath17as , which is much smaller than the scale of core shifts in j1222 + 0413 and ngc 4261 .
we do not take into account this north
south deviation for the final results of the core shift and the spectral index maps .
figure [ fig : cp ] reveals that the core position of shifts at all other observational frequencies relative to the 43 ghz core . by least - squares fitting , the core position is found to shift as a power - law function of frequency ; that is @xmath18 where @xmath19 , @xmath20 , @xmath21 , and @xmath22 are projected ( east west ) distance from the 43 ghz core , free - fitting parameters representing a scale factor , a power low index , and an asymptotic limit at infinite frequency , respectively ( solid curves in figure [ fig : cp ] ) . at the approaching jet side , these values are determined as @xmath23 , @xmath24 , and @xmath25 , respectively , with residuals @xmath26/n.d.f = 0.16 ( see table [ tab : cs ] ) .
this reduced @xmath26 is very small because we estimated the core positional errors conservatively .
the core positional errors could be actually smaller than our estimations . as a result ,
the free - fitting parameters , @xmath20 , @xmath21 and @xmath22 , are determined with an uncertainty of 10% , 5% and 20% , respectively .
the value of @xmath21 suggests a possible excess over the theoretical value of unity predicted for ssa under the equipartition condition and as frequently observed @xcite .
the asymptotic parameter @xmath22 was determined within a tolerance of 32 @xmath27as ( @xmath28 120 @xmath2 ) and separated by 82 @xmath27as ( @xmath28 310 @xmath2 ) from the 43 ghz core position .
we also measured and fitted the core shift at the counter jet side of . in this study ,
the core of the counter jet is defined as the closest east - side component to the core of the approaching jet , the region of maximum intensity yielded by modelfit . as shown in figure [ fig : cs ]
, the distance between the cores in the approaching and counter jets reduces as the frequency increases .
when the shifting core of the counter jet was fitted to a three - parameter model , we obtained @xmath29 , @xmath30 , and @xmath31 , respectively .
however , these errors have little contribution for confining the position of the central engine .
this suggests that a simple power - law is insufficient to represent the profile of counter - jet side .
hence , we applied two alternative model fitting procedures imposing further assumptions , which will be discussed in section [ sec : cs ] .
spectral index ( @xmath32 , @xmath33 ) maps were generated from final ( self - calibrated and cleaned ) images between two adjacent frequencies using the aips task comb . to ensure proper construction of spectral index maps , the images between the adjacent frequencies must spatially align , and their angular resolutions must match .
the map coordinates at each frequency were aligned using the core shift relative to the 43 ghz core in right ascension and the jet axis of in declination .
the images at adjacent frequencies were convolved by rendering both beams equivalent to the beam at the lower frequency . the intensity below 3@xmath34
was clipped out before calculating the spectral index .
all of the spectral index maps exhibit common features , in particular , the spectral index peaks in the regions between the cores of the approaching and counter jets .
as the jet extends further from the core , the spectral index becomes smaller .
figure [ fig : all_spx ] displays the spectral index maps of the 12 ghz and 58 ghz pairs . some regions show @xmath35 , which is the theoretical limit of ssa .
this result is consistent with other observations in which the circumnuclear structure favored ffa over ssa @xcite .
the spectral index gradients are clearly parallel to the jet axis for all frequency pairs .
the restored beams in each map are too large to decompose the absorption structures across the jet .
however , a transverse gradient marginally appear at 12 ghz and 25 ghz spectral index maps
. we can not rule out that the combination of the north south fluctuation in the jet structure ( section [ sec : coreshiftmeasurement ] ) and remaining position errors in east west direction causes the apparent transverse gradient .
figure [ fig : slice ] shows slice profiles of spectral index along the jet axis ( from the east to the west ) .
we estimated error components in the spectral index .
we tentatively displaced one image eastward and westward with respect to the other image by our conservative positional error , in order to evaluate a spectral index error due to the positional error in overlaying two images .
we also used the spectral error map , simultaneously produced during comb processing , and assumed a conservative flux calibration error of 5% at @xmath36 ghz and 10% at @xmath37 ghz .
the total error represented by an envelope is the square root of the summed squares of these error components . at 815
ghz and lower frequency pairs , a region of @xmath35 appears in the slice profiles , suggesting an ffa mechanism , although the error is too high for a conclusive assessment .
another feature is that the curve at @xmath38 5 ghz peaks at the counter jet side .
the peaks are located at the same positions , regardless of the error - induced displacement of the core position . at @xmath39 5
ghz , the peaks are located around the origin .
first , we consider the relationships between the black hole positions and limiting core positions .
@xcite showed that the central engine in is located @xmath40 distant from the observed core at 43 ghz in their analyses of core shift measurements .
this study assumed that the asymptotic line , representing the origin of a jet , spatially coincides with the central engine .
however , the core location in , indicated as the standing shock point @xcite , is actually @xmath3@xmath41 from the black hole , as determined from @xmath42-ray monitoring observations @xcite . on the other hand
, a core shift was discovered downstream of the jet .
core shift measurements of and other blazar sources revealed that the base of the jet is located @xmath43 from the 43 ghz core position @xcite .
recently , core shift data have accumulated at an accelerated pace for blazars @xcite . however , the above - cited examples highlight the difficulty of proving assumptions from one - sided jets . to overcome these limitations , we located the central engine in by investigating a double - sided system and measuring the core positions of approaching jet and counter jet .
as shown in figure [ fig : large ] , the core ( most upstream component ) at the counter jet side resides at @xmath44 mas ( @xmath45 ) from the approaching - jet core at 43 ghz .
the central engine and black hole must be situated at most 750 @xmath2 distant from this core to ensure its location between the cores of both jets . moreover , because the origin of the counter jet coincides with that of the approaching jet ( as discussed later ) , the assumption that the origin of a jet spatially coincides with the central engine within the error tolerance is reasonable for .
the center of activities were discussed using cross - identificatiton of the gaussian component models in multifrequency maps in other two - sided jet systems ( e.g. , ; @xcite and ; @xcite ) . for , a relatively distant source ,
the core shift in the counter jet has not been measured .
@xcite determined the position of the central engine in ngc 1052 with an uncertainty of @xmath46 pc , corresponding to @xmath47 assuming @xmath48 @xcite , from the both sides of jets .
our result for ngc 4261 is of higher accuracy and includes coordinates relative to 1222 + 0413 rather than relative to optically thin components in the source itself .
the central engine of is located at @xmath49as , corresponding to @xmath50 from the 43 ghz core .
central engine separation differs by orders of magnitude among , , and , being larger and smaller in and , respectively , than in .
the separation caused by a core shift is correlated with the strength of ssa , and depends on jet characteristics such as luminosity , velocity , structure , radial profiles of density and magnetic field , the geometric relationship to our line of sight , velocity gradients , and/or their time variation . according to @xcite ,
if a jet is continuous , uniform , stationary , and travels with constant velocity , and under equipartition condition , the ( projected ) core shift from the jet origin is given by @xmath51^{2/3 } , \label{equ : lob } \ ] ] where @xmath52 , and @xmath53 , @xmath54 , @xmath55 , and @xmath56 denote the lorentz factor , the intrinsic jet velocity , jet opening angle from an observer ( intrinsic opening angle @xmath57 ) , and viewing angle , respectively . @xmath58 and @xmath59 are the luminosity distance to the source ( parsec ) and synchrotron luminosity ( erg s@xmath60 ) . in this case
, @xmath61 depends on the jet structure , whether conical or parabolic , and does not usually exceed ln 3 for parsec - scale jets .
the frequency @xmath62 is given in hz .
in the present study , the separation between the central engine and core position is determined at 43 ghz . using equation ( [ equ : lob ] )
, we now compare the separation in to those estimated in and . in place of @xmath59 , we use the intrinsic luminosity at millimeter wavelengths , @xmath63 . in order to remove beaming effect from the observed luminosity @xmath64
, we calculate the beaming factor @xmath65 , where @xmath66^{-1}$ ] is the doppler factor , using the viewing angle and intrinsic velocity for each source . using @xmath67 , @xmath68 for @xcite and @xmath69 erg s@xmath60 @xcite , we estimated a beaming factor of 1.5 and @xmath70 erg s@xmath60 , assuming @xmath71 .
the parameters of , , and are provided in table [ tab : core43 ] . as a rough discussion , assuming that the jet opening angle is nearly constant under the equation ( [ equ : lob ] ) , the ratio of the core
central engine ( deprojected ) separation to that of in @xmath2 unit is given by @xmath72 for m87 case , the viewing angle of the jet is likely @xmath73@xmath74@xcite and apparent jet velocities of @xmath75@xmath76 have been reported for the innermost region at 43 ghz @xcite .
under these situation , the intrinsic velocity would be @xmath77@xmath78 .
assuming @xmath79 , @xmath77 and @xmath80 erg s@xmath60 @xcite , we obtain @xmath81 erg s@xmath60 . adopting a black hole mass of @xmath82 @xcite
, we obtain @xmath83 . for @xmath84 and @xmath85
, we obtain @xmath86 erg s@xmath60 and @xmath87 . on the other hand , based on the core shift measurements , the 43-ghz core
central engine distance ratio of @xmath88 ( ) to @xmath89 ( ) is 0.07 .
the difference between the two results is not so large .
the discrepancy may originate in uncertainties of the black hole mass or the intrinsic luminosity .
another possibility is the difference of opening angles between different sources . from equation ( [ equ : lob ] )
, the separation is inversely proportional to the opening angle . if the opening angle of is widely larger ( cf .
* ) than that of it can explain the difference of the separation between them .
by contrast , the core shift in substantially differs from that of .
the estimated jet velocity and inclination are @xmath90 and @xmath91@xmath92 , respectively @xcite .
such conditions should enhance the relativistic beaming effect , and indicate a beaming factor of @xmath93 .
since the observed luminosity was reported to be @xmath94 erg s@xmath60 , we estimate an intrinsic luminosity as @xmath95 erg s@xmath60 .
the mass of the black hole associated with is estimated to be @xmath96 @xcite .
substituting these value into equation ( [ equ : ratio ] ) , we obtain @xmath97 as an expected core shift , showing a large discrepancy from the result from @xmath42-ray monitoring observations ( @xmath983000 ) @xcite .
to explain the large discrepancy , a much larger intrinsic luminosity of @xmath99 erg s@xmath60 or a much smaller opening angle @xmath100 deg is necessary .
however , such a wide difference in intrinsic luminosity between and is quite unlikely if both fr - i radio galaxies and objects are regarded as a common agn population intrinsically @xcite .
more likely , the appearance of fr - i radio galaxies and objects differ because of the beaming effect imposed by jet inclination .
the jet of has a limb - brightened structure , which is probably associated with a fast spinal flow surrounded by a slow layer ( sheath ) @xcite .
if the jet axis inclines significantly from the line of sight , for such as and as fr - is , sheath emission dominates because spinal emission is deboosted by the relativistic effect . on the other hand ,
if the jet is viewed from nearly pole - on , for such as , spine emission from a small opening angle dominates .
this phenomenon can explain the varying distance between the central engine and the 43 ghz core between fr - i radio galaxies and objects .
the physical causes of the core shift are considered as ssa and/or ffa .
the core shift measurement for a two - sided system allow us to utilize a symmetry assumption for understanding the jet structure and surrounding circumstance , and offers a critical advantage over studies for one - sided systems .
only the counter jet itself could not well describe its core shift profile ( section [ sec : coreshiftmeasurement ] ) . in the following subsections
we establish the model that best explains the core shift of , focusing on the results of fitting parameters in equation ( [ equ : cs ] ) .
first , we assume a pure ssa model .
the physical conditions in the jet reflect the index of the core shift profile in equation ( [ equ : cs ] ) . in this model , @xmath101 ( representing @xmath21 for ssa case )
is given by @xmath102 where @xmath32 , @xmath103 and @xmath104 denote the spectral index at the optically - thin regime , the profiles of the magnetic field @xmath105 ( where @xmath106 ) , and the electron density @xmath107 ( where @xmath108 ) , respectively @xcite .
assuming @xmath109 and @xmath110 as the profile indices , corresponding to the stabilization of jet energy in equipartition , @xmath111 invariably becomes unity regardless of the spectral index value .
however , the fitting parameter at the approaching jet side is @xmath112 , which possibly exceeds that from previous studies for other sources ( @xmath113 ; i.e. , ; @xcite , ; @xcite , and ; @xcite ) , indicates that @xmath103 and @xmath104 in equation ( [ equ : k_ssa ] ) can take different values in .
we next predict the core shift profile at the counter jet side assuming the ssa model . if the jets are intrinsically symmetric , @xmath111 and @xmath22 of the counter and approaching jets side should be identical . on the other hand , comparing equations ( [ equ : cs ] ) and ( [ equ : lob ] ) , we find that @xmath20 depends on geometrical and intrinsic parameters of a jet .
the parameter @xmath20 differs between the approaching and counter jets only through the inclination , @xmath56 .
therefore , @xmath20 of the counter and approaching sides , denoted @xmath114 and @xmath115 , respectively , must satisfy @xmath116 .
in other words , the position of the core at the same frequency is farther from the black hole in the approaching jet than in the counter jet .
the parameter @xmath117 should be connected with @xmath118 by a difference of 180 in terms of viewing angle .
the curve predicted by the pure ssa model ( dashed line in figure [ fig : cs ] ) can explain the core position at both low frequency ( @xmath11 2 ghz ) and high frequency ( @xmath119 22 ghz ) .
on the other hand , the observed positions of the core at 5 , 8 , and 15 ghz are significantly larger than those predicted by this model , implying that the ssa model alone lacks sufficient absorption to explain the core shift in these regions .
another alternative is the ffa model .
the spectral index maps indicate that ffa plasma exists at least in regions near the putative black hole . for simplicity , and from the parameters of the core shift measurement , we assume a spherical ffa plasma ( see , appendix [ sec : app_ffa ] ) . in this case
, @xmath21 depends on the structure of the spherical ffa plasma . when the ffa opacity @xmath120 becomes unity , @xmath121 ( representing @xmath21 in equation ( [ equ : cs ] ) for the ffa model )
is given by @xmath122 where @xmath123 and @xmath124 represent the profiles of the electron density @xmath125 and temperature @xmath126 ( @xmath127 , @xmath128 ) in the ffa absorber , respectively .
these parameters can be determined from the opacity profile ; the denominator in the right side of equation ( [ eq : kf ] ) represents the index value of the opacity profile ( @xmath129 ) .
we obtain the radial profile of ffa opacity at the approaching jet side by fitting to a power - low function ; @xmath130 ( appendix [ sec : app_ffa ] ) . substituting this value into equation ( [ eq : kf ] ) ,
we obtain @xmath131 .
it is consistent with the result of the core shift measurements ( @xmath132 ) , in spite of independent measurements of the core shift and the opacity .
theoretical speculations have indicated that @xmath133 exceeds unity @xcite .
thus , the core shift at approaching also jet side can also be sufficiently interpreted by the spherical ffa model .
however , the spherical ffa is inconsistent with the radial profile of ffa opacity on the counter jet side .
the parameter @xmath133 should be identical in both the approaching and counter sides if the ffa absorber is symmetric .
@xmath20 reflects the path length of the absorber ; that is , the geometrical structure of the absorber ( for details , see appendix ) . if the absorber covers the counter jet side , @xmath20 must satisfy @xmath134 because the path length of the counter jet side is longer than that of the approaching one .
the core - shift curve predicted by the spherical ffa plasma model ( dot dashed line in figure [ fig : cs ] ) meets some of core positions at @xmath119 5 ghz on the counter jet side .
however , the core shifts at 1 and 2 ghz are inconsistent because their shifts are significantly smaller than those predicted by this model .
therefore , neither of these simple models can explain the core shifts in this two - sided system .
a possible solution is to combine the ssa core shift with an additional external absorber , which compensates for the inadequate core shift . indeed ,
our results show that the pure ssa model fitted fairly well on both approaching and counter sides at all frequency , except for the counter - jet components at 5 , 8 , 15 ghz . when the observed core shifts in the counter side is fitted without the cores at these frequencies , the best - fit values of @xmath135 and
@xmath136 ( representing the index and the asymptotic limit in this condition ) approximate those of the approaching side , as expected from ssa theory ( table [ tab : cs ] ) .
this suggests the presence of an absorber in a limited region .
thus , it appears that the cores at 5 , 8 , and 15 ghz are additionally affected by ffa .
ffa opacity @xmath120 at the positions of 222 ghz cores in the counter jet is determined by spectral analysis of multi - frequency images ( appendix [ sec : analysisffa ] ) and shown in figure [ fig : tau ] . on the 515 ghz core , @xmath120
is well represented by a power - law profile . on the other hand ,
@xmath120 at other frequencies , where ssa alone can describe the core shift profile , should be @xmath137 .
such a limited range of ffa regions is consistent with a gap found in the intensity profile in the counter jet side in previous studies ( e.g. , * ? ? ? * ) and an obscuring disk proposed by them .
thus , the geometrically thin disk predicted is a candidate to explain the core shift of the counter side .
we consider the structure of a disk to explain the core shift and the profile of @xmath120 at the counter jet side .
figure [ fig : disk ] shows the geometric relationships between the jet and disk , assuming @xmath138 and the rotation axis of the disk is aligned with the axis of pc - scale jet . as discussed in the combined model ,
if the disk plasma cause the core shift at 515 ghz , the density and temperature of the disk require the radial gradient ( @xmath139 , @xmath140 , where @xmath141 is the radius of the disk in @xmath2 unit , @xmath142 ) .
in addition , the thickness of the disk and the path length through the disk to observers are also described as @xmath143 and @xmath144 . by substituting @xmath125 , @xmath145 and @xmath146 into equation ( [ eq : tau_f ] )
, we obtain @xmath120 by the disk ( hereafter @xmath147 ) ; that is @xmath148 in the figure [ fig : tau ] , the least - squares fitting at at 515 ghz data reveals the profile of disk opacity to the power - law function of the disk radius ; that is @xmath149 , then we obtain @xmath150 to explain the small opacity at low frequency ( @xmath11 2 ghz ) , the low density , the short path length or high temperature are required because the ffa effects depend on frequency ( @xmath153 , see appendix [ sec : analysisffa ] ) . in these requirements , the condition of the path length and the temperature are physically unrealistic because the low frequency core positions correspond to the outer radii of the disk . on the other hand , it is realistic that the density is sufficiently small and/or the temperature of the plasma is too low to be ionized at the outer radii .
we consider the condition that the temperature at the radius corresponding to the 5 ghz core position and at the 2 ghz are higher and lower than @xmath3 k , respectively . assuming that @xmath154 predicted by standard disk model at the outer region @xcite and @xmath3 k at 5 ghz
, we obtain @xmath155 in addition , if @xmath156 ( wherever the path length is constant ) and @xmath157 pc , for simplicity , we obtain @xmath158 from equation ( [ equ : tau_ind ] ) .
it is very close to @xmath159 predicted by standard disk model . inserting these value and @xmath147 to equation ( [ equ : tau_disk ] ) , we obtain @xmath160 the density and temperature profile in this condition is shown in figure [ fig : disk ] .
we note that the assumptions of @xmath161 , and @xmath162 may be unsupported and thus other values are acceptable . substituting @xmath163 ( predicted by standard disk model ) into equation ( [ equ : tau_ind ] ) and @xmath164 pc ( corresponding to @xmath165 pc at the 5 ghz core ) , we obtain @xmath166 and @xmath167 @xmath168 .
although the index @xmath123 increases the gap to the value predicted by the model , @xmath125 at @xmath169@xmath170 is almost the same within a few times factor comparing to the case of @xmath156 . in the case of @xmath162 , @xmath171
is obtained from equation ( [ equ : tau_disk ] ) , indicating the thinner disk geometrically requires the higher density ( i.e. , @xmath172 @xmath168 at @xmath173 pc ) . however , if the disk thickness was much larger ( i.e. , @xmath174 pc ) , the disk would obscure the approaching jet as well as the counter jet .
thus , @xmath157 pc provide a reasonable upper limit , indicating the lower limit of @xmath125 .
on the other hand , this density profile can not explain the small opacity at the high frequency ( @xmath119 22 ghz , @xmath175 ) .
the interpretation under a simple pow - low profile requires a large @xmath124 such as @xmath176 , which leads to a sufficiently high temperature of plasma to reduce ffa effect at the inner region ( 22 ghz ) .
however , this is impractical because the disk temperature at 3 @xmath2 should be too high ( @xmath177 k ) .
the weak effects of ffa at the 22 ghz core indicate that the disk structure changes dramatically at a characteristic radius .
the riaf model is suitable for describing this situation of the low density at inner radii of an accretion .
in fact , x - ray observations suggest that a low luminosity and flux and spectral variations of the nuclear x - ray emission of ngc 4261 can be caused by an accretion disk with a low accretion rate with a low radiative efficiency such as riaf @xcite .
therefore , we propose the accretion flow of ngc 4261 has a two - component structure ; one is an inner , hot , optically thin , riaf and the other is an outer , truncated , cool , optically thick accretion disk .
this structure is considered to be not expedient for ngc 4261 but common for llagns based on broad spectral energy distribution ( sed ) and x - ray spectral studies @xcite . according @xcite , the transition radius , @xmath178 , in an accretion disk
is given by @xmath179 where @xmath180 and @xmath181 denote the black hole mass in units of solar masses , @xmath182 and the mass accretion rate in eddington unit , @xmath183 , @xmath184 g s@xmath60 .
inserting @xmath185 and @xmath186 to equation ( [ equ : xc ] ) , we obtain @xmath187 , requiring a significantly higher accretion rate compared to that of the typical llagns ( @xmath188@xmath189 ) .
recently , the equation of the transition radius is modified by the disk evaporation model taking account the effects of viscosity and magnetic field in accretion flow @xcite . in this case
, @xmath185 is easily acceptable for @xmath190 , which is an expected accretion rate for ngc 4261 .
consequently , this combined model of riaf and truncated disk is a possible solution to explain the core shift of at both side simultaneously .
figure [ fig : disk ] show this schematic illustration .
it is necessary to explain an observed large column density of @xmath191 , where @xmath192 represents the column density , toward the nucleus ( e.g. , * ? ? ?
* ) with an x - ray luminosity of @xmath193 ; the origin of the power - law x - ray emission is controversially associated either with the accretion flow or the base of the jet ( e.g. , * ? ? ?
* ; * ? ? ?
first , we consider the case that the x - ray absorber would be the outer thin disk . substituting @xmath194 in equation ( [ equ : n ] ) , the emission is passing through the region of a density of @xmath195 @xmath168 . in case of @xmath196 pc ,
we obtain @xmath197 @xmath198 indicating to be consistent with the x - ray observation .
however , the geometrically thin disk can obscure only the counter jet in this case .
it is inappropriate to make only the counter jet take responsibility for the x - ray emitting source .
second , we consider the case that the x - ray absorber would be the inner accretion flow . we adopt a simple adaf model @xcite with @xmath199 as a free - fall density profile and @xmath200 at @xmath201 .
the observed x - ray luminosity , suggesting an accretion rate of @xmath202 , roughly provides an electron temperature and restricts the amount of accreting material . in this condition , a resultant ffa opacity at @xmath175 at 22 ghz is significantly less than unity ( cf .
appendix ) , which is consistent with our vlbi result . however , total electron column density toward the nucleus ( down to 3 rs ) is at most @xmath203 through the accretion flow .
thus , only the accretion flow can not be responsible for the large column density observed at x - ray .
consequently , neither the outer thin disk or the inner riaf seems to be a sufficient x - ray attenuator ; an external absorber , such as a dusty torus , along our line of sight is required to explain x - ray observations .
we measured the origin of jets at the radio galaxy as the limit of the core position at infinite frequency using a core shift profile .
the origin is @xmath204 @xmath27as ( projected ) distant from the 43 ghz core position , corresponding to @xmath205 ( deprojected ) .
we also measured the core shift at the counter jet side and confirmed that the core positions at both sides asymptotically converge to the same limit .
therefore , we conclude that the jet origin estimated by core shift measurements spatially coincides with the central engine .
the spectral index maps reveal the possible presence of ffa , @xmath206 , in at least a limited area in the counter jet side .
the core shift profile of the counter jet is not solely attributable to synchrotron self - absorption and requires an additional absorbing component such as a geometrically thin disk .
moreover , the core shift and ffa opacity profiles from our result suggest a two - component accretion such as a combined model of riaf and truncated thin disk that is previously proposed for accretion flows of llagns to explain observed seds . the density and temperature profiles for the outer thin disk as free free absorber are estimated to be @xmath207 k and @xmath208 @xmath168 , respectively , in the case of @xmath157 pc . neither the outer thin disk or the inner riaf seems to be a sufficient x - ray attenuator ; an external absorber , such as a dusty torus , along our line of sight is required to explain an observed high column density toward the nucleus .
the very long baseline array is operated by the national radio astronomy observatory , a facility of the national science foundation , operated under cooperative agreement by associated universities , inc .
this work was partially supported by a grant - in - aid for scientific research ( b ; 24340042 , ad ) from the japan society for the promotion science ( jsps ) and the center for the promotion of integrated sciences ( cpis ) of sokendai .
lccccccc frequency [ ghz ] & 1.4 & 2.3 & 5.0 & 8.4 & 15 & 22 & 43 + phase - referencing observation & & & & & & & + beam width/ signal to noise & 52 & 24 & 10 & 6 & 4 & 3 & 2 + ionosphere & 2006 & 779 & 161 & 57 & 17 & 8 & 2 + troposphere & 13 & 13 & 13 & 13 & 13 & 13 & 13 + earth orientation & 5 & 5 & 5 & 5 & 5 & 5 & 5 + antenna position & 2 & 2 & 2 & 2 & 2 & 2 & 2 + apriori source coordinates & 1 & 1 & 1 & 1 & 1 & 1 & 1 + core identification analysis & & & & & & & + approaching jet & 349 & 60 & 17 & 40 & 3 & 5 & 3 + counter jet & 343 & 26 & 5 & 50 & 326 & 312 & 19 + total error ( rss ) & & & & & & & + approaching jet & 2055 & 792 & 170 & 77 & 30 & 23 & 19 + counter jet . &
2054 & 790 & 169 & 82 & 328 & 313 & 25 + cccccc @xmath62 & @xmath209 & @xmath34 & @xmath210&@xmath211 & @xmath212 + ( ghz ) & ( mjy / beam ) & ( mjy / beam ) & ( mas ) & ( mas ) & ( deg ) + ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) + 1.4 & 47.4 & 0.50 & 8.90 & 4.45 & 5.5 + 2.3 & 70.1 & 0.69 & 5.41 & 2.60 & 1.1 + 5.0 & 117.1 & 0.40 & 2.47 & 1.19 & 0.06 + 8.4 & 101.6 & 0.60 & 1.51 & 0.69 & 0.95 + 15 & 113.4 & 0.71 & 0.85 & 0.37 & -5.0 + 22 & 123.2 & 1.27 & 0.55 & 0.27 & -3.1 + 43 & 72.9 & 0.94 & 0.30 & 0.16 & -2.8 + ccc @xmath62 & @xmath213 & @xmath214 + ( ghz ) & ( mas ) & ( mas ) + ( 1 ) & ( 2 ) & ( 3 ) + 1.4 & @xmath215 & @xmath216 + 2.3 & @xmath217 & @xmath218 + 5.0 & @xmath219 & @xmath220 + 8.4 & @xmath221 & @xmath222 + 15 & @xmath223 & @xmath224 + 22 & @xmath225 & @xmath226 + 43 & @xmath227 & @xmath228 + ccccc & @xmath20 & @xmath21 & @xmath22 & @xmath26/n.d.f + ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) + @xmath19 & @xmath229 & @xmath230 & @xmath231 & 0.66/4 + cccccccc object name & @xmath232 & @xmath233 & @xmath234 & @xmath56 & @xmath235 & @xmath64 & @xmath236 + & [ @xmath2 ] & [ 10@xmath238 & & & & [ 10@xmath239 ergs@xmath60 ] & [ 10@xmath239 erg s@xmath60 ] + ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) & ( 8) + ngc 4261 & @xmath241 & 4.9@xmath242 & 0.46@xmath243 & 63@xmath244@xmath243 & 1.1 & 3.9@xmath245 & 2.5 + m 87 & @xmath246 & 60@xmath247 & 0.380.62@xmath248 & 15@xmath24425@xmath244@xmath249 & 1.41.7 & 5.2@xmath250 & 1.4 + bl lac & @xmath251 & 1.7@xmath252 & @xmath253 & 6@xmath24410@xmath244@xmath254 & 5.69.1 & 3.6@xmath255@xmath256 & 1.0 +
in this section , we estimate the distribution of ffa opacity along the jet axis ( from the east to the west ) . we determined an intensity profile along the jet by integrating the flux density of clean components in defined regions , at intervals of 0.25 mas and 1 mas inside and outside of 2 mas from the putative central engine , respectively . from the integrated flux densities , we then performed spectral fitting analysis in each area . the fitting function for spectral shape @xmath257 is described as @xmath258 where @xmath259 here , @xmath260 , @xmath261 and @xmath120 are the unabsorbed flux density at 1 ghz , the intrinsic spectral index of synchrotron emission , and ffa opacity , respectively ; @xmath125 , @xmath146 and @xmath145 are the electron density , path length , and temperature of the plasma , respectively . assuming that @xmath261 is fixed at some typical value ( @xmath262 ) , @xmath260 and @xmath120 were determined by least - squares fitting . figure [ fig : ffa_tau ] shows the determined distribution of ffa opacity @xmath120 . in addition , we determined the radial profile of ffa opacity at the approaching jet side by fitting to a power - law function ; @xmath263 where , @xmath264 , @xmath265 are free - fitting parameters .
the fitted parameters are listed in table [ tab : tau_ffa ] .
in this section , we consider an ffa - induced core shift using the radial profile of ffa opacity .
we assumed that the core position affected by ffa is emitting at the local maximum in the spectral shape .
the partial derivative of ( [ eq : s_nu ] ) with respect to @xmath62 is maximized at @xmath266 . substituting this @xmath267 into ( [ eq : tau ] ) and solving the equation for @xmath268
, we obtained the core shift profile at the approaching jet side . assuming the structure of the ffa plasma , the core shift profile at the counter jet side can be derived from the information of the approaching side . in this study , we presumed a simple spherical plasma ( see figure [ fig : ffa_model ] ) , which can obscure both the approaching jet and the counter jet .
if the electron density and temperature of the ffa plasma follow a power - law function , i.e. , @xmath269 and @xmath270 , the index of the opacity profile , @xmath265 should be identical in both the approaching and counter sides . when fitting the opacity profile at counter jet side , we fixed @xmath265 to the same value as that of the approaching jet and obtained @xmath264 .
table [ tab : tau_ffa ] provided with this fitting result .
let us consider the configuration as shown in figure [ fig : ffa_model ] .
substituting @xmath125 and @xmath145 into equation ( [ eq : tau_f ] ) , we integrate the path length in the plasma along the line of sight from the jet at @xmath271 .
the radius @xmath141 from the black hole is described in ( @xmath272 ) coordinate as @xmath273 .
therefore , the ffa opacity is given by adopting @xmath277 from the analysis of the opacity profile ( table.[tab : tau_ffa ] ) , the result of the integration is obtained in numerical calculations .
an asymmetric profile of the opacity between both of jets is caused by only the different viewing angles , corresponding to the integration ranges @xmath278 . even if @xmath279 and @xmath280 are unknown
, we can obtain the opacity profile of the counter jet by using the ratio of the opacity to that of the approaching jet .
acciari , v. a. , aliu , e. , arlen , t. , et al .
2009 , science , 325 , 444 agudo , i. , thum , c. , wiesemeyer , h. , & krichbaum , t. p. 2010
, , 189 , 1 albert , j. , aliu , e. , anderhub , h. , et al .
2007 , , 666 , l17 beasley , a. j. , & conway , j. e. 1995 , very long baseline interferometry and the vlba , 82 , 327 bietenholz , m. f. , bartel , n. , & rupen , m. p. 2004
, , 615 , 173 biretta , j. a. , sparks , w. b. , & macchetto , f. 1999 , , 520 , 621 birkinshaw , m. , & davies , r. l. 1985 , , 291 , 32 blandford , r. d. , knigl , a. 1979 , , 232 , 34 davies , m. l. , franzen , t. m. o. , davies , r. d. , et al .
2009 , , 400 , 984 ferrarese , l. , ford , h. c. , & jaffe , w. 1996 , , 470 , 444 gebhardt , k. , & thomas , j. 2009 , , 700 , 1690 geldzahler , b. j. , & witzel , a. 1981 , , 86 , 1306 ghisellini , g. , tavecchio , f. , & chiaberge , m. 2005 , , 432 , 401 gliozzi , m. , sambruna , r. m. , & brandt , w. n. 2003 , , 408 , 949 greisen , e. w. 2003 , information handling in astronomy - historical vistas , 285 , 109 hada , k. , doi , a. , kino , m. , et al .
2011 , , 477 , 185 jaffe , w. , ford , h. c. , ferrarese , l. , van den bosch , f. , & oconnell , r. w. 1993 , , 364 , 213 jones , d. l. , wehrle , a. e. , meier , d. l. , & piner , b. g. 2000 , , 534 , 165 jones , d. l. , wehrle , a. e. , piner , b. g. , & meier , d. l. 2001 , , 553 , 968 jorstad , s. g. , marscher , a. p. , mattox , j. r. , et al .
2001 , , 134 , 181 jorstad , s. g. , marscher , a. p. , lister , m. l. , et al .
2005 , , 130 , 1418 junor , w. , biretta , j. a. , & livio , m. 1999 , , 401 , 891 kaburaki , o. , nankou , t. , tamura , n. , & wajima , k. 2010 , , 62 , 1177 kadler , m. , ros , e. , lobanov , a. p. , falcke , h. , & zensus , j. a. 2004 , , 426 , 481 kameno , s. , sawada - satoh , s. , inoue , m. , shen , z .- q . , & wajima , k. 2001 , , 53 , 169 kharb , p. , shastri , p. , & gabuzda , d. c. 2005 , , 632 , l69 kovalev , y. y. , lobanov , a. p. , pushkarev , a. b. , & zensus , j. a. 2008 , , 483 , 759 krichbaum , t. p. , alef , w. , witzel , a. , et al .
1998 , , 329 , 873 lara , l. , alberdi , a. , marcaide , j. m. , & muxlow , t. w. b. 1994 , , 285 , 393 liu , b. f. , yuan , w. , meyer , f. , meyer - hofmeister , e. , & xie , g. z. 1999 , , 527 , l17 lobanov , a. p. 1998
, , 330 , 79 ly , c. , walker , r. c. , & junor , w. 2007 , , 660 , 200 mahadevan , r. 1997 , , 477 , 585 mannucci , a. j. , iijima , b. , sparks , l. , et al . 1999 ,
journal of atmospheric and solar - terrestrial physics , 61 , 1227 marscher , a. p. , jorstad , s. g. , darcangelo , f. d. , et al .
2008 , , 452 , 966 nandra , k. , oneill , p. m. , george , i. m. , & reeves , j. n. 2007 , , 382 , 194 napier , p. j. 1994 , very high angular resolution imaging , 158 , 117 narayan , r. , & yi , i. 1995 , , 452 , 710 nemmen , r. s. , storchi - bergmann , t. , yuan , f. , et al .
2006 , , 643 , 652 osullivan , s. p. , & gabuzda , d. c. 2009 , , 400 , 26 piner , b. g. , jones , d. l. , & wehrle , a. e. 2001 , , 122 , 2954 pradel , n. , charlot , p. , & lestrade , j .- f .
2006 , , 452 , 1099 pushkarev , a. b. , hovatta , t. , kovalev , y. y. , et al .
2012 , , 545 , a113 qiao , e. , liu , b. f. , panessa , f. , & liu , j. y. 2013 , , 777 , 102 shakura , n. i. , & sunyaev , r. a. 1973 , , 24 , 337 shepherd , m. c. , pearson , t. j. , & taylor , g. b. 1994 , , 26 , 987 sokolovsky , k. v. , kovalev , y. y. , pushkarev , a. b. , & lobanov , a. p. 2011
, , 532 , a38 sudou , h. , iguchi , s. , murata , y. , & taniguchi , y. 2003 , science , 300 , 1263 tingay , s. j. , & murphy , d. w. 2001 , , 546 , 210 tonry , j. l. , blakeslee , j. p. , ajhar , e. a. , & dressler , a. 2000 , , 530 , 625 urry , c. m. , padovani , p. , & stickel , m. 1991 , , 382 , 501 vermeulen , r. c. , readhead , a. c. s. , & backer , d. c. 1994 , , 430 , l41 walker , r. c. , romney , j. d. , & benson , j. m. 1994 , , 430 , l45 woo , j .- h .
, & urry , c. m. 2002 , , 579 , 530 worrall , d. m. , birkinshaw , m. , osullivan , e. , et al .
2010 , , 408 , 701 zezas , a. , birkinshaw , m. , worrall , d. m. , peters , a. , & fabbiano , g. 2005 , , 627 , 711 | we report multi - frequency phase - referenced observations of the nearby radio galaxy , which has prominent two - sided jets , using the very long baseline array ( vlba ) at 1.443 ghz .
we measured radio core positions showing observing frequency dependences ( known as `` core shift '' ) in both approaching and counter jets .
the limit of the core position as the frequency approaches infinity , which suggests a jet base , is separated by ( @xmath0 ) micro - arcsecond upstream in projection , corresponding to ( @xmath1 ) @xmath2 ( @xmath2 ; schwarzschild radius ) as a deprojected distance , from the 43 ghz core in the approaching jet .
in addition , the innermost component at the counter jet side appeared to approach the same position at infinity of the frequency , indicating that cores on both sides approach the same position , suggesting a spatial coincidence with the central engine . applying a phase - referencing technique
, we also obtained spectral index maps , which indicate that emission from the counter jet is affected by free free absorption ( ffa ) .
the result of the core shift profile on the counter jet also requires ffa because the core positions at 515 ghz can not be explained by a simple core shift model based on synchrotron self - absorption ( ssa ) .
our result is apparently consistent with the ssa core shift with an additional disk - like absorber over the counter jet side .
core shift and opacity profiles at the counter jet side suggests a two - component accretion : a radiatively inefficient accretion flow ( riaf ) at the inner region and a truncated thin disk in the outer region .
we proposed a possible solution about density and temperature profiles in the outer disk on the basis of the radio observation . |
the x - rays reprocessed on accretion disks to the optical , uv and x - ray bands in both x - ray binaries and agn could be one of our best probes of accretion disk structure and the environment around compact objects .
it is convenient to study an illuminated disk around neutron star low - mass x - ray binaries ( lmxb ) since the central continuum dominates the energetics of the disk atmosphere and accretion disk corona ( adc ) , and the system geometry is often determined .
measurements of the optical magnitudes and lightcurve amplitudes of lmxb yielded estimates of the angle subtended by the disk of @xmath0 12@xmath1 and of the average fraction of x - rays reprocessed by the disk into the optical of @xmath2 @xcite .
models that seem to correctly describe the accretion disk interior have been unable to produce disk geometries consistent with the optical data @xcite .
analytical models assuming an isothermal disk show better agreement with the data , yet produce illuminated disks that are significantly thinner than expected @xcite .
a pure continuum incident on the gaseous disk will photoionize and heat the gas .
radiative heating sustains a @xmath3 k adc and a smaller @xmath4-@xmath5 k atmosphere emitting x - ray lines on top of the optically thick disk .
x - ray spectra of lmxbs during dips show strong comptonized emission plus an excess flux at 0.65 kev consistent with the same covering fraction @xcite , positing that adc and disks may coexist .
the apparent size of the adc and the disk may not be equal @xcite , since the emission mechanisms have different radial dependence .
@xcite modeled the reprocessed continuum and fluorescence emission from the inner disk of agn with an adc , and @xcite modeled x - ray spectra of disk annuli in an lmxb with adc and normally incident radiation . in this paper
, we calculate the structure of an accretion disk illuminated by the grazing radiation of a neutron star in an lmxb with a compton - temperature adc , building on a model by @xcite .
this model , which we will detail in a future paper , includes a more accurate calculation of the vertical and radial disk structure than previous models , accounts for thermal instabilities , and incorporates recent atomic model calculations that predict x - ray spectra at a resolution superior to the _ chandra and _ xmm - newton spectrometers .
we obtain the first self - consistent flared disk geometry that matches the disk thickness expected from optical data , and we predict that broad x - ray lines will be observable and will provide , among other diagnostics , evidence for photo - evaporation , which is central to the origin of winds @xcite . _ _
we consider an lmxb with a @xmath6 @xmath7 primary radiating an eddington luminosity bremsstrahlung continuum with @xmath8 kev .
we use a set of fiducial system parameters for a bright lmxb , so application to a particular source will require using the observed x - ray continuum to improve accuracy .
the maximum radius of the centrally - illuminated disk is @xmath9 cm , so the orbital period @xmath10 day .
the minimum radius is @xmath11 cm , below which the omitted effect of radiation pressure will dominate .
we obtain the vertical disk atmosphere structure of each annulus by integrating the hydrostatic balance and 1-d radiation transfer equations for a slab geometry : @xmath12 @xmath13 @xmath14 while satisfying local thermal equilibrium : @xmath15 where @xmath16 is the pressure , @xmath17 is the density , @xmath18 is the incident radiation field , @xmath19 is the reprocessed radiation propagating down towards the disk midplane , @xmath20 the vertical distance from the midplane , @xmath21 the gravitational constant , @xmath22 the grazing angle of the radiation on the disk , @xmath23 the radius , and @xmath24 is the local absorption coefficient .
the difference between heating and cooling @xmath25 includes compton scattering , bremsstrahlung cooling , photoionization heating , collisional line cooling , and recombination line cooling from h , he , c , n , o , ne , mg , si , s , ar , ca , and fe ions . we assume cosmic abundances @xcite .
the code from @xcite computes the net heating and ionization equilibrium , compton scattering , and line scattering using escape probabilities .
a new disk structure calculation simultaneously integrates eqs .
[ eq : hydro]-[eq : rad2 ] by the runge - kutta method using an adaptive stepsize control routine with error estimation , and eq .
[ eq : thermal ] is solved by a globally convergent newton s method @xcite . at the adc height @xmath26 ,
the equilibrium @xmath27 is close to the compton temperature @xmath28 , from which we begin to integrate downward until @xmath29 .
the optically thick part of the disk , with temperature @xmath30 , is assumed to be vertically isothermal @xcite . to get @xmath30 , we assume that the viscous energy and illumination energy is all locally radiated as blackbody radiation . that is , for @xmath31 and @xmath32 : @xmath33 where @xmath34 is the neutron star radius , @xmath35 the accretion rate and luminosity , @xmath36 the stephan - boltzmann constant , and @xmath37 is the x - ray ` albedo ' derived from optical observations @xcite , such that @xmath38 is the fraction of x - rays absorbed at the photosphere . the height at which the integration ends is defined as the photosphere height @xmath39 .
thus , the assumption is that for @xmath40 viscous dissipation dominates heating .
the boundary conditions at the adc are set to @xmath41 , @xmath42 , and @xmath43 , where @xmath44 , @xmath45 is the boltzmann constant , and @xmath46 is the average particle mass in units of the proton mass @xmath47 .
the boundary conditions at @xmath39 for @xmath18 and @xmath19 are set free , and we use the shooting method @xcite with shooting parameter @xmath48 adjusted until we satisfy @xmath49 at the photosphere .
note @xmath50 is the viscosity ( @xmath51 ) dependent density calculated for an x - ray illuminated @xcite disk .
we calculate @xmath52 iteratively . to get @xmath30 self - consistently from equation [ eq : locdisk ] , we need @xmath53 .
we neglect the term @xmath54 , which is valid for @xmath55 cm . after an initial guess for @xmath56
, we define it as the height where @xmath57 is attenuated by @xmath58 .
we stop after @xmath52 and @xmath30 converge to @xmath59 % .
note @xmath60 for @xmath61 . since @xmath62 is not physically determined , but it is bound by @xmath63 , we ran a second model calculating @xmath52 from @xmath39 , which we will use to estimate systematic errors due to the 1-d radiation transfer calculation .
the disk atmosphere exhibits a thermally unstable region , which has direct consequences for the disk structure and x - ray spectrum .
the origin of the thermal instability is well understood , and its behaviour depends on metal abundances and the local radiation spectrum @xcite . within a range of pressure ionization parameters ( @xmath64 ) , thermal equilibrium
is achieved by three distinct temperatures , only two of which are stable to perturbations in @xmath27 ( fig .
[ fig1 ] ) .
@xmath65 , where @xmath66 is the radiation pressure and @xmath67 the gas pressure @xcite .
the thermally stable branches have positive slope @xcite .
previous spectral studies of illuminated accretion disks in lmxb had not explicitly selected the stable solutions @xcite , a choice which affects x - ray production .
the instability implies a large temperature gradient as the gas is forced to move between stable branches , requiring the formation of a transition region whose size may be determined by electron heat conduction , convection , or turbulence , depending on which dominates the heat transfer . for simplicity , we neglect emission from the transition region . upon calculation of the field length @xmath68 , the lengthscale below which conduction dominates thermal equilibrium @xcite , we estimate that conduction forms a transition layer @xmath69 times thinner than the size of the x - ray emitting zones
. nevertheless , x - ray line emission from the neglected conduction region may not be negligible in all cases @xcite .
conduction determines the dynamic behaviour of the gas .
the dynamics will allow us to attach a physical interpretation to the chosen solutions from fig .
roughly , a static conduction solution can only be found if the transition region splits the instability @xmath64-range in half ( fig .
[ fig1 ] ) .
this solution @xcite takes the low-@xmath27 stable branch at @xmath70 , and the high-@xmath27 stable branch at @xmath71 , separated by the transition layer at @xmath72 . a transition layer located away from @xmath72
will dynamically approach @xmath72 by a conduction driven mass flow .
a transition layer at @xmath73 is moving towards lower @xmath64 , and the gas is evaporating from the low-@xmath27 branch to the high-@xmath27 branch .
a transition layer at @xmath74 is moving towards higher @xmath64 , with the gas condensing to the low-@xmath27 branch @xcite .
we compute the disk structure for both condensing and evaporating solutions .
we assume a steady state , condensing or evaporating mass flow through the transition layer at @xmath75 or @xmath76 , respectively .
the static conduction solution is an intermediate case of the latter extreme cases .
we take a single - valued @xmath77 , since a two - phase solution would be buoyantly unstable , making the denser ( colder ) gas sink .
the evaporating disk corresponds to the low-@xmath27 branch , while the condensing disk corresponds to the high-@xmath27 branch ( fig .
[ fig1 ] ) .
this introduces spectral differences ( section [ sec : spec ] ) .
we do not know from first principles whether the disk atmosphere is evaporating , condensing , or static .
we estimate the speed of the conduction mass flow to be @xmath78 , by using the characteristic conduction time at the field length , where @xmath79 is the @xcite conductivity @xcite .
we obtain a conduction mass flow speed @xmath80 which is 1 - 2@xmath81 times the local sound speed .
thus , the phase dynamics will depend on the subsonic ( @xmath82 ) flow patterns in the disk atmosphere , and these flows will also determine the evaporation or condensation rates .
the only physical mechanism known to transport the necessary angular momentum for disk accretion involves a magneto - rotational instability ( mri ) which drives turbulent flow in the disk @xcite .
heat transfer due to this turbulent flow could quench the thermal instability and affect the disk structure .
the mri also favors our assumption of vertical isothermality in the optically thick disk .
with the disk structure @xmath83 , @xmath84 and ion abundances @xmath85 , we model the x - ray line emission from the disk atmosphere by using hullac ( hebrew university / lawrence livermore atomic code , @xcite ) .
the code calculates the atomic structure and transition rates of radiative recombination ( rr ) and radiative recombination continuum ( rrc ) for the h - like and he - like ions of c ,
n , o , ne , mg , si , s , ar , ca , and - .
the emissivities are calculated as described in @xcite .
the spectrum for each annulus is added to obtain the disk spectrum .
each annulus consists of a grid of zones in the vertical @xmath86 direction , and @xmath27 , @xmath17 and @xmath87 for each zone are used to calculate the rr and rrc emissivities .
the radiation is propagated outwards at inclination angle @xmath88 , including the continuum opacity of all zones above .
compton scattering is not included in the calculated spectrum because it is negligible in the vertical direction in the region where the line emission is formed .
the spectrum is doppler broadened by the projected local keplerian velocity , assuming azimuthal symmetry .
the self - consistent disk is thicker than would be expected from the local pressure scale height @xmath89 alone .
we find @xmath90 .
we fit the modeled photosphere and atmosphere height , @xmath39 and @xmath62 , with @xmath91 ( @xmath23 in cm ) , with fit parameters @xmath92 and @xmath93 . we find @xmath94@xmath95 @xmath96 , @xmath97+@xmath98 , @xmath99+@xmath100 @xmath101 , @xmath102 .
thus , @xmath103-@xmath104 @xmath89 and @xmath105-@xmath106+@xmath107 @xmath89 . estimated systematics are shown , if significant . @xcite
estimated @xmath108 , but in spite of the steeper radial dependence , the vrtilek disk is thinner , and it equals @xmath89 for @xmath109 cm . for comparison to the optically derived disk thickness , @xmath62 is the _ de facto disk boundary , since a fraction @xmath58 of the central x - rays have been absorbed there .
thus , previous theoretical studies severely underestimated the disk thickness . _ the atmosphere s ( @xmath110 ) maximum photoelectric opacity is always @xmath111 , although @xmath112 .
illumination heating dominates at @xmath113 @xmath114 cm ( eq . [ eq : locdisk ] ) , where only @xmath115 of the incident photons reach @xmath39 directly , while @xmath116 reach @xmath39 after reprocessing in the atmosphere .
thus , the atmospheric albedo is @xmath117 . to reproduce the disk ( atmosphere and photosphere ) albedo observed by @xcite ( @xmath118 ) , the photosphere albedo can be @xmath119 , which is a range closer to physical expectations
there is no significant change in the atmosphere structure by varying @xmath51 from 1 to 0.1 .
the lmxrb photon flux ( photon @xmath120 s@xmath121 kev@xmath121 ) is modeled by : @xmath122 where @xmath123 is the neutron star continuum , @xmath124 is the rr line and rrc modeled flux , @xmath125 and @xmath126 are the neutral hydrogen absorption column densities and @xmath127 are the @xcite absorption cross sections .
the system is assumed to be 10 kpc away . if we let @xmath128 in eq .
[ eq : spec ] with @xmath129 , the lines are swamped by the continuum .
thus , low inclination neutron star lmxbs are unlikely to have detectable x - ray lines from the disk .
instead , we consider @xmath130 @xmath120 and @xmath131 @xmath120 with @xmath132 , where an obscuring medium absorbs half of the continuum flux and is compact enough to leave the disk almost unobscured . with a partially obscured central continuum
, disk evaporation has an observable spectral signature .
we simulate 50 ks _ xmm - newton rgs 1 and _ chandra meg + 1 observations ( fig .
[ fig2],[fig3 ] ) .
some bright lines are shown on table [ tbl-1 ] .
the evaporating and condensing disks have contrasting / line ratios .
the evaporating disk contains gas at @xmath133-@xmath134 k , unlike the condensing disk .
the h - like ion line intensities are higher for the condensing disk since it has more gas at @xmath135 k. the spectral differences stem from the distinct differential emission measure distributions @xmath136 and from the recombination rate @xmath137 , where @xmath138-@xmath139 , @xmath140 is the emission measure , and @xmath141 are the ion and electron densities . _ _ the rrc and he - like line triplet can be used as temperature and density diagnostics , respectively .
the rrc width @xmath142 @xcite .
the forbidden line @xmath143 of at @xmath144 is absent due to collisional depopulation at high density , since @xmath145 @xmath146 .
the intercombination @xmath88 to resonance @xmath23 line ratio is @xmath147 , indicating a purely photoionized plasma @xcite .
the rrc broadening is doppler - dominated , resembling the rr lines , while the rrc has a component with @xmath148 that is produced at @xmath135 k.
using a model for a disk illuminated by a neutron star , we have produced a global spectrum of x - ray line emission from a self - consistent accretion disk atmosphere .
the flaring of the atmosphere matches optical observations @xcite . for fiducial parameters of a lmxb with a partially obscured central source with eddington luminosity
, observable spectral signatures would identify an evaporating disk atmosphere .
the / line ratios are a good tracer of evaporation from the disk .
the line - rich spectrum has line profiles resolvable by the _ xmm and _ chandra spectrometers .
the lines have broad wings and a double - peaked core with a velocity from the outermost radius .
only a single line may be resolved since the emission is well distributed throughout @xmath11 cm @xmath149 cm .
he - like triplet diagnostics show signatures of photoionization and high density ( @xmath150 @xmath146 ) , and h - like rrc are very broad , with a superposed narrow component in evaporating disks . _
_ the distinct spectral signature of evaporation is due to the presence of a thermal instability in the photoionized gas of the disk .
this instability can not be quenched by spectral variations , abundance variations , or optical depth effects @xcite .
conduction heating should not affect this result , though turbulent heat transfer or coronal heating might quench the instability if present . in the outer disk ,
the incident radiative power is @xmath151 times the viscous power , so coronal - like heating should be small .
the x - ray lines from the disk are weak and difficult to detect for a low - inclination lmxb .
only if the inclination is high and a cloudy environment exists do we expect them to be detectable . at press time , an _ xmm - newton observation of lmxb dipper exo 0748 - 67 has detected broad lines from photoionized gas with @xmath152 @xmath146 @xcite , which we interpret as the disk atmosphere emission , since it confirms the key predictions of our model .
_ allen , c. w. 1973 , astrophysical quantities ( 3rd ed . ; london : athlone press ) balbus , s. a. , & hawley , j. f. 1998 , rev .
phys . , 70 , 1 begelman , m. c. , & mckee , c. 1990 , , 358 , 375 church , m. j. 2000 , astro - ph/0012411 church , m. j. , balucinska - church , m. , dotani , t. , & asai , k. 1998 , , 504 , 516 cottam , j. , kahn , s. m. , brinkman , a. c. , den herder , j. w. , & erd , c. 2001 , , 365 , l277 de jong , j. a. , van paradijs , j. , & augusteijn , t. 1996 , , 314 , 484 dubus , g. , lasota , j. p. , hameury , j. m. , & charles , p. 1999 , , 303 , 139 field , g. b. 1965 , , 142 , 531 hess , c. j. , kahn , s. m. , & paerels , f. b. s. 1997 , , 478 , 94 klapisch , m. , schwab , j. l. , fraenkel , j. s. , & oreg , j. 1977 , opt
, 61 , 148 ko , y. , & kallman , t. r. 1994 , , 431 , 273 krolik , j. h. , mckee , c. f. , & tarter , c. b. 1981 , , 249 , 422 li , y. , gu , m. f. , & kahn , s. m. , submitted to morrison , r. , & mccammon , d. 1983 , , 270 , 119 liedahl d. a. , & paerels f. 1996 , , 468 , 33 mckee , c. f. & begelman , m. c. 1990 , , 358 , 392 nayakshin , s. , kazanas , d. , & kallman , t. r. 2000 , , 537 , 833 porquet , d. , & dubau , j. 2000 , , 143 , 495 press , w. h. 1994 , numerical recipes in fortran : the art of scientific computing ( cambridge : cambridge university press ) raymond , j. c. 1993 , , 412 , 267 sako m. , liedahl , d. a. , kahn , s. m. , & paerels , f. 1999 , , 525 , 921 shakura , n. i. , & sunyaev , r. a. 1973 , , 24 , 337 spitzer , l. 1962 , physics of fully ionized gases , ( new york : interscience ) vrtilek , s. d. , raymond , j. c. , garcia , m. r. , verbunt , f. , hasinger , g. , & kurster , m. 1990 , , 235 , 162 woods , d. t. , klein , r. i. , castor , j. i. , mckee , c. f. , & bell , j. b. 1996 , , 461 , 767 zeldovich , y. b. , & pikelner , s. b. 1969 , soviet physics jetp , 29 , 170 crr he@xmath51 triplet & 2.1 & 0.99 + ly@xmath51 & 3.4 & 4.7 + rrc & 1.3 & 0.61 + ly@xmath153 & 0.94 & 1.5 + l ( 0.815 kev ) & 0.45 & 0.91 + ly@xmath51 & 1.4 & 2.2 + h@xmath51 ( 1.28 kev ) & 0.74 & 0.86 + ly@xmath51 & 1.2 & 2.0 + he@xmath51 triplet & 1.2 & 2.4 + ly@xmath51 & 4.0 & 6.0 + ly@xmath51 & 15 & 16 + ly@xmath51 & 12 & 12 + | we model the x - rays reprocessed by an accretion disk in a fiducial low - mass x - ray binary system with a neutron star primary .
an atmosphere , or the intermediate region between the optically thick disk and a compton - temperature corona , is photoionized by the neutron star continuum .
x - ray lines from the recombination of electrons with ions dominate the atmosphere emission and should be observable with the _ chandra and _ xmm - newton high - resolution spectrometers .
the self - consistent disk geometry agrees well with optical observations of these systems , with the atmosphere shielding the companion from the neutron star . at a critical depth range ,
the disk gas has one thermally unstable and two stable solutions . a clear difference between the model spectra exists between evaporating and condensing disk atmospheres .
this difference should be observable in high - inclination x - ray binaries , or whenever the central continuum is blocked by absorbing material and the extended disk emission is not . _ _ |
motivated by the problem of local rigidity of homomorphisms from finitely generated groups to ( finite - dimensional ) lie groups , weil @xcite proved the following theorem .
[ thm : weil0](weil ) let @xmath0 and @xmath1 be finite - dimensional @xmath2-manifolds , @xmath3 and @xmath4 be @xmath2-maps .
let @xmath5 , @xmath6 .
suppose + ( 1 ) @xmath7 ; + ( 2 ) @xmath8 .
+ then there exists a neighborhood @xmath9 of @xmath10 in @xmath11 such that @xmath12 weil s proof relies heavily on the implicit function theorem .
if @xmath13 and @xmath1 in theorem [ thm : weil0 ] are banach manifolds , and @xmath14 and @xmath15 are closed complemented subspaces in the corresponding tangent spaces of the manifolds respectively , then weil s proof still works , based on the implicit function theorem for banach spaces . using the nash - moser inverse function theorem ,
hamilton @xcite proved a similar result in the setting of tame frchet spaces under more splitting assumptions ( see also @xcite ) .
such kind of results are useful in problems of deformation rigidity ( or local rigidity ) of certain geometric structures ( see e.g.@xcite,@xcite , @xcite , @xcite ) .
but such complementation or splitting conditions are not always satisfied and hard to verify in many concrete situations .
the first goal of this paper is to prove the banach version of theorem [ thm : weil0 ] without the assumption that the subspaces involved are complemented .
the only additional assumption we have to make , comparing with the finite - dimensional case , is that @xmath15 is closed .
more precisely , we will prove the following theorem .
for convenience , we state its local form .
[ thm : implicit0 ] let @xmath16 and @xmath17 be banach spaces , @xmath18 , @xmath19 be open @xmath20-neighborhoods .
let @xmath21 and @xmath22 be @xmath2-maps .
suppose + ( 1 ) @xmath23 ; + ( 2 ) @xmath24 ; + ( 3 ) @xmath25 ; + ( 4 ) @xmath26 is closed .
+ then there exists a @xmath20-neighborhood @xmath27 such that @xmath28 informally , we may think theorem [ thm : implicit0 ] as saying that @xmath29 determines some implicit function " from @xmath30 to @xmath9 ( although we do not even know a prori if it is continuous ) .
so we call theorem [ thm : implicit0 ] a generalized implicit function theorem "
. we will give three applications of the generalized implicit function theorem .
\(1 ) * parametrization of fibers of differentiable maps .
* suppose @xmath31 are finite - dimensional @xmath32-manifolds and @xmath4 is a @xmath32-map .
if @xmath33 is a regular value of @xmath34 in the sense that for each @xmath35 , the differential @xmath36 is surjective , then the implicit function theorem implies that @xmath37 is a @xmath32-submanifold of @xmath11 .
this fact can be generalized directly to banach manifolds under the additional assumption that @xmath38 is complemented in @xmath39 . without the complementation assumption
, we can not conclude that @xmath37 is a submanifold in general .
however , using the generalized implicit function theorem , we can prove that under the assumption of the existence of the map @xmath40 as in theorem [ thm : implicit0 ] , @xmath37 can be parametrized as a @xmath32-manifold in such a way that the manifold structure is compatible with the subspace topology , and the inclusion map is @xmath32 .
\(2 ) * lie subgroups of banach lie groups . *
if @xmath41 is a banach lie group , we call a closed subgroup @xmath42 of @xmath41 a lie subgroup if there is a banach lie group structure on @xmath42 compatible with the subspace topology ( there is at most one such structure ) .
it is well - known that closed subgroups of finite - dimensional lie groups are lie subgroups but this is false for banach lie groups ( cf .
@xcite ; @xcite , rem .
iv.3.17 ) . using the generalized implicit function theorem and a theorem of hofmann @xcite
, we will prove that certain isotropy groups are lie subgroups .
in particular , we will prove that if the coset space @xmath43 carries a banach manifold structure for which @xmath41 acts smoothly and the quotient map @xmath44 has surjective differential in some point @xmath45 , then @xmath42 is a banach lie subgroup of @xmath41 .
\(3 ) * local rigidity of homomorphisms .
* let @xmath41 be a banach lie group , @xmath46 a finitely generated group .
a homomorphism @xmath47 is locally rigid if any homomorphism from @xmath46 to @xmath41 sufficiently close to @xmath48 is conjugate to @xmath48 . using theorem [ thm : weil0 ] ,
weil @xcite proved that if @xmath49 , then @xmath48 is locally rigid .
we will generalize weil s result to the banach setting under the assumption that a certain linear operator has closed range .
theorem [ thm : implicit0 ] will be proved in section 2 .
the aforementioned three applications will be derived in sections 35 , respectively .
this work was initiated when the first author visited the department of mathematics at the tu darmstadt .
he would like to thank for its great hospitality .
in this section we prove the generalized implicit function theorem .
[ thm : implicit ] let @xmath16 and @xmath17 be banach spaces , @xmath18 , @xmath19 be open @xmath20-neighborhoods .
let @xmath21 and @xmath22 be @xmath2-maps .
suppose + ( 1 ) @xmath23 , @xmath50 ; + ( 2 ) @xmath24 ; + ( 3 ) @xmath25 ; + ( 4 ) @xmath26 is closed .
+ then there exists a @xmath20-neighborhood @xmath27 such that @xmath28 by the open mapping theorem , there exist positive constants @xmath51 such that for any @xmath52 and @xmath53 , there are @xmath54 and @xmath55 with @xmath56 , @xmath57 , and such that @xmath58 let @xmath59 choose a @xmath60 with @xmath61 , @xmath62 such that @xmath63 @xmath64 where @xmath65 , @xmath66 are the open balls in @xmath67 and @xmath68 with centers @xmath20 and radius @xmath69 . we claim that the @xmath20-neighborhood @xmath70 satisfies the request of the theorem .
in fact , we prove below that for @xmath71 , if @xmath72 , then there exists @xmath73 such that @xmath74 .
suppose @xmath75 such that @xmath72 .
choose a point @xmath76 , and define @xmath77 , @xmath78 inductively by @xmath79 here we use that @xmath80 .
we prove inductively that @xmath81 since @xmath82 @xmath83 and @xmath84 which implies that @xmath85 the inequalities hold for @xmath86 .
suppose that the inequalities hold for some @xmath87 .
we first have @xmath88 next we note that @xmath89 and that @xmath90 implies that @xmath91 holds for @xmath92 .
we thus obtain @xmath93 so we also have @xmath94 this proves the inequalities . by the definition of @xmath95 and the inequalities ,
the sequence @xmath96 is a cauchy sequence , hence converges to some @xmath97 , and we have @xmath98 we also have @xmath99 hence @xmath100 this proves the above claim , hence the theorem .
theorem [ thm : implicit ] does not hold without the assumption that @xmath26 being closed .
here is a counter - example .
let @xmath42 be a hilbert space with an orthonormal basis @xmath101 , @xmath102 be the @xmath2-map from @xmath42 to itself defined by @xmath103 , @xmath104 be the linear map from @xmath42 to itself determined by @xmath105 .
let @xmath106 , @xmath107 , @xmath108 , @xmath109 .
then @xmath110 , which is injective .
it is obvious that all the conditions in theorem [ thm : implicit ] hold except that @xmath111 is closed .
@xmath112 , but @xmath113 does not belong to @xmath114 for each @xmath115 .
let @xmath68 and @xmath17 be banach spaces , @xmath19 an open @xmath20-neighborhood , @xmath22 be a @xmath2-map and @xmath116 .
suppose @xmath117 and that @xmath26 is closed .
then there exists a @xmath20-neighborhood @xmath27 such that @xmath118
in this section we refine the implicit function theorem [ thm : implicit ] in such a way that it also provides a manifold structure on the @xmath20-fiber of @xmath34 . clearly , this is not a submanifold in the sense of bourbarki @xcite because its tangent space need not be complemented , but it nevertheless is a banach manifold . [
prop : open ] let @xmath67 , @xmath68 and @xmath17 be banach spaces . then the following assertions hold : the set @xmath119 of all surjective linear operators from @xmath67 to @xmath68 is open .
if @xmath120 and @xmath121 , then @xmath122 whenever @xmath123 .
the set @xmath124 of all closed embeddings of @xmath67 into @xmath68 is open .
if @xmath120 satisfies @xmath125 for some @xmath126 , then @xmath127 whenever @xmath123 .
the set of pairs @xmath128 with @xmath129 is an open subset in @xmath130 \(1 ) ( this proof is due to h. glckner
. another argument is given in @xcite , lemma and theorem 2.1 on pp.20/21 . )
let @xmath131 be open .
in view of the open mapping theorem , there exists an @xmath126 with @xmath121 .
we claim that @xmath132 is surjective for each @xmath133 with @xmath123 .
let @xmath134 and @xmath135 .
then there exists some @xmath136 with @xmath137 and @xmath138 .
then we have @xmath139 with @xmath140 now we pick @xmath141 with @xmath142 and @xmath143 and obtain @xmath144 iterating this procedure , we obtain a sequence @xmath145 in @xmath67 with @xmath146 , so that @xmath147 converges , and @xmath148 .
\(2 ) the open mapping theorem implies that @xmath120 is a closed embedding if and only if there exists an @xmath126 with @xmath149 for all @xmath150 .
if this is the case and @xmath133 satisfies @xmath123 , then @xmath151 implies that @xmath132 is a closed embedding .
\(3 ) let @xmath152 .
without loss of generality , we may assume that @xmath102 is an isometric embedding and that @xmath153 is a metric quotient map , i.e. , @xmath154 .
we show that if @xmath155 then @xmath156 and @xmath157 imply that @xmath158 .
let @xmath159 .
we put @xmath160 and assume that we have already constructed a sequence @xmath161 in @xmath162 with @xmath163 to find @xmath164 , we first observe that @xmath165 hence there exists @xmath166 with @xmath167 and @xmath168
. then @xmath169 and @xmath170 satisfies @xmath171 for @xmath172 we now have @xmath173 and @xmath174 we thus obtain sequences @xmath175 and @xmath176 satisfying the above conditions .
in particular , @xmath177 converges in @xmath178 and @xmath179 . with the preceding proposition we can strengthen theorem [ thm : implicit ] under the assumption that @xmath180 is injective and @xmath181 is surjective as follows .
[ rem - open ] let @xmath16 and @xmath17 be banach spaces , @xmath18 , @xmath19 be open @xmath20-neighborhoods .
let @xmath21 and @xmath22 be @xmath2-maps .
suppose + ( 1 ) @xmath23 , @xmath50 ; + ( 2 ) @xmath24 ; + ( 3 ) @xmath25 ; + ( 4 ) @xmath180 is injective , @xmath181 is surjective .
+ then there exists an open @xmath20-neighborhood @xmath182 such that @xmath183 is an open map into @xmath184 . for each @xmath185
the relation @xmath186 implies that @xmath187 so that proposition [ prop : open ] implies that @xmath188 holds for each @xmath189 in some open @xmath20-neighborhood @xmath182 .
applying theorem [ thm : implicit ] to the maps @xmath190 and @xmath191 , it follows that @xmath183 is an open map to @xmath184 .
[ lem : quasiiso ] let @xmath192 be an open subset containing @xmath20 , @xmath193 a @xmath2-map for which @xmath194 is a closed embedding
. then there exists an open @xmath20-neighborhood @xmath195 and constants @xmath196 with @xmath197 since @xmath194 is a closed embedding , there exists some @xmath126 with @xmath198 for @xmath150 . pick an open ball @xmath195 around @xmath20 with @xmath199 and note that this implies that @xmath200 we now obtain for @xmath201 : @xmath202 which in turn leads to @xmath203 with @xmath204 and @xmath205 the assertion now follows from @xmath206 let @xmath207 be a @xmath32-map from an open @xmath20-neighborhood @xmath208 in a banach space @xmath68 into another banach space @xmath17 with @xmath50 and @xmath181 is surjective .
assume that there is an open @xmath20-neighborhood @xmath9 in some banach space @xmath67 and a @xmath32-map @xmath209 such that @xmath210 , @xmath211 is injective , and @xmath212 . then by proposition
[ rem - open ] and lemma [ lem : quasiiso ] , @xmath184 a locally a banach @xmath32-manifold around @xmath20 , and a local @xmath32-chart is provided by @xmath40 . to get the global version of such result , we need more preparation . [
lem : cont ] the map @xmath213 is continuous .
fix @xmath214 with @xmath215
. we may w.l.o.g .
assume that @xmath216 is an isometric embedding @xmath217 .
then for each @xmath218 with @xmath219 the map @xmath220 is a closed embedding with @xmath221 we conclude that for any @xmath133 with @xmath222 we have @xmath223 for any @xmath224 , @xmath133 with @xmath225 and @xmath150 we thus obtain @xmath226 since @xmath216 is assumed to isometric , the map @xmath227 satisfies @xmath228 therefore the estimate above implies the continuity of @xmath229 . [ thm : ck ] let @xmath16 , and @xmath17 be banach spaces .
let @xmath230 be an open subset containing @xmath20 and @xmath231 be a @xmath32-map , @xmath232 , such that @xmath233 and @xmath181 is surjective .
further , let @xmath234 be open subsets containing @xmath20 and @xmath235 be @xmath32-maps such that @xmath236 , @xmath237 is injective , and @xmath238 for @xmath239
. then there exist open @xmath20-neighborhoods @xmath240 such that @xmath241 are homeomorphisms onto a @xmath20-neighborhood in @xmath242 and the map @xmath243 is @xmath32 .
we prove the theorem by induction over @xmath244 .
if it holds for each @xmath244 , it clearly holds for @xmath245 .
we first prove the theorem for @xmath246 . using proposition [ rem - open ] , we may w.l.o.g .
assume that @xmath247 is a homeomorphism onto an open @xmath20-neighborhood in @xmath242 and that @xmath248 is injective , @xmath249 is surjective , and @xmath250 hold for all @xmath251 .
next we use lemma [ lem : quasiiso ] to see that we may further assume that there exist constants @xmath196 such that for @xmath252 we have @xmath253 now fix @xmath254 with @xmath255 .
we have to show that @xmath256 is differentiable in @xmath257 .
let @xmath258 and @xmath259 .
since @xmath260 and @xmath261 are closed embeddings , the linear map @xmath262 is invertible .
we have @xmath263 with @xmath264 and @xmath265 with @xmath266 . for @xmath267
we then have @xmath268 so that @xmath269 this implies that @xmath270 and that @xmath271 from ( [ eq : esti ] ) we derive @xmath272 and hence @xmath273 therefore @xmath274 is differentiable in @xmath20 with @xmath275 .
we conclude that @xmath256 is differentiable in @xmath257 with @xmath276 the continuity of the differential @xmath277 follows from lemma [ lem : cont ] .
this finishes the proof of the theorem for @xmath246 .
now let us assume that @xmath278 and that the theorem holds for @xmath279-maps , showing that @xmath256 is a @xmath279-map .
it remains to show that the tangent map @xmath280 is also @xmath279 , which therefore proves that @xmath256 is @xmath32 .
the tangent map @xmath281 is a @xmath279-map with @xmath282 for which @xmath283 is surjective .
further , the maps @xmath284 are @xmath279-maps with @xmath285 our induction hypothesis implies that @xmath286 is a @xmath279-map , and this completes the proof .
[ thm : regular ] ( regular value theorem ; without complements ) let @xmath287 be a @xmath32-map between banach @xmath32-manifolds , @xmath288 , @xmath289 and @xmath67 be a banach space .
assume that @xmath290 is a regular value in the sense that for each @xmath291 the differential @xmath292 is surjective , and for every @xmath293 , there exists an open @xmath20-neighborhood @xmath294 and a @xmath32-map @xmath295 with @xmath296 , @xmath297 is injective , and @xmath298 .
then @xmath242 carries the structure of a banach @xmath32-manifold modeled on @xmath67 . in view of proposition [ rem - open ]
, the maps @xmath299 whose existence is assured by ( 2 ) yield local charts and theorem [ thm : ck ] implies that the chart changes are also @xmath32 , so that we obtain a @xmath32-atlas of @xmath242 .
if @xmath300 and @xmath301 are banach
lie groups , then the existence of canonical coordinates ( given by the exponential function ) implies that each continuous homomorphism @xmath302 is actually smooth , hence a morphism of lie groups .
this implies in particular that a topological group @xmath41 carries at most one banach
lie group structure .
if @xmath41 is a banach lie group , then we call a closed subgroup @xmath303 a _ banach lie subgroup _ if it is a banach lie group with respect to the subspace topology
. it is an important problem in infinite - dimensional lie theory to find good criteria for a closed subgroup of a banach lie group to be a banach lie subgroup . regardless of any lie group structure , we may defined for each closed subgroup @xmath42 of a banach lie group its lie algebra by @xmath304 and the product and commutator formula easily imply that @xmath305 is a closed lie subalgebra of the banach
lie algebra @xmath306 ( cf .
@xcite , lemma iv.3.1 ) . by restriction
we thus obtain an exponential map @xmath307 from @xcite , thm .
iv.4.16 we know that there exists a banach lie group @xmath308 with lie algebra @xmath305 and an injective morphism of lie groups @xmath309 for which @xmath310 is the subgroup of @xmath42 generated by the image of @xmath311 .
since @xmath305 is a closed subspace of @xmath306 , the map @xmath312 is a closed embedding , and @xmath42 is a banach lie subgroup if and only if @xmath313 is an open map , which is equivalent to the existence of an open @xmath20-neighborhood @xmath314 for which @xmath315 is an open embedding ( cf .
@xcite , thm .
this notion of a banach lie subgroup is weaker than the concept used by bourbaki ( @xcite ) , where it is assumed that @xmath305 has a closed complement , a requirement that is not always satisfied and hard to verify in many concrete situations . for closed
normal subgroups @xmath42 of banach lie groups one has the nice characterization that @xmath42 is a banach lie subgroup if and only if the topological quotient group @xmath43 is a banach lie group ( @xcite )
. it would be nice to have a similar criterion for general subgroups : [ conj ] a closed subgroup @xmath42 of a banach lie group @xmath41 is a banach lie subgroup if and only if @xmath43 carries the structure of a banach manifold for which the quotient map @xmath44 has surjective differential in each point and @xmath41 acts smoothly on @xmath43 .
the main difficulty arises from the possible absence of closed complements to the closed subspace @xmath305 in @xmath306 .
if @xmath42 is a banach lie subgroup of @xmath41 and @xmath305 has a closed complement , then the classical inverse function theorem provides natural charts on the quotient space @xmath43 , hence a natural manifold structure with all nice properties ( cf .
@xcite , ch . 3 ) .
without a closed complement for @xmath305 it is not known how to construct charts of @xmath43 .
however , the natural model space is the quotient space @xmath316 . with the aid of the `` implicit function theorem ''
, we can now prove one half of the conjecture above : [ thm : liesub ] let @xmath41 be a banach lie group and @xmath317 a smooth action of @xmath41 on a banach manifold @xmath11 , @xmath318 .
suppose that for the derived action @xmath319 the subspace @xmath320 of @xmath321 is closed .
then the stabilizer @xmath322 is a lie subgroup .
first we note that @xmath323 is a closed subgroup of @xmath41 with @xmath324 we consider the smooth maps @xmath325 then @xmath326 , @xmath327 , @xmath328 since @xmath329 is closed , theorem [ thm : implicit ] implies that there exists arbitrarily small open @xmath20-neighborhood @xmath330 such that @xmath331 is a @xmath332-neighborhood in @xmath323 .
by @xcite , thm .
iv.3.3 ( cf
. also @xcite , prop .
3.4 ) , @xmath323 is a banach lie subgroup of @xmath41 .
let @xmath41 be a banach lie group and @xmath303 a closed subgroup for which @xmath43 carries the structure of a banach manifold .
suppose the quotient map @xmath333 has surjective differential in some point @xmath334 and @xmath41 acts smoothly on @xmath43 .
then @xmath42 is a banach lie subgroup of @xmath41 .
the action of @xmath41 on @xmath43 is given by @xmath335 so that @xmath336 and the stabilizer of @xmath337 is @xmath338 . by theorem [ thm : liesub ]
, @xmath339 is a banach lie subgroup .
so @xmath42 is a banach lie subgroup .
[ rem : more - gen ] the preceding theorem has a natural generalization to the following setting . let @xmath42 be a closed subgroup of a banach lie group @xmath41 .
suppose that there exists an open @xmath332-neighborhood @xmath340 and a @xmath2-map @xmath341 to some banach manifold @xmath11 such that @xmath342 is surjective and that for @xmath343 we have @xmath344 then @xmath42 is a banach lie subgroup of @xmath41 .
suppose that the banach lie group @xmath41 acts smoothly on the banach manifold @xmath11 and @xmath345 .
is it always true that the stabilizer @xmath323 is a lie subgroup of @xmath41 ? from the theory of algebraic banach lie groups it follows that for each banach space @xmath68 , for each @xmath346 and each closed subspace @xmath347 , the subgroups @xmath348 and @xmath349 are banach lie subgroups .
since inverse images of lie subgroups are lie subgroups ( @xcite , lemma iv.11 ) the problem has a positive solution for linear actions and for stabilizers of closed subspaces .
if @xmath67 is not complemented in @xmath68 , then it is not clear how to turn the grassmannian @xmath350 into a smooth manifold on which @xmath351 acts .
we are therefore in a situation where we know that the closed subgroup @xmath352 is a banach lie subgroup , but nothing is known on the quotient @xmath353 .
therefore a solution of conjecture [ conj ] would in particular lead to a natural banach manifold structure on the grassmannian of all closed subspaces of @xmath68 .
let @xmath41 be a topological group , @xmath46 a finitely generated ( discrete ) group .
let @xmath354 be the set of all homomorphisms from @xmath46 to @xmath41 , endowed with the compact - open topology , which coincides with the topology of pointwise convergence .
a homomorphism @xmath355 is _ locally rigid _ if there is an open neighborhood @xmath9 of @xmath48 in @xmath354 such that for any @xmath356 , there exists some @xmath45 with @xmath357 for all @xmath358 .
if , moreover , @xmath41 is a banach lie group , then every homomorphism @xmath355 defines naturally a @xmath46-module structure on the lie algebra @xmath306 of @xmath41 , via @xmath359 , @xmath360 . for a fixed @xmath355 and a map @xmath361 the pointwise product @xmath362 is a homomorphism if and only if @xmath102 satisfies the cocycle condition @xmath363 for a smooth family of homomorphisms @xmath364 @xmath365 with @xmath366 , this observation directly implies that the differential of @xmath367 at @xmath368 is a cocycle of @xmath46 in @xmath306 , and if @xmath369 for a smooth curve @xmath370 in @xmath41 with @xmath371 , then the above differential is a coboundary .
so it is reasonable to expect some relation between the local rigidity of @xmath48 and the vanishing of @xmath372 .
the main ingredient in weil s proof of theorem [ thm : weil ] is the finite - dimensional version of theorem [ thm : implicit ] . using theorem [ thm : implicit ] , in this section we generalize weil s theorem to banach
lie groups .
to state our result , we first recall a definition of the first cohomology of groups , which is related to presentations of groups .
let @xmath46 be a group .
let @xmath374 be a presentation of @xmath46 , where @xmath178 is a free group on a subset @xmath375 , @xmath376 is a free group on a subset @xmath377 . for @xmath378 ,
@xmath379 can be expressed as a reduced word @xmath380 in @xmath242 , where @xmath381 , @xmath382 .
let @xmath208 be a @xmath46-module .
for a set @xmath104 , let @xmath383 denote the set of all maps from @xmath104 to @xmath208 .
then @xmath383 has a natural abelian group structure induced from that of @xmath208 .
we define the coboundary operators @xmath384 as follows : @xmath385 for @xmath386 and @xmath387 , where @xmath388 then @xmath389 and @xmath390 are homomorphisms of abelian groups , @xmath391 .
the first cohomology group of @xmath46 in @xmath208 is , by definition , @xmath392 .
[ r : coincide ] this definition of @xmath393 coincides with the usual one which is defined by @xmath394 , where @xmath395 @xmath396 a direct way to see this is to look at the bijection between @xmath397 and @xmath398 which sends @xmath399 to the element @xmath400 in @xmath398 , under which @xmath401 is mapped bijectively onto @xmath402 . a more intrinsic way to see the coincidence
is explained in remark [ r : hausdorff ] below .
now suppose that @xmath46 is a finitely generated group .
then the presentation of @xmath46 may be chosen in such a way that @xmath178 is finitely generated .
so we may assume that @xmath242 is finite , @xmath403 is countable .
suppose moreover that @xmath208 is a continuous banach @xmath46-module .
then for a countable set @xmath104 , @xmath383 a frchet space with respect to the semi - norms @xmath404 if , moreover , @xmath104 is finite , then @xmath383 is a banach space with respect to the norm @xmath405 in particular , @xmath406 is a banach space , @xmath407 is a frchet space .
it is easy to see that the coboundary operators @xmath389 and @xmath390 are continuous .
we say that @xmath390 _ has closed image _ if @xmath408 is a closed subspace of @xmath409 .
our generalization of weil s theorem can be stated as follows .
in general , the set of generators @xmath403 of the group @xmath376 is infinite .
so @xmath409 is only a frchet space .
the following lemma enable us to reduce the problem to the category of banach spaces so as to we can apply theorem [ thm : implicit ] .
[ lem : finite ] let @xmath46 be a finitely generated group , @xmath208 a continuous banach @xmath46-module , and let the notation be as above . if @xmath390 has closed image , then there exists a finite subset @xmath410 such that the map @xmath411 defined by @xmath412 satisfies @xmath413 , and @xmath414 is closed in @xmath415 . since @xmath390 is continuous , @xmath398 is a closed subspace of @xmath416 .
so @xmath417 is a banach space with respect to the norm @xmath418\|'=\inf_{z\in \ker(\delta^1)}\|c+z\|_s,\ ] ] where @xmath419=c+\ker(\delta^1)$ ] , and the linear injection @xmath420 induced by @xmath390 is continuous . since @xmath421 is closed in @xmath416 , @xmath422 is a frchet space with respect to the semi - norms @xmath423 , @xmath378 . by the open mapping theorem for frchet spaces ( see , for example , @xcite , section i.3 , corollary 1 ) , @xmath424 is an isomorphism of frchet spaces from @xmath67 onto @xmath408 .
by @xcite , section ii.1 , corollary 1 to proposition 5 , there is a finite subset @xmath410 such that @xmath425\| ' \leq & c\sup_{t\in t'}p_t(\widetilde{\delta^1}([c ] ) ) = c\sup_{t\in t'}p_t(\delta^1(c ) ) = c\sup_{t\in t'}\|\delta^1_{t'}(c)(t)\| = c\|\delta^1_{t'}(c)\|_{t'},\end{aligned}\ ] ] for all @xmath387 , where @xmath426 is a constant . if @xmath427 , the above inequality implies that @xmath428\|'=0 $ ] .
. then @xmath430 .
the converse inclusion is obvious .
let @xmath431 be the continuous linear map induced by @xmath432 .
the above inequality says nothing but @xmath428\|'\leq c\|\widetilde{\delta^1_{t'}}([c])\|_{t'}$ ] for @xmath419\in x$ ] .
so @xmath433 is closed in @xmath434 .
this proves the lemma .
* proof of theorem [ thm : rigidity ] .
* since the lie algebra @xmath306 of @xmath41 is a continuous banach @xmath46-module , and we have assumed that @xmath390 has closed image , by lemma [ lem : finite ] , there is a finite subset @xmath410 such that the map @xmath435 defined by @xmath412 satisfies @xmath413 , and @xmath414 is closed in @xmath436 . for a finite set @xmath104 ,
the set @xmath437 of all maps from @xmath104 to @xmath41 is isomorphic to the product @xmath438 of @xmath439 copies of @xmath41 , hence carries a natural banach manifold structure , and the tangent space of @xmath437 at the constant map @xmath440 may be identified with @xmath441 .
define the maps @xmath442 and @xmath443 by @xmath444 @xmath445 then it is easy to see that @xmath446 , @xmath447 , and that @xmath448 .
now we compute the differentials of @xmath29 and @xmath34 at @xmath449 and @xmath450 , respectively .
we have @xmath451 where @xmath452 , @xmath453 .
so @xmath454 we also have @xmath455 where @xmath456 , @xmath457 .
so @xmath458 since @xmath49 , we have @xmath459 we also have that @xmath460 is closed in @xmath461 .
this verifies all the conditions of theorem [ thm : implicit ] .
so by theorem [ thm : implicit ] , there exists a neighborhood @xmath462 of @xmath450 such that @xmath463 let @xmath9 be the neighborhood of @xmath48 in @xmath354 such that @xmath464 if and only if the map @xmath465 defined by @xmath466 is in @xmath467 .
then for any @xmath464 , @xmath468 . by what we have proved above , @xmath469 , which means that there exists @xmath45 such that @xmath470 for all @xmath453 , that is , @xmath357 for @xmath471 .
but @xmath472 generates @xmath46 .
so @xmath357 for any @xmath358 .
this shows that @xmath48 is locally rigid .
[ r : hausdorff ] the condition that @xmath390 has closed image in theorem [ thm : rigidity ] can be replaced by a more intrinsic condition .
we recall the definition of group cohomology by projective resolutions .
let @xmath46 be a group , let @xmath473 be a projective @xmath46-resolution of the trivial @xmath46-module @xmath474 .
then for a @xmath46-module @xmath208 , the cohomology @xmath475 of @xmath46 in @xmath208 is , by definition , the cohomology of the complex @xmath476 and @xmath477 is independent of the particular choice of the projective resolution .
the most usual definition of group cohomology that appeared in remark [ r : coincide ] corresponds to the standard resolution .
if each @xmath478 in the resolution is countable , and @xmath208 is a continuous banach @xmath46-module , then each @xmath479 , viewed as a closed subspace of @xmath480 , is a frchet space , and the coboundary operators @xmath481 are continuous . for some @xmath87 , to say that @xmath482 has closed image is equivalent to say that @xmath483 is hausdorff ( see @xcite ) . using a standard argument of homological algebra
, it can be shown that @xmath483 being hausdorff is independent of the particular choice of the countable projective resolution . the operator @xmath390 that we used in theorem [ thm : rigidity ] is just the first coboundary operator defined by the gruenberg resolution of @xmath46 associated with the presentation of @xmath46 ( see @xcite ) , and the modules @xmath478 in the gruenberg resolution are countable if the free group @xmath178 in the presentation of @xmath46 is countable .
so the condition that @xmath390 has closed image in theorem [ thm : rigidity ] is equivalent to the condition that @xmath484 is hausdorff , which is independent of the particular countable resolution that we use in the definition of the group cohomology . by @xcite , chapter ix , proposition 3.5 , if @xmath484 is finite - dimensional , then @xmath484 is hausdorff .
hofmann , k. h. , _ analytic groups without analysis _ , symposia mathematica * 16 * ( convegno sui gruppi topologici e gruppi di lie , indam , rome , 1974 ) ; 357374 , academic press , london , 1975 s. lang , `` fundamentals of differential geometry , '' graduate texts in math . * 191 * , springer - verlag , berlin , 1999 .
neeb , k .- h .
, _ infinite - dimensional lie groups and their representations _ , in `` lie theory : lie algebras and representations , '' progress in math . *
228 * , ed .
j. p. anker , b. rsted , birkhuser verlag , 2004 ; 213328 | we prove a generalized implicit function theorem for banach spaces , without the usual assumption that the subspaces involved being complemented
. then we apply it to the problem of parametrization of fibers of differentiable maps , the lie subgroup problem for banach lie groups , as well as weil s local rigidity for homomorphisms from finitely generated groups to banach lie groups . + keywords : implicit function theorem , banach manifold , banach lie group + msc2000 : 22e65 , 57n20 |
the formation of ripples and dunes at the surface of an erodible sand bed results from the interplay between the relief , the flow and the sediment transport .
the aim of these two companion papers is to propose a coherent and detailed picture of this phenomenon in the generic and important case of a unidirectional turbulent stream .
this first part is devoted to the study of the stationary flow over a wavy rough bottom . in the second part
we propose a common theoretical framework for the description of the different modes of sediment transport .
hydrodynamics and transport issues at hand , we then revisit the linear instability of a flat sand bed submitted to a water shear flow and show that , in contrast to ripples , subaqueous dunes can not form by a primary linear instability .
it has long been recognised that the mechanism responsible for the formation and growth of bedforms is related to the phase - lag between sediment transport and bed elevation ( @xcite ) .
it has been shown in the context of aeolian dunes that this lag comes from two contributions , which can be considered as independent as the time scale involved in the bed evolution is much slower than the hydrodynamics relaxation ( @xcite ) .
first there is a shift between the bed and the basal shear stress profiles .
this shift purely results from the hydrodynamics and its sign is not obvious _ a priori _
, i.e. the stress maximum can be either upstream or downstream the bed crest depending on the topography or the proximity of the free surface .
the second contribution comes from the sediment transport : the sediment flux needs some time / length to adapt to some imposed shearing .
this relaxation mechanism induces a downstream lag of the flux with respect to the shear . when the sum of these two contributions results in a maximum flux upstream the bed crest , sediment deposition occurs on the bump , leading to an unstable situation and thus to the amplification of the disturbance . in part 1
, we shall focus on the first of these contributions , the second one being treated in part 2 .
@xmath1 is perpendicular .
the bottom profile is @xmath2 . in that experiment ,
the wavelength and amplitude of the bumps are @xmath3 m and @xmath4 m , for a water depth @xmath5 m ) .
the point of maximum shear on the bump ( where the lines are squeezed ) is located upstream the crest .
[ wavyschematic ] ] we consider here the generic case of a flow over a fixed sinusoidal bottom of wavelength @xmath6 ( see figure [ wavyschematic ] for an illustration of the geometry and some notations ) . in order to obtain the basal shear stress and in particular its phase shift with respect to the topography
, the equations of hydrodynamics must be solved in this geometry .
the case of viscous flows has been investigated by @xcite .
the first attempts to model the high reynolds number regime in the context of ripples and dunes in rivers have dealt with potential flows ( @xcite ) , for which the velocity field does not present any lag with respect to the bottom .
the shallow - water approximation ( @xcite ) implies that the bedforms spread their influence on the whole depth of the flow .
however , patterns only have a significant influence within a vertical distance on the order of their wavelength .
it is then crucial to compute explicitly the vertical flow structure , taking into account the turbulent fluctuations . in order to overcome the flaws of the perfect flow
, constant eddy viscosity closures have been tried to improve kennedy s original model ( @xcite ) .
further progress has been made by @xcite , who used a more sophisticated modelling with an additional equation on the turbulent energy and a closure which involves a prandtl mixing length in the expression of the eddy viscosity .
@xcite made use of the same turbulent modelling , but in the case of an infinite water depth .
a mixing length approach was also used by @xcite to improve benjamin s laminar description . in the meteorological context of atmospheric flows over low hills ,
a deep and fundamental understanding of the physics of turbulent flows over a relief has been developed from the 70 s ( see the review by @xcite ) .
starting with the seminal work of @xcite , further refined by @xcite and @xcite , the gross emerging picture is that the flow can be thought of as composed of two ( or more ) layers , associated with different physical mechanisms and different length scales .
@xcite have been able to compute analytically the basal shear stress for asymptotically large patterns , under an infinite flow depth assumption .
their ideas have been discussed in a rather vast literature .
the predictions of these calculations , and in particular this layered structure of the flow , has been compared with experiments ( see e.g. @xcite ) , or field measurements on large scale hills ( see e.g. the review paper by @xcite ) , with a good degree of success , especially on the upstream side of the bumps . moreover , they have been tested against the results of the numerical integration , in various configurations , of navier - stokes equations closed with different turbulent closures ( @xcite ) .
the relevance of this approach for the description of the flow and the stresses around aeolian sand dunes has also been investigated ( see e.g. @xcite ) , and is amongst the current directions of research in that community ( @xcite ) . because the prediction of the stable or unstable character of a flat sand bed submitted to a turbulent shear flow is very sensitive to the way both hydrodynamics and transport issues are described and intermixed , we find it useful to discuss at length , in the two parts of this paper , the different mechanisms and scales involved at the different steps of the modelling .
it is indeed particularly revealing that , despite the fact that the approach presented here is very close to those of @xcite and @xcite and has been motivated by these works , we basically disagree with their conclusions , especially that river dunes are initiated by the linear instability of a flat bed .
the detail discussion we provide here gives also the opportunity to revisit the still debated question of the subaqueous ripple size selection ( see @xcite and references therein ) , as well as the important issue of the classification of bedforms ( @xcite ) .
this article is structured as follows . in the next section
, we briefly recall the equations for the base flow over a uniform bottom .
we then study the linear solution in the case of wavelengths much smaller than the flow depth .
importantly , the sensitivity of these linear results with respect to various changes in the modelling is tested in sections [ controlzzero ] and [ robust ] .
section [ nl ] is devoted to the derivation of the first non - linear corrections . in section [ fs ]
, we investigate the effect of the free surface in the case of wavelengths comparable or larger than the flow depth and interpret it in terms of topography induced standing gravity waves .
finally , we provide in section [ qualitativesummary ] a qualitative summary of the main results of the paper .
the most technical considerations are gathered in appendices .
we consider a turbulent flow over a relief . following reynolds decomposition between average and fluctuating ( denoted with a prime ) quantities , the equations governing the mean velocity field @xmath7
can be written as : @xmath8 where @xmath9 is the reynolds stress tensor ( @xcite ) . for the sake of simplicity ,
we omit the density factor @xmath10 in front of the pressure @xmath11 and the stress tensor .
the aim of this paper is to describe quantitatively the average flow over a fixed corrugated boundary within this framework .
the reference state is the homogeneous and steady flow over a flat bottom , submitted to an imposed constant shear stress @xmath12 .
the turbulent regime is characterised by the absence of any intrinsic length and time scales . at a sufficiently large distance @xmath1 from the ground ,
the only length - scale limiting the size of turbulent eddies
the so - called mixing length @xmath13 is precisely @xmath1 ; the only mixing time - scale is given by the velocity gradient @xmath14 .
as originally shown by @xcite , it results from this dimensional analysis that the only way to construct a diffusive flux is a turbulent closure of the form : @xmath15 where the mixing length is @xmath16 and @xmath17 is
the ( phenomenological ) von krmn constant . after integration
, one obtains that the velocity has a single non zero component along the @xmath18-axis , which increases logarithmically with @xmath1 ( @xcite ) : @xmath19 where @xmath20 is a constant of integration called the hydrodynamical roughness .
this expression does not apply for @xmath21 .
there should be layer of thickness @xmath22 close to the bottom , called the surface layer , matching the logarithmic profile to a null velocity on the ground .
the hydrodynamical roughness @xmath20 should be distinguished from the geometrical ( or physical ) roughness of the ground , usually defined as the root mean square of the height profile variations .
@xmath20 is defined as the height at which the velocity would vanish , when extrapolating the logarithmic profile to small @xmath1 .
the physical mechanism controlling @xmath20 can be of different natures .
if the ground is smooth enough , a viscous sub - layer of typical size @xmath23 must exist , whose matching with the logarithmic profile determines the value of @xmath20 . on the contrary
, if the geometrical roughness is larger than the viscous sub - layer , turbulent mixing dominates at small @xmath1 with a mixing length controlled by the ground topography . in the case of a static granular bed composed of grains of size @xmath24 , reported values of the hydrodynamical roughness are reasonably consistent ( @xmath25 in @xcite , @xmath26 in @xcite and @xmath27 ( @xcite ) . in section [ nl ]
, we will justify the connection between geometrical and hydrodynamical roughness on a rigourous basis and show that they are not simply proportional .
the situation is of course different in the presence of sediment transport , which may ( or not ) induce some negative feedback on the flow . in this case , the hydrodynamical roughness @xmath20 may directly be controlled by the transport characteristics ( e.g. mass flux and grain trajectories ) .
nature presents many other physical processes controlling the roughness : for instance , the flexible stems of wetland plants in low marshes or , for the wind , the canopy or the waves over the ocean . in all these cases , it can be assumed that the logarithmic law is a good approximation of the velocity profile above the surface layer , with a single known parameter @xmath20 .
we will first consider the asymptotic limit in which the typical relief length
say , the dune wavelength @xmath6 is much larger than the surface layer thickness @xmath22 .
the relief is locally flat at the scale @xmath22 , so that there must be a region close to the ground where the velocity profile shows a logarithmic vertical profile .
we will then discuss the case of moderate values of the ratio @xmath28 , for which the flow becomes sensitive to the details of the mechanisms controlling the roughness . in the logarithmic boundary layer
, the normal stresses can be written as : @xmath29 where @xmath30 is a second phenomenological constant estimated in the range @xmath31 .
note that @xmath30 does not have any influence on the results as it describes the isotropic component of the reynolds stress tensor , which can be absorbed into the pressure terms .
normal stress anisotropy is considered in section [ robust ] and appendix [ avi ] . introducing the strain rate tensor @xmath32 and its squared modulus @xmath33 , we can write both expressions ( [ tauxzprandtl ] ) and ( [ taullprandtl ] ) in a general tensorial form : @xmath34 in this paper
, we focus on 2d steady situations , i.e. on geometries invariant along the transverse @xmath35-direction , see figure [ wavyschematic ] .
as they are of permanent use for the rest of the paper , we express the components of the velocity and stress equations in the @xmath18- and @xmath1-directions .
the navier - stokes equations read : @xmath36 the stress expressions are the following : @xmath37 in these expressions , the strain tensor components are given by @xmath38 and the strain modulus by : @xmath39
we now consider the turbulent flow over a wavy bottom constituting the floor of an unbounded boundary layer . in rivers
, this corresponds to the limit of a flow depth @xmath40 much larger than the bed - form wavelength @xmath6 .
the solution is computed as a first order linear correction to the flow over a uniform bottom , using the first order turbulent closure previously introduced . for small enough amplitudes
, we can consider a bottom profile of the form @xmath41 without loss of generality .
@xmath42 is the wavelength of the bottom and @xmath43 the amplitude of the corrugation , see figure [ wavyschematic ] .
the case of an arbitrary relief can be deduced by a simple superposition of fourier modes .
we introduce the dimensionless variable @xmath44 , the dimensionless roughness @xmath45 and the function : @xmath46 we also switch to the standard complex number notation : @xmath47 ( real parts of expressions are understood ) .
we wish to perform the linear expansion of equations ( [ nscont])-([strainmod ] ) with respect to the small parameter @xmath48 .
the mixing length is still defined as the geometrical distance to the bottom : @xmath49 .
we introduce the following notations for the two first orders : @xmath50 , \label{defu}\\ u_z & = & u _ * k\zeta e^{ikx } w , \label{defw}\\ \tau_{xz } & = & \tau_{zx}= - u_*^2 \left[1+k\zeta e^{ikx } s_t\right ] , \label{defst}\\ p+\tau_{zz } & = & p_0+u_*^2 \left [ \frac{1}{3}\chi^2 + k\zeta e^{ikx } s_n\right ] , \label{defsn}\\ \tau_{zz } & = & u_*^2 \left [ \frac{1}{3}\chi^2 + k\zeta e^{ikx } s_{zz}\right ] , \label{defszz}\\ \tau_{xx } & = & u_*^2 \left [ \frac{1}{3}\chi^2 + k\zeta e^{ikx } s_{xx}\right ] . \label{defsxx}\end{aligned}\ ] ] the quantities @xmath51 , @xmath52 , etc , are implicitly considered as functions of @xmath53 .
an alternative choice is to consider functions of the coordinate @xmath54 .
such alternative functions are denoted with a tilde to make the distinction .
this important but somehow technical issue of the choice of a representation is discussed in appendix [ rep ] .
although the curvilinear and cartesian systems of coordinates are equivalent , the distinction between the two is of importance when it comes to the expression of the boundary conditions , and for the range of amplitudes @xmath43 for which the linear analysis is no more valid ( see section [ nl ] ) .
in particular , vertical profiles in the forthcoming figures will be mostly plotted as a function of the shifted variable @xmath55 .
the linearised strain rate tensor reads @xmath56 finally the navier - stokes equations lead to @xmath57 taking the difference of equations ( [ relaxsxx ] ) and ( [ relaxszz ] ) , one can compute @xmath58 to obtain four closed equations : @xmath59 introducing the vector @xmath60 , we finally get at the first order in @xmath61 the following compact form of the equation to integrate : @xmath62 the general solution of this equation is the linear superposition of all solutions of the homogeneous system ( i.e. with @xmath63 ) , and a particular solution @xmath64 .
four boundary conditions must be specified to solve the above equation ( [ systlinplaque ] ) .
the upper boundary corresponds to the limit @xmath65 , for which we ask that the vertical fluxes of matter and momentum vanish asymptotically .
this means that the first order corrections to the shear stress and to the vertical velocity must tend to zero : @xmath66 and @xmath67 . in practice , a boundary at finite height @xmath40 ( at @xmath68 ) is introduced , at which we impose a null vertical velocity @xmath69 and a constant tangential stress @xmath70 so that @xmath71 .
this corresponds to a physical situation where the fluid is entrained by a moving upper plate , for instance a stress - controlled couette annular cell .
then , we consider the limit @xmath72 , i.e. when the results become independent of @xmath40 . the lower boundary condition must be specified on the floor ( @xmath73 ) .
we consider here the limit in which the surface layer thickness @xmath22 is much smaller than the wavelength @xmath6 .
this allows to perform an asymptotic matching between the solution and the surface layer , whatever the dynamical mechanisms responsible for the hydrodynamical roughness @xmath20 are . indeed , focusing on the surface layer , we know that in the limit @xmath74 , the asymptotic behaviour of the local tangential velocity @xmath75 should be a logarithmic profile controlled by the local shear stress @xmath76 and the roughness @xmath77 .
the solution of ( [ systlinplaque ] ) should thus match this asymptotic behaviour as @xmath73 .
thus , @xmath20 is the only parameter inherited from the surface layer in the limit @xmath78 .
we will investigate the situation where this approximation is not valid anymore in section [ controlzzero ] . in the limit @xmath79 , the homogeneous solution of ( [ systlinplaque ] )
can be expanded in powers of @xmath53 and @xmath80 and expressed as the sum over four modes . adding the asymptotic behaviour of the particular solution @xmath81 , the full solution writes : @xmath82 the next terms in this expansion are @xmath83 . .
@xmath54 is the distance to the bottom , rescaled by the wavenumber . in all panels ,
the solid lines represent the real parts of the functions , whereas the dotted lines represent the imaginary ones .
dashed lines show the asymptotic behaviours ( [ uasymp]-[snasymp ] ) used as boundary conditions .
they match the solutions in the inner layer , which extends up to @xmath84 here .
we note @xmath85 and @xmath86 .
close to the boundary , a plateau of constant shear stress can be observed , which corresponds to the logarithmic zone .
it is embedded into a slightly larger zone of constant pressure in which the shear stress varies linearly . ] the values of the four coefficients @xmath87 , ... , @xmath88 are selected by the matching with the surface layer .
@xmath87 would correspond to a non vanishing normal velocity through the surface layer and should thus be null .
@xmath89 precisely corresponds to the logarithmic profile with a roughness @xmath20 and a basal shear stress modulation @xmath89 .
this gives @xmath90 .
@xmath91 would correspond to a modulation of the local roughness
more precisely of its logarithm .
we do not consider such a modulation so that @xmath92 .
@xmath88 corresponds to a sub - dominant behaviour associated to the basal pressure modulation ( @xmath93 ) . in summary
, the functions @xmath51 , @xmath52 , @xmath94 and @xmath95 should follow the following asymptotic behaviour : @xmath96 the region of thickness @xmath97 in which this asymptotic behaviour constitutes a good approximation of the flow field is called the inner layer .
equation ( [ snasymp ] ) means that the total pressure @xmath98 is constant across this boundary layer : @xmath99 and equation ( [ stasymp ] ) that the shear stress decreases linearly with height according to : @xmath100 the tangential pressure gradient is balanced by the normal shear stress , which means that inertial terms are negligible or equivalently that the fluid is in local equilibrium . in terms of energy ,
the space variation of the internal energy ( pressure ) is dissipated in turbulent `` friction '' .
these two equations correspond to the standard lubrication approximation for quasi - parallel flows . in practice , we solve the equations using a fourth order runge - kutta scheme with a logarithmic step . the integration is started at an initial value of @xmath53 inside the inner layer i.e. which verifies @xmath101 ) .
we write the solution as a linear superposition of the form @xmath102 , where the different terms verify : @xmath103 the boundary conditions on the bottom are then automatically satisfied , and the top ones give algebraic equations on the real and imaginary parts of @xmath104 and @xmath105 , which can be solved easily .
we have checked that the result is independent of the initial value of @xmath53 , as long as it remains in the announced range .
( aspect ratio @xmath106 ) , computed from the linearised equations ( @xmath107 ) .
the flow direction is from left to right .
note the left - right asymmetry of the streamlines around the bump in the inner layer ( grey lines ) .
note also the onset of emergence of a recirculation bubble in the troughs .
the thick line in the top right corner shows the positions that maximises the velocity along a streamline . ]
the velocity and stress profiles resulting from the integration of equation ( [ systlinplaque ] ) are displayed in figure [ modesverticaux ] .
looking at panel ( c ) , one can clearly see the region close to the bottom where the shear stress is constant , while the horizontal velocity component ( panel a ) exhibits a logarithmic behaviour .
this plateau almost coincides with the inner layer , which is the zone where the solution is well approximated by the asymptotic behaviour derived above .
the inner layer is embedded in a wider region characterised by a constant pressure ( panel d ) .
the estimate of the thickness @xmath97 is of crucial importance for the transport issue ( see next section and part 2 ) .
@xmath97 is the scale at which inertial terms are of the same order as stress ones in the reynolds averaged navier - stokes equations .
the original estimation of @xmath97 given by @xcite was further discussed in several later papers ( see e.g. @xcite ) .
our data are in good agreement with the scaling proposed by @xcite @xmath108 consistently , this scaling relationship is precisely that of the first neglected terms in the asymptotic expansion ( [ assymptotexpansion ] ) .
away from the bottom , all profiles tend to zero , so that one recovers the undisturbed flow field ( [ uzero ] ) at large @xmath53 .
the shape of these profiles are very consistent with the work of @xcite , who have compared the influence of the closure scheme on the linear flow over a relief , which means that the precise choice of the turbulent closure does not affect significantly the results . .
( a ) lin - lin plot in the shifted representation .
( b ) lin - log plot in the non - shifted representation .
the solid lines correspond to the real part and the dotted line to the imaginary one .
the velocity disturbance decreases exponentially over one wavelength ( dashed line ) . in this outer region
, the reynolds stress can be neglected . ] , @xmath109 , @xmath110 and @xmath111 as a function of @xmath112 .
these plots show the dependence of the basal shear and normal stresses with the number of decades separating the wavelength @xmath6 from the soil roughness @xmath20 , for a given bump aspect ratio .
the solid line corresponds to the results of the model , using the asymptotic matching with the surface layer .
the dashed lines represent the analytical formula deduced from @xcite by @xcite .
they agree well at very small @xmath113 . ] in order to visualise the effect of the bottom corrugation on the flow , the flow streamlines are displayed in figure [ streamlines ] ( see appendix [ streamlines ] for explanations about their computation ) .
it can be observed that the velocity gradient is larger on the crest than in the troughs as the streamlines are closer to each other .
the flow is disturbed over a vertical distance comparable to the wavelength .
a subtler piece of information concerns the position along each streamline at which the velocity is maximum .
these points are displayed in the right corner of figure [ streamlines ] .
away from the bottom , they are aligned above the crest of the bump . very close to it , however , they are shifted upstream . in other words ,
the fluid velocity is in phase with the topography in the upper part of the flow , but is phase advanced in the inner boundary layer where the shear stress tends to its basal value . in this inner layer , the profile is well approximated by its asymptotic expression ( [ uasymp ] ) .
an inspection of the velocity profile evidences two distinct regions ( see figures [ modesverticaux ] and [ perfectfluidinner ] ) , in which we recognise those at the basic partitioning of the flow in jackson & hunt work ( 1975 ) , and subsequent papers .
there is an outer region ( @xmath114 ) , where @xmath51 decreases exponentially with @xmath53 ( figure [ perfectfluidinner](b ) ) .
seeking for asymptotic solutions decreasing as @xmath115 , one has to solve the eigenvalue problem @xmath116 for asymptotically large values of @xmath53 .
at the two leading orders , the decrease rate @xmath117 is given by : @xmath118 the asymptotic behaviour is an oscillatory relaxation corresponding to @xmath119 .
however , the observed decrease corresponds to the intermediate asymptotic regime @xmath120 for which the solution is @xmath121 .
this behaviour is reminiscent from that of an inviscid potential flow . in other words ,
the effect of the turbulent shear stress on the flow disturbance can be neglected .
the intermediate region between the inner and the outer layers is responsible for the asymmetry of the flow as well as the upstream shift of the maximum velocity discussed above .
let us emphasise again that this is the physical key point for the formation of bedforms .
one can understand the reason of the phase shift with the following argument .
the external layer can be described as a perfect irrotational flow .
since the elevation profile is symmetric ,
the streamlines are symmetric too , as the flow is solely controlled by the balance between inertia and the pressure gradient induced by the presence of the bump . as a consequence ,
the velocity is maximum at the vertical of the crest .
now , inside the inner layer , this flow is slowed down by turbulent diffusion of momentum .
focusing on the region of matching between these outer and inner regions , the velocity needs some time to re - adapt to a change of shear stress , due to inertia .
thus , the shear stress is always phase - advanced with respect to the velocity .
one concludes that the basal shear stress is phase - advanced with respect to the bump . as mentioned in the introduction , we are especially interested in the shear stress and pressure distributions on the bottom .
we note @xmath85 and @xmath86 .
the ratio @xmath122 is the tangent of the phase shift between the shear stress and the topography .
it is positive as the shear stress is phase advanced .
the four coefficients @xmath123 , @xmath109 , @xmath110 and @xmath111 are displayed as a function of @xmath113 in figure [ abcdlin ] .
their overall dependence with @xmath113 is weak , meaning that the turbulent flow around an obstacle is mostly scale invariant .
more precisely , following jackson & hunt s work ( @xcite ) , it has been shown that , for asymptotically small @xmath113 , one expects logarithmic dependencies : @xmath124 where euler s constant is @xmath125 , @xmath126 is defined by the equation @xmath127 and with @xmath128 .
note that @xmath123 tends to @xmath129 and @xmath109 to @xmath130 as @xmath131 , as expected when the inner layer thickness @xmath97 vanishes . in this limit ,
the basal shear stress is directly proportional to the square of the velocity inherited from the outer layer , which is solution of the potential flow problem .
these expressions agree well with our numerical results for very small @xmath113 .
however , for realistic values of @xmath113 , e.g. @xmath132 , this approximation can not be accurately used as it leads to errors of order one note that jackson & hunt s expressions tend to diverge at larger @xmath113 . in comparison to @xmath123 and @xmath109
, we observe that the normal stress coefficients @xmath110 and @xmath111 are more robust with respect to the details of the model . in the limit of a perfect flow ,
the pressure varies as the square of the velocity . here
, one needs to consider the velocity at the scale @xmath6 of the perturbation , say @xmath133 , where the logarithmic factor @xmath134 should be evaluated for @xmath53 of order unity . from this argument
, we predict that the pressure coefficient @xmath110 should scale as the square of @xmath135 ( a parabola in figure [ abcdlin ] ) , which is very accurately verified . more precisely , @xmath136 ^ 2 $ ]
is an almost perfect approximation .
finally , it can be observed that the normal stress is also in phase advance with respect to the bottom profile .
the coefficient @xmath111 is positive and shows a linear variation with @xmath135 .
so far , the computation of the velocity and stress fields has been obtained without any specification of the physics at the scale of @xmath20 , as the integration of equation ( [ systlinplaque ] ) was started in the inner layer rather than on the bottom .
this is of course possible only if this layer is sufficiently thick , i.e. if @xmath97 ( or @xmath6 ) is much larger than the thickness of the surface layer @xmath22 introduced in section [ unif ] .
we now discuss several ways to describe the flow inside the surface layer , and investigate the subsequent effect on the shape of the stress coefficients as functions of @xmath113 .
these coefficients should be independent of the physics at work in this surface layer when @xmath113 is small enough , but we expect larger differences for larger values of @xmath113 .
we first present a convenient phenomenological model of geometrically induced roughness , which does not involve additional parameters .
because of its simplicity , it will be used in the next sections , as well as in the second part of the paper .
the results will be compared to a rigourous treatment resulting from the weakly non - linear hydrodynamical calculation in section [ nl ] .
we then consider the case of a viscous surface layer .
inspired from the aeolian transport properties , we finally discuss the focus point assumption as a possible way to describe the situation in which the surface layer is governed by the presence of sediment transport . and
@xmath109 ( dashed lines ) computed with the phenomenological model of geometrically induced roughness ( [ uzerogeom ] ) . for comparison
, the solid lines display the reference case shown in figure [ abcdlin ] . ] for an hydrodynamically rough bottom , the `` small scale '' roughness elements are larger than the viscous sub - layer . they are submitted to a turbulent drag from the fluid and reciprocally , their presence slows down the flow .
the exchanges of momentum in the surface layer are thus dominated by the turbulent fluctuations . following @xcite and others ,
a convenient phenomenological model is to define the mixing length involved in the turbulent closure ( [ tauijprandtl ] ) as @xmath137 . in this way
, @xmath13 is still essentially the geometrical distance to the bottom , except that it can not be smaller than the roughness length .
this choice reflects in a intuitive manner the physical picture one can infer from experiments or simulations where square - shaped roughness elements are glued on a flat wall ( see e.g. @xcite ) .
we will show in section ( [ nl ] ) that this picture must be refined when dealing with blunt roughness elements of moderate aspect ratio for instance the surface of a sand bed . with this expression for the mixing length ,
the integration of starting equations in the uniform and steady case gives @xmath138 where the lower boundary condition @xmath139 can now be taken in @xmath140 .
this expression is well approximated by the pure logarithmic profile ( eq . [ uzero ] ) as soon as @xmath1 is larger than , say , few @xmath20 .
in other words , for this model , @xmath141 . the above description of the linear analysis , and in particular the expression of the matrix @xmath142 and the vector @xmath143 involved in ( [ systlinplaque ] ) , in the case of a wavy bottom is still valid , but now with the following expression for the function @xmath144 the solution of ( [ systlinplaque ] ) can again be written as a linear superposition @xmath102 , where these three vectors are solutions of @xmath145 this decomposition ensures the requirement that both components of the velocity vanish on the bottom , leading to @xmath146 and @xmath147 . as in the previous section , the coefficients @xmath104 and @xmath105 are found by the upper boundary conditions .
the coefficients @xmath123 and @xmath109 resulting from this integration are displayed in figure [ abcdgeom ] .
one can see that , for @xmath148 , they are not very much different from those obtained in the previous section
. however , one can notice significant differences for @xmath149 . as the mixing length in the surface layer is larger in this case ( @xmath150 ) than in the asymptotic case ( @xmath151 ) , the turbulent ` diffusion ' is more efficient .
this results into a larger phase advance for the shear stress ( fig .
[ abcdgeom ] c ) . for practical purposes and for later use in the second part of this paper , a very good empirical fit of the coefficients @xmath123 and @xmath109
is obtained with @xmath152 with @xmath153 , @xmath154 and @xmath155 .
and @xmath109 computed with a viscous surface layer .
dotted line : @xmath156 ; dashed line : @xmath157 . as in figure
[ abcdgeom ] , the solid lines display the reference case of figure [ abcdlin ] . ] in hydraulically smooth situations , it is natural to expect that , very close to the bottom , the flow must be laminar and thus described by the equation @xmath158 whose solution is @xmath159 we thus neglect here the possibility of a phase shift across the viscous surface layer .
the transition from viscous to turbulent regime is governed by the reynolds number and occurs at a typical value @xmath160 .
the surface layer thickness can then be easily computed as @xmath161 . at @xmath162 , both viscous and turbulent expressions for the velocity must coincide : @xmath163 from this equality , we can deduce the hydrodynamical roughness seen from the inner layer , due to this viscous surface layer : @xmath164 in the case of a sand bed , the transition between the hydrodynamically smooth and rough regimes occurs when the viscosity induced roughness ( eq .
[ z0viscous ] ) is of the order of the geometrically induced roughness ( @xmath165 ) . with the corresponding value for @xmath112
, we solve equation ( [ systlinplaque ] ) in the usual manner , writing the solution in the form of the linear superposition as described above , except that the integration is started at the initial value @xmath166 , in which we impose that the velocity is parallel to the bed and equal to @xmath167 . at linear order ,
this leads to @xmath168 the resulting shear stress coefficients @xmath123 and @xmath109 are displayed in figure [ abcdhfufviscous ] . as one can expect , in comparison to the reference case , they are smaller for larger values of @xmath169 , and all different curves collapse as @xmath131 .
the viscous diffusion of momentum is less efficient than that induced by turbulent fluctuations .
moreover , in the stokes regime , for reynolds numbers much smaller than @xmath170 , the kinematic reversibility leads to a shear stress in phase with the topography .
consistently , it can be observed in figure [ abcdhfufviscous](c ) that the phase advance is reduced in the hydrodynamically smooth regime .
we will show in the second part of this article that this explains the fact that the wavelength at which ripples appear is larger as the reynolds number decreases .
experiments in the hydraulically smooth regime have been performed by @xcite , who measured the ionic mobility between two nearby electrodes .
this current is assumed to be related , without any spatial or temporal lag , to the basal shear stress .
the measured phase shift between the signal and the bottom topography could reach values as high as @xmath171 .
this would correspond to @xmath172 .
within the present model , the phase shift remains much lower than the measured @xmath171 .
this unexpected value has been interpreted as the signature of a lag of the laminar - turbulent transition with respect to the reynolds number criterion @xmath173 .
further experiments based on a different measure principle are needed to understand this discrepancy .
this viscous surface layer model is an effective way to take bedload transport into account . as a matter of fact , anticipating on the part 2 of this paper where the dynamical mechanisms governing the sediment transport are discussed , transported particles are not passive and exert a stress on the fluid .
close to the transport threshold , their influence on the flow is negligible .
however , as their density increases , transport induces a negative feedback on the flow , which should be taken into consideration in the hydrodynamics description .
the simplest model of multi - layer sheet flow would be a newtonian fluid whose viscosity increases with the concentration of moving sediments . in this large shear velocity regime , one thus expects a decrease of the phase - lag responsible for the ripples instability and possibly , a restabilisation of the bed . and @xmath109 computed in the presence of a ` focus point ' at height @xmath22 , where the velocity is @xmath167 , as a function of @xmath113 .
@xmath123 and @xmath109 are larger for larger values of @xmath174 ( @xmath170 , @xmath129 , @xmath175 and @xmath176 ) . however , the ratio @xmath122 is less sensitive to this parameter , up to @xmath177 .
again , the solid lines display the reference case of figure [ abcdlin ] . ]
an alternative manner to take the feedback of the transport on the flow into account can be achieved in analogy with the aeolian case , which provides the archetype of such a situation . in this case
, it has been shown that the moving grains slow down the flow in the transport layer , whose thickness @xmath22 is independent of the shear velocity @xmath178 .
note that in the subaqueous case , the transport layer thickness is observed to gently increase with the shear stress ( @xcite ) close to the threshold , in the erosion limited regime ( see part 2 ) . above @xmath22 , the effect of the particles on the flow is negligible and
one recovers the undisturbed logarithmic velocity profile , but with a roughness larger than that without transport .
below @xmath22 , the flow velocity is reduced and is independent of @xmath178 ( @xcite ) . as shown experimentally by @xcite
, the velocity vertical profiles measured for different shear velocities thus cross at the ` focus point ' @xmath162 and @xmath179 . at this point
we have @xmath180 which means that the effective roughness in the logarithmic region , due to this transport layer , is @xmath181 to determine the flow field in such a situation , the crucial point is to compare @xmath22 with the thickness of the inner layer @xmath97 , i.e. the size of the constant stress plateau ( see figure [ modesverticaux ] ) .
if @xmath22 is larger than @xmath97 , it means that one can not reduce the transport issue to a relationship between the sediment flux and the basal shear stress only . in that case , the whole vertical velocity profile , which depends on the entire bottom elevation , is involved .
conversely , for @xmath182 , one can account for transport by modifying the bottom boundary conditions as follows .
following what we have done in the previous sub - section , we can impose that the fluid velocity at @xmath183 is parallel to the bed and equal to @xmath167 . at the linear order
, we then get : @xmath184 the result of this choice is shown for the stress coefficients in figure [ abcdhfuf ] for various values of @xmath174 . @xmath123 and @xmath109 are larger for increasing focus velocities , or equivalently larger focus altitude . as in the viscous surface layer case , all curves collapse for @xmath131 because @xmath97 gets larger in this limit ( see equation ( [ ellscaling ] ) ) .
interestingly , as far as bedforms are concerned ( see part 2 ) , the ratio @xmath122 is much less sensitive to variations of @xmath174 , at least in the region @xmath185 .
finally , it should be noted that the focus point model only applies to the momentum limited transport regime ( see part 2 ) .
close to the transport threshold , in the erosion limited regime , the feedback of the particle transport on the flow is negligible and the transport should not to be taken into account in the hydrodynamical calculation .
this approach should be distinguished from that used by @xcite . in those articles ,
the flow boundary conditions are taken on the bottom , below the transport layer ( vanishing velocities ) , meaning that the feedback of the transport on hydrodynamics is neglected .
however , the stresses are evaluated in @xmath22 , above the transport layer .
this does not constitute a self - consistent choice .
moreover , this introduces a free parameter in the model which can be tuned to choose the phase shift at will .
we will present in the second part experimental evidences that this choice is not correct .
, which is a non dimensional parameter encoding the time lag between a change in the strain rate and that of the reynolds stress .
( a - c ) vertical profiles of the horizontal component of the velocity @xmath186 for @xmath107 for ( a ) @xmath187 , ( b ) @xmath188 and ( c ) @xmath189 respectively .
one can see that the profiles develop oscillations as @xmath190 increases , but that the behaviour close to the bottom ( in log scale ) remains the same . in panels ( d ) and ( e ) , we plot the coefficients @xmath123 , @xmath109 , @xmath110 and @xmath111 _ vs _ @xmath190 ( still for @xmath107 ) .
they are weakly affected , meaning again that the behaviour close to the bottom is almost unchanged . ] for these three dynamical mechanisms controlling the hydrodynamical roughness @xmath20 , we have seen that the asymptotic regime is recovered when the surface layer thickness @xmath22 is smaller than the inner layer thickness @xmath97 . as @xmath97 is much smaller than the wavelength @xmath6 ( for standard bedforms , @xmath191 ) , this constitutes a rather restrictive condition .
whenever @xmath22 is larger than @xmath97 , specific hydrodynamic models should be derived to determine the relations between stresses and topography . the phase shift between the basal shear stress and the topography originates from the interplay between inertia and shear stress .
the different models of surface layer correspond to different ways of mixing momentum in the direction normal to the wall .
although the argument is general , one sees that the precise value of the phase shift is rather sensitive to the physical origin of the momentum fluxes .
in particular , viscous diffusion leads to a much smaller phase advance than turbulent mixing .
in the same spirit as the previous section , where we have investigated the effect of different ways to treat the surface layer on the stress coefficients @xmath123 and @xmath109 , we would like now to show the robustness of the results when considering a second order turbulent closure , reynolds stress anisotropy or a moving bottom . as already stated , to solve quantitatively the ` dune problem ' , we need to take correctly into account the effects inducing a phase shift between the stresses and the relief . as a matter of fact , a first order closure assumes that the turbulent energy adapts instantaneously to the mean strain tensor . to take into account the lag between the stress and the strain tensors
, one needs to formulate a second order turbulent closure .
this can be achieved by deriving dynamical equations for @xmath192 , which relax the stresses towards their steady state expression prescribed by equation ( [ tauijprandtl ] ) ( see appendix [ 2ndorder ] ) .
@xmath193 . \label{2ndme}\ ] ] in this relation , the parameter @xmath190 encodes the time lag between an increase of the mean shear strain rate and the corresponding re - adjusment of the fluctuations of the shear stress .
we expect @xmath190 to be on the order of unity . in figure [ betaeffect ]
, we show the effect of this new parameter .
as expected for inertial effects in a relaxation process , finite values of @xmath190 generate oscillations in the vertical profiles of the velocities and stresses .
the example of the horizontal velocity is displayed in the panels ( a ) , ( b ) and ( c ) .
the amplitude and the frequency of these oscillations increase with @xmath190 .
interestingly , these oscillations do not affect much the behaviour of the modes close to the bottom . as a consequence , the coefficients @xmath123 , @xmath109 , @xmath110 and @xmath111
are weakly affected by @xmath190 , see panels ( d ) and ( e ) .
interestingly , both @xmath123 and @xmath109 decrease as @xmath190 increases and their ratio remains roughly constant .
@xmath190 has thus a negligible effect on the emergence of bedform , and we shall keep it to zero in the rest of the paper , as well as in part 2 . , @xmath109 , @xmath110 and @xmath111 as a function of @xmath113 for different values of @xmath194 ( @xmath195 , @xmath170 , @xmath175 and @xmath176 ) .
the dashed lines correspond to the isotropic case .
inset ( e ) : ratio @xmath122 .
arrows indicate increasing normal stress anisotropy . ]
it is an experimental fact that , in a turbulent boundary layer close to a rough wall , the reynolds stress tensor is _ not _ isotropic : @xmath196 is significantly larger than the other components ( @xcite ) .
besides , anisotropy seems less pronounced for a larger bottom roughness ( @xcite ) , an issue which is however still matter of debate ( @xcite ) . to account for this reynolds stress anisotropy
, it is easy to generalise the prandtl - like first order turbulent closure ( [ tauijprandtl ] ) with the following expression : @xmath197 where the value of @xmath198 now differs from that of @xmath199 . following the above - cited literature
, we expect @xmath200 to be around @xmath201@xmath202 .
the modification of the matrix @xmath142 due to this new closure is detailed in appendix [ avi ] .
it is shown that the relevant anisotropic parameter entering the equations is @xmath194 , for which a realistic value is on the order of unity . as evidenced in figure [ abcdanis ] , the corresponding values of the functions @xmath123 , @xmath109 , @xmath110 and @xmath111
are not much affected by this anisotropy in the relevant range of @xmath113 .
this is particularly true for the coefficients @xmath110 and @xmath111 , as well as for the ratio @xmath122 as soon as @xmath203 .
the normal stress anisotropy has thus a negligible influence on ripple and dune formation , and we will not take it into account in the rest of the paper , as well as in part 2 . and @xmath109 as a function of @xmath113 for different values of the bottom growth rate @xmath117 ( panels ( a ) and ( b ) ) , and different values of the bottom pulsation @xmath204 ( panels ( c ) and ( d ) ) .
arrows indicate increasing values of @xmath117 and @xmath204 . in panels ( a ) and ( b )
, grey dotted lines correspond to @xmath205 , @xmath175 , @xmath129 , @xmath170 and @xmath206 , the black dotted line being for @xmath207 .
the black dashed line corresponds to @xmath208 , the grey dashed lines being for @xmath209 , @xmath210 , @xmath211 , @xmath212 and @xmath213 . in panels ( c ) and
( d ) grey dotted lines correspond to @xmath214 and @xmath170 , the black dotted line being for @xmath215 .
the black dashed line correspond to @xmath216 , the grey dashed line being for @xmath217 and @xmath211 . for comparison , in all panels the solid lines correspond to the static case @xmath218 , @xmath219 . ] in order to investigate the effect of a moving bottom on the stress coefficients , we consider a bottom profile of wavevector @xmath220 like in ( [ bottomprofile ] ) , but which is now function of both position @xmath18 and time @xmath221 : @xmath222 in this expression , @xmath117 represents the growth rate of the profile , and @xmath223 its phase velocity . as discussed in @xcite ,
this investigation is important as we wish to use the present hydrodynamical study in the context of the formation and development of bedforms , which do have a ( small ) growth rate and a ( small ) velocity .
following expression ( [ movingbottom ] ) , we modify those for the functions @xmath51 , @xmath52 , @xmath94 and @xmath95 by inserting the extra - factor @xmath224 , along the lines of ( [ defu]-[defsn ] ) . in this new case , the linearised equations ( [ nsxlin ] ) and ( [ nszlin ] ) of section [ unboundedcase ]
must then be modified in the following manner : @xmath225 the linear equation ( [ systlinplaque ] ) is then the same , but now with the modified matrix @xmath226 with the surface layer model described in section [ geom ] , the non - slip boundary conditions on the bottom can be written as @xmath227 the result of the integration of this new system is shown in figure [ abcdsigmaomega ] for the shear coefficients @xmath123 and @xmath109 .
one can see that departure from the static case @xmath218 and @xmath219 is noticeable only for values of @xmath228 and @xmath229 of order one .
the effect of the wave propagation of the bedform can be understood by a simple argument . when the bedforms propagate upstream ( @xmath230 ) the relative flow velocity seen by the structure is larger so that it induces a larger shear stress modulation .
as @xmath231 is by definition the basal shear stress rescaled by @xmath232 , both @xmath123 and @xmath109 get larger .
reciprocally , when the bedforms propagate downstream ( @xmath233 ) , these coefficients are reduced . consistently with this argument , the ratio @xmath122 is only weakly affected by @xmath204 ( not shown ) .
the growth rate @xmath117 affects @xmath123 and @xmath109 in opposite ways and thus changes the phase shift between the shear stress and the topography . for @xmath234 , @xmath123
is increased while @xmath109 is reduced .
we have not been able to interpret this behaviour in a simple way . as discussed in part 2 , for ripples in water
flows these dimensionless numbers are respectively on the order of @xmath235 and @xmath236
. they would be even smaller for bedforms of larger wavelength . as a consequence ,
the assumption that the bottom can be treated as fixed for the study of bedforms is definitively valid ( see also the discussion of figure [ deltasigmaomega ] below ) .
in this section , we investigate the non - linear effects at finite values of the rescaled bottom corrugation amplitude @xmath61 .
in particular , we wish to address two issues : we wish to relate the hydrodynamic roughness to geometrical quantities ; we aim also to describe the first non - linear corrections to the basal stress coefficients @xmath123 , @xmath109 , @xmath110 and @xmath111 .
these results will be used in the second part of this work , to determine the equilibrium height of ripples . in this context , few authors have studied the influence the non - linear terms from hydrodynamics on the shape ( @xcite ) or the stability ( @xcite ) of the bedforms .
in contrast to the present paper , however , both of these works describe the turbulent closure with a constant eddy viscosity . in a very empirical manner , @xcite computed the flow over two - dimensional bedforms of arbitrary amplitude by the use of a wake function , as described in @xcite , coupled with a potential flow modified to take into account the velocity logarithmic law . finally , the linear results of @xcite have also recently been extended to the weakly non - linear situation ( @xcite ) . for this purpose ,
we perform an expansion with respect to the bottom corrugation amplitude , and introduce non - dimensional the following functions for the different orders : @xmath237 , \\ u_z & = & u _ * \left[(k\zeta ) e^{ikx }
w_1+(k\zeta)^2 w_{0}+(k\zeta)^2 e^{2ikx } w_{2}+(k\zeta)^3 e^{ikx } w_{3}\right ] , \\ \tau_{xz } & = & - u_*^2 \left[1+(k\zeta ) e^{ikx } s_{t1}+(k\zeta)^2 s_{t0}+(k\zeta)^2 e^{2ikx } s_{t2}+(k\zeta)^3 e^{ikx } s_{t3}\right ] , \\ p+\tau_{zz } & = & p_0+u_*^2 \left[(k\zeta ) e^{ikx } s_{n1}+(k\zeta)^2 s_{n0}+(k\zeta)^2 e^{2ikx } s_{n2}+(k\zeta)^3 e^{ikx } s_{n3}\right],\\ \tau_{zz}- \tau_{xx } & = & u_*^2 \left[(k\zeta ) e^{ikx } s_{d1}+(k\zeta)^2 s_{d0}+(k\zeta)^2 e^{2ikx } s_{d2}+(k\zeta)^3 e^{ikx } s_{d3}\right].\end{aligned}\ ] ] terms in @xmath238 , which do not contribute to the harmonic response ( i.e. to the stress coefficients ) , as well as terms of higher order than @xmath239 are neglected .
although the principle of the expansion in amplitude is simple , the actual calculations are painful , and the technical details have been gathered in appendix [ wnlc ] . in summary
, the non - linear effects result from the expansion of the mixing length ( terms in @xmath240 ) and from the self - interaction of the linear perturbations : in particular , the combination of terms @xmath241 generates second order terms in @xmath240 .
all involved functions are complex , except @xmath134 and those related to the second order homogeneous corrections ( index @xmath130 ) . to avoid the determination of the asymptotic behaviours in this case , we have chosen to treat the surface layer by the simple phenomenological model of section [ controlzzero ] . plugging the above expressions into the navier - stokes and turbulent closure equations , one eventually obtains a linear hierarchy of linear differential equations : @xmath242 where @xmath243 .
of course , @xmath244 and @xmath245 are the matrix and vector of expression ( [ systlinplaque ] ) .
we find that @xmath246 , whereas @xmath247 is slightly different and @xmath248 is very simple : @xmath249 @xmath250 all the heaviness of the method is encoded in the expressions of the right hand terms @xmath251 : the components of such vectors at a given order depend on the lower order functions @xmath252 and their derivatives .
the integration has thus to follow the hierarchy of the equations , one order after the other .
( a ) and @xmath253 ( b ) .
these curves have been computed with @xmath254 .
@xmath255 tends towards a negative constant value @xmath256 at large @xmath55 , which corresponds to an increased roughness at large distance from the wall .
notice also that , close to the bottom , @xmath253 has a constant value , reminiscent of the inner layer . ] the boundary conditions must be specified in order to perform the integration . following the geometrically induced roughness model ( section [ controlzzero ] )
, we require that both components of the velocity should vanish on the bottom .
these conditions express easily in the shifted representation , i.e. written in terms of the curvilinear coordinates ( see appendix [ rep ] ) : they simply read @xmath257 and @xmath258 . in terms of the cartesian functions , we get : @xmath259 , \\
u_2(0 ) & = & -\frac{1}{4 } \mu''(0 ) - \frac{1}{2 } \kappa \mu'^2(0 ) - \frac{1}{4 } \mu'(0 ) s_{t1}(0 ) , \\ u_3(0 ) & = & -\frac{1}{8 } \mu'''(0 ) + \frac{9}{8 } \mu'(0 ) - \frac{3}{4 } \kappa^2 \mu'^3(0 ) -\frac{1}{16 } \left [ \mu''(0 ) + 2 \kappa \mu'^2(0 ) \right ] \left [ 2s_{t1}(0)+s_{t1}^*(0 ) \right ] \nonumber \\ & & + \frac{1}{32 } \mu'(0 ) s_{t1}(0 ) \left [ s_{t1}(0)+2s_{t1}^*(0 ) \right ] - \frac{i}{16 } \mu'(0 ) \left [ 2s_{n1}(0)+s_{n1}^*(0 ) \right ] \nonumber \\ & & - \frac{1}{4 } \mu'(0 ) \left [ 2s_{t0}(0)+s_{t2}(0 ) \right ] , \end{aligned}\ ] ] and @xmath260.\end{aligned}\ ] ] for @xmath261 , @xmath262 and @xmath263 .
the extrapolation to @xmath130 of the upper part of the curves gives the value of the hydrodynamical roughness seen at a distance from the bottom .
( b ) coefficient @xmath264 as a function of @xmath113 .
a good fit of this function is given by @xmath265 . ] as in previous sections , for each order , the solution is expressed as a linear superposition of the form : @xmath266 , where the different vectors are solutions of the following equations : @xmath267 again , for the top boundary conditions , we introduce a lid at finite height @xmath40 , impose @xmath268 and @xmath269 , and look at the limit @xmath72 , i.e. when the results become independent of @xmath40 .
these conditions lead to two equations on @xmath270 and @xmath271 , whose solutions give @xmath272 and @xmath273 respectively .
( a ) and @xmath274 ( b ) .
the solid lines represent the real parts of the functions , whereas the dashed lines represent the imaginary ones .
these curves have been computed with @xmath254 .
notice again the inner layer where the stresses are constant .
we write @xmath275 and @xmath276 .
note that both @xmath277 and @xmath278 are negative . ]
we first consider the corrections to the homogeneous base solution ( index @xmath130 ) .
the corresponding velocity profile @xmath279 and shear stress profile @xmath253 are displayed in figure [ modenlzero ] .
the term in @xmath240 of the velocity decreases continuously from @xmath280 to @xmath281 and tends towards a negative constant @xmath256 far from the bottom
. correspondingly , the shear stress decreases and tends to @xmath130 far from the ground , as requested .
the calculation thus predicts an increase of the turbulent drag ( i.e. of the basal shear stress ) with the corrugation amplitude , due to the non - linearities . in terms of the velocity profile
, this translates into a hydrodynamic roughness @xmath282 of geometrical origin . identifying the expression of the velocity profile far from the bottom with @xmath283
, we simply get : @xmath284 as a consequence , @xmath282 increases with @xmath264 and with the aspect ratio @xmath48 . for the seek of illustration ,
several vertical profiles of the homogeneous part of the velocity @xmath285 are plotted in figure [ rugose](a ) for different values of @xmath48 . by definition , @xmath282 is the extrapolation of the asymptotic part of the curves to vanishing velocities .
interestingly , the large scale roughness @xmath282 can not be related to a single geometrical length , namely to the corrugation amplitude @xmath43 ( @xcite ) .
in particular ,
we predict that the macroscopic roughness @xmath282 associated to a wavy surface still depends on the microscopic roughness @xmath20 : as shown in figure [ rugose](b ) , for a given aspect ratio , the apparent roughness @xmath282 is larger for smaller @xmath113 . furthermore , expression ( [ defzg ] ) is consistent with numerical observations of @xcite and computations of @xcite , who respectively report linear ( in the range @xmath286 ) and quadratic variations for the coefficient @xmath264 as a function of @xmath287 . here ,
the best fit of this function gives @xmath288 .
, @xmath278 , @xmath289 and @xmath290 as a function @xmath113 .
comparing the signs with those of @xmath291 , @xmath292 , @xmath293 and @xmath294 , it can be inferred that the non - linearities oppose the linear effects . in particular , as the amplitude increases , the point of maximum shear stress drifts downstream , i.e. @xmath295 decreases . ]
the first non - linear corrections to the harmonic terms scale on @xmath239 . in figure
[ modenlthree ] , we show the corresponding profiles for the stresses . as in the first order case , there exists a layer close to the bottom where the stresses are almost constant ( inner layer ) . as requested , both components vanish far from the ground .
we note @xmath275 and @xmath276 the shear and normal stresses acting on the boundary .
these coefficients are plotted as a function of @xmath113 in figure [ abcdthree ] .
both @xmath277 and @xmath278 are negative while @xmath291 and @xmath292 are positive , which means that the first non - linearities oppose the linear effects .
we will show in part 2 that this is responsible for the selection of the height of current ripples .
the calculation of the non - linear corrections allows to determine the range of amplitude @xmath43 for which the linear approximation is valid .
the representation of the linear solution with the fixed system of coordinates ( @xmath296 ) is valid only when @xmath43 is much smaller than @xmath20 .
however , the representation of the same linear solution in the curvilinear coordinates ( @xmath297 ) is valid up to @xmath43 of the order of the inner layer thickness @xmath97 .
it is important to recall that all the descriptions of the flow equivalent at the linear order ( including the real solution of the fully non - linear problem ) can have very different domains of validity .
this weakly non - linear computation is illustrated in figure [ streamlinesnonlinear ] , which shows the streamlines for different aspect ratios .
it can be seen that the separation of streamlines and the subsequent formation a recirculation bubble occurs above an aspect ratio of @xmath298 , in agreement with observations . by comparison ,
the linear calculation , shown in figure [ streamlines ] , leads to the emergence of a recirculation bubble for an aspect ratio of @xmath299 i.e. twice larger .
the non - linear corrections are thus essentials to capture quantitatively the flow features .
panel [ streamlinesnonlinear]c ) , computed for an aspect ratio of @xmath300 shows a well - developed recirculation bubble .
the distortion of the separation streamline is not realistic , indicating the upper limit of validity of the model .
fortunately , the aspect ratio of ripples and dunes is typically smaller than @xmath0 , which falls into the domain of validity of the calculation .
several experiments in flumes or wind tunnels have been performed to measure velocity and reynolds stress profiles over two - dimensional fixed symmetric or asymmetric bedforms ( @xcite ) .
closer to our calculations , several experiments with a sinusoidal bottom are also reported in the literature ( e.g. @xcite )
. direct or large eddy numerical simulations of navier - stokes equations have also been performed in this geometry ( e.g. @xcite ) . for comparison with our model , data from @xcite
have been chosen .
they have been performed in a wind tunnel over sixteen waves with a wavelength @xmath301 mm and a trough - to - crest amplitude @xmath302 mm , covered with a carpet to make them aerodynamically rough .
this corresponds to an aspect ratio of @xmath299 much too large to be in the domain where the model is quantitative .
unfortunately , we have not found any better data - set for the seek of comparison .
the vertical profiles of the velocity measured at different locations are shown in figure [ nonlinearexperiment](a ) .
more precisely , these authors have measured the average of the instantaneous velocity modulus i.e. a quantity that is always positive and that mixes the average velocity and the fluctuations .
although a recirculation bubble is present , these data can not show it .
we compare these profiles to those computed at the upper limit of validity of the non - linear calculation ( @xmath263 ) . yet ,
the agreement with our computation is fair ; in particular , the way the flow is accelerated over crests and decelerated in troughs in qualitatively well reproduced .
the profiles at @xmath303 and @xmath304 from the crest are close to each other , indicating a re - symmetrisation of the flow by non - linearities .
note that the slight difference between these profiles is qualitatively reproduced by the model .
looking at the upper part of the experimental profiles , one sees that they would extrapolate to @xmath130 around @xmath305 mm while the ground roughness is slightly smaller than @xmath170 mm .
these two roughness are particularly visible on the profile measured on the crest .
the model is particularly successful in reproducing this feature .
the non - linear effects on the flow over obstacles are often described in terms of boundary layer separation .
it has been proposed by @xcite ( see also @xcite ) that one could still use the linear flow calculation in that case , introducing a fictive surface enveloping the obstacle and the recirculation bubble downstream of it .
as such an envelope creates a fictive bump maximum displaced downstream , it artificially moves the point of maximum shear stress on the bump in the same direction ( @xcite ) .
although this trick is of practical use to simulate dunes , this envelope technique is not based on any firm theoretical ground .
the weakly non - linear calculation performed here is thus of extreme interest to incorporate non - linear turbulent effects in dune numerical models in a more controlled way .
more generally , it can be used in any problem in which a good approximation of the mean flow is needed at low calculation cost , including separation .
for instance , it may find direct applications in the control of turbulence around vehicles .
an important limit of such reynolds averaged calculation is that they do not take vortex shedding into account . ,
@xmath306 and @xmath307 from top to bottom .
the fluid flows from left to right .
note the progressive development of a recirculation bubble for larger acpect ratios . ]
in this section , we investigate the effect of the additional presence of a free surface at a finite distance @xmath40 to the bottom .
this situation is relevant to the flow above river dunes ( see part 2 ) .
we follow the outline of the section [ unboundedcase ] , but staying for easiness with linear calculations in two - dimensional situations . in the case of a river inclined at an angle @xmath308 on the horizontal ,
the shear stress must balance gravity .
it thus varies linearly as @xmath309 and vanishes at the free surface . by definition of the shear velocity @xmath178
, we also write @xmath310 . in the context of a mixing length approach to describe turbulence ,
this length should vanish at the free surface .
for the sake of simplicity , following the discussion of section [ controlzzero ] , we take @xmath311 .
this choice results in a base flow that is logarithmic as in the unbounded situation : @xmath312 which is consistent with field and experimental observations .
the stress balance equation along the @xmath1-axis allows to get the pressure , which reads : @xmath313 we define the froude number as the ratio of the surface velocity @xmath314 to the velocity of gravity surface waves in the shallow water approximation : @xmath315 in the literature , the froude number is sometime defined as the ratio of the mean velocity to the velocity of gravity waves .
we will justify this choice in the next paragraph .
the froude number of natural sandy rivers lies in general between @xmath195 and @xmath306 as they flow on very small slopes .
larger froude numbers are reached in flume experiments .
( @xmath316 ) .
@xmath317 is the free stream reference velocity .
the symbols correspond different longitudinal locations , as shown in the schematics below the data ( crests @xmath318 , troughs @xmath319 and half ways up- @xmath320 and down - stream @xmath321 ) .
( b ) velocity profile predicted by the present model with a sinusoidal bottom of aspect ratio @xmath0 ( @xmath322 ) , plotted with the same symbol code . dotted line : base velocity profile . ] in the same manner as in section [ unboundedcase ] , we consider now a wavy bottom @xmath323 .
we note again @xmath324 and @xmath68 .
we write the first order corrections to the base flow as @xmath325 , \label{defufs } \\
u_z & = & u _ * k\zeta e^{ikx } w , \label{defwfs } \\
\tau_{xz } & = & \tau_{zx}= - u_*^2 \left[1 - \frac{\eta}{\eta_h } + k\zeta e^{ikx } s_t \right ] , \label{defstfs } \\ p+\tau_{zz } & = & p_0+u_*^2 \left[\frac{1}{\tan\theta } \left(1-\frac{\eta}{\eta_h}\right)+k\zeta e^{ikx } s_n \right ] , \label{defsnfs}\end{aligned}\ ] ] where , in accordance with equation ( [ uxfreesurface ] ) , the function @xmath134 is defined by the relation ( [ mu2etageom ] ) .
the free surface is also disturbed by the presence of the non - uniform bottom , and we denote @xmath326 the flow depth at the position @xmath18 .
the modified expression for the mixing length then reads @xmath327 linearising the free surface profile as @xmath328 , one can expand @xmath13 to the first order as @xmath329 \right \}.\ ] ] the shear stress closure as well as the reynolds averaged navier - stokes equations can be linearised in the same way as before , and we finally get at the first order in @xmath61 a system of differential equations which can be written under the following form : @xmath330 with @xmath331 again , making use of the linearity of the equations , we seek the solution under the form @xmath332 , where the vector @xmath333 is solution of equation : @xmath334 while @xmath335 , @xmath336 and @xmath337 are still solutions of equations ( [ equaxsgeom])-([equaxngeom ] ) .
the bottom boundary conditions @xmath338 and @xmath339 are then automatically satisfied . at the free surface , we impose the material nature of the surface , @xmath340 , and vanishing stresses : @xmath341 and @xmath342 .
these last three conditions select the coefficients @xmath343 and @xmath344 as well as the value of @xmath345 .
finally note that the analytical approximation of the solution close to the bottom in the limit @xmath131 is the same as in the unbounded case it does not depend on the position of the upper boundary and expressions ( [ uasymp])-([snasymp ] ) are thus still correct in the limit @xmath346 . and
@xmath109 as functions of @xmath113 for @xmath347 and @xmath348 ( solid line ) or @xmath349 ( dotted line ) . in the right part of the plots , the curves collapse on the shape displayed in panels ( a ) and ( b ) of figure [ abcdgeom ] .
they differ at smaller @xmath113 , showing a resonance peak and a divergence at @xmath131 . ] ( a)-(b ) , and @xmath350 ( c ) , with @xmath351 .
the velocity profiles ( bold lines ) are compared to their asymptotic behaviour in the inner and outer layers ( thin lines ) .
the solid lines represent the real part of the functions , and the dotted lines the imaginary ones .
( a ) and ( b ) show the very same profile , but with a logarithmic scale in ( a ) to emphasize the inner region .
the thin lines in ( a ) and ( c ) represent the asymptotic behaviour in the inner layer . those in ( b )
correspond to a sum of an increasing and a decreasing exponential of the form @xmath352 , as for an inviscid potential flow . ] in order to evidence the role of the free surface , we have plotted the stress coefficients @xmath123 and @xmath109 as functions of @xmath113 in figure [ stretchlimo ] , for different values of @xmath353 . for a large enough wave - number @xmath220 ( a small enough wavelength @xmath6 ) , one recovers the plots of the panels ( a ) and ( b ) of figure [ abcdgeom ] , independently of @xmath353 .
this means that for a bottom wavelength much smaller than the flow depth @xmath40 ( i.e. for subaqueous ripples ) , the free surface has a marginal effect and the results of section [ unboundedcase ] apply . for smaller @xmath113 , however , the curves exhibit a peak , whose position depends on the value of @xmath353 , followed by a diverging behaviour when @xmath131 . as discussed below
, this peak can be ascribed to a resonance of standing waves at the free surface , excited by the bottom topography , meaning that the proper scale is now @xmath40 and not @xmath20 . as the ratio @xmath354 is the key parameter separating ripples from dunes , we shall turn extensively in part 2 to this point .
the analysis of velocity profiles for different values of @xmath355 gives the following physical picture .
for @xmath356 , as for the unbounded case the flow can be thought of as being divided into two regions : an inner layer close to the bottom where it can be described by the equilibrium approximation and an outer layer behaving like an inviscid potential flow , where the profiles can be decomposed into the sum of decreasing and increasing exponentials @xmath357 . for smaller values of @xmath355
, this outer region progressively vanishes and the whole flow is controlled by the inner layer . ) .
the flow is from left to right .
note the squeezing of the lines _
downstream _ the crest of the bump .
amplitude @xmath358 ( b ) and phase @xmath359 ( c ) of the free surface as a function of @xmath355 for @xmath360 .
the peak in amplitude and the phase shift from @xmath130 to @xmath361 correspond to the resonance .
the two schematics illustrate the situations in phase or in antiphase . ] as a function of @xmath355 for @xmath362 ( dotted dashed line ) , @xmath363 ( dashed line ) , @xmath364 ( thin solid line ) , @xmath365 ( long dashed line ) , @xmath360 ( dotted line ) and @xmath366 ( solid line ) , and @xmath367 . crossing the resonance , the phase shifts from @xmath130 to @xmath361 .
( c ) and ( d ) , same for the friction force analytical model . ]
we display the phase and amplitude of the free surface as a function of @xmath355 in figure [ resonancefs](b - c ) .
the peak in amplitude accompanied by the phase shift of @xmath361 are the signature of a surface wave resonance .
the source of disturbances is of course the corrugation of the bed . for @xmath355 larger than its resonant value ,
the bottom and the free surface are in phase ; conversely , for @xmath355 below the resonance , they are in antiphase . in between , at the resonance
, the phase shift is @xmath368 ( figure [ resonancefs](a ) ) so that the streamlines are squeezed downstream to the crest .
this resonance is model - independent as it comes from a very robust physical mechanism .
as the fluid flows over the periodic bottom , gravity surface waves are excited at the wavelength @xmath6 .
the latter propagate at the velocity : @xmath369 with respect to the bottom ( see the friction force model derived in appendix [ ffc ] ) . as in the sound barrier phenomenon ,
the wave energy induced by the bottom disturbances accumulate when this velocity vanishes i.e. for : @xmath370 in the shallow water approximation ( @xmath371 ) , this resonant condition gives @xmath366 as standardly obtained in hydraulics . in the deep water approximation , it gives @xmath372 or equivalently @xmath373 ( @xcite ) .
so , for a bottom of wavelength @xmath6 , the flow is subcritical at low @xmath374 and low @xmath355 and supercritical at large @xmath374 and @xmath355 .
ignoring dissipation , the bernoulli relation states that the sum of the gravitational potential energy @xmath375 and the kinetic energy @xmath376 is constant along the free surface .
the subcritical regime corresponds to deep slow flows dominated by gravity : as the velocity increases over a bump , the corresponding increase of kinetic energy must be balanced by a loss of gravitational potential energy . as a consequence ,
the free surface is pinched over the bump ( @xmath377 , see figure [ resonancefs]c ) .
the supercritical regime corresponds to thin rapid flows dominated by kinetic energy . by conservation of the flow rate
, a pinch of the free surface would lead to an increased velocity . as the bump pushes up the free surface , the corresponding gain of potential energy
should be balanced by a decrease of kinetic energy which is achieved by a deformation of the free surface in phase with the bump ( @xmath378 , see figure [ resonancefs]c ) . in summary
, the free surface responds in phase with the excitation at small wavelength and becomes delayed as @xmath354 increases . as in a standard second order linear system , the disturbance and the system response are in quadrature at the resonance .
the phase @xmath379 and the rescaled amplitude @xmath358 of the free surface are displayed in figure [ deltasurfacelibre](a)-(b ) for different values of @xmath374 . in panels [ deltasurfacelibre](c)-(d ) , they are compared to the analytical predictions obtained using a much simpler closure ( appendix [ ffc ] ) .
one can see that the amplitude of the resonance increases with the froude number .
for very small @xmath374 , the phase curve is more complicated to interpret , but note that this corresponds to a vanishing amplitude @xmath345 : the resonance essentially disappears . for @xmath380 ,
the free surface amplitude seems to converge to some finite value , but the phase slowly goes back to @xmath130 .
this gentle crossover is indeed expected at very large wavelength , a situation for which the free surface must follow the bottom topography .
it is interesting to investigate how the resonance is affected by the fact that the bottom moves i.e. can grow or propagate . following the notations introduced in [ movingboundary ] , we display in figure [ deltasigmaomega ] the amplitude of the free surface @xmath358 as a function of @xmath355 for different values of the growth rate @xmath117 and the pulsation @xmath204 . for positive growth rates
@xmath117 , the q - factor of the resonance gets smaller but the resonant wavenumber is not affected .
a bottom propagating at the velocity @xmath223 moves the resonant peak along the @xmath355-axis . for a positive propagation velocity @xmath223 the surface velocity with respect to the bottom and thus the effective froude number
get reduced . as a consequence
, the resonant wavelength gets smaller and @xmath355 larger .
conversely , the peak moves to smaller wavenumber for an upstream moving bottom .
as in sub - section [ movingboundary ] , these effects are noticeable for values of @xmath228 and @xmath229 of order one , while realistic values are respectively on the order of @xmath236 and @xmath235 . for the ripples and dunes problem ( part 2 ) , the bedform motion can thus be safely ignored in the hydrodynamical calculation and the effect of free surface interpreted in terms of surface standing waves . as a function of @xmath355 for different values of the growth rate @xmath117 ( a ) and the pulsation @xmath204 ( b ) of the bottom boundary .
arrows indicate increasing values of @xmath117 and @xmath204 .
these graphs have been computed with @xmath349 and @xmath360 . in panel
( a ) , the grey dashed line is for @xmath381 , the back dashed line for @xmath382 , the black dotted line for @xmath195 and the two grey dotted lines for @xmath170 and @xmath129 . in panel ( b ) , the two grey dashed lines is for @xmath383 and @xmath211 , the black dashed line for @xmath382 , the black dotted line for @xmath195 and the three grey dotted lines for @xmath170 , @xmath129 and @xmath175 . for comparison , in all panels the solid lines correspond to the fixed case @xmath218 , @xmath219 . ]
the basal shear stress and pressure and subsequently the coefficients @xmath123 , @xmath109 , @xmath110 and @xmath111 are modified by the presence of the free surface when @xmath355 is of order one and below . in figure
[ abcdfreesurface ] , the coefficients are plotted as functions of @xmath355 for different values of the froude number .
one can see that the resonance peak is more pronounced for larger @xmath374 they are actually not visible when @xmath374 is too small . in agreement with the streamlines of figure [ resonancefs](a ) , which shows a squeezing downstream the bump crest , the peak of @xmath109 is negative , corresponding to a phase delay of the stress with respect to the bottom .
furthermore , the curves corresponding to the presence of a rigid lid at the same height @xmath40 do not exhibit these peaks .
finally , the diverging behaviour of @xmath109 as @xmath380 is also a free surface effect as , in the same limit , @xmath109 reaches a plateau in the case of a rigid top boundary .
the behaviours of @xmath123 , @xmath109 , @xmath110 and @xmath111 at small @xmath355 ( below the resonant condition ) can be determined analytically using the simple closure proposed in the appendix [ ffc ] . for small @xmath355 , we get @xmath384 , @xmath385 , @xmath386 and @xmath387 .
these scalings fit fairly well the solutions of the full equations . as a conclusion ,
there are two situations in which the excitation of standing waves by the topography affects significantly the characteristics of the inner layer : ( i ) around the resonance , since the surface wave amplitude is very large and ( ii ) for vanishing @xmath355 , when the distance @xmath40 between the topography and the free surface becomes so small that the inner layer invades the whole flow . , @xmath109 , @xmath110 and @xmath111 as functions of @xmath355 , for
@xmath388 ( dotted dashed line ) , @xmath360 ( dotted line ) and @xmath366 ( solid line ) .
the dashed lines correspond to a rigid boundary at the same height @xmath40 .
the plots have been computed for @xmath367 . comparing a free surface to a rigid boundary condition ( or to the case @xmath389 )
, it can be inferred that the hydrodynamics is controlled by the surface waves .
in particular , the resonance leads to a drop of the shear stress component @xmath109 i.e. to a downstream shift of the point of maximum shear stress . ) ]
as this article is based on a rather technical ground , it is useful to sum up , in a qualitative manner , our key results and to put them in perspective with respect to the second part of the paper . in the context of the formation of ripples and dunes from a flat sand bed submitted to a turbulent flow , a central issue is the description of the basal shear stress , which controls bed load transport . due to the scale separation between the typical evolution time of the bedforms and that of the flow , the bottom can be considered as quasi - static .
the hydrodynamics can be investigated independently of the transport issue . in the traces of the seminal work of @xcite
, three main regions can be evidenced in the turbulent flow over a wavy bottom . *
an outer layer , away from the bottom , in which the flow is well described by inviscid potential equations , i.e. where the perfect flow approximation is valid .
the streamlines follow the topography so that the velocity is in phase with the bottom . * an inner layer , which corresponds to the region where the inertial terms of the navier - stokes equation are negligible , and thus where the longitudinal pressure gradient is balanced by the transverse mixing of momentum due to turbulent fluctuations i.e. by the reynolds shear stress transverse gradient .
the thickness @xmath97 of the inner layer is related to wavelength by @xmath390 . at the transition between the inner and outer layers ,
the fluid velocity is slowed down by the shear stress . due to inertia , the velocity is always phase delayed with respect to the shear stress .
as the velocity is inherited from the outer layer , the shear stress is phase - advanced with respect to the topography . * a thin surface layer of thickness @xmath22 , which is responsible for the hydrodynamical roughness @xmath20 seen from the inner layer .
the dominant physical mechanism at work in this surface layer can be of different nature .
for instance , @xmath20 can result from the mixing due to roughness elements , the predominance of viscous dissipation , or the presence of bed - load transport .
specific scaling laws for @xmath20 and @xmath22 are obtained for each of theses cases .
the linear relationship between the stresses and the bottom profile can be encoded into two complex coefficients @xmath231 and @xmath391 . the tangent of the phase between the stresses and the bottom is given by the ratios @xmath122 and @xmath392 respectively .
these coefficients are key inputs for the sediment transport issue ( see part 2 ) , and are needed to compute the dispersion relation of the bedforms , i.e. the growth rate as a function of the bedform wavenumber , which tells whether a given wavelength is stable or unstable .
their dependencies with respect to several important parameters such as the length scale ratios @xmath393 or @xmath394 have been determined .
their sensitivity to different ways of modelling turbulence or of imposing the bottom boundary conditions have also been discussed .
our conclusions regarding these functions can be summed up as follows : * @xmath123 and @xmath109 are generically positive , whereas @xmath110 is positive and @xmath111 negative .
this means that the shear stress profile reaches its maximum before the crests of the bumps .
this effect can be visualised on the streamlines , which are squeezed at this pint of maximum shear . for the normal stresses ,
the situation is opposite : the pressure maximum is slightly delayed after the crests . as the pressure is almost constant across the inner boundary layer , this phase shift @xmath392 is less pronounced by an order of magnitude .
@xmath123 , @xmath109 , @xmath110 and @xmath111 have weak dependencies on the ratio @xmath395 .
the shear stress phase shift @xmath122 vanishes for asymptotically small @xmath396 and gently increases with @xmath397 .
as the inner layer thickness @xmath97 becomes of the order of the surface layer thickness @xmath22 , the phase shift drops .
* these features are robust to ( i ) the turbulent closure , ( ii ) the existence of a reynolds stress anisotropy , ( iii ) the motion ( growth or propagation ) of the bottom . * much more important is the role of the free surface in the case of a water depth @xmath40 on the order of the bottom wavelength @xmath6 .
the undulations of the bottom excite standing gravity waves at the free surface .
resonant conditions are reached when these waves precisely propagate at a velocity equal to that of the flow , i.e. when @xmath398 . at the resonance ,
the response of the free surface is in quadrature with the disturbance . when the froude number is large enough , the deformation @xmath358 of the free surface at the resonance is so large that it has a strong effect on the flow close to the bottom . in particular , the streamlines are squeezed downstream the crests of the bump , so that the shear stress becomes phase - delayed with respect to the topography ( @xmath399 ) . *
another effect due to the presence of a free surface is found for @xmath380 . in this limit
, the water depth becomes very thin in comparison to the wavelength , and the inner layer invades the whole flow ( i.e. @xmath400 ) .
in this situation , the shear stress and the bottom profiles tend to be in phase so that @xmath401
. moreover , the shear stress becomes phase delayed ( @xmath399 ) below a threshold value of @xmath355 that increases with the froude number .
* the shear stress profile is insensitive to the mechanisms at work in the surface layer provided that its thickness @xmath22 is smaller than the inner layer thickness @xmath97 : the hydrodynamical roughness @xmath20 is the single quantity inherited from the surface layer .
the asymptotic calculation performed by @xcite is recovered but only for asymptotically large @xmath402 , a limit hardly reached in real problems .
when @xmath22 is comparable or larger than @xmath97 , the shear stress coefficients @xmath123 and @xmath109 are smaller in the hydraulically smooth regime where the roughness is due to viscosity than in the hydraulically rough regime . in the later case ,
the mixing of momentum in the surface layer is dominated by the turbulent fluctuations induced by the roughness elements . in the ripples and dunes problem
, these results directly apply to the case where sediments are transported with negligible feedback on the flow ( erosion limited transport , see part 2 ) .
when the extra - stress due to bedload transport is significant ( momentum limited transport ) , we predict that @xmath123 and @xmath109 become larger .
but with an almost identical phase shift @xmath122 .
* the normal stress profile is almost independent of the surface layer model . beyond the linear case ,
we have expended the hydrodynamical calculation to the third order in @xmath61 .
in contrast to the linear calculation , the separation of streamlines , when the shear stress and the pressure gradient are antagonist , and the associated formation of a recirculation bubble are obtained for realistic values of the aspect ratio . at the quadratic order
, we get a correction to the mean velocity profile which corresponds to a roughness @xmath282 at large scale due to bottom corrugation .
this roughness of geometrical origin depends on the ` microscopic ' roughness @xmath20 inherited from the surface layer and is not simply proportional to the amplitude of the bottom profile .
we get the scaling law : @xmath403 , where the factor @xmath264 weakly depends on @xmath395 .
looking at the shear stress modulation , we have shown that the phase shift between the shear stress and the bottom profile is reduced by non - linear corrections and change sign at some particular aspect ratio . in the second part of this paper
, this non - linear analysis is used to predict the selection of the aspect ratio of mature ripples and dunes . ' '' '' the understanding of the qualitative reason for the upstream shift of the maximum shear stress on a bump has been meditated with a.b . murray .
this work has benefited from the financial support of the french minister of research .
we thank louv for hospitality , where part of this paper was written .
the calculations of section [ unboundedcase ] can be generalised in the case of the following anisotropic stress - strain relationship : @xmath404 at the linear order , the velocity , pressure and stress fields read : @xmath50 , \label{defu_avi}\\ u_z & = & u _ * k\zeta e^{ikx } w , \label{defw_avi}\\ \tau_{xz } & = & \tau_{zx}= - u_*^2 \left[1+k\zeta e^{ikx } s_t\right ] , \label{defst_avi}\\ p+\tau_{zz } & = & p_0+u_*^2 \left [ \frac{1}{3}\chi_z^2 + k\zeta e^{ikx } s_n\right ] , \label{defsn_avi}\\ \tau_{zz } & = & u_*^2 \left [ \frac{1}{3}\chi_z^2 + k\zeta e^{ikx } s_{zz}\right ] , \label{defszz_avi}\\ \tau_{xx } & = & u_*^2 \left [ \frac{1}{3}\chi_x^2 + k\zeta e^{ikx } s_{xx}\right ] , \label{defsxx_avi}\end{aligned}\ ] ] and the stress equations can be simplified into @xmath405 the normal stress difference is this time : @xmath406 so that on gets the following four closed equations : @xmath407 as before , they can be written in the usual compact matrix form ( [ systlinplaque ] ) , now with @xmath408
the dynamical equations governing the second - order moments @xmath192 can be derived rigorously . under the assumption of turbulence isotropy at the dissipative scale
, it can be written under the form : @xmath409 @xmath410 is the dissipation rate ; @xmath411 is the spatial flux of turbulent energy induced by fluctuations ; the pressure term @xmath412 conserves energy and is usually responsible for the isotropisation of fluctuations .
we wish to get a stress tensor that relaxes towards its steady state expression prescribed by equation ( [ tauijprandtl ] ) . for dimensional reasons ,
we write the relaxation rate under the form @xmath413 , where @xmath190 is a phenomenological constant , and keep the mixing length @xmath13 fixed by the geometrical distance to the wall .
the second moment equation then takes the form of a first order relaxation equation : @xmath193 .
\label{2ndme_2ndorder}\ ] ] setting @xmath187 , one recovers the stationary solutions ( [ tauijprandtl ] ) .
a finite value of @xmath190 introduces a lag between a change of the flow velocity field and the point / time at which the reynolds stress readapts to this change . for 2d steady situations ,
the stress relaxation equations are the following : @xmath414 , \label{relaxxz_2ndorder}\\ u_x \partial_x \tau_{xx}+u_z \partial_z \tau_{xx } & = & \frac{|\dot \gamma|}{\beta } \left[-\kappa^2 l^2 |\dot \gamma| \dot \gamma_{xx}+ \frac{1}{3 } \kappa^2 \chi^2 l^2 |\dot \gamma|^2-\tau_{xx } \right ] , \label{relaxxx_2ndorder}\\ u_x \partial_x \tau_{zz}+u_z \partial_z \tau_{zz } & = & \frac{|\dot \gamma|}{\beta } \left[-\kappa^2 l^2 |\dot \gamma| \dot \gamma_{zz}+ \frac{1}{3 } \kappa^2 \chi^2 l^2 |\dot \gamma|^2-\tau_{zz } \right ] .
\label{relaxzz_2ndorder}\end{aligned}\ ] ] at linear order , they simplify into : @xmath415 taking the difference of equations ( [ relaxsxx_2ndorder ] ) and ( [ relaxszz_2ndorder ] ) , one can compute @xmath416 to obtain four closed equations : @xmath417 as before , they can be written in the usual compact matrix form ( [ systlinplaque ] ) , now with @xmath418
recall that @xmath47 is the bottom profile whose wavelength is @xmath42 , and @xmath324 the dimensionless vertical coordinate . at the linear order ,
we write all the relevant quantities under the form : @xmath419 an alternative is to use the curvilinear coordinates @xmath420 and write the field @xmath421 as : @xmath422 we call these expressions respectively the ` non - shifted ' and ` shifted ' representations of @xmath421 .
they lead to the same linearised equations as they are related to each other at the linear order by @xmath423 in practice , this is especially relevant for @xmath51 , for which @xmath424 is not constant : @xmath425 is the shifted representation of the first order correction to the horizontal velocity .
however , @xmath426 , @xmath427 and @xmath428 .
importantly , note that the range in @xmath53 for which these two representations are valid is not the same _ a priori_. all fields are expanded up to the third order in @xmath61 , neglecting also non - harmonic terms in @xmath429 .
non - shifted representation of the streamwise velocity : @xmath430 shifted representation of the same quantity : @xmath431 expanding the functions @xmath432 with respect to @xmath61 , one can relate a representation to the other as : @xmath433 conversely : @xmath434 the passage from a representation to the other for the other fields works the same , except that there is no zeroth order ( function @xmath134 ) in the expressions .
to compute the streamlines , we introduce the so - called stream function @xmath435 , defined by @xmath436 and @xmath437 .
this function is such that @xmath438 , so that the iso - contours @xmath439 precisely show the streamlines . using the continuity equation ( [ nscont ] )
, it is easy to show that a solution is @xmath440 .
this integral is computed between @xmath441 ( the bottom ) and @xmath442 .
we note @xmath54 the rescaled distance to the bottom . restricting to the linear order , with the relation @xmath443 ( equation ( [ equaw ] ) ) , we end up with @xmath444 \right \},\ ] ] where the function @xmath134 is given by expression ( [ mu2etageom ] ) .
in the situation with a free surface , one can use the following representation for the field @xmath421 : @xmath445 this curvilinear variable @xmath55 vanishes on the bottom @xmath446 , and @xmath447 at the surface @xmath448 . the new function @xmath449 is related to those of the non - shifted representation @xmath450 and @xmath451 as : @xmath452 for @xmath453
, we have @xmath424 and @xmath454 .
consequently , the new stream function is @xmath455 \right \}.\ ] ] one can check that the free surface is indeed a streamline itself , as one of the top boundary conditions is @xmath456 . in the non - linear ( and unbounded )
case , we get : @xmath457
definition of the different functions involved in the expansion : @xmath237 , \\
u_z & = & u _ * \left[(k\zeta ) e^{ikx } w_1+(k\zeta)^2 e^{2ikx } w_{2}+(k\zeta)^3 e^{ikx } w_{3}\right ] , \\ \tau_{xz } & = & - u_*^2 \left[1+(k\zeta ) e^{ikx } s_{t1}+(k\zeta)^2 s_{t0}+(k\zeta)^2 e^{2ikx } s_{t2}+(k\zeta)^3 e^{ikx } s_{t3}\right ] , \\
p+\tau_{zz } & = & p_0+u_*^2 \left[(k\zeta ) e^{ikx } s_{n1}+(k\zeta)^2 s_{n0}+(k\zeta)^2 e^{2ikx } s_{n2}+(k\zeta)^3 e^{ikx } s_{n3}\right],\\ \tau_{zz}- \tau_{xx } & = & u_*^2 \left[(k\zeta ) e^{ikx } s_{d1}+(k\zeta)^2 s_{d0}+(k\zeta)^2 e^{2ikx } s_{d2}+(k\zeta)^3 e^{ikx } s_{d3}\right].\end{aligned}\ ] ] expansion of the mixing length : @xmath458 expansion of the strain tensor components : @xmath459 which gives for the strain modulus : @xmath460,\end{aligned}\ ] ] and then @xmath461\nonumber\\ & + & ( k\zeta)^2 \left[\frac{\kappa^2\mu'}{2}-\frac{\kappa}{2\mu'}(u_1'+u_1'^*+iw_1-iw_1^*)+\frac{1}{\mu'^2}\left(u_0'+\frac{u_1u_1^*}{\mu'}\right)\right]\nonumber\\ & + & ( k\zeta)^2 e^{2ikx}\left[\frac{\kappa^2\mu'}{2}-\frac{\kappa}{\mu'}(u_1'+iw_1)+\frac{1}{\mu'^2}\left(u_2'+2iw_{2}-\frac{u_1 ^ 2}{\mu'}\right)\right]\nonumber\\ & + & ( k\zeta)^3 e^{ikx}\left[\frac{\kappa^2}{4}\left(2u_1'+u_1'^*+2iw_1-iw_1^*\right)-\frac{\kappa}{\mu ' } \left(2u_0'+\frac{u_1(2u_1^*-u_1)}{\mu'}+u_2'+2iw_{2}\right)\right.\nonumber\\ & + & \left.\frac{1}{\mu'^2}\left(u_3'+iw_{3}-\frac{u_1 u_1^*}{\mu'^2}(u_1'+iw_1)+\frac{u_1 ^ 2}{2\mu'^2}(u_1'^*-iw_1^*)+\frac{4u_1^*u_2}{\mu'}\right)\right].\end{aligned}\ ] ] expressions of the different functions corresponding to the normal stresses @xmath462 : @xmath463 expressions of the different @xmath251 : @xmath464 $ \\ \quad & $ - \,\,\ , \frac{\mu'}{4}(u_1'^*-iw_1^*)\left[\kappa^2\mu'+\frac{2}{\mu'^2}(u_2'+2iw_{2})\right]$ \end{tabular } \\ $ 0 $ \\ \begin{tabular}{ll } $ i u_0 u_1 $ & $ + \,\,\ , \frac{i}{2 } u_2 u_1^ * + u_0 ' w_1 + \frac{1}{2 } w_1^ * u_2 ' + \frac{1}{2 } w_2 u_1'^*$ \\ \quad & $ + \,\,\ , \frac{4}{\mu'^2}u_2(u_1'^*-iw_1^*)-8\kappa u_2+\kappa^2\mu ' ( 2u_1-u_1^*)+\frac{6}{\mu'^3}u_1 ^ 2u_1^*$ \\ \quad & $ + \,\,\ , \frac{2\kappa}{\mu'}(u_1'+iw_1)(u_1^*-u_1)-\frac{2\kappa}{\mu'}u_1(u_1'^*-iw_1^*)+\frac{4}{\mu'^2}u_1u_0'$ \\ \quad & $ - \,\,\ , \frac{2}{\mu'^2}u_1^*(u_2'+2iw_{2})$ \end{tabular } \\ $ -i u_0 w_1 - \frac{3i}{2 } w_2 u_1^ * + \frac{3i}{2 } u_2 w_1^*$ \end{tabular } \right ) .\end{aligned}\ ] ]
several of the free surface effects can be recovered within a simple friction force model , for which analytical expressions of the linear solution of the flow can be derived .
in particular , the resonance condition as well as the behaviour of the basal stress coefficients @xmath123 , @xmath109 , @xmath110 and @xmath111 for @xmath380 can be found and interpreted .
we start from the navier - stokes equations for a perfect flow , with a crude additional turbulent friction term as an approximation of the stress derivatives : @xmath465 physically , the force applied to a fluid particle is directly related to the relative velocity with respect to the ground . at an angle @xmath308 ,
the following plug flow is an homogeneous solution of the above equations : @xmath466 in order to estimate the value of the friction coefficient , one can make use of the fact that typical turbulent velocity vertical profiles are logarithmic .
however , as the logarithm varies slowly when @xmath1 is much larger than @xmath20 , we write @xmath467 . identifying the shear stress on the bottom as @xmath468
, we finally get with the relation ( [ ux0_ffc ] ) @xmath469 for @xmath353 in the range @xmath470-@xmath471 , we get a typical value for @xmath472 on the order of few @xmath235 .
we now normalize quantities by @xmath473 and @xmath40 and get a single non - dimensional ( froude ) number : @xmath474 the starting equations can be linearised around the above reference state . looking a the flow over a corrugated bottom @xmath47 , it is easy to show that the solution is of the following form @xmath475 , \\ u_z & = & \overline{u}ik\zeta e^{ikx } \left [ a_+ e^{kz } + a_- e^{-kz } \right ] , \\ p & = & g \cos \theta ( h - z ) + \overline{u}^2 ( kh - i 2\omega ) \frac{\zeta e^{ikx}}{h } \left [ a_+ e^{kz } - a_- e^{-kz } \right ] , \end{aligned}\ ] ] where @xmath476 and @xmath477 must be determined by the boundary conditions .
this exponential form is characteristic of potential flows .
we require that the velocity normal to the bottom vanish . following the notations of the main part of the paper , we define @xmath478 such that the free surface is at the altitude @xmath479 .
it is a material line where the pressure vanishes .
the three boundary conditions are then : @xmath480 where , as before , @xmath345 is defined as @xmath328 .
the constants @xmath476 and @xmath477 , as well as @xmath345 are thus solutions of @xmath481 from which we get : @xmath482 , \\
a_- & = & \frac{1}{2 } \left [ 1 + \frac{(kh - i2\omega)\tanh kh - \frac{1}{{{\mathcal{f}}}^2 } } { ( kh - i2\omega ) - \frac{1}{{{\mathcal{f}}}^2 } \tanh kh } \right ] .\end{aligned}\ ] ] the shear stress is not part of the variables of this model , but we can consistently define it as @xmath483 .
looking at the shear stress @xmath484 and normal stress @xmath485 on the bottom , in accordance with the notations of the previous sections of the paper , we introduce the coefficients @xmath123 , @xmath109 , @xmath110 and @xmath111 as @xmath486 , \\
p_b & = & gh\cos\theta + \omega \overline{u}^2 ( c+id ) k\zeta e^{ikx},\end{aligned}\ ] ] which gives @xmath487 \tanh kh - \frac{1}{{{\mathcal{f}}}^2 } \ , kh [ \tanh^2 kh+1 ] } { \left ( kh-\frac{1}{{{\mathcal{f}}}^2}\tanh kh \right ) ^2 + 4\omega^2 } \ , , \\
b & = & \frac{2\omega}{{{\mathcal{f}}}^2 } \ , \frac{[\tanh^2 kh-1 ] } { \left ( kh-\frac{1}{{{\mathcal{f}}}^2}\tanh kh \right ) ^2 + 4\omega^2 } \ , , \\ c & = & \frac{1}{2\omega } \left ( -a-\frac{2\omega b}{kh }
\right ) , \\ d & = & \frac{1}{2\omega } \left ( -b+\frac{2\omega a}{kh } \right ) .\end{aligned}\ ] ] it is worth noting that the friction force model predicts negative values of @xmath109 for any @xmath355 .
this means that there is always a phase delay of the shear stress with respect to the bottom , which is a clear disagreement with the full solution . in order to fix this flaw
, one would need to empirically introduce an imaginary part to @xmath472 .
finally , this discrepancy shows that a precise description of the phase between the basal friction and the relief is a subtle and difficult issue that fully justify the use of a rigorous but heavy formalism .
ayotte , k.w .
, xu , d. & taylor , p.a .
1994 the impact of turbulence closure schemes on predictions of the mixed spectral finite - difference model for flow over topography .
_ boundary - layer met . _ * 68 * , 1 - 33 .
prandtl , l. 1925 bericht ber untersuchungen zur ausgebildeten turbulenz .
_ * 3 * , 136 - 139 . after ,
bradshaw , p. 1974 possible origin of prandt s mixing - length theory , _ nature _ * 249 * , 135 - 136 . | in the context of subaqueous ripple and dune formation , we present here a reynolds averaged calculation of the turbulent flow over a topography . using a fourier decomposition of the bottom elevation profile , we perform a weakly non - linear expansion of the velocity field , sufficiently accurate to recover the separation of streamlines and the formation of a recirculation bubble above the some aspect ratio .
the normal and tangential basal stresses are investigated in details ; in particular , we show that the phase shift of the shear stress with respect to the topography , responsible for the formation of bedforms , appears in an inner boundary layer where shear stress and pressure gradients balance .
we study the sensitivity of the calculation with respect to ( i ) the choice of the turbulence closure , ( ii ) the motion of the bottom ( growth or propagation ) , ( iii ) the physics at work in the surface layer , responsible for the hydrodynamic roughness of the bottom , ( iv ) the aspect ratio of the bedform and ( v ) the effect of the free surface , which can be interpreted in terms of standing gravity waves excited by topography .
the most important effects are those of points ( iii ) to ( v ) , in relation to the intermixing of the different length scales of the problem .
we show that the dynamical mechanisms controlling the hydrodynamical roughness ( mixing due to roughness elements , viscosity , sediment transport , etc ) have an influence on the basal shear stress when the thickness of the surface layer is comparable to that of the inner layer .
we evidence that non - linear effects tend to oppose linear ones and are of the same order for bedform aspect ratios of the order of @xmath0 .
we show that the influence of the free surface on the basal shear stress is dominant in two ranges of wavelength : when the wavelength is large compared to the flow depth , so that the inner layer extends throughout the flow and in the resonant conditions , when the downstream material velocity balances the upstream wave propagation .
leer.eps gsave 72 31 moveto 72 342 lineto 601 342 lineto 601 31 lineto 72 31 lineto showpage grestore |
it is well known that the momentum distribution @xmath3 of a system of interacting particles exhibits , at equilibrium , the scaling behavior @xmath4 at large momenta @xmath5 , where the coefficient @xmath6 defines the total tan contact @xcite , a universal quantity connecting the thermodynamic properties of the many - body system to the short range behavior of its wave function @xcite . in the case of a weakly interacting bose gas
this tail is physically associated with the presence of the quantum depletion of the condensate , a challenging feature well understood theoretically , but difficult to measure experimentally @xcite .
the @xmath0 tail of the momentum distribution has been measured in the unitary fermi gas , profiting of the possibility of turning off the interaction , just after the release of the trap @xcite . in this case
the momentum distribution is conserved during the expansion , so that experiments measure the value of the contact before expansion . despite the systematic investigations of the contact at equilibrium @xcite , its evolution in a dynamical context has been rarely explored @xcite .
the importance of this question is manifested by the fact that the momentum distribution is affected by interactions during the expansion and , for large expansion times , when it becomes time - independent , can be directly related to the measurable time - dependent density distribution @xmath7 through the ballistic relationship @xmath8 according to eq.([eq : ballistic ] ) , an asymptotic @xmath0 tail in the momentum distribution should show up in a tail of the density distribution , characterized by the @xmath9 dependence for large values of @xmath10 and @xmath11 .
a recent experiment @xcite , where the density distribution of an expanding gas of metastable @xmath1he atoms was measured with high accuracy , has indicated that the contact parameter @xmath6 is actually increased with respect to the initial value , and thus motivated us for the theoretical investigation of the dynamical evolution of the contact .
the mechanism of the expansion of a dilute atomic gas is well understood in the regime of distances satisfying the condition @xmath12 , where @xmath13 is the thomas - fermi radius at the expansion time @xmath11 , or momenta smaller than the inverse of the healing length @xcite . in this regime
approximate schemes based on scaling transformations @xcite or on the hydrodynamic picture @xcite provide an accurate description of the expansion and , in particular , have proven very successful in predicting the time dependence of the aspect ratio , a quantity measurable with high precision . the description of the expansion at a more microscopic level , involving the large-@xmath5 components of the momentum distribution , represents a more challenging problem , whose solution requires the implementation of time - dependent dynamic theories beyond the mean field level .
the purpose of the present work is to discuss some examples where such theories can be worked out explicitly and the question of the behavior of the tails of the momentum distribution can be addressed in a quantitative way at zero temperature .
these include the unitary fermi gas , the weakly interacting bose gas and the two - body problem with highly asymmetric mass imbalance which reduces the problem to the solution of the single - particle schrdinger equation .
the paper is structured as follows . in sec .
[ sec : manybody ] we investigate the evolution of the contact for the unitary fermi gas and the weakly repulsive bose gas . in sec .
[ sec : twobody ] we explicitly calculate the density and momentum distributions during the expansion for the two - particle system , both at unitarity and in the case of weakly repulsive interaction . in both cases the contact is found to decrease for large expansion times in the same way as in the corresponding many - body systems .
we also calculate the change of the contact after a sudden quench of the scattering length from a small value to infinity just before the expansion .
our concluding remarks are reported in sec .
[ sec : conclusion ] .
a remarkable feature exhibited by the fermi gas at unitarity is that an exact solution of the problem of the expansion from a three - dimensional ( 3d ) isotropic harmonic trap can be formulated in terms of the exact scaling transformation @xcite @xmath14 of the many - body wave - function . here
@xmath15 is the number of particles , @xmath16 is the normalization constant and @xmath17 with @xmath18 the dimensionless time and @xmath19 the trapping frequency .
consequently , the density distribution of the system during the expansion , is given by @xmath20 where @xmath21 is the equilibrium density evaluated at @xmath22 .
the ballistic relation ( eq .
( [ eq : ballistic ] ) ) then gives the following expression for the asymptotic momentum distribution @xmath23 which holds for large times satisfying the condition @xmath24 .
the scaling result shows that the occurrence of the @xmath0 tail in @xmath25 should be associated with a corresponding @xmath2 tail at large @xmath10 in the equilibrium density distribution before expansion .
this behavior is ruled out by the simple argument that no natural cut - off exists in coordinate space at large @xmath10 and that the @xmath2 law would imply an unphysical infrared divergent behavior in the ground state harmonic oscillator energy . an explicit expression for the time dependence of the contact
can be obtained by making the adiabatic ansatz @xcite .
this ansatz corresponds to assuming that , during the expansion , the large momentum component of the momentum distribution is given by the same expression @xmath26 holding at equilibrium , with @xmath27 evaluated using the time dependent value of the density profile .
the adiabatic ansatz holds if the two - body interaction , which dictates the short range behavior of the wave function , is kept constant or slowly varying in time . in the unitary fermi gas
the contact is given by @xmath28^{4/3}\}$ ] where @xmath29 is a universal parameter .
using the position and time dependence of the density predicted by the scaling law ( eq . ( [ castindensity ] ) ) , the contact
is then expected to evolve in time as @xmath30 where @xmath31 is the initial contact of the unitary fermi gas at equilibrium , @xmath32 is the bertsch parameter , and @xmath33^{1/3}$ ] is the initial local fermi momentum at the trap center .
it is worth noticing that the same analytic result ( eq . ( [ eq : asym_unitary ] ) ) for the asymptotic momentum distribution and for the time dependence of the contact can be also derived by studying the behavior of the one - body density matrix ( see appendix [ appdenx1 ] ) .
although in the case of 3d bose gas an exact scaling solution of the many - body wave function similar to eq .
( [ castin ] ) is not available , the adiabatic ansatz discussed above is still expected to hold , yielding the time dependence of the contact in the form @xmath34 $ ] where @xmath35 is the @xmath36-wave scattering length and the integrand is the local contact density at equilibrium predicted by bogoliubov theory .
we can then use the time dependence of the density profile during the expansion as predicted by the use of scaling transformations or hydrodynamic theory to predict the decay of the contact as @xmath37 where @xmath38 is the initial contact of the dilute bose gas with @xmath39 the initial thomas - fermi radius .
more complete and explicit results for the time evolution of the contact are available by solving the expansion dynamics of a two - body problem , which gives access to the exact time - dependence of the density and momentum distributions and provides explicit conditions for the applicability of the adibaticity ansatz . for the sake of brevity , we consider two interacting particles in an isotropic 3d harmonic trap with one of them being so heavy that the center - of - mass motion can be ignored .
this reduces the problem to the investigation of a one - body system described by the hamiltonian @xmath40 where @xmath41 is the mass of the light particle .
@xmath42 is the two - body interaction characterized by the value of the @xmath36-wave scattering length , a quantity which , in the presence of a fano - feshbach resonance , can be tuned by applying an external magnetic field @xcite . in this work , we will consider the regularized pseudopotential @xmath43 , whose eigenvalue problem can be solved analytically @xcite .
the eigenfunctions take the form @xmath44 where @xmath45 is the hypergeometric function , @xmath46 is the normalization constant , @xmath47 is the harmonic oscillator length and @xmath48 , with the eigenvalue @xmath49 determined by @xmath50 two different values of scattering length will be discussed in the following : at unitarity ( @xmath51 ) , the ground state energy is @xmath52 and the corresponding wave function takes the form @xmath53 .
the wave function decays exponentially at large @xmath10 whereas it approaches @xmath54 for small @xmath10 .
we notice that the size of the pair is reduced with respect to the non - interacting value according to @xmath55 , reflecting the attractive nature of the force at unitarity .
the ground state wave function in momentum space is readily obtained after a fourier transformation : @xmath56 where @xmath57 is the dawson function . using its asymptotic behavior @xcite
, one finds the result @xmath58 ^ 2 = \left\{\begin{array}{lll } \pi^{-5/2}a_{\text{ho}}^{3 } , & k\rightarrow 0 & \\
\frac{\pi^{-5/2}a_{\text{ho}}^{-1}}{k^4 } , & k\rightarrow \infty & \end{array}% \right . \\\ ] ] for the momentum distribution , yielding the value @xmath59 for the contact .
when the particle is released from the isotropic trap , the wave function evolves according to the exact scaling transformation , eq .
( [ castin ] ) ( with @xmath60 ) , thus the expanding density distribution is given by @xmath61\frac{1}{r^2a_{\text{ho } } } , \label{eq : unitaryrhot}\ ] ] where @xmath62 is the dimensionless time .
the density distribution exponentially decays at large @xmath10 and the absence of an additional @xmath2 tail indicates that the @xmath0 tail of the momentum distribution will disappear at large expansion time .
the time evolution of the momentum distribution can be explicitly obtained by calculating the wave function in momentum space .
one finds @xmath63 therefore , as a consequence of the asymptotic behavior of the dawson function @xcite , we find that , for any fixed value of @xmath11 , @xmath64 for @xmath65 with the contact @xmath66 vanishing like @xmath67 for large expansion times @xcite .
this is the same law obtained in the case of the unitary fermi gas discussed in the first part of the paper , using the adiabaticity ansatz for the evolution of the contact . and ( b ) momentum distribution @xmath68 in the presence of a regularized pseudopotential with unitary interaction applied to the two - body problem .
different colors are the distributions at different expansion time @xmath69 ( red ) , @xmath70 ( green ) , @xmath71 ( blue ) and the asymptotic momentum distribution(black ) .
the linear part of the density distribution @xmath72 at small @xmath10 exhibits the @xmath73 behavior and the linear part of the momentum distribution @xmath68 at large @xmath5 exhibits the @xmath0 behavior for finite expansion times.,width=325 ] the asymptotic momentum distribution , easily derivable by inserting eq.([eq : unitaryrhot ] ) into the ballistic formula ( eq.([eq : ballistic ] ) ) , takes the form @xmath74 the same expression for the asymptotic momentum distribution can be also obtained using the asymptotic @xmath75 of the dawson function , holding for @xmath76 and @xmath77 .
this result can be obtained using the relation between dawson function and the error function @xmath78 and equation ( 7.2.4 ) from ref .
@xcite .
figure [ fig : unitarity ] shows the density distribution @xmath79 and the momentum distribution @xmath80 at @xmath81 .
the figure shows that the value of @xmath5 , above which one can observe the @xmath0 behavior , becomes larger and larger as the increase of @xmath82 .
we find that for @xmath83 the @xmath0 tail is practically no longer visible and that the momentum distribution ( see fig .
[ fig : unitarity](b ) ) , is indistinguishable from the asymptotic behavior , except for small values of @xmath5 where reaching the asymptotic behavior requires even larger expansion times .
it is also worth to note that , for shorter times , where the @xmath0 tail is still visible in the momentum distribution , there is no trace of the @xmath2 tail in the density distribution ( see fig .
[ fig : unitarity](a ) ) because the ballistic relationship ( [ eq : ballistic ] ) is not applicable on these times . in this case ,
the energy and the normalization constant must be determined numerically .
choosing @xmath84 , one finds @xmath85 and @xmath86 @xcite . at small @xmath10 ,
the exact wave function can be approximated by @xmath87 , whereas the large @xmath10 behavior is dominated by the exponential factor .
the momentum distribution at large @xmath5 is determined by the small @xmath10 behavior of the wave function .
one finds the dependence @xmath88 for large @xmath5 , yielding the value @xmath89 for the contact which is proportional to @xmath90 like the case of the weakly interacting bose gas .
the problem of the expansion can be solved in a way similar to the one of ref .
@xcite , where the short time evolution of the contact was studied after quenching the scattering length to infinity in the presence of harmonic trap . after the release of the trap , the eigenfunctions of the continuous spectrum with energy @xmath91 are @xmath92 where the phase shift @xmath93 should be determined by the bethe - peierls boundary condition @xmath94 , and one finds @xmath95 . at unitarity , @xmath96 and thus @xmath97 . for ideal gas without interaction , @xmath98 and thus @xmath99 .
furthermore , the functions @xmath100 satisfy the orthogonality condition @xmath101 @xcite .
the projection of the initial wave function @xmath102 on such a basis gives @xmath103 which provides a useful quantity allowing for the calculation of the density and momentum distributions as a function of time .
the behavior of @xmath104 at large momenta satisfying the condition @xmath105 is related to the behavior of the wave functions at short distances satisfying the condition @xmath106 . substituting the approximated wave function @xmath107 into eq .
( [ eq : cp ] ) and keeping terms up to the first order in @xmath35 , we find @xmath108 if the interaction is turned off and @xmath109 if the interaction is present during the expansion . in the latter case , the _ approximate _ initial wave function @xmath110 is orthogonal to @xmath100 as it does not account for the effect of the harmonic trap .
a more accurate expression for @xmath104 is obtained by taking into account that the _ exact _ initial wave function is only approximately orthogonal to @xmath100 and satisfies the equation @xmath111 . on the other hand the wave function @xmath112 satisfies the equation @xmath113 .
multiplying the first equation by @xmath114 and the second one by @xmath115 , followed by an integration with respect to @xmath10 , and using the fact that _ both _ wave functions satisfy the bethe - pierels boundary condition at @xmath116 , we get @xmath117 note that at the value @xmath118 , the numerator in eq . (
[ eq : cp2 ] ) vanishes linearly and thus @xmath104 is well - defined for all values of @xmath119 . to find the behavior of @xmath104 at large @xmath119
, one can again replace @xmath102 with the small @xmath10 approximation .
a direct integration gives @xmath120 revealing a much faster decay of @xmath104 at large @xmath119 with respect to the case when the interaction is turned off simultaneously with the release of the trap . in particular @xmath121 if @xmath122 and @xmath123 if @xmath124 @xcite . in the absence of interaction ( solid blue line ) , in the presence of interaction ( dashed green line ) and after quenching the scattering length to infinity ( dotted orange line ) .
the absolute value is added because @xmath104 changes sign at the dips of the blue and orange lines .
the explicit @xmath119 dependence of @xmath104 at small and large @xmath119 is explained in the main text . ]
figure [ fig : cp ] shows the full numerical calculation of @xmath104 with the exact initial wave function ( eq . ( [ eq : eigenfunction ] ) ) . at large momenta ( i.e , in the region of @xmath105 ) , we find @xmath125 in the absence of interaction whereas it behaves like @xmath126 ( for @xmath127 ) and @xmath128 ( for @xmath124 ) in the presence of interaction , with the corresponding coefficients in excellent agreement with the above analytic estimate . as already pointed out the momentum distribution preserves its initial form if the interaction is turned off , whereas it dramatically changes during the expansion if the interaction is present .
such a different expansion dynamics is directly connected with the different behaviors exhibited by @xmath104 at large momenta . with the help of @xmath104 ,
the wave function during the expansion can be in fact calculated straightforwardly : @xmath129 and the wave function @xmath130 in momentum space can be obtained after a direct fourier transformation .
let us first consider the case where the interaction is turned off just before the expansion . since in this case
the initial @xmath0 large momentum tail is preserved during the expansion , a @xmath2 tail at large distance is predicted to develop in the density distribution during the expansion . as shown in fig .
[ fig : nr](a ) , a tail of the form @xmath2 does actually appear in @xmath79 at @xmath131 , beyond the exponential decay . in terms of @xmath132 ,
the wave function in momentum space takes the form @xmath133 since @xmath134 at large @xmath5 , thus the large momentum distribution @xmath135 which , as expected , coincides with the initial one .
( red ) @xmath136 ( green ) , @xmath137 ( blue ) .
the linear part of the density distribution @xmath72 at large @xmath10 exhibits the @xmath2 behavior if the interaction is turned off , whereas it exhibits a @xmath138 tail followed by a tail of @xmath139 if the interaction is present.,width=325 ] if the interaction is not switched off , the expansion dynamics is dramatically different . in fig . [
fig : nr](b ) we show the density distributions resulting from the direct calculation of eq .
( [ fullwavefunction ] ) . due to the presence of interaction
, the real space density always satisfies @xmath140 . beyond the exponential decay
, we find that a tail of the form @xmath141 followed by a tail of the form @xmath142 appears soon after a short expansion time and is preserved afterwards .
this can be understood by applying to eq .
( [ fullwavefunction ] ) the saddle point approximation which is valid at large @xmath82 . at the saddle point
@xmath143 , we find @xmath144 the absence of the @xmath2 tail in the density distribution indicates that the @xmath0 large momentum tail must disappear after a long enough expansion time . indeed , as shown in fig .
[ fig : nk ] , the large momentum distribution @xmath145 decreases and shrinks during the expansion , similarly to the case of the expansion in the presence of unitary interaction ( see fig . [ fig : unitarity ] ) .
since the momentum distribution reaches the ballistic regime at a different rate for different momenta , strong interference oscillations appear in the intermediate momentum regime connecting the low and large momentum sectors .
a careful analysis shows that the average of these oscillations results in a @xmath146 behavior with the increase of time .
therefore , when the ballistic relation is satisfied , the momentum distribution should exhibit a @xmath146 tail followed by a @xmath147 tail at @xmath148 , consistently with the large @xmath10 behavior of the density distribution ( see eq .
( [ eq : rhort ] ) ) .
( red ) @xmath136 ( green ) , @xmath137 ( blue ) and the asymptotic momentum distribution ( black ) .
the linear part of the momentum distribution at large @xmath5 exhibits a tail of @xmath0 whereas the asymptotic momentum distribution exhibits a @xmath146 tail followed by a @xmath149 tail at large @xmath5 . ] to investigate the time - dependence of the @xmath0 large momentum tail at large but finite @xmath82 , it is important to calculate the evolution of the wave function at small @xmath10 . in this case
only the small momentum behavior of @xmath104 is important for the integration in eq .
( [ fullwavefunction ] ) where @xmath150 ( see fig . [
fig : cp ] ) .
thus , the wave function is @xmath151 at small @xmath10 .
the exponential phase factor can be dropped at large time and the momentum distribution behaves like @xmath152 at large @xmath5 . in conclusion
the contact decreases as @xmath153 for @xmath154 which agrees with the numerical calculation of @xmath80 shown in fig .
[ fig : nk ] .
the time dependence of the decay is the same as in the case of the weakly interacting bose gas discussed in the first part of the paper , thereby providing a justification of the adiabatic ansatz employed to derive the @xmath155 decay law . in the last part of the work we consider the time evolution of the contact after a quench of the scattering length from a small value to @xmath51 just before the expansion .
this problem , already investigated both experimentally@xcite and theoretically@xcite in the presence of a harmonic trap , will be considered here by switching off the trap simultaneously with the quench of the scattering length . at unitarity ,
the continuum eigenstates are @xmath97 . using the approximated form of the initial wave function @xmath156 at small @xmath10 , where @xmath157 is the scattering length before the quench
, we get @xmath158 at large @xmath119 ( see fig . [
fig : cp ] ) . applying the saddle point approximation to eq .
( [ fullwavefunction ] ) at the saddle point @xmath143 , one then finds @xmath159 using the ballistic relation ( eq .
( [ eq : ballistic ] ) ) , we find that the asymptotic momentum distribution behaves like @xmath160 .
consequently , the @xmath0 tail of the momentum distribution should evolve into @xmath160 after long expansion time .
figure [ fig : quench ] shows the time - dependent evolution of the density and momentum distributions in the expansion after quenching the scattering length from @xmath161 to @xmath51 . a tail of the form @xmath162 is developed in the density distribution , in good agreement with the above analytical calculation .
the momentum distribution can be separated into two stages . at shorter times , the two - body correlations increase due to the quench and the @xmath0 tail increases and quickly reaches the in - trap equilibrium value , as investigated in ref .
@xcite . at longer times , the expansion dynamics dominates and the contact decreases , in agreement with our previous results for the unitary interaction . since the rate of the evolution is slow , a new @xmath160 tail is expected to appear after a long expansion time . at @xmath163 .
different colors are the distributions at different expansion time @xmath69 ( red ) , @xmath136 ( green ) , @xmath137 ( blue ) .
the black dotted line indicates the position of the large momentum tail for the in - trap equilibrium system with unitarty interaction .
the linear part of the density distribution at large @xmath10 exhibits a tail of @xmath162 at large times .
the linear part of the momentum distribution at large @xmath5 exhibits a tail of @xmath0 which first increases at short times and then decreases at larger expansion times.,width=325 ]
in this paper we have investigated the time - dependence of the density and momentum distribution of various expanding physical systems after the release from a harmonic trap , with special focus on the behavior of the contact parameter .
we have explicitly shown that the momentum distribution changes dramatically if the interaction is present during the expansion and the large momentum tail decreases and eventually disappears for large times .
on the other hand , differently from the case when the interaction is switched off before the expansion , the density distribution , which is the observable quantity in the time - of - flight experiments , does not exhibit any @xmath2 behavior , even at intermediate times where the @xmath0 tail in the momentum distribution is still visible .
so far , our investigations have been devoted to isotropically trapped systems .
the presence of anisotropy in the trapping potential is not expected to introduce major differences in the large @xmath5 behavior of the momentum distribution of the expanding gas , although its role might deserve further theoretical investigation .
on the basis of our results we conclude that the @xmath2 density tail , observed after the release of an anisotropic trap in the recent experiment by r. chang , _ et .
al _ carried out on a weakly interacting bose gas of metastable @xmath1he atoms @xcite , is still elusive and remains an open problem to understand .
_ note added. _ in the preparation of the manuscript , we noticed a preprint by s. e. gharashi and d. blume@xcite on the evolution of the contact for a smoothly released one dimensional harmonic trap .
we would like to thank david clemnt and alain aspect for stimulating our interest in the problem of the contact during the expansion and for informing us about their recent experiment on the expansion of metastable @xmath1he atoms .
useful correspondence and discussions with shina tan , stefano giorgini and christophe salomon are acknowledged .
c. qu thanks yangqian yan for helpful discussions .
this work was supported by the quic grant of the horizon2020 fet program and by provincia autonoma di trento .
using the scaling law ( eq .
( [ castin ] ) ) , an exact relationship can be obtained also for the time dependence of the one - body density matrix , @xmath164 introducing the sum and difference of these two position vectors , @xmath165 and @xmath166 , and taking the fourier transformation , one then obtains the exact result @xmath167 \label{eq : onebodynk}\ ] ] for the momentum distribution at time @xmath11 , in terms of the initial value of the one - body density matrix . to calculate the asymptotic momentum distribution @xmath168 it is convenient to introduce the rescaled variable @xmath169 . at large times , where @xmath170 ,
one then easily finds the result @xmath171 which coincides with the result ( eq . ( [ eq : asym_unitary ] ) ) derived in sec . [ sec : manybody ] employing the ballistic relation
. the time - dependence of the contact at finite expansion times can be also obtained by calculating the asymptotic behavior of eq .
( [ eq : onebodynk ] ) at large @xmath5 . for large momenta satisfying @xmath172 which implies @xmath173 , with @xmath174 the initial size of the trapped gas
, the variable @xmath175 can be ignored in the exponential factor of eq .
( [ eq : onebodynk ] ) and the momentum distribution approaches the scaling behavior @xmath176 as a consequence , the contact takes the form @xmath177 which coincides with the result obtained using the adiabatic ansatz for the unitary fermi gas .
99 s. tan , http://dx.doi.org/10.1016/j.aop.2008.03.004[ann .
phys . * 323 * , 2952 ( 2008 ) ; ] http://dx.doi.org/10.1016/j.aop.2008.03.005[*323 * , 2971 ( 2008 ) ; ] http://dx.doi.org/10.1016/j.aop.2008.03.003[*323 * , 2987 ( 2008 ) . ]
s. utsunomiya , l. tian , g. roumpos , c. w. lai , n. kumada , t. fujisawa , m. kuwata - gonokami , a. lffler , s. hfling , a. forchel , and y. yamamoto , http://dx.doi.org/10.1038/nphys1034[nature physics * 4 * , 700 ( 2008 ) . ]
we consider here the second energy level of eq .
( [ eq : eigenvalue ] ) to avoid the use of the first energy level of the pseudopotential interaction which corresponds to a bound state for @xmath183 in the absence of the harmonic trap .
it is worth to note that the regime @xmath184 does not exist in the hard sphere model where the range of interaction @xmath185 , because the whole theory is valid at the condition @xmath186 .
however , this regime can be reached if @xmath187 . | we study the expansion of an interacting atomic system at zero temperature , following its release from an isotropic three - dimensional harmonic trap and calculate the time dependence of its density and momentum distribution , with special focus on the behavior of the contact parameter .
we consider different quantum system , including the unitary fermi gas of infinite scattering length , the weakly interacting bose gas and two interacting particles with highly asymmetric mass imbalance . in all
cases analytic results can be obtained , which show that the initial value of the contact , fixing the @xmath0 tail of the momentum distribution , disappears for large expansion times .
our results raise the problem of understanding the recent experiment of r. chang _
et al . _
( arxiv:1608.04693 ) carried out on a weakly interacting bose gas of metastable @xmath1he atoms , where a @xmath2 tail in the density distribution was observed after a large expansion time , implying the existence of the @xmath0 tail in the asymptotic momentum distribution . |
among the most fundamental questions in cosmology are those relating to the natures of the geometry and topology of the universe . the geometry can be obtained from the matter - energy content of the universe , through einstein s field equations . in this general relativistic context , the universe seems to be well described by a locally homogeneous and isotropic friedmann - lematre - robertson - walker ( flrw ) metric , with a constant spatial curvature with sign @xmath3 , @xmath4 .
general relativity is however a local metrical theory , and as such does not determine the topology of spacetime . in general , the spatial section of a flrw spacetime can be a multiply connected manifold ( which we assume compact and orientable ) @xmath5 , where the covering space @xmath6 is either @xmath7 , @xmath8 or @xmath9 , depending upon @xmath10 ; and @xmath11 is a discrete and fixed - point free group of isometries of @xmath6 . thus ,
unless specified by a fundamental theory , the topology of the universe is expected to be inferred from observations.@xcite an immediate consequence of such multiple - connectedness , is that there will be multiple images of each radiating source , one for each isometry in @xmath11 .
therefore , searching for a non - trivial topology amounts to a search for multiple images or pattern repetitions . due to the existence of a cosmological horizon , however , only the subset of images within the observable universe ( i.e. a sphere of radius @xmath12 , where @xmath12 is the redshift distance relation evaluated at @xmath13 of the survey used , centered at the observer s position ) is detectable.@xcite most studies of cosmic topology so far have concentrated on particular manifolds .
given that there are an infinite number of possible topologies for each curvature sign , a more reasonable search strategy is to ask what is the set of all detectable topologies .
in general , each candidate manifold has a distinct isometry group @xmath11 .
since not all isometries give rise to detectable multiple images , it is reasonable to restrict our focus only to the subgroup generated by _
detectable _ isometries , which we refer to as the _ local shape of the universe_. as we shall show , this greatly simplifies the search for image repetitions , while making it possible to obtain some very general results . in the following we briefly outline our recent results for the spherical spaces,@xcite which show that in an appropriate limit typical manifolds have approximately the same local shape , and further discuss their observational implications .
recent high precision data by wmap@xcite are making it possible to comprehensively probe the topology of the universe for the first time , using pattern repetition .
in addition they have provided strong evidence suggesting that the universe is nearly flat , with @xmath14 . according to the most recent estimates of the density parameters ,
@xmath15 $ ] with a @xmath16 confidence,@xcite which still allows the 3-space to be spherical , hyperbolic or flat . furthermore ,
in a non - flat inflationary universe one would typically expect @xmath17 , implying @xmath18 in units of the curvature radius .
in addition , non - flat manifolds are rigid , in the sense that all metrical quantities , such as the length of closed geodesics , are fixed in units of the curvature radius .
hence , the inflationary limit imposes important constraints on the set of detectable isometries for non - flat manifolds , and hence on their local shape .
our aim here is to take a general point of view and ask what are the set of topologies that would be detectable in very nearly flat universes .
thus to proceed we shall , in addition to assuming @xmath19 , make two further physically motivated assumptions : ( i ) the observer is at a position @xmath20 where the topology is detectable , i.e. @xmath21 , where @xmath22 is the injectivity radius at @xmath20 defined as half the length of the smallest closed geodesic passing through @xmath20 ( see ref . for details ) , and ( ii ) the topology is not excludable , i.e. it does not produce too many images so as to be ruled out by present observations .
thus , our main physical assumption can be summarized as @xmath23 these assumptions severely restricts the set of detectable non - flat manifolds .
thus in the case of spherical manifolds , only lens spaces ( with @xmath24 ) and binary dihedrical spaces ( with @xmath25 ) of sufficiently high order of @xmath26 or @xmath27 are detectable . in the hyperbolic case ,
the only detectable manifolds are the so - called nearly cusped manifolds , which are sufficiently similar to the cusped manifolds ( cusped manifolds are not compact , and possess regions with arbitrarily small @xmath22 . ) in a recent study@xcite we considered both classes of manifolds and showed that a generic detectable spherical or hyperbolic manifold is locally indistinguishable from either a cylindrical ( @xmath1 ) or toroidal ( @xmath2 ) manifold , irrespective of its global shape . here
, we shall briefly review our results , and further discuss their observational implications . briefly the key arguments are the following . at any given point @xmath20
, an isometry @xmath28 will generate a closed geodesic with length @xmath29 . but
given the limit , ( [ hip ] ) we are only interested in geodesics short enough so that both the sources and their images lie within the cosmological horizon .
the question then is what do these detectable closed geodesic look like ?
in particular how much do they deviate from the clifford translations ( i.e. isometries with constant distance function @xmath30 ) within the observable universe ?
we have found that for a lens space @xmath31 with generator @xmath32 , the following results hold : \i ) the manifold @xmath31 is equivalent to the manifold @xmath33 , where @xmath34 is the second to last convergent in the continued fractions expansion of @xmath35 .
\ii ) the shortest geodesic at any point is generated by @xmath36 , for some convergent @xmath37 .
\iii ) @xmath38 .
thus by choosing @xmath26 to be sufficiently large , one can at each point of a lens space intersects a closed geodesic which is small in units of the curvature radius .
we employ this bound to divide the @xmath39 plane into detectable , undetectable and observationally excluded regions , as shown in ( fig .
thus any systematic search for a lens space - shaped universe needs only to concentrate on manifolds that lie in the white central region of the diagram .
\iv ) consider an observable sphere of radius @xmath12 , centered at an observer s position @xmath40 ( fig .
the isometry @xmath32 identifies points in two faces of the fundamental polyhedron ( dashed lines , shown edgewise for simplicity ) .
let @xmath41 be the distance @xmath42 and let @xmath43 and @xmath44 denote the maximum and minimum values of @xmath45 for a point @xmath20 within the detectable sphere .
we then find @xmath46 where @xmath47 and @xmath48 are position dependent , obeying @xmath49 .
a similar result holds for the cusp - like parts of nearly cusped hyperbolic manifolds , namely @xmath50 the above inequalities show clearly that subject to condition ( [ hip ] ) , the detectable isometries of lens spaces ( as well as cusp - like hyperbolic manifolds ) are very close to clifford translations for most observers .
we also note that all isometries in the isometry groups of the binary dihedral spaces are clifford translations , and therefore trivially obey the inequality above .
# 1#20.65 # 1 these results have important consequences for developing search strategies for cosmic topology .
they show that for a typical observer in a very nearly flat universe , the detectable part of the topology would be indistinguishable from either @xmath1 or @xmath51 manifold .
note that the detectable isometries can not have more than 2 independent generators because the fundamental domain has a far greater volume than the observable universe .
we also emphasize that no matter how flat the universe turns out to be , there are always an infinite number of candidate manifolds , both spherical and hyperbolic , that have detectable isometries .
furthermore , in the case of ( detectable ) spherical manifolds , we have also proven that through any point in the manifold there is a closed geodesic of length much smaller than the curvature radius , and therefore the fraction of the total manifold s volume where there are detectable isometries is indeed large .
so far our discussion has assumed the inflationary limit , @xmath52 .
however , the resolutions required to test such limits are not expected to be attainable in near future .
currently the best fit value for @xmath53 is @xmath54 to @xmath55.@xcite these bound are also compatible with other topologies not considered here , such as the recently proposed poincar dodecahedral space.@xcite it is therefore important to ask what the present results can tell us for less restrictive bounds .
results ( i)-(iii ) hold regardless of the value of @xmath53 .
the bounds in ( iv ) however rely on approximations that remain valid only if @xmath56 . for @xmath53 of this order , the local shape becomes distinguishable from @xmath1 or @xmath2 .
nevertheless , the bounds shown here still allow us to severely constrain the magnitude of the deviation.@xcite currently one of the most promising methods of searching for cosmic topology is the so called circles - in - the - sky method , where one looks for matching circles of anisotropies in cmb radiation due to topological identification
. for clifford translations in flat space these pairs must be strictly antipodal in the celestial sphere .
hence , in the inflationary limit , the circles due to isometries of spherical and hyperbolic manifolds should also be nearly antipolodal .
a preliminary search for nearly antipodal ( i.e. , with deviation @xmath57 ) and with radius larger than @xmath58 was undertaken in ref .
, with negative results .
it can be shown , however@xcite , that for @xmath59 the result ( iv ) still allows @xmath60 .
also , ref .
claims to have found evidence of antipodal circles of radius @xmath61 .
given the infinite number of possible candidate manifolds for cosmic topology , any realistic search strategy must be able to radically restrict the expected possibilities .
this is particularly important in the inflationary limit , where the order of any detectable cyclic subgroup as well as the number of candidate isometries which generate small geodesics are extremely large .
the results presented here are quite general , imposing severe constraints on the set of detectable isometries for all detectable very nearly flat manifolds . by severely restricting the expected topological signatures of detectable isometries
, we are able to provide an effective framework for searching for evidence of a non - trivial topology in cosmological observations .
more specifically , these results can be used to confine the parameter space which realistic search strategies such as the ` circles - in - the - sky ' method , need to concentrate on .
we thank cnpq , cbpf / mct and faperj for the grants under which this work was carried out .
bm thanks astronomy unit , qmul , for hospitality .
g. f. r. ellis , _ gen .
* 2 * , 7 ( 1971 ) ; d. d. sokolov and v. f. shvartsman , _ sov .
jetp _ * 39 * , 196 ( 1974 ) ; d. d. sokolov and a. a. starobinsky , _ sov .
* 19 * , 629 ( 1975 ) ; j. r. gott , _ mon . not .
r. astron .
soc . _ * 193 * , 153 ( 1980 ) ; h. v. fagundes , _ phys .
lett . _ * 51 * , 417 ( 1983 ) ; l - z .
fang and h. sato , _ gen .
* 17 * 1117 ( 1985 ) ; h. v. fagundes and u. f. wichoski , _ astrophys .
j. _ * 322 * , l52 ( 1987 ) ; w. oliveira , m. j. rebouas and a. f. f. teixeira , _ phys .
lett . _ * a188 * , 125 ( 1994 ) ; m. lachize - rey and j .- p .
luminet , _ phys . rep . _
* 254 * , 135 ( 1995 ) ; r. lehoucq , m. lachize - rey and j .- p .
luminet , _ astron .
astrophys . _
* 313 * , 339 ( 1996 ) ; a. de oliveira - costa , g. f. smoot and a. a. starobinsky , _ astrophys . j. _ * 468 * , 457 ( 1996 ) ; b. f. roukema and a. edge , _ mon . not .
r. astron .
soc . _ * 292 * , 105 ( 1997 ) ; a. bernui , g. i. gomero , m. j. rebouas and a. f. f. teixeira , _ phys .
rev . _ * d57 * , 4699 ( 1998 ) ; h. v. fagundes and e. gausmann , astro - ph/9811368 ; m. j. rebouas , r. tavakol and a. f. f. teixeira , _ gen .
grav . _ * 30 * , 535 ( 1998 ) ; n. j. cornish , d. spergel and g. starkman , _ class .
quantum grav . _
* 15 * , 2657 ( 1998 ) ; j. levin , e. scannapieco and j. silk , _ class .
grav . _ * 15 * , 2689 ( 1998 ) ; r. lehoucq , j .-
luminet and j .- p .
uzan , _ astron .
astrophys . _ * 344 * , 735 ( 1999 ) ; a. bernui and a. f. f. teixeira , gr - qc/9904180 ; j .-
uzan , r. lehoucq and j .- p .
luminet , _ astron .
_ * 351 * , 766 ( 1999 ) ; h. v. fagundes and e. gausmann , _ phys .
lett . _ * a261 * , 235 ( 1999 ) ; m. j. rebouas , _ int .
j. mod .
phys . _ * d9 * , 561 ( 2000 ) ; g. i. gomero , m. j. rebouas , a. f. f. teixeira and a. bernui , _ int .
j. mod .
phys . _ * a15 * , 4141 ( 2000 ) ; r. lehoucq , j .-
uzan and j .- p .
luminet , _ astron .
astrophys . _ * 363 * , 1 ( 2000 ) ; g. i. gomero , m. j. rebouas and a. f. f. teixeira , _ phys . lett . _ * a275 * , 355 ( 2000 ) ; j. r. bond , d. pogosyan and t. souradeep , _ phys .
rev . _ * d62 * , 043005 ( 2000 ) ; j. r. bond , d. pogosyan and t. souradeep , _ phys .
_ * d62 * , 043006 ( 2000 ) ; g. i. gomero , m.j .
rebouas and a.f.f .
teixeira , _ int .
j. mod .
phys . _ * d9 * , 687 ( 2000 ) ; d. muller , h. v. fagundes and r. opher , _ phys . rev . _ * d63 * , 123508 ( 2001 ) ; g. i. gomero , m. j. rebouas and a. f. f. teixeira , _ class .
quantum grav .
_ * 18 * , 1885 ( 2001 ) ; e. gausmann , r. lehoucq , j .-
luminet , j .-
uzan and j. weeks , _ class .
grav . _ * 18 * , 5155 ( 2001 ) ; d. muller , h. v. fagundes and r. opher _ phys .
rev . _ * d66 * , 083507 ( 2002 ) ; g. i. gomero , a. f. f. teixeira , m. j. rebouas and a. bernui , _ int .
phys . _ * d11 * , 869 ( 2002 ) ; g. i. gomero , m. j. rebouas and r. tavakol , _ int .
j. mod .
* a17 * , 4261 ( 2002 ) ; j. j. levin , _ phys . rep . _ * 365 * , 251 ( 2002 ) ; g. i. gomero and m. j. rebouas , _ phys .
* a311 * , 319 ( 2003 ) ; j. r. weeks , _ mod . phys .
_ * a18 * , 2099 ( 2003 ) ; j. weeks , r. lehoucq and j .- p .
uzan , _ class .
_ * 20 * , 1529 ( 2003 ) ; a. hajian and t. souradeep , astro - ph/0301590 ; c. j. copi , d. huterer and g. d. starkman , astro - ph/0310511 ; j .-
luminet , j. weeks , a. riazuelo , r. lehoucq and j .- p .
uzan , nature * 425 * , 593 ( 2003 ) ; m.j .
rebouas and g.i .
gomero , _ braz .
j. phys . _
* 34 * , 1358 ( 2004 ) ; b. mota , m. j. rebouas and r. tavakol , astro - ph/0403110 ; b. f. roukema et al .
, astron .
astrophys . * 423 * , 821 ( 2004 ) ; r. aurich , s. lustig and f. steiner , astro - ph/0412569 ; a. riazuelo , j. weeks , j .-
uzan and j .- p .
luminet , _ phys . rev . _ * d69 * , 103518 ( 2004 ) ; m. o. calvo , g. i. gomero , b. mota and m.j .
rebouas , astro - ph/0404536 ; n. j. cornish , d. n. spergel , g. d. starkman and e. komatsu , _ phys .
_ * 92 * , 201302 ( 2004 ) ; a. bernui , g. i. gomero , b. mota and m. j. rebouas , astro - ph/0403586 ; a. de oliveira - costa , m. tegmark , m. zaldarriaga and a. hamilton , _ phys .
* d69 * , 063516 ( 2004 ) ; a. bernui and t. villela , to appear in astron .
astrophys .
( 2005 ) ; j .-
p luminet , astro - ph/0501189 ; a. hajian and t. souradeep , astro - ph/05010001 ; t. souradeep and a. hajian , astro - ph/0502248 .
g. i. gomero , m. j. rebouas and r. tavakol , _ class .
quantum grav .
_ * 18 * , 4461 ( 2001 ) ; g. i. gomero , m. j. rebouas and r. tavakol , _ class .
quantum grav . _
* 18 * , l145 ( 2001 ) ; g. i. gomero , m. j. rebouas and r. tavakol , _ int .
j. mod.phys._ * a17 * , 4261 ( 2002 ) ; b. mota , m. j. rebouas and r. tavakol , _ class . quantum grav . _ * 20 * , 4837 ( 2003 ) . | recent high precision data by wmap and sdss have provided strong evidence to suggest that the universe is nearly flat .
they are also making it possible to probe the topology of the universe .
motivated by these results , we have recently studied the consequences of taking the inflationary limit , i.e. @xmath0 .
we have shown that in this limit a generic detectable spherical or hyperbolic topology is locally indistinguishable from either @xmath1 or @xmath2 , irrespective of its global shape . here
we briefly present these results and further discuss their observational implications . |
it is generally presumed that life on earth developed from an initial reservoir of simpler prebiotic organic material such as amino acids , fatty acids and sugars .
what is not clear is whether the chemical evolution to produce this initial reservoir occurred on the early earth , in the pre - solar nebula , or even earlier in interstellar molecular clouds .
in fact , a significant body of work exists which suggests that the necessary chemical evolution could not have taken place on the early earth itself ( joyce et al .
1984 ; bailey et al 1998 ) , suggesting that the chemical evolution took place within the pre - solar nebula , with the organic material being delivered to the earth by meteorites and comets during the phase of bombardment .
although there is expected to be some chemical and physical processing of material originating in the ism , through the pre - solar nebula , into cometary and meteoritic material , and on to the surface of a planet , a study of the most complex molecules in the ism is an important step in determining how likely it is that the reservoir of molecules from which life evolved originated in the pre - solar nebula .
indeed , a large number of surprisingly complex molecules , including amino acids , have been found in carbonaceous chondrite meteorites ( see e.g. wirick et al .
2006 ) , which are thought to be relatively unprocessed remnants of the presolar nebula .
in this paper we describe an australia telescope compact array ( atca ) search for two biologically important molecules : the simplest amino acid glycine ( @xmath5 ) , and the cyclic molecule propylene oxide ( @xmath6 ) which is chiral .
chiral molecules are those where asymmetry in carbon atom placement within a molecule leads to two distinct forms ( mirror images or enantiomers ) .
figure [ fig1 ] shows the chemical structure of proylene oxide .
( 80,100)(1,2 ) ( 30,47 ) ch@xmath7 ( 80,47 ) ch ( 125,47 ) ch@xmath8 ( 58,80 ) o ( 39,59)(1,1)19 ( 54,51)(1,0)20 ( 101,51)(1,0)20 ( 85,59)(-1,1)19 there have been many searches for glycine in molecular clouds , dating back over 20 years ( eg .
brown et al .
1979 , hollis et al . 1980 ,
snyder et al . 1983 ) with negative or ambiguous results .
miao et al .
( 1994 ) and snyder ( 1997 ) reported a possible detection of a 90 ghz transition of glycine in sgr b2 ( n ) with bima and ovro ( interferometers ) .
there were some caveats in snyder ( 1997 ) due to uncertainty in the glycine line rest frequency , the different velocity components in sgr b2 ( n ) , and the confusion with other lines .
combes , rieu & wlodarczak ( 1996 ) used the iram 30-m telescope ( single dish ) for the positions orion sio and sgr b2 ( oh ) and concluded that the confusion of lines in the millimetre spectra , particularly in the orion position , could make the detection of glycine impossible , as the glycine lines would be below the confusion limit .
kuan et al.(2003 , 2004a , 2004b ) have reported the detection of glycine in sgr b2 ( n - lmh ) , orion kl and w51 e1/e2 using the nrao 12-m telescope .
they used the detailed comparison of lines in the different sources to help solve the problem of `` interloper '' lines of other species . the use of rotation diagrams to provide a quantitative comparision of many detections and upper limits of the glycine lines for all three sources gives greater confidence in the results .
the total column density of glycine conformer i determined by kuan et al .
( 2003 ) in sgr b2 ( n - lmh ) is @xmath1 @xmath2 and rotational temperature @xmath9 k. this reported detection of kuan et al .
( 2003 ) has been disputed in a paper by snyder et al.(2005 ) , who use unpublished nrao 12-m data , and published sest data from nummelin et al .
( 1998 ) to show that upper limits from non - detections of glycine lines quite stringently rule out glycine in the three sources at the level reported by kuan et al .
( 2003 ) unless present as the higher energy conformer ii .
in addition , they argue how random coincidences of lines misidentified as glycine could lead to a plausible looking , but spurious , fit to lines in a rotation diagram .
investigations of primitive mechanisms of self - replication that involve rna only , indicate the need for a reservoir of homochiral ( having like chirality ) organic molecules on the early earth , before living organisms could evolve ( joyce et al .
an excess of about 10 % in one enantiomer over the other ( l over d in the case of amino acids on the earth ) is needed for rna template reproduction to work .
non - biological processes always create chiral organic molecules in a racemic mix ( equal numbers of both handednesses ) and there are no known abiotic mechanisms operating on the surface of the earth to generate the required enantiomeric excess .
therefore , an extraterrestrial origin for the prebiotic molecules needed to seed life on earth is likely if we accept the hypothesis that primitive reproduction involved only rna . an extraterrestrial mechanism for producing the requisite homochirality is necessary .
the most likely mechanism is circularly polarized uv radiation in the ism , which is known to be capable of generating an enantiomeric excess by selectively destroying one enantiomer over the other ( bailey et al 1998 ) .
regrettably radio astronomy can not distinguish the mirror image molecular forms , which in different concentrations would signal chiral imbalance , but it can potentially detect the presence of molecules capable of this property . to date , no molecules possessing a chiral centre have been detected in the ism .
glycine is not a chiral molecule , and our search for biologically important chiral molecules with the atca and mopra is focusing on propylene oxide .
the simpler , non - chiral , but chemically related cyclic species ethylene oxide ( @xmath10 ) has already been detected in sgr b2n ( dickens et al .
1997 ) .
two developments prompted this search for glycine and propylene oxide : * firstly , we have been able to use laboratory measured values of the rest frequencies of glycine and propylene oxide transitions between 85 and 115 ghz rather than the less accurate calculated values ( lovas et al .
1995 ) on which previous searched at 3 mm have had to rely .
the measurements were made at the school of chemistry , monash university , using a purpose - built stark - modulated free - jet microwave absorption spectrometer of a design similar to that described previously ( brown et al .
these measured values have an uncertainty of around only 0.07 km / s in the central velocity of a transition . * secondly , the new capabilities of the atca at millimetre wavelengths allows a search at high spatial resolution , minimising beam dilution effects and , as an interferometer , filtering out confusing lines from the extended gas within the beam .
our search has a multi - pronged approach , involving both single dish and interferometer observations of multiple frequencies within the 3-mm band .
the position sgr b2 ( lmh ) is of small diameter ( @xmath115 arcseconds ) and has so far provided detections of the related complex molecules methyl formate and acetic acid ( which is a precursor to glycine in some reaction networks ) .
these molecules seem to be confined to the lmh .
however , another similar molecule , glycolaldehyde ( hollis et al . 2001 ) , was found to be extended in the sgr b2 ( n ) molecular cloud , with very low concentrations in the lmh .
these three molecules are believed to have a common origin with glycine ( sorrell 2001 ) based on a model of chemical reactions in icy grain mantles .
it is therefore important to search for glycine in both extended and compact dense molecular gas .
we are thus using both the atca and the single dish mopra radio telescope .
the mopra telescope , jointly operated by the atnf csiro , and the university of new south wales , australia , has a 22-m diameter , making it the largest single dish telescope operating at 3-mm in the southern hemisphere .
mopra results will be discussed in a subsequent paper . in the laboratory
, glycine can adopt more than one conformation or spatial arrangement of the groups . since it is uncertain which conformer of glycine may be present in the ism , we searched for lines from both conformer i and conformer ii ( lovas et al . 1995 )
. conformer i has lower energy than ii by around @xmath12 @xmath13 ( lovas et al .
1995 ) , but conformer ii has a larger dipole moment and hence stronger lines .
there is a significant energy barrier between conformers i and ii , since to go from one to the other requires breaking a hydrogen bond rather than just an internal rotation .
the energy barrier is estimated theoretically at around @xmath14 @xmath13 or @xmath15 ( godfrey , brown & rodgers 1996 ) . at interstellar temperatures
the rate of tunnelling through the barrier will be low .
if glycine is formed at low temperatures on grain surfaces , it may be stuck in the original conformation , and if the formation chemical processes favoured conformer ii , it would relax only very slowly into the lower energy conformer i. thus it is worthwhile searching for both conformers .
observations were made with the australia telescope compact array , in 2002 and 2003 , during the southern winter period of conditions appropriate for 3-mm observations .
the log of observations is given in table 1 .
.log of atca observations , frequency bands and lines searched .
the frequencies quoted are from new laboratory measurements made at the school of chemistry , monash university .
the corresponding transitions are listed in table 4 .
the asterisk in column 4 denotes the line used as rest frequency for the velocity scale when more than one line was in the observed band .
i d specifies the molecule , propylene oxide ( po ) or glycine ( g ) , and conformer , i or ii ( glycine only ) , with this transition frequency .
strength is a measure of the intensity predicted at 10 k and 50 k ( 10 k if only one number is given ) , relative to a value of 100 for the strongest line of that species in the 80 - 115 ghz range .
[ cols="^,^,^ , < , < , < " , ] [ tab5 ] due to the large number of strong transitions of ethyl cyanide ( ch@xmath16ch@xmath17cn ) and methanol ( ch@xmath16oh ) in the 3-mm band , we have measured serendipitously 4 transitions of each molecule .
these two molecules are found in the small scale structure near continuum source n , at around 64 km / s , and also found at the other component at 73 km / s 5 arcsec to the north . to calculate the brightness temperature and column density ,
we assume a size @xmath18 arcsec , based on the median spatial size of line emission in our datacubes .
the upper level column densities are given in table 5 , along with some of the molecular parameters used .
given the detection of several transitions at different upper energy levels , we can fit rotation diagrams to derive the excitation temperature and total column density .
the fits give @xmath19 k and @xmath20 @xmath2 for ch@xmath16ch@xmath17cn in the main 64 km / s cloud , @xmath21 k and @xmath22 @xmath2 for ch@xmath16oh in the 64 km / s cloud and @xmath23 k and @xmath24 @xmath2 for ch@xmath16oh in the 73 km / s cloud .
these excitation temperatures are in good agreement with that expected for sgr b2 lmh , eg .
170 k for ch@xmath16oh from pei , liu & snyder ( 2000 ) .
the estimates of column density are dependent on the assumed source size , but our estimate of total column density for ch@xmath16ch@xmath17cn in sgr b2 lmh is about a factor of two higher than that obtained by miao & snyder ( 1997 ) , liu & snyder ( 1999 ) and liu , mehringer & snyder ( 2001 ) when corrected for the different assumed sizes .
similarly , our estimate of column density for ch@xmath16oh is in good agreement ( 30 % higher ) with that of pei et al .
( 2000 ) when corrected for source size .
we have used the atca during 2002 and 2003 to conduct a search for the simplest amino acid , glycine , and the simple chiral molecule propylene oxide , at 3-mm , using the three millimetre capable antennas available at that time .
we searched 15 portions of spectrum between 85 and 91 ghz , each of 64 mhz bandwidth .
the main results are as follows : * we have detected 58 emission features and 21 absorption features in the 15 portions of spectrum searched .
this gives a line density of 75 emission and 25 absorption lines per ghz stronger than the 5@xmath0 level of 110 mjy .
* we have tentatively assigned 23 of the detected spectral lines to transitions listed in the jpl on - line database ( table 3 ) but as many of these involve molecules not previously detected in the ism , these assignments can not be regarded with confidence .
* we did not confidently detect either glycine or propylene oxide , but can set 3@xmath0 upper limits for most transitions searched . * assuming that the glycine emission is extended with respect to the atca synthesised beam size of 17.0 x 3.4 arcsec@xmath3 , our 3@xmath0 upper limit of @xmath25 @xmath2 is consistent with that of @xmath1 @xmath2 of kuan et al .
( 2003 ) in their reported detection of glycine .
however , at the column density reported by kuan et al .
( 2003 ) we show that if glycine were confined to the scale of the lmh continuum source ( @xmath11 5 arcsec ) , it would have been easily detected in these atca observations .
these atca observations therefore put a strong upper limit on any small - scale glycine emission in sgr b2 , for both of conformers i and ii .
the monash authors acknowledge the financial support provided by the australia telescope national facility for the construction of the laboratory spectrometer , the spectrometer design , construction and technical services provided by jonathan g. crofts , and the collaboration with takeshi sakaizumi in the laboratory measurements of the glycine spectrum .
paj would like to acknowledge the support provided by an australia telescope national facility visiting fellowship .
we also thank the referee for very useful comments that improved the clarity and presentation of the paper .
kuan y .- j . , huang h .- c . , charnley s.b . ,
tseng w .- l . ,
snyder l.e . ,
ehrenfreund p. , kisiel z. , thorwirth s. , bohn r.k .
, wilson t.l .
2004b , in bioastronomy 2002 : life among the stars , iau symposium 213 , eds .
r. norris & f. stootman , p 185 | we have used the australia telescope compact array to conduct a search for the simplest amino acid , glycine ( conformers i and ii ) , and the simple chiral molecule propylene oxide at 3-mm in the sgr b2 lmh .
we searched 15 portions of spectrum between 85 and 91 ghz , each of 64 mhz bandwidth , and detected 58 emission features and 21 absorption features , giving a line density of 75 emission lines and 25 absorption lines per ghz stronger than the 5@xmath0 level of 110 mjy .
of these , 19 are transitions previously detected in the interstellar medium , and we have made tentative assignments of a further 23 features to molecular transitions .
however , as many of these involve molecules not previously detected in the ism , these assignments can not be regarded with confidence .
given the median line width of 6.5 km / s in sgr b2 lmh , we find that the spectra have reached a level where there is line confusion , with about 1/5 of the band being covered with lines .
although we did not confidently detect either glycine or propylene oxide , we can set 3@xmath0 upper limits for most transitions searched .
we also show that if glycine is present in the sgr b2 lmh at the level of @xmath1 @xmath2 found by kuan et al .
( 2003 ) in their reported detection of glycine , it should have been easily detected with the atca synthesized beam size of 17.0 x 3.4 arcsec@xmath3 , if it were confined to the scale of the lmh continuum source ( @xmath4 arcsec ) .
this thus puts a strong upper limit on any small - scale glycine emission in sgr b2 , for both of conformers i and ii .
[ firstpage ] ism : molecules - radio lines : ism - ism : individual : sgr b2 |
in recent articles we described calculations of the infinite volume , continuum limit of scalar and tensor glueball masses in the valence ( quenched ) approximation to lattice qcd @xcite . for a single value of lattice spacing and lattice volume ,
we reported also a calculation of the the decay coupling constants of the lightest scalar glueball to pairs of pseudoscalar mesons .
the mass and decay calculations combined support the identification of @xmath2 as primarily composed of the lightest scalar glueball @xcite .
evaluation of the mass of the lightest scalar quarkonium states and of quarkonium - glueball mixing amplitudes @xcite then yield a glueball component for @xmath2 of @xmath3 % . in the present article
, we describe the glueball mass data of ref .
@xcite in greater detail along with an improved evaluation of the mass predictions which follow from these data . for the scalar and tensor glueball masses
we obtain @xmath4 mev and @xmath1 mev , respectively .
the valence approximation , on which our results depend , may be viewed as replacing the momentum dependent color dielectric constant arising from quark - antiquark vacuum polarization with its zero - momentum limit @xcite and , for flavor singlet mesons , shutting off transitions between valence quark - antiquark pairs and gluons .
the valence approximation is expected to be fairly reliable for low lying flavor nonsinglet hadron masses , which are determined largely by the low momentum behavior of the chromoelectric field .
this expectation is supported by recent valence approximation calculations @xcite of the masses of the lowest flavor multiplets of spin 1/2 and 3/2 baryons and pseudoscalar and vector mesons .
the predicted masses are all within about 10% of experiment . for the lowest valence approximation glueball masses , the error arising from the valence approximation s omission of the momentum dependence of quark - antiquark vacuum polarization
we thus also expect to be 10% or less . refs .
@xcite show this error should tend to lower valence approximation masses below those of full qcd . for flavor singlet configurations
whose quantum numbers , if realized as quarkonium , require nonzero orbital angular momentum , it is shown in ref . @xcite that the additional error arising from the valence approximation s suppression of transitions between valence quark - antiquark pairs and gluons is likely to introduce an additional error of the order of 5% or less . for the lowest scalar glueball this error
is examined in detail in ref .
@xcite and found to shift the valence approximation mass by about 5% below its value in full qcd .
it is perhaps useful to mention that , for glueball masses , the valence approximation simply amounts to a reinterpretation of the predictions of pure gauge theory . in section [
sect : ops ] we define a family of operators used to construct glueball propagators . in section [
sect : lats ] we describe the set of lattices on which propagators were evaluated and the algorithms we used to generate gauge configurations and estimate error bars . in sections [ sect :
scalar ] and [ sect : tensor ] we present our results for scalar and tensor glueball propagators , respectively , and masses extracted from these propagators . in section [ sect : vol ] we estimate the difference between the scalar and tensor masses we obtain in finite volumes and the corresponding infinite volume limits . in section [ sect : cont ] we extrapolate scalar and tensor masses to their continuum limits . in section [
sect : comp ] we compare our calculations with work by other groups @xcite . for combined world average valence approximation scalar and tensor glueball masses
we obtain @xmath5 mev and @xmath6 mev , respectively .
we evaluated glueball propagators using operators built out of smeared link variables .
glueball operators built from link variables with an optimal choice of smearing couple more weakly to excited glueball states than do corresponding operators built from unsmeared links . as a consequence , the plateau in effective mass plots for optimally smeared operators begins at a smaller time separation between source and sink operators , extends over a larger number of time units , and yields a fitted mass with smaller statistical noise than would be obtained from operators made from unsmeared link variables .
examples of the improvements which we obtained by a choice of smeared operators with be given in section [ sect : scalar ] and [ sect : tensor ] .
initially , we constructed smeared operators from gauge links fixed to coulomb gauge .
this method gave adequate results for the largest values of lattice spacing we considered .
as the lattice spacing was made smaller , however , we found that the computer time required to gauge fix a large enough ensemble of configurations to obtain useful results became unacceptably large .
we then switched to a gauge invariant smearing method .
for the lattice sizes used in our extrapolation to the continuum limit , the gauge invariant mass results had statistical uncertainties typically a factor of three smaller than our earlier coulomb gauge results . in the remainder of the present article we discuss only the gauge invariant results .
a summary of our coulomb gauge mass calculations is given in ref .
@xcite . a family of gauge invariant smeared operators we construct following the adaptation in ref .
@xcite of the smearing method of ref . @xcite
. a related method of gauge invariant smearing is proposed in refs .
@xcite . for @xmath7 , @xmath8 , we define iteratively a sequence of smeared , space - direction link variable @xmath9 , with @xmath10 given by the unsmeared link variable @xmath11 .
let @xmath12 be @xmath13^{\dagger } + \nonumber \\ & & \epsilon \sum_j [ u_j^{n \epsilon}(x - \hat{j})]^{\dagger } u_i^{n \epsilon}(x - \hat{j } ) u_i^{n \epsilon}(x + \hat{i } - \hat{j}),\end{aligned}\ ] ] where the sum is over the two space directions @xmath14 orthogonal to direction @xmath15 .
the projection of @xmath12 into su(3 ) defines the new smeared link variable @xmath16 . to find @xmath16 we maximize over su(3 ) the target function @xmath17^{\dagger } \}.\end{aligned}\ ] ]
the required maximum is constructed by repeatedly applying an algorithm related to the cabbibo - marinari - okawa monte carlo method .
we begin with @xmath16 chosen to be 1 .
we then multiply @xmath18 by a matrix in the su(2 ) subgroup acting only on gauge index values 1 and 2 chosen to maximize the target function over this subgroup .
this multiplication and maximization step is repeated for the su(2 ) subgroup acting only on index values 2 and 3 , then for the subgroup acting only on index values 1 and 3 .
the entire three step process is then repeated five times .
five repetitions we found sufficient to produce a @xmath16 satisfactorily close to the true maximum of the target function in eq .
( [ projtarget ] ) .
iteratively maximizing the target function over su(2 ) subgroups turns out to be much easier to program than a direct maximization over all of su(3 ) . the additional computer time required for this iterative maximization , on the other hand , was a negligible fraction of the total time required for our calculation . from the @xmath9 we construct @xmath19 by taking the trace of the product of @xmath9 around the boundary of an @xmath20 square of links beginning in the @xmath21 direction .
the sum of @xmath19 over all sites with a fixed time value @xmath22 gives the zero - momentum loop variable @xmath23 .
for each triple @xmath24 , a field coupling the vacuum only to zero - momentum scalars states is @xmath25 where the sums are over space directions @xmath15 and @xmath14 .
a possible choice of the two independent operators coupling the vacuum only to zero - momentum tensor states is @xmath26 the optimal choice of @xmath24 for each operator and lattice spacing will be considered in the next section .
the set of lattices on which we evaluated scalar and tensor glueball propagators is listed in table [ tab : lats ] . on each lattice ,
an ensembles of gauge configurations was generated by a combination of the cabbibo - marinari - okawa algorithm and the overrelaxed method of ref .
@xcite . to update a gauge link we first performed a microcanonical update in the su(2 ) subgroup acting on gauge indices 1 and 2 .
this was then repeated for the su(2 ) subgroup acting on indices 2 and 3 , and the subgroup acting on indices 1 and 3 .
these three update steps were then repeated on each link of the lattice .
after four lattice sweeps each consisting of the three microcanonical steps on each link , we carried out one cabibbo - marinari - okawa sweep of the full lattice .
at least 10,000 sweeps were used in each case to generate an initial equilibrium configuration .
the number of sweeps skipped between each configuration used to calculate propagators and the total number of configurations in each ensemble are listed in the third and fourth columns , respectively , of table [ tab : lats ] .
although the number of sweeps skipped in each case was not sufficient to permit successive configurations to be treated as statistically independent , we found successive configurations to be sufficiently independent to justify the cost of evaluating glueball operators . for the propagators , effective masses and fitted masses to be discussed in sections [ sect : scalar ] and [ sect : tensor ] , we determined statistical uncertainties by the bootstrap method @xcite .
the bootstrap algorithm can be applied directly , however , only to determine the uncertainties in quantities obtained from an ensemble whose individual members are statistically independent .
we therefore partitioned each ensemble of correlated gauge configurations into successive disjoint bins with a fixed bin size .
bootstrap ensembles were then formed by randomly choosing a number of entire bins equal to the number of bins in the original partitioned ensemble . for bins sufficiently large , propagator averages found on distinct bins will be nearly independent .
it follows that for large enough bins , the binned bootstrap estimate of errors will be reliable .
it is not hard to show that once bins are made large enough to produce nearly independent bin averages , further increases in bin size will leave bootstrap error estimates nearly unchanged . the only variation in errors as the bin size
is increased further will come from statistical fluctuations in the error estimates themselves . to determine the required bin size for a particular error estimate to be reliable we applied the bootstrap method repeatedly with progressively larger bin sizes until the estimated error became nearly independent of bin size .
the final bin size we adopted for each lattice , chosen to be large enough for all of the error estimates done on that lattice , is given in fifth column of table [ tab : lats ] .
from the scalar operator of eq . [ defscalar ] , a propagator for scalars is defined to be @xmath27 to reduce statistical noise , @xmath28 is then averaged over reflections and time direction displacements of @xmath29 and @xmath30 . the collection of values of smearing iterations n , smearing parameter @xmath31 , and loop size @xmath32 for which propagators were evaluated for each lattice are given in table [ tab : nepsilons ] . at @xmath33 of 5.70 , and at @xmath33 of 5.93 on the lattice @xmath34 , we ran with relatively larger ranges of parameters to try to find values which coupled efficiently to the lightest scalar glueball . for other lattices ,
the parameter range was then narrowed to choices which , in physical units , were about the same as the value range which gave best results at @xmath33 of 5.93 . from the existence of a self - adjoint ,
positive , bounded transfer matrix for lattice
qcd , it follows that a spectral resolution can be constructed for @xmath35 , @xmath36 \ } , \nonumber \\ z_i & = & |
< i | s^{n \epsilon s } ( 0 ) | vacuum > |^2,\end{aligned}\ ] ] where the sum is over all zero - momentum , scalar states @xmath37 , @xmath38 is the energy of @xmath37 , and @xmath39 is the lattice period in the time direction . for large values of @xmath22 and @xmath39 ,
the sum in eq .
( [ scalarspec ] ) approaches the asymptotic form @xmath40 \ } \end{aligned}\ ] ] where @xmath41 is the smallest @xmath38 and thus the mass of the lightest scalar glueball and @xmath42 is the corresponding @xmath43 .
fitting @xmath44 to the asymptotic form in eq .
( [ scalarasym ] ) at @xmath22 and @xmath45 gives the scalar effective mass @xmath46 , which at large @xmath22 approaches @xmath41 . to extract values of @xmath41 from our data sets
, we began by examining effective mass graphs to find combinations of @xmath47 and @xmath32 for which @xmath48 shows a plateau at @xmath22 values for which we have data , and to determine which of these combination of @xmath47 and @xmath32 has the best plateaus . among the data sets used in our final extrapolation of the scalar mass to zero lattice spacing
, we included the largest range of values of @xmath47 and @xmath32 for the lattice @xmath34 with @xmath33 of 5.93 .
scalar effective masses obtained for this case with @xmath47 of @xmath49 and @xmath32 of @xmath50 are shown in figures [ fig : zx16b593n5s3 ] - [ fig : zx16b593n5s7 ] , respectively .
as the loop size @xmath32 is increased , initially the effective mass graphs become flatter , as shown , for example , by a decrease in the difference between @xmath51 and @xmath52 .
it follows that the relative coupling of the corresponding operators to the lightest scalar glueball increases with @xmath32 . beyond @xmath32 of 5 , however , this trend reverse .
thus , as might be expected , the relative coupling to the lightest state becomes weaker again when the loop is made too large . for @xmath32 of 7
the effective mass graph shows no sign of becoming flat even at the largest @xmath22 for which we have statistically significant data . for @xmath47 of 5 ,
the best coupling to the lightest state appears to occur with @xmath32 of 4 or 5 . for @xmath32
fixed at 4 , figure [ fig : zx16b593n5s4 ] and figures [ fig : zx16b593n6s4 ] - [ fig : zx16b593n8s4 ] show the variation in the effective mass graph as @xmath47 runs from 5 to 8 , respectively . the difference between @xmath51 and @xmath53 is least at @xmath47 of 6 and then grows again as @xmath47 is raised toward 8 . in figures [ fig : zx16b593n5s3 ] - [ fig : zx16b593n8s4 ]
the statistical uncertainty in effective masses grows as @xmath22 is made larger and tends to grow also if @xmath47 or @xmath32 is increased .
both of these phenomena are explained by the discussion in ref .
@xcite of the statistical uncertainty in propagators .
figures [ fig : zx16b570n6s2 ] - [ fig : zx32b640n8s10 ] show scalar effective masses for the each of the values of lattice size and @xmath33 listed in table [ tab : lats ] .
the parameters @xmath47 and @xmath32 for the data in figures [ fig : zx16b570n6s2 ] - [ fig : zx32b640n8s10 ] are chosen , for each lattice and @xmath33 , from among the set which couple best to the lightest scalar .
for each combination of lattice size and @xmath33 , we determined a final value of the scalar mass from the collection of propagators for which the effective mass graph showed at least some evidence of a plateau at large @xmath22 . for several different choices of @xmath22 interval ,
each of these propagators was fitted to the asymptotic form in eq .
( [ scalarasym ] ) by minimizing the fit s correlated @xmath54 .
the upper limit of each fitting interval @xmath55 we fixed at the largest @xmath22 for which we had statistically significant propagator data .
the lower limit of the fitting interval @xmath56 was then progressively increased from 1 to @xmath57 . as @xmath56 was increased , the fitted mass and the fit s @xmath54 per degree of freedom both generally decreased and the statistical error bar increased . for each @xmath47 and @xmath32 , the final choice of @xmath56 we took to be the smallest value for which the corresponding mass was within the error bars of all the fits with the same @xmath47 and @xmath32 and larger @xmath56 .
our intent in this procedure was to extract a mass from the largest time interval for which the propagator for each combination of @xmath47 and @xmath32 was consistent with the asymptotic form of eq .
( [ scalarasym ] ) .
the solid horizontal lines in figures [ fig : zx16b593n5s4 ] - [ fig : zx16b593n5s7 ] and figures [ fig : zx16b593n6s4 ] - [ fig : zx32b640n8s10 ] show the best fitted mass in each case and extend over the interval of @xmath22 on which these fits were made .
the dashed lines in these figures extend the solid lines to smaller @xmath22 to show the approach , with increasing @xmath22 , of effective masses to the final mass values . for the lattice @xmath34 with @xmath33 of 5.93 ,
tables [ tab:593zd1 ] - [ tab:593zd4 ] show the results of our fits for all the combination of @xmath47 , @xmath32 , @xmath56 and @xmath55 which we examined .
the best choice of @xmath56 and @xmath55 turned out to be 2 and 8 , respectively , for all @xmath47 and @xmath32 .
tables [ tab:570z ] - [ tab:640z ] show the fitted masses found with the best choice of @xmath56 and @xmath55 for all the lattice sizes , @xmath33 , @xmath47 and @xmath32 for which our effective mass data showed a plateau at large @xmath22 .
as expected , for each lattice size and @xmath33 , the fitted masses in tables [ tab:570z ] - [ tab:640z ] vary with @xmath47 and @xmath32 by an amount generally less than the statistical uncertainty in each mass . there is also a weak tendency for masses to fall initially with increasing @xmath47 and @xmath32 , as the corresponding operator s relative coupling to the lightest glueball increases .
then , in some cases , when @xmath47 and @xmath32 become too large the coupling to the lightest state decreases , the fitted masses show some tendency to rise again . to reduce this small remaining statistical uncertainty and systematic bias ,
our final value of mass for each lattice size and @xmath33 was obtained by an additional fit of a single common mass to a set of masses from a range of several @xmath47 and @xmath32 .
the common mass was chosen to minimize the correlated @xmath54 of the fit of the common mass to the collection of different best mass values .
the correlation matrix among the best mass values was determined by the bootstrap method .
the set of @xmath47 and @xmath32 used in each final fit was chosen by examining a decreasing sequence of sets , starting with all @xmath47 and @xmath32 , and progressively eliminating the smallest and largest @xmath47 and @xmath32 until a @xmath54 per degree of freedom below 2.0 was obtained .
the final fit was taken from the largest set of channels yielding a @xmath54 below 2.0 . if several sets of equal size gave @xmath54 per degree of freedom below 2.0 , we chose among these the set with smallest @xmath54 per degree of freedom .
tables [ tab:570zf]- [ tab:640zf ] show these combined fits and the set of @xmath47 and @xmath32 chosen for the final mass value for each lattice and @xmath33 . in all of these tables , it is clear that once enough of the largest and smallest @xmath47 and @xmath32 are eliminated to give an acceptable @xmath54 per degree of freedom , the fitted values vary only by small fractions of their statistical uncertainty as additional changes are made in the set of @xmath47 and @xmath32 . the final mass values are collected in table [ tab : finalz ] . at several points in tables [ tab:570zf]- [ tab:640zf ] ,
combined fits including several nearby values of @xmath47 and @xmath32 yield large @xmath54 while separate fits to smaller subsets of @xmath47 and @xmath32 give nearly equal masses and acceptable @xmath54 .
this phenomenon , we have found , does not indicate a problem with our data or our fits and arises instead because propagators with nearby values of @xmath47 and @xmath32 in some cases are very highly correlated and yield slightly different masses .
a similar problem would arise in trying to fit a single value @xmath58 to , say , a gaussian random variable @xmath59 with dispersion 1 , and a shifted copy @xmath60 .
for any choice of @xmath58 the fit s @xmath54 is infinite . nonetheless , for a monte carlo ensemble of 1000 @xmath59 values , taking @xmath58 as either @xmath61 or @xmath62 is a reliable estimate of the mean of @xmath59 with systematic error much smaller than the statistical error .
an alternative way to extract a single mass from glueball propagators for a range of @xmath47 , @xmath31 and @xmath32 uses the matrix of propagators @xmath63 for large @xmath22 and lattice time direction period @xmath39 , @xmath64 has the asymptotic form @xmath65 \}\end{aligned}\ ] ] where @xmath41 is the mass of the lightest scalar glueball and @xmath66 is a matrix independent of @xmath22 . in principle , @xmath64 can be extracted from our data and fitted to eq .
( [ masym ] ) to produce a value for @xmath41 . to find
the best @xmath41 and @xmath67 by minimizing the fit s @xmath54 , however , requires the statistical correlation matrix among the fitted @xmath68 .
if we fit , for example , to three choices of @xmath69 , three choices of @xmath24 and four values of @xmath22 , the correlation matrix has 1296 entries .
our underlying data set is too small to provide reliable entries for such a large correlation matrix . as a consequence
the value of @xmath41 determined this way will have a statistical error which can not be estimated reliably . in practice , we found that the value of @xmath41 produced by this method was not stabile as we varied the sets of @xmath70 and @xmath24 and the range of @xmath22 used in the fit .
a propagator for tensors is defined to be @xmath71.\end{aligned}\ ] ] where @xmath72 and @xmath73 are the tensor glueball operators of eq .
( [ deftensor ] ) , and @xmath74 is then averaged over reflections and time direction displacements of @xmath29 and @xmath30 to reduce statistical noise .
tensor propagators were found for gauge configuration ensembles and operator parameters listed in tables [ tab : lats ] and [ tab : nepsilons ] .
a tensor glueball mass was extracted from propagators by fitting the data to the tensor version of eq .
( [ scalarasym ] ) .
we obtained a satisfactory tensor glueball mass signal only for the lattices with @xmath33 of 5.93 , 6.17 and 6.40 .
we did not find an acceptable tensor signal at @xmath33 of 5.70 .
overall , the statistical errors in the tensor data are larger than those in the scalar data of section [ sect : scalar ] and , as a result , the fitting process encounters complications not present in the scalar fits
. tables [ tab:593st ] - [ tab:640 t ] list tensor masses for each gauge ensemble with @xmath33 of 5.93 and above , for each set of operator parameters in table [ tab : nepsilons ] , fitted on one or , in some cases , two choices of time interval . for all fits
the high end of the fitting range @xmath55 is chosen to be the largest value at which a statistically significant effective mass is found .
the low end of the fitting range @xmath56 is then progressively increased .
the smallest @xmath56 yielding a mass within one standard deviation of the masses for all larger @xmath56 is selected as the lower bound for an initial choice of the fitting range .
for the lattice @xmath34 at @xmath33 of 5.93 and for the lattice @xmath75 and @xmath33 of 6.40 , however , we found that for almost all choices of operator parameters a @xmath56 one unit larger than the initial choice yielded a noticeably lower mass .
these second values of @xmath56 and the corresponding masses are also listed in tables [ tab:593lt ] , and [ tab:640 t ] .
effective mass plots for tensors are shown in figures [ fig : tx12b593n6s6 ] - [ fig : tx32b640n8s10 ] , for the four lattices with @xmath33 of 5.93 and larger , for typical choices of operator parameters .
the solid line in each figure indicates the mass obtained from a fit over the time interval which the line spans .
the dashed line in each figures extend the solid line to smaller @xmath22 to show the approach of effective masses to the fitted value .
tables [ tab:593stf ] - [ tab:640tf ] list tensor masses found by combining , as discussed in section [ sect : scalar ] , the masses fitted to various sets of operators and choices of time interval .
table [ tab:593stf ] corresponds to the lattices @xmath76 at @xmath33 of 5.93 with fits using the single time interval given in table [ tab:593stf ] .
table [ tab:617tf ] corresponds to the lattice @xmath77 at @xmath33 of 6.17 with fits using the single time interval in table [ tab:617 t ] . in tables
[ tab:593stf ] and [ tab:617tf ] , all combined fits with acceptable @xmath54 per degree of freedom give masses consistent with each other to within statistical uncertainties .
in each case , the mass corresponding to the largest set with acceptable @xmath54 , marked with an arrow , is chosen as the final value .
table [ tab:593ltf ] for the lattice @xmath34 at @xmath33 of 5.93 shows combined fits using both choices of @xmath56 of table [ tab:593lt ] .
the combined fits using the smaller @xmath56 have unacceptably high @xmath54 per degree of freedom . for the fits using the larger @xmath56
the @xmath54 is acceptable , and the fitted masses are all consistent with each other within statistical uncertainties . the mass for the largest set of operators with the larger @xmath56 is chosen as the final number . table [ tab:640tf ] for the lattice @xmath75 at @xmath33 of 6.40 also gives combined fits for both @xmath56 in table [ tab:640 t ] .
most fits for both @xmath56 have acceptable @xmath54 per degree of freedom .
the masses obtained from the larger @xmath56 all lie one standard deviation or a bit more below the masses found with the smaller @xmath56 , however , and all have significantly better @xmath54 than the fits with the smaller @xmath56 .
the mass found from the largest set of operators for the larger @xmath56 is therefore chosen as the final result .
the collection of final tensor masses is listed in table [ tab : finalt ] .
we now consider an estimate of the difference between the scalar and tensor glueball masses in table [ tab : finalz ] and [ tab : finalt ] for finite lattice period @xmath39 and the infinite volume limits of these quantities
. for large values of @xmath39 , scalar @xmath78 and tensor @xmath79 glueball masses deviate from their infinite volume limits , @xmath80 and @xmath81 , respectively , by @xcite @xmath82 \}\end{aligned}\ ] ] where @xmath32 is 0 or 2 . in ref .
@xcite for @xmath33 near 6.0 , data for @xmath78 is shown to fit the two leading terms in eq .
( [ luscher ] ) reasonably well at 4 values of @xmath39 ranging from @xmath83 to @xmath84 .
this result is plausible since for @xmath39 ranging from @xmath83 to @xmath84 , the third term in eq .
( [ luscher ] ) is smaller than the second by a factor ranging from @xmath85 to @xmath86 . for our data at @xmath33 of 5.93 ,
table [ tab : finalz ] shows that @xmath80 is above 0.75 so that @xmath39 of 12 and 16 are larger than @xmath87 and @xmath84 , respectively .
thus we believe that for the data at @xmath33 of 5.93 , the leading two terms of eq .
( [ luscher ] ) likely provide a fairly reliable estimate of the @xmath39 dependence of @xmath78 and @xmath79 . fitting the @xmath88 data in table [ tab : finalz ] to the two leading terms of eq .
( [ luscher ] ) yields @xmath80 of @xmath89 and @xmath90 of @xmath91 .
in addition a bootstrap calculation yields with 95% probability @xmath92 . at @xmath88 , table [ tab : finalt ] combined with the leading two terms of eq .
( [ luscher ] ) gives @xmath81 of @xmath93 and @xmath94 is @xmath95 .
a bootstrap calculation yields with 95% probability @xmath96 .
overall , it appears to us safe to conclude that at @xmath33 of 5.93 the difference between scalar and tensor masses for @xmath39 of 16 and their infinite volume limits are of the order of 0.5% or less . in section [
sect : cont ] we show that the scalar and tensor glueball masses in tables [ tab : finalz ] and [ tab : finalt ] with @xmath33 of 5.93 and greater and @xmath97 fixed at about 13 are not far from asymptotic scaling .
we therefore expect the fractional volume dependent errors found in these masses to be about the same as the errors at @xmath33 of 5.93 .
thus the finite volume errors in all masses in tables [ tab : finalz ] and [ tab : finalt ] with @xmath33 of 5.93 and greater and @xmath97 of about 13 should be 0.5% or less .
the nonzero lattice spacing scalar and tensor glueball masses in lattice units given in tables [ tab : finalz ] and [ tab : finalt ] , respectively , we now convert to physical units and extrapolate to zero lattice spacing . to convert masses in lattice units to physical units , we divide by a known mass measured in lattice units .
one natural choice for this conversion factor is the rho mass @xmath98 .
values of @xmath98 for three of the four @xmath33 in tables [ tab : finalz ] and [ tab : finalt ] are given in ref .
@xcite . for the largest @xmath33 in tables [ tab : finalz ] and [ tab : finalt ] , ref .
@xcite does not report @xmath98 .
for the three @xmath33 considered in ref .
@xcite , however , the ratio @xmath99/[m_{\rho}(a ) a]$ ] is found to be independent of @xmath33 to within statistical errors . here
@xmath100 is obtained by the 2-loop callan - symanzik equation from @xmath101 found from its mean - field improved @xcite relation to @xmath33 . since @xmath99/[m_{\rho}(a ) a]$ ] is constant within errors , converting to physical units using @xmath100 then extrapolating to zero lattice spacing should give results nearly equivalent to those found using @xmath98 .
table [ tab : lambdas ] lists , for each @xmath33 , the corresponding mean - field improved @xmath101 and @xmath100 . the @xmath33 dependence of valence approximation glueball masses is determined entirely by the pure gauge part of the qcd action . the leading irrelevant operator in the pure gauge plaquette action has lattice spacing dependence of @xmath102 .
thus for scalar and tensor glueball masses @xmath80 and @xmath81 , respectively , we extrapolate to the continuum limit by @xmath103 ^ 2,\end{aligned}\ ] ] where @xmath32 is 0 or 2 . if @xmath100 in eq .
( [ contlim ] ) were replaced by @xmath98 , then since the leading irrelevant operator in the quark action has lattice spacing dependence of @xmath104 it might be argued that the quadratic @xmath102 term in the equation s right hand side should be a linear @xmath104 .
this in turn would contradict our claim that extrapolation using either @xmath98 or @xmath100 will give nearly equal results .
an answer to this objection is that the approximate constancy of @xmath99/[m_{\rho}(a ) a]$ ] implies that the @xmath104 irrelevant contribution to @xmath98 is quite small .
the constancy of @xmath99/[m_{\rho}(a ) a]$ ] as a function of @xmath105 or equivalently as a function of @xmath33 can not be explained by a cancellation of an @xmath104 term in @xmath100 with an @xmath104 term in @xmath98 since @xmath100 is defined to fulfill the true continuum two - loop callan - synamzik equation and itself has no @xmath104 corrections .
the leading correction to the @xmath33 dependence of @xmath100 is by a multiplicative factor of @xmath106 $ ] .
if @xmath100 is replaced by @xmath98 , any significant @xmath105 dependence which appears will come from the @xmath102 term in @xmath107 .
thus eq .
( [ contlim ] ) even with @xmath98 substituted for @xmath100 will remain correct . the scalar data of tables [ tab : finalz ] combined with @xmath100 of table [ tab : lambdas ] fitted to eq . ( [ contlim ] ) at the three largest @xmath33 is shown in figure [ fig : scalarquad ] .
the predicted continuum limit @xmath108 is @xmath109 .
the fit in figure [ fig : scalarquad ] has a @xmath54 of 0.6 over a range in which the term @xmath99 ^ 2 $ ] varies by more than a factor of 3.4 .
the variation of @xmath110/ [ \lambda^{(0)}_{\overline{ms } } a]$ ] over the fitting range , however , is only slight .
each of the three nonzero lattice spacing values of @xmath110/ [ \lambda^{(0)}_{\overline{ms } } a]$ ] is within 1.6 standard deviations of the extrapolated zero lattice spacing result .
thus we believe the extrapolation to zero lattice spacing is quite reliable and would expect the predicted continuum mass to be not very different from what would be obtained by any other reasonable , smooth extrapolation of the data .
the tensor data of tables [ tab : finalt ] combined with @xmath100 of table [ tab : lambdas ] fitted to eq .
( [ contlim ] ) at the three largest @xmath33 , the only @xmath33 for which tensor masses were found , is shown in figure [ fig : tensorquad ] .
the predicted continuum limit @xmath111 is @xmath112 .
the fit in figure [ fig : tensorquad ] has a @xmath54 of 0.8 , while , as before , the term @xmath99 ^ 2 $ ] in eq .
( [ contlim ] ) varies by more than a factor of 3.4 over the fitting range . to obtain scalar and tensor glueball masses in units of mev
, we combine the continuum limit @xmath113 of @xmath114 @xcite with @xmath115 of 770 mev to give @xmath116 of @xmath117 mev .
the scalar glueball mass becomes @xmath0 mev and the tensor mass becomes @xmath1 mev .
the continuum limit results are summarized in table [ tab : contlim ] . for @xmath113
we take the value given in ref .
@xcite for a lattice with period of about 2.4 fermi .
for the rho mass obtained at @xmath33 of 5.7 from a combination of propagators for rho operators with smearing parameters 0 , 1 and 2 , the 2.4 fermi result differs from the result for period 3.6 fermi by a bit over one standard deviation .
this difference appears to be largely a consequence of a slightly poorer separtion of the rho component of the propagator from excited state components in the 2.4 fermi rho mass calculation than in the 3.6 fermi calculation @xcite . for the rho operator with smearing parameter 4 , which couples more weakly to excited states
, the difference at @xmath33 of 5.7 between 2.4 fermi and 3.6 fermi predictions is much less than one standard deviation .
thus overall it appears to us reasonable to take the 2.4 fermi calculations as the infinite volume limit , within statistical errors .
the continuum limit values of @xmath113 for the the data combining smearings 0 , 1 and 2 and for the data from smearing 4 are nearly identical .
an independent calculation of the infinite volume , continuum limit of the valence approximation to several glueball masses is reported in ref .
@xcite . a second ,
more recent , calculation appears in ref .
@xcite . a comparison of ref .
@xcite with the original analysis @xcite of our results appears in ref .
@xcite .
the calculation of ref .
@xcite uses the same plaquette action we use but takes a different set of glueball operators .
the gauge field ensembles of ref .
@xcite range from 1000 to 3000 configurations . for the scalar and tensor masses ref .
@xcite reports @xmath118 mev and @xmath119 mev , respectively .
the predicted zero lattice spacing masses are not actually found by extrapolation to zero lattice spacing , but are obtained instead from calculations at @xmath33 of 6.40 of glueball masses in units of the square root of string tension , @xmath120 , then converted to mev using an assumed @xmath120 of 440 mev with zero uncertainty .
the uncertainties given in the masses are entirely the uncertainties in the @xmath33 of 6.40 calculations of masses in units of @xmath120 and are thus missing at least a contribution from the uncertainty in @xmath120 . a graph shown in ref .
@xcite suggests that the @xmath33 of 6.40 value of @xmath121/[\sqrt{\sigma ( a ) } a]$ ] is about 50 mev below the data s zero lattice spacing limit .
an additional error of @xmath122 mev in the scalar mass is therefore proposed in ref .
@xcite as a consequence of the absence of extrapolation to zero lattice spacing . since @xmath121/[\sqrt{\sigma ( a )
} a]$ ] of ref .
@xcite is clearly rising as lattice spacing falls , it does not appear to us that a symmetric error of @xmath122 mev an accurate representation of the effect of the absence of extrapolation .
if the statistical error and extrapolation error in the scalar mass are , nonetheless , taken at face value and combined the result is a prediction of @xmath123 mev .
no estimate is given for the extrapolation error in the tensor mass , which is found to be only weakly dependent on lattice spacing if measured in units of @xmath120 .
a scalar mass of @xmath123 mev is a bit over one standard deviations below the result @xmath0 mev in table [ tab : contlim ] , while the tensor mass of @xmath119 mev is in close agreement with our value of @xmath124 mev .
if the continuum limit of the ref .
@xcite data is found by extrapolation to zero lattice spacing of @xmath125/[\lambda^{(0)}_{\overline{ms } } a]$ ] , following section [ sect : cont ] , the result for @xmath126 is @xmath127 .
converted to mev using @xmath116 of @xmath128 mev , @xmath80 becomes @xmath129 mev .
this value is less than a standard deviation below the prediction @xmath0 mev in table [ tab : contlim ] .
the calculation of ref .
@xcite uses an improved action with time direction lattice spacing chosen smaller than the space direction .
the gauge field ensembles range in size from 4000 to 10000 configurations .
masses measured in units of the parameter @xmath130
@xcite are extrapolated to zero lattice spacing , then converted to mev using a value of @xmath130 found by extrapolation of @xmath131 to zero lattice spacing . as a result of working at relatively large values of lattice spacing ,
some ambiguity is encountered in matching the scalar mass s lattice spacing dependence to the small lattice spacing asymptotic behavior expected for the improved action . taking this uncertainty into account ,
the scalar mass is predicted to be @xmath132 mev .
the tensor mass , for which the extrapolation to zero lattice spacing encounters no problem , is predicted to be @xmath133 mev .
both numbers are a bit under one standard deviation above the predictions in table [ tab : contlim ] .
for the ratio @xmath134 ref .
@xcite predicts @xmath135 , in good agreement with the value @xmath136 in table [ tab : contlim ] .
thus the difference between table [ tab : contlim ] and ref .
@xcite is almost entirely a discrepancy in overall mass scale .
combining our extrapolation of @xmath127 for the data in ref .
@xcite with @xmath109 in table [ tab : contlim ] gives @xmath137 for @xmath126 , thus @xmath138 mev .
combining @xmath138 mev with @xmath132 mev of ref .
@xcite gives a world average valence approximation scalar mass of @xmath5 mev .
this number is consistent with the unmixed scalar mass of @xmath139 mev found in ref .
@xcite taking the observed states @xmath2 , @xmath140 and @xmath141 as the mixed versions of the scalar glueball and the two isoscalar spin zero quarkonium states , respectively .
the state @xmath2 in this calculation is assigned a glueball component of @xmath3 % . combining @xmath119 mev ,
@xmath1 mev and @xmath133 mev gives a world average valence approximation tensor mass of @xmath6 mev .
9 h. chen , j. sexton , a. vaccarino and d. weingarten , nucl .
b ( proc . suppl . ) * 34 * , 357 ( 1994 ) .
d. weingarten , nucl .
b ( proc . suppl . )
* 34 * , 29 ( 1994 ) .
j. sexton , a. vaccarino and d. weingarten , phys .
letts . 75 ( 1995 ) 4563 ; nucl .
b ( proc . suppl . ) * 47 * , 128 ( 1996 ) .
don weingarten , in continuous advances in qcd 1996,(edited by m. polikarpov ) , world scientific , singapore , 1996 .
w. lee and d. weingarten , nucl .
b ( proc.suppl . ) * 73 * , 249 ( 1999 ) .
d. h. weingarten , phys .
lett . 109b ( 1982 ) 57 ; nucl .
* b215 [ fs7 ] * , 1 ( 1983 ) . f. butler , h. chen , j. sexton , a. vaccarino and d. weingarten , phys .
letts . * 70 * , 2849 ( 1993 ) ; nucl .
phys . * b430 * ( 1994 ) 179 ; nucl .
* b421 * , 217 ( 1994 ) . k. kanaya , _ et al .
_ , nucl .
b ( proc .
* 63a - c * , 161 ( 1998 ) .
d. weingarten , nucl . phys .
b ( proc . suppl . )
* 53 * , 232 ( 1997 ) . g. bali , k. schilling , a. hulsebos , a. irving , c. michael , p. stephenson , phys . lett . * b 309 * , 378 ( 1993 ) . c. morningstar and m. peardon , phys . rev .
d * 56 * , 4043 ( 1997 ) ; hep - lat/9901004 , to appear in phys .
d. r. gupta _ et al .
_ , phys .
d * 43 * , 2301 ( 1993 ) .
m. falcioni _ et al .
_ , nucl .
* b251 * , 624 ( 1985 ) ; m. albanese _ et al .
_ , phys .
* b192 * , 163 ( 1985 ) . m. teper , phys .
* b 183 * , 345 ( 1987 ) .
f. r. brown and t. j. woch , phys .
lett . * 58 * , 2394 ( 1987 ) .
b. efron , the jackknife , the bootstrap and other resampling plans , society for industrial and applied mathematics , philadelphia , 1982 .
p. van baal and a. kronfeld , nucl .
b ( proc.suppl . ) * 9 * , 227 ( 1989 ) .
m. lscher , comm . math .
phys . * 104 * , 177 ( 1986 ) .
g. schierholz , nucl .
b ( proc . suppl . )
* 9 * , 244 ( 1989 ) .
g. p. lepage and p. mackenzie , phys .
d * 48 * , 2250 ( 1993 ) .
s. gottlieb , private communication .
r. sommer , nucl .
* b411 * , 839 ( 1994 ) . | we evaluate the infinite volume , continuum limit of glueball masses in the valence ( quenched ) approximation to lattice qcd . for the lightest scalar and tensor states we obtain masses of @xmath0 mev and @xmath1 mev , respectively . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.